International Journal of Remote Sensing Applications Volume 3 Issue 2, June 2013 www.ijrsa.

org

Delineation of Flood Prone Area using

Normalized Difference Water Index (NDWI)
and Transect Method: A Case Study of
Kashmir Valley
Hilal Ahmad Ganaie1, Haseena Hashia2 and Dheera Kalota3
Department of Geography,Jamia Millia Islamia, Neew Delhi-110025, India
[email protected]; [email protected]; [email protected]
1

Abstract prone areas depends upon the magnitude and

Floods of River Jhelum in Kashmir Valley are as old as the frequency of floods. Flood is a relatively high flow of
river itself. Historical accounts suggest that the river had water that overtops the natural and artificial banks.
witnessed the devastating floods since ages many among Water spreads over flood plain and generally causes
which have created havoc in terms their resultant damage to inhabitants, crops and vegetation. During
destruction. The area which is mainly affected by floods are extreme flood event it is important to determine
the Jhelum valley floor stretching from Anantnag in the quickly the extent of flood water and different classes
south to Baramulla in the north. The Valley is an ancient lake of land uses, which comes under water (Wang et al.
basin 140 kilometre long and 32 kilometre wide. In this
2002). Flood map can be applied to develop
paper delineation of flood plain of Kashmir Valley has been
carried out using Geographical Information System. The aim
comprehensive relief effort immediately after flooding,
of this paper is to determine the accuracy of using simple Various Expert Groups like Working Group for Flood
digital image processing techniques to map riverine water Control Programme for Xth Plan Rashtriya Barh Ayog
bodies with Landsat 5 TM and Shuttle Radar Topography (RBA) in 1980 in their reports have mentioned that
Mission (SRTM) Digital Elevation Model (DEM) data. This there should be authentic flood prone area figures and
paper is divided into two parts, part-I deal with the suggested a scientific approach to be used in
Normalized Difference Water Index of Kashmir Valley and identifying the flood prone areas.
part-II deals with Transect method of river Jhelum. The 3D
DEM and recent flood affected areas were used to A floodplain is a strip of relatively smooth land
supplement the above methodologies. bordering a stream and over flowed at a time of high
water (Leopold et al. 1964). Floodplains can be looked
Keywords
at from different perspectives. As a topographic
River Jhelum; Flood Plain Delineation; NDWI; Transect Method category it is a quite flat area and lies adjacent to a
river. Geomorphologically, it is a landform composed
I nt roduc tion
primarily of unconsolidated depositional material
Floodplains are hazardous to developmental activities derived from sediments being transported by the
if the vulnerability of those activities exceeds an related stream. Hydrologically, it is best defined as a
acceptable level. Delineating flood affected areas and land with different return periods of the parent stream.
water bodies in general are always the most critical A combination of these (characteristics) perhaps
concern to deal with flood management operation. It is comprises the essential criteria for defining the
also an important input for planning and efficient floodplain (Schmudde 1968). However, an important
management of flood affected areas. Accurate process resulting in the formation of Valley flats is a
information on the extent of water bodies is important combination of deposition on the inside of river curves
for flood prediction, monitoring and relief. Delineating and deposition from overbank flows. This process
non-flooded areas are also equally important because produced many of the flat areas adjoining river
these areas can serve as a temporary shelter for the channels and probably most of the flood plains of the
nearby affected areas. The management of the flood great rivers of the world.

53
www.ijrsa.org International Journal of Remote Sensing Applications Volume3 Issue 2, June 2013

As per Rashtriya Barh Ayog (RBA), the maximum area floods in Kashmir valley. One of the major rivers
damaged/affected in any one of the years is assumed which run across the length of valley also known as its
to be the area liable to floods in the state. The total of lifeline is the River Jhelum. The Jhelum is the main
such maxima of the various states is considered to be River of Kashmir valley with most of the towns and
the area liable to floods (Flood Prone Area) in the villages settled on its banks. The Jhelum and the host
country. The RBA, however, recognized that annual of streams constitute the drainage network of Kashmir
flooding is not coextensive and that different areas are valley. Every single drop of water, anywhere in the
often flooded in different years by different streams. valley has to merge in the Jhelum that means any
As per Centre Water Commission (CWC) records, stream, rivulets and Nallas that flow in the valley
based upon data of 1953-1978, the maximum area ultimately merges into the Jhelum. The total length of
affected in any year in the country is 17.50 million the river is 212Kms. Its basin lies between the Greater
hectares as against the area liable to floods which are Himalayas and Pirpanjal range.
40 million hectares. The figure of 40 million hectares
includes protected area of 10 million hectares, which
was once flood affected area (Purba et al. 2006).
The information regarding the delineation of flood
prone areas is difficult to collect using traditional
survey techniques particularly during the floods.
Remote Sensing technique is one of the efficient means
of delineating water boundaries over a large area at a
given point of time. The efficient methods for
delineating flood extent using Landsat Thematic
Mapper (TM) by distinguishing water and non-water
areas are based on reflectance characteristics of a pair
FIG. 1 LOCATION MAP OF KASHMIR VALLEY
of images before and after a flood event using
TM7+TM4 formula (Wang et al. 2002). For extracting Flood Area De line at ion
moist areas Hudson used band 7 and 5 of Landsat
Enhanced Thematic Mapper (ETM) image for The delineation of flood prone areas in the Kashmir
separating permanent water, flooded and non-flooded Valley has been carried out with various methods and
areas after a flood event in Mexico (Hudson et al. 2003). techniques, like: Transect Method and NDWI. These
methods have been done using Landsat TM and
St udy Are a Shuttle Radar Topography Mission (SRTM) Digital
Elevation Model (DEM) data. These methods are
Kashmir valley one of the significant parts of Jammu
economical and efficient for delineating flood prone
And Kashmir State (India) is located between the
areas. The best methodology for the flood hazard
Jammu and Ladakh. It is situated between 33°25 to
delineation is using the pre-flood, during flood and
34°30´ north latitude and 73°55´to 75°35´ east
past-flood satellite data. Remote Sensing data can be
longitude. Fig 1 shows the location of Kashmir Valley.
used to develop flood map in an efficient and effective
The Valley is an ancient lake basin 140 kms. long and
way. From the early era of passive Remote Sensing,
32 kms. wide. The average elevation of the valley is
special attention has been given to distinguishing
5,300 feet above sea level. The tall mountains that
water from dry surface. Multi Spectral Scanner (MSS)
surround the valley from south and southwest by Pir
band (0.8-1.1 Micrometer ‘μm’) has been found to be
Panjal range, in north by Kashmir Himalaya and in
particularly suitable for distinguishing water or moist
north and north east by Greater Himalayan range,
soil from dry surface due to its strong absorption of
which rising up to 16,000 feet ensure that the weather
water in the Near Infrared (NIR) range of the
here is pleasant for most of the year and also gives
spectrum (Smith 1997). Multi Spectral Scanner data
birth to number of glacier-fed streams.
has been used to deal with the flood-affected areas in
Climatologically, this area has semi-arid and cold Iowa (Rango and Solomonson 1974). For flood area
climate that includes cold winter associated with snow delineation two sets of satellite data are required, one
and frost and moderate summer. In most cases the set consisting of data acquired before the flood event
melting snow and rainfall are the main cause for and the other acquired during the flood occurrence.

54
International Journal of Remote Sensing Applications Volume 3 Issue 2, June 2013 www.ijrsa.org

But there are certain limitations which do not allow us imageries. The NDWI is used to reach the goal of
to go through this methodology for flood delineation isolating water and non-water features in Kashmir
as the satellite data is not readily available on various Valley. The water bodies have a unique spectral
flood events. Keeping in view the data limitations the response in the visible range when compared to the
feasible techniques of flood prone areas delineation surrounding land cover. The infrared band
are NDWI and Transect Method. classification gives a much better representation in the
water related features than other visible bands. There
M e thods are various definitions of NDWI that combine
different pairs of bands (normally of TM or ETM),
The delineation of flood prone areas in the Kashmir
which include, Green and Near Infrared (NIR) (band2
Valley has been generated by:
and band4) (McFeeters 1996), NIR and Short Wave
NDWI (Normalize Difference Water Index) and Infrared (SWIR) (band4 and band5) (Gao 1996) and
Transect Method. Red band and Middle Infrared (MIR) (band3 and
band5) (Xiao et al. 2002). There are several studies
Normalize Difference Water Index: The band Math employing these pairs of bands to delineate flood
function in ENVI 4.7 is used for computation of NDWI. prone areas. McFeeters NDWI is composed as:
The equation (2) is typed in the calculator and then the
values b1 and b2 are identified in the LANDSAT MSS.
NDWI=[(b1-b2)/(b1+b2)] (1) NDWI= OR (2)

b1 represents Green Where Green is a band that encompasses reflected

green light and NIR represents reflected near-infrared
b2 represents NIR (Near Infrared) radiation. The selection of these wavelengths was used
When equation (1) is used on Landsat MS satellite because of following criteria:
image that contains reflected visible Green band and  To maximise the typical reflectance of water
NIR. Water features have positive values, while soil features by using green light wavelengths.
 To minimise the low reflectance of NIR by
and terrestrial vegetation features have zero or
negative values.
water features.
Transect Method (Topographic profile): Transect  To take advantage of the high reflectance of
method or simply topographic profile has been used NIR by terrestrial vegetation and soil features.
on five main discharge gauge stations, where peak
When equation (2) is used to process a multi-spectral
water level data was available. Topographic profile
satellite image that contains a reflected visible green
has been done on Arc GIS 3D Analyst toolbar with the
band and NIR band, water features have positive
help of Interpolate Line tool, which clearly indicated
values, while soil and terrestrial vegetation features
the width of the flood plain on the SRTM DEM. The
have zero or negative values, owing to their typically
profile graph with distance along the profile route (in
higher reflectance of NIR than green light (McFeeters
map units) is on the X axis, and elevation on the Y axis.
1996).
Re sults a nd Disc ussion

Normalized Difference Water Index (NDWI)

There are numerous vegetation indices developed to
estimate vegetation cover with the Remote Sensing
imageries. A vegetation index is a number that is
generated by some combination of Remote Sensing
bands. The most common spectral index used to
evaluate vegetation cover is the Normalized
Difference Vegetation Index (NDVI). McFeeters
developed an index similar to the NDVI, which is
called the NDWI. The Normalized Difference Water
FIG. 2 THE NORMALIZED DIFFERENCE WATER INDEX AT FIVE
Index (NDWI) is being used to delineate open water DISCHARGE STATIONS OF RIVER JHELUM IN KASHMIR
features and enhance their presence in Remote Sensing VALLEY

55
www.ijrsa.org International Journal of Remote Sensing Applications Volume3 Issue 2, June 2013

Landsat TM image of the study area of October, 2010 water level and area of inundation. Alternatively, the
was used to generate NDWI. The NDWI of Kashmir relationship between stream flow, water level and
Valley is shown in Fig. 3, but due to the non visibility habitat area can be developed through the acquisition
of water bodies and the course of river Jhelum, we of aerial photography at different stream flows. Where
generated NDWI at five discharge stations of river the FTM gives a vertical two-dimensional perspective,
Jhelum. The Fig. 2 clearly shows the water and land Remote Sensing Method (RSM) provides a horizontal
masses of those five discharge stations. These stations two-dimensional perspective. The linking of both the
are Sangam, Pampore, Ram Munshibagh, Asham, and ground based FTM and remotely sensed information
Baramula. The availability of highest annual water through modelling RSM provides both the cross-
levels from 1980 to 2008 is the main reason for the sectional and channel length perspective for
selection of above stations. If NDWI technique is used determining habitat area and its relationship with
for the flood image, it would provide accurate results stream flow (Burack et al. 2009).
in the delineation of flood prone areas. In the absence
Generally, the active channel and floodplain are the
of such data, an alternative method, Transect Method
main types of transect. The active channel transect is
has also been used for delineation of flood prone areas
the sum of the widths of the actively flowing channels
in Kashmir Valley.
plus the bared bars. The floodplain transect is defined
as the Holocene Valley, which can be identified using
geological maps and DEM2. Often the computed water
surface elevations are manually plotted on maps in
order to delineate floodplains. Automatically this
manual plotting would result in significant savings of
both time and resources. A Geographic Information
System (GIS) offers the ideal environment for these
types of work through transect method. The Table 4.1
shows the highest water level above mean sea level at
different stations.
TABLE 1 HIGHEST ANNUAL WATER LEVELS (METERS) FROM
1980-2008

Ram
Sanga Pamp Baram
Station Munshi Asham
FIG. 3 THE NORMALIZED DIFFERENCE WATER INDEX OF m ore ula
KASHMIR VALLEY
bagh
Highest
Transect Method to Delineate Flood Prone Areas Water
Levels 1,594 1,588 1,587 1,582 1,578
A transect is a path which can count and record (meters
occurrences of the phenomena of the study. Transect AMSL)
measurements can be performed manually or through Year 1988 1995 1995 1995 1995
GIS application. In this method, the landscape source: irrigation and flood control department srinagar. j&k
positions (relative elevations) of riparian resources By using GIS based FTM for delineating flood prone
that are dependent on or sensitive to river water levels areas, the topographical profile of Kashmir Valley was
are surveyed along transects through the floodplain developed (Fig. 4.4). The topographical profile was
and river channel 1 . Observed water levels that generated on five main discharge stations which
inundate or saturate these landscape positions are depict the cross profile of the Jhelum flood plain. The
correlated with water flows recorded. flood plain of river Jhelum is narrow at Sangam
The Flood Transect Method (FTM) is a field intensive (Anantnag), therefore, the water level at this station is
method which establishes reference transects that span higher than the other discharge stations. The highest
the river and onto the surrounding riparian zone. level of Jhelum was recorded 1,594 meters AMSL (in
Through field visits and surveys at different stream year 1988) at Sangam from 1980 to 2008. The
flows, a relationship can be developed between flow, topographical profile of Sangam shows the cross

2Holocene is defined as the time period encompassing the last 8,000

1
years to the present

56
International Journal of Remote Sensing Applications Volume 3 Issue 2, June 2013 www.ijrsa.org

section of the river extends upto 4.5 kms on Jhelum 1,578 metres AMSL (1995) and the total width about
flood plain. The next site is at Pampore. The discharge 0.7 to 1.5 kms. After this station, river Jhelum leaves
data of this station shows that highest water level was Kashmir Valley and enters Muzaffarabad [Pakistan
1,588 meters AMSL (1995). The below topographical Occupied Kashmir (POK)].
profile of Pampore (Fig 4.4) shows that right side of
Finally three dimensional map (Fig. 5) has been
river Jhelum has not been affected by the floods due to
generated with the help of Arc Scene to show the flood
the Pampore Karewa. This shows that cross section of
plain areas of the Kashmir Valley. The 3 D Image
the river extends upto 6.5 kms on Jhelum flood plain.
clearly indicates that the flood plain is mostly on left
However, it extends when the embankment is
side of the river Jhelum. The usually flood affected
breached and in this condition the water level reaches
areas has been demarcated by the Irrigation and Flood
upto 10-15 kms and inundates those areas which were
Control Department of Srinagar. Fig. 6 shows the flood
not affected by the floods.
affected areas of the Kashmir Valley.

FIG. 5 THREE DIMENSIONAL MAP OF KASHMIR VALLEY

FIG. 4 THE TOPOGRAPHIC PROFILE OF KASHMIR VALLEY AT

FIVE MAJOR DISCHARGE STATIONS OF RIVER JHELUM
The third station is at Ram Munshibagh (Srinagar)
which is approximately situated at the middle of river
Jhelum. This area was highly flood prone but after the
construction of the Flood Spill Channel in 1903,
regulated discharge is flowing safely through the
Srinagar city. The topographic profile of Ram
Munshibagh station (Fig 4) shows that the left bank of
the river has large number of depressions which are
flooded during the peak flow discharge. The
maximum gauge height at this station during the past
flood was 1,587.76 meters AMSL (1995), while the FIG. 6 FLOOD PRONE AREAS OF KASHMIR VALLEY
cross section profile shows that it covers the width of
approximately 7 and 2 km on its left and right banks, Conc lusions
respectively. Srinagar city is located on its right bank, Delineating flood prone areas play an important role
where a large area is below the high water level but in flood mitigating measures. The NDWI and transect
with the help of embankments and good drainage method clearly shows that the lower Jhelum areas of
system the city was less affected. Asham (Bandipora) Kashmir valley are more affected then the upper
station is below the Shadi Pora (Bandipora) where the Jhelum areas. The transect method shows that the
width of the Jhelum is about 35 kms (Raza et al. 1978). width of lower Jhelum areas is 20-30 kms and the
The Asham falls in the lower course of the river upper Jhelum have 5-10 kms width. The flood
Jhelum and the highest water level was 1,581 meters management policy makers should mostly concern on
AMSL (1995). The topographic profile shows that the lower Jhelum areas and funds should be allocated
extent of the river is about 19 kms. The last station is at according to frequently flood affected areas for proper
Baramula, the highest water level at this station being managemen.

57
www.ijrsa.org International Journal of Remote Sensing Applications Volume3 Issue 2, June 2013

REFERENCES

Burack, T. S., et al., “Comparison of Floodplain Transect and data”. International Journal of Remote sensing,23(18): pp.
Remote Sensing Methods For the Lamprey River 3681–3696, 2002.
Instream Flow”. Report No. R-WD-09-9, Department of Xiao, X., et al. “Observation of Flooding and Rice
Environmental Services, New Hampshire. Available Transplanting of Paddy Rice Fields at the Site to
online: Landscape Scales in China Using Vegetation Sensor
http://www.des.nh.gov/organization/divisions/water/wmb/rive Data”. International Journal of Remote Sensing, 23(15): pp.
rs/instream/index (Assessed on 30/02/ 2011). 3009-3022, 2002.
Gao, B. C. “NDWI–A Normalized Difference Water Index
for Remote Sensing of Vegetation Liquid Water from Name: Ganaie., H. A
Place: University of Jamaia Millia Islamia,
Space”. Remote Sensing of Environment, 58: pp. 257–266, New Delhi
1996. DOB: 30/11/1981
Hudson, P. F. and Rene´ R. Colditz “Flood Delineation in a Educational Background: Ph. D from
University of Jamaia Millia Islamia, New
Large and Complex Alluvial Valley, Lower Pa´Nuco
Delhi, India in 2013.
Basin, Mexico”. Journal of Hydrology, 280: pp. 229–245. Specilization: Diasater Management,
2003. Remote Sensing and Hydrology.
1. “The Composition, Processing, Recycling and Dumping of
Leopold, L. B., et al., “Fluvial Processes in Geomorphology”.
E-waste in the National Capital Region, India, International
W.H. Freeman and Company San Francisco and London, Journal of Sustainability Development and Green Economics,
1964. (2012), Environment and Sustainable development, Issues
and challenges” 2.“Flood Disaster and its Management: An
McFeeters, S. K. “The Use of the Normalized Difference
Exploratory Overview of River Jhelum in Kashmir Valley”. Singh
Water Index (NDWI) in the Delineation of open Water P. J., (eds) “Keshav Publications, Ghaziabad (New Delhi
Features”. International Journal of Remote Sensing. 17(7): NCR) ( 2012), 3.The Case structure in Kashmir Valley. Jamia
Geographical Studies (ed). Manak Publications Pvt. Ltd.
pp. 1425-1432, 1996.
New Delhi, 110092. (2012).
Purba, G. S., et al., “Identification of Flood Affected Areas–
Need for a Scientific Approach”. Proceedings Indian Name: Hashia, H,
Place: University of Jamaia Millia Islamia,
Disaster Management Congress, Organized by National
New Delhi
Institute of Disaster Management and Central Water DOB: 25/01/1959
Commission, at Vigyan Bhawan, New Delhi, November 28- Educational Background: Ph. D from
30, 2006. University of Jamaia Millia Islamia, New
Delhi, India in 1991.
Rango A. and V. V. Solomonson “Regional Flood Mapping Specilization: Gender Geography
from Space”. Water Resources Research 10(3): pp. 473–484, She is working as Proffessor in Jamaia Millia Islamia
1974. University. She has number of publication and have
presented number of papers in International conferences.
Raza, M., et al., “The Valley of Kashmir: A Geographical
Interpretation”. Vikas Publishing House, New Delhi, Name: Kalota., D.,
1978. Place: University of Jamaia Millia
Islamia, New Delhi
Schmudde, T. H. “Floodplain”. In R.W. Fairbridge (ed.) “The
DOB: 30/11/1984
Encyclopedia of Geomorphology”. New York: Reinhold: pp. Educational Background: Ph. D from
359-362. 1968. University of Jamaia Millia Islamia, New
Delhi, India in 2013,
Smith L. C. “Satellite Remote Sensing of River Inundation
Research Scholar in Dept. of Geography,
Area, Stage, and Discharge: A Review”. Hydrological Jamia Millia Islamia, New Delhi, India.
Processes, 11: pp. 1427–1439, 1997. Specilization: Remote Sensing.
Wang, Y. et al., “An Efficient Method for Mapping Flood
Extent in a Coastal Flood Using Landsat TM and DEM

58