Prepared by

Barry Croke, Felix Andrews,

Susan Cuddy

www.toolkit.net.au/ihacres

WORKBOOK
Document History
Date Author Version Description of Change
15 Feb 2005 Susan Cuddy V1 First compilation of material

Copyright notice
© The Australian National University, Australia 2005

Legal Information
To the extend permitted by law, the Australian National University and the Cooperative Research Centre for
Catchment Hydrology (including its employees and consultants) accept no responsibility and exclude all
liability whatsoever in respect of any person's use or reliance on this publication or any part of it.

Availability
From early 2005, distribution of the IHACRES software package is via the CRC for Catchment Hydrology’s
Catchment Modelling Toolkit website under an agreement between the Integrated Catchment Assessment
and Management (iCAM) Centre of The Australian National University and the Cooperative Research
Centre for Catchment Hydrology.
The Product Manager for IHACRES is Dr Barry Croke of the iCAM Centre, The Australian National
University, [email protected].
Training workshops are run on an irregular basis, to meet demands.

Contact Details
CRCCH Toolkit website - http://www.toolkit.net.au
Integrated Catchment Assessment and Management Centre
http://icam.anu.edu.au/
School of Resources, Environment and Society
Building 48A Linneaus Way
The Australian National University
Acton ACT 0200

Revisions
IHACRES is under active development and has a planned delivery schedule of new and/or revised features
for the period 2004-2005.
Acknowledgements
IHACRES v2 is a major reimplementation of the original IHACRES model (hereafter known as
the Classic version).
It has been recoded to incorporate additional functionality by its ANU development team,
with the support of a team of beta testers who have extensive experience with the original
version of IHACRES. The development team would like to acknowledge the contribution
made by the following testers:
• Dr Ian Littlewood, Centre for Ecology and Hydrology, Wallingford, Oxfordshire, UK
• Dr Jeff Chanat, Univesity of Virginia, USA
• Dr Teemu Kokkonen, Helsinki University of Technology, Finland
• Dr Barbara Baginska, Environment Protection, Department of Environment and
Conservation, New South Wales, Australia.
Coding and core debugging have been by Alex Luddy and Felix Andrews of the ANU
development team within the Integrated Catchment Assessment and Management (iCAM)
Centre.
Tutorials have been designed and written by Jessica Spate, a PhD student at ANU and in-
house tested by Michael Kehoe.
This Workbook has been compiled by Susan Cuddy, iCAM Centre and CSIRO Land and
Water, from material written by Barry Croke.
 Contents

IHACRES Workbook
CONTENTS

1 Introduction...........................................................1
1.1 The workbook............................................................................................... 1
1.2 IHACRES...................................................................................................... 2
1.2.1 Overview .................................................................................... 2
1.3 Purpose ....................................................................................................... 2
1.4 Supporting Material....................................................................................... 2
1.4.1 Companion Documents................................................................ 2
1.4.2 Input Data .................................................................................. 2

2 Tutorial 1- Murrindindi River ..................................4

2.1 Catchment description ................................................................................... 4
2.2 Input data .................................................................................................... 4
2.3 Synopsis of calibration and simulation ............................................................. 4
Calibration settings ...................................................................... 4
2.4 Calibration results ......................................................................................... 5
2.5 Simulation results .......................................................................................... 5
2.6 Discussion .................................................................................................... 5
2.7 Further Reading ............................................................................................ 6

3 Tutorial 2 – Salmon Brook .....................................7

3.1 Catchment description ................................................................................... 7
3.2 Input data .................................................................................................... 8
3.3 Synopsis of calibration and simulation ............................................................. 8
Calibration settings ...................................................................... 8
Calibration results........................................................................ 8
Simulation results......................................................................... 8
3.4 Step 1 – load the data ................................................................................... 8
3.5 Step 2 – check the data ................................................................................. 9
3.6 Step 3 – define the calibration period(s) ........................................................... 9
3.7 Step 4 – calibrate the linear module ................................................................ 9
3.8 Step 5 – calibrate the non-linear module ......................................................... 9
Pre-Grid search ........................................................................... 9
Instrumental Variable ................................................................... 9
Post Grid search – find a suitable parameter set .............................. 9
Refine the parameter set ............................................................. 10

i
IHACRES Workbook

Accept the parameter set.............................................................10

3.9 Step 6 – simulation ......................................................................................10
3.10 Further Reading ...........................................................................................10

4 Tutorial 3 – Teifi Catchment................................. 11

4.1 Catchment description..................................................................................11
4.2 Input data ...................................................................................................11
4.3 Synopsis of calibration and simulation ............................................................12
Calibration settings.....................................................................12
Calibration results ......................................................................12
Simulation results .......................................................................12
4.4 Step 1 – load the data ..................................................................................12
4.5 Step 2 – check the data ................................................................................12
4.6 Step 3 – define the calibration period(s)..........................................................12
4.7 Step 4 – calibrate the linear module...............................................................13
4.8 Step 5 – calibrate the non-linear module ........................................................13
Pre-Grid search..........................................................................13
Instrumental Variable ..................................................................13
Post Grid search – find a suitable parameter set.............................13
Accept the parameter set.............................................................13
4.9 Step 6 – simulation ......................................................................................13
4.10 Further Reading ...........................................................................................14

ii
 Contents

TABLE OF FIGURES
Figure 3-1: Salmon Brook catchment in Western Australia. Modified from figures provided by
Richard Silberstein (CSIRO Land and Water, WA, Australia) and Department of
Environment, Water and Catchment Protection (Australia). .....................................................7

iii
This page intentionally left blank
Introduction

1 Introduction
IHACRES is a catchment-scale rainfall-streamflow modelling methodology. Its purpose is to
assist the hydrologist or water resources engineer to characterise the dynamic relationship
between basin rainfall and streamflow. Applications include:
• identification of unit hydrographs
• continuous time series streamflow modelling
• environmental change – hydrological regime studies
• runoff event modelling
• hydrograph separation (for example, to assist with water quality investigations)
• derivation of a Slow Flow Index
• derivation of Dynamic Response Characteristics
• investigation of relationships between Dynamic Response Characteristics and physical
catchment descriptors
• teaching unit hydrograph theory and its applications
• hydrometric data quality assurance/control (QA/QC)
• infilling missing streamflow data.

1.1 The workbook

This Workbook provides further tutorials (to that in the User Guide) on how to calibrate
IHACRES models for a range of catchments.
Input data for the examples are either distributed with the IHACRES software package, or can
be downloaded from its Product page on the Toolkit website.
Other examples of calibrated models are documented in journal articles which appear in the
scientific literature (For References, see User Guide, Section 1.6.2).

1
IHACRES Workbook

1.2 IHACRES

1.2.1 Overview
The emphasis in IHACRES is on modelling identifiable catchment-scale rainfall-runoff
behaviour rather than the small-scale hydrological processes by which rainfall causes
streamflow.
To run IHACRES, it must first be calibrated to optimise the performance over a known period.
In this way a catchment can be characterised by a small set of parameters. The model takes
as input, once calibrated, time series of rainfall and either temperature or potential evapo-
transpiration. The output is a time series of modelled stream flow. For the calibration period
and the validation period, observed stream flow is also available – this allows measures of
performance to be computed.
Calibration has two modules, in series:
• a non-linear loss module
• A linear unit hydrograph module.
The linear relationship between effective rainfall and streamflow allows the application of
well-known unit hydrograph theory which conceptualises the catchment as a configuration of
linear storages acting in series and/or parallel. All of the non-linearity commonly observed
between rainfall and streamflow is accommodated in the (non-linear loss) module which
converts rainfall to effective rainfall.
Conceptualisation of spatially distributed processes in both the non-linear and linear modules
is restricted. An advantage of this spatially “lumped” approach is that the model requires
only a small number of parameters, typically six (3 for the non-linear loss module and 3 for
the linear unit hydrograph module). Despite this structural simplicity, the IHACRES
methodology performs well for many types of catchment.

1.3 Purpose
The purpose of this Workbook is to support self-guided and workshop learning and to
illustrate a range of calibrations.
The User Guide contains details on the component parts and uses one tutorial to illustrate
how you move through the system. This by necessity omits a lot of functionality and is also
biased by the hydrological characteristics of the catchment under calibration.
This Workbook is the start of building a comprehensive ‘library’ of calibrated models to
inform the IHACRES user on selection of parameters and how to develop confidence in
calibration.

1.4 Supporting Material

1.4.1 Companion Documents

IHACRES User Guide

1.4.2 Input Data

Input Data for each tutorial is stored in a separate directory under the .\Tutorials directory of
IHACRES.

2
Introduction

It is anticipated that further tutorials will become available between releases. These tutorials,
and their data, will be available by download from the Toolkit website.

3
IHACRES Workbook

2 Tutorial 1-
Murrindindi River
Tutorial 1 is contained in Chapter 5 of the User Guide. It introduces you to a sub-catchment
of the Murrindindi River. It is an example of a smallish ephemeral catchment with strong
base flow.

2.1 Catchment description

Catchment name Murrindindi River subcatchment
Location Tributary of the Goulburn River, located north-east of
Melbourne, Victoria
Country Australia (south-eastern part)
Cover Almost entirely forested
Size 104.9 sq km
Hydrologic description A small catchment with a strong baseflow component

2.2 Input data

Input files .\Tutorials\Murrindindi\pq.dat
.\Tutorials\Murrindindi\t.dat
Streamflow pq.dat, column A
Measurement units = mm
Rainfall pq.dat, column B
Measurement units = ML/day
Temperature t.dat, column A
measurement units = degrees celsius
Length of input data 8453 records
Start and End Dates 8 June 1975 to 30 July 1998

2.3 Synopsis of calibration and simulation

Calibration settings
This table does not include default parameter settings, unchanged by the user.

4
Tutorial 1- Murrindindi River

Calibration Period 1 Days 366 to 1827

Warm up period 200 days
Delay 0 days
Grid search method Instrumental variable
Instrumental variable storages 2 exponential stores in parallel (2,1)
model
Drying rate at reference Start = 7, Finish = 20, Step = 2
temperature (tw)
Temperature dependence of drying Finish = 2.50, Step = 0.2
rate (f)Delay

2.4 Calibration results

R squared 0.855
Bias 9.724

2.5 Simulation results

Rest R squared 0.69

Rest Bias 37.884 mm/year
Individual years R squared Range from 0.85 to -.0.09
Individual years Bias -141 mm/yr

2.6 Discussion
1983 had a rainfall slightly above the average annual value, and followed the driest year on
record. This suggests that, at least in this data set, the dependence of streamflow generation
on antecedent conditions is not quite correct for periods following a drought.
This could be improved by calibrating the parameter ‘moisture threshold for producing flow’
(which was not calibrate in the tutorial), though this would require selecting a calibration
period that includes the years 1982 and 1983.

5
IHACRES Workbook

2.7 Further Reading

The calibrated model is the case study for Croke et al., 2005.
Croke, B.F.W., Andrews, F., Jakeman, A.J., Cuddy, S. and Luddy, A. (2005). Redesign of the
IHACRES rainfall-runoff model, to appear in the proceedings of the 29th
Hydrology and Water Resources Symposium, Engineers Australia, February 2005.

6
Tutorial 2 – Salmon Brook

3 Tutorial 2 – Salmon
Brook
Tutorial 2 introduces you to Salmon Brook, a Western Australian Benchmark catchment. It is
an example of a small ephemeral catchment with no base flow.

3.1 Catchment description

Figure 3-1: Salmon Brook catchment in Western Australia. Modified from figures provided by Richard
Silberstein (CSIRO Land and Water, WA, Australia) and Department of Environment, Water and Catchment
Protection (Australia).

Catchment name Salmon Brook

Location low – relief Darling Range of south-western Western Australia
Country Australia (south-west part)
Cover Dominated by jarrah forest (Eucalyptus marginata)
Size 0.82 sq km
Hydrologic description Small ephemeral catchment with no base flow
Mean annual rainfall 1200 mm
Evapotranspiration About 90% of annual rainfall (Ye et al, 1997)
Soils Surface predominantly highly permeable sands and gravels (Ye
et al, 1997)

7
IHACRES Workbook

3.2 Input data

Input files .\Tutorials\Salmon\salmon.csv
Streamflow salmon.csv, column A
Measurement units = mm
Rainfall salmon.csv, column B
Measurement units = mm
Potential evaporation salmon.csv, column C
measurement units = ET (mm)
Length of input data 4018 records
Start and End Dates 31 December 1973 to 30 December 1984

3.3 Synopsis of calibration and simulation

Calibration settings
This table does not include default parameter settings, unchanged by the user.

Calibration Period 1 1/01/1974 to 31/12/1978 (days 1-1826)

Warm up period None
Delay 0 days
Grid search method Instrumental variable
Instrumental variable storages model Single exponential store (1,0)
Moisture threshold to start flow (l) Start = 0, Finish = 300, Step value = 50

Calibration results

R squared >= 0.7

Bias < 20mm/year

Simulation results
R squared 0.78
Bias -0.25 mm/year

3.4 Step 1 – load the data

Salmon.csv contains the input data:
• Rainfall is in Column A. Its measurement unit is mm.
• Streamflow is in Column B. Its measurement unit is mm/timestep and, as a result, the
area of the catchment is not needed.
• An estimate of potential evaporation is in Column C. Its measurement unit is
mm/time step. Read this into the Temperature variable, with the units selected as ET
(mm).
Set the time period to 1973, December, 31 and default time step of 1 day.

8
Tutorial 2 – Salmon Brook

3.5 Step 2 – check the data

Use Data > Summary and Data > View to check that you have loaded the data correctly
and that it is synchronised.

3.6 Step 3 – define the calibration period(s)

Now, click on the Calibration tab. This time, we will neglect the warm-up period and specify
a calibration period of five years.
Enter the dates 1/01/1974 to 31/12/1978.
Press Add Calibration Period.

3.7 Step 4 – calibrate the linear module

Now, under Calibration, run the Cross Correlation.
You should see that IHACRES calculates the delay as a single time step as this is the peak in
the cross correlation function. However, a value of zero may be appropriate. (Experiment
with this – a delay of zero improves the model performance).
Check the Instrumental Variable box.
Proceed to the Non-Linear Module box.

3.8 Step 5 – calibrate the non-linear module

This tutorial uses the Classic module (only one available at time of writing).
Press Run Grid Search.

Pre-Grid search
The Pre-Grid search options for the five classic module parameters appears. For highly
ephemeral catchments, a threshold may be needed.
Change the search options for the moisture threshold to start flow (l) to Finish value = 300,
Step value = 50.

Instrumental Variable
Check the Single Exponential Store box, as befits a stream with no baseflow.

Post Grid search – find a suitable parameter set

When the grid search finishes, look at the statistics under the Post Grid Search tab. You
might want to sort the data by one or both of the target statistics, Bias and R2. Remember
that optimal values are zero and one respectively.
Look at some plots of the behaviour of R2 and Bias with the parameters, and decide on a
region for the finer grid search.

9
IHACRES Workbook

Refine the parameter set

When you have entered the appropriate maximum and minimum values in the Pre Grid
Search, decide how much finer your search should be in each dimension. Keep in mind that
when the number of grid points (displayed at top right) gets over a few thousand, your search
will be very time-consuming.
The new grid search should give you an appropriate parameter set to simulate over (with an
R2 ≥ 0.7 and a Bias < 20mm/year).
Analyse the model at this point in parameter space.

Accept the parameter set

Accept Calibration.

3.9 Step 6 – simulation

Your simulation will probably look something like this:

The R2 in simulation for this scheme is 0.78 and the Bias = -0.25 mm/year.

3.10 Further Reading

Ye, W., B. C. Bates, N. R. Viney, M. Sivapalan and A. J. Jakeman 1997, "Performance of
conceptual rainfall-runoff models in low-yielding ephemeral catchments", Water
Resources Research, vol 33, pp 153-16.

10
Tutorial 3 – Teifi Catchment

4 Tutorial 3 – Teifi
Catchment
Tutorial 3 introduces you to Teifi catchment in Wales, UK. It is an example of a largish
catchment with strong base flow.

4.1 Catchment description

Catchment name Teifi
Location
Country
Cover
Size 893.6 sq km
Hydrologic description Largish catchment with strong baseflow
Mean annual rainfall
Evapotranspiration
Soils

4.2 Input data

Input files .\Tutorials\Teifi\pq.dat and t.dat
Streamflow pq.dat column A
Measurement units = cumecs
Rainfall pq.dat, column B
Measurement units = mm
Potential evaporation t.dat, column A
Measurement units = degrees celsius
Length of input data
Start and End Dates 1 January 1961 to 30 December 1984

11
IHACRES Workbook

4.3 Synopsis of calibration and simulation

Calibration settings
This table does not include default parameter settings, unchanged by the user.

Calibration Period 1 1 August 1982 to 31 July 1985 (7882 – 8977)

Warm up period None
Delay 0 days
Grid search method Instrumental variable
Instrumental variable storages model 2 exponential stores in parallel

Calibration results

R squared >= 0.86

Bias -14 mm/year

Simulation results
R squared 0.765
Bias 114 mm/year

4.4 Step 1 – load the data

pq.dat and t.dat contain the input data:
• Rainfall is in pq.dat, Column A. Its measurement unit is mm.
• Streamflow is in pq.dat, Column B. Its measurement unit is cumecs.
• Temperature is in t.dat, Column A. Measurement unit is degrees Celsius.
Set the time period to 1961, January, 1 and default time step of 1 day.

4.5 Step 2 – check the data

Use Data > Summary and Data > View to check that you have loaded the data correctly
and that it is synchronised.

4.6 Step 3 – define the calibration period(s)

Now, click on the Calibration tab. This time, we will neglect the warm-up period and specify
a calibration period of three (3) years.
Enter the dates 1 Aug 1982 to 31 Jul 1985 (7882 – 8977).
Press Add Calibration Period.

12
Tutorial 3 – Teifi Catchment

4.7 Step 4 – calibrate the linear module

Now, under Calibration, run the Cross Correlation.
You should see that IHACRES calculates the delay as a single time step as this is the peak in
the cross correlation function.
Check the Instrumental Variable box.
Proceed to the Non-Linear Module box.

4.8 Step 5 – calibrate the non-linear module

This tutorial uses the Classic module (only one available at time of writing).
Press Run Grid Search.

Pre-Grid search
The Pre-Grid search options for the five classic module parameters appears. No changes are
suggested in this exercise.

Instrumental Variable
Check the 2 Exponential Stores in parallel box, as befits a stream with a significant baseflow
signal.

Post Grid search – find a suitable parameter set

When the grid search finishes, look at the statistics under the Post Grid Search tab. You
might want to sort the data by one or both of the target statistics, Bias and R2. Remember
that optimal values are zero and one respectively.
Look at some plots of the behaviour of R2 and Bias with the parameters, and decide on a
region for the finer grid search.
Calibrating with this period and with no warm-up period, you should be able to get an R2 of
≥ 0.86 with a Bias of –14 mm/year.

Accept the parameter set

Accept Calibration.

4.9 Step 6 – simulation

Your simulation will probably look something like this:

13
IHACRES Workbook

The R2 in simulation for this scheme is 0.765 and the Bias = 114 mm/year.

4.10 Further Reading

A calibration for the Teifi catchment using the Classic version of IHACRES is described in
Littlewood et al., 1997.
Littlewood, IG, K. Down, J.R. Parker and D.A. Post. (1997) IHACRES: Catchment-scale
rainfall-streamflow modelling (PC version) Version 1.0 - April 1997. Institute of
Hydrology, Centre for Ecology and Hydrology, Wallingford, Oxon, UK.
http://www.nwl.ac.uk/ih/www/products/mswihacres.html

14