Water Science and Technology Library

Ramakar Jha · Vijay P. Singh ·
Vivekanand Singh · L. B. Roy ·
Roshni Thendiyath   Editors

Water Resources
Management
and Reservoir
Operation
Hydraulics, Water Resources and Coastal
Engineering
Water Science and Technology Library

Volume 107

Editor-in-Chief
V. P. Singh, Department of Biological and Agricultural Engineering & Zachry
Department of Civil and Environmental Engineering, Texas A&M University,
College Station, TX, USA

Editorial Board
R. Berndtsson, Lund University, Lund, Sweden
L. N. Rodrigues, Brasília, Brazil
Arup Kumar Sarma, Department of Civil Engineering, Indian Institute of
Technology Guwahati, Guwahati, Assam, India
M. M. Sherif, Civil and Environmental Engineering Department, UAE University,
Al-Ain, United Arab Emirates
B. Sivakumar, School of Civil and Environmental Engineering, The University of
New South Wales, Sydney, NSW, Australia
Q. Zhang, Faculty of Geographical Science, Beijing Normal University, Beijing,
China
The aim of the Water Science and Technology Library is to provide a forum for
dissemination of the state-of-the-art of topics of current interest in the area of water
science and technology. This is accomplished through publication of reference
books and monographs, authored or edited. Occasionally also proceedings volumes
are accepted for publication in the series. Water Science and Technology Library
encompasses a wide range of topics dealing with science as well as socio-economic
aspects of water, environment, and ecology. Both the water quantity and quality
issues are relevant and are embraced by Water Science and Technology Library.
The emphasis may be on either the scientific content, or techniques of solution, or
both. There is increasing emphasis these days on processes and Water Science and
Technology Library is committed to promoting this emphasis by publishing books
emphasizing scientific discussions of physical, chemical, and/or biological aspects
of water resources. Likewise, current or emerging solution techniques receive high
priority. Interdisciplinary coverage is encouraged. Case studies contributing to our
knowledge of water science and technology are also embraced by the series.
Innovative ideas and novel techniques are of particular interest.
Comments or suggestions for future volumes are welcomed.
Vijay P. Singh, Department of Biological and Agricultural Engineering & Zachry
Department of Civil and Environment Engineering, Texas A &M University, USA
Email: vsingh@[Link]
All contributions to an edited volume should undergo standard peer review to
ensure high scientific quality, while monographs should also be reviewed by at least
two experts in the field.
Manuscripts that have undergone successful review should then be prepared
according to the Publisher’s guidelines manuscripts: [Link]
authors-editors/book-authors-editors/book-manuscript-guidelines

More information about this series at [Link]

Ramakar Jha · Vijay P. Singh · Vivekanand Singh ·
L. B. Roy · Roshni Thendiyath
Editors

Water Resources
Management and Reservoir
Operation
Hydraulics, Water Resources and Coastal
Engineering
Editors
Ramakar Jha Vijay P. Singh
Department of Civil Engineering Department of Biological and Agricultural
National Institute of Technology Patna Engineering
Patna, India Texas A&M University
College Station, TX, USA
Vivekanand Singh
Department of Civil Engineering L. B. Roy
National Institute of Technology Patna Department of Civil Engineering
Patna, India National Institute of Technology Patna
Patna, India
Roshni Thendiyath
Department of Civil Engineering
National Institute of Technology Patna
Patna, India

ISSN 0921-092X ISSN 1872-4663 (electronic)

Water Science and Technology Library
ISBN 978-3-030-79399-9 ISBN 978-3-030-79400-2 (eBook)
[Link]

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature
Switzerland AG 2021
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Contents

1 Integrated Water Resources Management of Thatipudi

Command Area, Vizianagaram, Andhra Pradesh . . . . . . . . . . . . . . . . 1
K. Kalyani and K. V. Jayakumar
2 Hydrological Modelling to Study the Impacts of Climate
and LULC Change at Basin Scale: A Review . . . . . . . . . . . . . . . . . . . . . 13
Dinu Maria Jose, Waleed Makhdumi,
and Gowdagere Siddaramaiah Dwarakish
3 Water Resource Management for Coal-Based Thermal Power
Plant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
S. A. Nihalani, Y. D. Mishra, and A. R. Meeruty
4 Evaluation of Reservoir Sedimentation Using Satellite
Data—A Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Beeram Satya Narayana Reddy and S. K. Pramada
5 Regionalisation of Watersheds Using Fuzzy C Means
Clustering Algorithm in the West Flowing River of Kerala . . . . . . . . 51
Thottungal Krishnankutty Drissia, Vinayakam Jothiprakash,
and Alayil Bahuleyan Anitha
6 Analysis of Relationship Between Landslides and Rainfall
in Karwar, Uttara Kannada District, Karnataka, India . . . . . . . . . . . 65
G. Thejashree, K. N. Lokesh, and G. S. Dwarakish
7 Optimal Cropping Pattern of Kulsi River Basin, Assam, India
Using Simulation and Linear Programming Model . . . . . . . . . . . . . . . 77
Jyotismita Taye, Bibhash Sarma, and Abhijit D. Lade
8 Comparison of Flux Footprint Models to a Mixed Fetch
Heterogeneous Cropland System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Shweta Kumari and K. B. V. N. Phanindra

v
vi Contents

9 Water Resources Assessment Issues and Application

of Isotope Hydrology in North East India . . . . . . . . . . . . . . . . . . . . . . . . 103
Prem Ranjan, Pankaj Kumar Pandey, Vanita Pandey,
and Pema Tshering Lepcha
10 Water Hammer Analysis for Pipe Line Network Using
HAMMER V8i . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Ajmal Hussain, Muhammad Mustafa, S. M. Ahbar Warsi,
and Sumit Kumar
11 Dam Break Flood Routing and Inundation Mapping Using
HEC-RAS and HEC-GeoRAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
A. Bharath, Anand V. Shivapur, and C. G. Hiremath
12 Suitability and Performance of Present Irrigation System
in Kokernag, Jammu and Kashmir . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Dar Himayoun, Roshni Thendiyath, and Jahangeer Saleem
13 Linking of Sediment Yield Pattern with Rainfall and Land-Use
Land-Cover Changes Within Burhanpur Sub-catchment,
India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
S. R. Resmi, P. L. Patel, and P. V. Timbadiya
14 Assessment of Probable Maximum Flood (PMF) Using
Hydrologic Model for Probable Maximum Precipitation
in Maithon Watershed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Bhanu Sharma and Kalyan Kumar Bhar
15 Simulating Failure of Indravati Dam Using Mike 11
and the Propagation of Breached Outflow . . . . . . . . . . . . . . . . . . . . . . . 177
Aditya Harikumar, Sachin Dhiman, and K. C. Patra
16 Optimization of Water Allocation for Ukai Reservoir Using
Elitist TLBO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
Vijendra Kumar and S. M. Yadav
17 Prediction of Reservoir Submerged Sediment Density . . . . . . . . . . . . 205
Y. C. Jabbar and S. M. Yadav
18 Micro-hydro Power Generation in India—A Review . . . . . . . . . . . . . . 219
Aparna M. Deulkar, Vivek S. Chavhan, and Pankaj R. Modak
19 Runoff Simulation and Irrigation Water Requirement
for Barman Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
A. Vishwakarma, M. K. Choudhary, and M. S. Chauhan
20 Nonlinear Regression Analysis Between Discharge and Head
for Piano Key Weirs with Increasing Developed Length (L/W)
Ratio and Constant Channel Width . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
Amiya Abhash and K. K. Pandey
Contents vii

21 Grey Water Characterization and Its Management . . . . . . . . . . . . . . . 251

Sarosh Alam Ghausi and Mohd Muzzammil
22 Intelligent Operation of Hirakud Reservoir Using
Metaheuristic Techniques (PSO and TLBO) . . . . . . . . . . . . . . . . . . . . . 263
Pooja Patnaik and Prakash Ch. Swain
23 Agricultural Water Management and Groundwater
Recharging Using Vadose Zone Modelling . . . . . . . . . . . . . . . . . . . . . . . 277
Anooja Thomas, Vivekanand Singh, and Brijesh Kumar Yadav
About the Editors

Ramakar Jha is Chair Professor in the Department of Civil Engineering and has
30 years of experience in the field of hydrology and water resources engineering.
He is presently working as Chair Professor in the Department of Civil Engineering,
National Institute of Technology (NIT) Patna, India, which is a premier institute in
India under the Ministry of Human Resource Development, Government of India.
He served at various levels from Scientist-B to Scientist-E1 at National Institute
of Hydrology (NIH), Roorkee, India, and as Professor in the Department of Civil
Engineering, NIT Rourkela. He has worked and working as Country Coordinator
of UNESCO-GWADI and Principal Investigator for many international (EU-FP7,
DAAD, ADB, AUS-Aid) and national research and consultancy projects (ISRO, DST,
MoWR, MHRD). Moreover, he served as Chair for many administrative positions and
received a couple of international and national awards for research papers. Presently,
he is working as Dr. Rajendra Prasad Chair for Water Resources under the Ministry
of Water Resources, Government of India, in the Department of Civil Engineering,
NIT Patna, Bihar.

Vijay P. Singh Texas A&M Professor of Indian origin is receiving a prestigious

award for his world-renowned work on water. He is receiving the 2013 Lifetime
Achievement Award from the American Society of Civil Engineers-Environmental
and Water Resources Institute, otherwise known as the ASCE-EWRI. The award is
in recognition of his work in the field of hydrology, which is the study of water in
all aspects, such as quality, distribution, preservation, and transportation. Some of
the work he has done has even created an entire new branch of hydrology—called
entropic hydrology—that is connected to the study of entropy, which means essen-
tially the study of order and disorder as it relates to the physical universe. His work
is considered fundamental for flood planning and water modeling around the world.
Since earning his doctorate degree, he has held teaching positions in some of the
most well-known universities in the USA. He was Associate Research Professor of
civil engineering at George Washington University from 1977 to 1978, Associate
Professor of civil engineering at Mississippi University from 1978 to 1981, and
Adjunct Professor as well as Coordinator of the Environmental and Water Resources
Systems Engineering Program at Louisiana State University from 1999 to 2006 and
ix
x About the Editors

2001 to 2006, respectively. He joined Texas A&M University in 2006, where he

currently wears a number of different hats. He is Professor of biological and agri-
cultural engineering, Professor of civil and environment engineering, and Caroline
and William N. Lehrer Distinguished Chair in Water Engineering (Hydrology). He
has authored or edited around 10 published works in the fields of engineering and
hydrology.

Vivekanand Singh is Professor in the Department of Civil Engineering and has 28

years of experience in the field of river hydraulics, groundwater, and water resources
engineering. He has published research papers in international journals including
ASCE Journal. He has done national projects and consultancy during his work at
NIT Patna. Prior to this, he was Scientist at National Institute of Hydrology, Roorkee.
He organized several summer courses, conferences, and workshops. He has been
working as Editor in some Indian journal.

L. B. Roy is Professor in the Department of Civil Engineering and has 33 years

of experience in the field of water resources engineering as well as geotechnical
engineering. He has been serving the Water Resources Engineering Department for
long time and carried out various field-based research activities for different river
systems of the region with special emphasis to floods. He has been involved in many
research projects and published several research papers in peer-reviewed journals.
He has been working for Intellectual Property Rights (IPR) and received awards for
various activities.

Roshni Thendiyath is Assistant Professor in the Department of Civil Engineering

and has 9 years of experience in the field of water resources engineering. She earned
her Ph.D. in civil engineering at the University of Pisa, Italy. After obtaining Ph.D.,
she had joined National Institute of Technology Calicut, India, as Adhoc Faculty
and later joined National Institute of Technology Patna as Assistant Professor. She
has supervised two Ph.D. and more than 30 master’s theses. Furthermore, she has
been awarded DST-SERB project for the topic Two-Phase Flows and Water Quality in
Rivers, funded by the Department of Science and Technology, MHRD, India. Results
obtained from her research have been published nearly 30 papers in international
journals and more than 15 papers in international conferences and three chapters.
She is active in a variety of professional bodies, and she has organized numerous
workshops and conferences in her academic career.
Chapter 1
Integrated Water Resources
Management of Thatipudi Command
Area, Vizianagaram, Andhra Pradesh

K. Kalyani and K. V. Jayakumar

Abstract IWRM provides ideas to help us make social choices about water alloca-
tion, the sustainability of water resources and the infrastructure use to manage the
resources establishment of an overall water policy and laws which use the basin as
the scale of management. It is widely recognised that better water management and
governance is the key to developing a sustainable, efficient and equitable water sector.
The present study area is Thatipudi reservoir and its command area. The command
area is characterised by highly fluctuating rainfall. This paper discusses on preparing
GIS maps, water requirement for command area using crop water requirements and
irrigation scheduling software, storage capacity of the reservoir and cropping pattern
to implement IWRM on command area. Highly fluctuating rainfall and corresponding
run-off water was analysed considering the irrigation requirements, conveyance effi-
ciency, domestic water requirements and water release into the sea during the floods.
The present study reveals that the current capacity of the reservoir is insufficient to
cater the allocation needs and thus necessitates the increase in storage capacity to
cater the water needs in dry season.

Keywords Integrated Water Resource Management (IWRM) · Irrigation

scheduling · Crop water requirements and reservoir capacity

1.1 Introduction

IWRM is a logical and appealing concept, its basis is that many different uses of water
resource are interdependent. Safe water supply and environmental sanitation are
essential for protecting the environment, improving health and reduction of poverty.
Despite increases in awareness of Integrated Water Resources Management (IWRM)

K. Kalyani (B)
Department of Civil Engineering, Andhra University, Visakhapatnam, AP 530003, India
K. V. Jayakumar
Civil Engineering, National Institute of Technology, Warangal, Telangana 506004, India
e-mail: kvj@[Link]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 1

R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
2 K. Kalyani and K. V. Jayakumar

and a number of studies that focus on the concept, still there exist a few in depth
scientific studies on the rationale behind the principles and concepts. Management is
used in its broadest sense. It emphasizes that one must not only focus on development
of water resources but that must manage water development in a way that ensures long
term sustainable use for future generations. IWRM is a process without formulas and
at the same time highlighted in guidelines produced by the Global Water Partnership.
IWRM attempts to put the Dublin Principles into practice, emphasizing the ideas of
integration, participation, and economic and financial sustainability and with the
basin as the unit for decision making.
Several large international conferences and agreements had particular influence
over its formation on Stockholm conference (1972) on the environment, Mar del Plata
Conference (1977) and the Dublin Conference (1992) held in preparation for the
United Nations Conference on Environment and Development (UNCED). Agarwal
et al. (2000) provides an extensive overview of established methods and approaches
that could be employed in water resources management. They comprise a wide range
of instruments used in assessing water resources, regulation, economic management,
conflict resolution, communication, and new technology.
IWRM provides ideas to help us consider how we can best make social choices
about water allocation and access as well as the sustainability of water resources and
the infrastructure use to manage those resources indicates by Giordano and Shah
(2014). Van Der Keuret al. (2010) identifies the characteristics and uncertainty as it
occurs in the different stages of the IWRM process with respect to sources, nature
and type of uncertainty and a common terminology that honour the most important
aspects from natural and social sciences and its application to the entire IWRM
process. Warwick and Haness (1994) observed that the conventional hydrological
data are inadequate for the purpose of design and operation of water resources system.
In such cases GIS are of great use for estimation of relevant hydrologic data. GIS
can be used for determination of catchment characteristics.
Integration across policy sectors is a core issue of IWRM. Grigg (2008) agree
that IWRM should coordinate the development and management of water, land and
other resources in order to improve economic and social welfare, social equity, and
environmental sustainability (GWP 2004). This recognizes the inherent interdepen-
dencies in nature and in the economic and social sectors. The GWP (2004) advocates
that the IWRM approach seeks to address a country’s key water-related development
problems, water for health, for food, for energy, for environment more effectively
and efficiently than is possible using traditional approaches.
Rees et al. (2008) argued that water governance reform need to actively recognize
and incorporate the financial dimension. Likewise, efforts to attract the additional
finance so badly needed by the sector have to be accompanied by governance reforms
to ensure that the funds are employed efficiently and that the financial flows can be
sustained.
1 Integrated Water Resources Management of Thatipudi Command … 3

1.2 Methodology

The study area details and the methodology adopted for the study area (Fig. 1.1)
were gathered and used in planning of water management programs.

1.2.1 Study Area

Thatapudi reservoir is constructed across the Gosthani River in Vizianagaram district

of Andhra Pradesh. The district is a part of the northern coastal plains of Andhra
Pradesh and it is situated within the geographical coordinates of 17°–15 and 19°–15
of the Northern latitude and 83°–0 to 83°–45 of the Eastern longitude. The reservoir
was built during the period 1963–68 to benefit an ayacut (culturable command area)
of 4868 hectares spread over three Mandals viz., Gantyada, Jami and S. Kota and
provide water for irrigation purpose and water supply to Visakhapatnam Municipal
Corporations at the rate of 22.70 MLD.

Fig. 1.1 Location and command area map of Thatipudi reservoir

4 K. Kalyani and K. V. Jayakumar

The Gosthani’s river water is diverted for agricultural and industrial purposes
and the river is the chief source of drinking water to the cities of Viziana-
garam and Visakhapatnam. Several infiltration wells have been sunk on the
Gosthani river bed to extract water especially during the summer months. Water
that released from canals is useful for villages and tribal population. It is also a
picturesque tourist location.

1.2.2 Data Collection

Data for the study were available from various sources like (i) Thatipudi reser-
voir (ii) Indian Meteorological Department and its Cyclone section (iii) Municipal
Corporation of Visakhapatnam, and (iv) Irrigation and Command Area Development
Department (I&CADD).
The required data (1997 onwards) were collected from various departments in
Vizianagaram and Visakhapatnam districts. This study has objectives like estimation
of storage capacity of the reservoir, water requirement for command area using crop
water requirements and irrigation scheduling software, various maps using GIS and
cropping pattern to implement IWRM on command area.

1.3 Analysis

The maps acquired from survey of India, are scanned and the final maps are presented
in ARCGIS as mandals and village’s maps, location and drainage maps. From these
maps water management phases are prepared from mandals and village’s maps crop-
ping pattern study proceeds to water requirement study than water demand for crop
seasonally can be identified. Now from location and drainage maps annual rain fall
data and storage capacity of reservoir are calculated. If water in the reservoir is excess
after meeting the demand management plan develops. If it is deficit after meeting
the demand then water developing and ground water extracting process should be
developed. The analysis is presented in flow chart as shown in Fig. 1.2.

1.3.1 Reservoir Storage Capacity

The current capacity of the reservoir is determined by Graphical method, using mass
curves. The capacity required for a reservoir depends upon the inflow available and
demand. Storage may not be required if the available inflow in the river is greater
than the demand. Larger capacity of the reservoir is required for higher demand if
the inflow into the reservoir is less.
1 Integrated Water Resources Management of Thatipudi Command … 5

Fig. 1.2 Flow chart of

management plan

   
S= It − Ot = (It − Ot ) (1.1)

where,
S = Net storage
I t = Volume of reservoir inflow
Ot = Volume of reservoir demand.

1.3.2 Irrigation Purpose

The village lands are irrigated by Thatipudi reservoir and a total of 155.54 ha (385
acres) are irrigated under it in the village. According to general public, the reservoir
6 K. Kalyani and K. V. Jayakumar

water is not sufficient for crops and also due to very poor maintenance of irrigation
canals. Major crops grown are paddy, vegetables and pulses.

1.3.3 Water Supply

From Thatipudi reservoir 22.70 MLD of water is supplied to Visakhapatnam. Munic-

ipal Corporation of Visakhapatnam apportions the water for industries and for
drinking water supply. Gosthani water supply scheme executed by the military in
1942 was taken over by the Municipality. In order to make further improvements
the authorities executed Gambhiram Gedda scheme to supply 15 MLD in 1957 to
supply water during one season. To meet the increasing water demand the authorities
executed a scheme in 1967 to supply 22.70 MLD from the Thatipudi reservoir.

1.3.4 Preparation of GIS Maps

The maps obtained from the Survey of India are scanned and then digitized in Auto
CAD and the digitized maps are further geo-referenced and the areas are calcu-
lated using ARCINFO and the final maps are presented in ARCGIS, which are the
command area map (Fig. 1.1) and drainage map (Fig. 1.3) respectively.

[Link] Delineation of Thematic Maps

The toposheets are used for determination of command area of all drains of Thatipudi
reservoir. In the present study command area is determined using GIS software. The
command area of Thatipudi reservoir is 9.61243 km2 and study area boundary is
107.398 km2 . The catchment area of Gosthani river areas is 3.47014 km2 and another
river found in these toposheets maps i.e. TakanunigeddaNala, a tributary of Gosthani
River has a catchment area of 1.19331 km2 . 37 villages are found under Thatipudi
command area.

1.4 Results

Annual graphical rainfall analysis was carried for Thatipudi command area with the
data from past year and presented.
The storage capacity is calculated by graphical method using mass curve and
graphically represented.
1 Integrated Water Resources Management of Thatipudi Command … 7

Fig. 1.3 Drainage network map of Thatipudi reservoir

1.4.1 Irrigation in Kharif Season

[Link] Water Requirement for Paddy Crop

Thatipudi Reservoir scheme which benefits an Ayacut of irrigated area spread over
three Mandals viz., Gantyada, Jami and S. Kota and major irrigation projects in
Vizianagaram District. Calculation of water requirements for Paddy irrigated area
8875 ha in the study area in Kharif season (Fig. 1.4; Table 1.1).

[Link] Water Requirement for Pulses Crop

Irrigated dry crop area is 460 ha in this command area in kharif season and most of
the crops are pulses and ground nuts. Water requirement is calculated for pulses over
an irrigated area of 300 ha (Table 1.2).
8 K. Kalyani and K. V. Jayakumar

140000

120000 Storage capacity

Mass d Inflow H-m
Mass Demand H-m

100000

80000

60000
Cum Inflow
Cum Demand
40000

20000

0
1996

1998

2000

2002

2004

2006

2008

2010

2012

2014
Year

Fig. 1.4 Mass curve of the Thatipudi reservoir

[Link] Water Requirement for Ground Nut Crop

Water requirement is calculated for ground nuts over an irrigated area of 160 ha in
kharif season in Thatipudi command area (Table 1.3).

1.4.2 Irrigation in Rabi Season

[Link] Water Requirement for Maize Crop

Calculation of water requirements in rabi season for irrigated dry area 4498 ha, in
that maize irrigated area is 2500 in this study area (Table 1.4).

[Link] Water Requirement for Vegetables

Calculation of water requirements in rabi season for irrigated dry area 4498 ha, in
that vegetables irrigated area is 1998 in this study area (Table 1.5)

1.5 Conclusions

Given the variation in the nature and complexity of the issues, the water and water-
related problems have to be approached at different levels. The study was undertaken
with a view to implement IWRM for Thatipudi reservoir command area and the
following conclusions are presented.
Table 1.1 Water requirement for paddy crop
Month Decadea Stageb Crop coefficient (K c ) E tc (mm/day) E tc (mm/decade) Effective rainfall (mm/dec) Irrigation required (mm/dec)
June 1 Initial 1.10 5.34 53.40 1.30 5.30
2 Initial 1.10 5.31 53.10 5.00 48.10
3 Development 1.10 5.27 52.70 18.70 34.00
July 1 Development 1.09 5.20 52.00 39.70 12.30
2 Development 1.08 5.10 51.00 53.40 0.00
3 Middle 1.07 5.03 50.30 45.50 9.90
Aug. 1 Middle 1.06 5.01 50.10 33.80 16.30
2 Middle 1.06 5.00 50.00 27.60 22.40
3 Middle 1.06 4.87 48.70 28.10 25.50
Sept. 1 Middle 1.06 4.74 47.40 30.50 16.90
2 Middle 1.06 4.61 46.10 30.90 15.20
3 Later 1.06 4.38 43.80 25.60 18.20
Oct. 1 Later 1.04 4.07 40.70 19.90 20.90
1 Integrated Water Resources Management of Thatipudi Command …

2 Later 1.01 3.75 37.50 15.30 22.20

3 Later 0.98 3.45 34.50 6.50 13.90
662.80 381.70 2810
Water requirement for paddy area = 24,938,750 m3
9
10 K. Kalyani and K. V. Jayakumar

Table 1.2 Water requirement for Pulsus

Month Decadea Stageb Crop E tc (mm/day) E tc (mm/dec) Effective Irrigation
coefficient rainfall Required
(K c ) (mm/dec) (mm/dec)
June 1 Initial 0.40 1.94 19.40 1.30 1.90
2 Initial 0.40 1.93 19.30 5.00 14.30
3 Development 0.40 1.93 19.30 18.70 0.60
July 1 Development 0.54 2.57 25.70 39.70 0.00
2 Development 0.75 3.56 35.60 53.40 0.00
3 Middle 0.97 4.60 46.00 45.50 5.00
Aug. 1 Middle 1.04 4.92 49.20 33.80 15.40
2 Middle 1.04 4.92 49.20 27.60 21.50
3 Middle 1.04 4.79 47.90 28.10 24.60
Sept. 1 Later 1.02 4.56 45.60 30.50 15.10
2 Later 0.75 3.25 32.50 30.90 1.60
3 Later 0.45 1.88 18.80 17.90 0.00
408.50 332.40 100.10
Water requirement for pulses area = 300,300 m3

Table 1.3 Water requirement for Ground Nut

Month Decadea Stageb Crop E tc (mm/day) E tc (mm/dec) Effective Irrigation
coefficient rainfall Required
(K c ) (mm/dec) (mm/dec)
June 1 Initial 0.40 1.94 19.40 1.30 1.90
2 Initial 0.40 1.93 19.30 5.00 14.30
3 Initial 0.40 1.92 19.20 18.70 0.50
July 1 Development 0.44 2.09 20.90 39.70 0.00
2 Development 0.61 2.89 28.90 53.40 0.00
3 Development 0.81 3.80 38.00 45.50 0.00
Aug. 1 Middle 0.99 4.69 46.90 33.80 13.10
2 Middle 1.05 4.93 49.30 27.60 21.60
3 Middle 1.05 4.80 48.00 28.10 24.70
Sept. 1 Middle 1.05 4.67 46.70 30.50 16.10
2 Middle 1.05 4.54 45.40 30.90 14.50
3 Later 0.97 4.00 40.00 25.60 14.40
Oct. 1 Later 0.75 2.95 29.50 19.90 9.70
2 Later 0.57 2.11 14.80 10.70 0.00
428.30 370.70 130.70
Water requirement for ground nuts area = 209,120 m3
1 Integrated Water Resources Management of Thatipudi Command … 11

Table 1.4 Water requirement for maize crop

Month Decadea Stageb Crop E tc (mm E tc (mm/dec) Effective Irrigation
coefficient /day) rainfall required
(K c ) (mm/dec) (mm/dec)
Dec 2 Initial 0.30 0.81 8.10 7.70 0.00
3 Initial 0.30 0.82 8.20 8.50 0.40
Jan 1 Development 0.36 0.99 9.90 0.20 9.70
2 Development 0.57 1.58 15.80 0.00 15.80
3 Development 0.79 2.30 23.00 0.00 25.20
Feb. 1 Middle 1.00 3.04 30.40 1.40 29.00
2 Middle 1.05 3.31 33.10 2.00 31.10
3 Middle 1.05 3.63 36.30 1.40 27.60
Mar. 1 Middle 1.05 3.94 39.40 0.10 39.40
2 Later 1.04 4.25 42.50 0.00 42.50
3 Later 0.88 3.74 37.40 0.50 40.70
Apr. 1 Later 0.64 2.81 28.10 14.00 14.10
2 Later 0.43 1.96 19.60 16.70 0.00
331.80 52.30 275.50
Water requirement for maize irrigated area = 6,887,500 m3

Table 1.5 Water requirement for vegetables

Month Decadea Stageb Crop E tc (mm/day) E tc (mm/dec) Effective Irrigation
coefficient rainfall Required
(K c ) (mm/dec) (mm/dec)
Dec. 2 Initial 0.70 1.88 18.80 7.70 4.90
3 Initial 0.70 1.90 19.00 8.50 12.40
Jan. 1 Development 0.73 1.99 19.90 0.20 19.80
2 Development 0.81 2.26 22.60 0.00 22.60
3 Development 0.91 2.63 26.30 0.00 28.90
Feb. 1 Middle 0.97 2.94 29.4 1.4 28
2 Middle 0.97 3.06 30.60 2.00 28.60
3 Middle 0.97 3.35 33.50 1.40 25.40
Mar. 1 Later 0.95 3.59 35.90 0.10 35.80
2 Later 0.89 3.60 36.00 0.00 32.40
272.00 21.20 238.80
Water requirement for vegetables irrigated area = 4,771,224 m3
12 K. Kalyani and K. V. Jayakumar

• The annual average rainfall in the study area is 1157.34 mm. The rainfall is
highly fluctuating in the study area and it is identified the need for Integrated
Water Resources Management. Thatipudi reservoir storage capacity was found to
be 97.93 [Link].
• It is observed from the above data that most the inflow into the reservoir is less than
the demand. So it is indicating that proper integrated water resources manage-
ment plan can be developed for extracting additional water for various other
developmental purposes.
• It can be seen that water requirement for irrigation throughout the year is
around 37.11 [Link] and considering conveyance efficiency also, it would be
53.01 [Link]. From the data available it can be seen that 51.46 [Link] of water
is released for irrigation purpose by I & CAD department.
• Water Supplied to Visakhapatnam city is found to be 18.00 [Link] against a
requirement of 97.93 [Link]
• By implementing various water resource techniques, water can be stored and
supplied to various users to fulfill the demand.
• By implementing IWRM in Thatipudi command area water resource management
can be developed and water utilisation will be developed in various factors in right
way.

References

Agarwal A, Delos Angeles MS, Bhatia R, Chéret I, Davila-Poblete S, Falkenmark M, Wright A

(2000) Integrated water resources management. Global Water Partner 64(8):109–117
Giordano M, Shah T (2014) From IWRM back to integrated water resources management. Int J
Water Resour Dev 36(6):1–13
Grigg NS (2008) Integrated water resources management: balancing views and improving practice.
J Water Int 33(3):279–292
Rees JA, Winpenny J, Hall AW (2008) Water financing and governance. Global Water Partnership,
No. 12
Van der Keur P, Brugnach M, Dewulf ARPJ, Refsgaard JC, Zorilla P, Poolman M, Mysiak J (2010)
Identifying uncertainty guidelines for supporting policy making in water management illustrated
for Upper Guadiana and Rhine Basins. J Water Resour Manag 24(14):3901–3938
Warwick JJ, Haness SJ (1994) Efficacy of ARC/INFO, GIS application to hydrological modeling.
J WRD and ASCE 120(3):366–379
Chapter 2
Hydrological Modelling to Study
the Impacts of Climate and LULC
Change at Basin Scale: A Review

Dinu Maria Jose, Waleed Makhdumi,

and Gowdagere Siddaramaiah Dwarakish

Abstract Two key factors which alter the hydrological system are human activity
and climate change. They affect the spatial and temporal distribution of water and
have unequal influence on hydrological system. Hence, the study of influence of
climate change and human activity is of immense importance for planning schemes
and strategies for water resources management. As a consequence of urbanization,
there has been a change in Land Use and Land Cover (LULC), which resulted in an
increase of surface runoff and degradation in water quality. Hydrological modelling
is an important tool which is used by researchers all over the world for analyzing the
influence of climate change and Land Use Land Cover Change (LULCC) on natural
water sources and for predicting potential future impacts from various possible
scenarios. Based on the future climate change scenarios, numerous basins may prob-
ably experience variation in the mean hydrological characteristics as well as in the
magnitude and frequency of extreme events. The main goal of this paper is to give a
brief review of studies done to evaluate the impacts of climate change and LULCC
on hydrological regime. There is a widespread usage of Landsat series of data for
hydrological modelling studies based on the period of analysis. And majority of the
hydrological modelling studies used semi-distributed models like SWAT for their
analysis as these aren’t much data intensive but gives reasonably accurate results.
Based on the results of various modelling studies, it can be concluded that, depending
on the characteristics of watershed and scenario assumptions, the combined effects
of climate change and LULCC may ameliorate or deteriorate each other’s influence.
These effects may vary with the season or land use classes involved in change. Hydro-
logical modelling integrating future climate change and LULCC scenarios can be an
effective tool in planning future water resource management strategies.

Keywords Climate change · Hydrological modelling · Urbanization · Land use

land cover change · Scenarios

D. M. Jose (B) · W. Makhdumi · G. S. Dwarakish

Department of Water Resources and Ocean Engineering, National Institute of Technology
Karnataka, Surathkal, Mangaluru, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 13

R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
14 D. M. Jose et al.

2.1 Introduction

Water is one among the crucial components of the environment and needs proper
management to accomplish its sustainable utilization in the future. Climate change
can affect the water availability on a large scale. Land use land cover changes like
urban growth can also alter the hydrological cycle affecting processes like evapo-
transpiration, infiltration, and surface and subsurface runoff (Dey and Mishra 2017).
These two influencing forces can affect the hydrological components at global, conti-
nental, regional and basin scales in all parts of the world (Praskievicz and Chang
2009). Hydrological modelling is an important tool which is used by researchers all
over the world for analyzing the influence of climate change and Land Use Land
Cover Change (LULCC) on natural water sources and for predicting potential future
impacts from various possible scenarios. Initially this paper gives a brief introduction
to various hydrological models. Later on, the studies on individual and combined
impacts of land use land cover and climate change are described. Finally, model
comparison studies and conclusions are presented.

2.2 Models

Hydrological models are widely used across the world for studying the effects of
various processes on water resources and thus for predicting potential future impacts
on the same (Praskievicz and Chang 2009). The models can be mainly classified
based on the physical and spatial structure (Krysanova et al. 2015). The models can
be grouped as distributed, semi-distributed and lumped based on the representation of
spatial heterogeneity at watershed scale (Dwarakish and Ganasri 2015). Distributed
and semi-distributed models capture the spatial variability better than the lumped
models (Bormann et al. 2009). Depending on the process description, models can
be categorized as conceptual, empirical and physical. Based on the consideration of
randomness, models are classified as stochastic and deterministic models. Former
involves randomness while latter doesn’t (Dwarakish and Ganasri 2015).

2.3 Hydrological Modelling in LULC and Climate Change

Impact Studies

Latitude, topography, geology, and land use affect the hydrological impacts of climate
in a basin. The rise in the surface runoff and undesirable high and low storm hydro-
graph are significant impacts of land use land cover change (Praskievicz and Chang
2009). The changes in climate and land use land cover (LULC) are interconnected and
are capable of modifying hydrological processes (Legesse et al. 2003; Li et al. 2009).
Nevertheless, the influence of climate and LULC change on hydrological regime are
2 Hydrological Modelling to Study the Impacts of Climate … 15

frequently researched separately. For example, there are researches done to find out
the influence of LULCC on evaporation losses (Cristina et al. 2015), infiltration rates
(Weatherhead and Howden 2009), and on runoff volume (Hundecha and Bárdossy
2004). And, there are studies which predict how climate change affects streamflow
(Mudbhatkal et al. 2017; Mudbhatkal and Amai 2017; Piao et al. 2010; Treesa et al.
2017), groundwater recharge (Soro et al. 2017), sediment load (Rodríguez-Blanco
et al. 2016) and water quality (Glavan et al. 2015).

2.3.1 Impacts of LULCC on Hydrology

There are many studies which evaluate the impact of LULC change on various
temporal and spatial scales (Fan and Shibata 2015; Zhou et al. 2013). As a result
of anthropogenic activities there has been subsequent increase in impervious area.
This leads to the alteration of the water balance of the catchment, with increase
in runoff, decrease in evapotranspiration and groundwater recharge (Leopold and
Dunne 1978). Factors like altitude, slope, distance from the river, type of agricultural
practices, type of soil and magnitude of erosion, frequency of drought and flood,
population density and distance from a built-up area affect the rate of LULC change
(Lin et al. 2009). The magnitude of impacts depend on the several factors like soil
depth (FAO 2008), precipitation events (FAO 2008), spatial layout of deforestation
areas (National Research Council 2008), area of watershed (Bi et al. 2014) etc.
Numerous researches are done to study the effects of LULC on hydrological regime
(Table 2.1).

2.3.2 Impact of Climate Change on Hydrology

Climate change has weakened the stationarity principle, a basic concept in water
resource engineering which says that future hydrological events will fluctuate within
the past variability (Milly et al. 2008). This is a result of human intervention in
the natural processes of earth leading to changes in the means and extremes of
evapotranspiration, precipitation and runoff. Studies are done to understand the trend
of extreme rainfall events using various statistical methods (Babar and Ramesh 2014).
Commonly used climate change scenarios were those which are in the Special
Report on Emission Scenarios (SRES) of Intergovernmental Panel on Climate
Change (IPCC) Fourth Assessment Report (AR4) (IPCC 2007) (Ficklin et al. 2009;
Tu 2009; Yoshimura et al. 2009; Praskievicz and Chang 2011). The Fifth Assessment
Report (AR5) of the IPCC brought in new scenarios called representative concentra-
tion pathways (RCPs). Table 2.2 shows the studies conducted to ascertain the effects
of climate change on various hydrologic parameters.
Table 2.1 Studies on impacts of land use land cover on hydrology
16

Author(s) Study area Study period Hydrological model DEM/Satellite Key conclusions Parameter
Babar and Ramesh Netravathi River 2001–2009 SWAT and RCRM ASTER/IRS 1D Sensitive parameters Surface runoff
(2015) Basin, India LISS-3, Landsat for the SWAT model
were ascertained.
RCRM which
requires few input
parameters can
predict stream flows
acceptably
Sinha and Eldho Netravathi river 1972, 1979, 1991, SWAT ASTER/Landsat An increase in Streamflow and
(2018) basin, India 2000, 2012 and 2030 sediment yield and sediment yield
streamflow was
observed due to the
increase in urban
and agricultural area
and a decrease in the
forest, leading to
changes in the
hydrological regime
(continued)
D. M. Jose et al.
Table 2.1 (continued)
Author(s) Study area Study period Hydrological model DEM/Satellite Key conclusions Parameter
Zhu and Li (2014) Little River 1984–2010 SWAT National Elevation A 3% increase in Streamflow,
Watershed, Tennesse Dataset DEM, streamflow was Sediment yield,
-Landsat observed in whole Total nitrogen, Total
watershed. 34.6% phosphorous
and 10% reduction
in sediment and
nutrient was
respectively
observed from 1984
to 2010, related to
the reduction in
agricultural land
Petchprayoon et al. Yom watershed, 1990–2006 MIKE 11 Landsat Due to urbanization, Peak river discharge
(2010) Thailand the discharge rate
was high at its
downstream than
that of other areas
Gyamfi et al. (2016) Olifants Basin, South 2000–2013 SWAT SRTM/ Landsat An increase of Surface runoff,
2 Hydrological Modelling to Study the Impacts of Climate …

Africa 46.97% in surface Lateral flow, Water

runoff was observed. yield, Groundwater
Urbanization was Lateral flow, ET,
identified as the Groundwater
major contributor to
increases in surface
runoff, water yield
and
evapotranspiration
17
Table 2.2 Studies on impacts of climate change on hydrology
18

Author(s) Study area Study period Hydrological GCM/RCM Key conclusions Scenario Parameter
model
Narsimlu et al. Upper Sind 1992–2098 SWAT PRECIS RCM A significant IPCC A1B Surface runoff
(2013) River Basin, increase in the Scenarios
India runoff and baseflow
is predicted. An
increase in
streamflow during
monsoon season and
a decrease in the
offseason is also
predicted
Ficklin et al. San Joaquin 1992–2005 SWAT – The 50 year IPCC A1F1 Evapotranspiration,
(2009) River watershed, averaged outcomes and B1 water yield,
US show that the basin streamflow
is highly sensitive to
future climate
changes
(continued)
D. M. Jose et al.
Table 2.2 (continued)
Author(s) Study area Study period Hydrological GCM/RCM Key conclusions Scenario Parameter
model
Treesa et al. Wainganga river 1971–2040 VIC CanESM2, Streamflow did not RCP 4.5 Streamflow
(2017) basin, India IPSL-CM5A-MR, show significant
MIROC-ESM, increase in the
ACCESS1-0, monsoon season but
GISS-E2-R and in the non-monsoon
GFDL- ESM2M season
Soro et al. (2017) Bandama Basin, 1986–2085 GR2M model HadGEM2-ES The results of the RCP 4.5 and Surface water and
West Africa model under simulated impacts RCP 8.5 groundwater
for RCP 4.5 and 8.5
were very different.
The study highlights
that there are huge
uncertainties
associated with
impacts studies done
with models
Rwigi et al. Sondu River 1961–2050 SWAT PRECIS The study predicts IPCC A2 Water yields
2 Hydrological Modelling to Study the Impacts of Climate …

(2016) basin, Kenya overall increase in

water yields with
wetter dry seasons
and drier wet
seasons in the future
19
20 D. M. Jose et al.

2.3.3 Combined Effects of Climate Change and LULC

Change

Hydrological models simulate the impacts of climate and human activities on the
hydrological regime when the hydro-meteorological data are fed to them. These can
be used for finding the individual impacts of the same on streamflow. This is done by
simulating the streamflow by assuming one of them constant and changing the other
(Dey and Mishra 2017). SWAT has been extensively used for studying the effects
of changing climate and land use land cover on streamflow (Dixon and Earls 2012).
SWAT is equally popular in water quality studies (Glavan et al. 2015). Models like
AVGWLF (Tu 2009), SWIM (Krysanova et al. 2015), PRMS (Legesse et al. 2003),
MIKESHE (Wang et al. 2013), GR2M (Ahn and Merwade 2014), GBHM (Ma et al.
2010) etc. for same purposes. Studies considering both climate change and land use
effects are shown in Table 2.3.

2.4 Model Comparison Studies

Tegegne et al. (2017) compared two conceptual models (GR4J, IHACRES) and a
semi-distributed model (SWAT) in four watersheds of Ethiopia. The results could not
conclude that any specific model was better than the other in all the watersheds. The
conceptual models performed better in smaller watersheds than in largest watershed
while SWAT performed better in largest watershed. Depending on the watershed,
the model performance varied. The study further combined the models using ANN
(Artificial Neural Network) and found reduction in RMSE (Root Mean Square Error)
values. Coupling semi-distributed models with Artificial Neural Network (ANN) can
improve the prediction of daily streamflow (Noori and Kalin 2016). An approach to
integrate different models can help in better understanding the response of watersheds
to LULCC and climate change (Kundu et al. 2017).
Krysanova and Hattermann (2017) did a comparison of impacts of climate change
on 12 basins distributed in important geographical zones. The hydrological models
used were ECOMAG, HBV, HYMOD, HYPE, mHM, SWAT, SWIM, VIC and
WaterGAP3. The climate scenarios were simulated using HadGEM2-ES, IPSL-
CM5A-LR, MIROC-ESM-CHEM, GFDL-ESM2M and NorESM1-M considering
RCP2.6, RCP4.5, RCP6.0 and RCP8.5. The study concludes that the uncertainty
associated with projections depend largely on selected GCM, followed by the selected
RCP and hydrological model selected is the least contributor.
Table 2.3 Studies on impacts of land use land cover and climate change on hydrology
Author(s) Study area Study period Model Results Parameter Emission scenario GCM/RCM
Chawla and Upper Ganga basin 1971 to 2100 VIC model Streamflow is Streamflow RCP 4.5, 8.5 ACC, CNR, CCS,
Mujumdar (2015) (UGB), India extremely sensitive to GFD, MPI and NOR
modifications in the
urban stretch while
moderately sensitive
to modifications in
croplands. Climate
change has a higher
impact on streamflow
than LULCC
Ahn and Merwade Indiana, New York, 1950–2010 GR2M The impact of human Streamflow – –
(2014) Arizona and Georgia influence on
area streamflow is greater
than that of climate
impact
(continued)
2 Hydrological Modelling to Study the Impacts of Climate …
21
Table 2.3 (continued)
22

Author(s) Study area Study period Model Results Parameter Emission scenario GCM/RCM
Kundu et al. Elbow River 1961–2070 MIKE SHE/ MIKE From the seasonal Streamflow A1B, B2 NCARPCM and
(2017) watershed in southern 11 assessment it was CGCM2
Alberta, Canada understood that the
LULC and climate
change scenario gave
higher streamflow in
spring. The study
emphasizes on the
fact that based on the
direction and
magnitude of impact
LULCC and climate
change can offset or
magnify each other’s
influence on the
hydrological regime
Kim et al. (2013) Hoeya River Basin, 2020–2050 SWAT The effect of LULC is Streamflow RCP 4.5, 8.5 HadGEM3-RA
Korea less than that of
climate change
Mango et al. Mara River Basin, 2002–2099 SWAT Large variations in Streamflow A1B Average of 21 GCMs
(2011) Kenya runoff was seen from in the MMD
small variations in (multi-model data set)
precipitation and land
use land cover
D. M. Jose et al.
2 Hydrological Modelling to Study the Impacts of Climate … 23

2.5 Discussion

This review deals with the studies that use hydrological modelling approach at basin
scale to evaluate the impacts of land use land cover and climate change on hydrology.
A brief note of model comparison studies is presented to detect the uncertainties asso-
ciated with modelling. Based on the review it was observed that the model perfor-
mance was dependent on the characteristics of the basin. Also, the impacts of land
use land cover and climate change may magnify or diminish each other’s effect on
hydrology based on the catchment characteristics. Studies predicting future impacts
of land use land cover and climate change on hydrology state significant changes
in the hydrological parameters in the future. These findings demonstrate possible
future water scarcity and related conflicts under the increasing urban development
and extreme climatic conditions.

2.6 Conclusions

There is a huge usage of hydrological models in the study of impacts of land use
land cover and climate change on water quality and quantity. Although there are
uncertainties associated with the modelling process, it can be useful in understanding
the complex interaction of various components and gives possible ranges of impacts
in the coming years. There is a necessity of more studies evaluating the effects of
climate change and land use land cover in many basins, as their interactive impacts
are likely to occur and are not well understood. There are several gaps in area of
hydrological modelling that could improve the accuracy of the outputs. Improvements
in the downscaling methods, choice of models, quality of input data can improve the
efficiency of modelling exercise. Quantification of uncertainties associated with the
modelling results could improve the reliability of modelled future projections.
In the land use land cover change studies, the sensors on the landsat series of
satellites are the popular ones because of the high temporal resolution and free
availability of data. And, SWAT is the most popular among the hydrological models
because it is less data intensive but gives reasonably accurate results. Based on
the results of various modelling studies, it can be concluded that, depending on
the characteristics of watershed and scenario assumptions, the combined effects of
climate change and LULCC may ameliorate or deteriorate each other’s influence.
These effects may vary with the season or land use classes involved in change.
Hydrological modelling integrating future climate change and LULCC scenarios
can be an effective tool in planning future water resource management strategies.
24 D. M. Jose et al.

References

Ahn K-H, Merwade V (2014) Quantifying the relative impact of climate and human activities on
streamflow. J Hydrol 515:257–266
Babar S, Ramesh H (2014) Analysis of extreme rainfall events over Nethravathi basin. ISH J Hydraul
Eng 20(2):212–221
Babar S, Ramesh H (2015) Streamflow response to land use—land cover change over the Nethravathi
River Basin, India. J Hydrol Eng 20(10):1–11
Bi C, Bi H, Sun G, Chang Y, Gao L (2014) Scale effects and variability of forest—water yield
relationships on the Loess Plateau, China. For Chron 90(2):184–191
Bormann H, Breuer L, Gräff T, Huisman JA, Croke B (2009) Assessing the impact of land use change
on hydrology by ensemble modelling (LUCHEM) IV: Model sensitivity to data aggregation and
spatial (re-) distribution. Adv Water Resour 32(2):171–192
Chawla I, Mujumdar PP (2015) Isolating the impacts of land use and climate change on streamflow.
Hydrol Earth Syst Sci 19(8):3633–3651
Cristina L, Dias P, Macedo MN, Heil M, Coe MT, Neill C (2015) Regional studies effects of land
cover change on evapotranspiration and streamflow of small catchments in the Upper Xingu River
Basin, Central Brazil. J Hydrol 4:108–122
Dey P, Mishra A (2017) Separating the impacts of climate change and human activities on
streamflow: a review of methodologies and critical assumptions 548:278–290
Dixon B, Earls J (2012) Effects of urbanization on streamflow using SWAT with real and simulated
meteorological data. Appl Geogr 35(1–2):174–190
Dwarakish GS, Ganasri BP (2015) Impact of land use change on hydrological systems: a review of
current modeling approaches. Cogent Geosci 1(1115691)
Fan M, Shibata H (2015) Simulation of watershed hydrology and stream water quality under land
use and climate change scenarios in Teshio River watershed, northern Japan. Ecol Indicators J
50:79–89
FAO (2008) Forests and Water: a thematic study prepared in the framework of the global forest
resources assessment 2005. FAO Forestry Paper 155
Ficklin DL, Luo Y, Luedeling E, Zhang M (2009) Climate change sensitivity assessment of a highly
agricultural watershed using SWAT. J Hydrol 374(1–2):16–29
Glavan M, Ceglar A, Pintar M (2015) Assessing the impacts of climate change on water quantity and
quality modelling in small Slovenian Mediterranean catchment - lesson for policy and decision
makers. Hydrol Process 29(14):3124–3144
Gyamfi C, Ndambuki JM, Salim RW (2016) Hydrological responses to land use/cover changes in
the Olifants Basin, South Africa. Water (switzerland) 8(12):1–16
Hundecha Y, Bárdossy A (2004) Modeling of the effect of land use changes on the runoff generation
of a river basin through parameter regionalization of a watershed model. J Hydrol 292:281–295
Kim J, Choi J, Choi C, Park S (2013) Impacts of changes in climate and land use/land cover
under IPCC RCP scenarios on stream flow in the Hoeya River Basin, Korea. Sci Total Environ
453:181–195
Krysanova V, Hattermann FF (2017) Intercomparison of climate change impacts in 12 large river
basins: overview of methods and summary of results. Clim Change 141(3):363–379
Krysanova V, Hattermann F, Huang S, Hesse C, Vetter T, Liersch S, Koch H, Kundzewicz ZW
(2015) Modelling climate and land-use change impacts with SWIM: lessons learnt from multiple
applications. Hydrol Sci J 60(4):606–635
Kundu S, Khare D, Mondal A (2017) An integrated modelling system to predict hydrological
processes under climate and land-use/cover change scenarios. Ecol Eng 105(10):1–23
Legesse D, Vallet-Coulomb C, Gasse F (2003) Hydrological response of a catchment to climate
and land use changes in Tropical Africa: case study south central Ethiopia. J Hydrol 275:67–85
Leopold LB, Dunne T (1978) Water in environmental planning. W.H. Freeman
Li Z, Liu W-Z, Zhang X-C, Zheng F-L (2009) Impacts of land use change and climate variability on
hydrology in an agricultural catchment on the Loess Plateau of China. J Hydrol 377(1–2):35–42
2 Hydrological Modelling to Study the Impacts of Climate … 25

Lin Y, Verburg PH, Chang C, Chen H, Chen M (2009) Landscape and urban planning developing and
comparing optimal and empirical land-use models for the development of an urbanized watershed
forest in Taiwan. Landsc Urban Plan J 92:242–254
Ma H, Yang D, Tan SK, Gao B, Hu Q (2010) Impact of climate variability and human activity on
streamflow decrease in the Miyun Reservoir catchment. J Hydrol 389(3–4):317–324
Mango LM, Melesse AM, McClain ME, Gann D, Setegn SG (2011) Land use and climate change
impacts on the hydrology of the upper Mara River Basin, Kenya: Results of a modeling study to
support better resource management. Hydrol Earth Syst Sci 15(7):2245–2258
Milly PCD, Betancourt J, Falkenmark M, Hirsch RM, Kundzewicz ZW, Lettenmaier DP, Stouffer RJ
(2008) Climate change: stationarity is dead: whither water management? Science 319:573–574
Mudbhatkal A, Amai M (2017) Regional climate trends and topographic influence over the Western
Ghat catchments of India. Int J Climatol 38(5):2265–2279
Mudbhatkal A, Raikar RV, Venkatesh B, Mahesha A (2017) Impacts of climate change on varied
river-flow regimes of Southern India. J Hydrol Eng 22(9):1–13
Narsimlu B, Gosain AK, Chahar BR (2013) Assessment of future climate change impacts on
water resources of Upper Sind River Basin, India Using SWAT Model. Water Resour Manage
27(10):3647–3662
National Research Council (2008) Hydrologic effects of a changing forest landscape
Noori N, Kalin L (2016) Coupling SWAT and ANN models for enhanced daily streamflow
prediction. J Hydrol 533:141–151
Petchprayoon P, Blanken PD, Ekkawatpanit C, Hussein K (2010) Hydrological impacts of land
use/land cover change in a large river basin in central-northern Thailand. Int J Climatol
30(13):1917–1930
Piao S, Ciais P, Huang Y, Shen Z, Peng S, Li J, Zhou L, Liu H, Ma Y, Ding Y, Friedlingstein P, Liu
C, Tan K, Yu Y, Zhang T, Fang J (2010) The impacts of climate change on water resources and
agriculture in China. Nature 467(7311):43–51
Praskievicz S, Chang H (2009) A review of hydrological modelling of basin-scale climate change
and urban development impacts. Prog Phys Geogr 33(5):650–671
Praskievicz S, Chang H (2011) Impacts of climate change and urban development on water resources
in the Tualatin River Basin, Oregon. Ann Assoc Am Geogr 101(2):249–271
Rodríguez-Blanco M, Arias R, Taboada-Castro M, Nunes J, Keizer J, Taboada-Castro M (2016)
Potential impact of climate change on suspended sediment yield in NW Spain: a case study on
the Corbeira catchment. Water 8(10):444
Rwigi SK, Muthama JN, Opere AO (2016) Simulated impacts of climate change on surface water
yields over the Sondu Basin in Kenya. Int J Innov Educ Res 4(8):160–172
Sinha RK, Eldho TI (2018) Effects of historical and projected land use/cover change on runoff and
sediment yield in the Netravati river basin, Western Ghats, India. Environ Earth Sci 77(111):1–19
Soro G, Yao A, Kouame Y, Bi T (2017) Climate change and its impacts on water resources in the
Bandama Basin, Côte D’ivoire. Hydrology 4(1):18
Tegegne G, Park DK, Kim YO (2017) Comparison of hydrological models for the assessment of
water resources in a data-scarce region, the Upper Blue Nile River Basin. J Hydrol: Reg Stud
14:49–66
Treesa A, Das J, Umamahesh NV (2017) Assessment of impact of climate change on streamflows
using VIC model. Euro Water 59(2013):61–68
Tu J (2009) Combined impact of climate and land use changes on streamflow and water quality in
eastern Massachusetts, USA. J Hydrol 379(3–4):268–283
Wang S, Zhang ZR, McVicar T, Guo J, Tang Y, Yao A (2013) Isolating the impacts of climate
change and land use change on decadal streamflow variation: assessing three complementary
approaches. J Hydrol 507:63–74
Weatherhead EK, Howden NJK (2009) The relationship between land use and surface water
resources in the UK. Land Use Policy 26(SUPPL. 1):S243–S250
26 D. M. Jose et al.

Yoshimura C, Zhou M, Kiem AS, Fukami K, Prasantha HHA, Ishidaira H, Takeuchi K (2009) 2020s
scenario analysis of nutrient load in the Mekong River Basin using a distributed hydrological
model. Sci Total Environ 407(20):5356–5366
Zhou F, Xu Y, Chen Y, Xu C, Gao Y, Du J (2013) Hydrological response to urbanization at different
spatio-temporal scales simulated by coupling of CLUE-S and the SWAT model in the Yangtze
River Delta region. J Hydrol 485:113–125
Zhu C, Li Y (2014) Long-term hydrological impacts of land use / land cover change from 1984 to
2010 in the Little River Watershed, Tennessee. Int Soil Water Conserv Res 2(2):11–21
Chapter 3
Water Resource Management
for Coal-Based Thermal Power Plant

S. A. Nihalani, Y. D. Mishra, and A. R. Meeruty

Abstract Water and energy are elementary requirements for well-being and pros-
perity of humans. They are mutually dependent on each other. Energy production
involves a huge quantity of water and in turn, water structures need a huge amount of
energy. This dependency is termed as ‘water-energy nexus’. As a result of population
and economic growth, and higher living standard, future demand for both water and
energy is anticipated to upsurge, particularly in the emerging economies like India.
The majority of energy-linked water usage is for power generation, which is led by
water-concentrated electricity generation from coal, natural gas, and nuclear mate-
rials. Majority of the water utilised at thermal power plants is for cooling purpose in
addition to ash handling, boiler feed water, flue gas desulphurisation and for other
applications in coal-fired power plants. Governing authorities shall impose limits on
freshwater usage by power plants by framing regulatory norms. The present stress on
freshwater supplies and increasing demand is compelling the power generating units
to search for alternative sources or supplementary sources of water. Utilising water
from such sources will help conserve fresh water for other uses, such as drinking and
agriculture. The current paper discusses water resource management for coal-based
thermal power plants. If further evaluates various alternatives or measures for water
conservation in thermal power plants.

Keywords Air cooled condenser · Evaporation loss · Makeup · DM plant · Zero

liquid discharge

S. A. Nihalani (B) · A. R. Meeruty

Civil Engineering Department, PIET, Parul University, Vadodara, Gujarat, India
e-mail: [Link]@[Link]
Y. D. Mishra
Mechanical Department, L&T-S&L, Vadodara, Gujarat, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 27

R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
28 S. A. Nihalani et al.

3.1 Introduction

Water is one of the unavoidable input needed for sustainable growth of the industrial
sector of any country. As the need for power increases, the demand for freshwater
supply for power generation will also grow. During past years, there is a decrease
in the availability of freshwater. Also, there is a rising realization in power plant
managers to preserve fresh water and decrease the use of freshwater by employing
various water conservation techniques including treatment, recycle and reuse. In addi-
tion, the need to leave water in receiving stream to meet environmental regulations
will further govern the future provision of freshwater supply.
Ministry of Environment and Forest has taken stringent steps to control environ-
ment and water pollution. MOEF has recently issued notification for consumption
of water usage and emission standards within thermal power plants. Based on the
notification issued by MOEF, all the existing thermal power plants have to reduce
specific water consumption up to 3.5 m3 /MWh. All the new plants that shall be
installed after 1st January 2017 are required to meet specific water consumption up
to a maximum of 2.5 m3 /MWh and must achieve zero liquid discharge (ZLD). This
new environment norm generates the need for every power generation company to
critically evaluate the water resource management of coal-based power plant. The
first step in water resource management is to evaluate the water requirement of
various consumers in power plant and understand the quality of water required at
various sources. This paper will evaluate the possible alternatives to reduce the fresh
makeup water requirement and show ways to achieve zero liquid discharge. This
paper will also discuss the adoption of air-cooled condenser technology in place of
the conventional wet cooling system to meet the new environment norms.

3.2 Water Resource Management in Thermal Power Plants

Water requirement in power generation is regulated by a number of factors like raw

water quality, condenser cooling system, coal quality, utilization of ash, disposal of
ash, wastewater handling etc. Based on water quality and end-user concerns, the
water utility in power plants can be divided into following types:
1. Ash Handling Water: water utilized to convert dry ash to slurry form for disposal.
2. Cooling Tower water: Water utilized for cooling towers.
3. Makeup Water: Water utilized for compensating evaporation losses in the
cooling tower.
4. Demineralised Water: Water utilized for generating steam in boilers.
5. Drinking Water: Water used for human consumption purpose.
6. Utility Water: Water used for various services like fire-fighting, dust suppression,
gardening, and sanitation purposes etc.
3 Water Resource Management for Coal-Based Thermal Power Plant 29

3.2.1 Cooling Water System Water Requirements

Cooling water system mainly uses water to condense the steam coming out from LP
turbine exhaust. The heat of condensed steam will again be rejected to the atmosphere
with the help of a wet cooling tower. In a closed cooling water system, cooling tower
rejects the heat of hot water coming out from the condenser by evaporating some
amount of water in the atmosphere. The process of evaporation inside cooling tower
increases the total dissolved solid levels of circulating water which in turn increases
scaling and corrosion problems in the system. To maintain dissolved solids levels
in the closed cooling water system, some amount of water needs to be drained out
from the system by making up fresh water to the system. The limit of dissolved
solids levels in the cooling water system is dictated by a cycle of concentration. The
cycle of concentration (COC) is how many times dissolved solids in water is getting
concentrated without increasing scaling/corrosion issues and the loss of water to
maintain COC of the system is termed as CW blow-down. In a closed cooling water
system, water is required to be makeup to compensate for evaporation loss, blow-
down loss and drift losses. Increasing COC of the system reduces the blow-down
loss from the system however; it will impact circulating cooling water quality and
affect the performance of the circulating cooling water system. Generally, river water
based power plants are designed to maintain COC as 5 and coastal power plants using
closed cooling water system are designed to maintain COC in range of 1.2 to 1.3.

3.2.2 Ash Handling System Water Requirements

Ash is generated due to the burning of coal inside the boiler, which needs to be
disposed of off to ash dyke. There are two types of ash generated from the boiler.
Ash collected at bottom of boiler and economizer is termed as Bottom ash and ash
collected at bottom of ESP is termed as Fly ash. Generally, bottom ash and fly ash are
generated in the ratio of 20% and 80% of the total ash generation during the burning
of coal. In ash handling system, water is required mainly to prepare ash slurry in
specific concentration for evacuating bottom ash and fly ash to ash dyke with the
help of slurry pumps. Water requirements in ash handling plant mainly depend on
the type of ash handling system and quantity of ash available in the coal. For bottom
ash disposal, mostly the practice is to use lean slurry disposal system in which 20–
22% concentrated slurry is made with the help of water and transferred to ash dyke
with the help of bottom ash slurry disposal system. However, a new trend of the dry
bottom ash handling system is finding a place in some of the recent tenders. For fly
ash disposal, industry proven practice is to use either of the following systems:
• The lean slurry disposal system
• High concentrated slurry disposal system (HCSD)
• Dry Fly ash disposal system.
30 S. A. Nihalani et al.

High concentrated slurry disposal system (HCSD) uses disposal of 50–55%

concentrated fly ash slurry through HCSD pumps. Hence, water requirements of
HCSD system are less as compared to the lean slurry fly ash disposal system. Further,
due to MOEF norms for 100% utilization of fly ash from the 4th year of plant oper-
ation, all the power plants since this notification have been designed to dispose fly
ash in wet mode (through Lean slurry or HCSD system) during initial period of plant
operation and then dispose of via dry fly ash disposal system. Power plants have
also been equipped with ash water recirculation system to recover clear water from
ash dyke area. With the installation of ash water recirculation system, approximately
70% water can be recovered and utilized in ash handling system for slurry preparation
after 1 year of plant operation.
With the help of water recovery from ash dyke after one year and 100% dry fly ash
utilization after 4 years of plant operation, water requirements in the ash handling
plant are reduced to a considerable extent. However, during initial periods of plant
operation, a considerable amount of water requirement is envisaged for wet bottom
ash and fly ash disposal system with no recovery from ash dyke. For a typical 2 ×
660 MW plant capacities with Indian coal quality that consists of high ash content, the
total makeup water requirement for ash handling plant is estimated in the following
range.
Table 3.1 shows that ash handling plant consumes 1.3–1.5 m3 of water for produc-
tion of 1 MWh electricity. For the plant using lean slurry disposal system for bottom
ash and fly ash disposal, above water requirement is reduced to 500–600 m3 /h after the
first year of plant operation when recovery of water from ash dyke area is available.
However, for a plant using the HCSD system for fly ash disposal, water requirement
remains in the range of 700 m3 /h because no water could be recovered from fly ash
HCSD system. However, during the fly ash utilization period, the water requirement
will be in the range of 220–250 m3 /h approximately.
Total makeup water requirement of ash handling plant during the initial period
of plant operation is generally being met by cooling tower blowdown water and
remaining by raw water. During the ash water recovery scenario, water requirement
can be met by cooling tower blowdown water and no additional raw water is needed.
However, during less water requirement in case of fly ash utilization after the 4th year
of plant operation, cooling tower blowdown water could not be fully utilized in ash
handling plant for slurry making and this creates a situation of excess availability of
cooling tower blowdown water. The scenario of existing power plants using dry fly ash
utilization and recovery water from ash dyke creates concern for power Generation
Company on how zero liquid discharge could be achieved.

Table 3.1 Water requirement for ash handling

Description Bottom ash and Fly ash disposal Bottom ash disposal in lean
the system in a lean slurry mode slurry mode and Fly ash disposal
with No ash water recovery in HCSD mode with No ash
water recovery
Makeup water requirement 1800–2000 m3 /h 1300–1400 m3 /h
3 Water Resource Management for Coal-Based Thermal Power Plant 31

3.2.3 Coal Handling System Water Requirements

In a coal-based thermal power plant, coal is used as working fuel in the boiler for
generating steam from water. The required amount of coal needs to be met from
nearby coal mines through conveyer/railway wagon/trucks. In the process of making
and transporting suitable size of coal in the boiler from Coal Storage area, water
is required to spray during unloading and transportation to avoid the emission of
coal particles in the atmosphere and to spray water at coal storage area to avoid
spontaneous ignition of coal. Water Requirements for coal handling plant varies
from plant to plant based on the location of coal storage, numbers of coal Stockpiles,
numbers of transfer points. For typical 2 × 660 MW unit, water requirement for coal
handling plant is estimated in the range of 100–150 m3 /h.

3.2.4 Demineralized Water System Water Requirements

Demineralized water is mainly used as power cycle makeup in condenser hot well
to compensate blowdown and other steam vent losses that occur in case of subcrit-
ical power plants. In supercritical power plants, cycle makeup is only required to
compensate for the steam vent losses and other losses due to leakages. Power cycle
makeup for supercritical power plants is generally considered as 0.5–1% of BMCR
steam flow. Further consumers of demineralized water in power plant are condensate
polishing units, demineralized water generation plant, closed auxiliary cooling water
system. Condensate polishing plant is installed in all supercritical power plants to
maintain the stringent quality of demineralized water in the power cycle as well as to
reduce power cycle makeup water requirement. Condensate polishing plant is placed
at the condensate extraction pump discharge to treat the condensate coming from
the condenser. In the condensate polishing plant, demineralized water is required to
regenerate the resins of condensate polishing units. DM water is also used as makeup
for the closed auxiliary cooling water system to compensate for water losses due to
leakages. For Typical 2 × 660 MW supercritical power plant, total demineralized
water requirement is estimated in the range of 50–60 m3 /h.

3.3 Water Balance for 2 × 660 MW Coal-Based Power

Plant

The first step in the development of optimized water resource management for the
coal-based power plant is to review the present scenarios of water balance in the
plant. For a review of present water resource management, the water balance of 2
× 660 MW coal-based supercritical power plant has been prepared considering all
32 S. A. Nihalani et al.

modern practices followed by recent power plants. Following design considerations

are adopted to develop the typical water balance of 2 × 660 MW plant:
• River water is considered a source of fresh raw water.
• Water cooled condenser is considered for circulating cooling water system and
the cycle of concentration for the system is considered as 5.
• Evaporation loss and drift loss inside the wet cooling tower is considered as 1.6%
and 0.005% of cooling tower flow respectively.
• Cooling tower blowdown is waste having high dissolved solids. However, this
waste will be utilized directly without treatment for ash slurry preparation in ash
handling plant.
• Total demineralized water requirement is considered as 1.5% of BMCR flow
considering regeneration water requirements and Conventional Ion exchange resin
based demineralization (DM) plant is considered to generate the DM quality water.
• Demineralization and condensate polishing plant waste is generated during regen-
eration of ion exchange resin. This waste is having high dissolved solids and will
be utilized in ash handling and coal handling plant after neutralization of waste.
• Pre-treatment plant sludge generated at bottom of clarifiers will further be treated
using thickener and dewatering centrifuge for generation of cake in semi-dry form.
• Floor wash along with oily waste generated will be treated in the clarifier and oily
water separator to recover clear water. Clearwater will then be utilized to meet
ash handling and coal handling plant water requirements.
• Lean slurry disposal system is considered for bottom ash disposal and high concen-
trated slurry disposal system (HCSD) is considered for fly ash disposal during
initial operation of the plant.
• Approximately 70% of water from ash dyke is considered to be recovered after
1 year of plant operation and 100% fly ash will be utilized after 4th years of plant
operation.
• Water requirement for future flue gas desulfurization (FGD) units is not considered
while reviewing the present scenario of water balance.
Based on the above design consideration, net raw water makeup requirement for
typical 2 × 660 MW coal-based supercritical plant has been estimated as 4230 m3 /h
which corresponds to specific water consumption of around 3.2 m3 /MWh. This water
requirement is established considering no recovery water received from an ash pond
during initial years of plant operation (Table 3.2).
The specific water consumption will further be reduced during ash water recovery
scenario as well as 100% dry fly ash utilization period. However, during 100% dry
fly ash utilization period there will be excess cooling tower blowdown in the range
of 220–250 m3 /h within the plant and zero liquid discharge could not be achieved.
Further, these values are estimated considering river water based plant and COC of 5.
For a plant which receives bore well water and seawater as raw water above specific
water consumption values increases significantly. The specific water consumption
values for current scenarios of river water and seawater-based plant using closed
cooling water system are shown in Table 3.3.
3 Water Resource Management for Coal-Based Thermal Power Plant 33

Table 3.2 Consumptive water requirement of 2 × 660 MW power plant

S. No. Systems Water consumption (m3 /h) Specific water consumption
(m3 /MWh)
1 Makeup water requirement 3000 2.27
in the condenser cooling
water system
2 Total makeup water 1400 1.06
requirement for ash
handling plant
3 Total makeup water 150 0.11
requirement for coal
handling plant
4 DM water requirements 60 0.05
5 Miscellaneous service 250 0.19
water & potable water
requirements including
HVAC, pump sealing
Total water requirements 4860 3.68
6 CW blowdown water 600
generated
7 Other waste generated 30
including DM plant waste,
CPU waste, floor washing
waste, and oily waste
Total waste collected 630
9 Reduction in ash handling 480
plant makeup water
requirement due to the
utilization of plant waste
10 Reduction in coal handling 150
plant makeup water
requirement due to the
utilization of plant waste
11 Net raw water makeup 4230 3.2
requirement

3.4 Analysis of New Environment Norms and Its Impacts

Study of new environment norms issued by MOEF reveal following facts with respect
to coal-based power plants:
• Once through cooling system is now discarded for all existing power plant as well
as for upcoming new power plants. Environment rules mandate to install a closed
cooling water system for all existing once-through power plant. Further, rules also
mandate to achieve specific water consumption of 3.5 m3 /MWh for all existing
once-through power plant within 2 years.
34 S. A. Nihalani et al.

Table 3.3 Water consumption for river and sea water based plant
S. No. Description Without ash water With ash water With ash water
recovery recovery recovery and 100%
fly ash utilization
A Seawater-based plant considering COC of 1.3
1 Specific water 8.7 m3 /MWh 8.7 m3 /MWh 8.7 m3 /MWh
consumption
2 With Ash water 7300 m3 /h 7800 m3 /h 8400 m3 /h
recovery and 100%
fly ash utilization
3 Is zero liquid No No No
discharge achieved
B River water-based plant considering COC of 5
1 Specific water 3.2 m3 /MWh 2.65 m3 /MWh 2.5 m3 /MWh
consumption
2 Excess CW 0 m3 /h 0 m3 /h 220–250 m3 /h
blowdown water
available in plant
3 Is zero liquid Yes Yes No
discharge achieved

• All new coal-based plants to be installed after 1st January 2017 should reduce
specific water consumption value to 2.5 m3 /MWh and should achieve zero liquid
discharge.
• All existing and new coal-based power plants have to meet the stringent require-
ments of SO2 and NOx emissions. To meet the new emission standards, most of
the power plant has to install Flue gas desulfurization units.
From above, it is observed that all the upcoming power plants irrespective of water
sources available (i.e. Sea water /river water /bore well) have to install closed cooling
system and flue gas desulfurization units. Further, flue gas desulfurization system
also uses water for removal of SO2 and other impurities from flue gases. There are
various technologies available for SO2 removal from the flue gas stream. Among
various technologies including sea water based FGD system, wet limestone FGD
system is considered in this assessment for large coal-based power plants. Makeup
water requirement for Wet limestone FGD system has been estimated in the range
of 220–250 m3 /h based on supplier’s inputs for 2 × 660 MW units.
In view of above scenario, it is observed that new environment norms could
not be achieved with the help of present water resource management adopted in
coal-based power plants using sea water and plant where higher COC could not
be possible. The implementation of new environment norms must need to evaluate
various alternatives for further optimization of water requirements and achievement
of zero liquid discharge in coal-based power plants.
3 Water Resource Management for Coal-Based Thermal Power Plant 35

3.5 Adoption of Dry Cooling System (Air Cooled

Condenser)

In this alternative, water resource management using an air-cooled condenser is eval-

uated with no change in ash handling system operation. In air-cooled condenser heat
is rejected to the atmosphere by sensible cooling only. It does not involve evaporative
cooling. Hence makeup water requirement for condenser cooling reduces to zero.
The water resource management for a plant using an air-cooled condenser with fly
ash HCSD system and plant using wet cooling tower with fly ash HCSD system can
be correlated as Table 3.4.
Above table shows that specific water consumption value of this alternative is
almost 50% lower than the wet cooling system. With this alternative, new environ-
ment norms of specific water consumption, as well as zero liquid discharge, could
easily be achieved in all operating scenarios of plant irrespective of the quality of
water.

Table 3.4 Consumptive water requirement of 2 × 660 MW power plant with WCC and ACC
S. No. Systems Water Specific water Water Specific water
consumption consumption consumption consumption
(m3 /h) (m3 /MWh) (m3 /h) (m3 /MWh)
Water-cooled condenser Air-cooled condenser
1 Makeup water 3000 2.27 0 0
requirement in
the condenser
cooling water
system
2 Makeup water 200 0.15
requirement in
the auxiliary
cooling system
3 Total makeup 1400 1.06 1400 1.06
water
requirement for
ash handling
plant
4 Total makeup 150 0.11 150 0.11
water
requirement for
coal handling
plant
(continued)
36 S. A. Nihalani et al.

Table 3.4 (continued)

S. No. Systems Water Specific water Water Specific water
consumption consumption consumption consumption
(m3 /h) (m3 /MWh) (m3 /h) (m3 /MWh)
Water-cooled condenser Air-cooled condenser
5 DM water 60 0.05 60 0.05
requirements
6 Miscellaneous 250 0.19 250 0.19
service water
and potable
water
requirements
including
HVAC, pump
sealing
7 FGD system 300 0.22 300 0.22
makeup water
requirements
8 Total water 5160 3.9 2360 1.8
requirements
9 CW blowdown 600 70
water generated
10 Other waste 30 30
generated
including DM
plant waste,
CPU waste,
floor washing
waste, and oily
waste
11 Total waste 630 100
collected
12 Reduction in ash 480 100
handling plant
makeup water
requirement due
to the utilization
of plant waste
13 Reduction in 150 0
coal handling
plant makeup
water
requirement due
to the utilization
of plant waste
(continued)
3 Water Resource Management for Coal-Based Thermal Power Plant 37

Table 3.4 (continued)

S. No. Systems Water Specific water Water Specific water
consumption consumption consumption consumption
(m3 /h) (m3 /MWh) (m3 /h) (m3 /MWh)
Water-cooled condenser Air-cooled condenser
14 Net raw water 4530 3.4 2260 1.7
makeup
requirement

3.6 Additional Water Conservation Techniques Used

in TPP

The utility of tremendously large volumes of water in the thermal power plants
demands that the thermal power plant companies undertake numerous water conser-
vation measures. Apart from the use of an air-cooled condenser, following water
conservation techniques may also help to achieve new environmental norms of water
consumption for many power plants.

3.6.1 Adoption of 100% Utilization of Fly Ash in Dry Mode

Present water balance of river water based plant using HCSD system for fly ash
disposal estimates specific water consumption in the range of 3.2 m3 /MWh during
the initial period of plant operation. This meets the criteria of new environment norms
specified for many existing power plants. These power plants could achieve specific
water consumption of 3.5 m3 /MWh by utilizing 100% dry fly ash disposal in place
of conventional lean slurry disposal system and using recovery water from ash dyke.
For river water based plants to be installed after January 2017, new environment
norms of 2.5 m3 /MWh could be achieved using 100% fly ash disposal in dry mode
after 1 year of plant operation when recovery of water also starts from ash dyke
area. However, during the above scenario, there was surplus CW blowdown water
in the range of 220–250 m3 /h which could now be utilized to meet the makeup
water requirement of recently added FGD system. Further, a wet FGD system also
generates waste in the range of 10–15 m3 /h for 2 × 660 MW capacity plant. As
waste from FGD system is less, this can be disposed to ash dyke after the removal of
suspended solids and heavy metals through a suitable treatment process. With above
consideration, zero liquid discharge could be possible in the plant. This alternative
helps to achieve new environment norms for river water based plant after 1 year of
plant operation without a substantial change in present systems of the plant. However,
for plants where 5 COC of water could not be achieved due to the quality of raw water
(in case of bore well and seawater), it is not possible to achieve new environmental
norms of 2.5 m3 /MWh and zero liquid discharge with the adoption of 100% dry fly
ash utilization.
38 S. A. Nihalani et al.

3.6.2 Increasing Cycle of Concentration for Circulating

Cooling Water System

Makeup water requirement in the CW system is a summation of cooling tower evap-

oration loss, CW blowdown and cooling tower drift loss. Cooling tower evaporation
loss and drift loss are a function of cooling tower performance and could not be
reduced. CW blowdown loss is a function of the cycle of concentration being main-
tained in the CW system. The cycle of concentration for CW system will depend
on the quality of makeup water. Generally, river /canal water having low dissolved
solids may achieve a higher cycle of concentration in the range of 6 to 7. Most of the
recent river water based plant has been designed to achieve 5 COC. Increasing cycle
of concentration will reduce the amount of CW blowdown waste which in turn reduce
net CW make-up water requirement. For typical 2 × 660 MW plant, increasing COC
from 5 to 6 will reduce CW blowdown quantity in the range of 60–80 m3 /h.

3.6.3 Installation of Ash Water Recovery System from Ash

Dyke

Ash dyke receives the ash slurry from bottom ash and fly ash disposal system. In
Ash dyke, heavy particles of ash will get settled at the bottom. Over a period of time,
clear water from the ash slurry will overflow from ash dyke area. To reuse clear water
received from ash dyke area, recovery system will be installed. In Ash water recovery
system, clear water overflow from ash dyke area will be passed through clarifiers for
further clarification. Clarified water is then recycled/reused for ash slurry preparation.
Generally, overflow from ash dyke area starts after one year of plant operation based
on ash dyke volume. Installation of ash water recovery system recovers approx. 70%
of water. Hence, Ash handling plant water requirement during ash water recovery
scenario will reduce to 30% of water being supplied at initial plant operation.

3.6.4 Recycling of CW Blowdown to Other Systems

CW blowdown is waste generated from circulating cooling water system to maintain

dissolved solids levels. CW blowdown can directly be reused for ash slurry prepa-
ration as well as for coal dust suppression system in power plant. Reusing of CW
blowdown in ash handling plant and coal handling plant also reduces fresh makeup
water requirements. However, dry fly ash utilization, as well as ash water recovery
scenario, creates a situation of excess CW blowdown waste available within the plant.
This excess CW blowdown water needs to be treated and recycled suitably to meet
the zero liquid discharge condition.
3 Water Resource Management for Coal-Based Thermal Power Plant 39

3.7 Conclusion

New environment norms for water reduction and zero liquid discharge was reviewed
along with existing plant water resource management and possible alternative options
are discussed and compared to optimize the water requirements for existing as well
as future power plants. The outcome of various alternative options discussed in this
paper are summarized below:
• For Indian conditions, adopting dry cooling system shall result in a reduction of
plant output by about 7% and similarly, the gross heat rate shall increase by about
7%. The capital cost of the plant per MW shall increase by about 10%.
• For future river water based plants, installation of wet limestone FGD system,
HCSD system for fly ash disposal and achieving 100% utilization of dry fly ash
from 1st year could achieve the specific water consumption of 2.5 m3 /MWh with
zero liquid discharge after 1 year of plant operation when recovery of water from
ash dyke starts.
• For existing river water based plants, installation of wet limestone FGD system
and achievement of
• 100% dry fly ash utilization could meet the new environment norms.
• For seawater based once through the plant, Installation of the closed cooling
system, as well as FGD system, is now necessity of new environment norms.
Since seawater FGD plant requires a huge amount of seawater, wet limestone
FGD system is the probable solution to reduce the FGD water requirements in
seawater-based plants. Further, Installation of the wet cooling tower for seawater
based plants also end up with a huge amount of excess CW blowdown water in the
range of 7300–8400 m3 /h. Achievement of zero liquid discharge with such huge
amount of excess blowdown does not seem practical with the help of available
water treatment technologies in the market.
• Specific water consumption of 3.5 m3 /MWh could also not be achieved for existing
seawater based plant with the installation of wet cooling towers. At present, the dry
cooling system (air cooled condenser) is only probable alternatives to meet the new
environment norms for seawater-based plants. However, installation of the air-
cooled condenser to existing seawater based plants using a water-cooled condenser
is challenging in terms of meeting rated plant output as well as retrofitting of the
air-cooled condenser in place of the water-cooled condenser.
• For new as well as existing plants using bore well water which operates on
low COC of 3–4, specific water consumption along with zero liquid discharge
could be achieved by utilizing 100% fly ash and installing wastewater treatment
plants for excess blowdown water. However, in some of the cases, costly thermal
desalination technology may be required to achieve zero liquid discharge.
• Industry and Academia are working towards reducing the operating performance
and capital cost gap between dry cooling and the wet cooling. With stringent
environmental norms and scarcity of water favours use of the dry cooling system
and would be the prominent technology in future.
40 S. A. Nihalani et al.

Acknowledgements The study is part of Industry Defined Research Project titled “To Reduce
Water Consumption and Optimize Efficiency of Thermal Power Plants by Using Air Cool
Condensers” funded by Royal Academy of Engineering under Newton Bhabha Fund in collaboration
of L&T Power with Parul University.

References

Notification issued by Ministry of Environment, Forest and Climate Change dated 7 December
2015
Water Use and Efficiency in Thermal Power Plants, Report by FICCI
Water balance diagram for 2 x 660 MW plant engineered by L&T-Sargent & Lundy Limited
Chapter 4
Evaluation of Reservoir Sedimentation
Using Satellite Data—A Case Study

Beeram Satya Narayana Reddy and S. K. Pramada

Abstract India is an immense nation with high spatial and temporal inconsistency
of precipitation. The end target is to tap the available water assets and to use the water
as per the necessity. Numerous dams have already constructed to tap the available
water resources. But the storage capacity of these reservoirs is gradually reducing
due to sedimentation. To decide the useful existence of reservoir and to evaluate the
sedimentation rate in the reservoir, it is fundamental duty to conduct the surveys inter-
mittently. If the actual sedimentation process evaluated in the reservoirs, measures
can propose ahead of time, and reservoir operational plan can arrange for typical
usage of water. Some traditional procedures of measurement of reservoir sedimen-
tation like hydrographic surveys and inflow-outflow techniques are expensive and
also tedious. Remote sensing method turns into a helpful instrument for assessment
of reservoir sedimentation because of its spatial, spectral and time-related features,
which give succinct, monotonous and auspicious information concerning the area
of water spread of the reservoir. Because of the deposition of sediments in a reser-
voir, the water spread at a height continuously alters. Presently by looking at the
decreasing area of water spread with time, the sedimentation rate in a reservoir can
be determined easily. In the current research, the sedimentation rate and volume
ascertain within the Thenmala reservoir were carried out utilising satellite data. The
data of IRS-P6 satellite with LISS-III sensor for 6 different date of pass was used to
extract the area of water spread. The initial elevation-area-capacity curves and the
reservoir levels on the six date of pass were acquired from the authorities of the dam.
Utilising the cone formula, the capacity in zone of live storage is calculated. Because
of silt deposition inside the reservoir, 44.063 (M Cu.m) storage capacity lost in a
traverse of 30 years, i.e. 1985–2014.

Keywords Linear Imaging Self Scanning (LISS-III) · Normalised Difference

Water Index (NDWI) · Area of water spread · Reservoir sedimentation · Satellite
remote sensing

B. S. N. Reddy (B) · S. K. Pramada

Civil Engineering, NIT, Calicut, Kerala, India
e-mail: satyanarayana_p170016ce@[Link]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 41

R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
42 B. S. N. Reddy and S. K. Pramada

4.1 Introduction

Dam is designed across river aiming in irrigation, power generation, regulation and
supervision of discharge. Nowadays many reservoirs are not performing their design
operation because the sediment is occupying the operational capacity of storage
(Ijamand Al-Mahamid 2012). Once the stream enters a storage reservoir, attributable
to the small flow rate in the reservoir, they tend to the settlement of sediment. As
settlement of silt happens bit-by-bit, operational capacity in the reservoir reduced
which makes inefficient of performing actual purpose of impoundment of dam as
time passes.
Due to the construction of dams on natural streams which forms a reservoir will
subject to some extent of silt influx and accumulation, which in turn reduces the reser-
voir life. Periodical capacity surveys of reservoirs facilitate in judging the speed of
sediment deposition and depletion in capacity of storage (Jeyakanthan and Sanjeevi
2013). Bathymetry and sediments spatial diffusion modelling, geographical infor-
mation system (GIS) can be adapted for sedimentation studies. Many researchers
had used remote sensing technology in the assessment of sediment accumulation
(Goel et al. 2002; Jain et al. 2002; Pandey et al. 2016; Yeo et al. 2014; Narasayya
et al. 2013). Usage of satellite image recently becomes economical and expedient to
figure out the sediment deposition and to determine its dispersal and accumulation
pattern (Narasayya et al. 2013).
Satellite image techniques offer information obtainment over an extended amount
of duration which provides comprehensive, monotonous and timeous information
about reservoir sediment deposition nature. Reservoir area of water spread at specific
elevation may be acquired meticulously with satellite imagery. Curtailment if any,
within the region of water spread at particular elevation is due to sediment deposition.
Multi-date of the pass of satellite imagery related to the study area is required, from
which area of water spread extracted will permit assessment of storage volume lost
due to sediment deposition.
The present aim of the study is to assess the capacity loss in Thenmala reservoir,
Kollam, Kerala, India due to sedimentation through satellite data, to update elevation-
area-capacity curve and to estimate the loss of storage due to sediment deposition.

4.2 Description of Study Area

Thenmala dam was constructed in the year 1985 on the river Kallada in Kollam district
in Kerala. Kallada River is a west-flowing river which originates from Kulathupuzha
ranges. After flowing through 120 km in Kollam district, it falls into Ashtamudi Lake.
The study area lies between latitudes 08° 49 N to 09° 17 N and longitudes 77° 16
E to 76° 24 E. It covers portions of Thiruvananthapuram, Kollam and Alleppey
districts. The location of the dam is 08° 57 N and 77° 4 E. Kallada River forms by
the confluence of three major tributaries of river viz., Kulathupuzha, Senthuri and
4 Evaluation of Reservoir Sedimentation Using Satellite … 43

Kalthuruthi. The map of the study area is shown in Fig. 4.1. Area of watershed is
549 km2 . Main benefits of reservoir are irrigation and power generation. The salient
features of the dam are shown in Table 4.1.

Fig. 4.1 Location of study area

Table 4.1 Salient features of Thenmala dam

Official name Kallada dam
Location Kollam District, Kerala
Coordinates 08° 57 00 N 77° 04 20 E
Purpose Irrigation, hydro power
Status Operational
Opening date 1986
Dam and spillways
Height 85.35 m
Length 335 m
Spillways 3
Spillway type Ogee with radial gate auxiliary-Labyrinth
Reservoir
Catchment area 549 km2
Surface area 23 km2
Gross storage 504.92 Mm3
Live storage 487.92 Mm3
Dead storage 17 Mm3
Storage between RL 88 m to FRL 115.82 423.953 Mm3
44 B. S. N. Reddy and S. K. Pramada

Table 4.2 Satellite image collected from NRSC Hyderabad

Satellite Sensor Path Row Date of pass Elevation in m
IRS-P6 LISS-III 100 67 22 February 2008 109.3
IRS-P6 LISS-III 100 67 20 January 2012 112.22
IRS-P6 LISS-III 100 67 05 October 2013 115.15
IRS-P6 LISS-III 100 67 16 December 2013 114.05
IRS-P6 LISS-III 100 67 26 February 2014 106.72
IRS-P6 LISS-III 100 67 04 January 2015 114.25

4.3 Data and Software Used

LISS-III sensor data of IRS-P6 satellite with a spatial resolution 23.5 m used for the
analysis. The data was purchased from National Remote Sensing Centre (NRSC)
Hyderabad. Data collected from NRSC Hyderabad of LISS-III was available in 4
bands of different wavelengths regions like Band-2 (Green), Band-3 (Red), Band-4
(NIR) and Band-5 (SWIR) (Short-wave Infrared). Each band will have 23.5 m as
spatial resolution. Date of the pass of satellite images are given in Table 4.2. False
colour composite (FCC) images which prepared from satellite images with available
time of pass are shown in Fig. 4.2.
Original Elevation-area-capacity curve and reservoir level on satellite date of pass
were acquired from Kallada dam authorities.
Software used for image processing and area of water spread calculation are as
ERDAS IMAGINE (Image processing software) and ArcMap (Area calculation).

4.4 Methodology

The main intention involved here is to extract the area of water spread at different
elevation in a reservoir, i.e. between full reservoir level (FRL) and minimum draw-
down level (MDDL) with the available cloud-free satellite imageries. Flowchart of
the methodology adopted in the present study is shown in Fig. 4.3.
Firstly, select the satellite imagery which is free from cloud cover. The data
was acquired from National Remote Sensing Centre (NRSC), Hyderabad. The
geo-referenced images were utilised to get normalised difference water index
(NDWI).
In the visible region (0.4–0.7 µm) of spectrum the significance of transmission,
absorption and coefficient of reflection in the water is very less. In the near-IR band,
absorption of wavelengths in water increases briskly, whereas both the factor of
reflection and transmission is less. With the usage of different signatures of spec-
tral land use types, water pixel extracted from other land use pixels such as soil,
vegetation, barren land, etc.
4 Evaluation of Reservoir Sedimentation Using Satellite … 45

Fig. 4.2 False colour composite of Thenmala reservoir

The normalised difference water index (NDWI) will be accustomed to determine

the pixels of water within the imagery. Water absorbs maximum radiation in the NIR
region.
The NDWI image processing done as below:
46 B. S. N. Reddy and S. K. Pramada

Reservoir data
Satellite data Other survey details
(Original E-A-C table)

Selection of date of pass

Preparation of NDWI image

Using Erdas imagine, keeping

threshold of a pixel value for
water for the area of water
spread extracted

Removal of discontinuous pixels

and the tail end of a reservoir

Area of water spread calculation

using ArcMap

Delineate multi-date area of water spread

Capacity assessment

Storage loss assessment

Fig. 4.3 Flowchart of the methodology

Green − NIR BAND2 − BAND4

i.e., NDWI = OR .
Green + NIR BAND2 + BAND4

Selection of the above wavelengths enhances the features of water as mentioned

below.
1. Intensifies the distinctive reflectivity in water attributes with the usage of the
green band;
2. Curtailment the less reflectivity of Near-IR in water attributes; and
3. Considering the benefit of the high reflectivity of Near-IR by soil and vegetation
attributes.
NDWI image can visualise for water pixels and nonwater pixels. So from this
NDWI image, all water pixels of the study area are extracted and processed and the
4 Evaluation of Reservoir Sedimentation Using Satellite … 47

area of water spread is estimated. The same process is repeated with the satellite
imageries of the different date of pass.
The calculated area unit brought in this operation are used to calculate the volume
between 2 consecutive elevations by cone formula. The volume of water (V ) between
2 sequential spread A1 and A2 and elevation difference (H) is calculated as
 √ 
V = H ∗ A1 + A2 + A1 ∗ A2 /3

The calculated accumulated capability of the reservoir has obtained by adding

calculated volumes between consecutive intervals (Ninija Merina et al. 2016).

4.5 Results and Discussion

Extracted areas of water spread at different date of pass are given in Fig. 4.4 and the
calculated area of water spread from satellite imagery is given in Table 4.3.
Extensive care should be taken in the removal of discontinuous water pixels and
the water pixels at the downstream of the reservoir, specifying the tail end of the
reservoir. There may be a chance of misinterpretation of area estimation if careful
editing is not done.
Using the water spread area extracted from imagery, capacity of reservoir is calcu-
lated using cone formula and cumulative of calculated capacity is given in Table
4.4.
Live capacity lost due to sedimentation for the survey year 2014 is 44.063 Mm3
for a span of 30 years, i.e. from 1985 to 2014 between elevation (106.72) and FRL
(115.82). Live capacity lost due to sedimentation is given in Table 4.5.

4.6 Conclusions

The accurate estimation of reservoir sedimentation is very much essential to deter-

mine the storage capacity of reservoirs and there by effiecient operation of reservoirs.
Satellite data is very much useful in this aspect. In this study, Tenmala Dam, Kollam,
Kerala, India is taken as the case study to estimate the reservoir sedimentation using
satellite data and to estimate the revised elevation capacity curve. In the year 2014,
Thenmala dam live storage is calculated as 379.890 Mm3 . In a span 30 years, i.e.
from 1985 to 2014 loss of reservoir capacity is 10.39% observed in the zone of live
storage. The calculated rate of average siltation was 2.676 mm/year.
The major limitations of remote sensing technique are as follows:
• As the water flucutation in a reservoir can take place from penstock level to full
reservoir level. So area of water spread extraction from imagery is restricted only
to the zone of live storage, below that is not possible.
48 B. S. N. Reddy and S. K. Pramada

Fig. 4.4 Extracted area of water spread with date of the pass of Thenmala Reservoir
4 Evaluation of Reservoir Sedimentation Using Satellite … 49

Table 4.3 Area of water spread estimated from satellite data

Date of pass Elevation (m) Area (Mm2 )
22 February 2008 109.3 16.555441
20 January 2012 112.22 16.854026
05 October 2013 115.15 17.977752
16 December 2013 114.05 17.339904
26 February 2014 106.72 15.389784
04 January 2015 114.25 17.508384

Table 4.4 Comparsion of calculated capacity with the original capacity

Elevation Original Calculated Original Calculated original Calculated
(m) WSA WSA (Mm2 ) volume volume cum. cum.
(Mm2 ) (Mm3 ) (Mm3 ) Capacity capacity
(Mm3 ) (Mm3 )
106.72 18.828 15.390 227.142 227.142
109.3 20.325 16.555 50.681 41.159 277.823 268.301
112.22 21.683 16.854 61.337 48.728 339.160 317.028
114.05 22.455 17.340 40.341 31.255 379.501 348.284
114.25 22.516 17.508 4.518 3.481 384.019 351.765
115.15 22.834 17.978 21.149 15.952 405.168 367.717
115.82 23.239 18.397 18.785 12.173 423.953 379.890

Table 4.5 Live capacity lost due to sedimentation

Original data (1985) Survey year (2014) SRS
Live capacity in mm3 between RL and 423.953 379.890
FRL
Sediment deposited in two consecutive 44.063
surveys in Mm3
Period in years 30
Silt deposition between two consecutive 1.469
surveys Mm3 /year
Capacity loss
– Mm3 44.063
– % to original capacity 10.39

• Lack of accessibility in cloud free imageries throughout the operation of reservoir.

50 B. S. N. Reddy and S. K. Pramada

References

Goel MK, Jain SK, Agarwal PK (2002) Assessment of sediment deposition rate in Bargi Reservoir
using digital image processing. J Hydrol Sci 47(5):81–92
Ijam AZ, Al-Mahamid MH (2012) Predicting Sedimentation at Mujib Dam Reservoir in Jordan.
Jordan J Civ Eng 6(4):448–463
Jain SK, Singh P, Seth SM (2002) Assessment of sedimentation in Bhakra Reservoir in the western
Himalayan region using remotely sensed data. Hydrol Sci J 47(2):203–212
Jeyakanthan VS, Sanjeevi S (2013) Capacity survey of Nagarjuna Sagar reservoir, India using linear
mixture modal (LMM) approach. Int J Geomat Geosci 4(1):186–194
Narasayya K, Roman UC, Sreekanth S, Jatwa S (2013) Assessment of Reservoir sedimentation
using remote sensing satellite imageries. Asian J Geoinform 12(4):172–180
Ninija Merina R, Sashikumar MC, Rizvana N, Adlin R (2016) Sedimentation study in a reservoir
using remote sensing technique. Appl Ecol Environ Res 14(4):296–304
Pandey A, Chaube UC, Mishra SK, Kumar D (2016) Assessment of reservoir sedimentation
using remote sensing and recommendations for desilting Patratu Reservoir India. Hydrol Sci
J 61(4):711–718
Reservoir sedimentation assessments method a Handbook, Bureau of Indian Standards, IS
13665:1993
Yeo I-Y, Lang M, Vermote E (2014) Improved understanding of suspended sediment transport
process using multi-temporal landsat data: a case study from the old woman Creek Estuary
(Ohio). IEEE J Sel Top Appl Earth Obs Remote Sens 7(2):636–647
Chapter 5
Regionalisation of Watersheds Using
Fuzzy C Means Clustering Algorithm
in the West Flowing River of Kerala

Thottungal Krishnankutty Drissia, Vinayakam Jothiprakash,

and Alayil Bahuleyan Anitha

Abstract Regionalisation of the watersheds, defined as the grouping of statistically

similar watersheds is generally carried out for regional flood frequency analysis. This
also helps in rainfall-runoff modelling. In this study, the catchment area of 43 stream
gauge in the west flowing rivers of Kerala is grouped using fuzzy c means (fcm)
clustering algorithm. The attributes in the feature vector and number of clusters are
two deciding factors, which has no particular guidelines. Hence, three sets of feature
vectors from the catchment characteristics formed for analysis. The first set includes
catchment area (A), main channel length (Lc), the circulatory ratio (Rc), average
rainfall (P), the total length of streams (Lu), the total number of streams (Nu). Lc, A,
Rc, P, Lu, Nu, L-CV (L coefficient of variance) forms the second group and Lc, A,
Rc, P, Lu, Nu, L-CV, latitude and longitude forms the third group. The fuzzy c means
clustering was carried out with 4, 5 and 6 cluster numbers for each feature vector. The
FCM clustering partitions the rescaled feature vector by minimising the objective
function. Clusters were validated using cluster validity indices such as partition
coefficient (PC), classification entropy (CE), partition index (SC), separation index
(S) and Xie and Beni’s index (XB). L moments heterogeneity test (H Test) was used
to test the homogeneity of each cluster. It was found that five clusters formed from the
second group of feature vector with attributes Lc, A, Rc, P, Lu, Nu, L-CV produced
homogeneous regions.

Keywords Fuzzy c means cluster · L moment · Heterogeneity test · Validity

index · Regionalisation

T. K. Drissia (B) · V. Jothiprakash · A. B. Anitha

Indian Institute of Technology Bombay, Powai, Maharashtra 400076, India
V. Jothiprakash
e-mail: vprakash@[Link]
T. K. Drissia · A. B. Anitha
Centre for Water Resources Development and Management, Kozhikode, Kerala 673571, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 51

R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
52 T. K. Drissia et al.

5.1 Introduction

Regionalisation of watersheds is the first step in regional flood frequency analysis,

which identifies regions behaving in a hydrologically similar fashion (Nathan and
McMahon 1990). In addition, regionalisation helps in developing the rainfall-runoff
model for that particular region. Amongst the many statistical techniques used for
regionalisation includes L moment method (Hosking 1990), the region of influence
approach (Burn 1990), canonical correlation analysis (Ouarda et al. 2001), ward’s
cluster method (Jingyi and Hall 2004), hierarchical and trellis method (Chérif and
Bargaoui 2013).
Clustering is defined as the classification of data into groups so that the elements
in that group process some similar characteristics. Many clustering techniques have
been used in regionalisation. Some of the clustering algorithms are the k means
algorithm and the Fisher algorithm (Hartigan 1989). From a hydrological point of
view, clustering attempts to identify a cluster or group of catchments with similar
catchment characteristics and are dissimilar with other clusters (Burn and Goel 2000).
Two types of clustering algorithm were defined in the literature. One is partitional
clustering and the other hierarchical clustering (Rao and Srinivas 2008). Hierarchical
clustering gives nested sequence of partitions, whereas partitional clustering produce
partition of the data (Rao and Srinivas 2008). Soft computing techniques employed
in regionalisation are self-organising map (SOM), artificial neural network (ANN)
(Kumar and Chatterjee 2015), fuzzy c means (fcm) algorithm (Rao and Srinivas
2006; Budayan et al. 2009). A comparison of the artificial neural technique and fuzzy
clustering technique shows that fuzzy clustering techniques were better (Kumar and
Chatterjee 2015). Two fuzzy clustering techniques are available, Fuzzy c means
(fcm) clustering algorithm and subtractive algorithm (Bataineh et al. 2011). In fuzzy
clustering of watersheds, one watershed can belong to more than one cluster. In
the real case , one watershed may resemble more than one group of catchments.
Hence, fuzzy c means (fcm) clustering gives more fruitful results for catchment
classification.
Though many works of literature were reported on the application of soft
computing techniques in rainfall-runoff modelling, few were reported in regional
flood frequency analysis. In this context, the application of the fuzzy c means algo-
rithm (fcm) in the regionalisation of watersheds in west -flowing rivers (WFR) of
Kerala was attempted. WFR consists of numerous small and medium rivers that have
high variation in topography and climate. Hence, identifying homogeneous regions
in this part of the country is challenging.
5 Regionalisation of Watersheds Using Fuzzy C Means Clustering … 53

5.2 Study Area and Data

West flowing rivers (WFR) of Kerala is the study area considered for the anal-
ysis. WFR of Kerala forms a major part of the zone ‘WFR from Tadri to Kanyaku-
mari’, a zone defined by the Central Water Commission (CWC), Govt. of India.
This region has 41 WFR that originates from Western Ghats and drains to Arabian
Sea. Kerala receives an average annual rainfall of 3000 mm mainly by two monsoon
seasons, southwest and northeast.
43 watersheds in WFR were selected based on the catchment area of stream gauge
stations maintained by the CWC, (20 stations) and Water Resources Department,
Government of Kerala (23 stations) (Table 5.1). The boundary and drainage of the

Table 5.1 Details of watersheds

S. No. Rivergauging Catchment area River basin
1 Manjeswaram 25.44 Manjeswar
2 Anakkallu 166.25 Uppala
3 Shiriya d/s 322.56 Shiriya
4 Shiriya u/s 348.00 Shiriya
5 Madhur 66.04 Mogral
6 Moonnamkadavu 216.80 Chandragiri
7 Erivanjipuzha 957.00 Chandragiri
8 Kakkadavu 276.50 Karingode
9 Mangara 109.60 Kuppam
10 Irude 189.63 Valapattanam
11 Palapuzha 237.25 Valapattanam
12 Perumannu 1070 Valapattanam
13 Kannavam 60.75 Anjarakandy
14 Meruvambai 180.00 Anjarakandy
15 Kuttiyadi 238 Kuttiyadi
16 Kollikkal 34.82 Korapuzha
17 Chaliyar 386.00 Chaliyar
18 Kanjirapuzha 64.00 Chaliyar
19 Kuniyil 1876 Chaliyar
20 Karathodu 750 Kadalundi
21 Ambarampalayam 950 Bharathapuzha
22 Kumbidi 5755 Bharathapuzha
23 Mankara 2,775 Bharathapuzha
24 Pudur 1313 Bharathapuzha
(continued)
54 T. K. Drissia et al.

Table 5.1 (continued)

S. No. Rivergauging Catchment area River basin
25 Pulamanthole 940 Bharathapuzha
26 Arangaly 1342 Chalakudy
27 Karuvannur 725.00 Karuvannur
28 Ambalakadavu 1160.00 Chalakudy
29 Neeleeswaram 4234 Periyar
30 Kalampur 405 Moovattupuzha
31 Ramamangalam 1208 Moovattupuzha
32 Kidangoor 615 Meenachil
33 Teekoy 57.00 Meenachil
34 Manimala 490.00 Manimala
35 Kallooppara 731 Manimala
36 Malakkara 1713 Pamba
37 Thumpamon 810 Achankoil
38 Kalleli 419.00 Achenkoil
39 Kollakkadavu 952.71 Achenkoil
40 Punalur 870.00 Kallada
41 Pattazhi 1210 Kallada
42 Ayilam 540 Vamanapuram
43 Ottasekharamangalam 247.35 Neyyar

watersheds were digitised. This was used for estimating the catchment characteris-
tics. The details of watersheds are given in Fig. 5.1. The weighted average rainfall
of the watersheds was also estimated.

5.3 Methodology

Clustering of watersheds was carried out using fcm clustering algorithm. The
methods are explained in following sections.

5.3.1 Fuzzy Clustering Algorithm

In fuzzy c means clustering the attributes are selected to make a matrix of gauging
stations and attributes (Rao and Srinivas 2008). Based on the correlation between
the topographic parameters and peak discharge eleven attributes were selected for
creating feature vector. The parameters are the area of catchment up to stream gauge
5 Regionalisation of Watersheds Using Fuzzy C Means Clustering … 55

Fig. 5.1 Value of validity indices for the clusters formed out of feature vector Lc, A, Rc, P, Lu, Nu

station, main channel length, maximum basin length, shape factor, the circulatory
ratio, relief ratio, the total length of streams, average annual rainfall, the total number
of streams, latitude and longitude. The details of topographic parameters are given
in Sect. 5.4.1. The coordinates included in the feature vector accounts the proximity
of the stations. The feature vector is rescaled for bringing the values in 0–1 scale.
For rescaling the values of the parameters, Eq. 5.1 is used. Rescaling is necessary,
as there is difference in magnitude, unit and importance.

x − x(min)
x= (5.1)
x(max) − x(min)

This feature vector is used for clustering using fuzzy c means algorithm. It is given
by Yk = [y1k + y2k + · · · + ynk ] where yik is denotes the value of attribute i in yk .
The rescaled feature vectors represented as a matrix X provided in Eq. 5.2.
⎡ ⎤
x11 . . . x1N
X = ⎣ ... ... ⎦ (5.2)
xn1 xn N

The centroids of the clusters (c) are represented by V = (v1 , …, vc ) in which v1 is

the centroid of cluster 1. The following objective function is minimised to partition
the data X to c clusters.


c 
N
J (U V : X ) = (u ik )m d 2 (xk , vi ) (5.3)
i=1 k=1
56 T. K. Drissia et al.

Subjected to following constraints


c
Uik = 1 ∀k ∈ ({1, . . . , N }) (5.4)
i=1


N
0< Uik < N ∀i ∈ {1, . . . , c} (5.5)
k=1

where uik ∈ [0, 1]–membership of the kth rescaled feature vector x k in the ith fuzzy
cluster.
U is the fuzzy partition matrix which contains the membership of each rescaled
feature vector in each fuzzy cluster.
The parameter μ ∈ [1, ∞] refers to the weight exponent for each fuzzy
membership.
⎡ ⎤
u 11 . . . u 1k . . . u 1N
⎢ . . . ⎥
⎢ ⎥
⎢ ⎥
U = ⎢ u i1 u ik ui N ⎥ (5.6)
⎢ ⎥
⎣ . . . ⎦
u c1 u ck u cN

The distance from kth rescaled feature vector x k to the centroid of ith cluster vi

d 2 (xk , vi ) = (xk − vi )T Ai (xk − vi ) (5.7)

where the norm Ai is a positive definite symmetric matrix associated with cluster i.
For estimation of Euclidean distance between x k and vi , Ai = I ∀i, where I is a unit
matrix.

5.3.2 Validity Index

Many cluster validity index are available in literature. Some of the indices relevant to
fuzzy c means clustering are applied in this study. This include partition coefficient
(PC), classification entropy (CE), partition index (SC), separation index (S) and Xie
and Beni’s index (XB).
The partition coefficient formulated by Bezdek (1974) uses only membership
values and is given by Eq. 5.8.

1 
c N
VPC (U ) = (u ik )2 (5.8)
N i=1 k=1
5 Regionalisation of Watersheds Using Fuzzy C Means Clustering … 57

Partition entropy or classification entropy (CE) (Bezdek et al. 1984)

c N

1 
VCE (U ) = − u ik loga (u ik ) (5.9)
N i=1 k=1

Partition index (SC) (Sote and Pande 2014)

 N  m

c
j=1 u i j x j − vi2
SC = c (5.10)
i=1
Ni k=1 vk − vi2

Separation Index (S)

c  N  m
i=1 j=1 u i j x j − vi2
S= (5.11)
N mini j x j − vi2

Xie and Beni’s index (XB) (1991)

c n
u 2 xi − vk2
XB(c) = k=1 i=1
 ik (5.12)
nx mini= j vi − v j

5.3.3 L Moments Heterogeneity H Test

Once the clusters are formed, the homogeneity in each cluster is checked with hetero-
geneity test (H test) formulated by Hosking and Wallis ( 1997). Before H test, discor-
dancy test was conducted to identify any discordant sites. For performing both the
tests L moment ratio such as L-covariance (L-CV), L skewness (L-skew) and L
kurtosis (L-kurt) were estimated based on the equations given in Hosking and Wallis
(1997). The heterogeneity test (H test) estimated is as follows.

(V − μv)
H= (5.13)
σv

where V is the weighted standard deviation of L-CV of all the stations.

μv and σv are mean and standard deviation of simulated weighted standard deviation
of L-CV.
For this, 500 realisations were simulated using four- parameter kappa distribution.
The four parameters representing mean (χ ), variance (α) and shape (k, h) were
estimated by fitting kappa distribution to regional average L moment ratios 1, t R , t3R ,
t4R . The regions with H < one regarded as “acceptability homogeneous”, H value in
the range 1–2 as “possibly homogeneous” and H > two “definitely heterogeneous”.
58 T. K. Drissia et al.

5.4 Results and Discussion

Initial screening was carried out by discordancy test and was found that two stations
were discordant. The stations are Manjeswaram and Madhur which are removed
before regionalisation.

5.4.1 Parameter Selection for Feature Vector

The catchment characteristics are correlated with peak discharge of gauging station
to find out the parameters that influence the peak discharge. The linear aspects of
catchment characteristics include mainstream length (Lc), maximum basin length
(Lb), length of overland flow (Lo), bifurcation ratio (Rb), stream length ratio (Lsr),
total length of streams (L) and number of streams (Nu). The areal aspects as area
of catchment upto gauging station (A), form factor (Ff), shape factor (Fs), circula-
tory ratio (Cr), elongation ratio (Re), drainage density (Dd), stream frequency (Sf),
constant of channel maintenance (Cc). Apart from the above parameters, relief ratio
(Rr) weighted average rainfall (P) and L moment variance (L-CV) are also considered.
All the parameters except rainfall and L-CV are correlated with peak discharge. The
parameters with high-correlation coefficient as given in Table 5.2 taken for creating
feature vector.

Table 5.2 Regression

Catchment characteristics R
coefficient
Maximum basin length (Lb, km) 0.621
Main stream length (Lc, km) 0.637
Length of overland flow (Lo, km) 0.131
Mean bifurcation Ratio (Rb) 0.092
Mean stream length ratio (Lsr) 0.048
Total stream length, km (Lu) 0.773
Number of streams (count) (Nu) 0.696
Catchment area (A, km2 ) 0.711
Form factor (Ff) 0.098
Shape factor (Fs) 0.391
Circularity ratio (Rc) 0.337
Elongation ratio (Re) 0.092
Drainage density (Dd, km/km2 ) 0.093
Stream frequency (Sf) 0.007
Constant of channel maintenance (Cc) 0.131
Relief ratio (Rr) 0.194
5 Regionalisation of Watersheds Using Fuzzy C Means Clustering … 59

The parameters with high-regression coefficient as in Table 5.1 are catch-

ment area (A), maximum basin length (Lb), mainstream length (Lc), total stream
length (Lu), number of streams (Nu), shape factor (Fs) and circulatory ratio (Rc).
The coordinates (lat and long) are also considered to understand the influence of
proximity on cluster formation.

5.4.2 Fuzzy C Means Clusters

The input required for fcm clustering are attributes for feature vector and number
of clusters. There are no guidelines for the selection of attributes and the number of
clusters. Initially, all the parameters were used as the feature vector, but no homoge-
nous regions formed. Thereafter, the feature vector was created out of the rescaled
value of the ten parameter based on the regression coefficient, A, Lb, Lc, L, Nu,
Fs, Rc, lat, long, P (Table 5.2). The maximum basin length (Lb) and main stream
length (Lc) are similar characteristics and hence, only Lc is selected for the clus-
tering. Likewise, only total stream length (Lu) was selected between Lu and Nu.
Shape factor (Fs) and circulatory ratio (Rc) are related to perimeter and catchment
area and hence only one is selected. Based on these three sets of feature vectors are
[Lc, A, Rc, P, Lu, Nu], [Lc, A, Rc, P, Lu, Nu, L-CV] and [Lc, A, Rc, P, Lu, Nu, L-CV,
lat, long]. Another input required is optimum number of clusters, which is explained
in following section.

5.4.3 Validity Index

Validity indices such as partition coefficient (PC), classification entropy (CE), parti-
tion index (SC), separation index (S) and Xie and Beni’s index (XB) were esti-
mated for three cluster centres 4, 5 and 6. Partition coefficient measures the quantity
of overlap between clusters. The high value of PC and low value of CE indicate
minimum overlap between the clusters. The values of validity index estimated for
the three set of feature vectors are given Figs. 5.1, 5.2 and 5.3. But the values estimated
from the clusters says that there are some overlapping amongst the clusters.

5.4.4 L Moments H Test

The H test was estimated for three sets of feature vectors with various combinations
of the number of clusters. The H value of each cluster is given in Table 5.3. In the
first set of the feature vector [Lc, A, Rc, P, Lu, Nu], the H value was mostly above
2 indicating a heterogeneous region. The third feature vector, [Lc, A, Rc, P, Lu,
Nu, L-CV, lat, long] also produced heterogeneous regions. However, the H value
60 T. K. Drissia et al.

Fig. 5.2 Value of validity indices for the clusters formed out of feature vector [Lc, A, Rc, P, Lu,
Nu, L-CV]

Fig. 5.3 Value of validity indices for the clusters formed out of feature vector [Lc, A, Rc, P, Lu,
Nu, lat, long]

of second feature vectorwith five number of clustersis either below one or between
one and two indicating“acceptability homogeneous” and “possibly homogeneous”.
The inclusion of the L Coefficient of variance has a good impact on the formation of
clusters of watersheds. Interestingly, the inclusion of coordinates does not provideho-
mogeneous clusters. It can be stated that in WFR nearness of the watersheds is not an
influencing factor in creating homogeneous regions. Five homogeneous clusters were
formed with Lc, A, Rc, P, Lu, Nu, L-CV as attributesin the feature vector.
The location of the watersheds in each region (zone) is given in Fig. 5.4. However,
there are few nearby watersheds that form the cluster, there is no continuity in the
clustering. In addition, the watersheds in each cluster are distributed all over the
study area. Defuzzification was not carried out as it was stated by (Rao and Srinivas
2008) that considerable effort should be taken for adjusting hard clusters for forming
homogeneous regions. It was suggested to take the final outcome of the fcm clustering
algorithm as fuzzy clusters.
5 Regionalisation of Watersheds Using Fuzzy C Means Clustering … 61

Table 5.3 H value for each cluster and feature vector

S. No Feature vector Cluster name Number of cluster
4 5 6
1 Lc, A Rc, P, Lu, Nu C1 3.29 2.51 2.7
C2 NC 3.76 NC
C3 2.93 2.89 3.55
C4 NC −0.83 NC
C5 NC 0.88
C6 2.68
2 Lc, A, Rc, P, Lu, Nu, C1 NC 1.79 2.13
L-CV C2 2.24 0.78 1.82
C3 0.53 0.59 2.58
C4 2.2 1.36 NC
C5 1.54 0.9
C6 NC
3 Lc, A, Rc, P, Lu, Nu, C1 NC 2.68 2.19
L-CV, lat, long C2 2.09 2.16 NC
C3 2.39 4.76 1.52
C4 1.38 1.59 2.60
C5 1.57 1.56
C6 2.13
NC-kappa parameter not converged

5.5 Conclusion

Regionalisation of watersheds was carried out using the fuzzy c means algorithm.
The major inputs needed for fcm clustering are the number of clusters and param-
eters required for the feature vector. Out of 20 catchment characteristics selected,
seven parameters have a good correlation with peak discharge. Combinations of the
parameters are attempted and found that feature vector with catchment area (A),
mainstream length (Lc), circulatory ratio (Rc), weighted average rainfall (P), total
stream length (Lu), the total number of streams (Nu) and L moment variance (L-
CV) produce five homogeneous regions. The regions are not continuous though few
watersheds are nearby in each cluster. The selection of attributes for feature vector
has not followed any particular guidelines, hence, further analysis on procedures for
selection of suitable attributes may be attempted in future studies. Flood frequency
relationships can be developed for each cluster using the index flood method. This
will help in estimating flood frequency for the ungauged sites.
62 T. K. Drissia et al.

Fig. 5.4
Homogeneous Regions
formed using fuzzy c means
clustering algorithm

Acknowledgements Authors are thankful to the Central Water Commission, Government of India
and Water Resources Department, Government of Kerala for sharing the discharge data.

References

Bataineh KM, Naji M, Saqer M (2011) A comparison study between various fuzzy clustering
algorithms. Jordan J Mech Indust Eng 5(4):335–343
Bensaid AM, Hall LO, Bezdek JC, Clarke LP, Silbiger ML, Arrington JA, Murtagh RF (1996)
Validity-guided (Re)Clustering with applications to image segmentation. IEEE Trans Fuzzy Syst
4:112–123
Bezdek JC, Ehrlich R, Full W (1984) FCM: The fuzzy c-means clustering algorithm. Comput
Geosci 10(2–3):191–203
Budayan C, Dikmen I, Birgonul MT (2009) Comparing the performance of traditional cluster
analysis, self-organizing maps and fuzzy C-means method for strategic grouping. Expert Syst
Appl 36(9):11772–11781. [Link]
5 Regionalisation of Watersheds Using Fuzzy C Means Clustering … 63

Burn DH (1990) Evaluation of regional flood frequency analysis with a region of influence approach.
Water Resour Res 26(10):2257. [Link]
Burn DH, Goel NK (2000) The formation of groups for regional flood frequency analysis. Hydrol
Sci J 45(February):97–112. [Link]
Chérif R, Bargaoui Z (2013) Regionalisation of maximum annual runoff using hierarchical and
Trellis methods with topographic information. Water Resour Manage 27:2947–2963. [Link]
org/10.1007/s11269-013-0325-0
Hartigan JA (1989) Clustering algorithms. Wiley, pp 331–335. [Link]
Hosking JRM (1990) L moments: analysis and estimation of distributions using linear combination
of order statistics. J Roy Stat Soc 52(1):105–124
Hosking JRM, Wallis JR (1997) Regional frequency analysis: an approach based on L-moments.
Cambridge University Press, New York
Jingyi Z, Hall MJ (2004) Regional Flood Frequency Analysis for the Gan-Ming River Basin in
China 296:98–117. [Link]
Kumar R, Chatterjee C (2015) Regional flood frequency analysis using soft computing techniques.
Water Resour Manage, pp 1–7. [Link]
Nathan RJ, McMahon TA (1990) Identification of homogeneous regions for the purposes of
regionalisation. J Hydrol 121:217–238. [Link]
Ouarda TBMJ, Girard C, Cavadias GS, Bobée B (2001) Regional flood frequency estimation with
canonical correlation analysis. J Hydrol 254(1–4):157–173. [Link]
4(01)00488-7
Rao AR, Srinivas VV (2006) Regionalization of Watersheds by Fuzzy Cluster Analysis 318:57–79.
[Link]
Rao AR, Srinivas VV (2008) Regionalization of watersheds. Springer, Berlin
Sote AM, Pande SR (2014) Web based fuzzy clustering analysis. Res Inventy: Int J Eng Sci 4(11):51–
57
Xie X, Cui Y (2011) Development and test of SWAT for modeling hydrological processes in
irrigation districts with paddy rice. J Hydrol 396(1–2):61–71. [Link]
2010.10.032
Chapter 6
Analysis of Relationship Between
Landslides and Rainfall in Karwar,
Uttara Kannada District, Karnataka,
India

G. Thejashree, K. N. Lokesh, and G. S. Dwarakish

Abstract Landslide is a hazardous geological phenomenon and a common occur-

rence in hilly regions. It causes loss of life and destruction of property depending on
the intensity of its occurrence. Landslide is the movement of mass of rock, earth or
debris down the slope. One of the important triggering factors for landslide occur-
rence is rainfall and is usually considered for slope failure predictions, yet the rela-
tionship between rainfall and landslide occurrences is very complex. An attempt
is made to understand the relationship between rainfall and landslide in the study
area. A devastating landslide occurred in October 2009 in Kadwad village of Karwar
taluk in Uttara Kannada district, killing 19 people and burial of several houses. On
the same ill fated day, there occurred landslides at 21 more locations in Karwar.
The landslide incidence was triggered due to heavy rainfall on 2nd and 3rd October
2009 which resulted in mass movement of debris. For this study, daily precipita-
tion data from Karnataka State Natural Disaster Monitoring Centre/District Disaster
Professional, Karwar and landslide records from media and other sources of that
particular year were considered for the statistical analysis to understand landslide
mechanism. Analysis shows a strong correlation between landslide occurrence and
rainfall pattern. The landslide incidence is determined by the magnitude of the short
episode of rainfall intensity with the threshold value of above 260 mm in that partic-
ular year. Thus the rainfall with less than the minimum threshold value did not cause
triggering of landslide whereas landslide occurred above the threshold value. The
possible cause for the triggering of Kadwad landslide is not due to saturation of soil
caused by antecedent precipitation but it is due to short term intensive rainfall.

Keywords Rainfall · Kadwad landslide · Rainfall pattern · Threshold · Statistical

analysis

G. Thejashree (B) · K. N. Lokesh

Department of Civil Engineering, National Institute of Technology Karnataka (NITK), Surathkal,
Srinivasnagar Post, Karnataka 575025, India
G. S. Dwarakish
Department of Applied Mechanics & Hydraulics, National Institute of Technology Karnataka
(NITK), Surathkal, Srinivasnagar Post, Karnataka 575025, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 65

R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
66 G. Thejashree et al.

6.1 Introduction

Landslide is one of the disastrous geological phenomena posing grave threat to

life and property. Rainfall is one of the significant factors which causes landslide
occurrence especially in regions with extreme seasonal precipitation. Other factors
such as climatic factors, geology, geomorphology of the area and human intervention
also play a major part in inducing landslides. Much previous research has been carried
out by researchers to correlate rainfall and landslide occurrence in hilly terrains. The
most important and difficult task for landslide research is to infer how rainfall-induced
landslide process works. On the basis of previous studies on information of rainfall
data and landslides incidences, the threshold required to provoke landslide initiation
can either be related to the significant increasing amount of rainfall continuously
or to the intensity of rainfall. The most commonly used method globally for such
threshold was developed by Caine (1980) taking into consideration the intensity and
duration of rainfall event. Earlier, threshold studies have been carried out by many
researchers to establish empirical relation between landslide incidences and total
amount of rainfall occurred (Brand et al. 1984; Claessens et al. 2007 and Bai et al.
2014).
Various models have been suggested by several researchers for thresholds of land-
slide initiation due to rainfall, such as statistical analysis by correlating relationship
among intensity of rainfall (I), rainfall duration (D) and cumulative rainfall (E),
Logistic regression model, hydrological modeling and quantitative antecedent soil
water status (ASWS) models as shown by Caine (1980), Terlien (1998), Lan et al.
(2003), Tsai (2008), Sengupta et al. (2010), Rosi et al. (2012) and Bai et al. (2014).
The rainfall threshold is calculated by correlating the rainfall amount necessary for
landslide initiation with highest intensity of rainfall (Govi et al. 1985). Glade et al.
(2000) considered the effect of antecedent rainfall in rainfall threshold for initia-
tion of landslide. The rainfall threshold calculation can be formulated to adapt it
in mitigating the consequences of landslide (Jakob and Weatherly 2003). Sengupta
et al. (2010) calculated rainfall threshold based on cumulative rainfall and rainfall
duration, intensity of rainfall and Mean Annual Precipitation (MAP). In yet another
case study analysis on the relationship of landslide occurrence and rainfall was made
using antecedent rainfall saturation and regression model. The resulting threshold
values can be helpful to those institutions and organizations which deal with failures
of slope by evaluating the possibility of occurrence of landslide for diverse rainfall
events (Bai et al. 2014). Peruccacci et al. (2017) also attempted to quantify rainfall
thresholds for the landslide initiation using cumulative rainfall amount and dura-
tion. Landslide triggering phenomena such as rainfall, groundwater conditions like
permeability, antecedent degree of saturation and soil characteristics are not well
understood. Moreover, prediction of landslides due to rainfall in various regions is
difficult owing to lack of availability of requisite data and continuous monitoring of
small scale landslides events.
The present study aims to examine the relationship between rainfall event and
landslide incidents by statistical analysis of rainfall threshold in landslide initiation.
6 Analysis of Relationship Between Landslides and Rainfall … 67

The analysis was carried out on Kadwad landslide which occurred on October 2nd&
3rd 2009 in Karwar taluk of Uttara Kannada district, Karnataka. The current work
is based on the rainfall data and landslide inventory available for Kadwad landslide
event in Karwar.

6.2 Study Area

Karwar taluk in Uttara Kannada district of Karnataka forms a part of Western Ghats
where the range dips into the sea with peaks emerging as picturesque islands. The
geographical area is situated at 14° 50 2.12 N and 74° 10 44.50 E; 14° 50 34.25
N and 74° 10 35.24 E which lies in Survey of India (SOI) Toposheet no. 48 J/1
(Fig. 6.1). The area receives average annual precipitation of approximately 3000 mm.
On October 2nd, 2009 between 4:15 PM and 4:30 PM a massive landslide occurred
in Zariwad locality of Kadwad village (Fig. 6.3), east and south of Karwar town and
Kali River estuary respectively. The aggressive landslide that took place killed 19
people and devastated 9 houses. Also, high intensity precipitation of two consecutive
days resulted in subsidiary landslides in 21 locations all along the hills obstructing
traffic movement on NH-17. The massive slide in Kadwad village was characterized
as debris slide. At some locations, it consisted of debris flow, weathered materials
and lateritic boulders while at some places rock fall was also witnessed (Fig. 6.2).
As reported by Geological Survey of India, the landslide of Kadwad is considered
as deep seated of more than 20 m depth of failure, large rotational landslide which
had a fast rate of movement and having slope angle of approximately 30°.

Fig. 6.1 Location of the study area

68 G. Thejashree et al.

Fig. 6.2 Spatial distribution of rainfall in Uttara Kannada District

Fig. 6.3 a & b Landslide photographs in Kadwad Village

6.2.1 Geology and Geomorphology

Regionally, the area comprises rocks of Peninsular Gneissic Complex (PGC), older
granitoids and rocks belonging to the Chitradurga Group under the DharwarSuper-
group of Archaean age as well as dolerites of Palaeoproterozoic age intruding older
rocks. The rocks have undergone intensive chemical weathering resulting in the
formation of laterites, a common rock type in humid tropical environment. In the
study area, laterites form a thick mantle capped over underlying granite. Along with
granite, the area also possesses saprolite at places and lateritic soil with lithomargic
6 Analysis of Relationship Between Landslides and Rainfall … 69

clay is also present. Geomorphologically, the area is low-lying valley region affected
by percolation of groundwater. The area is recognized as paleo river channel that
flowed into the Kali River in the past which is in north of Zariwad of Kadwad village.
This old river channel is still linked to the Kali River and the level of groundwater
percolation in the channel fluctuates at regular intervals in respect to high and low
tides of the river. Konkan Railway track goes through this paleo-river valley connects
between Karwar town with Goa and other places. A number of lineaments are present
in the area with high density dendritic drainage pattern where higher stream orders
are structurally controlled (Hegde et al. 2014).

6.2.2 Climate and Rainfall

The area lies on a coastal strip known as the Monsoon Coast. It is surrounded by the
Western Ghats and Arabian Sea on east and west respectively. The area possesses
tropical climate with hot summers from March to May where temperature can reach
as high as 37 °C. Winter starts from December to February and has mild temperatures
of 24–32 °C. The monsoons are experienced in between June and October and receive
average annual precipitation of over 3000 mm. Most of the rainfall occurs during the
peak monsoon period from June to August .
The rainfall in the area includes both Southeast and Northeast monsoons. During
peak monsoon period, rainfall is received from southeast monsoon. Retreating
monsoon in the area accounts from October to December and has an average annual
rainfall of around 250 mm. In the year 2009, the amount of rainfall in retreating
monsoon was more when compared to other years. The spatial distribution of rainfall
for Uttara Kannada district is shown in Fig. 6.2.

6.3 Materials and Methodology

Rainfall and landslides are closely associated processes since rainfall is one of the
chief driving factors for landslide occurrences in Western Ghats. For present studies,
daily precipitation data was obtained from Karnataka State Natural Disaster Moni-
toring Centre/District Disaster Professional, Karwar for over ten years from 2001
to 2010 was used for statistical analysis. Reported landslide of October 2009 from
media and other sources corresponding to rainfall data of that year was analyzed and
calculated for rainfall threshold. Threshold is a level either maximum or minimum at
which a process tends to experience change. Two thresholdstypes have been defined
in the literature:
i. Minimum threshold: it is the minimum stage below which a process does not
experience change
70 G. Thejashree et al.

ii. Maximum threshold: it is the stage over which a process always experiences
change.
To study the interrelationship between rainfall and landslide occurrences threshold
can be derived from the amount of rainfall where if it is below the threshold landslide
does not occur and if it’s above the threshold landslide always occurs with respect
to other variable conditions of the area. Empirical models are used to define rainfall
threshold as suggested by many researchers (Crosta and Frattini 2001; Aleotti 2004;
Guzzetti et al. 2007). Threshold values for rainfall triggered landslides are marked
based on the rainfall measurements such as rainfall intensity (I), duration (D) and
rainfall event (E) as seen in many literatures. The ratio of the total rainfall received
for a given duration and the corresponding duration of time is defined as the Rainfall
Intensity and expressed in depth per unit time e.g. mm/hr or cm/hr. Caine (1980)
proposed intensity-duration threshold for landslide by taking 73 records globally
and derived an empirical form.

I = 14.82 D −0.39 (6.1)

In addition to I-D relationships, thresholds based on rainfall event (E) and Dura-
tion (D) have also been proposed by (Cannon and Ellen 1985) to predict rainfall
induced landslides. As proposed by Caine (1980) a global E-D threshold for shallow
landslides as:

E = 14.82 D 0.61 (6.2)

Antecedent rainfall measurements along with duration have also been combined
to derive thresholds for landslide initiation by many authors (Kim et al. 1991; Terlien
1998; De Vita 2000; Heyerdahl et al. 2003; Gabet et al. 2004; Cardinali et al. 2006).
A derivative index called E MAP was proposed considering the cumulative rainfall
for the event normalized by the Mean Annual Precipitation (MAP) (Sengupta et al.
2010) for local extent of area.

E
E MAP = (6.3)
MAP

6.4 Results and Discussion

Multiple landslides were reported for this area during 2nd and 3rd October 2009.
The local rainfall data for the years 2001–2010 was considered for this analysis.
Coastal talukas of the district receive comparatively higher rainfall (Fig. 6.2). The
mean annual precipitation for the area is around 3277 mm, almost 85% of which is
obtained during the southwest monsoon period (June to September), 5–6% during
the premonsoon period (January to May) and 8–9% during the northeast monsoon
6 Analysis of Relationship Between Landslides and Rainfall … 71

season (October to December). The rainfall records for the study area are available for
24 h intervals. The rainfall variation over the years 2001–2010 is shown in Fig. 6.4a
which clearly shows the concentration of rainfall over the monsoon season every
year.
A detailed plot for rainfall (mm) for the year 2009 is also presented in Fig. 6.4b
with major landslide events marked with red asterisks. Records indicate several
occurrences of high daily rainfall but the concentration of landslide occurrences
during October indicates a role of antecedent rainfall also. While there are at least
18 days where the daily rainfall exceeded 50 mm, only two of these resulted in
landslides. There are number of reported landslides during 2nd to 3rd October 2009
of which major slides were reported in Kadwad village locality. While the rainfall
on 3rd October was the peak rainfall for the year and can explain the landslides

Fig. 6.4 a Daily rainfall in Karwar District shown from January 2001 to December, 2010. b Daily
rainfall in Karwar District shown from 01st January 2009 to 31st December, 2009 with landslide
events marked with red asterisks
72 G. Thejashree et al.

observed, the rainfall for 2nd October, 2009 was not nearly high enough to account
for the landslides in the area. This indicates a significant relation between antecedent
rainfall and landslide initiation.
To investigate the relationship between antecedent rainfall and landslide initi-
ation, the total rainfall for all events (E) and the corresponding duration (D) was
derived from daily rainfall records and is presented in Table 6.1. All the extreme
rainfall events and durations from 2001 to 2010 along with the corresponding E MAP
are listed in Table 6.1. Cumulative rainfall (E) is defined as the total rainfall for an
uninterrupted series of rainy days. This measure i.e. E combined with the duration
of event provides useful insights into the mechanism of Landslides at Karwar and
rainfall intensity/duration. When the relationship between E and D is plotted as in
Fig. 6.5, it is clear that the rainfall events resulting in landslide occurrence involved
a minimum of 11 days of continuous rainfall. Also, the minimum threshold for
cumulative event rainfall (E) for landslide initiation is around 260 mm. On normal-
izing E with the Mean Annual Precipitation (MAP) of Karwar district i.e. 3277 mm, a
minimum E MAP of 0.08 is obtained. Therefore, for events having E MAP values greater
than or equal to 0.08 and having duration of more than 11 days, a landslide incidence
is very likely. The events on 10-07-2001 and 29-07-2001 are the only anomalies to
the above derived threshold. However, these can be explained due to the fact the
effect of human induced disturbances such as quarrying, construction etc. during the
early years of 2000s was not as severe as during the latter part of the decade.
Thus, this study indicates that for the coastal region of Western Ghats of Karnataka,
both peak daily rainfall as well as antecedent rainfall is important factors in land-
slide initiation. A minimum of 11 days of continuous rainfall with cumulative rain-
fall above 260 mm is necessary to trigger landslides and can be attributed to the
combination of saturation of soil caused by antecedent precipitation and short-term
high intensity rainfall events. This method provides an early warning signal to local
administrative authorities and can be used for advanced preparedness and relief
actions.

6.5 Conclusions

The study shows a strong interdependence between landslide triggering and rainfall
pattern. The landslide incidence is a direct function of the magnitude of the event
rainfall depth. Events having E MAP value above the threshold of 0.08 (above 260 mm)
and continuous rainfall duration of 11 days or more caused triggering of landslide
in Karwar that particular year. Thus, the rainfall events with intensity less than the
minimum threshold value did not cause triggering of landslide whereas landslides
occurred when the twin thresholds of rainfall and duration are crossed. There were
some rainfall events for which the landslide occurrence criteria was true but land-
slide did not occur. This shows that the slope was stable enough preventing slope
failure even though there was required amount of rainfall for landslide initiation. The
possible cause for the triggering of Kadwad landslide is a combination of saturation
6 Analysis of Relationship Between Landslides and Rainfall … 73

Table 6.1 Cumulative rainfall and their Durations along with corresponding E MAP
Date Duration of event, D Cumulative rainfall event, E (in mm) EMAP
(in hours) (in days)
Events with no landslide
27-05-2001 192.0 8 249.6 0.08
11-06-2001 120.0 5 157.3 0.05
24-06-2001 168.0 7 213.9 0.07
02-09-2001 384.0 16 177.2 0.05
15-10-2001 216.0 9 77.4 0.02
22-05-2002 72.0 3 127.2 0.04
07-07-2002 240.0 10 75.4 0.02
20-07-2002 288.0 12 223.2 0.07
30-07-2002 216.0 9 130.8 0.04
07-08-2002 144.0 6 104.9 0.03
11-08-2002 72.0 3 264.5 0.08
27-08-2002 288.0 12 135.8 0.04
15-10-2002 168.0 7 293.9 0.09
17-06-2003 192.0 8 358.7 0.11
27-08-2003 384.0 16 238.7 0.07
24-10-2003 96.0 4 47.1 0.01
18-05-2004 336.0 14 195.7 0.06
19-06-2004 168.0 7 75.7 0.02
17-07-2004 216.0 9 241.2 0.07
13-08-2004 48.0 2 171.9 0.05
05-10-2004 336.0 14 244.1 0.07
19-06-2005 312.0 13 220.1 0.07
05-07-2005 216.0 9 152 0.05
27-08-2005 312.0 13 145.7 0.04
27-09-2005 192.0 8 195.8 0.06
29-05-2006 168.0 7 36.5 0.01
28-07-2006 192.0 8 152.6 0.05
06-10-2006 120.0 5 182.5 0.06
04-06-2007 168.0 7 178.3 0.05
23-07-2007 240.0 10 145.4 0.04
16-08-2007 192.0 8 96.3 0.03
25-09-2007 264.0 11 171.7 0.05
05-10-2007 168.0 7 112.2 0.03
05-06-2008 456.0 19 141.8 0.04
(continued)
74 G. Thejashree et al.

Table 6.1 (continued)

Date Duration of event, D Cumulative rainfall event, E (in mm) EMAP
(in hours) (in days)
22-09-2008 264.0 11 152.6 0.05
27-05-2009 216.0 9 52.8 0.02
27-06-2009 192.0 8 165.6 0.05
26-07-2009 144.0 6 81.2 0.02
31-08-2009 216.0 9 174 0.05
03-06-2010 72.0 3 53.4 0.02
17-06-2010 144.0 6 204.2 0.06
09-10-2010 144.0 6 33.1 0.01
21-10-2010 168.0 7 99 0.03
Events with landslide
02-10-2009 264.0 11 280.15 0.09
03-10-2009 288.0 12 620.1 0.19
Anomalous events
10-07-2001 360.0 15 521.2 0.16
29-07-2001 312.0 13 306.8 0.09

Fig. 6.5 Cumulative event rainfall (E) versus event duration (D)

of soil caused by antecedent precipitation and short term high intensity rainfall along
with human interventions.

Acknowledgements The authors greatly acknowledge National Institute of Technology, Surathkal,

Karnataka for providing financial assistance and facilities for carrying out the research work.
6 Analysis of Relationship Between Landslides and Rainfall … 75

References

Aleotti P (2004) A warning system for rainfall-induced shallow failures. Eng GEol 73(3):247–265
Bai S, Wang J, Thiebes B, Cheng C, Yang Y (2014) Analysis of the relationship of landslide
occurrence with rainfall: a case study of Wudu County, China. Arab J Geosci 7(4):1277–1285
Brand EW, Premchitt J, Phillipson HB (1984) Relationship between rainfall and landslides in Hong
Kong. In: Proceedings of the 4th international symposium on landslides. BiTech Publishers,
Vancouver, vol 1, pp 377–384
Caine N (1980) The rainfall intensity-duration control of shallow landslides and debris flows.
Geografiskaannaler: Ser A, Phys Geogra 62(1–2):23–27
Cannon SH, Ellen SD (1985) Rainfall conditions for abundant debris avalanches in the San Francisco
Bay region, California. Calif Geol 38(12):267–272
Cardinali M, Galli M, Guzzetti F, Ardizzone F, Reichenbach P, Bartoccini P (2006) Rainfall induced
landslides in December 2004 in south-western Umbria, central Italy: types, extent, damage and
risk assessment. Nat Hazards Earth Syst Sci 6(2):237–260
Claessens L, Schoorl JM, Veldkamp A (2007) Modelling the location of shallow landslides and their
effects on landscape dynamics in large watersheds: an application for Northern New Zealand.
Geomorphology 87(1–2):16–27
Crosta GB, Frattini P (2001) Rainfall thresholds for triggering soil slips and debris flow. In: Proceed-
ings of the 2nd EGS Plinius conference on mediterranean storms: Publication CNR GNDCI, vol
2547, pp 463–487
De Vita P (2000) Phenomena of instability of pyroclastic coverings of the Monti Lattari, Sarno
and Salerno (Campania) and analysis of the pluviometric determinants. Quaderni Appl Geol
7(2):213–235
Gabet EJ, Burbank DW, Putkonen JK, Pratt-Sitaula BA, Ojha T (2004) Rainfall thresholds for
landsliding in the Himalayas of Nepal. Geomorphology 63(3–4):131–143
Glade T, Crozier M, Smith P (2000) Applying probability determination to refine landslide-
triggering rainfall thresholds using an empirical Antecedent Daily Rainfall Model. Pure Appl
Geophys 157(6–8):1059–1079
Govi M, Mortara G, Sorzana PF (1985) Hydrological events and landslides. GeologiaApplicata e
Idrogeologia 20:359–375
Guzzetti F, Peruccacci S, Rossi M, Stark CP (2007) Rainfall thresholds for the initiation of landslides
in central and southern Europe. Meteorol Atmos Phys 98(3–4):239–267
Hegde VS, Krishnaprasad PA, Shalini R, Nilamwar D, Tejaswini B, Girish KH, Malewadi CS (2014)
Land-slide hazards of October 2009 at Karwar, Karnataka: A lesson for planning developmental
activities in the tropical Ghat regions. Int J Earth Sci Eng 7:260–268
Heyerdahl H, Harbitz CB, Domaas U, Sandersen F, Tronstad K, Nowacki F, Engen A, Kjekstad O,
Devoli G, Buezo SG, Diaz MR (2003) Rainfall induced lahars in volcanic debris in Nicaragua
and El Salvador: practical mitigation. In: Proceedings of international conference on fast slope
movements—prediction and prevention for risk mitigation, IC-FSM2003. Patron Pub, Naples,
pp 275–282
Jakob M, Weatherly H (2003) A hydroclimatic threshold for landslide initiation on the North Shore
Mountains of Vancouver, British Columbia. Geomorphology 54(3–4):137–156
Kim SK, Hong WP, Kim YM (1991) Prediction of rainfall-triggered landslides in Korea. In: Bell
DH (ed) Landslides, 2nd edn. A.A. Balkema, Rotterdam, pp 989–994
Lan H, Zhou CH, Lee CF, Wang S, Wu FQ (2003) “Rainfall-induced landslide stability analysis
in response to transient pore pressure-A case study of natural terrain landslide in Hong Kong” .
Science in China Ser E Technol Sci 46:52–68
Peruccacci S, Brunetti MT, Gariano SL, Melillo M, Rossi M, Guzzetti F (2017) Rainfall thresholds
for possible landslide occurrence in Italy. Geomorphology 290:39–57
Rosi A, Segoni S, Catani F, Casagli N (2012) Statistical and environmental analyses for the definition
of a regional rainfall threshold system for landslide triggering in Tuscany (Italy). J Geog Sci
22(4):617–629
76 G. Thejashree et al.

Sengupta A, Gupta S, Anbarasu K (2010) Rainfall thresholds for the initiation of landslide at
LantaKhola in north Sikkim, India. Nat Hazards 52(1):31–42
Terlien MT (1998) The determination of statistical and deterministic hydrological landslide-
triggering thresholds. Environ Geol 35(2–3):124–130
Tsai TL (2008) The influence of rainstorm pattern on shallow landslide. Environ Geol 53(7):1563–
1569
Chapter 7
Optimal Cropping Pattern of Kulsi River
Basin, Assam, India Using Simulation
and Linear Programming Model

Jyotismita Taye, Bibhash Sarma, and Abhijit D. Lade

Abstract Irrigation potential of an area can be benefitted if we thrust on maximizing

crop production by suitable allocation of limited resources to get optimal productivity.
The Kulsi River Basin, located in the Southern bank of the Brahmaputra River, is
the area of study. The gross command area and net irrigated area in the basin are
37,908 Ha and 23,882 Ha, respectively. The river basin is deficit of regular irrigation
facilities, and only traditional cropping pattern (rain-fed conditions only) has been
followed by the cultivators of the area. Hence, to promote the areas potential, the Kulsi
Multipurpose Project Authority of Brahmaputra Board has proposed a multipurpose
reservoir for the generation of hydropower, flood control, and irrigation. The capacity
of the reservoir is 525.64 MCM, plant capacity is 55 MW, tail water level is 63 m,
and dead storage of the reservoir is 85.86 MCM. Keeping in mind the benefits to be
received by the proposed reservoir, the objective of the paper is mainly emphasized to
compute the crop water requirement of the Kulsi Basin and to formulate a simulation
and linear programming (LP) model to find the best possible cropping arrangement
in order to maximize the net benefit of the crops.

Keywords Irrigation potential · Optimal productivity · Crop water requirement ·

Cropping pattern · Simulation and linear programming model

7.1 Introduction

With the increasing demand due to residents’ growth and rapid industrialization,
water is becoming valued and scarce. The rising demand and scarcity of water makes
it important to use the available water in the most economical way. Crops grow with
an adequate amount of water. Proper planning of crops to be grown in an area can be

J. Taye (B) · A. D. Lade

Indian Institute of Technology, Guwahati, India
e-mail: taye176104022@[Link]
B. Sarma
Assam Engineering College, Guwahati, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 77

R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
78 J. Taye et al.

assured with sufficient amount of water present there. Higher crop productivity can
only be attained by proper soil, water, and crop management in the areas with assured
irrigation. The key to effective irrigation water management lays in proper estimation
of crop water requirement, which are primarily based on cropping pattern, rainfall
in the area, and other climatology factors. The objective of the paper is emphasized
to study the water requirements of different crops and to obtain the most favorable
cropping pattern in order to maximize the net profit of crops using simulation and
LP model.
Planning of optimal cropping pattern needs evaluation of crop water require-
ments (CWR) of individual crops. Each crop has different water requirements.
Food and Agriculture Organization of the United Nations, FAO (Doorenbos 1977)
has given procedures to calculate crop water requirements of crops by a software
called CROPWAT. It is a tool to carry out standard calculations for reference evapo-
transpiration, crop water requirements and crop irrigation requirements, and more
specifically the design and management of irrigation schemes (Karamouz et al. 2010).
CROPWAT has been used by various researchers to evaluate the water requirements
of crops. Kuo and Liu (2002) presented a CROPWAT model where they evaluated
the water requirements of crops in Taiwan. Similar work has been carried out by Kia
(2013) in Iraq, Saravanan and Saravanan (2014) in Tamil Nadu, India, and Surendran
et al. (2015) in Palakkad district, Kerala, India. The Kulsi River having the potential
of supplying irrigation water needs proper planning for the distribution of water.
Sarma et al. (2013) have shown how optimization and simulation can be combined
for sizing Kulsi reservoir system. Optimal cropping pattern and production of food
crops with maximum profit are important information for irrigation planning which
is achieved by optimization methods. In this study, linear programming optimiza-
tion model is used. Many researchers have carried out various works with the use
of linear programming. Singh et al. (2001), Vedula (1992), Raman et al. (1992),
Mainuddin et al. (1997), Sarker et al. (1997), Kangrang et al. (2010) made use of
linear programming technique for optimum utilization of resources (land and water)
considering various resource and requirement constraint. The objective of the paper
includes:
• To compute crop water requirements of Kulsi Basin
• To formulate a simulation and linear programming model to find
(a) The power potential of the proposed Kulsi Multipurpose Project
b) The irrigation potential of the proposed Kulsi Multipurpose Project
(c) The optimal cropping pattern for maximization of net benefit
(d) The area under different crops for maximization of net benefit.
7 Optimal Cropping Pattern of Kulsi River Basin, Assam, India … 79

Fig. 7.1 Study map of the Kulsi River Basin

7.2 The Study Area

The Kulsi River Basin is a tributary (south bank) of the river Brahmaputra situated
in Kulsi-Deosila sub-basin. The river is formed mainly by three tributaries, Khri,
Krishniya, and Umsiri. They all emerge from west Khasi Hills ranges (Fig. 7.1).
It covers an overall area of 3770 km2 within Assam and Meghalaya. Out of the
total catchment areas, 685 km2 is plain catchment in Assam and 3085 km2 is hill
catchment in Meghalaya and Assam. The total length of Kulsi is about 196 km from
its source to its outfall.

7.3 Methods and Materials

7.3.1 Crop Water Requirement

The depth of water consumptively used by a crop (plus unavoidable irrigation appli-
cation losses) may be defined as crop water requirement. The Food and Agricul-
ture Organization of the United Nations (FAO) has given procedures to calculate
crop water requirements of crops by a software called CROPWAT. The nature and
stages of growth of the crops including the environmental conditions influence the
water requirement of crops. Figure 7.2 shows the change in crop coefficient during
different growth stages of crop. Higher temperatures will increase transpiration from
vegetation, boosting water requirements for both rain-fed and irrigated crops.
80 J. Taye et al.

Fig. 7.2 Schematic representation of crop coefficient during different growth stages of crop
(Courtesy: Pokorny 2018)

In this study, computation of crop water requirement has been carried out in
CROPWAT 8.0 by calling up successively the suitable climatic and rainfall data,
together with the crop data and corresponding planting dates of crops. For the
calculation of crop water requirement, the model requires the following:
• Reference crop evapo-transpiration (ETo ): The ETo of Guwahati has been calcu-
lated by the FAO Penman–Monteith method in CROPWAT 8.0. For the calcula-
tion of ETo , the FAO Penman–Monteith method needs the records of radiation,
air temperature, air humidity, wind speed, and sunshine hours of the area selected.
• Rainfall data: The rainfall data have been collected from the Meteorological center
at Gopinath Bordoloi International Airport site maintained by IMD for the year
1951 to 2000 to find out the effective rainfall. Effective rainfall is the total amount
of rainwater useful for meeting the water need of the crops.
• Crop data: This part of calculation in CROPWAT requires the value of crop coeffi-
cient (K c ) Table (7.1), growth period of crops, planting date, rooting depth, yield
response factor, critical depletion of crop, and crop height.
In the absence of regular irrigation facilities in the area, only traditional cropping
pattern has been followed till now by the cultivators of the area and the crops are
grown in rain-fed condition only. The existing cropping intensity in the area is found
to be 146.56%.

7.3.2 The Simulation Model

Simulation models are considered as a valuable tool for understanding the process
of a system including the operators’ knowledge and decision (Khadr and Elshemy
2017). In this paper, in order to observe the behavior of the proposed reservoir by
7 Optimal Cropping Pattern of Kulsi River Basin, Assam, India … 81

Table 7.1 Crop coefficient (K c ) of some crops during different stages of crop growth
Crops Initial stage Crop development stage Mid-season stage Late season stage
Rice 1.05 1.2 0.8 1
Jute 0.4 1.15 0.5 2
Onion 0.5 0.75 1.05 0.85
Potato 0.45 0.75 1.15 0.85
Pea 0.45 0.8 1.15 1.05
Radish 0.45 0.6 0.9 0.9
Maize 0.4 0.8 1.15 0.7
Sugarcane 0.3 1.2 0.5 2
Lentil/Pulses 0.45 0.75 1.1 0.65

Kulsi Multipurpose Project, a simulation model is prepared. With 40 years of input

data, the simulation model was run. The simulation model is assumed to have an
initial storage of the reservoir as equal to dead storage. May is considered as the start
of each year. The best suited equations for elevation versus capacity (Fig. 7.3) and
area versus capacity (Fig. 7.4) are obtained.
Collection of data and processing are essential steps to formulate the model. The
following are the hydrological data of the River Kulsi collected from Brahmaputra
Board to formulate the simulation model:
• Inflow data at monthly time step (Table 7.2).
• Mean monthly evaporation data.
• Area–capacity–elevation characteristics of the reservoir.
• Irrigation demand.

Fig. 7.3 The elevation

versus capacity curve with
best suited equation
82 J. Taye et al.

Fig. 7.4 The area versus

capacity curve with best
suited equation

Table 7.2 Average inflow

Months Average inflow Months Average inflow
(MCM) at each month of the
(MCM) (MCM)
year
May 47.46 November 52.3
June 94.87 December 36.53
July 150.48 January 25.57
August 132.85 February 21.56
September 117.43 March 18.81
October 93.97 April 29.77

The elevation–area–capacity values have been worked out from reservoir area
maps developed by the Kulsi Multipurpose Project Authority. Original elevation–
area–capacity values have been considered for the purpose of the study.
The full reservoir level (FRL) for the project has been fixed at El 115.00 m. The
elevation–capacity and area-capacity curves are generated using the data given in
Table 7.3.
The irrigation demand considered for the purpose of the study is taken from the
computation of crop water requirement. The irrigation demand of all the months
is shown in Fig. 7.5.
As proposed by the Brahmaputra Board, the reservoir capacity is taken as 525.64
MCM, plant capacity as 55 MW, and tail water level as 63 m. Initially, firm power
is assumed as 0.5 MW. From the data available and using it with the help of the
simulation model, the power and irrigation potential are obtained by satisfying the
minimum target reliability.
7 Optimal Cropping Pattern of Kulsi River Basin, Assam, India … 83

Table 7.3 The elevation–area–capacity values of the reservoir of the Kulsi Multipurpose Project
Elevation (m) Area (ha) Capacity (ha m)
58 0 0
60 18 18
70 471 2483
80 660 8118
90 899 15,913
100 1323 27,023
110 1813 42,703
115 2131.5 52,564.25
120 2450 73,879.25

Fig. 7.5 Monthly

distribution of irrigation
demand (MCM)

7.3.3 Linear Programming Model

An important factor of agricultural benefit is its planning of cropping pattern. Land,

water, labor, and capital are some of the resources on which cropping pattern depends
(Sarker and Quaddus 2002). Optimal cropping pattern and production of food crops
with maximum profit is important information for irrigation planning using optimiza-
tion methods. In this study, the LP model prepared using Lingo is used to formulate
the optimal cropping pattern in the basin. The model is prepared with the aim of
maximization of net benefit of the crops in a given area. The objective function in
the model is formulated to determine the cropping pattern (optimal) in the area of
study that will maximize the net benefit of the crops. The decision variable in the
84 J. Taye et al.

formulated LP model is the area under the crops (X k ). The constraints that are consid-
ered are area constraint, water constraint, and affinity constraint. The prepared LP
model is run once by considering the affinity constraint and once without considering
affinity constraint.
Objective Function:
• Maximization of net benefit
Maximize Z
Subject to


NC
Z= bk Ck X k
k−1

where k denotes particular crop from number 1 to NC.

NC Number of crops
bk Net benefit of crop k per ton
Ck Crop yield in ton per hectare of crop k
Xk Area under crop k in hectares

Constraints:
The resource availability constraints which the model should satisfy while achieving
the desired objectives are:
• Land availability constraints
The total area assigned to different crops in a particular season cannot surpass the
culturable command area (CCA) of the area. Thus, the area assigned to Kharif
and perennial crops should not exceed the CCA in summer and the area assigned
to Rabi and perennial crops also should not exceed CCA in winter. This can be
expressed as:

X k ≤ CCA
k

where X k = Area under crop k in hectares

CCA = Culturable command area of the region
• Water availability constraints
The water requirement of the crops should be fully met during the entire crop
period. It should not exceed the available water. The total irrigation demand of
a crop k for a given growth period of the crop is taken as the water available for
crop. It is expressed as:
7 Optimal Cropping Pattern of Kulsi River Basin, Assam, India … 85


NC
Wk,t X k ≤ Ot
k=1

where W k,t = Water depth of crop (crop water requirement) k at time t

Ot = Total irrigation demand of crop k at time t
• Affinity constraint
Affinity constraint is considered to bind the area under the crops within their upper
limit or lower limit. It is expressed as:

LLk ≤ X k ≤ ULk

where LLk = Lower limit of area for crop k in hectares

ULk = Upper limit of area for crop k in hectares

7.4 Results and Discussions

(1) Crop water requirement (CWR)

The study of crop water requirement is carried out by CROPWAT 8.0. Before the
calculation of crop water requirement of each crop, it is necessary to calculate the
reference evapo-transpiration (ETo ) and the effective rainfall of the area. The evapo-
transpiration rate from a reference surface (a hypothetical grass reference crop with
an assumed crop height), not short of water, is called the reference crop evapo-
transpiration or reference evapo-transpiration and is denoted as ETo (Allen et al.
1998). The FAO Penman–Monteith method is suggested as the only method for
determining ETo . Table 7.4 shows the ETo measured using CROPWAT.
Effective rainfall is the total amount of rainwater useful for meeting the water need
of the crops. The rainfall data have been collected from the Meteorological center
at Gopinath Bordoloi International Airport site maintained by IMD for the year
1951 to 2000, which has been considered for the study of the basin. For the study
of crop water requirement, the effective rainfall is calculated by CROPWAT. The
FAO CROPWAT 8.0 uses the USDA Soil Conservation Service method by default to
calculate the effective rainfall. The only input here is to provide the average rainfall
data of every month in millimeters. Table 7.5 shows the effective rainfall measured.
The crop data and soil data are entered into the CROPWAT to find the crop water
requirements of each crop. The soil data and crop data are different for different
crops, and it is to be found locally. The result shows that the CWR for onion (306.9
mm/year) and garlic (306.9 mm/year) is the maximum in the Rabi season. Cauliflower
and radish require less water than the other crops. The CWR for sugarcane, which
is an annual crop, is 352.8 mm/year. The maximum CWR is for summer paddy, i.e.,
local boro (432.4 mm/year) and minimum for radish (5.9 mm/year). The CWR is
86 J. Taye et al.

Table 7.4 Reference crop evapo-transpiration (ETo ) per month

Month Min temp Max temp Humidity Wind speed Sunshine Radiation ETo
°C °C % km/day Hours MJ/m2 /day mm/day
January 10.2 23.6 89 1 10.4 17.5 2.17
February 11.9 26.2 83 1 11.3 21.1 2.89
March 15.8 30 64 3 11.6 24.4 3.78
April 20 31.2 72 3 12.4 27.7 4.99
May 22.6 31.1 77 2 13.2 29.8 5.71
June 25 31.7 82 2 13.4 30.1 6.13
July 25.5 31.8 86 2 13.4 30 6.22
August 25.5 32.1 83 1 13 28.7 5.9
September 24.6 31.6 86 1 12.2 25.8 5.2
October 21.9 30.2 86 1 11.3 21.8 4.06
November 16.6 27.6 85 1 10.5 18.1 2.85
December 11.7 24.6 90 1 10.3 16.6 2.19
Average 19.3 29.3 82 2 11.9 24.3 4.34

Table 7.5 Effective

Months Rain (mm) Effective rainfall (mm)
rainfall (mm) at each month
of the year January 11.8 11.6
February 17.2 16.7
March 55.1 50.2
April 147.1 112.5
May 248.9 149.8
June 316.8 156.7
July 351.2 160.1
August 269.4 151.9
September 186.9 131
October 90.9 77.7
November 17.8 17.3
December 6.8 6.7
Total 1719.9 1042.3

maximum in the November month (1089.2 mm) and minimum in the month of June
(36.8 mm).
7 Optimal Cropping Pattern of Kulsi River Basin, Assam, India … 87

(2) Simulation model

The application of simulation model is started with the assumption that the initial
storage in the reservoir is taken equal to the dead storage 85.86 MCM. After devel-
oping and solving the model, the power potential and irrigation potential of the
reservoir are obtained. The result shows that the project has a potential of supplying
934 MCM of water annually for irrigation needs. Considering the annual irrigation
demand as 934 MCM, it was found that a firm power of 0.7 MW can be produced
with the target reliability, i.e., irrigation reliability of 75.05% and power reliability
of 98.93%.
(3) Linear programming model

The LP model is prepared using Lingo. The model is prepared with the objective of
maximization of net benefit. The model is run once considering affinity constraint
and once without affinity constraint and the results are observed for each run. When
considering affinity constraints, the model takes 10% of the area of proposed cropping
pattern as minimum area for maximization of net benefit. The result shows that
for maximization of net benefit without considering affinity constraint, the linear
programming model provides a net benefit of |19,105.6 lakhs considering only one
crop (sugarcane), while considering affinity constraint the model provides a net
benefit of | 19,624.7465 lakhs (Table 7.6), where sugarcane takes the maximum area
among the other crops. Moreover, an optimal cropping pattern with variety of crop
distribution within the area is proposed with the presence of affinity constraint.

7.5 Conclusion

In this paper, an attempt has been made to plan an optimal cropping pattern for
Kulsi River Basin in Assam, India. The first study carried out was to find out the
crop water requirements of the existing crops in the basin. The second study was to
formulate a simulation model to study the behavior of the reservoir as proposed by
the Kulsi Multipurpose Project Authority. The third study was to find out the optimal
cropping pattern for maximization of net benefit using linear programming model.
Crop water requirement is an important parameter to study as it allows us to know
how effective the crop root zone is in storing water. Maximizing the net benefit of
the crops will benefit the cultivators in choosing the appropriate crops. This optimal
cropping pattern may serve as a guideline to adopt a new cropping pattern in this
area or any other similar area nearby.
88 J. Taye et al.

Table 7.6 Crop area distribution to maximize net benefit in presence of affinity constraint
Crop area distribution to maximize net benefit (Presence of affinity constraint)
Crops Area under Yield (MT/ha) Total yield in Rate per MT Benefit (| in
crops for MT lakh)
maximizing net
benefit (ha)
Summer 70 4 280 8000 22.4
paddy local
paddy
Summer 24 2.5 60 8000 4.8
paddy local
early ahu
Autumn paddy 35 2.4 84 7000 5.88
local regular
ahu
Bottle gourd 10 15 150 8000 12
Jute 229.3 1.5 343.95 15,000 51.5925
Winter paddy 380 3 1140 7500 85.5
local
Winter paddy 1200 4 4800 7000 336
HYV Ranjit
Bahadur
Black and 157.5 0.6 94.5 30,000 28.35
green gram
Sesame 109.2 0.8 87.36 40,000 34.944
Potato 1964.5 8 15,716 10,000 1571.6
Maize 85.7 3 257.1 10,000 25.71
Onion 71.9 8 575.2 5000 28.76
Garlic 92.9 3 278.7 30,000 83.61
Sugarcane 21,667 40 866,680 2000 17,333.6
Total benefit 19,624.7465

References

Allen RG, Pereira LS, Raes D, Smith M (1998) Crop evapotranspiration-Guidelines for computing
crop water requirements—FAO Irrigation and drainage paper 56. FAO, Rome 300(9):D05109
Doorenbos J (1977) Guidelines for predicting crop water requirements. FAO Irrigation and Drainage
Paper, vol 24, pp 1–179
Kangrang A, Compliew S (2010) An application of linear programming model for planning dry-
seasonal Irrigation system. Trends Appl Sci Res 5(1):64–70
Karamouz M, Semsaryazdi M, Ahmadi B, Ahmadi A (2010) Climate change impacts on crop
water requirements: a case study. In: 1st IWA Malaysia young water professionals conference,
international water association, pp 1–11
Khadr M, Elshemy M (2017) Data-driven modeling for water quality prediction case study: the
drains system associated with Manzala Lake, Egypt Ain Shams Eng J 8(4):549–557
7 Optimal Cropping Pattern of Kulsi River Basin, Assam, India … 89

Kia DR (2013) Water requirements for major crops in different agro-climatic zones of Iraqi Kurdistan
using by CROPWAT 8.0. IOSR J Agric Veter Sci 6(3):30–36
Kuo SF, Liu CW (2002) Simulation and optimization model for irrigation planning and management.
Hydrolog Proc 17(15):3141–3159
Mainuddin M, Gupta AD, Onta PR (1997) Optimal crop planning model for an existing groundwater
irrigation project in Thailand. Agric Water Manage 33(1):43–62
Pokorny J (2018) Evapotranspiration. Encyclopedia of Ecology. Ref Module Earth Syst Environ
Sci 2:292–303
Raman H, Mohan S, Rangacharya NCV (1992) Decision support system for crop planning during
droughts. J Irrigat Drainage Eng 118(2):229–241
Saravanan K, Saravanan R (2014) Determination of water requirements of main crops in the tank
irrigation command area using CROPWAT 8.0. Int J Interdisc Multidiscipl Stud 1(5):266–272
Sarker RA, Quaddus MA (2002) Modeling a nationwide crop planning problem using a multiple
criteria decision making tool. Comput Ind Eng 42(2–4):541–553
Sarker RA, Talukdar S, Haque AFMA (1997) Determination of optimum crop mix for crop
cultivation in Bangladesh. Appl Math Model 21(10):621–632
Sarma B, Das KK, Patwari NN (2013) Sizing of Kulsi reservoir system using non-linear optimization
and simulation. In: Proceedings of Assam international water conference - 2013, 21–22 February,
2013, Guwahati
Singh DK, Jaiswal CS, Reddy KS, Singh RM, Bhandarkar DM (2001) Optimal cropping pattern in
a canal command area. Agric Water Manage 50(1):1–8
Surendran U, Sushanth CM, Mammen G, Joseph EJ (2015) Modelling the crop water require-
ment using FAO-CROPWAT and assessment of water resources for sustainable water resource
management: a case study in Palakkad district of humid tropical Kerala, India. Aquatic Procedia,
4, 1211–1219
Vedula S, Mujumdar PP (1992) Optimal reservoir operation for irrigation of multiple crops. Water
Res Res 28(1):1–9
Chapter 8
Comparison of Flux Footprint Models
to a Mixed Fetch Heterogeneous
Cropland System

Shweta Kumari and K. B. V. N. Phanindra

Abstract Flux footprint models delineate the source area of micrometeorological

fluxes as measured by the eddy covariance (EC) system. An accurate knowledge of
flux footprint is crucial for effective interpretation of measured fluxes and upscaling
to a regional scale with the help of satellite data. Flux footprint models are derived
based on the assumption that “measured fluxes are originated from a homogeneous
cropland system.” For small and fragmented land-holdings of the Indian agricul-
tural system, this assumption is often violated, hence questioning the applicability
of existing models. The objective of this paper is to evaluate the performance of
various footprint models to a mixed fetch (heterogeneous) cropland system. Field-
scale experiments were conducted in continuous cotton (C3) sugarcane (C4) plots
located in Nandikandi village, Telangana, India. Two low height EC towers capturing
homogeneous fluxes from individual fields and one tall EC tower capturing hetero-
geneous fluxes from the combinations were used in the study. Half hourly fluxes are
calculated as the covariance between vertical wind velocity and scalar concentra-
tion of interest (water vapor, air temperature, or carbon dioxide). Three analytical
models (viz. Hsieh, Schuepp, and Kormann and Meixner) were applied along with
the Lagrangian particle dispersion model, considering unstable (daytime) and stable
(night time) flux measurements. Our results conclude that, for homogeneous crop-
lands, the Schuepp model performs best during unstable atmospheric condition and
Hsieh model performs best in stable atmospheric conditions. Hseih model is suitable
in representing the source area fluxes in a mixed fetch condition for all atmospheric
stability conditions.

Keywords Flux footprint · Eddy covariance · Analytical model · Lagrangian

stochastic model · Cropland

S. Kumari (B) · K. B. V. N. Phanindra

Department of Civil Engineering, Indian Institute of Technology Hyderabad, Hyderabad 502284,
India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 91

R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
92 S. Kumari and K. B. V. N. Phanindra

8.1 Introduction

Flux footprint represents the upwind region from where the atmospheric fluxes
measured by an instrument are generated. Several atmospheric (atmospheric stability,
surface roughness, and measurement height), surface (homogenous, heterogeneous),
and phenological properties can influence the shape and extent of footprint (Kolari
et al. 2004). Eddy covariance (EC) methodology allows for the quantification of
the fluxes of interest (water vapor, air temperature, carbon dioxide, and methane) by
co-variating with vertical wind speed, at high frequency. A footprint model is a mathe-
matical representation of the spatial distribution of flux sources and the corresponding
magnitude with respect to the measurement location (Arriga et al. 2017). Footprint
models help in visualizing the spatial distribution of source area fluxes as measured
by the EC flux towers. Methods for modeling flux footprint include: (a) analyt-
ical models (computationally least expensive, solves advection–diffusion equation,
and simple to use), (b) Lagrangian particle models (accounts for three-dimensional
turbulent diffusion and non-Gaussian inhomogeneous turbulence), and (c) large-eddy
simulations (address spatial heterogeneity and non-topography explicitly).
Analytical models consider atmospheric and crop phenological characteristics to
develop flux footprint and are largely constrained by a number of assumptions. These
include (a) horizontal homogeneity, (b) stationarity, and (c) zero mean vertical wind
speed during the averaging period. These assumptions are a major limitations of the
Eddy covariance system application to real world fields and question the accuracy
of the predicted flux footprint.
One of the primary assumptions, on which the analytical models were built is that
the contributing source area to the flux measurements has homogeneous cropland
and surface roughness characteristics. This assumption can be violated under two
conditions: (i) for small and fragmented croplands, wherein multiple crop fields exist
within the fetch of a tower, and (ii) stable atmospheric conditions with calm wind,
where fetch is considerably large. For heterogeneous surfaces (that are commonly
seen in the Indian agricultural setting), there is a dependency of measured signal on
how strongly a certain part of the surface influences the sensor, and thus the location
and size of its footprint do affect the measured signal.
Applicability of flux footprint models to either homogeneous or heterogeneous
croplands of the Indian sub-continent was never explored in literature. This could be
due to the non-availability of flux datasets for model validation. The objective of this
work is to evaluate the performance of analytical models applied to homogeneous and
heterogeneous cropland systems considering daytime unstable and nighttime stable
atmospheric conditions. Data from three EC towers with two capturing homogeneous
cotton (C3) and sugarcane (C4) fluxes, and one capturing their combined effect were
used to evaluate the footprint models.
8 Comparison of Flux Footprint Models to a Mixed Fetch Heterogeneous … 93

8.2 Study Area and Data Analysis

Field-scale experiments were performed in privately owned agricultural fields located

at Nandi Kandi Village, Telangana, India. The annual average precipitation of the
study area is 795 mm with the month July receiving the most precipitation of
170.2 mm and January receiving the least precipitation of 5.1 mm. The study area has
the BSh (Köppen-Geiger climate classification) and receives an average of 48.8 days
of precipitation, with the most precipitation occurring in July with 10.5 days and the
least precipitation occurring in January with 0.4 days. The annual average tempera-
ture of the study area is 26.4 °C with May (32.6 °C) as the warmest and December as
the coolest month (21.3 °C). Two contiguous fields comprising of cotton (C3 plant)
and sugarcane (C4 plant) were selected for the experiments. The cotton field has an
area of 3 acres with a canopy height of 1.0 m and was planted in June 2018. Sugarcane
has an area of 2.5 acres with a canopy height of 2.2 m and was planted in January
2018. Both the crops are drip irrigated at an average frequency of 7 days. Experiments
were performed using three EC flux system is composed of a 3D sonic anemometer
and an open-path fast response infrared gas analyzer (IRGASON-EB-IC, Campbell
Sci. Inc., USA) to measure CO2 and H2 O fluxes. Raw data were collected with a
logger (CR1000, Campbell Sci. Inc., USA) at 10 Hz frequency and averaged over
half-hour intervals for use with computations. Additionally, slow response meteoro-
logical variables including precipitation (TE525-L-PTL, Tipping Bucket, Campbell
Sci. Inc., USA), soil heat flux (HFP01SC-L-PT-L, Campbell Sci. Inc., USA), solar
radiation (CNR 4, Campbell Sci. Inc., USA), soil moisture (CS616-L-PT-L, Camp-
bell Sci. Inc., USA), and photosynthetically active radiation (LI-190R, LI-COR Inc.,
USA) were obtained at the 30-min interval. Two EC flux towers are placed at the
center of C3 and C4 croplands at heights of 3 m and 4 m, respectively, and a third
EC flux tower is placed at a height of 7 m to capture their combined effect (Fig. 8.1).
Flux data was collected during the period 08–12 October 2018.
Primary data processing was performed on half-hour mean fluxes using
Eddypro post-processing software (Version 6.2.0, LI-COR, USA). A number of stan-
dard flux corrections including (a) removal of bad data, (b) tilt corrections on sonic
measurements, (c) frequency response corrections, and (d) Webb–Pearman–Leuing
(WPL) corrections were applied on fast response measurements. The corrected half-
hour fluxes have spiked due to unrealistic meteorological measurements, leading to
the application of secondary corrections to remove spurious data and fill the gaps
with reasonable estimates.
We used the REddyproc package developed in an open-source ‘R’ environment
along with MATLAB script to perform secondary data processing that includes: (a)
flux spike removal, (b) removal of negative nighttime CO2 fluxes, (c) friction velocity
correction, and (d) gap filling and uncertainty analysis. Windrose diagram (Fig. 8.2)
of the study region observed during the data-period confirms that a westerly (270°
from North) flowing wind predominates in the study area, and hence, all the sensors
were placed accordingly, to capture the maximum possible fluxes.
94 S. Kumari and K. B. V. N. Phanindra

Fig. 8.1 a [Left] Distribution of agricultural farms within the study area, and position of three flux
towers used in the analysis. Towers ‘1’ and ‘2’ (3 and 4 m height) capture homogeneous fluxes from
cotton and sugarcane, respectively, whereas tower ‘3’ (7 m height) captures the combined fluxes from
the two fields. [Right] Typical layout and footprint from contrasting in a mixed fetch environment.
b Eddy covariance flux tower installed in (col: 1) homogeneous cotton field (measurement height of
3 m), (col: 2) homogeneous sugarcane field (measurement height of 4 m), and (col: 3) heterogeneous
sugarcane-cotton fields (measurement height of 7 m) in the Nandi-Kandi village

8.3 Methodology

The 10 Hz frequency diurnal flux data from the three EC flux systems were processed
and averaged for half-hourly interval. After the quality checks (Foken et al. (1996))
only, the good data were considered for the flux footprint estimation. Flux foot-
prints for both unstable and stable atmospheric conditions were estimated using
three analytical models (Schuepp model, Korman and Meixner model, and Hsieh
model) and the experimental data collected. The flux footprint predicted by the
analytical models were then compared with the results of the flux footprint estimated
8 Comparison of Flux Footprint Models to a Mixed Fetch Heterogeneous … 95

Fig. 8.2 Windrose diagram showing predominant wind direction within the study area during the
monitoring period

by Lagrangian particle distribution model. The models used to predict flux footprint
include:
(1) Lagrangian (Kljun) model considers release of a point source, and tracts the
trajectories downwind of the source toward the measurement location, with
footprint function given by
 b  
∗ −c
F̂ y∗
= a X̂ − d exp
X̂ ∗ − d


y∗
where a, b, c, and dare fitting parameters, F is the crosswind-integrated footprint
function, and X ∗ is along with wind distance.
(2) Schuepp model estimates weighing factors of footprint function for scalar flux
and concentration as an analytical solution of advection–diffusion equation,
given by
 
U zm U zm
f (x) = exp
ku ∗ x 2 ku ∗ x
96 S. Kumari and K. B. V. N. Phanindra

where U is mean horizontal wind speed, zm is measuring height, u* is frictional

velocity, k is von Karman constant, and x is the horizontal upwind distance of sources
from the measurement system.
(3) Korman and Meixner model considers homogeneous and stationary flow
conditions over homogeneous terrain, given by

1 ξμ (ξ )
f (x) = ex
(μ) x 1+μ

where  is the Gamma function, μ is a parameter that depends on atmospheric

stability conditions and ξ is a coordinate incorporating measurement height.
(4) Hsieh model, an approximate analytical solution to advection–dispersion
equation, largely suited to thermally stratified flows, given by

 
1 −1
f (x, zm) = D Z u p
|L|1−P
exp D Z u p
|L|1−P
k2 ∗ x 2 k2 ∗ x

u∗
3

L=− gQ
k Tρcp

where k = 0.4 is the von Karman constant, u* is friction velocity, Q is turbulent heat
flux, T is surface temperature, cp heat capacity, and z is the measuring height.
Atmospheric stability (ζ ) is defined as:

Z −d
ζ =
L

where Z is sensor height above the ground, D is zero plane displacement height, and
L = Obukhov length (Table 8.1).

Table 8.1 Classification of

Category (Z − d/L)
stability of atmosphere based
on calculated atmospheric Extremely unstable ≤1
stability (Z-d/L). Unstable −1 to −0.05
Atmospheric stability
Neutral −0.05 to 0.05
condition defines the size of
the flux footprint Stable 0.05–1
Extremely stable >1
8 Comparison of Flux Footprint Models to a Mixed Fetch Heterogeneous … 97

8.4 Results and Discussion

Table 8.2 shows detailed meteorological and surface roughness conditions measured
by eddy covariance stations during stable and unstable atmospheric conditions. The
partition of fluxes were done using the parameter (Z − d/L), with a value <0
representing unstable (Daytime), and >0 representing stable (Nighttime) conditions.
Figure 8.3 shows the variation in the meteorological parameters during the experi-
ment period. Frictional velocity (u*) varies with the surface roughness, which varies
with crop type and can be observed from Table 8.2.
Figure 8.4 represents the crosswind-integrated footprint for CO2 fluxes obtained
from three different analytical flux footprint models (Hsieh, K & M and Schuepp)
and Lagrangian particle distribution Kljun (LPDM-B) flux footprint model for two
different atmospheric stability condition (unstable and stable) in three different
canopies) C3 cotton cropland, (b) C4 sugarcane cropland, and (v) C3–C4 mixed
cropland. It can be observed that the CO2 flux footprint or fetch varies with the
change in atmospheric stability condition and approximately 75% of the flux source
is confined within a range of 30–40 times the effective measurement height (Zm −
d). The flux footprint in the stable atmospheric condition is too large with respect
to the unstable atmospheric condition as in the stable condition, i.e., during a clear
and calm night, the air within a stable layer is not turbulent and in unstable condition
parcel of air are warmer than the surrounding air and together with turbulent air
enhances the vertical motion of the CO2 eddies.
From Fig. 8.4, it can be observed that Hsieh and K&M model are underestimating
the upwind peak flux footprint distance (X max ) and the magnitude of peak flux (f max ).
Schuepp model is underestimating f max with respect to backward Lagrangian particle
distribution model (Kljun (LPDM-B) for C3 and C4 crops during unstable conditions.
K&M model was observed to overestimate X max , as the model neglects along with
wind turbulent diffusion. Our results conclude that ignoring the along-wind turbulent
diffusion can lead to an underestimation of flux contribution close to flux tower
(Rannik et al. 2000) and increases the extent of flux footprint fetch (Schmid 2002).
The performance of analytical models is evaluated using three residual statistical
parameters including RMSE, NSE, and correlation coefficient R2 . The model that
results in lowest RMSE, highest NSE, and R2 was chosen to accurately represent the
flux footprint. Schuepp model performs best to the Lagrangian model in both C3 and
C4 crop field in unstable condition and Hsieh model performs best in mixed fetch
condition (C3–C4) by considering the heterogeneity in the flux source in both stable
and unstable atmospheric stability condition. Hsieh model also performed best for
the homogenous cotton crop (C3) and gives a good estimate of flux footprint in stable
condition (Table 8.3).
 98

Table 8.2 List of meteorological variables mean wind speed (u [ms−1 ]), frictional velocity (u* [ms−1 ]), Obukhov length (L[m]), roughness length (zo[m]),
displacement height (d[m]), atmospheric stability (Z − d/L), evapotranspiration (ET [mm/hr]), and vapor pressure deficit (VPD [hPa]) measured by eddy
covariance station (Zm = 3, 4, and 7 m) considering unstable (day time) and stable (night time) atmospheric conditions at Nandi Kandi flux site
Day time
Crop DOY u (ms−1 ) u* (ms−1 ) L(m) z0(m) d(m) (Z − d/L) ET VPD
Cotton (C3) 282.000 1.597 0.344 −154.149 0.220 1.020 −0.019 0.058 6993.525
Sugarcane (C3–C4) 282.000 0.430 0.083 −0.982 0.500 3.000 −0.489 −0.112 2443.264
Mixed fetch (C3–C4) 282.000 0.957 0.032 −0.630 0.600 2.680 −3.682 −0.004 2643.263
Night time
Crop DOY u (ms−1 ) u* (ms−1 ) L (m) z0(m) d(m) (Z − d/L) ET VPD
Cotton (C3) 282.000 1.289 0.343 78.845 0.220 1.020 0.038 0.024 5595.206
Sugarcane (C3–C4) 282.000 0.716 0.024 0.275 0.500 3.000 1.744 0.009 1784.401
Mixed fetch (C3–C4) 282.000 0.733 0.012 0.159 0.600 2.680 14.572 0.003 2197.845
S. Kumari and K. B. V. N. Phanindra
8 Comparison of Flux Footprint Models to a Mixed Fetch Heterogeneous … 99

Fig. 8.3 Boxplots of a wind speed, b wind direction, c friction velocity u*, d atmospheric stability
parameter (Z − d/L), and e the standard deviation of the lateral wind component during October
08–12, 2018

Fig. 8.4 Crosswind-integrated footprint for CO2 fluxes (Fc) for Kljun (LPDM-B) footprint curve
(black), Hsieh footprint curve (green), K & M footprint curve (red) and Schuepp footprint curve
(blue) for the Cotton (C3) (col: 1), Sugarcane (C4) (col: 2) and mixed fetch (C3–C4) (col: 4)
considering stability conditions (day time-row:1, night time-row 2) with measurement height (Zm)
= 3 m for C3, 4 m for C4 and 7 m for mixed fetch (C3–C4) cropland

8.5 Conclusion

Flux measurement and footprint modeling in the heterogeneous canopy with vari-
ation in atmospheric stability and surface roughness conditions is challenging. In
this study, three analytical flux footprint models are compared with the Lagrangian
flux footprint model applicable to typical fragmented croplands of the Indian setting.
100 S. Kumari and K. B. V. N. Phanindra

Table 8.3 Statistics of flux footprint model performance considering residual statistical parameters
for C3 at (Zm = 3 m), C4 (Zm = 4 m), and mixed cropland (C3–C4) (Zm = 7 m) in different
atmospheric stability conditions
Model comparison w.r.t. KLJUN( LPDM-B)
Analytical Model RMSE NSE Correlation RMSE NSE Correlation
Cotton(C3): Day time Cotton(C3): Nighttime
(Unstable condition) (Stable condition)
Hsieh 0.0126 0.1648 0.2528 0.0053 0.2914 0.7649
K&M 0.0084 0.0521 0.7824 0.0054 0.2895 0.7400
Schuepp 0.0073 0.0951 0.9541 0.0063 0.1929 -0.0103
Sugarcane(C4): Day time Sugarcane(C4): Nighttime
(Unstable condition) (Stable condition)
Hsieh 0.0119 0.6244 0.5759 0.0125 0.5351 0.5510
K&M 0.0140 0.4261 0.0497 0.0143 0.3246 0.0925
Schuepp 0.0121 0.5334 0.6692 0.0119 0.4530 0.7992
Mixed fetch(C3-C4): Day time Mixed fetch(C3-C4): Nighttime
(Unstable condition) (Stable condition)
Hsieh 0.0247 0.4299 0.3530 0.0060 -2.2641 0.8233 Good
K&M 0.0239 -0.3994 -0.9333 0.0063 -2.2837 0.5200 Average
Schuepp 0.0240 -1.0154 -0.5678 0.0066 -2.3682 -0.0157 Bad

We have experimented in C3 (cotton), C4 (sugarcane), and mixed (C3–C4) crop-

lands having different homogeneity and surface roughness conditions. Flux foot-
print estimated using Lagrangian particle distribution method that considers the
heterogeneity of fetch was used as a reference. To select the suitable model for
given agro-meteorological conditions, the performance of three analytical models
was evaluated with respect to the Lagrangian model. All analytical models were
observed to be sensitive to surface roughness, atmospheric stability, and measure-
ment height. We conclude that the conventional thumb rule of fetch measurement
(1:100) is not applicable to the study area as about 75% of the fluxes comes from
source, i.e., 30–40 times the effective measurement height. For homogenous crop-
land condition, Schuepp model can be used in unstable condition and Hsieh model
in stable condition. However, for heterogeneous condition, Hsieh model can be used
in both unstable and stable atmospheric conditions.

References

Arriga N, Rannik Ü, Aubinet M, Carrara A, Vesala T, Papale D (2017) Experimental validation

of footprint models for eddy covariance CO2 flux measurements above grassland by means of
natural and artificial tracers. Agric For Meteorol 15(242):75–84
Aubinet M, Grelle A, Ibrom A, Rannik Ü, Moncrieff J, Foken T, Kowalski AS, Martin PH, Berbigier
P, Bernhofer C, Clement R (1999) Estimates of the annual net carbon and water exchange of
forests: the EUROFLUX methodology. In: Advances in ecological research, vol 30, pp 113–175.
Academic Press
Finnigan JJ (2004) A re-evaluation of long-term flux measurement techniques part II: coordinate
systems. Bound-Layer Meteorol 113(1):1–41
8 Comparison of Flux Footprint Models to a Mixed Fetch Heterogeneous … 101

Foken T, Leclerc MY (2004) Methods and limitations in validation of footprint models. Agric For
Meteol 127(3–4):223–34
Foken T, Wichura B (1996) Tools for quality assessment of surface-based flux measurements. Agric
for Meteorol 78(1–2):83–105
Horst TW, Weil J (1992) Footprint estimation for scalar flux measurements in the atmospheric
surface layer. Bound-Layer Meteorol 59(3):279–296
Hsieh CI, Katul G, Chi TW (2000) An approximate analytical model for footprint estimation of
scalar fluxes in thermally stratified atmospheric flows. Adv Water Resour 23(7):765–772
Kaimal JC, Finnigan JJ (1994) Atmospheric boundary layer flows: their structure and measurement.
Oxford university press
Kljun N, Calanca P, Rotach MW, Schmid HP (2004) A simple parameterisation for flux footprint
predictions. Bound-Layer Meteorol 112(3):503–523
Kolari P, Pumpanen J, Rannik Ü, Ilvesniemi H, Hari P, Berninger F (2004) Carbon balance of
different aged Scots pine forests in Southern Finland. Glob Change Biol 10(7):1106–1119
Pasquill F, Smith FB (1983) Atmospheric diffusion: study of the dispersion of windborne material
from industrial and other sources. Wiley, New York
Rannik Ü, Aubinet M, Kurbanmuradov O, Sabelfeld KK, Markkanen T, Vesala T (2000) Footprint
analysis for measurements over a heterogeneous forest. Bound-Layer Meteorol 97(1):137–166
Schmid HP (1994) Source areas for scalars and scalar fluxes. Bound-Layer Meteorol 67(3):293–318
Schmid HP (2002) Footprint modeling for vegetation atmosphere exchange studies: a review and
perspective. Agric For Meteorol 113(1):159–183
Van de Boer A, Moene AF, Schüttemeyer D, Graf A (2013) Sensitivity and uncertainty of analytical
footprint models according to a combined natural tracer and ensemble approach. Agric For
Meteorol 15(169):1–1
Wilson JD, Swaters GE (1991) The source area influencing a measurement in the planetary boundary
layer: the “footprint” and the “distribution of contact distance.” Bound-Layer Meteorol 55(1–
2):25–46
Chapter 9
Water Resources Assessment Issues
and Application of Isotope Hydrology
in North East India

Prem Ranjan, Pankaj Kumar Pandey, Vanita Pandey,

and Pema Tshering Lepcha

Abstract The freshwater resources are likely to be severely impacted by climate

change as reported by many researchers around the globe. India already facing a
shortage of freshwater resources due to rapid urbanization and industrialization and
acceleration of economic development activities. Assessment of water resources
plays a key role in the development of the economy of any nation. Northeastern
states of India are blessed with huge water resources. The region faces inconsistent
circumstance as devastating floods in monsoon months and water scarcity in non-
rainy periods at many places. Most of the areas having high reliefs or undulating
terrains where rainwater does not get sufficient time to infiltrate into the soil. As a
result, quick runoff is dominant in the region, and ultimately reduces the recharge
of springshed, and results in the reduction of discharge of many springs and streams
during the non-rainy season. To overcome freshwater stress in the region, a proper
understanding of the water cycle, its management, and development of various infras-
tructure is highly required. Though some problems could not be solved due to the
lack of hydrological data like spring discharge data, water quality, etc. The alter-
nate use of environmental isotope technologies aids researchers to estimate origin,
recharge source age, and its movement within the hydrologic cycle can be of ultimate
solutions. The stable isotopes (Deuterium and Oxygen-18) are excellent indicators of
the water circulation, whereas radioisotopes (Tritium and Carbon-14) distinct value
in detecting the mean residence time (MRT). This methodology has special value in
terms of its cost-effectiveness and the investigative encouragement of the specialists.
This paper presents the use of isotopic hydrology for sustainable development of the
water resources of northeastern region of India.

Keywords Springshed · Environmental isotopes · Groundwater · Spring water ·

Hydrogen and oxygen

P. Ranjan (B) · P. K. Pandey · V. Pandey · P. T. Lepcha

Department of Agricultural Engineering, North Eastern Regional Institute of Science and
Technology, Arunachal Pradesh, Nirjuli 791109, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 103
R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
104 P. Ranjan et al.

9.1 Introduction

In the recent few years, every part of India is facing the problems of hydrological
extremes such as floods and droughts frequently. More than 100th districts of the
country are still witnessing a drought like-situation, due to failure of monsoon rain-
fall (Jitendra 2015). According to the Indian Meteorological Department, a mild to
extremely dry conditions in 404 Indian districts have witnessed due to poor rainfall
in past few years. It is alarming the coming water crises in the country. Many regions
of India facing water crisis in the peak of summer. The Himalayas are known as the
“water towers of the Earth.” The Himalayan range provides water to the millions
of people at the downstream end through the perennial rivers (Vashisht 2015). It
is reported that in the Himalayan region rainfall will be slightly increased along
with greater intensity (Rees and Collins 2006; Singh et al. 2008; Tambe et al. 2012;
Agarwal et al. 2012). In the past five decades, it has been witnessed that the rain-
fall pattern has changed from low-intensity longer-duration events to high-intensity
smaller-duration events (Vashisht 2017; Vashisht and Ranjan 2020). Kusre et al.
(2017) reported that in Sikkim more than 90% average annual rainfall is concen-
trated during April to September. On the other hand, they have witnessed that less
than 2–8 mm rainfall occurring at some stations in the non-rainy period. On the
monthly basis, 70–85% of the rainfall is reported from April to September, whereas
the remaining 15–30% of the rainfall is observed from October to March. Apart from
rainfall, change in land use and land cover is another challenge in water resource
management. Recharging capacity of catchments is reducing due to deforestation,
cutting of hills for new construction (such as a house, roads, and mining), cleaning of
the forest for jhoom cultivation, forest fires, overgrazing and trampling by livestock.
So, one of the main reason behind the reduced in groundwater recharge is the change
in land use and land cover that results in higher runoff (Negi and Joshi 2002).
Nowadays, new methodologies and techniques have been developed in the field of
water resource management sector. In the context of computer-aided research, remote
sensing technique with GIS can be incorporated for any water-related studies. This
can reduce the complexity involved in studying water resources management studies
at any level. Isotope hydrology and hydrogeology is an interdisciplinary science
and emerging discipline with expanding investigation tools for many environmental
problems (Barbieri 2019). It is an alternate solution to many problems and its impor-
tance felt more in recent years. The water sector has facing different problems such as
depleting groundwater, deterioration in water quality, and many other erratic natural
events which affect the hydrological cycle. The traditional standard methods often
fail to provide proper information to many of these problems, whereas use of isotope
method provides a clear picture and help in finding a suitable solution (Barbieri
2019; Kumar 2013). It is now possible, by using of isotopes as conservative tracers
of groundwater flow paths, its sources of regions, resident time in the watershed, iden-
tifying the recharge areas, chemical reactions in surfaces and another environmental
situation. Use of isotopes in hydrology is based on the general concept of “tracing”
9 Water Resources Assessment Issues and Application of Isotope … 105

in which either manually injected or naturally occurring isotopes are utilized. Envi-
ronmental isotopes referred to the naturally occurring elements in waters contain
about seventeen hundred stable and unstable or radioactive isotopes. Environmental
isotopes either stable or unstable (radioactive) have a distinct advantage over artifi-
cially tracers in the sense they encourage the investigation of different hydrological
measures on a larger temporal and spatial scale through their characteristic circula-
tion in a hydrological system. Hence, environmental isotopes technique is unique in
regional studies to find time and space integrated attributes, while artificial tracers are
effective for site-specific and local applications. The most habitually utilized envi-
ronmental isotopes contain deuterium (2 H), tritium (3 H), oxygen (18 O), and carbon
(13 C, 14 C) occurring in water as constituents of dissolving organic and inorganic
compounds (Clark and Fritz 2013). Hydrogen and oxygen (stable isotopes), being
building blocks of the water molecule and hydrological cycle, show a characteristic
isotope signature in water that can be procured as tracers of groundwater recharge
(Price and Swart 2006; Lee et al. 1999; Clark and Fritz 2013) and its signature is
not modified by other biosphere factors like rock–water interaction (Sidle 1998).
In mountainous areas, δ18 O and δ2 H are used to distinguish precipitation derived
groundwater as a result of altitude effect. For determining the recharge area of the
groundwater (Fontes et al. 1967; Stahl et al. 1974), the origin of springs (Dhakal et al.
2014), delineation of spring catchment, to estimate the recharge from precipitation
groundwater flow from higher slopes (Jeelani and Shah 2014).

9.2 Water Resources Assessment Issues in Northeastern

Hilly States

Northeastern states of India face water scarcity despite receiving intensive orographic
rainfall, because of both the physical and anthropogenic reasons. People’s general
perception has water resource management practices that are relevant to water scares
region, but this a wrong opinion need of these practices which has been also felt in
high precipitation regions. Cherrapunji has the highest rainfall in the world, but this
place suffers from drought for 9 months of the year (Agarwal 2000; Kumar et al.
2005). Most of the hilly states of the northeastern part of India has high rainfall
likewise Meghalaya, but in the non-monsoon period, they felt water-scarce situation
(Kusre et al. 2018). Highly undulating land is one of the main reasons for the shortage
of water in these hilly states. These hilly states suffer from the shortage of water due
to highly undulating terrain, generally in the non-monsoon period when there is little
rain or no rain (Lepcha et al. 2021). Also, the other main cause of the declining of
springs discharge is steep terrain of these hilly state. In Sikkim, the elevation range
is shown in Fig. 9.2 (DEM of Sikkim) which indicated that steep terrain, it is also
the main cause of less opportunity of recharge of springs and water conservation
practice is tough tasks, this scenario exists in all states of Northeast India. Fig. 9.3
shown the DEM of Meghalaya. The water-scarce situation is causes of forest fires,
106 P. Ranjan et al.

drop in productivity, incidences of pest and diseases, and shortage of drinking water;
it directly affects the economy of the states. Two springs located in Lekhi village and
Barapani of Papumpare district of Arunachal Pradesh is shown in Fig. 9.1.
Springs (shown in Fig. 9.1) are the main source water, most of the communities are
dependent on this important source (NITI Ayog 2017). It is an essential source for the
existence of human beings in this region. Apart from human and domestic animals, it
has also important for wildlife, ecosystem (Tambe et al. 2012). They are not just part
of the heritage, but the health of the rivers also depends on the health of the springs.
It formed when geologic, hydrologic, or living forces cut the underground layers of

Fig. 9.1 Springs located at a Lekhi village, b Barapani, Papumpare district, Arunachal Pradesh

Fig. 9.2 Digital elevation model of Sikkim

9 Water Resources Assessment Issues and Application of Isotope … 107

Fig. 9.3 Digital elevation model of Meghalaya

soil and rock where water is in movement (Giberson et al. 2013). Each spring is not
quite the same due to the slope and geological structure of the springshed such as
its type, recharge area, and discharge rate. Based on these characteristics, spring has
different classifications such as depression, contact, seepage, gravity, and fracture or
fault spring (Mahamuni and Upasani 2011).
A questionnaire survey is conducted in Sikkim by Kusre et al. (2017), and they
observed that most of the people have felt water scarcity situation in the non-monsoon
period. Ninety percent of respondents expressed that from December to March, water
availability is very less. Also, it has been reported that the discharge of springs is
declining or drying in the December to March period (Tambe et al. 2012). The main
reason of declining of springs are climatic vulnerability (especially rainfall), change
in land use/land cover, socioeconomic and demographic changes (Tambe et al. 2012;
Ranjan and Pandey 2020). In some areas, where villages are situated at hilltop, the
problem has become so acute that peoples are leaving or moving their villages.
It was noticed by local natives that from the last four to five decades, the flow
rate of spring and streams is sharply declined. Now the situation is so worst that
the springs, lakes are dry in non-rainy period, and also perennial rivers’ flow rate is
diminished. To sustain in this water-scarce situation, they have no other choice. To
collect water, peoples travel long distances (Vashisht 2015). In rural areas, women
and children wake up early morning to travel long distances to collect water to fulfil
the daily household requirement. According to a report, 72% women and 14% of
108 P. Ranjan et al.

children have to bear the responsibility of collecting potable water. About, 60% of
the women have to walk half km, and 10% of them have to walk 4 km o fetch
water (Vashisht and Sharma 2007). In this water-scarce scenario, farmers are leaving
abandon the agricultural land due to an insufficient quantity of water for irrigation.

9.3 Springshed Development and Challenges

To survive in this water-scarce scenario, there is a need to adopt different water

management practices. Such as the installation of water harvesting structure, reju-
venation of springs, recharging of groundwater, transferring of water from excess to
shortage area, and spring development practices (Vashisht 2015; Agrawal et al. 2015;
Tambe et al. 2013). Implementing of springshed development programs has many
challenges such as inventory of spring in remote areas, accurately identification of
recharge areas, in-situ rainwater harvesting in agricultural land, developing the local
capacity building, and public financing. The major goals of springshed development
are
• Stop deforestation, planting trees, increase vegetation cover and soil organic
matter, reduce soil erosion, and reduce runoff velocity,
• In-situ rainwater harvesting by bunds, trenches, ponds, and check dams,
• Increase infiltration opportunity time to naturally recharge the aquifer,
• Protection of springshed to allow ecological restoration and prevent contamination
of the groundwater and runoff water,
• Develop suitable water storage structure,
• Aware and train the community about springshed development and management
practices.
Information of reliability of springs is essential to adopt any water management
practices. There are various methods, models, and indices which are available to
determine the reliability of springs such as (i) Time-lag from the incidence of partic-
ular rainfall to consequent change in the spring discharge, (ii) the amount of rainwater
stored in the aquifer; (iii) maximum to minimum discharge ratio of spring in a year;
and (iii) prediction of spring discharge rate and yield for upcoming recession period.
So, to assess the spring reliability rainfall, the discharge rate is required. To quantify
the spring reliability based on historical data, different mathematical techniques are
available (Vashisht 2015, 2017).
The discharge rate of any springs depends on several factors, such as natural and
man-made. But the main factor is the amount of water stored in an aquifer, which
depends on rainfall occurrence in spring catchment and locally recharged (Ranjan
and Pandey 2020). Spring will eventually dry up, if there is no rainfall or recharge
in spring catchment or if the recharge rate is lesser than pumping rate. About 90%
of the total rainfall occurred from June to October or April to September, in this
duration aquifer is recharged. Precipitation in the non-monsoon period is negligible
or sometimes it completely absent. So, recharge non-rainy season is negligible or if
9 Water Resources Assessment Issues and Application of Isotope … 109

some amount of rainfall occurs in this duration it may evaporate. The best condition
of recharging of springs is when temperature and evaporation are low and the ground
is more absorbent.

9.3.1 In-Situ Spring Water/Rainwater Harvesting

Springs are drying due to less recharge or no recharge of the aquifer. The major
reason for not recharging or replenishment of aquifers are deforestations, change in
precipitation pattern, land cover, and soil property. Due to deforestation and change
in precipitation (such as high density and short duration), rainwater is not getting
sufficient time for infiltration or deep percolation, only a few quantities of water
infiltrate, and most of the flushes in valleys and streams. The high quantity of runoff
is one of the major reason for landslides.
Construction of trenches, pit and check-dam, plantation of local tree species,
minimizing grazing, cutting of fuelwoods and grasses and installing the fencing
around the springshed are the common measures for development of springs and
groundwater. Developing trenches (shown in Fig. 9.4a), CCT, pits, and check dams
(shown in Fig. 9.4b) is generally used in rejuvenation programs. These structures
break the runoff velocity and increase the infiltration opportunity time to recharge
the aquifer. Apart from the rejuvenation program, harvest available water when it is
excess and construct storage structures. Apart from these techniques, local commu-
nities of the Himalayan region has also developed many in-situ water harvesting
practices, depending upon location specifics and named them accordingly.

Fig. 9.4 a Trenches excavated at hill slope, and b Bamboo check dam constructed in stream
110 P. Ranjan et al.

9.3.2 Springshed Protection

Protect the springshed catchment area, grow preferred natural vegetation, and disre-
gard the human activities like agricultural uses which increase the chances to the
inflow of pesticides and erosion risks. Adopt different traditional and modern tech-
niques to protect the recharge area. Such as install fences surrounding of recharge
area to avoid man and animal interference, avoid any garbage dumping sites, wastew-
ater discharge in springshed, avoid plant and hill cutting for new construction (roads,
hotels, picnic spots, and industries) in springshed, avoid pesticides and insecticides
in agricultural filed, plant local, and less water consumption trees. Local communi-
ties and water user groups are also using traditional knowledge to protect the spring.
Communities of this reason assume water sources are holy places. They worship
these sources as a goddess. These practices are also helpful to protect spring from
contamination because people think this is a holy place. So, they not throw any litters
and garbage at that site. But these practices are limited at the only source point.

9.3.3 Major Challenges, Conflicts and People Participation

in Spring Development and Management Practices
in Northeast Himalayas Region

In Himalayan region, travelling and fieldwork are a very hectic task due to the uneven
topographic. Without proper climbing equipment and local guide, many places are
inaccessible due to high relief. The complex geology and repeated structural distur-
bances and thick forest coves make it more difficult to develop a geological map
and its study. Also, disasters such as landslides and floods add more difficulties in
research work. Unavailability of spring data is one of the major challenges of devel-
opment practices. Lack of historical data of springs such as location, discharge rate,
types of spring availability, water quality, and other aquifer properties. So, there is
no database for any planning, development, and management practices. The major
cause of the lack of historical data is the lack of awareness about this important
source. Most of the communities living in this region are depended on these impor-
tant sources, but data collection is never a priority. Communities are not aware of
the upcoming water-scare situation. So, these are the major reason for such data
gaps. As per the geohydrology of the area, the recharge area of springs may fall in
the forest, in administrative boundaries of beneficiary communities or maybe in any
other villages or maybe in any other private land. Apart from geohydrology land,
use of recharge area is also a major challenge, and it may be agricultural and horti-
cultural land, reserved forest, private forest, and wastelands. So, this is one of the
causes of conflicts between two villages or communities or landowner. In the case of
government property or forest land, there is a need for permission from the concerned
department or forest department for any development practices. The methodology
for executing the revive measures cannot be uniform in all the cases referenced, and
hence, it poses challenges for the implementing organizations.
9 Water Resources Assessment Issues and Application of Isotope … 111

9.4 Isotope Hydrology Applications in Water Resource

Management

Investigation of groundwater from conventional techniques fail to provide proper

information such as recharge sources, mechanism, and deposition. Adopting the
new technologies such as isotope hydrology helps to investigate the clear picture or
proper information of groundwater flow paths, its sources of regions, resident time
in the watershed, identifying the recharge areas, chemical reactions in surfaces, and
another environmental situation. Environmental isotopes, both stable and unstable
(radioactive) isotopes, are used in water resources assessment. Stable isotopes usually
measured by Isotope Ratio Mass Spectrometer (IRMS), in terms of the isotope
ratios of the less abundant to more abundant. Radioactive (or unstable) isotopes are
measured by counting the number of atoms (using accelerator mass spectrometer) or
counting their radioactive decays (by liquid scintillation counter) in a given sample.
It is an effective technique to obtain the necessary critical hydrological information
such as (1) Determination of age, velocity, and path of flow; (2) Origin or source of
recharge; (3) Possible interconnections between different aquifers, and surface water
and groundwaters; and (5) Aquifer properties such as porosity, transmissivity, and
dispersivity. The isotopic concentrations are expressed as the variations between the
measured ratios and the standard sample. The isotopic compositions of water are
denoted by deviations (δ) and expressed in per mil (‰) from a reference standard
mean ocean water (SMOW) as shown below:
⎧   ⎫
⎪ Abundnce of less (heavier) isotope
⎨ Abundance more abundant (lighter) isotope ⎪
⎬
Sample
δ( /00) = 
0
 − 1 × 1000
⎪ Abundance of less abundant (heavier) isotope
⎩ Abundance ⎪
⎭
of more abundant (lighter) isotope
Standard

or
 
Rsample
δ( /00) =
0
− 1 × 1000 (9.1)
Rstandard

As the defined δ values are small numbers, they are expressed in ‰ (per mill).

9.4.1 Global Meteoric Water Line (GMWL) and Local

Meteoric Water Line (LMWL)

GMWL Eq. (9.2) is recognized by Craig (1961) based on global relationship between
δD and δ 18 O in natural meteoric water. Water samples of rainfall, snow, and river
are collected from all over the world and analyzed for their isotopic composition.
Based on Eq. (9.2), deviations of isotopic composition from equilibrium process
to local or regional line can be defined. The meteoric water line may vary with a
112 P. Ranjan et al.

place to place depending on weather condition. Further, Rozanski et al. (1993) have
developed Eq. (9.3) based on Eq. (9.2) by incorporating the monthly average values
from different stations belonging to Global Network of Isotopes in Precipitation
monitored by World Meteorological Organization and Atomic Energy Agency.

. O + 10
δD = 8 × 18 (9.2)

δ D = 8.1 (± 0.07) × δ18O + 11.27 (± 0.65) (9.3)

The GMWL represents the variation of δ2 H or δD to δ18 O in precipitation and can

be utilized to identify the source of water, mixing, and other hydrological measures.
Almost, precipitation in every part of the world follows this relationship. But in some
specific areas which have different temperature and humidity or unique climate,
create their special LMWL with a different slope and intercept. The slope and
intercept of LMWLs depend on hydrological parameters.

9.4.2 Identify the Springs Origin and Their Recharge Area,

Altitude Effect, Mean Residence Time

Springs originating naturally from unconfined aquifers were several factors such as
changes in climatic conditions, land use, and demography. are affecting the discharge
of springs. Communities of the northeastern hilly region are fully depending on
springs. Springs are also responsible for ecological balance and baseflow of flowing
rivers. The variation in the δ18 O, δD (or δ2 H) isotopic signatures of catchment waters
(streams, springs, precipitation, and snow) have been used to identify the hydrological
source areas under different flow conditions based on a spatio-temporal variation of
stable water isotopes which is pre-requisite for understanding the identification of
recharge problem.
Dhakal et al. (2014) conducted an isotope-based study in Sikkim to identify the
origin and recharge area to understand the spring’s dynamics. They found that the
initial results reveal that the spring discharges are strongly dependent on the rainfall
pattern. The stable isotope studies of the spring samples show that the spring samples
fall above the global meteoric water line as shown (Fig. 9.4) by Himalayan aquifers
and evaporation enrichment during the recharging process. Electrical conductivity
analysis shows that the aquifers are recharged on a local level. All spring samples
fall above the Global Meteoric Water which is generally observed in the Himalayan
region because rainwater is depleted. Linear regression line equation for samples
collected in March 2013 is δ2 H = 5.5δ18 O–5.9 (r = 0.82, n = 12) which indicates
evaporative enrichment of stable isotopes during the recharge process. The highest
enrichment has been found in the spring named Lower Changey Source. Linear
regression line equation for samples collected in May 2013 is δ2 H = 4.1 δ18 O–16
which indicates more enrichment than March samples with a linear regression line
9 Water Resources Assessment Issues and Application of Isotope … 113

of δ2 H = 5.5.δ18 O–5.9. The study shows that the springs are local origin and the
possible recharge zone of the aquifers is within the local geographical area. They
also show that the rainfall-spring interdependence which clears that the source of
recharge is rainwater (Fig. 9.5).
There are several studies on the identification of origin and recharge area delin-
eation in different Himalayan regions. In western Himalya, Jeelani (2008) observed
an inverse relationship between spring and stream discharge with precipitation in
Anantnag District of Jammu & Kashmir. He also observed that springs of Anantnag
district are recharged by glaciers and are under threat due to the reduction in glacial
coverage. Jeelani et al. (2010), Bhat and Jeelani (2015) delineated the origin of
streams and springs of Liddar and Bring Bringi catchment of Kashmir Himalayas
using δ18 O. They also analyzed the effect of altitude on springs and observed multiple
sources during warm months and common source during colder months. Shah et al.
(2017) estimated mean residence time (MRT) using tritium method, sine wave model,
and artificial tracers and reported that cold karst springs are fed with recent recharge
and on the other hand, warm karst springs with a mixture of old and recent recharge.
A comparison of MRT has been made between tritium and the sine wave model,
longer MRT was observed in warm karst springs and shorter in cold springs. They
reported that short travel time of karst spring in the region indicates the water quality
deterioration and variation in the magnitude of flux. So, identified catchment area
must be protected, preserved, and managed from any anthropogenic activities.

Fig. 9.5 δ 18 O versus δ 2 H plot of spring water samples collected in March and May 2013. Source
Dhakal et al. (2014)
114 P. Ranjan et al.

9.5 Conclusion

The northeastern hilly states communities are mostly dependent on springs and
streams for freshwater. It is reported that the springs are drying up due to the uncer-
tainty of rainfall and land use/land cover change. Water harvesting is also a tough
challenge due to the steep topographic profile exist in the NE region of India. Water
scarcity is increased in the winter season. At some places, women and children travel
long distances to get water. To sustain in this water-scarce situation, rejuvenate or
recharge springs/groundwater and protect this important source. Use traditional and
modern techniques to harvest water. Now, it needs to the management of springs
using new scientific methods or technique which can help in enhancing the climate
change resilience of the local communities. Environmental isotopes such as Oxygen-
18, Deuterium (2 H), and tritium (3 H) and isotopes are useful to identify the origin of
recharge, recharge area. It is also helpful to identify the source of contamination of
spring water.

Acknowledgements Authors are grateful to the Department of Agricultural Engineering, North

Eastern Regional Institute of Science and Technology, Nirjuli, Arunachal Pradesh, for providing
the support and valuable suggestions for bringing this paper.

References

Agarwal A (2000) Drought? Try capturing the rain. Center Sci Environ 1–16
Agarwal A, Agarwal NK, Rathi VK (2015) Water stress on springs of Lesser Himalayan region.
British J Appl Sci Technol 9(3):243–255
Agarwal A, Bhatnaga NK, Nema RK, Agrawal NK (2012) Rainfall dependence of springs in the
Midwestern Himalayan Hills of Uttarakhand. Mt Res Dev 32(4):446–455
Barbieri M (2019) Isotopes in hydrology and hydrogeology. Water 11:291
Bhat NA, Jeelani G (2015) Delineation of the recharge areas and distinguishing the sources of
karst springs in Bringi watershed, Kashmir Himalayas using hydrochemistry and environmental
isotopes. J Earth Syst Sci 124(8):1667–1676
Clark ID, Fritz P (2013) Environmental isotopes in hydrogeology. CRC Press
Craig H (1961) Isotopic variations in meteoric waters. Science 133(3465):1702–1703
Dhakal D, Tiwari A, Tambe S, Sinha UK, Arrawatia ML (2014) Isotope studies to identify the
origin and recharge area of Himalayan springs as a climate change adaptation initiative: a case
study from Sikkim, Eastern Himalaya. Int J Earth Sci Eng 07(01):136–140
Fontes JC, Letolle R, Olive P, Blavoux B (1967) Oxygene 18 et tritium dans le basin d’Evian.
In: Proceedings of symposium on isotopes in hydrology. International Atomic Energy Agency,
Vienna, p 401
Giberson D, Knysh K, MacDonald E (2013) Springs of the clearest and purest water. Island Mag
73:2–11
Jeelani G (2008) Aquifer response to regional climate variability in a part of Kashmir Himalaya in
India. Hydrogeol J 16(8):1625–1633
Jeelani GH, Bhat NA, Shivanna K (2010) Use of δ18O tracer to identify stream and spring origins
of a mountainous catchment: A case study from Liddar watershed, Western Himalaya. India. J
Hydrol 393(3–4):257–264
9 Water Resources Assessment Issues and Application of Isotope … 115

Jeelani G, Shah RA (2014) Distinguishing and estimating the sources of spring recharge: a case
study of Martand Karst Spring of Kashmir Valley
Jitendra (2015). Down to earth. [Link]
cts-have-suffered-drought-conditions-for-15-years-50993.
Kumar CP (2013) Hydrological studies using isotopes. Int J Innov Res Dev 2(13)
Kumar R, Singh RD, Sharma KD (2005) Water resources of India. Curr Sci 794–811
Kusre BC, Bora PK, Rai D, Adhikari K, Niranjit Khuman C, Tabing J (2018) Developing water
resources management strategies for South Sikkim District, India. Water Sci Technol Water
Supply 18(2):648–659
Kusre BC, Bora PK, Rai D, Suranjoy Singh S, Tana A, Subba S (2017) Combating water scarcity
through roof water harvesting: planning and design with stakeholders’ perception in Sikkim
(India). Water Sci Technol Water Supply 17(3):799–810
Lee KS, Wenner DB, Lee I (1999) Using H-and O-isotopic data for estimating the relative contri-
butions of rainy and dry season precipitation to groundwater: example from Cheju Island. Korea.
J Hydrol 222(1–4):65–74
Lepcha P T, Pandey PK, Ranjan P (2021) Hydrological significance of Himalayan surface water
and its management considering anthropogenic and climate change aspects. In: IOP conference
series: materials science and engineering, vol 1020, No. 1, p 012013
Mahamuni K, Upasani D (2011) Springs: a common source of a common resource. In: 13th Biennial
conference of the international association for the study of the commons (IASC), Sustaining
commons: sustaining our future.
Negi GCS, Joshi V (2002) Drinking water issues and development of spring sanctuaries in a
mountain watershed in the Indian Himalaya. Mt Res Dev 22(1):29–31
NITI Aayog (2017) Inventory and revival of springs in the himalayas for water security. Depertment
of Science and Technology, Government of India, New Delhi
Price RM, Swart PK (2006) Geochemical indicators of groundwater recharge in the surficial aquifer
system, Everglades National Park, Florida, USA. Geol Soc Am Spec Pap 404:251–266
Ranjan P, Pandey PK (2020) Reviving, development and protection of springs to increase water secu-
rity in the Himalayan Region. In: Proceedings of the 5th international conference on computers
& management skills, pp 25–29
Rees GH, Collins DN (2006) Regional differences in response of flow in glacier-fed Himalayan
rivers to climate warming. Hydrol Process 20:2157–2167
Rozanski K, Araguás-Araguás L, Gonfiantini R (1993) Isotopic patterns in modern global
precipitation. Clim Change Continental Isotopic Rec 78:1–36
Shah RA, Jeelani G, Jacob N (2017) Estimating mean residence time of karst groundwater in
mountainous catchments of Western Himalaya, India. Hydrol Sci J 62(8):1230–1242
Sidle WC (1998) Environmental isotopes for resolution of hydrology problems. Environ Monit
Assess 52(3):389–410
Singh P, Kumar V, Thomas T, Arora M (2008) Changes in rainfall and relative humidity in river
basins in northwest and central India. Hydrol Process 22:2982–2992
Stahl W, Aust H, Dounas A (1974) Origin of artesian and thermal waters determined by oxygen,
hydrogen and carbon isotope analyses of water samples from the Sperkhios Valley, Greece. In:
Proceedings of a symposium on isotope techniques in groundwater hydrology, vol I
Tambe S, Kharel G, Arrawatia ML, Kulkarni H, Mahamuni K, Ganeriwala AK (2012) Reviving
dying springs: climate change adaptation experiments from the Sikkim Himalaya. Mt Res Dev
32(1):62–72
Tambe S, Kharel G, Subba S, Arrawatia ML (2013) Rural water security in the sikkim Himalaya:
status, initiatives and future strategy. India Mountain Initiative, Summit in Kohima, Nagaland
Vashisht AK (2015) Adaption of seepage spring development technique to manage the water scarcity
in Himalayan region–a practical approach. Water Util J 11:93–98
Vashisht AK (2017) Use of mathematical techniques for determining the reliability of springs. In:
Development of water resources in India. Springer, Cham, pp 113–124
116 P. Ranjan et al.

Vashisht AK, Ranjan P (2020) Intermittent Multi-column sand filter: a unique solution to multiple
applications. J Inst Eng (India) Ser A 101(1):69–75
Vashisht AK, Sharma HC (2007) Study on hydrological behaviour of a natural spring. Curr Sci,
837–840
Chapter 10
Water Hammer Analysis for Pipe Line
Network Using HAMMER V8i

Ajmal Hussain, Muhammad Mustafa, S. M. Ahbar Warsi, and Sumit Kumar

Abstract Hydraulic transients occur as a direct result of rapid variations of flow field
in pressurized (closed-conduit) systems. The change in velocity from valve closures
or pump operations causes pressure surges that are propagated away from thesource
throughout the pipeline. If the maximum pressures exceed the bar ratings(mechanical
strength) of the piping material, different types of failure such as pipe bursts canoccur.
Similarly, if the minimum pressure drops below the vapour pressure of the fluid, cavi-
tationcan occur and can be detrimental to the pipeline system. The purpose of present
study is to asses and simulate the hydraulic transients in a pipe line network of treated
effluent rising main of Mpophomeni sanitation scheme using Bentley HAMMER
V8i. A total of five scenarios were simulated using different combinations. The
simulation results shows that the transient pressures in the pipeline exceeded the bar
rating of the pipe where the burstsor cavitation may occur for the simulated scenarios.
This study shows that the transient pressures in pipe line system were reduced to safe
limit after providing water hammer protection devices.

Keywords Hydraulic transient · Water hammer · Cavitations · Bentley HAMMER

V8i

10.1 Introduction

Water hammer commonly occurs when a valve closes suddenly at an end of a pipeline
system, and a pressure wave propagates in the pipe (Chaudhry 1979). Under steady
state conditions in a pipeline system, flow variables like discharge remainconstant.
However, if a sudden change occurs in the system through a change in control oper-
ations such asthe closure of an outlet valve or the sudden shutdown of a pump due

A. Hussain (B) · S. M. A. Warsi · S. Kumar

Department of Civil Engineering, Zakir Hussain College of Engineering and Technology, AMU,
Aligarh 202002, India
M. Mustafa
Design Section, SMEC India Private Limited, Gurgaon, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 117
R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
118 A. Hussain et al.

to power failure, a transientstate is initiated, and it takes a finite amount of time

before another (new) steady-state conditionis established in the pipeline system. The
flow phenomenon associated with such rapid changesis called a hydraulic (or fluid)
transient. The main concern during a hydraulic transient in asystem is the rapid fluctu-
ations in the pressure since dramatic changes in the pressure can resultin catastrophic
damage to pipelines and hydraulic machinery.
By closing the valve rapidly,the valve converts the kinetic energy carried by the
fluid particles into strain energy in the pipewalls. This results in a “pulse wave” of
abnormal pressure to travel from the disturbance into thepipe system. Energy losses
due to mainly friction cause the transient pressure waves todecay until a new steady
state is established. Hydraulic transient events in water distribution system can cause
significant damage, disruption,and expense (Huo et al. 2007). In general, transient
events are usually most severe at control valves, pump stations, in high-elevation
areas, and in remote locations that are far from overhead storage tanks. However, all
systems have to start up, switch off, undergo flow changes, and so on. In addition,
water systems are not immune from human errors, malfunction and break downof
mechanical devices, and other risky events. Emadi and Solemani (2011) investigated
the effect of parameters such as pipe diameter, thickness and type, moment of inertia
and temperature on maximum water hammer in the context of Kuhrang Pumping
Station. Fluid pipeline failures due to water hammer effects are described in detail
by Schmitt et al. (2006).
During a hydraulic transient state, a pipeline may be subjected to objectionable
high and lowpressure cycles. The high pressures can damage the pipeline system
components, such as valves,pumps, and other pipeline components, as discussed
earlier. The change in the fluid velocity (more correctly discharge) in the pipeline
systems is the first step that leads to a hydraulictransient. The resulting change in
pressure is directly proportional to the change in velocity. Hence, as much as possible,
sudden changes in the velocity should be avoided to minimize theoccurrence of
pressure transients in the system. Bergant et al. (2012) presented a comprehensive
water hammer analysis of pumping system for control of water in underground mines.
Deshmukh (2014) presented the hydraulic transient analysis of Kolar water pipeline
using using Bentley HAMMER V8i.
The present study was conducted using the popular surge software Bentley
HAMMER V8i. Bentley HAMMER V8i is a versatile program capable of modeling
any type of surge protection device and its powerful graphical results presentation
and interpretation capability has helped thousands of engineers worldwide design
large complex transmission mains, small branching networks as well as large distri-
bution networks for over 30 years. Bentley HAMMER V8i is based on technology
originally created by Environmental Hydraulics Group (HAMMERTM 2005). Water
hammer equations for elastic pipes produce a 1-D partial equation and may be solved
by Methods of Characteristics (MOC). Hammer Software uses the Method of Char-
acteristicto solve non-linear differential equations which have the following form,
Evangelisti (1969), Fox (1977), Streeter (1967, 1972):
10 Water Hammer Analysis for Pipe Line Network Using HAMMER V8i 119

dV 1 dP f |V |V
+ · + =0 (10.1)
dt ρc dt 2D
∂V 1 ∂ρ
a2 + =0 (10.2)
∂s ρ ∂t

Solution of the above equations using MOC will be

C+:
a f x
(Qp − Qi−1 ) + (Hp − Hi−1 ) + Qp |Qi−1 | = 0 (10.3)
gA 2gDA2

C-:
a f x
(Qp − Qi+1 ) − (Hp − Hi+1 ) + Qp |Qi+1 | = 0 (10.4)
gA 2gDA2

For the solution there are many boundary conditions are considered requiring
like reservoirs, pumps, pipeline branches, dead ends etc. This method not only saves
times but also reduces likelihood of mistakes which may occur while copying date
to the software.

10.2 Problem Statement and Procedure of Analysis

The present study is carried for a pipe line network of treated effluent rising main
of Mpophmeni sanitation scheme, South Africa which consists of following major
units:
a. Pump: Pump working with the capacity of supplying a flow rate of 250 m3 /h up
to a head of 76 m. Pump has a rated power of 90 kW at 1450 rpm.
b. Rising Main Network: Rising main consists of PVC class 12 pipe of diameter
400 mm for first 3000 m and then the pipe diameter reduces to 315 mm upto
5423 m,then pipe diameter increases to 400 mm till 5503 m,then again reduces
to 315 mm and continues till the tail end discharging into the receiving chamber.
The whole network configuaration is shown in Fig. 10.1.
Profile for the elevation with chainage of the rising main network is shown in
Fig. 10.2. The pump station is located at zero chainage of the profile and the rising
main culminates in free fall at a receiving chamber at the tail end. As evident from
Fig. 10.2; three major peaks and low points are observed along the network path and
in the last leg of the network i.e. after chainage of around 5500 m, network will be
acting under gravity. The undulating topography of the rising main path exposes the
network to the risk of surge occurrences in events of power failure, sudden closure
of valves etc. The main objective of the study is to identify transient issues for this
system and recommend surge protection alternatives.
120 A. Hussain et al.

Fig. 10.1 Steady state model for problem 2

Fig. 10.2 Profile of hydraulic grade line and elevation with chainage of the rising main network

10.3 Results and Analysis

10.3.1 Baseline Scenario: Steady State Conditions

A baseline run of the network is conducted to identify the baseline scenario that is
network running under steady state without any transient event. This is conducted
to identify the steady state conditions for the network. Baseline scenario model is
shown in Fig. 10.2.
Figure 10.2 indicates thatHGL is significantly below the ground elevation after
around 4300 m chainage. So, to counter negative pressure heads in this region, a
10 Water Hammer Analysis for Pipe Line Network Using HAMMER V8i 121

Fig. 10.3 Reservoir location between J-50 and J-52

Fig. 10.4 Hydraulic Grade Line after providing the reservoir for the network

reservoir at the highest elevation point(1105 m) at J51 is provided which is shown

in Fig. 10.3.
After providing the reservoir; Hydraulic Grade Line (HGL) for the network is
shown in Fig. 10.4. As a result the negative pressure heads are well within the limits
ie −10 m H2 O (or −98.1 kPa) and the positive pressure heads are also within the
safe limits as shown in Fig. 10.5.

10.3.2 Surge Analysis on Baseline Network Without Surge

Protection

In the next stage the impact of a power failure is simulated without any surge protec-
tion device. Hydraulic Grade Line for baseline network without surge protection
device is shown in Fig. 10.6. For the analysis it is assumed that the check valves
installed at the pump closes after 5 s of power failure, which is below the critical
122 A. Hussain et al.

Fig. 10.5 Pressure diagram for steady state analysis after providing reservoir

Fig. 10.6 Hydraulic grade line for baseline network without surge protection device
10 Water Hammer Analysis for Pipe Line Network Using HAMMER V8i 123

Fig. 10.7 Pressure diagram for transient analysis for the network without surge protection device

time period. After running transient simulation, it’s found that the transient effect is
only in the portion before the reservoir at high point as shown in Figs. 10.6 and 10.7.
Also negative pressures are below −98 kPa in many parts of the pipeline. So, surge
protection devices are required for protection from water hammer in this region only.

10.3.3 Analysis with Surge Protection Device

[Link] With Application of Four Air Valves

To minimize negative pressure heads, double acting Air Valves, with inflow orifice
dia. 80 mm and out flow orifice dia. 2.0 mm, are adopted, at different locations shown
in Table 10.1 and Fig. 10.8.
The results obtained after adding valves in the pipe line network are shown in
Fig. 10.9. The locations at which valves were added, negative pressures are reduced

Table 10.1 Location of air

Air valve no Chainage
valve at different chainage
along the pipe line AV-1 1080
AV-2 2728
AV-3 4785
AV-4 5260
124 A. Hussain et al.

Fig. 10.8 Air valve location (AV-1, AV-2, AV-3 & AV-4) at different locations of the network

Fig. 10.9 Pressure diagramafter addingfour air valves to the network

but they are more than the permissible limits at many locations. So, some other
combination of surge protection devices is to be used.

[Link] With Application of 4 Air Valves and 1 Hydropneumatic Tank

To further reduce the negative pressures, a hydropneumatic tank to the network at

J1 is provided as shown in Fig. 10.10. The properties of the hydropneumatic tank
are:Volume = 2000 L,Liquid Volume(Initial) = 1600 L, Tank Calculation Model =
Gas Law Model, Dia. (Tank inlet Orifice) = 175 mm and HGL (initial) = 1140 m.
The results obtained after adding four valves and a hydropneumatic tank in the
pipe line system is shown in Figs. 10.11 and 10.12. After adding hydropneumatic
10 Water Hammer Analysis for Pipe Line Network Using HAMMER V8i 125

Fig. 10.10 Hydropneumatic tank at J1

Fig. 10.11 Hydraulic grade line with elevation and chainage for the network with hydropneumatic
tank and four air valves

tank, the negative pressure values have been reduced significantly and are within safe
limits as shown in Fig. 10.12. Positive pressures also do not exceed the safe limits.
Therefore, it can be said that the pipe line network of treated effluent rising main of
Mpophmeni sanitation scheme is safe with the present mitigation measures.

10.4 Conclusion

Based on the findings, application of four air valves and one hydropneumatic tank is
recommended for the safe operation of treated effluent rising main of Mpophmeni
sanitation scheme. Without surge protection device, the negative pressure was greater
126 A. Hussain et al.

Fig. 10.12 Pressure diagram for transient analysis with hydropneumatic tank and four air valves

than −98 kPa on several points, but after providing surge protection, negative pres-
sures are well below the safe limit. The hydropneumatic tank may be provided at the
immediate downstream of the pump. A minimum of above four air valves shall be
provided to contain the effect of downsurge in the network, however any additional
air valves provided in the network will further improve the network performance. It
can be concluded that the present developed model for the pipeline system reduces the
risk of damages associated with water hammer and consequently increase the safety
and as well as reduce the failure rate for the present pipe line system. It also reduces
wearing and tearing effects of water hammer in pumping and pipeline systems, and
increase lifetime of the infrastructure.

References

Bergant A, Simpson AR, Sijamhodzic E (2012) Water hammer analysis of pumping system for
control of water in underground mines. International mine water association, pp 9–19
Chaudhry MH (1979) Applied hydraulic transients. Van Nostrand Reinhold Company, New York
Deshmukh TS (2014) Hydraulic transient analysis of kolar water pipeline using Bentley HAMMER
V8i. Int J IJERT 3(9)
Emadi J, Solemani A (2011) Maximum water hammer sensitivity analysis. IJERT
Evangelisti G (1969) Waterhammer analysis by the method of characteristics. L’Energia Elettrica,
Nos. 10–12
Fox JA (1977) Hydraulic analysis of unsteady flow in pipe networks. The Macmillan Press Ltd.,
London and Basingstoke
10 Water Hammer Analysis for Pipe Line Network Using HAMMER V8i 127

HAMMERTM User Guide, Bentley Systems Inc. (2005)

Huo J, Eckmann DH, Morris KE (2007) Surge protections: review, analysis, and engineering design.
In: Proceedings of AWWA/ACE07, Toronto, Ont., Canada
Schmitt C, Pluvinage G, Hadj-Taïeb E, Akid R (2006) Water pipeline failure due to water hammer
effects. Fatigue Fract Eng Mater Struct 29(12):1075–1082
Streeter VL (1967) Water-hammer analysis of distribution systems. J Hydraulics Div Proc ASCE
93(HY5):185–201
Streeter VL (1972) Unsteady flow calculations by numerical methods. Proc ASME J Basic Eng
94(series D, No. 2):457–466
Chapter 11
Dam Break Flood Routing
and Inundation Mapping Using
HEC-RAS and HEC-GeoRAS

A. Bharath, Anand V. Shivapur, and C. G. Hiremath

Abstract Dam break analysis helps to predict the breach flood hydrograph and its
time of arrival at desired locations along with depth and velocity of flow. This paper
illustrates a case study on dam break analysis of Hidkal Dam using one-dimensional
hydraulic model HEC-RAS with the help of river geometry generated by DEM. The
present study illustrates the simulation of flood hydrograph under PMF scenario for
both piping and overflow failure modes. The peak flow values as obtained from the
model are 72,020.57 m3 /s and 78,384.21 m3 /s at the downstream of the dam and
the corresponding inundation area estimated to be of 74.32 km2 and 78.19 km2 for
piping failure and overtopping failure, respectively. The flow depth is found to vary
from 30.39 m to 13.09 m for piping failure and 34.98 m to 13.86 m for overtopping
failure at the downstream of the dam and the end of river reach, respectively.

Keywords Dam break analysis · Breach flood hydrograph · Breach parameters ·

HEC-RAS · Flood inundation

11.1 Introduction

A dam is a barrier constructed across a river, which obstructs or directs the flow
and often creates a reservoir or lake. The water impounded by the dam is used for
many purposes like drinking water supply, irrigation water supply and hydroelec-
tric power generation. Though the dams have many benefits for our society, their
failures cause most devastating disasters. Dam failures are very rare events, but it
causes immense damage to life and property, hence it generally reminds engineers
the probability of dam failure (Abhijith et al. 2017). Dam break can be summarized as
partial or catastrophic failure of a dam resulting in uncontrolled flood wave of greater

A. Bharath (B)
Department of Civil Engineering, GITAM University, Bangalore, India
A. V. Shivapur · C. G. Hiremath
Department of Water and Land Management, Centre for PG Studies, VTU, Belagavi, Karnataka,
India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 129
R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
130 A. Bharath et al.

height traveling at a very high speed which can cause damage to life, property and
infrastructure in the downstream (Xiong 2011).
Piping, extreme storm, landslide, earthquake, malfunctioning of equipment, foun-
dation failure, sabotage, structural damage are the mechanisms that can cause dam
failures (USACE 2016). Regardless of the reason, majority of the dam failures begin
with the breach formation (Mohamed 2018; Wang et al. 2015).
A breach can be defined as the opening formed in the dam which results in dam
failure and this event produces a flood wave from the water impounded behind the
dam. Though the main failure modes have been recognized as overtopping or piping,
the real failure mechanics for both earthen and concrete dams are not well understood
(Xiong 2011; Zhang et al. 2009).
Dam break analysis helps in predicting the peak flood and areas affected by flood
in the downstream due to dam breach which helps in the estimation of rehabilitation
cost. Such analysis predicts the potential of precautionary measures which can be
taken to avoid the dam break which minimizes or avoid damage (Kulkarni et al.
2016). Dam break modeling helps to simulate dam break flood wave propogation
along the downstream of the dam (Khosravi et al. 2019).
In the present study, a dam break analysis has been carried out using a hydraulic
model HEC-RAS for different failure mode to predict the breach flood, floodplain
mapping as suggested by Abhijith et al. (2017), Ackerman et al. (2008), Xiong
(2011).

11.2 Study Area

Hidkal dam is selected for dam break analysis as no such studies have been reported
on it. It is constructed across Ghataprabha River, and is located in Hidkal village,
Belagavi District, Karnataka, India. The Hidkal dam has a main section of length
4481 m and total length is 10,183 m. The spillway section is of masonry structure
whereas the remaining section is of earthen embankment. The height of the dam is
53.34 m above the foundation and it has a storage capacity of 51 TMC ft irrigating
an area of 317,435.41 ha.

11.3 Methodology

In the present study, hydraulic model HEC-RAS has been used for dam break analysis
and extraction of river geometry data and floodplain mapping has been done by
HEC-GeoRAS. HEC-RAS uses St. Venant’s equations for flood routing for unsteady
flow (Ackerman and Brunner 2008; Derdous et al. 2015). Figure 11.1 represents the
workflow processes for dam break analysis and Flood Plain mapping using HEC-RAS
and HEC-GeoRAS.
11 Dam Break Flood Routing and Inundation Mapping … 131

Fig. 11.1 Workflow

processes for dam break
analysis and flood plain
mapping

Data requirement and processing

The following data are used for the analysis:
• Salient features of the dam in study reach of the river.
• Probable Maximum Flood.
• Elevation-storage/area relationship of the reservoir.
• Cross sections of the river from dam site to the most downstream location of
interest.
• Manning’s roughness coefficient for the river.
• Breach Data.
• Boundary conditions.
Salient features of the dam, probable maximum flood and elevation-storage/area
relationship of the reservoir are collected from dam authority. River geometry is
extracted with the help of HEC-GeoRAS tool using DEM data (Source Earth Explorer
Website ([Link]) 32 m resolution). Through visual inspection of the
site using web imagery, the manning’s roughness coefficients are chosen based on
land use characteristics (Bharath 2020a, b). The breach parameters are final bottom
breach width, depth, side slope and formation time. Froehlich (2008) regression
equations are used to determine the breach parameters (Bharath 2020a, b). Boundary
conditions greatly influences the downstream floodwater, hence they must be care-
fully selected and defined. In this analysis, the boundary conditions defined are; flow
hydrograph (i.e., PMF) at upstream, normal depth at downstream and time-series of
gate opening at inline structure.
132 A. Bharath et al.

Legend
680
EG 03JUL2018 1000
WS 03JUL2018 1000

670 Ground

Bank Sta

660
Elevation (m)

650

640

630

620

610
0 1000 2000 3000 4000 5000

Station (m)

Fig. 11.2 Dam section at the end of the breach simulation

Plan: p3 River: Ghataprabha Reach: Reach 1 RS: 23304.79

670 80000
Legend

70000 Stage HW
660
Stage TW
60000
Flow
650
50000 Gate Flow - Radial_gate

Flow (m3/s)
Gate Flow - Sluice
Stage (m)

640 40000

30000
630

20000

620
10000

610 0
0600 1200 1800 2400 0600 1200 1800 2400 0600 1200
01Jul2018 02Jul2018 03Jul2018
Time

Fig. 11.3 Stage-flow hydrograph at the inline structure for piping failure

11.4 Results and Discussions

The dam break analysis is performed for unsteady flow condition for piping and over-
topping failure considering reservoir full condition. The river geometry is extracted
for 22 km stretch in the downstream; the cross section is extracted and interpolated
to a closer interval of 150–200 m. The simulation is done considering that the PMF
starts at 1 July 2018 at 9 am and ends at 3 July 2018 at 10 am and hence results
obtained are corresponding to the simulation period.
Figure 11.2 shows the dam profile at the end of the simulation, it shows the breach
formed on the dam and depleted water level at the end of the simulation.
Figures 11.3 and 11.4 show the headwater and tail water stage hydrograph, flood
hydrograph and discharge through the gates at the inline structure for piping and
11 Dam Break Flood Routing and Inundation Mapping … 133

Plan: P_OT River: Ghataprabha Reach: Reach 1 RS: 23304.79

670 80000 Legend

Stage HW
660
Stage TW
60000
Flow
650
Gate Flow - Radial_gate

Flow (m3/s)
Gate Flow - Sluice
Stage (m)

640 40000

630
20000
620

610 0
0600 1200 1800 2400 0600 1200 1800 2400 0600 1200
01Jul2018 02Jul2018 03Jul2018
Time

Fig. 11.4 Stage-flow hydrograph at the inline structure for overtopping failure

overtopping failure modes, respectively. Table 11.1 shows the maximum values of
headwater stage, tailwater stage and flow rate and their arrival time.
Flood inundation maps
Floodplain maps are produced in HEC-GeoRAS platform using HEC-RAS results.
Figure 11.5 shows the area inundated due to piping and overtopping mode of failure.

Table 11.1 Maximum stage-flow details

Particular Piping failure Overtopping
Maximum Time at maximum Maximum Time at maximum
HW Stage (m) 665.91 02-07-18 20:14 667.85 02-07-18 18:40
TW Stage (m) 646.51 02-07-18 23:48 651.17 02-07-18 21:14
Flow (m3 /s) 72,020.57 02-07-18 23:36 78,384.21 02-07-18 21:50

Fig. 11.5 Flood inundation map for a Piping failure, b Overtopping failure
134 A. Bharath et al.

It is found that the area of inundation is 74.32 km2 and 78.19 km2 due to piping and
overtopping failure, respectively.
Figure 11.5 depicts the major villages that gets affected due to dam failure. Totally,
twenty villages get affected among which seventeen are severely affected. These
villages have a great threat of loss of life and property. The majority of the area that
is being affected is agricultural land hence there may be a huge loss of crops which
in turn may affect the economy.
Flow variation at different river stations
Flood hydrographs at different downstream stations (downstream of the dam, 5, 10,
15, 20 and 22 km away from the dam) are compared as shown in Figs. 11.6 and 11.7.
The results obtained depicts that the peak flow value is very high at the station closer
to the dam and it gradually decreases as the distance increases. The kinks formed

80000
Q-DS

60000 Q-5km
Flow (m3/s)

Q-10km
40000 Q-15km

Q-20km
20000
Q-22km

0
7-2-18 4:48 PM 7-2-18 9:36 PM 7-3-18 2:24 AM 7-3-18 7:12 AM 7-3-18 12:00 PM
Time

Fig. 11.6 Flood hydrographs for piping failure

100000
Q-DS
80000
Q-5km
Flow (m3/s)

60000 Q-10km

Q-15km
40000
Q-20km
20000 Q-22km

0
7-2-18 2:24 PM 7-2-18 7:12 PM 7-3-18 12:00 AM 7-3-18 4:48 AM 7-3-18 9:36 AM
Time

Fig. 11.7 Flood hydrographs for overtopping failure

11 Dam Break Flood Routing and Inundation Mapping … 135

in the peak flood hydrographs are may be due to surge created in the flow due to
meandering of streams and valley cross section.
Stage variation at different river stations
Stage hydrographs at different stations (upstream & downstream of the dam, 5, 10,
15, 20 and 22 km away from the dam) are compared as shown in Figs. 11.8 and
11.9. The results obtained depicts that the peak stage value is very high at the station
closer to the dam and it gradually decreases as the distance increases. The flow depth
is found to vary from 30.39 m to 34.98 m for piping failure and 34.98 m to 13.86 m

670

660

650
Stage (m)

640

630

620

610
7-1-18 9:36 AM 7-1-18 9:36 PM 7-2-18 9:36 AM 7-2-18 9:36 PM 7-3-18 9:36 AM
Time

Stage HW stage-DS stage-5km stage-10km

stage-15km stage-20km stage-22km

Fig. 11.8 Stage hydrographs for piping failure

670

660

650
Stage (m)

640

630

620

610
7-1-18 12:00 PM 7-2-18 12:00 AM 7-2-18 12:00 PM 7-3-18 12:00 AM 7-3-18 12:00 PM
Time
Stage HW stage-DS stage-5km stage-10km
stage-15km stage-20km stage-22km

Fig. 11.9 Stage hydrographs for overtopping failure

136 A. Bharath et al.

100000
Piping
80000 Overtopping

Flow (m3/s) 60000

40000

20000

0
DS 5km 10km 15km 20km 22km
River stations

Fig. 11.10 Peak flow magnitudes at different river stations

for overtopping failure at the downstream of the dam and the end of river reach,
respectively. The kinks formed in the stage hydrographs at the downstream station
are may be due to surge formed behind the dam during failure.
Figure 11.10 highlights the peak flow values of piping and overtopping failures
at different river stations. This comparison depicts that peak flow due to overtopping
failure is higher than that of piping failure at all the river station hence overtopping
is considered as the most critical failure mode for dam break analysis.

11.5 Conclusions

Dam break model is simulated for the worse scenario, i.e., unsteady flow condi-
tion (PMF) and reservoir full condition for both piping and overtopping failure
considering reservoir full condition. HEC-RAS gives an easy platform to route the
dam breach flood along the downstream of the river and helps in determining the
hydraulic conditions at critical locations and different river stations. Floodplain maps
are generated in HEC-GeoRAS platform using HEC-RAS results. The results show
that the area of inundation is 74.32 km2 and 78.19 km2 due to piping and overtopping
failure, respectively. Totally, twenty villages get affected among which seventeen
are severely affected. The obtained results can be utilized in planning dam safety
measures, flood mitigation measures, land use planning in the downstream and to
develop an emergency response plan during the flood.
Limitations of the study
DEM data used for generating the cross section data was of coarser resolution which
affects the accuracy of results given from the model hence DEM of finer resolution or
cross sections from field survey should be used. HEC-RAS one-dimensional model
was used which gives one-dimensional results hence it will not work where multi-
dimensional modeling is required. HEC-RAS model assumes flow is perpendicular
to the cross sections which might result in improper inundation area when the flow
increases.
11 Dam Break Flood Routing and Inundation Mapping … 137

Acknowledgements The authors would like to take this opportunity to recognize the efforts and
assistance from every individual on all aspects in the preparation of this paper.

References

Abhijith R, Amrutha G, Vijayaraj G, Rijisha TV (2017) Dam break analysis of Idukki Dam using
HEC RAS. Int Res J Eng Technol (IRJET) 4(7):3410–3415
Ackerman CT, Brunner GW (2008) Dam failure analysis using HEC-RAS and HEC-GeoRAS. In
World environmental and water resources congress 2008: Ahupua’A, 8
Ackerman CT, Fleming MJ, Brunner GW (2008) Hydrologic and hydraulic models for performing
dam break studies. In :World environmental and water resources congress 2008: Ahupua’a—
Proceedings of the world environmental and water resources congress 2008, vol 316, pp 1–11.
Retrieved from [Link]
Bharath A, Manjunatha M, Tangadagi RB, Preethi S (2020a) J Green Eng (JGE). Environmental
Assessment for Rainwater Harvesting at GITAM Campus 10(4):1776–1785
Bharath A, Preethi S, Manjunatha M, Tangadagi RB (2020b) Prediction of temperature data for
Ghataprabha Sub- 26:140–144
Derdous O, Djemili L, Bouchehed H, Tachi SE (2015) A GIS based approach for the prediction of
the dam break flood hazard—A case study of Zardezas reservoir ‘skikda, Algeria’. J Water Land
Dev 27(1):15–20. Retrieved from [Link]
Froehlich DC (2008) And their uncertainties. Environ Prot 134:1708–1721
Khosravi K, Rostaminejad M, Cooper JR, Mao L, Melesse AM (2019) Dam break analysis and
flood inundation mapping: The case study of Sefid-Roud Dam, Iran. In: Extreme hydrology and
climate variability: monitoring, modelling, adaptation and mitigation. Elsevier Inc. Retrieved
from [Link]
Kulkarni SR, Ukarande SK, Jagtap S (2016) Dam break analysis—a case study. Int J Eng Res
5(Special 1):207–209
Mohamed MMA (2018) Overtopping breach peak outflow approximation of embankment dam
by using Monte Carlo method. Beni-Suef Univ J Basic Appl Sci 7(4):724–732. Retrieved from
[Link]
USACE (2016) HEC-RAS river analysis system hydraulic reference manual version 5.0. Hydrologic
Engineering Center, p 547
Wang B, Zhang T, Zhou Q, Wu C, Chen YL, Wu P (2015) A case study of the Tangjiashan land-
slide dam-break. J Hydrodyn 27(2):223–233. Retrieved from [Link]
8(15)60476-0
Xiong Y (2011) A dam break analysis using HEC-RAS. J Water Resour Prot 3(06):370–379.
Retrieved from [Link]
Zhang L, Xu Y, Jia JS (2009) Analysis of earth dam failures: a database approach. Georisk 3(3):184–
189. Retrieved from [Link]
Chapter 12
Suitability and Performance of Present
Irrigation System in Kokernag, Jammu
and Kashmir

Dar Himayoun, Roshni Thendiyath, and Jahangeer Saleem

Abstract This study deals with the suitability and performance of the irrigation
system in the Kokernag, a hilly and mountainous area in the state of Jammu and
Kashmir. Based on the variation in the soil texture, soil depth and output of paddy
fields, the whole study area was divided into two zones, namely zone A, covering the
middle portion of the Kokernag valley constituting the main paddy fields and zone B,
covering the area around the nallah Bringi. The suitability of this irrigation system
was assessed using parametric approach. This parametric evaluation was based on
soil texture, soil depth, slope and drainage of that area. These factors were rated
(0–100) based on their characteristics and importance for surface irrigation. Then,
capability index was developed for each zone using the rating of above mentioned
factors. On comparing this capability index with various suitability classes, it was
found that the zone A was highly suitable, whereas zone B was marginally suitable
for surface irrigation system. Further, the suitability analysis was carried on the
basis of the availability of water and socio-economic background of that area. It
was concluded that the existing system of irrigation is suitable in both the zones.
The performance was evolved using physical and financial performance indicators
in both the zones. Based on these performance indicators, the irrigation system in
zone A was performing better than that of zone B.

Keywords Surface irrigation · Nallah Bringi · Capability index · Suitability and

performance · Parametric evaluation

D. Himayoun (B) · R. Thendiyath

Department of Civil Engineering, National Institute of Technology Patna, Patna, India
R. Thendiyath
e-mail: roshni@[Link]
J. Saleem
Department of Civil Engineering, National Institute of Technology Srinagar, Srinagar, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 139
R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
140 D. Himayoun et al.

12.1 Introduction

Plants like human beings require air and water for their survival. Water is naturally
supplied to these plants by nature through direct rain or through flood water. But
there are some places where the distribution of rain and flood water is not uniform,
in that case we need some artificial measures to feed these crops and this technique
of applying water to the crops artificially is called irrigation.
A suitable method of irrigation is necessary for efficient use of water, as irrigation
is considered as a vehicle for economic and agricultural development. It also plays
a vital role in employment creation, improved nutrition, income generation, food
security and raising the living standard of farmers. There is scarcity of fresh water
resources in the world and irrigation is the major user of the fresh water supplies.
With limited resources and increasing competition for these resources, there should
be optimal utilisation of irrigation water (Elshaikh et al. 2018). How best an irrigation
system performs its designated function, may depend on various variables such as
climatic conditions, price, infrastructure design, socio-economic settings, availability
of inputs and more important the management of the irrigation system (Ahmad et al.
2017; Elshaikh et al. 2018; Fang et al. 2018; Ghahroodi et al. 2015; Moreno et al.
2016; Salahou et al. 2018).
Kashmir valley has vast perennial fresh water sources. These resources have been
utilised by mankind right from primitive time for various purposes. For irrigation
purposes the sources of water supply are springs, nallahs, ground water, river trib-
utaries and local streams. Due to availability of its vast fertile land, Kokernag has
a very good potential of agricultural production. Amongst all the tehsils of district
Anantnag in Jammu and Kashmir, Kokernag has the highest agricultural production.
There are various crops grown in this area in kharief season like paddy, maize, pulses,
fodder and other vegetables. Amongst them paddy is the main crop for which irriga-
tion is needed. The irrigation employed there is direct irrigation method. In order to
take care of diurnal variations in direct method of irrigation, some pondage is created
by constructing a weir or barrage across the river. This helps in rising the water level
in the river, then water is directed into the off taking canal which flows under gravity.
The method of application of irrigation water is contour basin type. Since for paddy
crop, ponding of water is needed up to around five centimetre from ground surface,
which on sloping land is very difficult that is why the fields in the Kokernag area are
contour basin type.
The present irrigation process in this area is surface irrigation gravity flow type.
This method involves more losses and hence is less efficient as compared to other
irrigation method like drip irrigation or sprinkler irrigation. In order to increase
the output of its fertile lands, there is considerable need to increase the efficiency
of its network of irrigation canals. So that this area comes to the position of its
self-sufficiency.
Evaluation of suitability of surface irrigation on the basis of parametric approach
depends on topographical characteristics like slope, physical properties of soil and
drainage properties of that area. This suitability can also be discussed on the basis of
12 Suitability and Performance of Present Irrigation System … 141

socio-economic background and water availability in that area (Alonso et al. 2019;
Antonio et al. 2018; Dang et al. 2018; Morris et al. 2015; Repository 2005; Usman
et al. 2015). For proper use of water through irrigation process, it is also important
to assess the performance of that irrigation scheme. Performance evaluation can be
done using physical performance indicators and financial performance indicators.
The aim of this study is to evaluate suitability and performance of irrigation system
in Kokernag using parametric approach and performance indicators, respectively.
Evaluation of suitability and performance of irrigation system is needed to achieve
an increase in overall efficiency and to improve water management practises. This
study contributes to the available literature by presenting the aggregated diagnosis
of suitability and performance of present irrigation system to the decision-makers.
This is useful to make them aware of the overall performance and steps to be taken
on achieving sustainable balance between irrigation management and environmental
preservation for an irrigation scheme.

12.2 Methodology

The methodology for evaluating suitability of this surface irrigation system in the
study area was discussed on the basis of three approaches.

12.2.1 Parametric Approach

In the parametric evaluation approach, land is evaluated on the basis of numerical

indexes (Albaji et al. 2001; Dengiz et al. 2006). This evaluation approach is based
on physical and chemical properties of soil. In this approach, soil is classified by
rating the land characteristic from 0 to 100 and comparing with the tables of soil
requirement. These ratings are used to calculate the multiplicative index of the soil.
Another classification of land resources is based on land capability. This classification
depends on climatic parameters, soil and topography and does not take into account
socio-economic conditions and yield. In this classification, negative features of land
are highlighted, whilst allotting capability classes to the lands of different types.

12.2.2 On the Basis of Socio-Economic Background

Socio-economic status of a society is the combined measure of both economic and

sociological position in relation to the other society based on the income, educa-
tion and occupation. Here in this area, most of the families are of lower class and
agriculture is their main source of income. Most of the people are unskilled and
142 D. Himayoun et al.

unemployed. The idea of taking this factor into account was that the surface irri-
gation scheme require large number of labours and they need not be skilled. To
some extent, this irrigation system solves the unemployment problem there, since
there are large number of unskilled labours available. Also the other systems like-
drip irrigation system, sprinkler irrigation system and sub-surface irrigation system
require high-initial cost, skilled labour, more electricity and proper management and
maintenance, which in majority of cases is not available in this area.

12.2.3 On the Basis of Usage and Availability of Water

Type of crop affects the choice of selecting a particular irrigation method since
different crops require different amount of water for their growth. Paddy and sugar-
cane require a large quantity of water as compared to other crops. For paddy, crop
ponding of water has to done in the field continuously for about seventy to eighty
days, this requirement is cheaply fulfilled by surface irrigation if there is abundance
of water available. Here in the present case, the availability of water is much more
than the total requirement there. Since in surface irrigation, there are more losses as
compared to other methods of irrigation that means choice of selecting a particular
type of irrigation also depends upon the water availability. If there is plenty of water
available then we can go for surface irrigation and if there scarcity of water, then we
have to go for other methods of irrigation that are more conservative.

12.3 Suitability and Performance Indicators

12.3.1 Capability Index

The capability index (CI) for surface irrigation was calculated on the basis of
Eq. (12.1) given below

A×B C D
CI = × × (12.1)
100 100 100

where ‘A’ is the soil texture rating, ‘B’ is the soil depth rating, ‘C’ is the slope rating
and ‘D’ is the drainage rating. The ranges of capability index and the corresponding
suitability classes are presented in Table 12.1.
12 Suitability and Performance of Present Irrigation System … 143

Table 12.1 Suitability

Capability index Class Suitability class Symbol
classes for various capability
indices >80 1 Highly suitable S1
60–80 2 Moderately S2
suitable
45–60 3 Marginally suitable S3
30–45 4 Currently not N1
suitable
<30 5 Permanently not N2
suitable

12.3.2 Relative Water Supply (RWS)

RWS is a measure of adequacy as well as seasonal timeliness of an irrigation system.

It is a dimensionless parameter and is defined as the ratio of total water supply to the
total demand at scheme level (Eq. 12.2). The numerator is the irrigation water supply
and the denominator consists of non-beneficial evapotranspiration (ET), losses to
drains, consumptive use and net flow towards groundwater.

(Irrigation water)
RWS = (12.2)
(ET + Seepage + Percolation)

There are three ranges for RWS—low RWS (<1), medium RWS (1–2) and high
RWS (>2). In low RWS, the water supply is below actual requirements, requires strict
control and higher degrees of management. In medium RWS, a variety of operational
strategies can be successfully employed, moderate amounts of management effort
are needed and high RWS has more likely the farmer controlled demand schedules
and greatest flexibility in operation.

12.3.3 Benefit—Cost Ratio (BCR)

The financial performance of irrigation system is evaluated on the basis of benefit—

cost ratio (Eq. 12.3). This indicator gives the relation between the total input and
total output in terms rupees. The ratio should be greater than one for all irrigation
systems. The benefit—cost ratio was worked out for the irrigation system in both the
zones and it was greater than one for both the zones, zone ‘A’ having higher valve.

Output cost
Benifit − Cost ratio = (12.3)
Input cost
144 D. Himayoun et al.

12.4 Description of Study Area

12.4.1 Location

Kokernag is a sub-district town and a notified area in Breng Valley. This area lies in
district Anantnag in state of Jammu & Kashmir. Kokernag is located towards southern
part of the Kashmir valley in the foothills of famous Pir-Panjal mountain range and is
at a distance of 25 km from Anantnag and about 80 km from Srinagar city. Kokernag
area is the catchment area of nallah ‘Bringi’. Bringi is one of the tributaries of the
river Jhelum, hence this catchment is considered as one of the upland catchments
of Jhelum river. Bringi catchment lies between 33.20° N and 33.45° N latitude and
75.10° E and 75.30° E longitude covering a total area of 595 km2 . The elevation
varies from 1650 to 4000 m above mean sea level. Figure 12.1, shows the location
map of study area.

Fig. 12.1 Location map of the study area

12 Suitability and Performance of Present Irrigation System … 145

12.4.2 Topography

Kokernag is hilly and mountainous towards the southeast and southwest with broad
inter-mountain valley on the northern side. The altitude of the hill ranges up to 3000 m
above mean sea level. The valley in the central part has mildly undulating topography
with its elevation ranging from 1800 to 2000 m above mean sea level. The soils are
fertile mostly silty and sandy silt in texture and the main crop grown is paddy. The
land features of the study area are presented in Table 12.2.

12.4.3 Precipitation Characteristics

The rainfall characteristics in Jammu and Kashmir are different from rest of India. In
India, most of the rainfall (about 75%) occurs in monsoon season, that is, from June
to October. In Jammu and Kashmir most of the rainfall occurs in winter season and
spring season, that is, from January to April. Table 12.3, shows the 12 year averaged
monthly rainfall at station Kokernag.

Table 12.2 Land features of the study area

Irrigated area Unirrigated area (ha) Orchards Grazing area Total
(ha) (ha) (ha) (ha)
1627 2015 843 1114 5609

Table 12.3 Average monthly

Month Rainfall in mm
rainfall at station Kokernag
January 64
February 125
March 113
April 106
May 92
June 71
July 85
August 69
September 66
October 25
November 26
December 54
Total 895
146 D. Himayoun et al.

12.4.4 Irrigation Details

[Link] Source of Irrigation

Nallah Bringi is the main source of irrigation in Kokernag. It is located in district

Anantnag of Jammu and Kashmir and starts from Wandevalgam village which is
located at a distance of 30 km from Anantnag main town and ends near Larkipora
village which is located in the out skirts of Anantnag. Bringi catches snowmelt water
and becomes active during summer season when there is high-intensity rainfall in the
upper catchments. The average discharge in the Bringi nallah varies from 12 to 75
cumecs. This is sufficient to meet the requirement of that area. The weekly discharge
variation in Bringi nallah is presented in Fig. 12.2.

[Link] Type of Irrigation

The irrigation scheme adopted in Kokernag is direct irrigation method flowing under
gravity. The method of application of irrigation water is contour basin type. Since
for paddy crop, ponding of water is needed up to around five centimetre from ground
surface which on sloping land is very difficult that is why the fields there are contour
basin type and hence the need of contour basin irrigation.

[Link] Types of Crops Grown

There are various crops grown in this area in kharief season like—paddy, maize,
pulses, fodder and other vegetables. Among them paddy is the main crop for which
irrigation is done. Table 12.4, shows the area covered by the crops grown in the study
area.
The total irrigated area in Kokernag is around 1627 ha. Out of this, 1499 hectares
of area is under paddy crop, whilst remaining is under the rest of crops grown there.

Fig. 12.2 Weekly variation in Bringi discharge

12 Suitability and Performance of Present Irrigation System … 147

Table 12.4 Crops grown in the study area

Crop Paddy Pulses Maize Vegetables Fodder
Area in ha 1499 74 1855 129 56

Fig. 12.3 Irrigation requirement in Kharief season

Figure 12.3, shows the total irrigation requirement in kharief season in Kokernag
area.
The water requirement for paddy crop starts early in the month April for the devel-
opment of nurseries for paddy crop, but that is in small quantities. It is about one-tenth
of that required for normal paddy crop after transplantation. After transplantation,
the water requirement rises and is at peak during the month of June and July. Then,
water requirement starts decreasing and reaches zero at the mid of September.

12.5 Data Collection

The data was collected from the various departments like agricultural department
division Achabal Anantnag, irrigation department division Dooru Anantnag, water-
shed management department Anantnag and Indian meteorological department
Srinagar. Due to non-availability of some data in these departments, some data
was collected locally and some on interview with different farmers. The different
types of data collected are; soil data, topographical data, crop production data,
water availability and requirement data and some data related to socio- economical
background.
148 D. Himayoun et al.

12.5.1 Soil Data

This includes the information regarding the soil texture and soil depth. Different
sites were chosen for this data requirement. In the middle areas of the Kokernag,
the soil depth varies from 100 to 200 cm, but in the area around the nallah Bringi,
it was ranging from 25 to 50 cm with an average of 40 cm. Beyond this soil cover
depth, there is mixture of gravel, sand, silt and some stones. For determination of soil
texture, two different soil samples were taken. One from Bringi region and another
from the main paddy field region, i.e. from middle of the Kokernag area. From sieve
analysis and hydrometric analysis, particle size distribution curves of both the soil
samples were drawn as shown in Figs. 12.4 and 12.5, respectively.

12.5.2 Water Availability and Water Requirement Data

The discharge data of the Bringi region and the water requirement for paddy crop
were obtained from the irrigation department Vailoo. Table 12.5, presents the average
weekly variation in discharge of Bringi at Wayul gauge, Table 12.6, presents the
monthly requirement of water for paddy Crop in middle fields of Kokernag and
Table 12.7, presents the monthly requirement of water for paddy crop in area around
nallah Bringi Kokernag.
Now from the above data, we see that the middle portion of the valley considerably
differ from the area around the nallah Bringi keeping in view the soil texture, soil
depth and water requirement in the fields. It is because of these variations that the
whole study area was divided into two zones: Zone A—this includes the area covering

Fig. 12.4 Particle size distribution curve for soils in middle fields of Kokernag
12 Suitability and Performance of Present Irrigation System … 149

Fig. 12.5 Particle size distribution curve for soils around Bringi zone

Table 12.5 Average weekly variation in discharge of Bringi at Wayul gauge

Month Discharge (m3 /s) Discharge (m3 /s) Discharge (m3 /s) Discharge (m3 /s)
(1st week) 2nd week 3rd week 4th week
April 14.82 14.79 19.78 23.52
May 26.7 29.07 27.43 27.29
June 20.51 19.3 18.36 17.54
July 18.02 16.38 16.09 18.76
August 16.24 17.09 14.71 14.16
September 13.43 12.1 – –

Table 12.6 Showing the

Month Delta for paddy in mm
monthly requirement of water
for paddy crop in middle January –
fields of Kokernag February –
March –
April –
May 200
June 300
July 300
August 225
September 150
October –
November –
December –
Total 1175
150 D. Himayoun et al.

Table 12.7 Showing the

Month Delta for paddy in mm
monthly requirement of water
for paddy crop in area around January –
nallah Bringi Kokernag February –
March –
April –
May 280
June 420
July 420
August 315
September 210
October –
November –
December –
Total 1645

the middle portion of the Kokernag valley that constitutes the main paddy fields.
This area is about the 90% (1350 ha) of the total paddy field area. And Zone B—this
includes the area around the Bringi stream constituting the Bringi paddy fields. This
covers an area of about 10% (150 ha) of the paddy field area.

12.6 Results and Discussions

12.6.1 Suitability

Based on the variation in soil texture, soil depth and output from the fields, the whole
study area was divided into zones. Zone A consists of the main paddy fields which lie
in the middle of the Kokernag valley. The other zone was the area which lie around
the Bringi Nallah named as zone B. Zone A constitute about 1350 ha of paddy field
area where as zone B constitute an area of around 150 ha of paddy land. In order
to determine the soil texture in these zones, soil samples were taken from both the
zones and particle size distribution curves of both the samples were drawn using sieve
analysis and hydrometric analysis. The soil from zone A was found to be silty soil
and the soil from the zone B was found to be sandy silt soil. Also the permeability
tests were done on the soil samples, taken from zone A and zone B, at optimum
moisture content. This could not represent the actual field conditions, but it gave us
relative permeability valves. The permeability in zone B was found to be twice that
of zone A. For calculating the soil depth in both the zones, some pits were dug up in
both the zones. In zone A, the average soil cover depth was found to be more than
100 cm and in zone B it was around 40 cm with minimum thickness of about 25 cm.
12 Suitability and Performance of Present Irrigation System … 151

Table 12.8 Suitability and

Soil type
performance indices
Zone A Zone B
Performance indices
RWS 1.6 1.71
BCR 1.83 1.4
Suitability index
CI 85.5 46

The crop in both the zones, for which suitability of irrigation system is to be evolved,
was chosen paddy crop and both the paddy field zones get irrigation from the same
source—‘Bringi stream’. Since the paddy fields require large quantities of water and
increases the chances of water logging but in both the sites there was no area where
these conditions could occur. It was because of natural slope present there which
could drain off these fields quickly and easily.
The suitability of this surface irrigation system was evolved using parametric
approach (Sys et al. 1991). Capability index for both the was calculated and based
on this index the land in zone A was found to be highly suitable for the surface
irrigation with the CI value of 85.5, whereas the land in zone B was found to be
marginally suitable for the same with the value of 46. Moreover, the type of irriga-
tion employed also depends on the factors like—capital cost, operating cost, ease of
use and efficiency potential. The surface irrigation system there reduces the oper-
ating costs of irrigation—principally electricity, which was cited as a major cause
of failure in that area. The surface irrigation system is the least expensive to imple-
ment, but does require more labour than the other systems. However, with the high
unemployment levels in the area, this had become a positive factor for that economy.
Also other systems of irrigation requires more skilled labours and the availability of
these was difficult in that area that is why the present irrigation system has become
easier to operate. A comparison was made between the crop water requirement and
availability of water there. It was found that the water available for irrigation during
peak demands in both the zones was much more than the requirement. The suitability
and performance indices of the irrigation system are presented in Table 12.8.

12.6.2 Performance

The performance of this irrigation system in these two zones was evaluated using
the physical performance indicators and financial performance indicator. Physical
performance indication was based on the relative water supply and irrigation water-
yield. The financial performance indication was based on the benefit–cost ratio.
In zone A, the relative water supply was around 1.6, this means the water supplied
there is higher than the water requirement. The reason behind this are various losses
152 D. Himayoun et al.

like-seepage, application losses and tail water losses. Based on the irrigation water-
yield indication, it was estimated that 1 kg of rice at an average requires 4500 L of
water which is very high as compared to the average requirement of about 2000–
3000 L based on the International Water Management Institute standard. For financial
performance, benefit–cost ratio was calculated, and its value was found around 1.83.
This means that the net return from the system is greater than the initial investment.
In zone B, the relative water supply was around 1.71 which is higher as compared
to that of zone A. The reason behind this is the presence of sand content in that soil,
which increases seepage and other losses. Based on the irrigation water yield, it was
estimated that 1 kg of rice requires about 9800 L of water which when compared
with that of zone A is very large. The possible reason for this high-water use has
been already discussed, that is, the presence of sandy soil there and less soil depth.
This high-water use increases the percolation losses and other seepage losses, and
decreases the output of this area. The financial performance in this zone, based on
benefit cost ratio, was about 1.41 that means output from this area is very low as
compared to that of zone A. The possible reason for this low output may be the thin
soil cover. Since the root zone depth is greater than the soil depth and the presence
of gravel and stones below the soil cover limit the potential growth of the paddy root
which ultimately affected the output of this area.

12.7 Conclusions

Although based on parametric approach, the land in zone A was highly suitable
whereas the land in zone B was marginally suitable for surface irrigation, but when
suitability was evaluated on the basis of socio-economic background and availability
of water in the study area, it was found that surface irrigation system is suitable in
both the zones. Also area under zone A (90%) is much more than zone B (10%) that
means the additional water losses that occur in zone B are negligible as compared to
the total water use in both the zones. Hence from the above study it was concluded that
the existing system of irrigation is suitable in the study area. Now for performance,
we have observed that the performance of irrigation system in zone A is better than
that of zone B because of low-percolation losses and higher yield in zone A. Also
the performance on the basis of relative water supply and irrigation water-yield is
very low in both the zones, but it is on the basis of benefit–cost ratio that makes the
performance acceptable in both the zones.

References

Ahmad E, Brar AS, Mishra SK, Singh KB (2017) Field crops research simulating response of wheat
to timing and depth of irrigation water in drip irrigation system using CERES-wheat model. Field
Crops Res 214(04):149–63. [Link]
12 Suitability and Performance of Present Irrigation System … 153

Albaji M, Boroomand Nasab S, Hemadi J (2001) Comparison of different irrigation methods

based on the parametric evaluation approach in West North Ahwaz Plain. Agric Water Manage
208(05):193–203. [Link]
Alonso A, Feltz N, Gaspart F, Sbaa M, Vanclooster M (2019) Comparative assessment of irrigation
systems’ performance : case study in the Tri Ff a agricultural district, NE Morocco. Agric Water
Manage 212(01):338–48. [Link]
Antonio D, Nicotra A, Mateos L, Marcello S (2018) Improvement of the irrigation performance
in water users associations integrating data envelopment analysis and multi-regression models.
Agric Water Manage 205(05):38–49. [Link]
Dang T, Pedroso R, Laux P, Kunstmann H (2018) Development of an integrated hydrological-
irrigation optimization modeling system for a typical rice irrigation scheme in Central Vietnam.
Agric Water Manage 208(05):193–203. [Link]
Dengiz O (2006) Comparison of different irrigation methods based on the parametric evaluation
approach. Agric Water Manage 30(3):21–29
Elshaikh AE, Jiao X, Yang SH (2018) Performance evaluation of irrigation projects: theories,
methods, and techniques. Agric Water Manage 203(01):87–96. [Link]
2018.02.034
Fang Q, Zhang X, Shao L, Chen S, Sun H (2018) Assessing the performance of different irrigation
systems on winter wheat under limited water supply. Agric Water Manage 196(02):133–43.
[Link]
Ghahroodi EM, Noory H, Liaghat AM (2015) Performance evaluation study and hydrologic and
productive analysis of irrigation systems at the Qazvin Irrigation Network (Iran). Agric Water
Manage 148(06):189–95. [Link]
Moreno MA, Castillo A, Montero J, Tarjuelo JM (2016) ScienceDirect optimisation of the design
of pressurised irrigation systems for irregular shaped plots. Biosyst Eng 151(05):361–73. https://
[Link]/10.1016/[Link].2016.10.005
Morris MR, Hussain A, Gillies MH, Halloran NJO (2015) Inflow rate and border irrigation
performance. Agric Water Manage 155(04):76–86. [Link]
Repository, Institutional (2005) Institutional repository performance assessment of irrigation water
management of heterogeneous irrigation schemes: 1. A framework for evaluation this item was
submitted to Loughborough’s Institutional Repository by the Author and Is Made Available Un
Salahou MK, Jiao X, Lü H (2018) Border irrigation performance with distance-based cut-off. Agric
Water Manage 201(01):27–37. Retrieved ([Link]
Usman M, Liedl R, Awan UK (2015) Spatio-temporal estimation of consumptive water use for
assessment of irrigation system performance and management of water resources in irrigated
Indus Basin, Pakistan. J Hydrol 51(05):361–73. [Link]
Chapter 13
Linking of Sediment Yield Pattern
with Rainfall and Land-Use Land-Cover
Changes Within Burhanpur
Sub-catchment, India

S. R. Resmi, P. L. Patel, and P. V. Timbadiya

Abstract In present study, the temporal changes in the suspended sediment concen-
tration and sediment yield pattern of Burhanpur sub-catchment, which is a part of
Upper Tapi basin, India, are investigated. The variation of sediment yield during
three decades, viz. 1980–1989, 1990–1999 and 2000–2010 is correlated with the
changes in the land-use land-cover and climatic pattern within the sub-catchment. The
changes in the pattern of monsoon rainfall intensity were investigated using modi-
fied Mann–Kendall (MMK) test while land-use land-cover within the sub-catchment
was investigated by classifying the decadal images of the sub-catchment. The anal-
ysis of observed data revealed that sediment concentration is very high, even for
small-to-moderate flows during the onset on monsoon, due to availability of uncon-
solidated sediments after a long dry period of October–May months. The monsoon
rainfall intensity has shown decreasing trend over the sub-catchment during the
period 1980–2010. Also, agricultural areas are found to be increasing with time at
the expense of forest cover, fallow and scrub lands. The waterbodies, in the form of
minor storage structures, are increased during the period 1990–2010. The excessive
trapping of sediments within the sub-catchment due to the development of minor
storage structures (waterbodies) was responsible for reduction of sediment yield
within the Burhanpur sub-catchment.

Keywords Sediment yield · Land-use land-cover · Modified Mann–Kendal test ·

Monsoonal rainfall intensity · Waterbodies

S. R. Resmi · P. L. Patel (B) · P. V. Timbadiya

Sardar Vallabhbhai National Institute of Technology Surat, Surat 395007, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 155
R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
156 S. R. Resmi et al.

13.1 Introduction

The geomorphic characteristics of a watershed affect the energy fluxes, mass move-
ment and dispersion of water and sediment within the watershed. Soil erosion induces
detachment and transport of sediments and is treated as one of the dominant geomor-
phic processes on earth’s land surface (Toy et al. 2002). The sediment is detached from
the soil surface by impacts of raindrops and by the shearing force of surface runoff
(Jain and Kothyari 2000). In case the sediments available for transport are greater than
transport capacity of flow, the sediment gets accumulated on the land surface (Trimble
1975; Toy et al. 2002). The amount of sediment that arrives at the outlet of a catch-
ment is termed as sediment yield (Glymph 1954; Parsons et al. 2006; Fryirs 2013;
Vanmaercke et al. 2014). This sediment yield within a catchment is determined by its
environmental conditions, such as climate, soil, topography, land use, other human
disturbances, which can affect the sediment supply, transport, storage, residence time
and the connectivity of sediment sources to the catchment outlet (Mingguo et al. 2008;
Fryirs 2013; Shi et al. 2013; Yan et al. 2013). The quantification of sediment yield
and its determinant factors is of utmost importance for sustainable management of
catchments (Parsons et al. 2006). Sediment yields are sensitive to the factors like
reservoir construction, land clearance and land-use changes as well as other forms
of land disturbances, such as mining activity, soil and water conservation measures,
sediment control programmes and climate change (Walling and Fang 2003). Human
activities and climate change have profound impacts on geomorphic characteristics
of catchments (Fryirs 2013). Human-induced activities, such as construction of dams
and reservoirs, decrease the sediment yield, whereas clearance of natural vegetation
and mining activities is responsible for reduction in the flow resistance over the soil
surfaces, thereby increasing the sediment yield. Land cover is one of the major factors
that reduces soil erosion due to improvement in soil structure, increase in the surface
roughness and infiltration rate, trapping the sediments and breaking the landscape
connectivity (Burylo et al. 2007; Hudek et al. 2010, Ouyang et al. 2010; Preti et al.
2011; Lü et al. 2012; Rey and Burylo 2014). Human land-use activities can funda-
mentally change the hydrogeomorphology of rivers (James and Lecce 2013). Such
changes in land use can have significant impact on the runoff, sediment delivery,
water quality, aquatic habitat and channel morphology. Human impacts on rivers
can be classified as ‘direct changes’, which occur within channels and the riparian
vegetation and ‘indirect changes’, which occur in the form of agricultural expan-
sion, urbanization, mining and changes in the vegetation. Apart from anthropogenic
changes within the catchment, the climatic factors like changes in rainfall pattern are
also responsible for changing the sediment yield pattern at the outlet of the catch-
ment. The present study is an attempt to quantify the changes in suspended sediment
concentration and sediment yield at Burhanpur stream gauging station over a period
of 30 years, i.e. 1980–2010. The study has been carried out to quantify changes on
decadal basis, i.e. during 1980–1989, 1990–1999 and 2000–2010. The variations in
sediment yield pattern due to changes in monsoonal rainfall intensity and land-use
land-cover are linked together for Upper Tapi sub-basin (Burhanpur sub-catchment).
13 Linking of Sediment Yield Pattern with Rainfall and Land-Use … 157

The study contributes significantly in understanding of the effects of climatic and

land-use land-cover changes on sediment yield pattern within the sub-catchment.

13.2 Study Area

The Tapi river is the second largest west flowing river in India, after Narmada river.
Originating from the Tapi talab near Multai in Betul district of Madhya Pradesh, the
river traverses around 784 km till it joins the Arabian Sea at the Gulf of Cambay
in Gujarat. The study area of interest in present study, Burhanpur sub-catchment of
the Upper Tapi basin, is located between latitudes 21° 00 00 N to 22° 00 00 N
and longitudes 75° 45 00 E to 78° 30 00 E and drains an area of 10,612 km2
to the Hatnur reservoir (Fig. 13.1). The Upper Tapi river, which has a length of
around 350 km upto Hatnur dam, is perennial in nature and has two stream gauging
stations, viz. Dedtalai and Burhanpur, located at CH: 236.34 km and CH: 307.89 km,
respectively. The main Tapi river is monsoon dominated, with significant contribu-
tion from south-west monsoon season that starts in mid-June and subsides in the
mid-October. The July and August months are the rainiest months of the year, which
contributes to around 50% of the total rainfall. The monsoon period is followed by a

Fig. 13.1 Study area: Burhanpur sub-catchment

158 S. R. Resmi et al.

long dry season, which spans over 8 months (October–May). There are seven rain-
gauge stations in the sub-catchment, which are operated and maintained by India
Meteorological Department (IMD). The Upper Tapi river passes through a variety
of topography, such as gullied ravenous terrains, hilly areas and alluvial plains.
The Upper Tapi basin (Burhanpur sub-catchment) stationed at an average altitude of
487 m from the mean sea level (MSL) with an average gradient of 6.26°, encompasses
main Tapi river of 6th order. The Upper Tapi river lies in close proximity to Satpura
ranges (average distance = 38 km), accordingly, northern tributaries lie in steep
sloped areas, thereby exhibiting flashy streamflow characteristics. The Burhanpur
sub-catchment is predominantly occupied by Deccan trap basaltic lava formations,
followed by quaternary alluvial deposits towards the downstream of Dedtalai stream
gauging station. Also, rock formations like granite-gneiss, quartz mica and sand-
stone deposits are present in minor quantities, especially, in the north-eastern part
of the sub-catchment near the Betul region. The Deccan trap basaltic lava forma-
tion, granite-gneiss deposits and sandstone traces are the main sources of coarser
sediments, viz. cobbles and boulders, in the headwater region of the sub-catchment,
whereas the quaternary alluvial deposits in Burhanpur and Purna sub-catchments
contribute to the silt and clay sediments in the river, especially, during the monsoon
period (Hire 1999).

13.3 Methodology

The steps followed in quantifying the temporal sediment yield pattern within the
Burhanpur sub-catchment are described in Fig. 13.2.

13.4 Analysis of Data, Results and Discussions

The observed suspended sediment concentrations at Burhanpur stream gauging

station during 1980–2010 were used to obtain the sediment load and sediment yield.
The analysis was performed on a decadal basis, i.e. 1980–1989, 1990–1999 and
2000–2010.
Preliminary inspection of the data reveals that, more than 99% of the sediment
load is transported during the monsoon months (June–October) at Burhanpur stream
gauging station. Accordingly, in present study, the monsoonal period has been taken
into consideration for further analyses. For approximately 92% of the time, the
concentration of suspended sediments is below 4 g/L at Burhanpur stream gauging
station. This points to the fact that there is no adequate source of unconsolidated
sediments in the catchment area upstream of the station. The presence of rocky bed
stratum, upstream of Burhanpur stream gauging station, is responsible for scarcity
of fine-grained material in the stream flow.
13 Linking of Sediment Yield Pattern with Rainfall and Land-Use … 159

Data Input

Suspended Daily Daily

sediment rainfall discharge
concentration

Preparation of time series of sediment load, sediment yield,

rainfall intensity and total discharge at Burhanpur station

Preliminary evaluation of the general characteristics of suspended sediment

concentration in Burhanpur sub-catchment (Burhanpur station)

Evaluating the spatio-temporal variation in total sediment yield along

with the variation in rainfall intensity during monsoon period

Performing trend analysis on rainfall intensity for IMD raingauge

stations in Burhanpur sub-catchment during 1980-2010

Performing Level-I Supervised Land-use land-cover (LULC)

classification for Burhanpur sub-catchment and generation of maps for
the decades 1980-1989, 1990-1999 and 2000-2010

Evaluate and discuss the variations in sediment

yield pattern in relation to LULC changes and
rainfall trend

Fig. 13.2 Methodology adopted in present study

Table 13.1 summarizes the primary characteristics of the suspended sediment

concentration and suspended sediment load at the outlet of Burhanpur sub-catchment
of Upper Tapi river. The mean suspended sediment concentration values were
observed to be 0.32 g/L during 1980–1989, 0.33 g/L during 1990–1999 and 0.24 g/L
during 2000–2010.
The mean suspended sediment concentration values were observed to be
decreasing over the decades. The maximum recorded suspended sediment concen-
tration was 9.01 g/L during the period 1980–1989, which was in the year 1988,
corresponding to a discharge of 4707 m3 /s and a rainfall of 85 mm (22/07/1988).
160 S. R. Resmi et al.

Table 13.1 Statistical characteristics of the suspended sediment at Burhanpur stream gauging
station
Period Suspended sediment Suspended sediment Average annual
concentration (g/L) load (t/day) sediment
2
Mean Min Max Mean Min Max (10 ) yield (t/km /day)
6

1980–1989 0.32 0.001 9.01 31,168 0 3.65 3.21

1990–1999 0.33 0 12.72 48,878 0 9.04 5.29
2000–2010 0.24 0 3.92 21,773 0 5.55 2.33

Similarly, during 1990–1999, a maximum observed suspended sediment concen-

tration was 12.72 g/L on 16/07/1993, corresponding to a discharge of 8268 m3 /s
and a rainfall of 126 mm over the sub-catchment. This indicates clearly that daily
suspended sediment concentration is maximum during the onset of monsoon period,
due to availability of unconsolidated sediment deposits, after a long dry span of eight
months, i.e. from October to May in the sub-catchment. Even though the sediment
sources are limited in Burhanpur sub-catchment due to its rocky topography in the
upstream reaches, higher relief and steeper gradients are responsible for generating
geomorphically effective flows for transporting the sediments downstream.
For Burhanpur sub-catchment, the average annual rainfall intensity (RI) during
1980–2010 was 20.3 mm/day for an average sediment yield of 3.57 t/km2 /day.
The average sediment yield for the period 1980–1989 was 3.21 t/km2 /day, with
the higher yields obtained in years 1983 (RI = 19.3 mm/day) and 1988 (RI =
22.1 mm/day). In the second decade, the average sediment yield observed was
very high (5.29 t/km2 /day). The average sediment yield during 1990–1994 was
8.19 t/km2 /day, wherein RI during monsoon season ranges between 21 mm/day
and 25.2 mm/day. On the other hand, during 2000–2010, for a rainfall intensity of
20.6 mm/day, average sediment was reported to be 2.33 t/km2 /day. While comparing
Fig. 13.3a–c, it is seen that, in general, there is a reduction in the sediment yield from
Burhanpur sub-catchment with effect from year 1994, with an average sediment yield
of 2.40 t/km2 /day and 2.33 t/km2 /day during the period 1995–1999 and 2000–2010,
respectively.

Fig. 13.3 Decadal variation in total sediment yield and rainfall intensity during monsoon period
at the outlet of Burhanpur sub-catchment
13 Linking of Sediment Yield Pattern with Rainfall and Land-Use … 161

Fig. 13.4 Trend analysis results of rainfall intensity for Burhanpur sub-catchment

The pattern of total sediment yield was analysed while performing trend analysis
of monsoonal rainfall intensity for seven raingauge stations in the sub-catchment
using modified Mann–Kendall (MMK) test during the period 1980–2010. The
results revealed that the monsoonal rainfall intensity in the sub-catchment exhibited
decreasing trend during 1980–2010, especially, at stations Burhanpur and Raver,
which fall in the alluvial reaches of the sub-catchment (Fig. 13.4).
The land-use land-cover statistics showed that there has been an increase in the
agricultural area in the sub-catchment, at the expense of its forest cover (Fig. 13.5).
Forest cover plays a major role in holding the soil in place, thereby preventing
the soil erosion. The increase in agricultural activities like ploughing and tillage
disturbs the soil surface, thereby increasing the sediment delivery to the river.
However, this increasing effect is not observed in the Burhanpur sub-catchment.
The spatio-temporal variations of waterbodies within the sub-catchment were also
assessed and presented, along with the drainage network in Fig. 13.6. The analysis
revealed that the waterbodies have expanded considerably during 1980–2010, even
though they occupy only a small fraction of the total catchment area. The area of
waterbodies existed within Burhanpur sub-catchment is 6.83 km2 , 17.72 km2 and
20.05 km2 , respectively, during the decades 1980–1989, 1990–1999 and 2000–2010.
This implies that construction of minor storage dams, irrigation weirs and minor
check dams was responsible in trapping of sediments within the sub-catchment,
162 S. R. Resmi et al.

Fig. 13.5 Land-use land-cover pattern in Burhanpur sub-catchment

Fig. 13.6 Decadal variation in spatial distribution of waterbodies over the decades in Burhanpur
sub-catchment
13 Linking of Sediment Yield Pattern with Rainfall and Land-Use … 163

which has further culminated in the reduction of sediment yield at the outlet of
sub-catchment. Also, the reduction in fallow land and scrub lands, which are major
sources of unconsolidated sediments, has reduced over the decades and responsible
for reduction in the sediment yield.

13.5 Conclusions

The key findings from present study are summarized below:

(a) More than 99% of the sediment load is transported during monsoon months
(June–October), wherein, about 92% of the time, the concentration of
suspended sediments is below 4 g/L at Burhanpur stream gauging station.
(b) Concentration of the suspended sediment is maximum during the early
monsoon season, i.e. end of the June month and throughout the July month.
(c) The mean suspended sediment concentration values were computed as 0.32 g/L
during 1980–1989, 0.33 g/L during 1990–1999 and 0.24 g/L during 2000–
2010. However, due to the occurrence of high magnitude flows, the average
annual sediment yield was significantly higher during the period 1990–1999.
(d) The results of trend analysis of monsoonal rainfall indicated that Burhanpur
sub-catchment exhibited decreasing trend during the period 1980–2010. The
decreasing trend of rainfall with added influence of land-use land-cover
changes within the sub-catchment resulted in lesser sediment yield at the
catchment outlet in the recent decade.
(e) The land-use land-cover statistics showed that there has been an increase in the
agricultural area in the Burhanpur sub-catchment at the expense of its forest
cover, fallow and scrub land. The temporal–spatial distribution of waterbodies,
in the form of minor storage structures, were found to be increasing during the
period 1980–2010. The reduction of fallow and scrub land and increase in the
areas of waterbodies within the sub-catchment were responsible for reduction
in sediment yield at the outlet of the catchment during the decade 2000–2010.

Acknowledgements The authors are thankful to the Central Water Commission (CWC), Ministry
of Jal Shakti, Government of India for providing necessary financial support through project titled
“Morphological Study of Tapi River using Remote Sensing Technique” and data required for the
study. The authors are also thankful to Centre of Excellence (CoE) on ‘Water Resources and Flood
Management’, SVNIT-Surat under TEQIP-II for the resourceful help and support.
164 S. R. Resmi et al.

References

Burylo M, Rey F, Delcros P (2007) Abiotic and biotic factors influencing the early stages of vege-
tation colonization in restored marly gullies (Southern Alps, France). Ecol Eng 30(3):231–239
Fryirs K (2013) (Dis) Connectivity in catchment sediment cascades: a fresh look at the sediment
delivery problem. Earth Surf Proc Land 38(1):30–46
Glymph LM (1954) Studies of sediment yields from watersheds. Int Union Geod Geophys Int Assoc
Hydrol 10th Gen Assem Rome, Italy, Part 1:178–191
Hire SP (1999) Geomorphic and hydrologic studies of floods in the Tapi basin. Doctoral dissertation,
University of Pune, Pune, India
Hudek C, Burylo M, Rey F (2010) Root system traits of Mahonia aquifolium and its potential use
in soil reinforcement in mountain horticultural practices. Sci Hortic 125(3):504–511
Jain MK, Kothyari UC (2000) Estimation of soil erosion and sediment yield using GIS. Hydrol Sci
J 45(5):771–786
James LA, Lecce SA (2013) Impacts of land-use and land-cover change on river systems. Treatise
Geomorphol 9:768–793
Lü Y, Sun R, Fu B, Wang Y (2012) Carbon retention by check dams: regional scale estimation.
Ecol Eng 44:139–146
Mingguo Z, Qiangguo C, Qinjuan C (2008) Modelling the runoff-sediment yield relationship using a
proportional function in hilly areas of the Loess Plateau, North China. Geomorphology 93(3):288–
301
Ouyang W, Skidmore AK, Hao F, Wang T (2010) Soil erosion dynamics response to landscape
pattern. Sci Total Environ 408(6):1358–1366
Parsons AJ, Wainwright J, Brazier RE, Powell DM (2006) Is sediment delivery a fallacy? Earth
Surf Proc Land 31(10):1325–1328
Preti F, Forzieri G, Chirico GB (2011) Forest cover influence on regional flood frequency assessment
in Mediterranean catchments. Hydrol Earth Syst Sci 15(10):3077
Rey F, Burylo M (2014) Can bioengineering structures made of willow cuttings trap sediment in
eroded marly gullies in a Mediterranean mountainous climate? Geomorphology 204:564–572
Shi ZH, Ai L, Li X, Huang XD, Wu GL, Liao W (2013) Partial least-squares regression for linking
land-cover patterns to soil erosion and sediment yield in watersheds. J Hydrol 498:165–176
Toy TJ, Foster GR, Renard KG (2002) Soil erosion: processes, prediction, measurement, and control.
Wiley, New York, pp 25–43
Trimble SW (1975) A volumetric estimate of man-induced soil erosion on the southern Piedmont
Plateau. Prospective Technol Predicting Sediment Yields sources 142–152
Vanmaercke M, Poesen J, Broeckx J, Nyssen J (2014) Sediment yield in Africa. Earth Sci Rev
136:350–368
Walling DE, Fang D (2003) Recent trends in the suspended sediment loads of the world’s rivers.
Global Planet Change 39(1):111–126
Yan B, Fang NF, Zhang PC, Shi ZH (2013) Impacts of land use change on watershed streamflow and
sediment yield: an assessment using hydrologic modelling and partial least squares regression. J
Hydrol 484:26–37
Chapter 14
Assessment of Probable Maximum Flood
(PMF) Using Hydrologic Model
for Probable Maximum Precipitation
in Maithon Watershed

Bhanu Sharma and Kalyan Kumar Bhar

Abstract Hydrological modeling is a widely used technique to simulate the hydro-

logical response in a basin due to precipitation for the purpose of basin water manage-
ment by water resource planners. In hydrological studies, estimating surface runoff
in a watershed based on the rate of precipitation received and quantifying discharge at
the outlet is important. An analysis has been done for developing hydrologic model
for the calculation of probable maximum flood (PMF) using probable maximum
precipitation (PMP), over a watershed. PMP is defined as “the greatest depth of
precipitation for a given duration that is physically possible over a given storm area
at a particular geographical location at a certain time of the year.” The focus of the
study is firmly on PMP estimates derived through statistical approach (Hershfield’s
method). The probable maximum flood (PMF) is the flood that can be predicted to be
measured using a hydrologic HEC-HMS model from the most extreme combination
of critical meteorological and hydrologic conditions that are reasonably possible in
particular drainage area. To compute loss, rainfall excess conversion to runoff, and
flow routing, methods like initial and constant rate, Snyder’s unit hydrograph, and
Muskingum routing were chosen, respectively. The study presents a methodology
for portraying the PMP estimation for Maithon watershed by analyzing individual
steps within the PMP derivation procedure. Calculated PMP value has been used in
HEC-HMS model to compute the PMF for Maithon catchment.

Keywords Probable maximum precipitation · Probable maximum flood ·

Hydrologic model · Hydraulic model · Maithon

B. Sharma (B)
Department of Hydrology, Indian Institute of Technology, Roorkee 247667, India
K. K. Bhar
Department of Civil Engineering, IIEST, Shibpur 711103, India
e-mail: kalyan@[Link]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 165
R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
166 B. Sharma and K. K. Bhar

14.1 Introduction

Probable maximum precipitation (PMP) is referred as “the greatest depth of the

precipitation for a given duration meteorologically possible for a given size storm
area at a particular location at a particular time of year, with no allowance made for
long-term climatic trends” (WMO 1986). This term is used by many hydrologists all
over the world to assess the probable maximum flood (PMF) through which hydraulic
structures can be designed. The predominant factor in designing of spillway of a
dam is PMP. Machhu II dam located in Gujrat, India, failed in 1979 due to the
insufficient spillway capacity and killed 1419 people. There are so many methods
related to statistical, dynamic, and empirical approach to find PMP. The magnitude of
spatial and temporal distribution has considered for a catchment to find the probable
maximum flood at a dam. The designing of dam on the basis of probable maximum
flood using probable maximum precipitation is widely used in the world. For different
regions of USA, National Weather Service (NWS) published hydrometeorological
reports to find PMP. The PMP has various methods to calculate based on the location
of watershed, availability of data, and other factors. The long-term data for rain gauge
stations has been used in statistical method. This method gives the point value of
PMP and needs area reduction curve to vary the area spatially over the area. This
paper has used Hershfield’s methods to assess PMP, which is a statistical method
based on the previous rainfall data.
This PMP value for different location in the watershed is used to assess the PMF
for that particular watershed. The peak discharge is used for the design of a hydraulic
structure and river headwork after considering hydrological factors. Spillway design
flood is used to design the spillway of storage structure which is generally known as
the maximum discharge crossing over the spillway without any damages and warning
to the stability of structure. The flood that is the result of the most severe combi-
nations including rare combinations of meteorological and hydrological factors is
known as probable maximum flood. In hydrologic studies, estimation of surface
runoff in a watershed on the basis of precipitation at outlet is the prominent thing.
Runoff assessment is a very tedious task due to the variations in parameters. In
this study, HEC-HMS model has been used to simulate the rainfall–runoff for a
watershed. Simulation of rainfall–runoff model will estimate the runoff. The inap-
propriate assessment of runoff leads to some problems in adequate management
of water resources and reservoir. The peak of the runoff can be assessed based on
the rational method, empirical method, and flood frequency method. The watershed
models and unit hydrograph techniques calculate flood hydrograph as well as peak
flow of the runoff.
14 Assessment of Probable Maximum Flood (PMF) Using Hydrologic … 167

Fig. 14.1 Study area

14.2 Study Area

14.2.1 Geographical Description of Study Area

The present study deals with the hydrologic aspects of an existing dam in India,
namely the Maithon dam, which is one of the four dams of the Damodar valley
(DV) system. Maithon dam is constructed across river Barakar and the location of
dam is at Latitude 23°47 13.06 north and longitude 86°49 01.44 east. Barakar
river is a tributary of Damodar river. The catchment area upstream of the dam has
been estimated as 6391.7 km2 , which spreads over the districts of Dhanbad, Jamtara,
Giridih, Kodarma, Hazaribag, and Chatra of the state of Jharkhand. It extends between
latitude 23°46 34.12 north and longitude 85°09 16.26 east, to latitude 24°32 09.80
north and longitude 86°53 19.20 east (Fig. 14.1).

14.2.2 Physiographic Description of Maithon Dam

Maithon dam had completed in 1957 under the supervision of Mr. [Link], a civil
engineer of Tennessee Valley Authority (TVA). Barakar is a seasonal river, 82% of
the reservoir filled up by the rainfall in July–September and stored water is used for
irrigation by Burdwan area (Table 14.1).
168 B. Sharma and K. K. Bhar

Table 14.1 Physiographic

Description Value
description of Maithon dam
River Barakar, major tributary of
damodar river
Height 56.08 m
Length 4064.35 m—Earthen
embankment
362.41 m concrete overflow
section
Gross storage capacity 1093.54 mm3
Live storage capacity 441.64 mm3
Full reservoir level 152.40 m
Minimum draw down level 132.59 m
Dead storage level 132.59 m
Dam top level 156.06 m
Crest level 140.21 m
Spillway type Ogee
Type of gates Radial
Size 12.5 m (h) × 12.19 m (w)
No. of bays 12 nos

14.3 Materials and Methodology

In this present study, the rainfall data (1994–2015) collected from Indian Meteoro-
logical Department (IMD) was used to prepare the rainfall distribution map. Digital
elevation model (DEM) and the satellite images are used as the primary source
of data. The digital elevation model (DEM) has been taken from Advanced Space
borne Thermal Emission and Reflection Radiometer (ASTER) that is a free product
of United States Geological Survey. It has a horizontal spatial resolution of 30 m and
a vertical accuracy of more than 10 m. The study area comprises of Barakar river
catchment, which was delineated by automatic catchment delineation from DEM in
ArcGIS 10. Different rainfall distribution map and PMP distribution maps are created
using ArcGIS software.

14.3.1 Delineation of Catchment

Catchment is an area or region bounded peripherally by a divide and eventually

drained to a single water source or body of water. A series of operations has been
carried out to extract the drainage network and the catchment with the help of the
14 Assessment of Probable Maximum Flood (PMF) Using Hydrologic … 169

Fig. 14.2 Maithon watershed

software ArcGIS 10.0. There are so many steps like fill, flow direction, flow accu-
mulation, stream network, stream feature, and snap pour point had been performed
(Fig. 14.2).

14.3.2 Preparation of Sub-catchment

Each of the tributaries of the main river or lake system has its own drainage area,
known as a sub-catchment. To delineate the sub-catchment, previous procedure
is used with inclusion of multiple pour point. Figure shows the delineated sub-
catchment. The following table shows the area of different sub-catchments (Fig. 14.3;
Table 14.2).

14.3.3 Rainfall Distribution Maps

The first step is to identify the rain gauge stations covering the entire catchment.
The name of rain gauge stations is Barkisuriya, Maithon, Nandadih, Palganj, Giridh,
Jamtara, Padma, Jamua, Barhi, Sillai Chalk, Tilaiya, and Dumri under the Maithon
catchment. On the basis of rainfall data, the rainfall map has been made, which
shows the average value of rainfall occurring in the years from 1992 to 2017. After
identifying the rain gauge station, the spatial variation of rainfall has been shown
with the help of rainfall distribution map over the catchment area made with the help
of inverse distance weighting (IDW) interpolation technique (Fig. 14.4).
170 B. Sharma and K. K. Bhar

Fig. 14.3 Sub-catchments of Maithon watershed

Table 14.2 Area of

Sub-catchment Area (Sq. Km.)
sub-catchments
1 1031.80
2 1451.95
3 1225.45
4 1768.83
5 910.66
Total area 6388.69

14.4 Probable Maximum Precipitation

PMP is defined as “the greatest depth of precipitation for a given duration that is
physically possible over a given storm area at a particular geographical location at
a certain time of the year.” The statistical method, often called Hershfield method
(1961), is more commonly used and can be applied, if long-term precipitation data
is available.
The Hershfield method is based on average precipitation and standard precipitation
deviation, similar to the frequency factor method of Chow (1951), which is expressed
as (Fig. 14.5; Table 14.3):

P = X̄ + K m Sn (14.1)

where
14 Assessment of Probable Maximum Flood (PMF) Using Hydrologic … 171

Fig. 14.4 Rainfall map

Fig. 14.5 PMP maps

1
n
X̄ = Xi (14.2)
n i=1

n  2
i=1 X I − X̄
Sn = (14.3)
n−1
172 B. Sharma and K. K. Bhar

Table 14.3 PMP values

Station PMP values (cm)
Barkisuriya 32.81
Maithon 46.78
Nandadih 47.96
Palganj 51.08
Giridh 65.16
Jamtara 58.09
Padma 65.14
Jamua 53.10
Barhi 59.10
Sillai Chalk 64.09
Tilaiya 47.06
Dumri 54.04

X i − X n−1
Km = (14.4)
σn−1

X̄ Sample mean of n annual maximum precipitation.

Sn Sample standard deviation.
Km Frequency factor.
Xi Highest observed annual maximum rainfall in the series.
X n−1 Mean of the annual maximum, excluding the highest value.
σn−1 Standard deviation of the annual precipitation, excluding the highest value.

14.5 Probable Maximum Flood (PMF)

Probable maximum flood (PMF) is calculated by HEC-HMS hydrologic model which

is being used. To compute loss, rainfall excess conversion to runoff and flow routing,
methods like initial and constant rate, Snyder’s unit hydrograph and Muskingum
routing were chosen, respectively.

14.5.1 Physiographic Parameter of the Sub-catchments

The physiographic parameters of the mentioned above were extracted for each of the
five sub-catchments using DEM by ArcGIS (Table 14.4).
14 Assessment of Probable Maximum Flood (PMF) Using Hydrologic … 173

Table 14.4 Sub-catchment centroid, longest stream

Sub-catchment Area (km2 ) Length of longest Length of stream to
stream (km) centroid (km)
R1 910.66 65.97 32.80
R2 1768.83 92.34 39.96
R3 1225.45 75.56 30.36
R4 1451.95 93.19 38.47
R5 1031.80 79.61 29.42

6388.69

Table 14.5 Basin lag time

Sub-catchment Length of longest Length of stream to Basin lag time (h)
stream (km) centroid (km)
R1 65.97 32.80 55.071
R2 92.34 39.96 64.634
R3 75.56 30.36 56.045
R4 93.19 38.47 64.078
R5 79.61 29.42 56.39

14.5.2 Snyder’s Method

Snyder (1938) developed a synthetic UH based on a study of watersheds in basins

ranging from 10 to 10,000 mi2 (Table 14.5).

tlag = CCt (L L c )0.3 (14.5)

tlag Basin lag time.

C Conversion constant (2.75).
Ct UH peaking coefficient (1.8–2.2).
L Length of the main stream from the outlet to the divide (km).
Lc Length along main stream to appoint nearest the watershed centroid (km).

14.5.3 Design Loss Rate

For assessment of probable maximum floods, a loss rate of 0.1 cm/hour is currently
used, as the influence of soil, land use, etc., have the minimal influences under such
extreme conditions. So, a loss rate of 0.1 cm/hour was adopted for the current study.
174 B. Sharma and K. K. Bhar

14.5.4 Base Flow

For the purpose of this study, the base flow rate has been adopted as 0.045 m3 /s/km2 .

14.5.5 Muskingum Parameters

See Table 14.6.

14.5.6 HEC-HMS Model

See Fig. 14.6.

Table 14.6 Muskingum

Reach Proportionality coefficient Weighing factor (X)
coefficient
(K)
R1 8.2 0.2
R2 5.9 0.2
R3 3.4 0.2
R4 5.8 0.2

Fig. 14.6 Basin model

14 Assessment of Probable Maximum Flood (PMF) Using Hydrologic … 175

Fig. 14.7 Final outflow hydrograph

14.5.7 Output

See Fig. 14.7.

14.6 Conclusion

The one-day probable maximum precipitation for Maithon catchment is 43.02 cm

using Hershfield’s method for 12 rain gauge stations. We can also see the point PMP
values for different stations given in the table above. Maithon, Nandadih, and Tilaiya
are having the value in the range of 40–50 cm. Palganj, Jamtara, Jamua, Barhi, and
Dumri are having the value in the range of 50–60 cm. Giridh, Padma, and Sillai chalk
having the values in the range of 60–70 cm. statistical method has the advantage of
taking into account actual precipitation data and its application is simple and fast.
A long record will yield generally more reliable PMP estimates than will a short
record of comparable quality. The peak outflow at the outlet of Maithon catchment
is 18751 m3 /s. The spillways are originally designed to sustain the discharge of
13,592 m3 /s. Here, we can clearly see that the design flood has increased by 37%
for Maithon watershed. The variation in rainfall due to climate change and due to
176 B. Sharma and K. K. Bhar

the silt charge accumulated in the reservoir is the prominent reason for increment
in design flood. The increment in value of probable maximum flood evolves the
problem like dam break which leads to the flooding in downstream area of dam that
causes financial loss, loss of life, and decrement in soil fertility. This study can have
the prominent role in dam break scenario and evacuation policies during dam break.

References

Chow VT (1951) A general formula for hydrologic frequency analysis. Trans Am Geophys Union
32(2):231–237
Snyder FF (1938) Synthetic unit-graphs. Eos. Trans Am Geophys Union 19(1):447–454
WMO (1986) Manual on estimation of probable maximum precipitation (PMP). ISBN 978-92-63-
11045-9
Chapter 15
Simulating Failure of Indravati Dam
Using Mike 11 and the Propagation
of Breached Outflow

Aditya Harikumar, Sachin Dhiman, and K. C. Patra

Abstract Dams are invariably used for multipurpose benefits to the society such
as generation of hydropower, irrigation, water supply, flood control, and others, but
with associated high risk. As it retains huge amount of water, there is always a risk of
formation of breach. There is a need to study the breaching aspects of a dam. In this
paper, a hydraulic model named MIKE11 developed by Danish Hydraulic Institute
(DHI) is used to carry out simulation of flood resulting from the failure of Indravati
multipurpose dam in the Odisha state. Propagation of the flood wave at its downstream
reaches is studied, and the consequent inundation map of the downstream submerged
areas is prepared. Digital elevation map of the study area is obtained from USGS site
and is converted into ARCII with the help of ArcGIS software. Simulations carried
out posed a challenge due to the availability of low-resolution topographic data, stiff
slope of the channel, obstructions at the downstream side of dam, and other factors.

Keywords ArcGIS · Indravati dam · Dam breach · Flood inundation · MIKE11

15.1 Introduction

Dams are the basic piece of frameworks to the general public that add to social
advancement and success. They serve too many beneficial purposes that are essen-
tial for the growth of the society; at the same time, dams hold a potential danger of
disappointment because of numerous specialized security issues and threats. Disap-
pointments are viewed as one of the real “low-likelihood” occasions. The surges
coming due to dam failures can prompt catastrophes with gigantic death toll and
property, particularly in thickly populated territories.
Past calamities resulting from dam disappointments are specifically identified
with the clearing time when a dam failure happens (Wahl 1998). It is along these
lines critical to comprehend the procedure of dam rupturing, and if conceivable, to
consider key breaking parameters required to display the dam rupturing process. Be

A. Harikumar (B) · S. Dhiman · K. C. Patra

Department of Civil Engineering, NIT, Rourkela, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 177
R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
178 A. Harikumar et al.

that as it may, many existing dams still posture expanding risks to the downstream
territories because of basic crumbling, lacking outline, defective development, and
poor activity and upkeep. These dams are alluded to as upset dams.
Hanson et al. (2005) clarified that the breach development rate has a critical
effect on the pinnacle release from a dam failure. In order to understand the breach
development rate, one has to keep a check on the breach initiation and formation
times.
Wahl (1998) had given a clear cut explanation about two types of times associ-
ated with dam breach. The breach initiation time, which can be utilized to start an
early cautioning, is the span from the beginning of overtopping stream over the dam
dissolving downstream slope to the beginning of bringing down of the peak of the
dam interfacing stored water to the broke outpouring. Breach development time is
the span beginning from the bringing down of the bank peak and consummation with
the total draw off the reservoir water.
Over the last two decades, some software have been developed for modeling of
dam break that includes HEC-RAS, MIKE11, MIKE 21, and MIKE FLOOD. Indu-
bitably, quality, and dependability of the counterfeit processes for every component
of the river flood boost the quality and authenticity of the created river flood danger
maps.
For successful running of any software model, it requires a correct representation
of channel of the downstream corresponding floodplains and other data as accurately
as possible to predict the flood magnitude and level of water along the channel after
the dam break (Gichamo et al. 2012).
For this purpose, DEM file is required for the study area. Apart from the topo-
graphical data, satellite images also provide the exact information along with the
different structures present in the area of interest. This is very much useful to get
an idea of the area that can get inundated after the burst of a particular dam on the
upstream (Qi and Altinakar 2011).
Many researchers have worked on this field to get an idea of the peak outflow
from the dams while connecting it to the different parameters [(SCS 1981), Singh
and Snorrason (1984), Froehlich (1995 and 2008)]. Froehlich (1995) had collected
data from 63 dam failure conditions to get an idea of the average width of breach.
In order to get an idea about the formation time of the breach, he used data from 21
real cases of dam failure. Majority of the works are focussed to two parameters, i.e.,
average width of breach and formation time of the breach, as these are more sensitive
to failure peak during breaching of a dam.
In this paper, a numerical simulation of the failure of Indravati dam is carried
out by employing Danish Hydraulic Institute’s (DHI)—MIKE 11. In this study, the
effect of overtopping failure is given attention through Saint–Venant equations. A
reliable and accurate mathematical model needs to be developed for breach evolution
for analyzing the impact of dam-break floods on downstream regions (Dhiman and
Patra 2018).
15 Simulating Failure of Indravati Dam Using Mike 11 … 179

15.2 Study Area

15.2.1 Salient Features of Indravati Dam

River Indravati originates in the eastern mountainous regions of India in the state
of Odisha. It is an east flowing river that enters into Chhattisgarh state, finally
joining the Godavari river. Indravati river basin lies between latitudes 18°43 25 N
to 19°26 46 N and longitudes 80°16 19 E to 83°07 10 E. Major tributaries of this
river are Bhaskal, Narangi, Nibra, Kotri, Bandia, and Nandiraj rivers. Figure 15.1
shows the Indravati basin and dam while the other feature of Indravati dam is given
in Table 15.1. Figure 15.2 depicts a schematic diagram of the Indravati river system.

Fig. 15.1 Map showing Indravati river basin and the dam

Table 15.1 Salient features of Indravati dam

Dam location and properties Reservoir capacity
River Indravati Gross storage capacity at 2307 M cumec
F.R.L
State Odisha Dead storage capacity 851.94 M cumec
Latitude & longitude 19o 16 34.8 N and Live storage capacity 1455.76 M cumec
82o 49 42.4 E
Catchment area 2630 sq. km Full reservoir level R.L. 642 m
Length 539 m Maximum water level R.L. 643 m
Height 45 m Min draw down level R.L. 625 m
Dam top crest level R.L. 645 m Maximum flood discharge 23,030 cumec
Deepest bed level R.L. 600 m
180 A. Harikumar et al.

Fig. 15.2 Schematic diagram of Indravati river basin

15.3 Dam Breach Parameters

It is one of the most important inputs which have to be carefully considered in a

dam-break analysis. Usually earthen dams take more time for its complete failure
as compared to a masonry dam. The failure time of earthen dam is 0.1 to 1 h and
that of masonry dam is 0.1 to 0.2 h (Fread 2006).Table 15.2 shows the guidelines
followed during dam-break analysis as recommended by the UK and the federal
energy regulatory commission (FERC) of the USA.
There are various ways in which breaches can be formed on embankment dams,
and it is pretty difficult to find out the extent of the erosion with the help of stringent
mathematical formulas. Breach formation in embankment dams due to overtopping
15 Simulating Failure of Indravati Dam Using Mike 11 … 181

Table 15.2 Guidelines followed during dam break by UK and federal energy regulatory
commission (FERC) of USA
Dam type Average breach Failure time hrs Breach side slope Agency
width H:l V
Eaithen/Rockfill (0.5–5.0) × HD 0.5–4.0 0–1.0 USACE (2007)
(1.0–5.0) × HD 0.1–1.0 0–1.0 FERC (1988)
(2.0–5.0) × HD 0.1–1.0 0–1.0 NWS(Fread 2006)
Concrete gravity Multiple 0.1–0.5 Vertical USACE (2007)
Monoliths 0.l–0.3 Vertical FERC
Usually ≤ 0.5 L 0.1–0.2 Vertical NWS (Fread 2006)
Usually ≤ 0.5 L

Fig. 15.3 Dimensions of a trapezoidal dam breach approximation, (height Hb , width B̄ and ratio
of side slope z (H:V), water in the reservoir Y f )

floodwaters has been counterfeit using complex 2D depth averaged flow models
connected with slope failure method and soil erosion (Froehlich 2004; Faeh 2007).
Usually in practice, breach is presumed to take the shape of a trapezium as shown in
Fig. 15.3 (Fread 1984; USACE 1978).

15.4 Methodology

The DEM file of the study area is converted into arcii format to import it into the
mike hydro setup. The first step in mike hydro is to create a simulation file to save
the progress of the work. In most of the dam-break cases, there is a single or multiple
channels, a reservoir, the dam and its structures like spillways and others. At first,
182 A. Harikumar et al.

Table 15.3 Dam breach

Dam crest 645 m Breach width at 250 m
parameters
elevation end
Reservoir level at 644.3 m Breach slope 0:1 (vertical)
breach
Breach width at 10 m
start

the network file is created, then the branches of the reservoir and downstream river
are created. While creating/giving the cross section, it is mandatory to give the first
chainage section of the reservoir to mark it as the storage. After that the boundary
file is created, inflow at the upstream is required for the simulation. Time series for
water level and discharge is to be made, after which HD parameters are to be created.
The HD module makes use of an implicit method for the calculation of unsteady
flow in rivers. After all this procedure, the setup is run and for this a simulation
editor is created. While considering the dam breach modeling, one of the greatest
uncertainties is the simulation of the breach (Wurbs 1987). Depending upon the size
of the reservoir, the parameters importance varies. The dam breach parameters used
are shown in Table 15.3.

15.5 Model Setup

The first step is the creation of Indravati river in the network editor. The river extends
up to a stretch of 138.9 km at its downstream. Up to a length of 50,000 m, the river
cross sections are provided at spacing of 500 m and after that and till 138,900 m it
is provided at a spacing of 1000 m. The dam-break arrangement is defined at 130 m
chainage from the starting point. At chainage “0”, the reservoir is modeled. The
PMF is considered as the inflow to the reservoir. At chainage point of 138,900 m,
boundary conditions are applied. The q/h values are generated automatically from
manning’s formula. Table 15.4 gives the q/h values, and Fig. 15.4 shows the rating
curve obtained from MIKE 11. Manning’s roughness coefficient for the whole river
is taken as 0.022 following Chow and V.T. (1959) guidelines.

Table 15.4 Stage discharge

Stage (h) in m Discharge (q) Stage (h) in m Discharge (q)
values
in m3 /s in m3 /s
537.6 0 541.7 6069
538 22 542 7636
539 335.6 545.7 29,791.3
540 1879.4 550 61,702.7
541 3765.7
15 Simulating Failure of Indravati Dam Using Mike 11 … 183

Fig. 15.4 Rating curve

15.6 Results and Analysis

During dam-break analysis, the most critical state is when the water in the reservoir
is at full reservoir level (FRL) and a PMF infringes the reservoir. The value of PMF
considered is 23,030 cumec. Dam breach is initiated at 25.3 h from the beginning of
PMF, and the maximum water level attained at that time is 645.201 m. The maximum
value of discharge flowing out of the dam is 33,218.4 cumecs, which is 1.44 times
more than the PMF. This maximum discharge is attained at 6 h from the beginning
of the dam break. Velocity of water coming out at this point is 10.6 m/s. The corre-
sponding value of breach parameters is given in Table 15.5, while Table 15.6 shows
the statistics of dam breach. Figures 15.5 and 15.6 show the discharge versus the
time graph and reservoir water level vs the time graph, respectively.

15.6.1 Flood Routing

For the purpose of flood routing at its downstream reaches, 8 chainage points are
considered at locations of 1.5, 5, 7.5, 10.5, 25, 55, 100, and 120 km d/s of the
reservoir. At the dam, the value of peak discharge coming out after breaching process
commences is 33,218.4 cumec. This discharge comes out from the dam after a time

Table 15.5 Breach

Level of breach (m) 620
parameters at maximum
discharge Depth in breach (m) 11.97
Breach bottom width (m) 250
Breach width at crest (m) 300
184 A. Harikumar et al.

Table 15.6 Statistics of dam breach

Time (h) Q in V in Reservoir Level of Depth in Breach Breach
breach (m3 /s) breach (m/s) water breach (m) breach (m) bottom width at
level (m) width (m) crest (m)
25.3 12.2 2.215 645.201 644.3 0.525 10 11.4
25.5 38.4 2.806 645.205 643.7 0.844 15.3 17.8
26 332.1 4.355 645.217 641.7 2.038 35.3 41.9
26.5 1034.2 5.474 645.211 639.6 3.221 55.4 66.0
27 2237.3 6.387 645.174 637.67 4.386 75.4 90.1
27.5 3999.5 7.168 645.09 635.64 5.522 95.5 114.2
28 6346.9 7.849 644.938 633.6 6.618 115.5 138.3
28.5 9272.3 8.447 644.697 631.5 7.661 135.6 162.4
29 12,736.7 8.974 644.346 629.51 8.638 155.6 186.5
29.5 16,669.3 9.434 643.864 627.51 9.538 175.7 210.67
30 20,968.2 9.833 643.229 625.41 10.347 195.7 234.7
30.5 25,503.2 10.17 642.423 623.42 11.054 215.8 258.6
31 30,112.8 10.447 641.422 621.42 11.646 235.8 282.9
31.3 33,218.4 10.597 640.551 620 11.967 250 300
31.4 33,008.8 10.576 640.468 620 11.917 250 300.0

Fig. 15.5 Discharge versus time at the dam site

of 6 h from the start of the breach. At 1.5 km d/s the value of maximum discharge
is 33,067.48 cumec. The time of arrival of flood is just 1 min from the beginning of
flood. At 5 km d/s, the value of maximum discharge is 32,899.44 cumec, while the
arrival time of peak value of flood is 9 min. At 7.5 km d/s, the value of maximum
15 Simulating Failure of Indravati Dam Using Mike 11 … 185

Fig. 15.6 Reservoir w.l

versus time

discharge is 32,704.8 cumecs, while the arrival time of peak value of flood is 12 min.
At 10.5 km d/s, the value of maximum discharge is 32,453.44 cumecs, and its arrival
time is 17 min. At 25 km d/s, the value of maximum discharge is 29,239.14 cumecs,
while its arrival time is 1 h 3 min. At 55 km d/s, the value of maximum discharge
is 23,895.56 cumecs, while its arrival time is 3 h 23 min. At 100 km d/s, the value
of maximum discharge is 16,526.59 cumecs, and its arrival time is 8 h 30 min. At
120 km d/s, the value of maximum discharge is 13,693.69 cumecs, and its arrival
time is 11 h 3 min.

15.6.2 Longitudinal Profile of the Bed

Figure 15.7 depicts the longitudinal profile of the bed of river Indravati after dam
break along with the water level attained at different downstream locations resulting

Fig. 15.7 Longitudinal profile of the bed of river Indravati after dam break
186 A. Harikumar et al.

from the dam break. Figure 15.8 represents the time series discharge of the river at
midpoint of selected sections.
Level of water at all the 8 chainages is shown in Table 15.7. Figure 15.9 depicts
the time series water level of all the 8 chainages.

Fig. 15.8 Time series discharge

Table 15.7 Level of water along with the time of occurrence at each chainages
Distance d/s (km) w.l. (m) Max. w.l. Distance d/s (km) w.l. (m) Max. w.l.
time time
1.5 606.87 29/09/2018 55 569.67 29/09/2018
[Link] [Link]
5 596.69 29/09/2018 100 556.34 30/09/2018
[Link] [Link]
7.5 593.94 29/09/2018 120 552.69 30/09/2018
[Link] [Link]
10.5 589.08 29/09/2018
[Link]
25 581.45 29/09/2018
[Link]

Fig. 15.9 Time series water level

15 Simulating Failure of Indravati Dam Using Mike 11 … 187

Fig. 15.10 Flood inundation map of Indravati river (a) before and (b) after flooding

15.7 Flood Maps

Based on the contour maps available, the area which comes under the effect of flood
can be obtained. This information is shown in Fig. 15.10a and b as before and after
images. Figure 15.11a and b shows the before and after image of one of the area
along the course of the river where the flood advances into the adjacent flooding
plains.

15.8 Emergency Action Plan (EAP)

EAP is a written document given by the dam authorities specifying the plan of action
that are required to be taken for the protection of a dam or to control the loss of life
and property. It is not like a provisional plan which stresses on the maintenance and
corrective construction of the dam rather it helps in the simplification of the problems
of dam safety. The central dam safety organization (CDSO) under the central water
188 A. Harikumar et al.

Fig. 15.11 Flood map at Khatiguda which is 5 km d/s of dam (a) before and (b) after flooding

commission (CWC) aims in simplifying the problems in dam safety. In the current
situation, a severe weather condition is taken in which heavy rainfall is also taken
into consideration. Breaching of the Indravati dam takes place when the water level
reaches a height of 645.201 m which is above the top of the dam. Dam breach is
initiated at 25.3 h from the beginning of PMF. The maximum value of discharge
flowing out of the dam is 33,218.4 cumec which is 1.44 times more than the PMF
and it occurs at 31.3 h which signifies the breach in the dam within 6 h. From Table
15.8, we can get to know the discharge at all the 8 chainages and their respective time
of arrival. There are three levels of danger signal such as emergency level-1(yellow),
emergency level-2 (orange), and emergency level-3 (red). These are described on the
basis of breach time available and the locality affected by the flood.
From Table 15.8, it is evident that some parts of the river reach area come under
extreme danger zone, i.e., locations varying from 1.25 to 25 km from dam site. These
15 Simulating Failure of Indravati Dam Using Mike 11 … 189

Table 15.8 Discharge along with the time of occurrence and emergency level at each chainages
Distance d/s (km) Discharge (m3 /s) Max. Q. Time Emergency level
1.25 33,067.48 29/09/2018 [Link] Red
5.25 32,899.44 29/09/2018 [Link] Red
7.25 32,704.80 29/09/2018 [Link] Red
10.25 32,453.44 29/09/2018 [Link] Red
24.75 29,239.14 29/09/2018 [Link] Red
54.5 23,895.56 29/09/2018 [Link] Orange
99.5 16,526.56 30/09/2018 [Link] Yellow
119.5 13,693.69 30/09/2018 [Link] Yellow

areas are very much affected because during the propagation of the wave there is a
fair chance that the community living in these areas would be affected a lot.

15.9 Conclusions

Here, in this paper, a hypothetical failure of Indravati reservoir is simulated using

MIKE 11. The dam has a height of 45 m with gross storage capacity of 2300 m
cumecs. Effects of dam break at the downstream side of the dam are studied with
the help of flood hydrograph, duration of the flood, water level, and velocity of
propagation of flood wave. From the modeling, the following conclusions can be
drawn.
• Maximum flood discharge from the failure of the Indravati dam is 33,218.4 cumec,
which is 1.44 times more than the PMF.
• The dam-break results can be utilized to prepare maps for an emergency action
plan that can help communities in arranging future improvements in zones that
are flooding prone.
• The time series water level and discharge at different c/s (1.5, 5, 7.5, 10.5, 25, 55,
100, and 120 km) are obtained that suggests the flood-prone areas resulting from
the dam breach.

References

Chow VT (1959) Open-channel hydraulics. McGraw-Hill, New York, pp 680

DHI (2008) MIKE 11- A modeling system for rivers and channels, user’s manual. Danish Hydraulic
Institute, Horsholm, Denmark
Dhiman S, Patra KC (2018) Experimental study of embankment breach based on its soil properties.
ISH J Hydraul Eng. [Link]
190 A. Harikumar et al.

Faeh R (2007) Numerical modeling of breach erosion of river embankments. J Hydraul Eng
133(9):1000–1009
Federal Power Commission (1984) Hydroelectric power resources of the United States, developed
and undeveloped. Federal Power Commission
Fread DL (1984) DAMBRK—the nws dam-break flood forecasting model. National weather service
hydrological technical note no. 4, hydrologic research laboratory, silver spring, md
Fread DL (2006) Dam failure analysis. ASDSO Adv Technical Sem
Froehlich DC (1995) Peak outflow from breached embankment dam. J Water Resour Plann Manage
121(1):90–97
Froehlich DC (2004) Two-dimensional model for embankment dam breach formation and flood
wave generation. In: Proceeding of the dam safety 2004, 21st annual conference of the association
of state dam safety officials, association of state dam safety officials, Lexington, ky
Froehlich DC (2008) Embankment dam breach parameters and their uncertainties. J Hydraul Eng
134(12):1708–1721
Gichamo TZ, Popescu I, Jonoski A, Solomatine D (2012) River cross-section extraction from the
aster global dem for flood modelling. Environ Model Softw 31:37–46
Hanson GJ, Cook KR, Hunt SL (2005) Physical modelling of overtopping erosion and breach
formation of cohesive embankments. Trans ASAE 48(5):1783–1794
Qi H, Altinakar MS (2011) A GIS-based decision support system for integrated flood management
under uncertainty with two dimensional numerical simulations. Environ Model Softw 26(6):817–
821
SCS (1981) Simplified dam-breach routing procedure. Technical release no. 66 (rev. 1). USDA-
NRCS, Washington, D.C.
Singh KP, Snorrason A (1984) Sensitivity of outflow peaks and flood stages to the selection of dam
breach parameters and simulation models. J Hydrol 68(1):295–310
U.S. Army Corps of Engineers (USACE) (1978) Flood hydrograph package HEC-1 user’s manual
for dam safety investigations. Hydrologic Engineering Center, Davis, Calif
U.S. Army Corps of Engineers (USACE) (2007) Performance evaluation of the New Orleans and
southeast Louisiana hurricane protection system: final report of the interagency performance
evaluation task force, vol III
Wahl TL (1998) Prediction of embankment dam breach parameters—a literature review and
needs assessment. Dam safety rep. No. Dso-98–004, U.S. department of the interior, bureau
of reclamation, Denver
Wurbs RA (1987) Dam-breach flood wave models. J Hydraul Eng 113(1):29–46
Chapter 16
Optimization of Water Allocation
for Ukai Reservoir Using Elitist TLBO

Vijendra Kumar and S. M. Yadav

Abstract The optimal reservoir policy is a complex problem to optimize. This

paper introduces the improved teaching learning-based optimization (TLBO) by
introducing the elitist concept, in order to improve the convergence, global search-
ability and faster evolution process. The objective of the present study is to maxi-
mize the water allocation for Ukai reservoir, India, to supply water for irrigation,
domestic and industrial uses at different dependable inflow. Elitist teaching learning-
based optimization (ETLBO) algorithm has been used to optimize water alloca-
tion, using four different models having dependable inflow as 60, 65, 70 and 75%.
The results from ETLBO are compared with ordinary TLBO, differential evolu-
tion (DE), particle swarm optimization (PSO) and linear programming (LP). It was
observed that ETLBO performed better in terms of better global searchability and
faster convergence than TLBO, DE, PSO and LP.

Keywords Optimization · Reservoir operation · Teaching learning-based

optimization · Elitist algorithm · Ukai dam

16.1 Introduction

Complexities in water allocation due to increasing demands and shortage of water

resources have increased the importance of obtaining the sustainable development
and maximizing benefits (Bahrami et al. 2018). Due to the disparity between the
demand and supply, it causes pressure on water supply, conflict among other water
users, competition and impact on the environment (Chang and Wang 2013). It is a
challenging and important for the policymaker to make effective water allocation
and water management issues. In order to solve these problems, optimization of

V. Kumar (B)
Civil Engineering Department, G. H. Raisoni College of Engineering and Management, Pune,
Maharashtra, India
S. M. Yadav
Civil Engineering Department, Sardar Vallabhbhai National Institute of Technology, Surat,
Gujarat, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 191
R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
192 V. Kumar and S. M. Yadav

an existing project is an alternative solution (Kumar and Yadav 2020a). Either tradi-
tional or evolutionary-based algorithm can be applied to solve the reservoir operation
problems. The traditional techniques used in reservoir operation are linear program-
ming (LP), nonlinear programming (NLP) and dynamic programming (DP). Despite
the effectiveness of these methods, it has its own limitation. Such as LP need linear
objective function and constraints, DP is cursed by dimensionality and NLP is not
able to solve the non-convex problem effectively (Kumar and Yadav 2020b). These
are the few limitations of the traditional methods, apart, it is observed to stick in local
optimum solution.
An evolutionary algorithm is free from these limitations and able to provide a
global solution. Over the past few decades, evolutionary algorithms have gained
interest among researchers. Different algorithms have been used to solve the reser-
voir operation problem, for example, particle swarm optimization (PSO) (Kumar and
Reddy 2007; Afshar 2013; Bai et al. 2017), genetic algorithm (GA) (Chang et al.
2010), differential evolutionary (DE) (Regulwar et al. 2010), weed optimization algo-
rithm (WOA) (Asgari et al. 2016), cuckoo search (CS) (Ming et al. 2015), artificial
bee colony (ABC) (Hossain and El-shafie 2014) and shark algorithm (Ehteram et al.
2017). The existing algorithms are having few limitations. Evolutionary algorithms
are having two type of parameters, first, common controlling parameters, i.e. popu-
lation size and a number of iterations, and second is individual algorithm-specific
parameters, such as PSO used internal parameters like social and cognitive param-
eters and inertia weight. GA required mutation parameters, crossover parameters
and reproduction parameters. DE used crossover rate and scaling factor. ABC neces-
sitates the number of employed bees, scout bees and onlookers. These parameters
are called algorithm-specific parameters, and it is different for different algorithms.
Before executing the algorithm, it is required to tune these parameters, improper
selection of these parameters affect the overall performance of the algorithm. Apart,
other algorithms such as cuckoo search, bat algorithm, ant colony optimization, weed
optimization, etc., used in reservoir operations have their own parameters.
Teaching learning-based optimization (TLBO) is a newly developed algorithm
developed based on human behaviour of teaching and learning process (Rao et al.
2011). TLBO do not have any algorithm-specific parameters. Rao et al. (2012) used
TLBO to optimize the continuous nonlinear problem, and it performed better than
other algorithms. Baykasoǧlu et al. (2014) used TLBO for flow shop and job shop
scheduling problem and was found to have considerable potential compared to other
algorithms. Bayram et al. (2015) applied TLBO over dissolved oxygen problem.
Baghlani et al. (2017) used TLBO to solve steel truss sizing optimization problem.
Kumar and Yadav (2018) used TLBO to solve a multi-reservoir operation problem.
Kumar and Yadav (2019) used ETLBO to optimize the cropping pattern over two
different models and compared the results with the elitist jaya algorithm.
This paper presents the improved version of TLBO, by introducing the elitist
concept in order to improve the convergence, faster evolutions process, and global
search ability. The objective of the study is to maximize the water allocation for
Ukai reservoir to supply water for irrigation, domestic and industrial uses at different
dependable inflow. Elitist teaching learning-based optimization (ETLBO) has been
16 Optimization of Water Allocation for Ukai Reservoir … 193

used to optimize water allocation, for four different models with different dependable
inflow, namely 60, 65, 70 and 75%. The results from ETLBO are compared with
ordinary TLBO, DE, PSO and LP. The following section describes methods and
materials.

16.2 Methods and Materials

16.2.1 Differential Evolution (DE)

DE was proposed by (Storn and Price 1996) as an effective heuristic technique used
for solving optimization problems. DE optimized the problem based on a set of
population and updating the population with existing ones according to the simple
formula. And, then keeping the best solution out of the two which score better
functional value or fitness to hand the optimization problem. The algorithm follows
the following steps.
Step 1: Decide the parameters and generate the initial population vector solution,
j
i.e. Yk,G = {yk,G
1
, yk,G
2
, . . . , yk,G
D
} using the expression as per Eq. (16.1).
 
j j j
yk,0 = ymin + Rk, j ∗ ymax
j
− ymin (16.1)

j
where yk,0 is the initial population at kth component for jth vector at generation
G = 0, k = 1, 2, . . . , n, n is the population size, j = 1, 2, . . . , p, p is dimensional
j j
parameter vector, ymin and ymax are the lower and upper limit, Rk, j is the random
number.
j
Step 2: Produce mutation vector, i.e. Vk,G = {Vk,G 1
, 2k,G , . . . , Vi,G
D
} for each target
j
vector Yk,G , mutation operator is used.
 
j
Vk,G = Yr1k , G + F ∗ Yr2k , G − Yr3k , G (16.2)

where r1k , r2k , and r3k are the integers randomly generated between [1, NP], F is the
scaling factor parameter varies between [0, 2].
j
Step 3: Next crossover operation is applied wherein each target vector Yk,G
j j
and its corresponding mutation vector Vk,G to obtained a trail vector Uk,G =
{Uk,G , Uk,G , . . . , Uk,G }. The basic binomial crossover is defined as Eq. (16.3).
1 2 D


j
j Vk,G , if (rand[0, 1] ≤ CR) or ( j = jrand )
Uk,G = j (16.3)
Yk,G , otherwise
194 V. Kumar and S. M. Yadav

where jrand is the integers randomly generated between [1, NP], CR is the crossover
rate parameter varies between [0, 1].
Step 4: Selection operation
    
j j j
j Uk,G , iff Uk,G ≤ f Yk,G
Yk,G+1 = j
(16.4)
Yk,G , otherwise

Step 5: Increase the generation to G = G + 1. The algorithm reiterates until the

maximum iterations are accomplice.

16.2.2 Particle Swarm Optimization (PSO)

PSO was proposed by Eberhart and Kennedy (1995). It is a metaheuristic-based

swarm intelligence established optimization technique, similar to other population-
based algorithm. PSO works similar to the natural grouping as birds flock and fish
school. It provides the optimal solution based on the individual and social behaviour
(Shourian et al. 2008). The algorithm follows the following steps.
Step 1: Decide the parameters and generate the initial group of random particles,
for each particle, generate positions Yk and velocity vk .
Step 2: Evaluate the fitness and obtained Pbestk , i.e.

Pbestk (i − 1), if F(Pbestk (i)) ≤ F(Pbestk (i − 1))
Pbestk = (16.5)
Pbestk (i), if F(Pbestk (i)) > F(Pbestk (i − 1))

where Pbestk represents the best position of the individual, I represent the iteration,
i.e. I = 1, 2, . . . , i, of the kth individual.
Step 3: Next is find the global best position from the swarm, i.e.

Gbestk (I ) = max{F(Gbest1 (i)), F(Gbest2 (i)), . . . , F(Gbestn (i))} (16.6)

Step 4: Update the velocity using Eq. (16.7) if vk > vmax then vk = vmax ; if
vk < vmin then vk = vmin .

vk (i + 1) = w(i) ∗ vk (i) + c1 ∗ r1 (i) ∗ (Pbestk (i) − Yk (i))

+ c2 ∗ r2 (i) ∗ (Gbestk (i) − Yk (i)) (16.7)

where vk and Yk are the velocity and position of kth individual, respectively. The
internal parameters are w(i), c1 and c2 , term w(i) represent the internal weight, c1
is the cognitive parameters and c2 is the social parameters.
Step 5: Update the velocity using Eq. (16.8).
16 Optimization of Water Allocation for Ukai Reservoir … 195

Yk (i + 1) = Yk (i) + vk (i + 1) (16.8)

This complete one cycle, the algorithm reiterates until the maximum iterations
are accomplice.

16.2.3 Teaching Learning-Based Optimization (TLBO)

TLBO is a metaheuristic optimization technique similar to other population-based

algorithm proposed by Rao et al. (2011). It works on the human behaviour of teaching
and learning process. It consists of a two-phases teaching and learning. The algorithm
follows the following steps.
Step 1: Decide the common controlling parameters and generate the initial
population random between the bounds, using Eq. (16.9).
 
Randomly Generated Population yk = yk,min + Rk ∗ yk,max − yk,min (16.9)

where yk is the initial population at kth component, k = 1, 2, . . . , n, n is the popula-

tion size, yk,min and yk,max are the lower and upper limit, Rk is the random number
between [0, 1].
Step 2: From the randomly generated population identify the best solution.
Step 3: Teaching phase, wherein teacher teaches to improve the results, as
presented in Eqs. (16.10) and (16.11).

Difference mean = r ∗ (ybest − ymean ) (16.10)

ynew = (yold + Difference mean) (16.11)

where ybest and ymean are the best and mean solution, ynew and yold are the modified
and old solution. Better functional value obtained by ynew and yold is selected at the
end of the teaching phase.
Step 4: Learning phase, wherein any two learners tries to exchange their knowl-
edge for better results. For example, any two random learners, i.e. y1 and y2 , where
1 = 2

iff(y1 ) < f (y2 ); Then, ynew, 1 = yold,1 + ri ∗ (y2 − y1 ) (16.12)

iff(y1 ) > f (y2 ); Then, ynew, 1 = yold,1 + ri ∗ (y1 − y2 ) (16.13)

Again, the better functional value is selected between the old and new solution.
Step 5: This complete one cycle, the algorithm reiterates until the maximum
iterations are accomplice.
196 V. Kumar and S. M. Yadav

16.2.4 Elitist Teaching Learning-Based Optimization

(ETLBO)

ETLBO is the improved version of TLBO, wherein elitist concept has been applied
with ordinary TLBO. The elitist concept has been applied in the different evolu-
tionary algorithm; the same concept is applied in TLBO, wherein it preserves the
best solution. Elitism helps to remove the worst solution with the best elite solu-
tion; the replacement depends on the number of elite sizes. The working steps are
similar with TLBO up to learning phase. At the end of the learning phase, an elite
solution is identified. The greedy selection process is adapted to update the solution.
This complete one cycle, the algorithm reiterates until the maximum iterations are
accomplice.

16.3 Study Area and Data Collection

Tapi River is an interstate River of India originates from Betul district at an elevation
of 752 m above MSL and having a total length of 724 km. It drains an area of
65,145 km2 into Gulf of Khambhat, Arabian Sea. Ukai dam was constructed across
the Tapi River in 1972. It was constructed for multipurpose such as irrigation, water
supply, power generation and flood control. It is having a total capacity of 8510
million cubic metres (MCM), which include live storage of 7079.19 MCM and
dead storage of 1430.82 MCM. Table 16.1 shows the salient features of the Ukai
reservoir. The reservoir water is allocated for domestic purpose, industrial purpose
and irrigation purpose. Downstream of the Ukai dam is Surat city having a population
of approximately 6.6 million. Figure 16.1 shows the index map of the study area,

Table 16.1 Salient features

General ıtem Details
of Ukai reservoir
Mean annual rainfall in the 785 mm
watershed
Maximum annual rainfall in the 1191 mm
watershed
Coordinates 21° 14 53.52 N
73° 35 21.84 E
Catchment area 62,225 Km2
Top of dam 111.252 m
Area 65,145 Km2
State in the basin Maharashtra, Madhya
Pradesh, and Gujarat
Road width on the spillway 6.706 m
Type of spillway Radial
16 Optimization of Water Allocation for Ukai Reservoir … 197

Fig. 16.1 Index map a India map showing Tapi basin, b Tapi basin and c Cultivated land from
Ukai reservoir

wherein Fig. 16.1a represents the India map showing Tapi basin, Fig. 16.1b represents
the Tapi catchment area, and Fig. 16.1c shows the location of the dam and cultivated
land from Ukai reservoir. Four different canal networks are present in different colour
combinations, namely Ukai right and left bank main canal and Kakrapar right and left
bank main canal, which was used for irrigation. Figure 16.2 shows the line diagram
of Ukai dam, and it represents the various demands and losses.
The data required for the present study are water allocation for irrigation, domestic
demands, industrial demands, storage of dam, the inflow of the reservoir and evap-
oration losses. The data were collected from Surat Irrigation Circle (SIC) and Ukai
Left Bank Division (ULBD).

16.4 Mathematical Models

The objective function and the constraints are listed as follows.

198 V. Kumar and S. M. Yadav

Fig. 16.2 Line diagram of Ukai dam

16.4.1 Objective Function

The aim of the present study is to maximize the water allocation for irrigation demand,
domestic demand and industrial demands, over a year from Ukai reservoir, in consid-
eration of hydropower benefits. The objective function is mathematically expressed
as Eq. (16.14).


12 
12 
12
Max F(B) = Ri,t + Rd,t + Rin,t (16.14)
t=1 t=1 t=1

where F(B) is the maximum water allocated for different purpose, Ri,t is the water
allocated for irrigation purpose in MCM, Rd,t is the water allocated for domestic
purpose in MCM, Rin,t is the water allocated for industrial purposein MCM and t is
the time period, i.e. t = 1, 2, . . . , 12.
To satisfied reservoir constraints, the penalty is applied to the reservoir. The
objective function is modified as per Eq. (16.15).
16 Optimization of Water Allocation for Ukai Reservoir … 199


12 
12 
12
 2
Max F(B) = Ri,t + Rd,t + Rin,t − p(n) ∗ abs p(n,t) (16.15)
t=1 t=1 t=1

where p(n) is the penalty parameters p(n) = 5, and p(n,t) is the penalty function.

16.4.2 Constraints

Following are constraints that are subjected to the reservoir.

Storage capacity constraint
Monthly storage of the reservoir should not be less than the minimum monthly
capacity at which hydropower is generated and should not be more than the maximum
monthly capacity, for time t month, i.e. t = 1, 2, . . . , 12.

St,Min ≤ St ≤ St,sMax (16.16)

where St,min is the monthly minimum storage capacity of the reservoir in MCM,
St,max is the monthly maximum storage capacity of the reservoir in MCM and St is
the monthly storage in MCM.
Reservoir storage continuity constraint
The reservoir storage is governed by the continuity equation, represented by
Eq. (16.17).
 
St+1 = St + It − Ri,t + Rd,t + Rin,t − E t (16.17)

Where St is the active storage during month t, It is the reservoir inflow during month
t in MCM, E t is the evaporation loss during month t in MCM and St+1 is the storage
during month (t + 1).
Irrigation demand constraint
Monthly irrigation demand should be equal to or less than the maximum water
allocated for irrigation and it should be equal to or more than the minimum water
allocated for irrigation.

Ri,min,t ≤ Ri, t ≤ Ri,max,t (16.18)

Where Ri,min,t is the minimum water allocated for irrigation and Ri,max,t is the
maximum water allocation for irrigation, during month t in MCM.
200 V. Kumar and S. M. Yadav

Domestic water demand constraint

Monthly domestic water demand should be equal to or less than the maximum water
allocated for domestic demand, and it should be equal to or more than the minimum
water allocated for domestic demand.

Dd,min,t ≤ Rd,t ≤ Dd,max,t (16.19)

where Dd,min,t is the minimum water allocated for domestic demand and Dd,max,t is
the maximum water allocation for domestic demand, during month t in MCM.
Industrial water demand constraint
Monthly industrial water demand should be equal to or less than the maximum water
allocated for industrial demand, and it should be equal to or more than the minimum
water allocated for industrial demand.

I Din,min,t ≤ Rin,t ≤ I Din,max,t (16.20)

where I Din,min,t is the minimum water allocated for industrial demand and I Din,max,t
is the maximum water allocation for industrial demand, during month t in MCM.
Non-negativity constraint

Ri,t > 0, Rd,t > 0, Rin,t > 0 (16.21)

16.5 Results and Discussion

The aim was to maximize the water allocation from Ukai reservoir, at various depend-
able inflows. The Weibull method was used to find the different dependable inflow
(Subramanya 2013). The different population sizes were used, i.e. 10, 20, 40, 60,
80, 100 and 150 and the iteration size as 50,000. The internal parameters in DE, i.e.
the lower and upper scaling factors were taken as 0.2 and 0.8, respectively, and the
crossover probability was taken as 0.2. The internal parameters in PSO, i.e. the social
and cognitive parameters were taken as 2.0 and 1.5, respectively, and inertia weight
as 1.
Table 16.2 shows the comparative results obtained using ETLBO, TLBO, PSO
and DE, in the form best, worst and average solution. The results were averaged
over 15 different runs. A very small standard deviation and coefficient of variation
have been observed for all the algorithms. Table 16.2a shows the best results over
15 different runs, wherein ETLBO, TLBO and PSO were having same results, and
DE and LP results were almost similar but less than other three algorithms. Table
16.2b shows the average results obtained over 15 different runs, and it was found
16 Optimization of Water Allocation for Ukai Reservoir … 201

Table 16.2 Comparative results (a) best, (b) average and (c) worst, at various dependable inflows
Results Algorithm Maximum allocation in MCM
60% 65% 70% 75%
(a) Best ETLBO 3224.520 4023.190 4672.770 5351.020
TLBO 3224.520 4023.190 4672.770 5351.020
PSO 3224.520 4023.190 4672.770 5351.020
DE 3219.339 4022.210 4672.160 5350.420
LP 3223.920 4022.579 4672.170 5350.420
(b) Average ETLBO 3224.520 4023.190 4672.770 5351.020
TLBO 3224.515 4023.187 4672.764 5351.018
PSO 3224.513 4023.183 4672.747 5351.014
DE 3219.249 4022.149 4672.123 5350.331
(c) Worst ETLBO 3224.520 4023.190 4672.770 5351.020
TLBO 3224.490 4023.160 4672.730 5351.010
PSO 3224.480 4023.150 4672.600 5350.980
DE 3219.039 4022.010 4672.000 5350.130
Bold value shows the best solution for (a), (b) and (c) separately

that ETLBO results were better as compared to other algorithms. Table 16.2c shows
the worst results obtained over 15 different runs, and it was observed that ETLBO
results were better as compared to other algorithms. Overall it was found that ETLBO
performed better when compared to TLBO, PSO, DE in term of average results and
worst solution. The global optimum solution obtained using ETLBO for 60%, 65%,
70% and 75% dependable inflow were 3224.520, 4023.190, 4672.770 and 5351.020,
respectively, in MCM, and similar were obtained using TLBO and PSO. But it was
found that ETLBO had better uniformity in the solution as compared to TLBO and
PSO. DE and LP results were inferior than other algorithms. Figure 16.3a–d shows
the convergence curve for water allocated having 60, 65, 70 and 75% dependence
flow. It was found that ETLBO had better convergence rate as compared to other
algorithm in all the dependable inflow. Note, iteration up to 5000 is only shown.
Next, the best optimum solution obtained from ETLBO at different dependable
inflow water is compared with the actual water releases. Figure 16.4 shows the
comparison between the actual and computed water releases. Figure 16.4, it was
observed that at 60% dependable inflow much variations were observed with actual
releases of water than other dependable inflow. At 75%, least variation has been
observed as compared with actual releases of water.
202 V. Kumar and S. M. Yadav

Fig. 16.3 Convergence curve (a) 60% (b) 65% (c) 70% (d) 75% dependable inflow

Fig. 16.4 Comparison of actual and computed release for different dependable inflow
16 Optimization of Water Allocation for Ukai Reservoir … 203

16.6 Conclusion

The present study introduces an improved version of TLBO, by introducing the elitist
concept. The objective was to maximize the water allocation from Ukai reservoir to
supply water for various needs at different dependable inflow. ETLBO and TLBO
were chosen because it does not require internal parameters. To turn the algorithm as
it is commonly observed in other algorithm, it saves a lot of time. The results from
ETLBO were compared with ordinary TLBO, PSO, DE and LP. It was found that
ETLBO performed better when results are compared as average and worst solution.
ETLBO results were uniform, faster convergence rate and better global search ability.
When comparison was done with actual releases, it was found that 75% dependable
inflow variations were less compared with other dependable inflows. Thus, it is
suggested to operate the reservoir as per the 75% dependable inflow level model.

References

Afshar MH (2013) Extension of the constrained particle swarm optimization algorithm to optimal
operation of multi-reservoirs system. Int J Electr Power Energy Syst 51:71–81
Asgari HR, Bozorg Haddad O, Pazoki M, Loáiciga HA (2016) Weed optimization algorithm for
optimal reservoir operation. J Irrig Drain Eng 142(2):04015055
Baghlani A, Makiabadi MH, Maheri MR (2017) Sizing optimization of truss structures by an
efficient constraint-handling strategy in TLBO. J Comput Civ Eng 31(4):04017004
Bahrami M, Bozorg-haddad O, Chu XF (2018) Application of Cat Swarm Optimization Algorithm
for Optimal Reservoir Operation. J Irrig Drain Eng 144(1):04017057. [Link]
(ASCE)IR.1943-4774.0001256
Bai T, Kan Y, Chang J, Huang Q, Chang FJ (2017) Fusing feasible search space into PSO for
multi-objective cascade reservoir optimization. Appl Soft Comput 51:328–340
Baykasoǧlu A, Hamzadayi A, Köse SY (2014) Testing the performance of teaching-learning based
optimization (TLBO) algorithm on combinatorial problems: flow shop and job shop scheduling
cases. Inf Sci 276:204–218
Bayram A, Uzlu E, Kankal M, Dede T (2015) Modeling stream dissolved oxygen concentration
using teaching–learning based optimization algorithm. Environ Earth Sci 73(10):6565–6576
Chang FJ, Wang KW (2013) A systematical water allocation scheme for drought mitigation. J
Hydrol 507:124–133
Chang LC, Chang FJ, Wang KW, Dai SY (2010) Constrained genetic algorithms for optimizing
multi-use reservoir operation. J Hydrol 390(1–2):66–74
Eberhart R, Kennedy J (1995) A new optimizer using particle swarm theory. In: MHS’95. Proceed-
ings of the sixth international symposium on micro machine and human science. IEEE, pp
39–43
Ehteram M, Karami H, Mousavi SF, El-Shafie A, Amini Z (2017) Optimizing dam and reservoirs
operation based model utilizing shark algorithm approach. Knowl -Based Syst 122:26–38
Hossain MS, El-shafie A (2014) Performance analysis of artificial bee colony (ABC) algorithm in
optimizing release policy of Aswan high dam. Neural Comput Appl 24(5):199–1206
Kumar ND, Reddy JM (2007) Multipurpose Reservoir Operation Using Particle Swarm Optimiza-
tion. J Water Resour Plan Manag 133(3):192–201
Kumar V, Yadav SM (2018) Optimization of reservoir operation with a new approach in evolutionary
computation using TLBO algorithm and Jaya algorithm. Water Resour Manag 32(13):4375–4391
204 V. Kumar and S. M. Yadav

Kumar V, Yadav SM (2019) Optimization of cropping patterns using Elitist-Jaya and Elitist-TLBO
algorithms. Water Resour Manage 33(5):1817–1833. [Link]
04-z
Kumar V, Yadav SM (2020a) Self-adaptive multi-population based Jaya algorithm to optimize the
cropping pattern under a constraint environment. J Hydroinformatics 1–18. [Link]
2166/hydro.2019.087
Kumar V, Yadav SM (2020b) Optimization of water releases from Ukai reservoir using Jaya algo-
rithm. Advances in intelligent systems and computing, vol 949. Springer, Singapore. [Link]
org/10.1007/978-981-13-8196-6_29
Ming B, Chang JX, Huang Q, Wang YM, Huang SZ (2015) Optimal operation of multi-reservoir
system based-on cuckoo search algorithm. Water Resour Manag 29(15):5671–5687
Rao RV, Savsani VJ, Vakharia DP (2011) Teaching-learning-based optimization: a novel method for
constrained mechanical design optimization problems. CAD Comput Aided Des 43(3):303–315
Rao RV, Savsani VJ, Vakharia DP (2012) Teaching-learning-based optimization: an optimization
method for continuous non-linear large scale problems. Inf Sci 183(1):1–15
Regulwar DG, Choudhari SA, Raj PA (2010) Differential evolution algorithm with application to
optimal operation of multipurpose reservoir. J Water Resour Prot 02(06):560–568
Shourian M, Mousavi SJ, Tahershamsi A (2008) Basin-wide water resources planning by integrating
PSO algorithm and MODSIM. Water Resour Manag 22(10):1347–1366
Storn R, Price K (1996) Minimizing the real functions of the ICEC’96 contest by differential
evolution. In: Proceedings of IEEE international conference on evolutionary computation. IEEE,
pp 842–844
Subramanya K (2013) Engineering hydrology. Tata McGraw-Hill Educ 45–60
Chapter 17
Prediction of Reservoir Submerged
Sediment Density

Y. C. Jabbar and S. M. Yadav

Abstract This study proposes a density prediction model for reservoir sediment
deposition using artificial neural networks (ANNs). To compute the reservoir capacity
loss, it is necessary to estimate the weight per unit submerged sediment volume as
the sediment transported in a river is measured in gravimetric terms as sediment load.
Understanding the sensitivity of the estimated density to relate catchment sediment
yield with the reservoir deposition rate, ANNs are utilized to precisely compute
submerged sediment density. A dataset with 262 field observed densities for the
reservoirs which always remain submerged is prepared and used. Three input vari-
ables, sand, silt, and clay proportion are selected. Then, the method of training and
validation of the ANNs for the density prediction is presented. The model results
show that the ANN model is flexible and robust to capture the complex physical
process and is better than the Lara Pemberton empirical relationship. The proposed
trained network, having the best predictive capability, is given as a MATLAB code.

Keywords Sediment volume · Reservoir · Deposition · Density · ANN

17.1 Introduction

To determine the economic life of a dam in a sediment-bearing river, the most impor-
tant challenge is to estimate the dead storage space. The dead storage allocation is
done based on the estimated sediment deposition in the course of the useful life of
a reservoir. The gradual reduction in reservoir capacity, with time, depends on the
sediment load trapped by the reservoir and on the density of deposition (Jabbar and
Yadav 2019a, b). Sediment load can be precisely estimated in gravimetric terms from
the hydrometric observations, whereas the density of deposited sediments is variable

Y. C. Jabbar (B)
Department of Civil Engineering, Sitarambhai Naranji Patel Institute of Technology & Research
Centre, Umrakh, Surat, Gujarat 394601, India
S. M. Yadav
Department of Civil Engineering, S. V. National Institute of Technology, Surat, Gujarat 395007,
India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 205
R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
206 Y. C. Jabbar and S. M. Yadav

in the areal extent (of the reservoir submergence-area) and increases with the depth
of deposition. The submerged density of the deposited sediment may range from 363
to 2000 kg/m3 (Strand and Pemberton 1982).
In an existing built dam-reservoir, the sediment distribution and subsequent vari-
ation of the density can be studied by sampling the sediments. Point sampling repre-
sents the density of the sediment deposited at a point. In practice, however, capacity
loss determination requires the knowledge of bulk-density over the submerged area
of reservoir (Jabbar and Yadav 2019a, b). To convert the point sampled density into
a reservoir-representative bulk-density, huge number of sampling points should be
established. Such multi-point samplings are not feasible in large reservoirs, as the
reservoir can be deep up to hundreds of meters (Annandale et al. 2016). Estimating
a value of density from secondary data is carried out to predict the reservoir sedi-
mentation and capacity loss from the hydrometric observations (Bussi et al. 2013;
Marineau and Wright 2017). Sand, silt, and clay content of sediments are associated
with the initial weight per unit volume for different operational characteristics of the
reservoirs.
Lane and Koelzer (1943) first highlighted the importance of the density of sedi-
ments deposited in reservoirs. They have analyzed the data from different rivers and
reservoirs and developed a relationship to predict initial density. Considering the
effect of consolidation of clay and silt for a period of T-years, the density computa-
tion was proposed. For sediment composed of sand, silt, and clay material, the unit
weight for each material is multiplied to their respective fraction (Eq. 17.1).
   
Pc Pm
Wi = Wc + k log10 T + Wm + k log10 T
100 100
 
Ps
+ Ws + k log10 T (17.1)
100

where, Wi = Initial density in kilograms per cubic meter.

K c , K m and K s are constants of clay, silt, and sand, respectively, based on type
of reservoir operation.
Pc , Pm and Ps are percentages of clay, silt, and sand, respectively, of the incoming
sediment.
Wc , Wm and W S are coefficients of clay, silt, and sand, respectively of the incoming
sediment as given in Table 17.1. Thus, initial density, i.e., density of a year of
deposited sediment can be expressed as Eq. 17.2.

Pc Pm Ps
Wi = Wc + Wm + Ws (17.2)
100 100 100
Lara and Pemberton (1965) modified the coefficients of density (Ws , Wm and Wc )
with the use of additional data on the density of deposited material from reservoir
surveys (Table 17.2). They have recommended Miller (1953) consolidation equation
to compute density for a period of T-years. As a higher number of data is utilized to
17 Prediction of Reservoir Submerged Sediment Density 207

Table 17.1 Initial weight as per Lane and Koelzer (1943) for different reservoir operations
Reservoir operation Wc Kc Wm Km Ws Ks
(kg/m3 ) (kg/m3 ) (kg/m3 )
Type-1a 480 256.3 1041 91.3 1490 0
Type-2b 787 171.4 1185 43.3 1490 0
Type-3c 961 96.1 1265 16.0 1490 0
Type-4d 1250 0 1314 0 1490 0
Note The values under the table are originally (Lane and Koelzer 1943) in the lbs./cu. Ft. which are
converted to kg/m3
a Sediment always submerged or nearly submerged (type-1)
b Normally moderate to considerable reservoir drawdown (type-2)
c Reservoir normally empty (type-3)
d River bed sediments (type-4)

Table 17.2 Lara and

Reservoir operation Wc Wm Ws
Pemberton (1965) modified
(kg/m3 ) (kg/m3 ) (kg/m3 )
the initial weight for different
reservoir operations Type-1a 416 1121 1554
Type-2b 561 1137 1554
Type-3c 641 1153 1554
Type-4d 961 1169 1554
a Sediment always submerged or nearly submerged (type-1)
b Normally moderate to considerable reservoir drawdown (type-2)
c Reservoir normally empty (type-3)
d River bed sediments (type-4)

obtain modified initial weight coefficients (Table 17.2), Lara and Pemberton (1965)
advised the use of the modified coefficients for the determination of initial weight
values. No further modification in Lara and Pemberton (1965) initial weight coeffi-
cients has been done, and the same are used up to date to determine the initial unit
weight (Marineau and Wright 2017). With the complex relationship between the sedi-
ment size fraction and initial density, empirically derived (regression) relationships
have shortcomings in depicting the nonlinearity. In addition to it, the proportion
of sediments are interrelated to each other (multicollinearity). Thus, identifying a
relationship from them using the regression approach is not advised (Donald and
Glauber 1967). The fraction of sand, silt, and clay has to add up to be one (100%); an
increase in the proportion of one may consequently decrease the other two fractions
of sediments. This is problematic; it can be hard to unravel which of the sediment
fraction best describes its shared variance with density. On the other hand, artificial
neural networks (ANNs) have proven fit to model the behavior of extremely complex
processes in numerous fields (Yegnanarayana 2009).
In the present work, an attempt is made to model the impact of particle size
fraction on the initial density of deposition for the sediments, which always remain
submerged. Besides, no study exists in the literature yet to model the density estimates
208 Y. C. Jabbar and S. M. Yadav

using ANN techniques. After several trials with field data, the architecture of ANN
model is determined. The network training is performed by the inputs of particle size
fractions (sand, silt and clay) and targets as observed/sampled densities of different
reservoirs.

17.2 Materials and Methods

17.2.1 Data

In this study, field data published by Lara and Pemberton (1963) are used. These
data were collected from different reservoirs having sediment always remained
submerged. The statistics of the sampled density is presented in Table 17.3. It is
observed from the sampled density statistics that the standard deviation of the density
observed is between 35 and 493 kg/m3 . The standard deviation of sand, silt, and
clay was observed to be 3.19, 3.32, and 1.30, respectively, for the reservoir with a
minimum standard deviation of observed density.

17.2.2 ANN Model Design

A neural network has proved itself as an elite soft computing tool and can be applied
to analyze, predict, or forecast the problems pertaining to curve fitting (Liu et al. 2013;
Noori and Kalin 2016; Safari et al. 2016; Yang et al. 2009), pattern recognition, clus-
tering, and dynamic time series (Raman and Sunilkumar 1995). With the advance-
ment of computing power, deep and shallow neural networks have gained importance
in the field of mapping complex and nonlinear relationships. Curve fitting problem is
addressed in the present study by ANN using MATLAB scripts. Sand, silt, and clay
proportion of the reservoir deposited sediments are given as inputs to estimate the
output of field density. The network is developed to predict the deposited sediments
density for the type of reservoir in which sediment always remain submerged. A
shallow feed-forward neural network of two layers, consisting hidden, and output
layer is designed (Basheer and Hajmeer 2000). If data grouping is not done correctly,
the network may result in a very rigid memory-based network (Bowden et al. 2002;
Maier and Dandy 2000). The efficiency of the network to map complex nonlinear
relationships is directly influenced by the number of nodes selected in network design
(Tokar and Johnson 1999). Thus, neural networks are designed in the present study by
changing two parameters. One, grouping of data in different proportions for training,
calibration, and validation. Second, number of nodes in the hidden layer is varied.
The network is trained with Levenberg–Marquardt backpropagation algorithm. The
transfer function computes the net layer output from its net input. Without the use
of a transfer function, the neural network may merely reduce to a group of linear
Table 17.3 Statistics of the sampled density observations
Sr. No Reservoir Location Surveyed by Sample no Observed density statistics Average of
Max Min Average Standard Clay Silt Sand
deviation
1 Harry strunk Nebraska USBRa 22 1590.63 480.55 1144.45 339.31 14.39 61.81 22.67
2 Leavenworth Kansas SCSb 9 1169.35 592.68 923.73 213.11 35.18 61.71 3.11
country state
lake
3 Lake Calhoun Illinois SCS 2 1089.26 760.88 925.07 232.20 32.15 67.60 0.25
4 Lake Bracken Illinois SCS 3 776.90 584.67 664.23 100.30 41.90 57.30 0.80
5 Lanecaster South Carolina SCS 7 1263.86 631.13 1011.45 209.90 28.77 62.69 8.54
reservoir
6 Conchas New Mexco CEc 8 1345.55 331.58 755.67 334.46 39.63 47.25 13.13
reservoir
7 Tarryall Colorado FSd 3 1400.01 499.78 1066.30 493.21 8.53 32.67 58.63
reservoir
17 Prediction of Reservoir Submerged Sediment Density

8 Lake McBride Iowa SCS 5 993.14 897.03 948.29 34.73 17.80 80.00 2.20
9 Lake Olathe Kansas SCS 1 961.11 961.11 961.11 – 22.40 22.10 55.50
10 Lake EI Kansas SCS 1 1057.22 1057.22 1057.22 – 15.20 27.10 57.70
Darado
11 Fort supply Oklahoma CE 26 1920.61 488.56 990.80 472.93 47.85 23.19 29.35
reservoir
12 Lake Texoma Oklahoma and CE 59 1729.99 560.65 1197.31 316.21 24.76 41.92 33.02
Texas
13 Lake Illinois ISWSe 16 1349.40 439.87 739.05 263.26 81.56 13.81 4.63
Bloomington
(continued)
209
Table 17.3 (continued)
210

Sr. No Reservoir Location Surveyed by Sample no Observed density statistics Average of

Max Min Average Standard Clay Silt Sand
deviation
14 Lake Decatur Illinois SCS 58 1129.46 319.89 665.57 188.79 50.89 46.64 2.38
15 La Von Texas CE 29 1177.36 352.41 799.65 253.26 63.89 32.14 3.97
reservoir
16 Hulah Oklahoma CE 13 1382.39 584.67 815.83 239.12 49.46 38.77 11.77
reservoir
a Bureau of Reclamation (BR), US Department of the Interior Coast
b Soil Conservation Service (SCS), US Department of Agriculture
c Corps of Engineers (CE), US Department of the Army
d Forest Service (FS), US Department of Agriculture Geological
e Illinois state water survey
Y. C. Jabbar and S. M. Yadav
17 Prediction of Reservoir Submerged Sediment Density 211

Fig. 17.1 Flowchart schematically presenting the neural network

functions (Jankowski and Duch 2001; Karlik and Olgac 2011; Yonaba et al. 2010).
The transfer function of the hidden layer is selected as a tan-sigmoidal function
(Eq. 17.3), and for the outputlayer, it is set as a linear function (Eq. 17.4).

1
sig(x) = (17.3)
1 + e−x

lin(x) = mx (17.4)

where x will be the output from the neurons, which is to be transferred, and m is
the slope of the linear transfer function. The transferred value will be an input to the
next layer or can be the output itself. The flowchart of the developed neural network
is presented in Fig. 17.1.

17.3 Results and Discussion

For training, calibration, and testing process, data is to be divided into three partitions.
In the present study, ANNs are developed using three sets of such partitioned data.
Partitioning is carried by random selection of data. The first set has training data of
60%, 20% of calibration, and 20% of testing data (60, 20, and 20). The second set
contains 70%, 15%, and 15% (70, 15, and 15) of training, calibration, and testing
data. The third set comprises 80%, 10%, and 10% (80, 10, and 10) of training testing
and calibration data. Number of neurons (n) is varied as three, four, five, and six
for the hidden layer. The developed neural networks with varying data partition and
number of neurons are presented in Table 17.4. The performance of the networks has
been evaluated by comparing the predicted and observed densities. Nash–Sutcliffe
model efficiency factor (E), Index of agreement (I), mean percentage error (P), and
coefficient of determination (r 2 ) are the error statistics computed for all the stages
(training, validation, and testing) and are presented in Table 17.5. Analyzing the data,
212 Y. C. Jabbar and S. M. Yadav

Table 17.4 Developed ANN details

Sr. No. Number of data points No. of neurons Network Id
Training Calibration Testing
1 60 20 20 3 ANN1
2 60 20 20 4 ANN2
3 60 20 20 5 ANN3
4 60 20 20 6 ANN4
5 70 15 15 3 ANN5
6 70 15 15 4 ANN6
7 70 15 15 5 ANN7
8 70 15 15 6 ANN8
9 80 10 10 3 ANN9
10 80 10 10 4 ANN10
11 80 10 10 5 ANN11
12 80 10 10 6 ANN12

it is found that all the developed neural networks performed well; the best network
based on the performance is ANN5. The regression plot of the observed and ANN
predicted density is presented in Fig. 17.2.
Performance of the empirically derived density (Lara and Pemberton 1965) esti-
mate and neural network predicted density estimate can be noted from the coeffi-
cient of determination. The coefficient of determination between Lara and Pemberton
(1965) derived density and observed density is found as 0.6980, while the observed
and ANN5 derived density has a coefficient of determination of 0.7285 (Fig. 17.3).
The Nash-Suctliff model efficiency factor (E) and Index of agreement (I) obtained by
comparing the observed and ANN5 model derived density are 0.91 and 0.73, respec-
tively, while E and I obtained by comparing the observed and Lara and Pemberton
(1963) model are 0.68 and 0.89 respectively. Thus, the developed ANN5 model is
recommended to be used further to predict sediment density (for the reservoirs in
which sediments always remained submerged). ANN model as a neural network
code is given in Annexure 1. The code contains the weights and biases of the trained
network, which is a ready to use MATLAB script.

17.4 Conclusion

In the present study, artificial neural network (ANN) models were examined to esti-
mate reservoir deposited sediment density. For achieving this objective, 262 observed
field densities and their constituent sand, silt, and clay proportions were consid-
ered. Comparison of the ANN model results with Lara and Pemberton (1963) model
indicated that the ANN model (ANN5) is better than Lara and Pemberton (1963)
Table 17.5 Comparison of observed and predicted submerged reservoir sediment densities
Network Nash Sutcliffe model efficiency, Index of agreement, I Coefficient of determination, R2 Mean percentage error, P
ID E
Training Validation Testing Training Validation Testing Training Validation Testing Training Validation Testing
ANN1 0.704 0.815 0.672 0.90 0.94 0.90 0.730 0.645 0.744 6.1 5.6 10.6
ANN2 0.588 0.717 0.466 0.91 0.94 0.87 0.736 0.577 0.679 6.1 5.7 11.6
ANN3 0.600 0.783 0.597 0.90 0.95 0.90 0.733 0.658 0.756 7.7 7.3 11.7
ANN4 0.562 0.720 0.538 0.89 0.93 0.88 0.725 0.644 0.764 10.1 9.7 14.2
ANN5 0.732 0.775 0.668 0.91 0.93 0.90 0.734 0.666 0.760 5.6 5.2 10.2
ANN6 0.571 0.699 0.537 0.90 0.92 0.89 0.709 0.648 0.768 8.7 8.4 13.1
17 Prediction of Reservoir Submerged Sediment Density

ANN7 0.613 0.650 0.572 0.92 0.92 0.90 0.732 0.646 0.779 5.3 5.0 8.2
ANN8 0.623 0.701 0.413 0.91 0.92 0.87 0.733 0.611 0.745 7.5 7.1 12.8
ANN9 0.631 0.626 0.642 0.88 0.88 0.90 0.634 0.627 0.713 5.6 4.8 11.2
ANN10 0.334 0.489 0.619 0.83 0.88 0.90 0.557 0.651 0.754 9.9 10.0 14.9
ANN11 0.140 0.459 0.385 0.80 0.86 0.82 0.492 0.633 0.647 12.7 11.0 20.3
ANN12 0.319 0.312 0.676 0.84 0.87 0.91 0.562 0.622 0.738 5.0 5.7 10.3
213
214 Y. C. Jabbar and S. M. Yadav

Fig. 17.2 Comparison of observed and ANN predicted densities (ANN1) R2 = 0.7222, (ANN2)
R2 = 0.7176, (ANN3) R2 = 0.7279, (ANN4) R2 = 0.7208, (ANN5) R2 = 0.7285, (ANN6) R2 =
0.7068, (ANN7) R2 = 0.7284, (ANN8) R2 = 0.7215, (ANN9) R2 = 0.6515, (ANN10) R2 = 0.7191,
(ANN11) R2 = 0.6166, (ANN12) R2 = 0.7210; the black dotted line is the best fit line and the blue
line is the linear regression fit

model. The Nash–Sutcliffe model efficiency factor (E) and coefficient of determi-
nation (r 2 ) for Lara and Pemberton (1963) derived density estimates are 0.68 and
0.6980, respectively, which increased for ANN derived density estimates to 0.91 and
0.7285, respectively. Thus, the results confirm that the ANN5 model is flexible and
robust to capture the complex sediment deposition process and is better than the Lara
Pemberton empirical relationship.
17 Prediction of Reservoir Submerged Sediment Density 215

Fig. 17.3 Predictability of ANN5 derived density with the Lara Pemberton (1965) derived density

Annexure 1

function [y1] = DesnityPredictionFunction(x1)

%Density prediction function.
% [y1] = DensityPredictionFunction(x1) takes these arguments:
% x = 3xQ matrix, input #1
% and returns:
% y = 1xQ matrix, output #1
% where Q is the number of samples.
% ===== NEURAL NETWORK CONSTANTS =====
% Input 1
x1_step1.xoffset = [0;1;0];
x1_step1.gain = [0.0217391304347826;0.0233918128654971;0.0208333333333333];
x1_step1.ymin = -1;
% Layer 1
216 Y. C. Jabbar and S. M. Yadav

b1 = [-1.2768300157015601837;-2.012427608471574203;-2.5463928311399506299];
IW1_1 = [-0.38545642814567215861 3.6978354365474639387 -0.32068398027300082376;-
1.8172773609952270757 0.20324942901802645068 0.62662087339392069651;-
0.5754955583970926547 -0.78067914007961214384 -2.6582090817266075256];
% Layer 2
b2 = -0.059137336364163739511;
LW2_1 = [0.05781172921120955599 0.47204386618357141447 -0.65591831490209306921];
% Output 1
y1_step1.ymin = -1;
y1_step1.gain = 0.00124943775301115;
y1_step1.xoffset = 319.89;
% ===== SIMULATION ========
% Dimensions
Q = size(x1,2); % samples
% Input 1
xp1 = mapminmax_apply(x1,x1_step1);
% Layer 1
a1 = tansig_apply(repmat(b1,1,Q) + IW1_1*xp1);
% Layer 2
a2 = repmat(b2,1,Q) + LW2_1*a1;
% Output 1
y1 = mapminmax_reverse(a2,y1_step1);
end
% ===== MODULE FUNCTIONS ========
% Map Minimum and Maximum Input Processing Function
function y = mapminmax_apply(x,settings)
y = bsxfun(@minus,x,[Link]);
y = bsxfun(@times,y,[Link]);
y = bsxfun(@plus,y,[Link]);
end
% Sigmoid Symmetric Transfer Function
function a = tansig_apply(n,~)
a = 2 ./ (1 + exp(-2*n)) - 1;
end
% Map Minimum and Maximum Output Reverse-Processing Function
function x = mapminmax_reverse(y,settings)
x = bsxfun(@minus,y,[Link]);
x = bsxfun(@rdivide,x,[Link]);
x = bsxfun(@plus,x,[Link]);
end

References

Annandale GW, Morris GL, Karki P (2016) Extending the life of reservoirs: sustainable sediment
management for dams and run-of-river Hydropower. The World Bank. [Link]
978-1-4648-0838-8
Basheer IA, Hajmeer M (2000) Artificial neural networks: fundamentals, computing, design, and
application. J Microbiol Methods 43(1):3–31. [Link]
17 Prediction of Reservoir Submerged Sediment Density 217

Bowden GJ, Maier HR, Dandy GC (2002) Optimal division of data for neural network models
in water resources applications. Water Resour Res 38(2):2–1–2–11. [Link]
1WR000266
Bussi G, Rodríguez-Lloveras X, Francés F, Benito G, Sánchez-Moya Y, Sopeña A (2013) Sediment
yield model implementation based on check dam infill stratigraphy in a semiarid Mediterranean
catchment. Hydrol Earth Syst Sci 17(8):3339–3354. [Link]
Donald F, Glauber R (1967) Multicollinearity in regression analysis: the problem revisited author(s).
In: Farrar DE, Robert R. Glauber source : the review of economics and statistics, vol 49. no. 1.
The MIT Press Stable UR, pp 92–107. [Link]
Jabbar, Yazad C, Yadav SM (2019b) Development of reservoir capacity loss model using bootstrap-
ping of sediment rating curves. ISH J Hydraulic Eng 1–13. [Link]
2019.1665483
Jabbar YC, Yadav SM (2019a) Development of a relationship between hydrometric and hydro-
graphic observations to predict reservoir capacity loss. Acta Geophysica 67(5):1451–1469. https://
[Link]/10.1007/s11600-019-00323-0
Jankowski N, Duch W (2001) Transfer functions: hidden possibilities for better neural networks.
In: 9th European symposium on artificial neural networks, vol 7. pp 81–94
Karlik B, Olgac V (2011) Performance analysis of various activation functions in generalized MLP
architectures of neural networks. Int J Artif Intell Expert Syst (IJAE) 1(4):111–122
Lane EW, Koelzer VA (1943) Density of sediments deposited in reservoirs. St Paul US Engineer
District Sub-Office, Hydraulic Laboratory, University of Iowa, Iowa. [Link]
docs/Report_9.pdf
Lara, Pemberton EL (1963) Initial unit-weight of deposited sediments. In: Federal inter-
agency sedimentation conference. U.S.D.A., USA, pp 818–845. [Link]
Record/000212230%0A [Link]
Lara JM Pemberton EL (1965) Initial unit weight of deposited sediments. In: Federal inter-agency
sedimentation conference. U.S. Agriculture Research Service Publ 970, 818–845
Liu QJ, Shi ZH, Fang NF, Zhu HD, Ai L (2013) Modeling the daily suspended sediment concentra-
tion in a hyperconcentrated river on the Loess Plateau, China, using the Wavelet-ANN approach.
Geomorphology 186:181–190. [Link]
Maier HR, Dandy GC (2000) Neural networks for the prediction and forecasting of water resources
variables: a review of modelling issues and applications. Environ Model Softw 15(1):101–124.
[Link]
Marineau MD, Wright SA (2017) Daily reservoir sedimentation model: case study from the Fena
Valley reservoir, Guam. J Hydraul Eng 143(9):05017003. [Link]
1943-7900.0001344
Miller CR (1953) Determination of the unit weight of sediment for use in sediment volume
computations. US Department of the Interior, Bureau of Reclamation, Project Planning Division
Noori N, Kalin L (2016) Coupling SWAT and ANN models for enhanced daily streamflow
prediction. J Hydrol 533:141–151. [Link]
Raman H, Sunilkumar N (1995) Multivariate modelling of water resources time series using artificial
neural networks. Hydrol Sci J 40(2):145–163. [Link]
Safari MJS, Aksoy H, Mohammadi M (2016) Artificial neural network and regression models for
flow velocity at sediment incipient deposition. J Hydrol 541:1420–1429. [Link]
[Link].2016.08.045
Strand RI, Pemberton EL (1982) Reservoir sedimentation technical guidelines for Bureau of
reclamation. Denver, Colorado
Tokar AS, Johnson PA (1999) Rainfall-runoff modeling using artificial neural networks. J Hydrol
Eng 4(3):232–239. [Link]
Yang CT, Marsooli R, Aalami MT (2009) Evaluation of total load sediment transport formulas using
ANN. Int J Sedim Res 24(3):274–286. [Link]
Yegnanarayana B (2009) Artificial neural networks. PHI Learning Pvt. Ltd
218 Y. C. Jabbar and S. M. Yadav

Yonaba H, Anctil F, Fortin V (2010) Comparing sigmoid transfer functions for neural network
multistep ahead streamflow forecasting. J Hydrol Eng 15(4):275–283. [Link]
(ASCE)HE.1943-5584.0000188
Chapter 18
Micro-hydro Power Generation
in India—A Review

Aparna M. Deulkar, Vivek S. Chavhan, and Pankaj R. Modak

Abstract Hydropower technology has been around for more than a century.
Hydropower comes from converting the energy in flowing water—using a water
wheel or a turbine—into useful mechanical power. This power is then converted into
electricity by an electric generator. Micro-hydropower systems are small hydropower
plants that have an installed power generation capacity of less than 100 kilowatts
(KW). Many micro-hydropower systems operate “run of river,” which means that
no large dams or water storage reservoirs are built, and no land is flooded. Deple-
tion of fossil fuel and the inability to meet the rising demand of electricity is some
drawbacks for the economic development of India. This paper presents the study to
investigate the possibility of the micro-hydro power generation and its advantages in
India.

Keywords Micro-hydro power · Economic development · Installed power

generation capacity · Renewable source

18.1 Introduction

Energy is very important for all human activity. Development of country is depending
on available energy resources. Water is a natural source to generate the energy. In the
past decades, there is a world-wide problem of fossil fuel depletion, climatic change,
and increased electricity demand. (Anandh and Vinoth 2018) There is a fast depletion
of renewable sources that was used in the past for the generation of electricity, and the
difficulty in reachability of the grid supply to the remote villages was a big challenge

A. M. Deulkar (B) · V. S. Chavhan · P. R. Modak

Department of Civil Engineering, AISSMS COE, Kennedy Road, Near R. T. O. Office, Pune
411001, India
V. S. Chavhan
e-mail: vschavhan@[Link]
P. R. Modak
e-mail: prmodak@[Link]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 219
R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
220 A. M. Deulkar et al.

faced. The best possible remedial measure in this scenario is to make use of some
renewable energy source like hydro power, wind, biomass, etc., so that it will be first
step to reduce fossil fuel depletion. The aim of present work is to focus on micro-
hydro power generation in India as renewable energy source to fulfill energy demand.
Study of micro-hydro power plant focuses on three main folds such as technical as
well as economical feasibility studies, design of civil works, and selection of electro
mechanical components. There is a huge potential to develop a micro-hydro power
plant which would meet the energy demand of the tribal settlement in India and
thereby improving their living condition.
By considering advantages of micro-hydro power plant like no air pollution, no
waste product like nuclear power plant, water left from the dam or cannel can be
reuse, it provide provision for flood, and it is a clean source of energy. For rural
development and employment, micro-hydro power plant is the best solution. Since
small hydro power plants do not require long-term planning, less expensive, for rural
electrification over a complete lifetime, they are cheapest technology available. It is
noted every time that the development of large hydro power plant is resisted by local
community, environmentalist, and NGOs due to issues such as deforestation and
resettlement of the community (Höffken 2014). Another advantage of micro-hydro
power plant is they use available flow of river, and no storage is required.
India is blessed with large number of river flowing over a country. All these rivers
have great potential for hydro power which can be harnessed. Figure 18.1 shows the
basin wise potential of Indian River system (Chauhan and Vig 2017).
India is ranked third biggest in the USA, with 5.5% global share in 2016 (Anandh
and Vinoth 2018). Since India is largely dependent on fossil fuel imports to meet its
energy demands, about 70% of India’s electricity generation capacity is from fossil
fuels (Anandh and Vinoth 2018). To meet the future energy demands, it is essential
to tap all possible sources of small hydro energy using decentralized power gener-
ation (Anandh and Vinoth 2018). Micro-hydro power is the small-scale harnessing
of energy from falling water which converts hydraulic energy to electric energy. It
is cheapest solution for poor communities in rural areas with an affordable, easy

Fig. 18.1 Estimated hydro potential

18 Micro-hydro Power Generation in India—A Review 221

to maintain, and long-term solution to their energy needs. In India, the total hydro-
electrical potential is about 69% (including both large and small hydro projects
and those installed/under installation). Anandh and Vinoth (2018) According to the
international energy agency (IEA), large-scale hydroelectric plants currently supply
16% of the world’s electricity. However, such kind of projects requires tremendous
amounts of land impoundment, dams and flood control, and often they produce envi-
ronmental impacts (Singh 2009). Among the renewable energy source, small hydro
power contributes 13% of the total grid connected power generation, thereby consti-
tuting second largest grid connected system after wind power, as per the report by
Ministry of New and Renewable Energy. Michael, and Jawahar (2017) Micro-hydro
power is a type of hydroelectric power that typically produced up to 5–100 KW of
electricity using the natural flow of water. This type of power plant can provide power
to a small community.
Up to, November 2016, the installed capacity of hydropower in India is
43133.43 MW. This installed capacity of hydropower is about 13.97% of the total
installed capacity of the country. The current potential of hydropower in India is
about 1,48,701 MW. This potential is 3.45 times the installed capacity (Chauhan and
Vig 2017). Indian power is generally based on fossil fuel to move toward renewable
energy source and as 13.69% (Chauhan and Vig 2017) is the contribution by hydro
power plant. To promote micro-hydro power plant by considering advantage above
mentioned, it is necessary to take step toward micro-hydro power plant. Hence, this
paper gives a review of micro-hydro power generation in India the water resources,
current status, potential, and future of hydro energy in India.

18.2 Literature Review

This part is compiled with a review of past research work in the field of micro-hydro
in India. Purpose of this literature review is to find key for further research in this
field.
1. Anandh and Vinoth (2018) studied current scenario and future potential of
small hydropower plant India. Share of hydro power in our country is declining
persistently, from 46% in the year 1960 to 16.33% in the year 2015. Small-
scale hydropower is a concentrated energy source with a long-lasting and robust
technology, and its systems’ life can extend to up to 50 years or more without
any major new investments.
2. Dinesh Chauhan and Sunny Vig present paper gives a complete review of the
water resources, current status, potential, and future of hydro energy in India.
India ranks third after China and USA in the world in terms of numbers of
dams. But large hydro power plant required many years for full construction
and its commissioning of plant. Author suggest that besides the large hydro
potential, small hydro power plant is the best solution to fulfill rural energy
222 A. M. Deulkar et al.

demand by utilizing present hydro potential, and it also provide employment

to rural population.
During the study, author observed that hydro generating unit sizes has been
increased from 22 MW (from the independence) to 250 MW till today. Till
august 2016, only 34.29% of the northern region of India potential and near
about 2% of the north eastern region of India potential has been utilized. If
the hydro power potential in these region is utilized completely, then nation’s
dependence on thermal power plants can be reduced significantly.
3. Adamantia et al. (2017) studied about climate change with respect to growing
demand of energy. Globally averaged temperatures in 2016 were 0.99 °C
(1.78 °F). By studying advantages of renewable energy, i.e., micro-hydro power
plan authors provide a comprehensive assessment of the environmental impacts
of small hydropower plants (SHPs). Considered the problem of assessing the
environmental cost of a micro-hydro power (MHP) by using the approach of
external cost. The external cost was derived from the combination of life cycle
air pollutants emission factors for the three main construction components
(concrete, steel, aggregates). The external cost was estimated to play a non-
negligible part in the total investment cost that can affect the feasibility of the
scheme. Also, more applications of the method on renewable energy sources
will better demonstrate its benefits and drawbacks through additional practice.
4. Michael and Jawahar (2017) studied energy scenario. According to study, at
the end of the Aug 2014, the power generation reached around 20,000 Mw only
by small hydro power plant with more than 15,000 MW potential remaining
unexploited. Authors consider Kerala state of India for study, which is having
seasonal discharge but more discharge in rainy season. Tribal village in Kerala
state does not have access to electricity. They use kerosene for lighting purpose.
To fulfill energy demand of tribal village, first energy requirement and energy
potential are calculated. Available discharge is reduces by 80% to account
for various losses. By considering losses, electro mechanical equipment (like
Turbine, Generator) is selected for power generation of 120 family. This
study resulted with micro-hydro power plant is found to be technically and
economically viable as operation cost is 60 Lakhs.
5. Manual (2015) gives brief idea, advantages, and challenges associated to the
micro-hydropower. India is endowed with rich hydropower potential to the
tune of 148 GW. India ranks fifth in the world in terms of usable hydro power
potential. India has around 36 GW of installed hydropower capacity, whereas
an additional 13 GW is under construction.
6. Saxena (2013) studied potential of small hydro power plant in India. According
to his study potential of small hydro (upto 25 MW station capacity) in India is
of about 20,000 MW of which about 3632 MW has been exploited. In India,
hydro projects up to 25 MW station capacity have been categorized as SHP
projects and responsibility of small hydro development rests with Ministry of
New and Renewable Energy (MNRE). India has a history of about 110 years
of hydropower. The first small hydro project of 130 KW commissioned in the
hills of Darjeeling in 1897 marks the development of hydropower in India.
18 Micro-hydro Power Generation in India—A Review 223

The total hydroelectric power potential in the country is assessed at about

150,000 MW, equivalent to 84,000 MW at 60% load factor. The potential of
SHP projects is estimated at about 20,000 MW. Of this, 6474 potential sites
with an aggregate capacity of 19,749 MW have been identified. Today the SHP
program in India is essentially private investment driven. In fact, 329 private
sector SHP projects of about 1748 MW capacity have been setup. Private sector
entrepreneurs are finding attractive business opportunities in small hydro and
state governments also feel that the private participation may be necessary for
tapping the full potential of rivers and canals for power generation. Author
concludes that appropriate selection of sites and sizing of projects to give
higher plant load factors are considered important to further improve economic
viability of commercial SHP projects.
7. Bhoi and Ali (2014) studied hydro power plant in India and its environment
impact also studied HYDRO power plant projects in India. Hydro power can be
classified as: large hydro power, medium hydro power, and small hydro power.
These are classified according to the power generation capacity. Large hydro
power: >100 MW, medium hydro power: 30–100 MW, small hydro power: 1–
30 MW. Very small-scale hydro power plant classified as: Mini-hydro power
whose capacity is between 100 KW and 1 MW and micro-hydro ranging up
to 100 KW. Authors studied component part of hydro power plant along with
turbine types according to head. Kaplan and Propeller 2 < H < 40 Francis,
10 < H < 350 Pelton, 50 < H < 1300 (head in m). In environmental impact,
author explains advantages and problem associate with hydropower like plant
require suitable site, weather condition, local habitat, climatic condition, flow
of water, and head. A large part of the land area is required to install a hydro
project so may create disturbance to the local habitat. Requirement of large
area is meeting by the afforestation programs which disturb the ecosystem.
Sometime many aquatic animals are also affected by the construction of dam
across water mass.
8. Saxena and Kumar (2010) studied hydropower development in India. In which
they studied whole country is divided into five power regions, and planning is
done on a regional concept. At the time of independence in the year 1947, only
1362 MW of electricity was produced in India. After that installed capacity of
power generation had grown to 164,509 MW of which hydro is 37086 MW
(25%), thermal is 106,433 MW (65%), nuclear is 4560 MW (2.9%), and renew-
able energy sources 16429 MW (7.7%). The share of small-scale hydropower
(SHP) is 2820 MW. The potential of small hydro power projects is estimated
at about 15,384 MW with 5718 potential identified sites.
9. Baidya (2006) studied Indian scenario of small hydro power and its advantages
as a renewable resource. Author explains importance of small hydro power
(SHP) and how it is optimally and viably source of energy to improve economic
condition of the people in the village area and overall development of the
country. This in turn help in capacity addition to the hydro power generation,
and shortage of electricity can be avoided to some extent.
224 A. M. Deulkar et al.

10. Vyas et al. (2015) studied micro-hydel power system design and its implemen-
tation in Rajasthan. In this study, authors observed that though India blessed
with hydro potential, only marginal amount of power is so far tapped from the
renewable source, because these potential sites are located in difficult terrains
and remote access. Considering the design and the cost estimation, it can be
concluded that the small hydro projects are feasible in the arid and semi-
arid regions of Rajasthan. Author suggested that by considering important
aspects like proper maintenance, providing silt excluders we can improved
micro-hydro potential as renewable source.
11. Choudhury and Ghosh (2013) focus on responsible hydro power development
in India. As hydro power is important and renewable source of energy, but this
article is focus on “responsible” hydropower development. Means hydropower
development should more stable and sustainable investment for medium-to-
long term.
12. Souza and Donald (2015) studied need for hydropower in India, hydro poten-
tial, and growth and share of hydroelectric installed capacity and generation.
Share of hydro power in India is 16.9% of the total installed capacity of
237,742.94 MW as on February 28, 2015, and 4.4% global install capacity
and ranked 6th in the list of global nations. The hydropower generation for
2012–12 and 2013–14 stood at 12.5 and 14% of the total energy generation.
As against the power generation target of 122,263 MU for 2013–14, genera-
tion from hydroelectric power stations (above 25 MW Installed Capacity) was
134,847.52 MU, which was 10.29% more than the target.

18.3 Conclusion

During literature study, it was observed that micro-hydropower is the best solution
to overcome growing energy demand. Some work has been carried out related to
the same, but there is need to more focus on micro-hydro power development in
sustainable way in state of dependency on hydropower. It is observed that to improve
livelihood of villagers and also remote villages’ micro-hydro power is the solution.
Hydro is a renewable resource that is replenished by the environment over a
relatively short period of time. Water is neither depleted nor its composition altered
during the generation cycle. A run of river plant stores water in the weir for a short
duration, and the water is returned to water cycle on each day. Small hydro projects
(90%) efficient in utilization of the resource than solar (15–20%), wind (35%) and
other renewable energy sources. In long term, small hydro schemes have the least
impact among the environmental indicators like acidification, climate change, ozone
layer depletion, photo chemical oxidation (smog), etc. Capital subsidy for small hydro
power should be increased, and better mechanism to deliver the subsidy should be
devised. Incentive available to small hydro projects up to 25 MW may be extended to
the hydro power projects up to 100 MW. Small hydro power plants have advantage
of life span almost 50 years, i.e., more than twice the life span of other renewable
sources like wind, solar, etc.
18 Micro-hydro Power Generation in India—A Review 225

References

Adamantia ZV, Didaskalou E, Georgakellos D (2017) Article on financial appraisal of small hydro-
power considering the cradle-to-grave environmental cost: a case from Greece
Anandh T, Vinoth R (2018) A comprehensive assessment of small hydro power in India—current
scenario and future potential. J Mech Prod Engg 8: 413–424
Baidya G (2006) Development of small hydro. Himalayan small hydropower summit. pp 34–43
Bhoi R, Ali SM (2014) Potential of hydro power plant in India and its impact on environment. J
Eng 10:114
Chauhan D, Vig S (2017) A review of present status and potential of hydro power In India. Int J
Mod Trends Eng 4:2349–9745
Choudhury N, Ghosh A (2013) Responsible hydropower development in India. Manual by the
Council on Energy, Environment and Water for the Independent Power (EEW)
Höffken JI (2014) A closer look at small hydropower projects in India: Social acceptability of two
storage-based projects in Karnataka. pp 155–166
Hydro power in India key enables for a better tomorrow (2015) Pwc manual. [Link]
Maual Saxena P (2013) Renewable energy akshayUrja 6:24
Michael PA, Jawahar CP (2017) Design of 15 KW micro hydro power plant for rural electrification at
valara. In: 1st International conference on power engineering, computing and control, PECCON,
pp 163–171
Saxena P, Kumar A (2010) Hydropower development in India. In: Conference alternate hydro energy
centre, Indian Institute of Technology Roorkee
Singh D (2009) Micro-hydro-power, resource assessment handbook, an initiative of the Asian and
Pacific Center for Transfer of Technology
Souza SMD, Donald J (2015) Green growth and hydropower in India. Draft by The Energy and
Resources Institute (Teri)
Vyasa A, Guptab NK, Guptac SK, Gautamd P, Jehooe AS (2015) Mini/micro hydel power system
design and its implementation in Rajasthan. In: International conference on water resources,
coastal and ocean engineering (Icwrcoe), Aquatic Procedia 4:1537–1544
Chapter 19
Runoff Simulation and Irrigation Water
Requirement for Barman Command

A. Vishwakarma, M. K. Choudhary, and M. S. Chauhan

Abstract The study is based on runoff simulation using the SIMHYD and AWBM
model of the Narmada River at Barman. The NSE and R2 values of AWBM for cali-
bration are 0.751 and 0.821 and for the validation are 0.797 and 0.862, respectively,
while the NSE and R2 values of SIMHYD for calibration are 0.814 and 0.731, and
for validation period 0.700 and 0.755, respectively. AWBM accuracy values indicate
a better agreement between the observed and simulated runoff than the SIMHYD
model. It is concluded that AWBM is more suitable for Barman command to simu-
late the basin’s hydrological response to the rainfall and predict daily runoff with a
better degree of accuracy. Further study is carried out to estimate the irrigation water
requirement based on rainfall pattern knowledge to achieve the proper crop calendar
of Barman command. For estimating the irrigation water requirement of a wheat
crop, 12 years (2001–2012) average climatic data, including 12 years (2001–2012)
of annual rainfall data, is used. The probability analysis of rainfall was (at 20, 50, and
80%) conducted to better understand rainfall behavior. CROPWAT model is used to
calculate the reference evapotranspiration (ETO ) by using climatic parameters (like
maximum and minimum temperature, relative humidity, wind velocity, and sunshine
hours). It is found that the rainfall for dry, wet, and at normal probability year are
598 mm, 1168 mm, and 866 mm, respectively, and corresponding irrigation water
requirement estimated by CROPWAT are 274 mm, 264.7 mm, and 269.5 mm.

Keywords American water balance model (AWBM) · Coefficient of determination

(R2 ) · Nash–Sutcliffe efficiency (NSE) · Reference evapotranspiration · Irrigation
water requirement · CROPWAT 8.0

A. Vishwakarma (B) · M. K. Choudhary · M. S. Chauhan

Department of Civil Engineering, Maulana Azad National Institute of Technology, Bhopal, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 227
R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
228 A. Vishwakarma et al.

19.1 Introduction

Estimation of direct runoff is a complicated and challenging process. Nowadays,

a large number of rainfall-runoff models are available with different levels of
complexity. A rainfall–runoff model is a mathematical model that describes catch-
ment and gives the relationship between precipitation and runoff (Podger 2004).
However, models with simple structure and least input providing adequate results are
most liked by the hydrologists (Boughton 2006). Many studies had been performed
to evaluate the applicability of various rainfall–runoff models (Jones et al. 2006;
Thimme et al. 2013; Haque et al. 2015; Li et al. 2015; Yu 2015; Sachan et al.
2016; Chouhan et al. 2016). In this study, two conceptual models, namely AWBM
(Boughton 2004; Boughton and Chiew 2007) and SIMHYD (Chiew and McMahon
1991; Chiew and Siriwardena 2005) are applied for the runoff simulation of the
Barman command of Madhya Pradesh, India.
Rainfall is not a constant phenomenon. It occurs as normal, deficit, and heavy as
its magnitude may be daily, monthly, or yearly (Manivannan et al. 2016). Hence, for a
complete analysis of rainfall, statistical approach is preceded for the Barman command
of Madhya Pradesh. CROPWAT model (Smith 1992) is used to calculate the irrigation
water requirement of the rabi crop of the study area for variable rainfall conditions.
Hence, it is concluded that how much water should be released from the proposed
dam upstream of the study area as per the rainfall condition of a particular year.

19.2 Study Area and Data

19.2.1 Study Area

The Barman is a town on the Narmada River banks located at a longitude of 79°01 19
E and latitude of 23°01 33 N in the Narsinghpur district, Madhya Pradesh, India,
as shown in Fig. 19.1. In the present study, runoff simulation and irrigation water
requirements of the Basin of Narmada River of Barman command have been assessed.

19.2.2 Data Collection

Daily rainfall data of 12 years from 2001 to 2012 has been collected. Reference
evapotranspiration (ETo) has been calculated using climatic data of Barman for the
period from years 2001–2012 by the CROPWAT model. It was assumed that the
average climatic data of Barman station represents its whole area. The ETo was
low in December (2.21 mm/day) and maximum (6.76 mm/day) during May, and the
average ETo was found to be 3.84 mm/day. The average climatic data is shown in
Fig. 19.2.
19 Runoff Simulation and Irrigation Water Requirement … 229

Fig. 19.1 Location of Barman in Madhya Pradesh, India

Fig. 19.2 Average climatic data of the study area

230 A. Vishwakarma et al.

19.3 Methodology

19.3.1 Net Irrigation Water Requirement

As per FAO manual no. 56 (Allen et al. 1998), “The irrigation water require-
ment represents the difference between the crop water requirement and effective
precipitation. The irrigation water requirement also includes additional water for
leaching of salts and compensating for non-uniformity of water application.” It can
be summarized as,


T
I Rn = (K c · E T o − Peff ) (19.1)
t=0

where:
IRn Net irrigation requirement (mm).
ET o Reference crop evapotranspiration (mm).
Peff Effective dependable rainfall (mm).
Kc Crop coefficient.
T Total growing period of crops.

19.3.2 Rainfall–Runoff Modeling

Rainfall–runoff models give the relationship between precipitation and overland flow.
Here, two models, AWBM and SIMHYD, were used to evaluate runoff simulation
of Barman command in Narmada. The essential data required for the setup of the
AWBM and SIMHYD model are listed below:
1. Area of the catchment in km2
2. Rainfall data daily time series, mm/day
3. Potential evapotranspiration data daily time series, mm/day
4. Observed discharge data daily time series, m3 /s.

19.3.3 Probability Analysis

For proper planning and management of irrigation water, rainfall of normal, dry, and
wet years can be calculated by using the statistical probability approach. The steps
involved are:
1. Calculation of yearly rainfall as per the available rainfall records.
2. Arrangement of yearly rainfall in descending order.
19 Runoff Simulation and Irrigation Water Requirement … 231

3. Calculation of the plotting position or can say the probability of exceedance of

rainfall by using Weibull’s formula of statistical probability, given as:

Fa = m/(N + 1) (19.2)

where
N number of records,
m rank number,
Fa plotting position.

4. Preparation of a graph on a semi-log scale between rainfall (mm/year) and its

probability of exceedance.
5. Calculation of yearly rainfall values at 20, 50, and 80% probability.
6. Calculation of the monthly values of the normal year as per the following
relationship:

Pnorm
Pinorm = Piav ∗ (19.3)
Pav
where
Piav Average monthly rainfall of the ith month.
Pinorm Monthly rainfall for the normal year of the ith month.
Pav Average yearly rainfall.
Pnorm Yearly rainfall at 50% probability of exceedance.
Similarly, values for dry and wet years can be calculated.

19.3.4 Performance Evaluation of Model

[Link] Coefficient of Determination, R2

Coefficient of determination is a method to evaluate the reliability of the model

between the simulated and observed runoff data. Mathematically, it is expressed as
follows:
⎛ ⎞2
n − −
⎜ (O i − O )(P i − P ) ⎟
R2 = ⎜
⎝
i=1
 ⎟
⎠ (19.4)
n − 2 n − 2
i=1 (O i − O ) i=1 (P i − P )

where O and P are observed and predicted values, respectively.

232 A. Vishwakarma et al.

[Link] Nash–Sutcliffe Criteria

It is defined as “one minus the sum of the absolute squared differences between the
calculated and observed values normalized by the variance of the observed values”
during the study (Nash and Sutcliffe 1970).
Mathematically, the formula is expressed as:
n
(O i − Pi )2
NSE = 1 − 1
(19.5)
n − 2
1 (O i − O )

where
Oi is the observed discharge,
Pi is the modeled or predicted discharge,
−
O is the mean of the observed discharge.

19.4 Results and Analysis

19.4.1 AWBM Calibration and Validation Charts

The AWBM model was calibrated for five years period from 2001 to 2005 and then
validated for the remaining period of three years from 2006 to 2008. The NSE graph
presenting a comparison between observed and simulated discharge during model
calibration (NSE = 0.751) and validation (NSE = 0.821) is shown in Fig. 19.3.
The R2 value obtained during calibration and validation is shown in Fig. 19.4 and
Fig. 19.5, respectively.

19.4.2 SIMHYD Model

[Link] SIMHYD Model Calibration and Validation Charts

The SIMHYD model was calibrated for five years period from 2001 to 2005 and then
validated for the remaining period of three years from 2006 to 2008. The NSE graph
presenting a comparison between observed and simulated discharge during model
calibration (NSE = 0.814) and validation (NSE = 0.731) is shown in Fig. 19.6.
The R2 value obtained during calibration and validation is shown in Fig. 19.7 and
Fig. 19.8, respectively.
19 Runoff Simulation and Irrigation Water Requirement … 233

Fig. 19.3 Nash–Sutcliffe efficiency of AWBM calibration and validation

Fig. 19.4 R2 chart for

calibration

Fig. 19.5 R2 chart for

validation
234 A. Vishwakarma et al.

Fig. 19.6 Nash–Sutcliffe efficiency of SIMHYD model calibration and validation

Fig. 19.7 R2 chart for

calibration

Fig. 19.8 R2 chart for

validation
19 Runoff Simulation and Irrigation Water Requirement … 235

19.4.3 Accuracy of Models

The accuracy of both the models (AWBM and SIMHYD) for calibration and valida-
tion has been summarized in Table 19.1. AWBM performed better than the SIMHYD
model.

19.4.4 Effective Rainfall

There are different methods exist to estimate the effective rainfall in the CROPWAT
8.0 model. USDA Soil Conservation Service Method is one of the most commonly
used methods. The average monthly effective rainfall of the Barman is shown in
Fig. 19.9.

19.4.5 Probability Analysis of Yearly Rainfall

Using Weibull’s formula, 12 years of rainfall records of Barman arranged in

descending order as shown in Table 19.2.

Table 19.1 Accuracy parameter values for calibration and validation

Model Accuracy criteria Calibration Validation
AWBM Nash–Sutcliffe efficiency (NSE) 0.751 0.821
AWBM Coefficient of determination (R2 ) 0.797 0.862
SIMHYD Nash–Sutcliffe efficiency (NSE) 0.814 0.731
SIMHYD Coefficient of determination (R2 ) 0.700 0.755

Fig. 19.9 Average monthly effective rainfall distribution of study area

 236

Table 19.2 Rainfall records

Year 2012 2011 2008 2003 2009 2010 2007 2002 2005 2006 2001 2004
Rank no. 1 2 3 4 5 6 7 8 9 10 11 12
Rain (mm/year) 1267.6 1156.7 1033.6 993.8 949.1 936.2 931.3 836.1 725.7 500.0 460.5 444.5
Fa (%) 7 14 21 29 36 43 50 57 64 71 79 86
A. Vishwakarma et al.
19 Runoff Simulation and Irrigation Water Requirement … 237

Fig. 19.10 Probability yearly rainfall

A logarithmic regression equation is obtained by preparing a semi-log graph

between yearly rainfall and probability, which gives yearly rainfall values at 20, 50,
and 80% probability of exceedance as 1168 mm, 866 mm, and 598 mm, respectively.
The graph is shown in Fig. 19.10.

19.4.6 Probability Analysis of Monthly Rainfall

It is clear from the probability analysis that the maximum deficiency in rainfall in
July is 95.1 mm, and the minimum deficit of 0.017 mm rainfall has been observed
in April month as shown in Fig. 19.11.

Fig. 19.11 Monthly rainfall under different probability level for Barman command
238 A. Vishwakarma et al.

Fig. 19.12 IWR of wheat crop in Barman command

19.4.7 Irrigation Water Requirement

The irrigation water requirements in the Barman area have been computed using
CROPWAT 8.0 software. The model calculates the irrigation water requirement of
rabi crop (Wheat) on a ten-daily basis, as shown in Fig. 19.12.
As per the average annual rainfall data and average climatic data, the irrigation
water requirement is 270.7 mm. Similarly, it can be calculated for rainfall at normal
(50%), dry (80%), and wet (20%) year of probability.

19.5 Conclusions

The models were tested and evaluated based on accuracy criteria Nash–Sutcliffe effi-
ciency and coefficient of determination (R2 ). The NSE and R2 values from AWBM
for calibration are 0.751 and 0.821 and for the validation is 0.797 and 0.862, respec-
tively, while NSE and R2 values from SIMHYD for calibration are 0.814 and 0.731,
and for validation are 0.700 and 0.755, respectively. AWBM accuracy values indicate
a better agreement between the observed and simulated runoff than the SIMHYD
model. Based on the accuracy values, it is concluded that AWBM is more suitable
for the Barman command in simulating the hydrological response of the command
to the rainfall and predicting daily runoff with the greater degree (13–14% higher
for R2 value) of accuracy.
19 Runoff Simulation and Irrigation Water Requirement … 239

It is concluded that the maximum contribution of rainfall is in July (around 30.8%)

and then in August (about 28%). Maximum deficiency in monthly rainfall from
average to dry year of probability is observed in July and minimum in April. Around
27% of the yearly rainfall deficiency is observed between average rainfall and rainfall
in a dry year. It is found that the rainfall for dry, wet, and at normal probability year
are 598 mm, 1168 mm, and 866 mm, respectively, and corresponding irrigation
water requirement estimated by CROPWAT are 274 mm, 264.7 mm, and 269.5 mm,
respectively. This study provides the idea related to irrigation water requirements
at variable rainfall pattern satisfactorily. Hence, it becomes very flexible and easy
to calculate rainfall at different probability levels by using statistical probability
analysis and irrigation water requirements by using CROPWAT 8.0 model to achieve
the proper irrigation planning.

Referencess

Allen RG, Pereira LS, Raes D, Smith M (1998) Crop evapotranspiration—guidelines for computing
crop water requirements-FAO Irrigation and drainage paper 56. FAO Rome 300(9):D05109
Boughton W (2004) The Australian water balance model. Environ Model Softw 19(10):943–956
Boughton W (2006) Calibrations of a daily rainfall-runoff model with poor quality data. Environ
Model Softw 21(8):1114–1128
Boughton W, Chiew F (2007) Estimating runoff in ungauged catchments from rainfall, PET and
the AWBM model. Environ Model Softw 22(4):476–487
Chiew FH, McMahon TA (1991) The Applicability of Morton’s and Penman’s evapotranspiration
estimates in rainfall-runoff modeling 1. JAWRA J Am Water Resources Association 27(4):611–
620
Chiew FHS, Siriwardena L (2005) Estimation of SIMHYD parameter values for application in
ungauged catchments 1
Chouhan D, Tiwari HL, Galkate RV (2016) Rainfall runoff simulation of Shipra river basin using
AWBM RRL toolkit. Int J Civil Eng Tech Res (IJETR) 5(3):73–76
Haque MM, Rahman A, Hagare D, Kibria G (2015) Parameter uncertainty of the AWBM model
when applied to an ungauged catchment. Hydrol Process 29(6):1493–1504
Jones RN, Chiew FH, Boughton WC, Zhang L (2006) Estimating the sensitivity of mean annual
runoff to climate change using selected hydrological models. Adv Water Resour 29(10):1419–
1429
Li L, Lambert MF, Maier HR, Partington D, Simmons CT (2015) Assessment of the internal
dynamics of the Australian water balance model under different calibration regimes. Environ
Model Softw 66:57–68
Manivannan S, Khola OPS, Dinesh D (2016) Probability analysis of weekly rainfall for crop
planning in Nilgiris hills of Tamil Nadu. J Agromet 18(1):163–164
Nash JE, Sutcliffe JV (1970) River flow forecasting through conceptual models part I—a discussion
of principles. J Hydrol 10(3):282–290. [Link]
Podger G (2004) Rainfall runoff library user manual. CRC for Catchment Hydrology, Melbourne
Sachan S, Chandola VK, Lohani AK (2016) Probability analysis of rainfall and crop water require-
ment using CROPWAT model for crop planning in a canal command of upper Bhima Basin of
Maharashtra. Int J Agric Environ Biotechnol 9(1):123
Smith M (1992) CROPWAT: a computer program for irrigation planning and management (No.
46). Food & Agriculture Organization
240 A. Vishwakarma et al.

Thimme Gowda P, Manjunaththa SB, Yogesh TC, Satyareddi SA (2013) Study on water requirement
of Maize (Zea mays L.) using CROPWAT model in northern transitional zone of Karnataka. J
Environ Sci Comput Sci Eng Technol 2(1):105–113
Yu B (2015) How would peak rainfall intensity affect runoff predictions using conceptual water
balance models? Proc Int Assoc Hydrol Sci 371:109–115
Chapter 20
Nonlinear Regression Analysis Between
Discharge and Head for Piano Key Weirs
with Increasing Developed Length (L/W)
Ratio and Constant Channel Width

Amiya Abhash and K. K. Pandey

Abstract Piano key weir (PKW) is becoming increasingly popular for dissipation
of excess discharge from spillways in old and newly constructed hydraulic structures.
The relationship between discharge and head thus warrants necessary attention. The
relationship between discharge and head for piano key weirs with increasing devel-
oped ratio for free flow over PKW for a constant channel width is investigated.
Further, the variation of increasing developed length ratio in constant channel width
with the power exponent of the head is analyzed. Nonlinear regression analysis was
performed using the Levenberg–Marquardt algorithm. The parameter estimate and
ANOVA table were prepared for flow over piano key weirs from data available in the
literature. The analysis shows a decreasing dependence on the exponent of the head
with increasing developed length ratio for the same channel width.

Keywords Piano key weir · Free flow · Discharge · Head · Developed length ratio

20.1 Introduction

With increasing floods and to avoid the danger of overtopping, many spillways
require replacement or an increase in their discharge capacity. This led to research
for increasing the crest length of the weir to improve their discharge capacity. Piano
key weir serves as an excellent alternative for increasing the overflow capacity of
existing and new dams (Leite Ribeiro et al. 2009, 2011).
The piano key weir is an improvement over rectangular Labyrinth type weir,
which has cantilevered apexes so that the allowable weir developed length can be
fit in the previous spillway channel width (Anderson and Tullis 2012). This enables
improved hydraulic capacity while reducing the structural footprint (Lempérière and
Ouamane 2003; Lempérière and Vigny 2011). Piano key weir’s geometric feature is

A. Abhash (B) · K. K. Pandey

Department of Civil Engineering, Indian Institute of Technology (BHU), Varanasi, India
e-mail: [Link].civ16@[Link]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 241
R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
242 A. Abhash and K. K. Pandey

Fig. 20.1 Fundamental parameters of a PKW Pralong et al. (2011)

easily adaptable than a Labyrinth weir for direct placement on the dam crest (Pralong
et al. 2011).
A naming convention was formulated for piano key weirs by Pralong et al. (2011).
The basic unit of PKW consists of two half inlets and one outlet, as shown in Fig. 20.1.
As shown in Fig. 20.1, the main geometric parameters of the PKW consist of the
total crest length L of weir, weir height P, channel width W, the transverse weir crest
length B, the total upstream and downstream cantilever (overhang), lengths Bo and
Bi , and the upstream and downstream widths W o and W i . Important fluid parameters
include density ρ, dynamic viscosity μ, and surface tension σ.

20.2 Literature Review

The piano key weirs are hydraulically more efficient than linear overflow weirs or
classical Labyrinth weirs for similar head and spillway width for flow in low heads
(Laugier et al. 2012; Tiwari and Sharma 2015).
There is no generally accepted standardized design procedure of PKW currently
available. This is due, in part, to a large number of PKW geometric parameters,
and a limited understanding of their influences on the head-discharge relationship
20 Nonlinear Regression Analysis Between Discharge and Head … 243

(Anderson and Tullis 2011). The discharge from PKW is affected by more than 20
geometrical parameters (Sharma and Tiwari 2013; Tiwari and Sharma 2017), which
are interlinked among each other.
Many researchers have given the discharge capacity of piano key weirs for unit
width of channel and for a given head. They expressed the discharge head relationship
as:

Q = K H n,

where Q is discharge, H is head over the crest or the head at some distance upstream
from piano key weir, n is the exponent, L is the developed length, and K is the system
constant.
Lempérière and Ouamane (2003)obtained a head discharge relationship for a
“Standard model” of PKW with all the characteristics defined by the only parameter
Pm (Fig. 20.1). They also proposed a formula relating the specific discharge q with
upstream total upstream head Hu. The value of exponent n was found to be unity.
Kabiri-Samani and Javaheri (2012) proposed dependence of the discharge over a
PKW in case of free flow on the following factors

Q = f (H, L , P, W, Wi , Wo , B, Bi , Bo , g, ρ, σ, μ) (20.1)

where f is a function. Performing dimensional analysis, Eq. 20.1 was written as

Cd = φ (H/P, L/W, B/P, Wi /Wo , Bi /B, Bo /B, W, R) . . . (20.2)

where φ is another functional symbol,

W = (ρ H/σ ) 0.5Q/[(H + P) W ] and R = ρ Q H/[(H + P) W μ] . . .

(20.3)

where R and W are the Reynolds and Weber numbers, respectively.

H > 30 mm was ensured so that the surface tension effect on discharge values is
small. Since the flow is turbulent around PKW, the viscosity effect is considerably
smaller compared with the gravity effect and hence discharge dependence on the
Weber and Reynolds number can be omitted.
The discharge capacity of a PKW is given by the same formula as that of a
sharp-crested weir as below:-

2  
Q= Cd 2g W H 1.5 . . . (20.4)
3
where C d is the coefficient of discharge, g is the acceleration due to gravity, and W
is the channel width.
The exponent “n” in the formula above is taken as 1.5 subject to the limitations
H > 30 mm, 0.1 ≤ H/P ≤ 0.6, 2.5 ≤ L/W ≤ 7, 1 ≤ B/P ≤ 2.5, 0.33 ≤ W i /W o ≤
244 A. Abhash and K. K. Pandey

1.22, 0 ≤ Bi /B ≤ 0.26, 0 ≤ Bo /B ≤ 0.26, and H d /H ≤ 0.6. The data points from

the discharge versus head relationship for different values of developed length ratio
L/W were extracted from literature with plot digitizer. Exponent n was observed
to be 1.5 in various literature (Anderson 2011; Laugier 2011; Leite Ribeiro et al.
2011; Kabiri-Samani and Javaheri 2012; Mehboudi et al. 2016). Head over exponent
was interestingly observed to be 0.75 in the experiments done by Tiwari and Sharma
(2017) with a developed length ratio (L/W ) equal to 4.76. Discharge head relationship
was observed as below:
Q
= C H 3/4 . . . (20.5)
LW
The head-to-weir height ratio (H/P) was kept below 0.1 for the experiments
while greater than 0.1 for experiments done by Kabiri-Samani and Javaheri (2012).
However, the discharge intensity was observed to be greater in this case for the same
head against the discharge intensity observed for 0.1 < H/P < 1.
The present study explores the exponent b over head variation for increasing values
of developed length ratio (L/W ) against the fixed value of 1.5 used in literatures taking
constant width of the channel for the case H/P greater than 0.1.

20.3 Methodology

20.3.1 Nonlinear Regression Analysis

Nonlinear regression is a curve-fitting tool intended to minimize the squared sum

(SS) of the difference between observed data and the possible fit. However, it is an
iterative or cyclical process where an initial guess of the parameters’ values will
have to be provided. In the successive iterations, parameter values are changed by a
small amount and SS recalculated. This process is repeated until the changes in the
parameter values result in minimizing the value of SS. Since in nonlinear regression,
the second and higher derivatives are not zero, an iterative process is required to
calculate the optimal parameter values. Nonlinear regression algorithm includes the
Gauss–Newton, the Marquardt–Levenberg methods, gradient descent methods, etc.
They all require an initial guess value of parameters and use them to estimate the
parameters in an iterative process.

[Link] The Marquardt–Levenberg Algorithm

The Levenberg–Marquardt algorithm (LMA), also known as the damped least

squares (DLS) method, is the most widely used method for solving nonlinear least
square minimization problems. The LM algorithm interpolates between the Gauss–
Newton algorithm (GNA) and the method of gradient descent. It is more robust
than the GNA as it gives a solution even if the initial guess is very far from the
20 Nonlinear Regression Analysis Between Discharge and Head … 245

final minimum. The LM algorithm provides a solution for nonlinear least squares
minimization. This implies that the function has the following special form:

1 2
j=m
f (x) = r j (x)
2 j=1

where x = (x 1 , x 2 , … x n ) is a vector, and each r j is a function from i to . The r j

is referred to as a residual, and it is assumed that m ≥ n. Now, f can be rewritten as
f (x) = 21 ||r (x)||2 . The derivatives can be written using the Jacobian matrix (J) of r
∂r
with respect to x defined as J (x) = ∂ xij Where 1 ≤ j ≤ m and 1 ≤ i ≤ n. Taking the
sum of square deviation with respect to small increase δ and setting the residual to
zero, we get:

(J T J )δ = J T (y − f (β))

where f and y are vectors with ith component f (x i, β) and yi .

For the general nonlinear case, we have:


m
∇ f (x) = r j (x)∇r j (x) = J (x)T r(x)
j=1


m
∇ 2 f (x) = r j (x)∇ 2 r j (x) + J (x)T J (x) and hence
j=1

((J T J ) + μI )δ = J T (y − f (β))

where μ, is the dampening factor, I the identity matrix and δ is the small increment
given to a parameter β.
The distinctive property of least squares problems is that from the given Jacobian
matrix (J), the Hessian ∇ 2 f (x) is easily found, and it is possible to approximate the
r j s through linear functions (∇ 2 r j (x) are small) or the residuals (r j (x) themselves
are small). The Hessian simply becomes∇ 2 f (x) = J (x)T J (x). The non-negative
dampening factor is updated after each iteration. A smaller value is used if the
reduction in residual value is rapid, while if the iteration leads to an insufficient
reduction in residual value, the dampening factor is increased.

[Link] ANOVA

ANOVA, also called analysis of variance, is a statistical tool to test differences

between two or more means. ANOVA yields levels of variability within a regression
model.
246 A. Abhash and K. K. Pandey

SST = SSM + SSE, where SS is a notation for the sum of squares and T, M, and
E is a notation for total, model, and error, respectively.
R-squared or coefficient of determination is a statistical tool to measure data
points’ closeness to the fitted regression line. It is the percentage of variation in the
response variable which can be explained through a linear model.
R2 = (Explained variation)/(Total variation) and is between 0 and 100%:
• 0% value suggests that the model cannot explain the variability in the response
data about its mean.
• 100% value suggests that the model can explain all the variability in the response
data about its mean.
• Higher the R2 , the better the regression model fits your data.

20.4 Results and Discussion

A nonlinear regression analysis for Q/LW, where Q is the discharge, L the developed
length, and W the width of the channel, was done from the experiments done by
Kabiri-Samani and Javaheri (2012) to best fit the curve taking Q/LW as a function
of a constant multiplied by head raised to an exponent variable b.

Q/L W = a H b

The results of nonlinear regression analysis for various developed length are
presented in Tables 20.1, 20.2, 20.3, 20.4, and 20.5. Table 20.6 summarizes the
constant “a” and exponent “b” values for increasing L/W ratio for a fixed channel
width of piano key weir (Fig. 20.2).

Table 20.1 a Parameter estimate for L/W = 2.5. b ANOVA table for L/W = 2.5
(a)
Parameter Estimate Std. error 95% confidence interval
Lower bound Upper bound
a 0.678 0.035 0.594 0.761
b 1.193 0.024 1.135 1.251
(b)
Source Sum of squares df Mean squares
Regression 466.881 2 233.440
Residual 0.156 7 0.022
Uncorrected total 467.037 9
Corrected total 82.096 8
Dependent variable: Q/LW a
R2 = 1 − (Residual sum of squares)/(corrected sum of squares) = 0.998
20 Nonlinear Regression Analysis Between Discharge and Head … 247

Table 20.2 a Parameter estimate for L/W = 4.0. b ANOVA table for L/W = 4.0
(a)
Parameter Estimate Std. error 95% confidence interval
Lower bound Upper bound
a 0.512 0.025 0.453 0.570
b 1.186 0.022 1.134 1.238
(b)
Source Sum of squares df Mean squares
Regression 299.359 2 149.679
Residual 0.082 7 0.012
Uncorrected total 299.440 9
Corrected total 52.378 8
Dependent variable: Q/LW a
R2 = 1 − (residual sum of squares)/(corrected sum of squares) = 0.998

Table 20.3 a Parameter estimate for L/W = 6.0. b ANOVA table for L/W = 6.0
(a)
Parameter Estimate Std. error 95% confidence interval
Lower bound Upper bound
A 0.590 0.018 0.545 0.635
B 1.015 0.015 0.978 1.052
(b)
Source Sum of squares df Mean squares
Regression 148.421 2 74.210
Residual 0.015 6 0.002
Uncorrected total 148.435 8
Corrected total 14.909 7
Dependent variable: Q/LW a
R2 = 1 − (residual sum of squares)/(corrected sum of squares) = 0.999

20.5 Conclusion

The data is taken from literature, and as such, the same limitations apply here also.
The results as such are subject to the present tests limitations h > 30 mm, 0.1 ≤ H/P
≤ 0.6, 2.5 ≤ L/W ≤ 7, 1 ≤ B/P ≤ 2.5, 0.33 ≤ W i /W o ≤ 1.22, 0 ≤ Bi /B ≤ 0.26, 0 ≤
Bo /B ≤ 0.26, and H d /H ≤ 0.6
The nonlinear regression analysis shows that as we bring more and more fold in
the weir geometry to increase the crest length for more significant discharge release
through PKW in a constant width of channel width, the exponent (b) of head (h)
decreases from the analytical value of 1.5 to almost 0.9.
248 A. Abhash and K. K. Pandey

Table 20.4 a Parameter estimate for L/W = 7.0. b ANOVA table for L/W = 7.0
(a)
Parameter Estimate Std. error 95% confidence interval
Lower bound Upper bound
a 0.584 0.021 0.535 0.633
b 0.975 0.019 0.932 1.018
(b)
Source Sum of squares df Mean squares
Regression 109.120 2 54.560
Residual 0.030 8 0.004
Uncorrected total 109.150 10
Corrected total 17.117 9
Dependent variable: Q/LW a
R2 = 1 − (Residual sum of squares)/(corrected sum of squares) = 0.998

Table 20.5 a Parameter estimate for L/W = 8.5. b ANOVA table for L/W = 8.5
(a)
Parameter Estimate Std. error 95% confidence interval
Lower bound Upper bound
a 0.616 0.026 0.556 0.676
B 0.879 0.022 0.830 0.929
(b)
Source Sum of squares df Mean squares
Regression 90.386 2 45.193
Residual 0.036 8 0.004
Uncorrected total 90.422 10
Corrected total 11.644 9
Dependent variable: Q/LW a
R2 = 1 − (residual sum of squares)/(corrected sum of squares) = 0.997

For H/P < 0.1, we have observed that exponent b drops to 0.75 though the discharge
intensity that can be passed through PKW increases for the same head as opposed to
0.1 < H/P < 0.6. This may be likely due to the effect of surface tension force which
was excluded from analysis by Kabiri-Samani and Javaheri (2012) by taking H >
30 mm.
Since in confined channel width, as we increase the developed length ratio, the
width of inlet and width of outlet will automatically decrease. To accommodate this,
we have to increase the overhangs, which are made structurally stable by increasing
the weir height. The entire analysis was carried out about a single parameter, i.e., the
developed length ratio (L/W ), since all the geometrical parameters are interrelated.
20 Nonlinear Regression Analysis Between Discharge and Head … 249

Table 20.6 Values of “a” and “b” for increasing L/W ratio
S. no. L/W a B R2 value
1 2.5 0.678 1.193 0.998
2 4.0 0.512 1.186 0.998
3 6.0 0.590 1.015 0.999
4 7.0 0.584 0.975 0.998
5 8.5 0.654 0.879 0.997

Exponent (b) y = -0.0839x + 1.3013

1.4
R² = 0.9381
1.193 1.186
1.2
1.015
0.975
1 0.879

0.8

0.6

0.4

0.2

0
2.5 4 6 7 8.5
Exponent (b) 1.193 1.186 1.015 0.975 0.879

Exponent (b) Linear (Exponent (b)) Linear (Exponent (b))

Fig. 20.2 Change in exponent “b” of head over increasing L/W value for flow over PKW in a
channel of constant width

References

Anderson RM (2011) Piano key weir head discharge relationships, Utah State University
Anderson R, Tullis B (2011). Influence of piano key weir geometry on discharge. In: Proceedings
of international conference labyrinth and piano key weirs Liège B
Anderson R, Tullis B (2012) Piano key weir hydraulics and labyrinth weir comparison. J Irrig Drain
Eng 139(3):246–253
Kabiri-Samani A, Javaheri A (2012) Discharge coefficients for free and submerged flow over piano
key weirs. J Hydraul Res 50(1):114–120
Laugier F, Blancher B (2011) Influence of structural thickness of sidewalls on PKW spillway
discharge capacity
Laugier F, Vermeulen J, Pralong J (2012) Achievement of new innovative labyrinth piano key
weir spillways (PKW). In: Proceedings of piano key weir for in-stream storage and dam safety
(pKwIsD-2012), New Delhi, pp 25–42
Leite Ribeiro M, Bieri M, Boillat J-L, Schleiss A, Delorme F, Laugier F (2009) Hydraulic capacity
improvement of existing spillways-design of a piano key weirs. In: Proceedings (on CD) of the
23rd congress of the international commission on large dams CIGB-ICOLD
250 A. Abhash and K. K. Pandey

Leite Ribeiro M, Bieri M, Boillat J-L, Schleiss A, Singhal G, Sharma N (2011) Discharge capacity
of piano key weirs. J Hydraul Eng 138(2):199–203
Lempérière F, Ouamane A (2003) The Piano Keys weir: a new cost-effective solution for spillways.
Int J Hydropower Dams 10(5):144–149
Lempérière F, Vigny J (2011) General comments on labyrinth and piano keys weirs—the future.
In: Labyrinth and piano key weirs–PKW 2011, pp 289–294
Mehboudi A, Attari J, Hosseini S (2016) Experimental study of discharge coefficient for trapezoidal
piano key weirs. Flow Meas Instrum 50:65–72
Pralong J, Vermeulen J, Blancher B, Laugier F, Erpicum S, Machiels O, Pirotton M, Boillat J-L,
Leite Ribeiro M, Schleiss A (2011) A naming convention for the piano key weirs geometrical
parameters. In: Labyrinth and piano key weirs, pp 271–278
Sharma N, Tiwari H (2013) Experimental study on vertical velocity and submergence depth near
piano key weir. In: Labyrinth and piano key weirs II-PKW, pp 93–100
Tiwari H, Sharma N (2015) Developments to improve hydraulic competence of spillways. Aquatic
Proc 4:841–846
Tiwari H, Sharma N (2017) Empirical and Mathematical modeling of head and discharge over piano
key weir. In: Development of water resources in India. Springer, pp 341–354
Chapter 21
Grey Water Characterization and Its
Management

Sarosh Alam Ghausi and Mohd Muzzammil

Abstract Water, the driving force of nature, is the life sustaining resource present
on the earth. Depleting groundwater tables and wells running dry has forced the
researchers and policymakers to come up with an efficient solution to optimize water
use and save water. With rainfall behaviour being uncertain and diminishing surface
water sources, there is a critical need to look out for alternate water source. In
this scenario, grey water use provides an efficient option. Grey water is the waste
water generated from showers, laundry washing, washbasins and sinks. It does not
come in contact with organic impurities which result in lower BOD as compared
to black water (water generated from toilet flushing). The basic idea is to separate
grey water from domestic sewage. Depending on the type of grey water and its level
of treatment, it can be reused for various purposes like irrigation, flushing, floor
washing, watering gardens, automobile washing, etc. Studies have suggested that
recycling grey water can save up to 70% of fresh water consumption. To control the
growing water imbalance, grey water use has huge potential to be used as sustainable
alternate resource. Present study characterizes different types of grey water in terms
of water quality parameters and suggests a low-cost treatment system for its further
use.

Keywords Grey water · Characterization · Low-cost treatment system · Water

quality parameters

S. A. Ghausi (B)
Civil Engineering Department, IIT BOMBAY, Mumbai 400076, India
M. Muzzammil
Civil Engineering Department, Z.H. College of Engineering & Tech, AMU, Aligarh, U.P. 202002,
India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 251
R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
252 S. A. Ghausi and M. Muzzammil

21.1 Introduction

As per United Nations department of economic and social affairs (UNDESA) report,
almost one-fifth of world that is around 1.2 billion people live in areas of physical
water scarcity with another 1.6 billion people approaching this situation (FAO 2007).
It is projected that by 2025 about 3.5 billion people will live in water stressed countries
(UNESCO 2012). More than one in every six people in the world do not have access
to potable water. Around 700 million people in 43 countries suffer from water scarcity
(Global Water institute 2013). A third of the world’s biggest groundwater systems
are already in distress (Richey et al. 2016). There are several other studies depicting
current water crisis and thus it is evidently clear to look out for new advances and
technology for water conservation and its management. There is an urgent need to
critically look into the alternate sources of water. Talking about alternate sources,
rainwater is the first option coming into the mind but the major problem with rain is
that it is uncertain and non-homogeneous and could not fulfil the constant demand
of water. Expanding industrialization has highly reduced the surface water source.
Problem with desalination is the mineral decomposition of potable water, thus in this
scenario grey water use is a highly efficient option to be used as an alternate resource.
While rainwater harvesting and Integrated Water Resource Management (IWRM) are
water conservation measures, grey water use is an important undeveloped technology.
Grey water is the waste water generated from showers, laundry washing, wash-
basins and sinks. Depending on the type of grey water and its level of treatment, it
can be reused for various purposes like irrigation, flushing, floor washing, watering
gardens, automobile washing, etc. Studies have suggested that recycling grey water
can save up to 70% of fresh water consumption (Alam et al. 2012; Ghausi et al.
2017). Since it does not come in contact with human waste, its water quality param-
eters like BOD are much lesser than usual sewage and it require much lower degree
of treatment. Thus, grey water use is important because it restricts water demand
and reduces stress on treatment system. Table 21.1 clearly depicts the difference
between water quality parameters in grey and sewage water (Rana et al. 2014). It can
be observed that biochemical oxygen demand (BOD) and chemical oxygen demand
(COD) are much lower for grey water as compared to sewage water.

Table 21.1 Comparison of

Parameter Unit Grey water Sewage water
water quality parameters of
grey and sewage water PH – 6.4–8.1 5.5–8.5
Suspended solids Mg/l 40–340 100–350
BOD5 Mg/l 40–50 100–300
COD Mg/l 100–300 175–600
Nitrogen (total) Mg/l 2–23 30–60
Phosphorus Mg/l 0.1–0.8 7–20
Turbidity NTU 15–270 –
21 Grey Water Characterization and Its Management 253

Grey water can be used for various purposes that include irrigating plants, watering
fields, toilet flushing, floor washing, car washing, recharging of dry ponds, etc. Sepa-
rate grey water lines from sewer lines in domestic households could be a very good
concept in designing of new cities. Grey water use is a huge success in USA with
around 20 states using it. However, there are conflicts to permit its use for irrigation.

21.2 Characterization of Grey Water

The sources of grey water generated from domestic households mainly include water
from laundry, washbasin and bathing. The water generated from kitchen is considered
as dark grey water due to comparatively high content of organic impurities and hence
not taken up in grey water characterization. Characterization is done on basis of
various water quality parameters that are Conductivity, alkalinity, total hardness,
BOD5 , COD, chloride, turbidity and PH.
Water is characterized as
1. Laundry grey water
2. Bathing grey water
3. Washbasin grey water
Sample collection for laundry grey water was done by collecting the grey water
from drain of a washing machine. For washbasin, grey water sample was collected
by storing the water from washbasin pipe into a large container. Soap used was
an antiseptic soap, whereas for bathing, grey water was collected by storing the
water before it reaches the drain, and the soap used was a commercial bathing soap.
Different water quality parameters of these samples are shown in Table 21.2.
From the following data broadly it can be inferred that grey water generated from
laundry is most polluted while that in washbasin is least polluted. It was observed that
laundry water was more turbid and dark grey in colour with moderate smell whereas
washbasin and bathing sample were greyish white and have slight smell. Grey water
in washbasin sample seems to become clearer with time due to sedimentation whereas
for laundry sample it requires skimming of foam and lather from the top surface. PH
of all the three samples was found to be almost with in neutral range; however, laundry
sample seems to have slightly alkaline nature as it also shows the highest alkalinity
value. BOD5 was maximum for laundry sample and minimum for washbasin sample
and so was the case of COD. The COD for laundry sample was much larger than
for washbasin and bathing sample; however, it should be noted that COD value is
directly related to dilution of sample. Concentrated grey water sample may lead to
much higher value of COD (>1000).
Following graph depicts the comparison of different parameters of grey water
samples.
254 S. A. Ghausi and M. Muzzammil

Table 21.2 Water quality

Laundry Washbasin Bathing
parameters of grey water
Alklanity 668 384 410
(mg/l)
Hardness 270 332 254
(mg/l)
Chloride 236 92.4 60
(mg/l)
BOD5 (mg/l) 49.12 30 35
COD (mg/l) 960 360 380
Conductivity 1400 1127 1275
(µS/cm)
Turbidity 165 93 108
(NTU)
PH 8.5 7.41 7.76
TDS (mg/l) 1475 1073 1167
Colour Dark grey Greyish white Greyish white
Odour Moderate Slight Slight
21 Grey Water Characterization and Its Management 255

21.3 Treatment of Grey Water

The main objective in treatment of grey water is the removal of soapy colloidal
impurities which are responsible for the turbidity in it. Sand bed filters are often used
to remove the particles for grey water. Study proposes a model of low-cost treatment
tank for treating grey water.

The treatment tank is subdivided into three compartments. Compartment 1 is

provided with layers of fine sand, coarse sand and gravels (10 mm) resting on a
porous stand. All these materials are cheaply and readily available. Grey water is
allowed to pass through this compartment and then it enters compartment 2 where
it rises against the gravity. Particles if any will settle down in this compartment. In
this compartment, some disinfectant may also be added if required or bleaching can
be done to prevent odour problems. Some flocculating agent may also be added if
colloidal impurities still persist. From compartment 2, water is allowed to fall in
compartment 3 increasing the turbulence and hence more air–water interfaces which
will subsequently result in increased aeration.
Experiments were performed to check the efficiency of this tank, and similar
conditions were simulated and tested for compartment 1.
256 S. A. Ghausi and M. Muzzammil

Grey water samples of laundry, washbasin and bathing were passed through this
filter and there water quality was checked before and after passing.

21.3.1 Laundry Sample

Comparison between water quality parameters of direct and filtered sample is made
clear in Table 21.3.
The desirable and permissible limits for drinking water standards are as per IS
10500: 1991. For drinking water standards, there is no standard value for BOD and
COD mentioned but BOD <30 mg/l for inland surface water and <100 mg/l for
irrigation water and COD <250 mg/l. It can be inferred from the table and the graph
that after filtration the water quality of grey water is highly improved. BOD value
of filtered sample is less than 10. There is huge decrease in total dissolved solids
which came into the permissible range of less than 500. Also, the turbidity is highly
reduced and the water becomes clear. All the water quality parameters are with in
permissible limits (Kumar and Avinash 2012).
21 Grey Water Characterization and Its Management 257

Underlying figure shows the direct laundry grey water sample (left) and filtered
sample (right).

Table 21.3 Water quality of laundry grey water and filtered water
Direct sample Filtered sample desirable limits Permissible limits
Total hardness (mg/l) 270 220 300 600
Alkalinity (mg/l) 668 440 200 600
Chloride (mg/l) 236 98 250 1000
BOD5 (mg/l) 49.12 9.7 30 100
COD (mg/l) 960 430 – –
TDS 1475 470 500 2000
Turbidity (NTU) 165 7 10 10
PH 8.5 7.66 6.5–8.5 6.5–8.5
Colour Dark grey Colourless
Odour Moderate No
258 S. A. Ghausi and M. Muzzammil

Table 21.4 Water quality of bathing grey water and filtered water
Direct sample Filtered sample Desirable limits Permissible limits
Total hardness 254 178 300 600
Alkalinity 410 326 200 600
Chloride 60 50 250 1000
BOD 35 8.9 30 100
COD 380 244 – –
TDS 1167 443 500 2000
Turbidity 108 5 10 10
PH 7.76 7.48 6.5–8.5 6.5–8.5
Colour Greyish white Colourless
Odour Slight No

21.3.2 Bathing Sample

Comparison between water quality parameters of direct and filtered sample is made
clear in Table 21.4.
From Table 21.4 also, similar behaviour can be inferred as in Table 21.3. There
is a considerable decrease in BOD5 , COD, turbidity and TDS of the sample.
21 Grey Water Characterization and Its Management 259

Following graph shows how the water quality of filtered sample is enhanced.
Underlying figure shows direct bathing grey water sample (left) and filtered sample
(right) (Table 21.5).

Table 21.5 Water quality of direct and filtered washbasin sample

Direct sample Filtered sample Desirable limits Permissible limits
Total hardness (mg/l) 332 268 300 600
Alkalinity (mg/l) 384 310 200 600
Chloride (mg/l) 92.4 72.8 250 1000
BOD5 (mg/l) 30 6.55 30 100
COD (mg/l) 360 220 – –
TDS (mg/l) 1073 411 500 2000
Turbidity (NTU) 93 3 10 10
PH 7.41 7.38 6.5–8.5 6.5–8.5
Colour Greyish white Colourless
Odour Slight No
260 S. A. Ghausi and M. Muzzammil

21.3.3 Washbasin Sample

Washbasin sample is the purest form of grey water and on further treatment it can
be observed that BOD5 has decreased to mere 6 mg/l whereas turbidity has fallen to
3 NTU making water almost perfectly clear. Underlying figure shows the washbasin
grey water sample (left) and filtered sample (right).
21 Grey Water Characterization and Its Management 261

21.4 Conclusion

Present study focuses on the use of grey water as an alternate resource to solve
the increasing problem of water scarcity. It characterizes grey water on the basis of
source of generation from domestic households as laundry grey water, washbasin
grey water and bathing grey water. On basis of water quality experiments done, it
was concluded that washbasin grey water is the purest form of grey water. A model
of low-cost treatment tank was also proposed. Experiment results showed that there
is vast improvement in water quality of grey water by passing it through the proposed
filter. With growing imbalance between water demand and water, grey water has a
huge potential to serve as an alternate water resource and it will not be wrong to say
that it is high time and grey water use is the need of the hour. It is only the matter of
time of how early we realize its importance.

References

Alam J, Muzzamil M (2012) Grey water use: a need of hour. India water week, Ministry of water
Resources, New Delhi
General standards for discharge of environmental pollutants (Schedule VI) (1986) Environment
protection rules
Ghausi SA, Muzzammil M, Alam J (2017) Grey water use as an alternate resource. In: Proceedings of
37th IAHR world congress held at Kualalumpur, Malaysia w.e.f 13th to 18th Aug, pp 4722–4729
Indian standard for drinking water—Specifications IS 10500 (1991)
Kumar M, Puri A (2012) A review of permissible limits of drinking water. Indian J Occupational
Environ Med
Rana K, Shah M, Upadhyay A (2014) Integrated approach towards grey water management. Int J
Eng Sci Res Technol 3(1):239–242
Richey AS, Thomas BF (2016) Quantifying renewable ground water stress with GRACE
Sarkar P, Sharma B, Malik U (2014) Energy generation from grey water in high rised buildings: the
case of India. Renew Energy 69:284–289
Chapter 22
Intelligent Operation of Hirakud
Reservoir Using Metaheuristic
Techniques (PSO and TLBO)

Pooja Patnaik and Prakash Ch. Swain

Abstract In water resources engineering, the real-life problems are mostly involved
with the nonlinear formulations. It is becoming a difficult task for large-scale
nonlinear optimization problems to obtain the optimal solutions. Since various
conflicting demands such as irrigation, power production, industrial water supply,
municipal water supply, etc., should be satisfied with water available in reservoir,
the optimal operating policy for multipurpose reservoir is a necessity. In this study,
the metaheuristic techniques like particle swarm optimization (PSO) and teaching–
learning-based optimization (TLBO) approaches are developed to overcome the limi-
tations of conventional techniques. The main objective of this paper is to develop
a policy for optimizing the total release of water for irrigation, power generation
and industries during non-monsoon period with the case study of the multipurpose
Hirakud reservoir of Odisha state in India. The TLBO and PSO models are imple-
mented in MATLAB, and the developed programs are executed on a 4 GB RAM,
64-bit operating system, Quad core processor, and it is observed that the computa-
tional time required for iteration of TLBO and PSO is 10 h 33 min and 11 h 37 min,
respectively, whereas in PARAMsavak supercomputer of 64 GB RAM, Ubuntu-
operating system, 28 logical core processor the computational time required for 300
iterations of TLBO and PSO are 1 h 22 min and 1 h 54 min, respectively. Therefore,
supercomputer is used for running the model as it saves computational time. On
comparing the policy developed by PSO and TLBO methods for reservoir operation
problems, it is observed that the release of water for irrigation, power generation,
and industrial purpose are more by using TLBO than that of the release by other
techniques, keeping strict surveillance on arriving at the dead storage level of the
reservoir at the end of the non-monsoon period.

Keywords Reservoir operation · Particle swarm optimization ·

Teaching–learning-based optimization · Reservoir management · Hirakud dam

P. Patnaik (B) · P. Ch. Swain

Civil Engineering Department, VSSUT, Burla 768018, India

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 263
R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
264 P. Patnaik and P. Ch. Swain

22.1 Introduction

In India, there are several single and multipurpose reservoir projects. The reservoir
water is often used to meet the requirements for various purposes such as municipal
and industrial water supply, water supply for irrigation, hydropower generation, etc.
Nowadays, it is a complex task for the optimal planning and appropriate manage-
ment of multipurpose reservoir. For optimizing the reservoir operation problems,
formerly many strategies were developed. Based on the idea of characteristics of
reservoir, statistics, constraints and objective characteristic, the selection of the opti-
mization techniques are done. Because of each problem contains its precise physical
and operating characteristics, there is no standard algorithm available which can
be relevant to all problems since four decades reservoir operation is one of the
important research area for water resources field. Many mathematical models have
been purposed during this period. Some of the research works are described below.
Eberhart and Kennedy (1995) proposed PSO for continuous nonlinear functions.
They discussed two paradigms that put in force the concept, one was globally orien-
tated (gbest), and other was locally oriented (lbest), accompanied by way of effects
received from applications and tests upon which the paradigms have been shown
to perform correctly. Reddy and Kumar (2009) used elitist-mutated multi-objective
particle swarm optimization (EM-MOPSO) method for integrated water resources
management. Their objective was to minimize flood risk, maximize hydropower
production, and minimize irrigation deficits in a year, subject to various physical and
technical constraints. Swain and Nanduri (2009) used neuro-fuzzy inference system
for finding the reservoir management policy of a multipurpose reservoir system,
i.e., Hirakud reservoir. This model was developed for forecasting inflows into the
reservoir and operation of the reservoir during the monsoon period. Recently, a new
metaheuristic technique that is teaching–learning-based optimization (TLBO) has
been recently developed and is becoming more popular for solving the optimiza-
tion problems (Rao and Teja 2015) as it uses only common controlling parameters
such as number of generations and population size and no specified control parame-
ters. It causes difficulties and complexity when dealing with the reservoir operation.
So a proper and effective operation policy should be done which can deal with all
associated problem and produce a proper standard policy for the reservoir.

22.2 Study Area and Data Details

The Hirakud reservoir is situated in Orissa state, India. The dam is built across river
Mahanadi at about15 km upstream of Sambalpur town in the state of Orissa and dam
is in operation since 1957. The geographical location of Hirakud dam in global grid is
at latitude 21°32 N and longitude 83°52 E. The location map of Hirakud reservoir
is shown in Fig. 22.1. Hirakud reservoir is one of the biggest artificial lake in Asia
with reservoir spread of 743 km2 at full reservoir level. The water surface area at
22 Intelligent Operation of Hirakud Reservoir Using Metaheuristic … 265

Fig. 22.1 Satellite image of Hirakud reservoir. (Source Google Earth)

dead storage level is around 274 km2 . The reservoir has gross, live, and dead storage
capacities of 7189, 5375, and 1814 million cubic meter (Mm3 ), respectively. The
reservoir elevations at gross storage and dead storage are 192.024 m and 179.830 m,
respectively.
The Hirakud reservoir system serves multiple purposes such as flood control, irri-
gation, and hydropower. In monsoon season, the first priority of the reservoir is to
minimize the adverse effects of floods. From the analysis of past historical data and
practical experience, the Hirakud dam authority follows some guidelines to maintain
the reservoir levels at different time periods to protect the downstream region from
flooding. Thus, when formulating mathematical model for reservoir operation, these
flood control restrictions need to be satisfied at various time periods. Then, the project
provides water for irrigation in the districts of Sambalpur, Bargarh, Bolangir, and
Subarnapur for an area of 155,635 ha in Kharif season (June–October) and 108,385 ha
in Rabi season (November–April). The water released through the power house irri-
gates the Mahanadi Delta area of 436,000 ha. The project provides hydropower gener-
ation through two power houses, namely Burla and Chiplima. Data of 27 years are
taken on ten daily basis from 1990 to 2016. Flow data are collected from odisha water
planning organization, Secha Sadan, Bhubaneswar; office of Chief Engineer-cum-
Basin Manager, Upper Mahanadi, Burla; office of the executive engineer, Hirakud
dam circle, Burla; master control room, Burla; central water commission, Burla
and Bhubaneswar; Indian Meteorological Department, Hirakud and Bhubaneswar;
Survey of India, Bhubaneswar. Data are processed before feeding to the models.
266 P. Patnaik and P. Ch. Swain

22.3 Methodology

Optimization techniques are used in every sector such as water demand management,
construction program management, industrial management, power system control,
etc. Main objective of using optimization is determining variables and determining
the optimum result. Word ‘optimization’ refers to perfection. It defines as choosing
best from set available. It is a process of making perfection and more effective
results. It deals with the best solution in a simple way. With progression of time,
new methods are introduced in optimization from classical methods to evolutionary
algorithm. Particle swarm optimization and teaching–learning-based optimization
are used. They are broadly discussed below.

22.3.1 Particle Swarm Optimization Algorithm

The practical swarm optimization (PSO) is one of the metaheuristic methods that are
widely used in water resource management. It was initially proposed by Eberhart
and Kennedy (1995). It has been inspired by the social behavior of animals like bird
flocking, fish schooling, and insect swarming. The PSO used the idea that social
sharing of information among members may have some evolutionary advantage.
PSO has been applied to many real-world problems. A standard PSO algorithm was
initialized with a population (swarm) of random potential solutions (particles). Each
particle iteratively moves across the search space and was attracted to the position
of the best fitness historically achieved by the particle itself (local best) and by the
best among the neighbors of the particle (global best). In PSO, instead of using more
traditional genetic operators, each particle adjusts its flying according to its own
flying experience and its companions’ flying experience.
Assume the place of the goal is in the M dimension. Individual particle is denoted
 T
by, i.e., X j = x j1 , x j2 , . . . x j M . The velocity of each particle is represented by
 T
V j = V j1 , V j2 , . . . V j M . The position of each particle with personal best posi-
tion is represented as P Bid n
= ( pbi1 , pbi2 , . . . pbi D )T . The global best value of
 T
individuals is denoted as P Bgd n
= pbg1 , pbg2 , . . . pbg D .
The following equations give the updating rules for velocity and position of the
particle:
   
n
P Bid −X (P Bgd
n
− X)
Vidn+1 =χ ωVidn + c1r1n + c2 r2
n
(22.1)
t t

n+1
X id = X id
n
+ (t)Vidn+1 (22.2)

where χ is the constriction coefficient, x is the inertial weight, c1 and c2 are the
cognitive and social parameters of PSO, d = 1, 2, …, D, the index for decision
22 Intelligent Operation of Hirakud Reservoir Using Metaheuristic … 267

variables, i = 1,2,…, NS, the index for swarm population, NS is the size of the
swarm, g is the index of best particle among all particles in the population, n is the
generation number, t is time step (considered as unity) and r 1 and r 2 are uniformly
generated random numbers between 0 and 1. By the effect of these two equations,
individual particle changes its velocity and position with respect to its previous best,
the best position of its neighbor. For getting the faster convergence, proper tuning c1
and c2 is needed. The inertia weight is used to control the effect of previous velocity in
new one. The larger value of inertia weight refers to the global exploration, whereas
the smaller value of inertia weight refers to the local exploration. So choosing the
value of the inertia weight is very important that will help to reduce the number of
iterations.

22.3.2 Teaching–Learning-Based Optimization

The TLBO algorithm is a teaching–learning process inspired algorithm proposed

by Rao et al. (2011; 2012a, b) and Rao and Savsani (2012) based on the effect of
influence of a teacher on the output of learners in a class. The algorithm describes
two basic modes of the learning: (i) through teacher (known as teacher phase) and
(ii) through interaction with the other learners (known as learner phase). In this
optimization algorithm, a group of learners is considered as population, and different
subjects offered to the learners are considered as different design variables of the
optimization problem and a learner’s result is analogous to the ‘fitness’ value of the
optimization problem. The best solution in the entire population is considered as the
teacher. The design variables are actually the parameters involved in the objective
function of the given optimization problem, and the best solution is the best value
of the objective function. The working of TLBO is divided into two parts, ‘Teacher
phase’ and ‘Learner phase.’ Working of both the phases is explained below.

[Link] Teacher Phase

It is the first part of the algorithm where learners learn through the teacher. During
this phase, a teacher tries to increase the mean result of the class in the subject
taught by him or her depending on his or her capability. At any iteration i, assume
that there are ‘m’ number of subjects (i.e., design variables), ‘n’ number of learners
(i.e., population size, k = 1, 2, . . . n) and M j,i be the mean result of the learners in a
particular subject ‘j’ ( j = 1, 2, . . . m) The best overall result X total−kbest,i considering
all the subjects together obtained in the entire population of learners can be considered
as the result of best learner kbest. However, as the teacher is usually considered as
a highly learned person who trains learners so that they can have better results, the
best learner identified is considered by the algorithm as the teacher. The difference
between the existing mean result of each subject and the corresponding result of the
teacher for each subject is given by,
268 P. Patnaik and P. Ch. Swain
 
DifferenceMean j,k,i = ri X j,kbest,i − TF M j,i (22.3)

where X j,kbest,i is the result of the best learner in subject j; TF is the teaching factor
which decides the value of mean to be changed, and ri is the random number in range
[0,1]. Value of TF can be either 1 or 2. The value of TF is decided randomly with
equal probability as,

TF = round[1 + rand(0, 1){2 − 1}] (22.4)

TF is not a parameter of the TLBO algorithm. The value of TF is not given as

an input to the algorithm, and its value is randomly decided by the algorithm using
Eq. (22.4). After conducting a number of experiments on many benchmark functions,
it is concluded that the algorithm performs better if the value of TF is between 1 and
2. However, the algorithm is found to perform much better if the value of TF is either
1 or 2, and hence to simplify the algorithm, the teaching factor is suggested to take
either 1 or 2 depending on the rounding up criteria given by Eq. (22.4). Based on the
Difference_Mean j,k,i , the existing solution is updated in the teacher phase according
to the following expression.

X j,k,i = X j,k,i + Difference_Mean j,k,i (22.5)

 
where X j,k,i is the updated value of X j,k,i . X j,k,i is accepted if it gives better function
value. All the accepted function values at the end of the teacher phase are maintained,
and these values become the input to the learner phase. The learner phase depends
upon the teacher phase.

[Link] Learner Phase

It is the second part of the algorithm where learners increase their knowledge by
interacting among themselves. A learner interacts randomly with other learners for
enhancing his or her knowledge. A learner learns new things if the other learner has
more knowledge than him or her. Considering a population size of ‘n,’ the learning
phenomenon of this phase is explained below.
 
Randomly select two learners P and Q such that X total−P,i = X total−Q,i (where,
   
X total−P,i and X total−Q,i are the updated function values of X total−P,i and X total−Q,i
of P and Q, respectively, at the end of teacher phase)

 
  
  
X j,P,i = X j,P,i + ri X j,P,i − X j,Q,i , if X total−P,i < X total−Q,i (22.6)

 
  
  
X j,P,i = X j,P,i + ri X j,Q,i − X j,P,i , if X total−Q,i < X total−P,i (22.7)
22 Intelligent Operation of Hirakud Reservoir Using Metaheuristic … 269


X j,P,i is accepted if it gives a better function value.
Equations (22.6) and (22.7) are for minimization problems. In the case of
maximization problems, Eqs. (22.8) and (22.9) are used.

 
  
  
X j,P,i = X j,P,i + ri X j,P,i − X j,Q,i , if X total−Q,i < X total−P,i (22.8)

 
  
  
X j,P,i = X j,P,i + ri X j,Q,i − X j,P,i , if X total−P,i < X total−Q,i (22.9)

Teaching–learning-based optimization (TLBO) is a population-based algorithm

which simulates the teaching–learning process of the class room. This algorithm
requires only the common control parameters such as the population size and
the number of generations and does not require any algorithm-specific control
parameters.

22.4 Reservoir Operation Model

Main aim of any water resources project is to get minimum irrigation deficit with
maximum hydropower production. These two objectives are conflicting to each
other. For minimizing irrigation deficit, more release of water is required, and for
hydropower generation, more water should be stored for maintaining high water level
for production of more water levels. But nowadays, industrialization is one of the
most important factor to improve the economic growth of a country. So the supply of
water to the industries cannot be over ruled. Hence, the demand and release of water
for industrial purpose are another parameter for the specific multipurpose reservoir
operating model.
• For irrigation release operation

The square deficit should be minimized


T
Minimize Z = (IRD − IR)2 (22.10)
t−1

where Z is squared deviation of irrigation demand and releases; IRD is irrigation

demands in period t(106 ) m3 ; IR is water release in period t(106 ) m3 at t = 1,2,3,….
• For hydropower production operation
Maximizing the production of energy


T
Maximize E = p(P)t (H )t (22.11)
t−1
270 P. Patnaik and P. Ch. Swain

where E is Energy produced (KW); p is power production coefficient; Pt is release

made to river bed turbine in period (106 ) m3 ; H t is net heads available (m).
• For industrial release operation

The square deficit should be minimized


T
Minimize F = (IND − INR)2 (22.12)
t−1

where F is squared deviation of irrigation demand and releases; IND is industrial

demands in period t(106 ) m3 ; INR is water release in period t(106 ) m3 at t = 1,2,3,….
• For combining irrigation release and hydropower generation operation

According to the basis of priority of the objective of the multipurpose reservoir, the
final fitness function is
T

T
IRD − IR 2 E max − P(P)t (H )t
Minimize Z = c f 1 + c f2
t−1
IRD t−1
E max

T
+ c f3 (IND − INR)2 (22.13)
t−1

where cf 1, cf 2 , and cf 3 is constants are used according to the priority; IRD is irri-
gation demands in period t(106 ) m3 ; IR is water release in period t(106 ) m3 ; E max is
maximum energy produced (KW); p is power production coefficient; Pt is release
made to river bed turbine in period (106 ) m3 ; H t is net heads available (m); IND is
industrial demands in period t(106 ) m3 ; INR is water release in period t(106 ) m3 .

22.4.1 Constraints for Irrigation Release, Hydropower

Generation and Industrial Release

(a) Continuity constraints of the reservoir

S(t + 1) = St(t) + In(t) − P(t) − IRR(t) − Evp(t) − Ovf(t) − IND(t)

(22.14)

For all t = 1,2,3,…,T.

Where, St is reservoir storage at time t(106 ) m3 ; S(t + 1) is reservoir storage at next
time period t(106 ) m3 ; In(t) is inflow to the reservoir at time period t(106 ) m3 ; P(t)
22 Intelligent Operation of Hirakud Reservoir Using Metaheuristic … 271

is release made to river bed turbine in period t(106 ) m3 ; IRR(t) is water release in
period t(106 ) m3 ; Evp(t) is evaporation at time period t(106 ) m3 ; Ovf(t) is overflow
at time period t(106 ) m3 .
(b) Constraints of release from the reservoir

Dmin (t) ≤ IRt ≤ Dmax (t) (22.15)

where Dmax (t) is maximum irrigation demand (106 ) m3 ; Dmin (t) is minimum irrigation
demand (106 ) m3 .
(c) Maximum power production limits

p P(t)Ht ≤ E max (22.16)

where p is power production coefficient; P(t) is release made to river bed turbine in
period t(106 ) m3 ; H t is net heads available (m); E max is maximum energy produced
(KW).
(d) Constraints of industrial release from the reservoir

INDmax ≤ INR ≤ INDmin (22.17)

where INDmax (t) is maximum industrial demand (106 ) m3 ; INDmin (t) is minimum
industrial demand (106 ) m3 .
(e) Canal capacity limits

IRRt ≤ Camax (22.18)

where Camax = Maximum canal capacity (106 ) m3 .

(f) Discharge from the reservoir through penstock

Q = A∗V


= d 2 2g H (22.19)
4

where Q is discharged from the reservoir for hydropower production (106 ) m3 ; d

is diameter of the penstocks (m); g is acceleration due to gravity; H is head for
hydropower generation, and it is function of time (m).
272 P. Patnaik and P. Ch. Swain

22.5 Results and Discussion

A. PSO
At first, the numbers of iterations are arbitrarily chosen to start. By changing the
iterations time to time, we are reaching to a point where optimality is maximum.
For different iterations, water release for irrigation, power, and industry is found out
(units-Million cubic meter), and corresponding efficiency (in %) is determined. For
different constant values, the optimum values are checked by using MATLAB code
(Fig. 22.2).
This graph shows the fitness value with respect to number of iterations. It is seen
that after 167th iteration, the fitness value become constant. Particle swarm optimiza-
tion has an advantage over DE since it is converged very fast as its convergence rate
is high and time taken for convergence is less. After 167th iteration, its fitness value
is converged to 2.280272.
B. TLBO
This figure represents of the fitness value for various iterations. With iteration, the
fitness value is updated, and after some iteration, it attains to a constant value. For
each iteration, the fitness value is recorded on the basis of that coding is progressed.
From this figure, it is clearly seen that by using teaching–learning-based optimiza-
tion, after 72nd iteration, it converges to one value. On the basis of fitness value,
the effectiveness of algorithm is decided. Its fitness value varies from 256.2233 to
0.065159. After 72nd iteration, its fitness value is converged to 0.065159 (Fig. 22.3;
Table 22.1).
For irrigation release for non-monsoon season is 134.28 Million m3 and
137.26 Million m3 by using PSO and TLBO, respectively. For hydropower generation
for non-monsoon season are 189.515 and 207.315 MW by using PSO and TLBO.

Fig. 22.2 Fitness value over iteration with particle swarm optimization
22 Intelligent Operation of Hirakud Reservoir Using Metaheuristic … 273

Fig. 22.3 Fitness value over iteration with teaching–learning-based optimization

Table 22.1 Comparison table showing the global optimum release by TLBO and PSO
Methods TLBO(MCum) PSO(MCum)
Water release for irrigation 137.26 134.28
Water release for power 697.00 692.6753
Water release for industries 82.156 78.481
Global optima 916.41 867.31

For industrial release for non-monsoon season are 78.481 and 82.156 Million m3 by
using PSO and TLBO (Tables 22.2, 22.3, and 22.4; Figs. 22.4, 22.5, 22.6, and 22.7).
This graph clearly represents that by using the teaching–learning-based optimiza-
tion, we can conserve more end storage then compare to the actual end storage for
a particular ten daily basis after the release of water from the reservoir for irri-
gation, power generation, and industrial purpose. Thus, the policy developed by
teaching–learning-based optimization is far better than the policy of other techniques.

Table 22.2 Irrigation release for the historical years for the non-monsoon season from 1990 to
2016 for Hirakud reservoir system from TLBO, PSO and DE methods for 10-daily time periods
(in Million m3 )
Statistical parameters Models for reservoir operation
TLBO PSO
Maximum value (Million m3 ) 137.26 134.11
Minimum value (Million m3 ) 25.78 20.69
Average value (Million m3 ) 79.34 75.95
274 P. Patnaik and P. Ch. Swain

Table 22.3 Hydropower production for the historical years for the non-monsoon season from 1990
to 2016 for Hirakud reservoir system from TLBO, PSO, and DE methods for 10-daily time periods
(in MW)
Statistical parameters Models for reservoir operation
TLBO PSO
Maximum value (MW) 207.315 189.515
Minimum value (MW) 1.369 1.178
Average value (MW) 32.387 29.441

Table 22.4 Industrial release for the historical years for the non-monsoon season from 1990 to
2016 for Hirakud reservoir system from TLBO, PSO and DE methods for 10-daily time periods
(in Million m3 )
Statistical parameters Models for reservoir operation
TLBO PSO
Maximum value (Million m3 ) 82.156 78.15
Minimum value (Million m3 ) 11.25 10.25
Average value (Million m3 ) 80.09 79.73

Fig. 22.4 Comparison of irrigation release for 27 years by using optimization techniques

22.6 Conclusions

This study developed optimal operation policies by using soft computing methods
for Hirakud reservoir system. On applying evolutionary algorithms such as TLBO,
PSO, and DE for solving reservoir operation problems, it is found that the TLBO
method is resulting in better performance for reservoir optimization by providing
good quality optimal solutions in fewer numbers of iterations as compared to other
two methods. The following conclusions are drawn from various experiments done
for 27 years that for irrigation release for non-monsoon season are 134.28 Million m3
and 137.26 Million m3 by using PSO and TLBO, respectively, for hydropower gener-
ation for non-monsoon season are 189.515 MW and 207.315 MW by PSO and TLBO,
22 Intelligent Operation of Hirakud Reservoir Using Metaheuristic … 275

Fig. 22.5 Comparison of hydropower generation for 27 years by using optimization techniques

Fig. 22.6 Comparison of industrial release for 27 years by using optimization techniques

Fig. 22.7 Comparison of actual end storage to end storage water level by teaching–learning-based
optimization technique for non-monsoon period of 2016

respectively, and for industrial release for non-monsoon season are 78.481 Million m3
and 82.156 Million m3 by using PSO and TLBO, respectively. For the single year
2016, the actual release for the irrigation is 130.683 Million m3 , but for PSO and
TLBO is 133.9175 Million m3 and 134.4254 Million m3 , respectively, whereas
the actual hydropower generation are 176.3101 MW, but for PSO and TLBO is
276 P. Patnaik and P. Ch. Swain

181.3101 MW and 185.949 MW, respectively. And the actual releases of water for
industries are 72.683 Million m3 , but for PSO and TLBO are 75.9175 Million m3 and
80.4254 Million m3 , respectively. Therefore, on comparing the actual release and the
policy developed by PSO and TLBO methods for reservoir operation problems, it
is obtained that the release of water for irrigation, power generation, and industrial
purpose is more by using TLBO than that of the release by other techniques.

Acknowledgements I would like to express my heartfelt gratitude to Prof. Prakash Chandra Swain,
my supervisor for rendering their valuable guidance and suggestions, encouragement, comments
and constructive criticisms given then and there which has made this research work a successful
one. Without his kind help, I would not even have a chance to work on this area. Furthermore I am
greatly indebted to EE, Hydrology Division, Hirakud Dam Circle for sharing the real-life data.

References

Eberhart RC, Kennedy J (1995) A new optimizer using particle swarm theory. In: Proceedings of
6th symposium on micro machine and human science. IEEE Service Center, Piscataway, NJ, pp
39–43
Kumar DN, Raju KS, Ashok B (2006) Optimal reservoir operation for irrigation of multiple crops
using genetic algorithms. J Irrig Drainage Eng 132:123–129
Rao RV, Savsani VJ, Vakharia DP (2011) Teaching–learning-based optimization: a novel method
for constrained mechanical design optimization problems. Comput Aided Des 43:303–315
Rao KP, Savsani VJ (2012) Mechanical design optimization using advanced optimization tech-
niques. Springer, London
Rao RV, Savsani VJ, Balic J (2012a) Teaching–learning-based optimization algorithm for uncon-
strained and constrained real parameter optimization problems. Eng Optim 44(12):1447–1462
Rao RV, Savsani VJ, Vakharia DP (2012b) Teaching–learning-based optimization: a novel
optimization method for continuous non-linear large scale problems. Inf Sci 183(1):1–15
Rao KP, Teja RB (2015) Teaching learning based optimization for short term hydro-thermal
scheduling. Conference on power, control, communication and computational technologies for
sustainable growth (PCCCTSG) 978-1-4673-6890-2/15
Reddy MJ, Kumar DN (2009) Performance evaluation of elitist-mutated multi-objective particle
swarm optimization for integrated water resources management. J Hydroinformatics 79–88
Swain PC, Nanduri UV (2009) Neuro-fuzzy inference system for operation of a multi-purpose
reservoir. Hydroinformatics in hydrology, hydrogeology and water resources. International
Association of Hydrological Sciences Publication 331
The project report on Hirakud dam (2016)—by master control station, Burla
Zhang J, Wu Z, Cheng C, Zhang S (2011) Improved particle swarm optimization algorithm for
multi-reservoir system operation. Water Sci Eng 4(1):61–73
Chapter 23
Agricultural Water Management
and Groundwater Recharging Using
Vadose Zone Modelling

Anooja Thomas, Vivekanand Singh, and Brijesh Kumar Yadav

Abstract Temporal and spatial soil moisture status in variable-saturated zone plays
a vital role in agricultural water management, groundwater recharge and solute trans-
port of that region, and it is fundamentally regulated by plant growth and its water
extraction rate along with soil properties and weather conditions. The soil moisture
distribution pattern at soil element scale is regulated by Richard’s equation integrated
with a sink term representing water uptake rate by plants. For obtaining soil moisture
dynamic across the root zone of a site-specific area, Richards equation-based model
coupled with sink term is numerically solved using finite-difference method. The
developed model is tested for simplified case(s) and applied over a wheat-cropped
site in Patna region of Bihar, India, with pedotransfer functions obtained from soil
sample analysis. Stimulation results for a period of 120 days show that out of 24 cm
irrigation applied and 10.61 cm rainfall occurred, 47.2% has been lost as evapotran-
spiration, 28.9% of the water has been stored in the soil column as soil moisture, and
23.9% of the water has recharged to groundwater table. Based on these results, an
irrigation schedule is recommended for an optimal utilization of both rainwater and
irrigation water for a better management of available water resources.

Keywords Groundwater recharge · Root water uptake · Soil moisture modelling ·

Numerical modelling

A. Thomas (B) · B. K. Yadav

Department of Hydrology, Indian Institute of Technology Roorkee, Roorkee, Uttarakhand
247667, India
B. K. Yadav
e-mail: brijkfhy@[Link]
V. Singh
Department of Civil Engineering, National Institute of Technology Patna, Patna, Bihar 800005,
India
e-mail: vsingh@[Link]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 277
R. Jha et al. (eds.), Water Resources Management and Reservoir Operation,
Water Science and Technology Library 107,
[Link]
278 A. Thomas et al.

23.1 Introduction

The compilation of World Bank experiences and strategies stated that water manage-
ment of rain-fed agriculture is one of the major challenges of irrigated agriculture.
The cultivation practices in India cannot vary dependent on unpredictable monsoon.
Excessive irrigation results in water logging and soil salinity in addition to pesticides
and nutrient leaching groundwater contamination (Yadav and Junaid 2014) and water
stress affects yield during its sensitive growing stage. Moreover, the study area makes
use of arsenic-contaminated groundwater for the irrigation purpose. This may lead to
major health risk in near future (Singh and Ghosh 2011). Soil moisture status serves
as a tool for the management of the irrigation and thereby for a cost effective, better
crop yield and for considering other agriculture-related environmental issues.
The soil moisture flow through the partially saturated soil is a complex system
governed by highly nonlinear processes described by means of physical–mathemat-
ical models. Modelling progress is achieved by one dimensional Richard’s equation
for describing infiltration through vadose zone which is coupled with sink term for the
account of evapotranspiration (ET) and plant growth stage (Yadav et al. 2009). Root
water uptake is a spatial and temporal process influenced by plant type, soil char-
acteristics and the prevailing climatic conditions, plant growth stage, rooting depth
and density distribution (Yadav and Mathur 2008) and availability of soil water.
Numerous methods have proposed with various degree of accuracy for the estima-
tion of root water uptake. There are major three approaches for computing root
water uptake, microscopic, mesoscopic and macroscopic approaches. Microscopic
approach simulates individual root considering individual root as a uniform radius,
long cylinder with water-absorbing properties (Feddes and Raats 2004; Šimůnek and
Hopmans 2009). The mesoscopic approach simulates the root water uptake through
diffusion and is the function of matrix flux potential difference of the soil and the
root (van Lier et al. 2006).
Numerous water extraction models are proposed over macroscopic empirical
model (Feddes et al. 1978; Ojha and Rai 1996; Yadav et al. 2009). Water uptake
patterns proposed can categorize as constant (Feddes et al. 1978), linear, (Prasad
1988), nonlinear (Shankar et al. 2012) or exponential (Li et al. 1999; Kang et al.
2001). Molz and Remson (1970) proposed a hypothetical linear pattern from four
layer of equal depth with water uptake of 40, 30, 20 and 10% top from each layer.
Groundwater recharge through the unsaturated zone of a cropped site has simu-
lated considering that the excess water after the soil moisture holding capacity of soil,
which drains below the root zone and reaches the groundwater table (Rushton 1988;
Wohling et al. 2012). Numerous studies are focused on the potential grounder esti-
mation through unsaturated zone the focus of study (Jimenez-Martinez et al. 2009;
Min et al. 2015) as it required less parameterization effort calculation efficiency,
cost effective and its predicting capability compared to saturated zone and surface
water-based models.
23 Agricultural Water Management and Groundwater Recharging … 279

This study focused on developing a simplified and least parameterized soil mois-
ture model for the recharge estimation along with agriculture water management and
its field application over a wheat-cropped site at Patna, Bihar.

23.2 Methodology

Special and temporal movement of soil moisture through the unsaturated zone is
governed by the Richard’s equation coupled with the sink term for the account of
the out flux of soil moisture from the system through plants. The mixed form of
Richard’s 1D equation is written as,
  
∂θ ∂ ∂ψ
= K (ψ) − 1 + S(z, t) (23.1)
∂t ∂z ∂z

where θ = volumetric soil water content (volume of water per unit volume of soil);
ψ = soil matrix potential; S(z, t) = sink term representing water extraction by plant
roots from a unit volume of soil per unit time; z = vertical coordinate (taken positive
upwards); K = unsaturated hydraulic conductivity of the soil; and t = time. For
solving mixed form of Richard’s equation, it requires functional relation between soil
moisture content θ and hydraulic conductivity K in term of soil moisture potential,
ψ. Soil moisture characteristics relationship proposed by Haverkamp et al. (1977) is
used.

23.2.1 Initial and Boundary Conditions

For the initialization of the simulation, the initial condition of the soil moisture
condition or water potential corresponding to field condition at the starting time
of simulation is specified at each node. The actual field condition of soil moisture
status at individual is difficult to measure. For this reason, an arbitrary value of soil
moisture content is selected. Corresponding pressure head and unsaturated hydraulic
conductivity (computed using the characteristic relationships) are also specified at
all the nodes as the initial conditions,

i.e. θ (z, t) = θini for z ≥ 0, t = 0 (23.2)

The depth of irrigation and rainfall and evaporation from the surface is defined
as the upper boundary condition. For the subsurface flow resulting from irrigation
and rainfall infiltration, the upper BC will change with time. During the initial stage
of irrigation/rainfall, a specified depth of ponding ψ (x) = h(x) (= depth of water)
is considered. Let the imposed pressure head at the surface of ground for the flow
280 A. Thomas et al.

domain be ψ b . The above pressure head ψ b is used along with the pressure head of
first node ψ 1 and pressure head at the second node ψ 2 in the direction of z-axis for
the determination of the ground surface flux as given below,
 
(−8ψb + 9ψ1 − ψ 2 )
Vz |z=0 = −K (ψb ) (23.3)
3z

As time progresses, the water gets infiltrated into soil, which starts decreasing of
flow depth at the surface. When all water gets infiltrated into soil, then the upper
boundary is changed to the flux-type boundary. The flux boundary is taken as the
irrigation and rainfall infiltration and actual rate of evaporation at the first node near
the ground surface with a condition that it cannot falls below ψ min , where

RT (t)
ψlim = ln[ f (t)] (23.4)
Mg

where ψ min = ψ b is the pressure head at the surface (bar); R = the universal gas
constant (8.31 × 107 eg/mole/K); T (t) is the absolute temperature (0 K); M =
molecular weight of water (18 gm/mole); g = acceleration due to gravity.

23.2.2 Root Growth Function and Potential

Evapotranspiration

Atmospheric boundary condition requires specifying rate of potential evapotranspi-

ration, irrigation and precipitation. Potential evapotranspiration is calculated using
pan evaporation data. For more accurate rather than considering pan evaporation
method, evaporation is calculated using Christiansen (1968). Pan evaporation data
is related to the potential evapotranspiration (E vpt ) using equation E vpt = K E p
as suggested by Allen et al. (1998), where E p is the pan evaporation, and K is the
consumptive use coefficient, which depends upon the type of crops and growth stages
and characterizes evaporation and plant water uptake relative to pan evaporation data.
The distribution of the potential evapotranspiration alone the root depth is given as,

E pdis (z) = Crt E vpt (0, t) For 0 ≤ z ≤ d (23.5)

where E pdis (z) = distribution of potential evapotranspiration based on root density;

Crt = coefficient of the root density at depth z; and d = depth of the root zone.
The root growth function or coefficient of root density is considered as the function
defined by Molz and Remson (1970).
The actual evapotranspiration is calculated based on the available soil moisture
at that node and is given as (Zaradny 1993):
23 Agricultural Water Management and Groundwater Recharging … 281
⎧ 
⎪
⎪ E θ−θw
For −15800cm ≤ ψ ≤ −400cm
⎨ pdis θ f −θ w

E a (z, t) = E pdis  For −400cm ≤ ψ ≤ −50cm (23.6)

⎪
⎪
⎩ E pdis θs −θ For −50cm ≤ ψ ≤ 0cm
θs −θan

where E a (z, t) is the actual evapotranspiration from the root of the crop (cm) at
particular node; θw is the soil moisture at wilting point; θ f is the soil moisture
content at field capacity; θan is the soil moisture content at anaerobiosis point; and θs
is the saturated moisture content. The variation of actual evapotranspiration (AET)
along the root zone depth is given by (Michael 1978)
⎧
⎪ 0.4 × E a (0, t) For 0d ≤ z ≤ 4d
d
⎪
⎨
0.3 × E a (0, t) For 4 ≤ z ≤ 2
St (0, t) = (23.7)
⎪
⎪ 0.2 × E a (0, t) For d2 ≤ z ≤ 3×d
⎩ 4
0.1 × E a (0, t) For 3×d
4
≤z≤d

23.2.3 Laboratory Experiment and Parameter Estimation

The model simulation requires to specify the parameters θs , θr and K s . Five soil
samples were collected from the field and tested in the equipment, pressure plate
apparatus for the soil moisture retention curve. Soil moistures corresponding to
applied pressure are recorded for the laboratory observed soil moisture retention
curve. Pedotransfer function model, RETC Code (van Genuchten et al. 1991), anal-
yses the hydraulic conductivity functions and soil water retention of unsaturated soil
from subset of the measured pressure head and water content data. The observed
pressure head vs. soil moisture data provided as the input to the RETC Code and
obtained the parameters soil moisture retention curve θs , θr and K s . The simulated soil
moisture retention curve fitted in MATLAB and obtained the Haverkamp constants,
m and n. The soil samples are subjected to particle size analysis for the soil type.
The estimated parameters are given in Table 23.1.

Table 23.1 Hydraulic properties of soil sample

Hydraulic properties θs θr A m B n
Loam soil 0.3460 0.038 121.1 2.446 29.45 0.8011
282 A. Thomas et al.

23.2.4 Simulation Parameters

Daily climatic data including pan evaporation, rainfall, mean temperature, wind
velocity, relative humidity and mean sunshine hour have been collected from ICAR
Patna near wheat-cropped site from 21th November to 20th March for the duration
of 120 days. Accurate pan evaporation value calculated using Christiansen (1968)
and rainfall event is shown in Fig. 23.1. The simulation domain parameters and the
crop parameters are explained in the Tables 23.2 and 23.3 respectively.

23.2.5 Numerical Method

Soil moisture dynamics of vadose zone is simulated using 1D Richard’s equation,

and water extraction pattern is calculated according to the macroscopic approach

Number of Days from sowing (nov 21th 2013 to march 20th 2014)

105
113
121
17
25
33
41
49
57
65
73
81
89
97
1
9

0 5.0
4.5

Pan Evaporation (mm)

0.5
4.0
1 rainfall
3.5
Rainfall (cm)

Pan Evaporation
1.5 3.0
2.5
2 2.0
2.5 1.5
1.0
3
0.5
3.5 0.0

Fig. 23.1 Daily pan evaporation, precipitation and irrigation for the entire crop period

Table 23.2 Domain

Domain parameters Details
parameters used in simulation
of soil moisture regime in Node size 2 cm
wheat-cropped site Depth of simulation 200 cm
Initial condition −40 cm (suction head at field
capacity)
Upper boundary condition Both flux and head boundary
condition
Lower boundary condition Groundwater table
Field capacity −40 cm
Permanent wilting point −15,800 cm
23 Agricultural Water Management and Groundwater Recharging … 283

Table 23.3 Crop parameters

Crop parameters Details
used in simulation for wheat
crop Crop type Wheat
Rooting depth 60 cm
Crop duration 120 days
(Initial growth stage 0–20 days
Development stage—20–40 days
Mid-season 40–90 days
Late season 90–120 days)
Irrigation details Four irrigations with 6 cm depth each
(On 21th, 60th, 90th and 110th days)

suggested by Molz and Remson (1970) due to its simplicity and requirement of less
number of model parameters and solved numerically along with an appropriate initial
and boundary conditions. The solution of the partial difference equation is obtained
using the strongly implicit finite-difference scheme (Hong et al. 1994; Singh and
Bhallamudi 1998). The equation is specified at each node, and numerical technique
is used to solve the nonlinear equation in each time steps. At each iteration process,
a set of algebraic equations in tridiagonal matrix form are formulated. The iterative
process will continue to obtain soil moisture content until all nodes reach a reasonable
degree of convergence. If the soil moisture of a particular node is more than the
field capacity, it will transfer to the next node below it and finally added up to get
groundwater recharge. A programming code is developed in FORTRAN for solving
the governing equations.

23.3 Results and Discussion

The model is simulated for a simplified case without root uptake and compared well
with the available literature data. The model is then applied to a wheat field for
predicting soil moisture regime sand subsequent groundwater recharge.

23.3.1 Soil Moisture Flow Without Root Water Uptake

The model is first validated over the soil sample simulated by Haverkamp et al.
(1977) without sink term. The input parameters for the characteristics relationship
were saturated hydraulic conductivity, K s = 34 cm/hr; saturated moisture content, θ s
= 0.287; residual moisture content, θ r = 0.075; the constants A = 1.175 × 106 ; m =
4.74; B = 1.611 × 106 ; and n = 3.96. The depth of the soil column was 80 cm. The
initial condition was specified as the volumetric moisture content, θ ini = 0.1. The
boundary condition at the top was considered as the constant flux = 13.69 cm/hr, and
the bottom boundary condition was specified as the constant moisture contents 0.1
284 A. Thomas et al.

Fig. 23.2 Verification of the Volumetric Moisture Content

model with experimental 0.0 0.1 0.2 0.3
data: soil moisture profiles 0
dashed lines for
0.1 hr
experimental and solid lines -10
for present model 0.2 hr
-20
0.3 hr

-30 0.4 hr

Depth (cm)
0.5 hr
-40
0.6 hr
-50
0.7 hr
-60 0.8 hr

-70 Present Model

Haverkamp et al. 1977
-80

(Haverkamp et al. 1977). Figure 23.2 shows the movement of moisture content with
different times computed with present model and Haverkamp et al. (1977). Results
of both the models are very close.

23.3.2 Soil Moisture Flow Through the Vadose Zone

The soil moisture movement through the soil after the application of the irrigation on
21th day of crop period is presented in Fig. 23.3. Soil moisture moves slowly down
through the soil to reach the entire soil column into its field capacity. Soil moistude
content above the field capacity add as the recharge to the groundwater. Eventually,
the soil moisture decreased due to the out flux in the form of root water uptake.

23.3.3 Prediction of Irrigation Scheduling

During the early days of irrigation, the soil moisture is abundant, and the water uptake
is at its optimal condition. The loss of soil moisture from top layer by evaporation and
water uptake by plant roots decreases the soil moisture status eventually depending
on the root density distribution. At this stage, root water uptake will be in its potential
rate. Once the soil moisture content falls down to wilting point, further extraction from
that node is zero, and evapotranspiration reduces than its potential value. Prolonged
water-stressed condition results in the wilting of the plants, and gradually, permanent
wilting occurs. This will affect the yield of the crop.
23 Agricultural Water Management and Groundwater Recharging … 285

Soil moisture content

0 0.2 0.4
0

-50

-100
Depth (cm)

-150 day 20
day 21
day 22
-200 day 23
day 30

-250

Fig. 23.3 Movement of soil moisture flux through the soil column under saturated bottom boundary
condition

Due to continuous loss of soil moisture from the top layer, the availability
decreased, and the actual evapotranspiration becomes less than its potential value.
From Fig. 23.4, it can be observed that actual evapotranspiration is less than its poten-
tial during the 38th–60th day and from 76 to 84th day. Irrigation is recommended
during this period.

0.30
0.25
PEt(cm/day)
PEt/AEt (cm)

0.20 AEt(cm/day)
0.15
0.10
0.05
0.00
0 20 40 60 80 100 120
Number of Days from sowing

Fig. 23.4 Potential evapotranspiration (PEt) and actual evapotranspiration (AEt) of the wheat crop
286 A. Thomas et al.

45
40
35 Water applied
Depth of water (cm)

30 Recharge
25
20
15
10
5
0
-5
-10
-15
0 20 40 60 80 100 120
Days after sowing

Fig. 23.5 Recharge of groundwater over the crop peroid

23.3.4 Groundwater Recharged for the Wheat Crop Region

The developed model is used for recharge estimation in an agricultural field with
loam soil. Over the entire crop period of 120 days (21st November 2013–20th March
2014), total 10.614 cm of rainfall has occurred, and 24 cm of irrigation water has
been applied. Thus, total water application is 34.614 cm. In early irrigation, all the
water was built as the soil moisture in the soil column, and some water has added
from the groundwater to soil. It is shown as negative sign in Fig. 23.5. During the
entire crop period, total 8.26 cm of water has been recharged, 16.25 cm has been lost
as evapotranspiration, and 10.01 cm of water has been stored in the soil column as
soil moisture.

23.4 Conclusions

A 1D Richard’s equation-based soil moisture model has been developed through

this study for the prediction of temporal and spatial movement of soil moisture after
irrigation or rainfall and the contribution to underlying groundwater. Sink term is
coupled with mixed form of Richard’s equation for the account of root water uptake
by plants variation. The nonlinear Richard’s equation is solved numerically using
highly implicit finite-difference method, and iteration technique is used for predicting
unknown soil moisture and suction head. The developed model is validated with
results of Haverkamp et al. (1977). Results of present model are very close to the
Haverkamp’s results.
23 Agricultural Water Management and Groundwater Recharging … 287

The model has been used to simulate a wheat-cropped site with loam soil in Patna
district Bihar. The daily soil moisture regime, suction head and actual evapotran-
spiration are simulated for the entire crop period. During the crop period from 21th
November to March 20th, total 10.614 cm of rainfall has occurred, and 24 cm of irri-
gation water has been applied. Out of 34.6 cm of total water applied, 23.9% of water
has recharged, 47.2% has been lost as evapotranspiration, and 28.9% of water has
been stored in the soil column as soil moisture. The model is strictly following water
balance, and hence, it is justifiable. From the simulation, it is noted that the actual
evapotranspiration is less than its potential rate during 38th–60th days and from 76
to 84th days. Thus, additional irrigation is recommended during this periods.
The model developed can be applied at any land area if the weather data, soil
type and cultivation details are available. Thus, the advantage of this model is that
it can be used for recharge estimation of a cultivated area, and it can be obtained
without visiting a field or detailed data. It can also give the information of irrigation
requirements of the crop. The irrigation schedule can be planned with the help of
soil moisture profile, and the depth of water required for the irrigation also can be
determined.

References

Allen RG, Pereira LS, Raes D, Smith M (1998) Crop evapotranspiration. Guidelines for computing
crop water requirements. Irrigation and drainage paper no. 56, FAO. Rome, Italy
Christiansen JE (1968) Pan evaporation and evapotranspiration from climatic data. J Irrig Drainage
Eng, ASCE 94(2):243–265
Feddes RA, Raats PAC (2004) Parameterizing the soil–water–plant root system. In: Feddes RA, de
Rooij GH, van Dam JC (eds) Proceedings of the unsaturated zone modelling: progress, challenges
and applications. Wageningen UR Frontis Series, vol 6. Kluwer Academic Publishers, Dordrecht,
The Netherlands, x–xi (Chapter 4), pp 95–141
Feddes RA, Kotwalik PJ, Zaradny H (1978) Simulation of field water use and crop yield. Centre
for Agricultural Publishing and Documentation, Wageningen, Netherlands
Haverkamp R, Vauclin M, Touma J, Wieranga PJ, Vachaud G (1977) A comparison of numerical
simulation models for one dimensional infiltration. Soil Sci Soc Am J 41:285–294
Jimenez-Martinez J, Skaggs TH, van Genuchten MT, Candela L (2009) A root zone modelling
approach to estimating groundwater recharge from irrigated areas. J Hydrol 367(1–2):138–149
Kang S, Zhang F, Zhang J (2001) A simulation model of water dynamics in winter wheat field and
its application in a semiarid region. Agric Water Manage 49:115–129
Li KY, Boisvert JB, De Jong R (1999) An exponential root water uptake model. Can J Soil Sci
79:333–343
Michael AM (1978) Irrigation: theory and practice. Vikas Publishing House PVT Ltd
Min L, Shen Y, Pei H (2015) Estimating groundwater recharge using deep vadose zone data under
typical irrigated cropland in the piedmont region of the North China plain. J Hydrol Eng 527:305–
315
Molz FJ, Remson I (1970) Extraction term models of soil moisture use by transpiring plants. Water
Resour 6(5):1346–1356
Ojha CSP, Rai AK (1996) Nonlinear root water uptake model. J Irrig Drainage Eng 122:198–202
Prasad R (1988) A linear root water uptake model. J Hydrol 99:297–306
288 A. Thomas et al.

Rushton K (1988) Numerical and conceptual models for recharge estimation in arid and semi-arid
zones. Estimation of natural groundwater recharge. Springer, pp 223–238
Shankar V, Hari Prasad K, Ojha C, Govindaraju R (2012) Model for nonlinear root water uptake
parameter. J Irrig Drainage Eng 10:1061
Singh V, Bhallamudi SM (1998) Conjunctive surface subsurface modelling of overland flow. Adv
Water Resour 21(7):567–579
Singh SK, Ghosh AK (2011) Entry of arsenic into food material—a case study. World Appl Sci J
13(2):385–390
Šimůnek J, Hopmans JW (2009) Modeling compensated root water and nutrient uptake. Ecol Model
220(4):505–521
van Genuchten MTh, Leij FJ, Yates SR (1991) The RETC code for quantifying the hydraulic
functions of unsaturated soils. USEPA, Washington, DC
van Lier QDJ, Metselaar K, van Dam JC (2006) Root water extraction and limiting soil hydraulic
conditions estimated by numerical simulation. Vadose Zone J 5:1264–1277
Wohling DL, Leaney FW, Crosbie RS (2012) Deep drainage estimates using multiple linear
regression with percent clay content and rainfall. Hydrol Earth Syst Sci 16(2):563–572
Yadav B, Junaid S (2014) Groundwater vulnerability assessment to contamination using soil
moisture flow and solute transport modeling. J Irrig Drainage Eng 10:1061
Yadav BK, Mathur S (2008) Modeling soil water uptake by plants using nonlinear dynamic root
density distribution function. J Irrig Drainage Eng 134(4):430–436
Yadav BK, Mathur S, Siebel MA (2009) Soil moisture flow modeling with water uptake by plants
(wheat) under varying soil and moisture conditions. J Irrig Drainage Eng 135(3):375–381
Zaradny H (1993) Groundwater flow in saturated and unsaturated soil. Zeidler RB (ed) Balkema,
Rotterdam, The Netherlands