Source hydrology calibration and
dissolved oxygen modelling
Yarra River Application Project
Application Report
Pierotti, S., Carty, R, Woodman, A
29 June 2012
Copyright Notice
© 2012 eWater Ltd
Legal Information
This work is copyright. You are permitted to copy and reproduce the information, in an
unaltered form, for non-commercial use, provided you acknowledge the source as per the
citation guide below. You must not use the information for any other purpose or in any
other manner unless you have obtained the prior written consent of eWater Ltd.
While every precaution has been taken in the preparation of this document, the publisher
and the authors assume no responsibility for errors or omissions, or for damages
resulting from the use of information contained in this document. In no event shall the
publisher and the author be liable for any loss of profit or any other commercial damage
caused or alleged to have been caused directly or indirectly by this document.
Citing this document
Pierotti, S., Carty, R, Woodman, A (Melbourne Water, Australia and eWater Cooperative Research
Centre 2011) Source hydrology calibration and dissolved oxygen modelling; Victoria – Yarra River
Application Project
Simon Pierotti (Melbourne Water)
Richard Carty (Melbourne Water)
Amanda Woodman (Sinclair Knight Merz)
Publication date: June 2012 (Version 1)
ISBN: 978-1-921543-64-7
Acknowledgments
eWater CRC acknowledges and thanks Melbourne Water and partners to the CRC and
individuals who have contributed to the research and development of this publication.
The authors also wish to acknowledge David Waters (Department of Environment and
Natural Resources, Queensland) and Rhys Coleman (Melbourne Water) for reviewing
and editing the report
For more information:
Simon Pierotti
Melbourne Water
100 Wellington Parade
East Melbourne, VIC 3002
E: [email protected]
OR
Innovation Centre, Building 22
University Drive South
Bruce, ACT, 2617, Australia
T: 1300 5 WATER (1300 592 937)
T: +61 2 6201 5834 (outside Australia)
E: [email protected]
www.ewater.com.au
ii
Source hydrology calibration and dissolved oxygen modelling
Disclaimer
The following disclaimer relates to all Melbourne Water maps and data used in the preparation
of the Source model and/or this report.
Whilst all due skill and attention has been taken in collecting, validating and providing the attached
data, Melbourne Water shall not be liable in any way for loss of any kind including damages, costs,
interest, loss of profits or special loss or damage, arising from any error, inaccuracy, incompleteness or
other defect in this information.
In utilising this information the recipient acknowledges that Melbourne Water makes no representations
as to the accuracy or completeness of this information and the recipient ought to carry out its own
investigations if appropriate.
Source hydrology calibration and dissolved oxygen modelling
iii
Contents
1 Executive Summary .................................................... 1 2 Introduction ................................................................. 3 2.1 Background ........................................................................................................................... 3 Catchment Description.......................................................................................................... 3 Victoria in Drought ................................................................................................................ 4 Dissolved Oxygen and River Flow ........................................................................................ 5 2.2 Project Objectives ................................................................................................................. 6 2.3 Description of the Model ....................................................................................................... 6 Source................................................................................................................................... 6 QUASAR ............................................................................................................................... 6 3 3.1 Methods ...................................................................... 8 Model Inputs ......................................................................................................................... 8 Sub-catchments .................................................................................................................... 8 Rainfall and Potential Evapotranspiration (PET)................................................................... 9 Land use and Pervious Fraction Determination .................................................................... 9 Storages and Releases....................................................................................................... 10 Diversions ........................................................................................................................... 11 Yering Gorge Pumping Station ........................................................................................... 12 Determination of Hydrological Parameters ......................................................................... 13 3.2 Hydrology Calibration and Validation .................................................................................. 16 3.3 Dissolved Oxygen Modelling in QUASAR........................................................................... 16 Model Inputs ....................................................................................................................... 16 Quasar Calibration .............................................................................................................. 17 3.4 Scenario Testing ................................................................................................................. 17 Altered Flow Regimes ......................................................................................................... 18 Climate Change .................................................................................................................. 18 4 Results and Discussion ............................................. 19 4.1 Hydrology Calibration Results............................................................................................. 19 4.2 Quality of Observed and Predicted Data ............................................................................ 22 4.3 Scenario Testing ................................................................................................................. 24 5 Conclusions............................................................... 27 6 References ................................................................ 28 iv
Source hydrology calibration and dissolved oxygen modelling
Source hydrology calibration and dissolved oxygen modelling
v
1
Executive Summary
The Yarra River catchment encompasses an area of some 4,000km2 and provides habitat for
various aquatic species including Macquarie Perch, Murray Cod, and various frogs, birds
and eels. The catchment is also highly regulated with a number of irrigation diversions and
extensive water harvesting to supply the city of Melbourne.
The state of Victoria has experienced a severe drought for over 10 years. The resultant
effects of the drought have led to an increase in the competing demands of harvesting
sufficient water to supply Melbourne while maintaining enough flow to ensure the ecological
health of the river. During the drought the Victorian government explored ways to reduce the
environmental flows to the river by 10 GL/annum. However concerns were raised regarding
what impact the lower flows would have upon dissolved oxygen (DO) levels within the mid to
lower Yarra. Low DO levels in the river system can have a serious ecological impact upon
the various aquatic organisms that rely on oxygen to breathe. The ongoing drought has also
caused a gradual decline in both the overall total flow and the number of high flow events
within the river also impacting on river DO levels. Modelling provides a means of exploring a
range of alternative flow regimes and their subsequent impact on DO.
Through the eWater CRC’s Applications projects, the eWater Source water quantity and
quality model was applied by the Yarra River Application Project team to generate hydrologic
time series under current and future flow management scenarios. The modelled output was
then used as an input into the QUASAR DO model applied by the projects ecological team to
test the ecological impacts of various current and future flow management scenarios.
Source model was successfully calibrated and validated. Key findings from the hydrology
calibration and validation process include:
• The runoff generated by Source generally represented the system well, with predicted
(modelled) flow within ±5% of the observed runoff for gauges used in DO modelling.
• The Nash Sutcliffe objective function (E) was used to assess modelled and observed
flow estimates. Daily E values ranged from 0.4 – 0.6 for the calibration gauges with
the validation gauges having daily E values ranging from 0.54 – 0.69. Monthly E
values were acceptable ranging from 0.51- 0.89 and 0.87- 0.89 for calibration and
validation gauges respectively. Percentage volume differences for the validation
gauges were within 5% of observed flows
• Poor quality gauging station data, particularly in low flow periods for many of the
tributaries meant that calibration was difficult. Manual adjustments to calibration
parameters were required to match up peak flows
• The modelling of the pumping regime at the Yering pump station proved difficult due to
the variable nature of the operation of these pumps and generally resulted in
overestimation of extractions from the Yarra River when compared to observed
pumping amounts.
1
Source hydrology calibration and dissolved oxygen modelling
Five scenarios were tested including the base case which represented current conditions.
Three altered flow regime scenarios and two climate change scenarios were tested. The
hydrological outputs from Source were used in the DO modelling. The scenarios included;
• No diversions and No Yering Pumping
• Diversions and No Yering Pumping
• Increased Environmental Flow
• Climate Change 2030
• Climate Change 2060
The Climate Change scenarios significantly reduced flows in the lower reaches of the Yarra
River. It was also evident that diversions only play a minor role in reducing flows in the lower
Yarra River, the major contributor being the Yering Gorge pumping station extractions.
DO modelling was undertaken for two reaches of the Yarra River – Chandler Highway and
Banksia Street. The Quasar water quality modelling package was used for this task. The five
scenario flow inputs were modelled over four separate years to represent current conditions
(2008), dry conditions (2006), average conditions (2004) and wet conditions (1996).
Measured daily DO data at the Chandler Highway site matched well to predicted (modelled)
DO for Base scenario flows. Only monthly DO data was available for the Banksia Street
reach, and this matched well to modelled data. Results show that climate (dry versus
average and wet years) has more influence on DO in the Yarra rather than specific altered
flow scenarios. Climate change is predicted to reduce flow in the river, which results in more
frequent and longer periods of low DO.
The development of this base Source model for the Yarra River will provide a simulation tool
which will enable stakeholders of the Yarra catchment to assess a range of current and
future scenarios of interest. Further work may need to be carried out to fine tune the DO
modelling as well as improving the calibration of Source to low flows to help assist with
decision making. The modelling of the Yering extraction also needs to be refined in future
versions of the model.
Source hydrology calibration and dissolved oxygen modelling
2
2
Introduction
2.1
Background
Catchment Description
The Yarra River catchment is located to the north east of Melbourne and encompasses an
area of some 4,000 km2 (Figure 1). The average annual rainfall across the catchment
ranges from 680mm in Burnley (near Melbourne) to 1,080mm in the river headwaters around
the Upper Yarra Reservoir (above Warburton) (MWC and PPWPCMA 2004a).
•
Melbourne
Yering
Gorge
Figure 1 -Yarra River Catchment indicating monitoring sites, reservoirs, streams and broad land use classes.
3
Source hydrology calibration and dissolved oxygen modelling
Between the headwaters and mouth of the river, the Yarra passes through many different
classes of land use. The top of the catchment is located on the southern slopes of the Great
Dividing Range within the Yarra Ranges National Park. This area contains steep forested
slopes that have been protected for more than 100 years due to its extensive use for
Melbourne’s water supply harvesting. Further downstream, the mid Yarra flood plains have
been largely cleared and developed for agricultural purposes, while in the lower reaches, the
Yarra passes through heavily urbanised environments. Overall 50% of the catchment is used
for agriculture, 22% is urbanised, and 21% is forested.
In addition to direct water harvesting through the placement of water supply reservoirs at a
number of locations throughout the catchment, water is extracted from the various rivers and
creeks via weirs. A large pumping station is located in the mid Yarra (Yering Gorge Pumping
Station), which can pump to a maximum of 1,000 ML/d. There are also numerous farm dams
and diversions throughout the catchment making the Yarra River catchment a highly
regulated system (MWC and PPWCMA 2004a).
Victoria in Drought
For over 10 years, the state of Victoria has been experiencing a severe drought with rainfall
significantly lower than the long term average (DSE 2007). This has led to an increase in the
competing demands of supplying sufficient water to Melbourne while maintaining appropriate
environmental flows. As a result, investigations were carried out to determine how
environmental flows could be reduced by 10 GL/annum while still maintaining an appropriate
level of ecological health in the river (MWC 2007 and DSE 2007).
The outcome of the investigation recommended that the environmental flows downstream of
the Yering Gorge Pumping Station (as measured by the downstream gauge at Warrandyte)
could be reduced from 245 ML/d to 200 ML/d if the drought continued. This meant that
pumping at Yering Gorge was required to cease once flows reached 200 ML/d at
Warrandyte (or 150 ML/d at the gauge further downstream at Chandler HWY) Figure 1. A
flow reduction to 200ML/d was however subject to some additional restrictions in order to
mitigate the ecological risk to the river. If stream flows at the gauging station upstream of the
pumping station (at Yarra Glen) reached 1,500 ML/d during April-May, or 2,000 ML/d JuneSeptember then this became the minimum environmental flow downstream of the pumping
station for 7 days. Therefore, only flows higher than this could be extracted during those
periods. It is important to note however that the restriction could only be activated if the flow
upstream actually reaches the trigger points (flows), only once per calendar year, and only
need be applied if the event did not occur in the previous calendar year.
The Victorian Government adopted these reduced environmental flow targets with the
requirement of a yearly review on the ecological impact on the river. The level of water
restrictions in Melbourne has since eased due to decent rain, and the original 245ML/d flow
entitlements has been re-established. Future drought conditions may mean that these
reductions are re-imposed.
Modelling scenarios will attempt to address the implications of some of these reduced or
increased environmental flow targets on river flow and dissolved oxygen in the lower Yarra
River.
Source hydrology calibration and dissolved oxygen modelling
4
Dissolved Oxygen and River Flow
A diverse range of species including Macquarie Perch, Murray Cod, eels, frogs and bird life
are supported by the River. Concerns have recently been raised regarding the ecological
health of the river system after consistently observing low dissolved oxygen (DO)
concentrations in the mid and lower reaches of the river over the past few years (eg. Figure
2). Low DO concentrations in the river can have serious ecological implications for aquatic
life that rely on oxygen to breathe (eWater 2009).
Chandler HWY Dissolved Oxygen (Minimum Values) Time Series
16
180,000
14
160,000
DO Concentration (mg/l)
120,000
10
100,000
8
80,000
6
60,000
4
Monthly Flow (ML/Month)
140,000
12
40,000
2
20,000
0
1/08/1998
14/12/1999
27/04/2001
9/09/2002
22/01/2004
5/06/2005
18/10/2006
1/03/2008
0
14/07/2009
Date
DO Min
Flow
Figure 2 - Decreasing trend in minimum Dissolved Oxygen concentration over a 10 year period. Plotted against
monthly flow at Kew (Chandler HWY Gauge)
There has been a decreasing trend of DO minimum concentrations between 1998-2009
(Figure 2). Additionally, there are indications of a significant increase in the variability of DO
concentrations over the past 5 years along with some very low readings of DO (eg. 0.51
mg/L during March 2007). There has also been a consistent decrease in the monthly flow at
this site and a noticeable drop in the number of high flow events.
The concerns around the low dissolved oxygen concentrations during these drought periods
has led to the need to assess the impacts different flow regimes have upon the level of DO in
the river, and to help reduce their impact on aquatic fauna. Previous catchment modelling on
the Yarra River has not looked at DO levels or concentrated on low to medium sized flow
events. The Ports E2 (BMT WBM, 2008) model was developed as a decision support system
(DSS) for the development of the Better Bays and Waterways Plan (BB&W) – Water Quality
Improvement Plan (WQIP) for Port Phillip and Western Port catchments and bays. The main
focus of Ports E2 was to model gross pollutant loads in the bays and was calibrated with this
in mind and made no effort to model DO. The focus of the Yarra River modelling in this
project was on calibrating low to mid-range daily flows a key component to DO levels in
streams. Modelling offers a way of assessing the impacts of various flow regimes on DO.
5
Source hydrology calibration and dissolved oxygen modelling
2.2
Project Objectives
The Yarra River eWater application project is a partnership project between Melbourne
Water, Sinclair Knight Merz (SKM), Monash University, EPA Victoria, Department of
Sustainability and Environment (DSE), Southern Rural Water and eWater.
The objectives of the application project were to;
• Build on previous catchment modelling work utilising the Source modelling software to
a real world project, in this case generating hydrological time series inputs for DO
modelling in the Yarra River in a regulated catchment.
• From field measurements determine correlations between flow, DO and temperature
and how native fish (specifically Macquarie Perch) respond to changing DO conditions
• Correlations may then be used to try and establish trigger levels, which could be used
to trigger management actions, such as refreshing water in poor DO zones by
augmenting river flows
This report focuses on the model development in Source, hydrological calibration and
generation of hydrological time series of various scenarios to be tested in the DO modelling.
The development of a Quasar model to model DO is also documented in this report along
with the DO modelling results.
2.3
Description of the Model
Source
Source is a water quality and quantity modelling framework that supports decision-making
and a whole-of-catchment management approach. The model structure operates as a nodelink network and gives access to a collection of models, data and knowledge that simulate
the effects of climatic characteristics (like rainfall and evaporation) and catchment
characteristics (like land-use or vegetation cover) on runoff and contaminant loads from
unregulated catchments. Source can operate at a daily time step and can be used to predict
the flow and load of constituents at any location in the catchment over time (Delgado et al
2011).
The model development of the Yarra River Source Model including input data used is
described in the following section.
QUASAR
QUASAR (QUAlity Simulation Along River systems) is a model that describes the timevarying (i.e. dynamic) transport and transformation of solutes, including dissolved oxygen, in
branched river systems using 1D ordinary, lumped parameter differential equations of mass
conservation (Cox 2003). QUASAR is capable of modelling DO on large branched river
systems with multiple influences such as effluent discharges, abstractions and weirs. Simple
flow and load addition equations are used at the top of each reach (i.e. in the first
computational element) for all flows entering (or being abstracted from) that reach and so all
influences such as discharges and abstractions are considered to enter at the beginning of
the reach. The solution of the mass-balance equations for flow and the determinants being
simulated is then made in each element and the results of these calculations are used as
Source hydrology calibration and dissolved oxygen modelling
6
input for the next element. At the end of a reach the results are stored and then used as the
upstream influence on the next reach. In dynamic mode, this is performed once per timestep (Cox 2003). The basic hydraulic model used in Quasar reduces the data requirements
and also simplifies the process of calibration and results in relatively quick runtimes.
However, it has a temporal program constraint in which the maximum duration of a run is
one year. QUASAR is not a current commercial product but is available free of charge. The
model is not supported apart from the provision of a user manual, and it will not be further
developed.
7
Source hydrology calibration and dissolved oxygen modelling
3
Methods
The most recent available data was used to construct the Source model. A modelling period
of 32 years was chosen (June 1978-2009) to represent both wet and dry periods. This
section provides a description of the input data sets and spatial layers used in the model and
how they were generated. An outline of the verification and validation of the model are also
detailed and a description of the scenarios tested in the model to be implemented into the
DO modelling.
3.1
Model Inputs
Sub-catchments
Source has the capability of internally delineating sub-catchment boundaries based on an
input Digital Elevation Model (DEM). For this study however, the sub-catchments were
created in a GIS environment prior to input into the model, to gain better user control of
where boundaries should be located, particularly with respect to stream gauge locations.
Sub-catchments were derived manually from an existing GIS layer. The sub-catchment
boundaries in the model were up scaled from a GIS Drainage layer (DR_MWC_Catchment)
that represents the tributaries of major waterways captured using the drainage scheme and
hydraulic model information. The sub-catchments were segmented to coincide with
streamflow gauges in the Yarra catchment and also to represent storages in the catchment.
Figure 3 shows the final sub-catchment map and node link network that was used in
modelling the Yarra River Catchment. Note the modelling area for this project only captures
the Yarra Catchment above the Chandler Highway Gauge on the Yarra River (229143).
N
Figure 3 – Sub-catchments used in Source and location of gauges (orange dots) used to delineate sub-catchments
Source hydrology calibration and dissolved oxygen modelling
8
Rainfall and Potential Evapotranspiration (PET)
Data for rainfall and potential evapotranspiration (PET) was sourced from the SILO Data Drill
database. Grids of interpolated data are accessed by the SILO Data Drill from the Bureau of
Meteorology’s (BOM) station records. Interpolation of the rainfall and PET data occurs on a
5km grid and is provided as individual daily grid sets. When imported into Source each daily
grid is interrogated to produce a single time series data set for each sub-catchment (SILO
2009 and Waters 2008). The time series for each calibrated sub-catchment was then
extracted and used for hydrologic calibration external to Source.
Land use and Pervious Fraction Determination
The focus of this Source model development was on the hydrology. However a range of land
use categories were included to enable future users have the option to model constituents.
The PortsE2 (WBM 2008) land use map and codes (Table 1) were used for this project for
consistency and enabled the use of the PortsE2 parameter sets for certain sub-catchments.
In total, 16 land use categories were created. These are summarised in Table 1 and can be
seen spatially in Figure 4.
Table 1 - Land Use Classification - Based upon BMT WBM Report (2008)
9
Classification
% of Total Catchment
Annual Horticulture
1.2
Forest
26
Pasture Cropping - Irrigated
0.02
Pasture Cropping – Non Irrigated
33
Perennial Horticulture
1.7
Plantation Forest
17
Rural Green Space
0.35
Rural Industrial
0.26
Rural Roads
1.9
Rural Township
1.1
Urban
12
Urban Commercial
0.95
Urban Green Space
1.6
Urban Industrial
0.98
Urban Roads
1.0
Water
0.71
Source hydrology calibration and dissolved oxygen modelling
Figure 4 - Land Use Classification Map
The land use classification was used to determine the pervious fraction of each subcatchment. From Table 1, urban, urban commercial, urban industrial and urban roads were
all classified as predominantly impervious surfaces. A pervious fraction of 0.7 was applied to
these four urban areas as per PortsE2 report, Argent (2006). .
Storages and Releases
There are a number of water storages in the Yarra Catchment which impact on catchment
flows. Source has the ability to model these storages via a Storage link model. Models must
use a set of parameters that describe initial conditions and the release curves of the
storages.
The Yarra catchment contains 6 major reservoirs. One of the six storages were included in
the model as per Argent (2006). Table 2 describes how each of the storages were
represented in the model.
Source hydrology calibration and dissolved oxygen modelling
10
Table 2 - Representation of major reservoirs
Storage
Representation
Upper Yarra
Not Modelled, rarely spills. Used an observed
flow node to represent environmental flow release
of 10ML/day
O’Shannassy
Not Modelled. Used an observed flow node to
represent environmental flow release of 8ML/day
Maroondah
Storage link model calibrated to represent typical
behaviour via an appropriate release curve –
1ML/day release, with additional spills in
moderate to wet years
Silvan
Not Modelled, seasonal storage for water from
major harvesting sites. Observed flow of 2ML/day
to represent environmental flow releases
Sugarloaf
Not Modelled. Off-stream storage that harvests
water from the Yarra River via the Yering Gorge
Pumping Station
Yan Yean
Not Modelled. Off-stream storage. Observed flow
node of zero representing current average
releases
Diversions
The Yarra River catchment contains approximately 1480 diversion licences and 560
registered farm dams; this includes some licences that have been discontinued prior to 2009.
Total annual licensed irrigation diversion volume data was available for the majority of
reaches.
A similar method as applied in Argent (2006) for the PortsE2 model was used to account for
the diversion. Diversions were represented by a number of loss nodes on the main stem of
the Yarra River, with losses set for flows above 200ML/day to reproduce the licensed
diversion volumes on an annual basis (Table 3). Diversions also occur on three major
tributaries, Armstrong Creek, Starvation Creek, and McMahons Creek. These diversions
were lumped together and modelled via loss nodes in a similar way to the irrigator
diversions. Losses were aggregated above certain gauges to facilitate exploration of
changes to diversion conditions under future scenarios.
Table 3 - Average Annual Diversions
11
Node in Sub-catchment
Mean Annual Flow
(ML)(Gauge)
Diversion
(ML)
Diversion
(%)
Yarra River@ Healesville
149,860 (#229212)
8,160
5.5
Yarra River@ Yering
364,400 (#229653)
8,360
2.3
Yarra River@ Warrandyte
440,000 (#229200)
3,500
0.8
Yarra River@ Chandler Hwy
471,700 (#229143)
9,000
1.9
Yarra Tributaries
46,881
28,765
61.4
Source hydrology calibration and dissolved oxygen modelling
Yering Gorge Pumping Station
One of the major extractions in the Yarra River is the Yering Gorge pumping station. Four
(250ML/day maximum) variable pumps are able to extract water from the main stretch of the
Yarra River into Sugarloaf Reservoir. Two 70 ML/day low flow pumps can also extract water.
Operators use their discretion on when to operate these pumps depending on;
• Environmental flow constraints downstream of the pumps
• Storage in Sugarloaf Reservoir
• Demands in the system
Currently Source is unable to model the varying pumping constraints of the Yering Gorge
extraction. Therefore the pumps were represented in a similar fashion to the diversions as
described in the section above. A loss curve (Figure 5) was calibrated by analysing observed
extraction amounts with respect to upstream flow at Yering Gorge and then implementing a
trial and error method in obtaining reasonable extractions compared to that of the observed
extractions. This method should be considered preliminary and it is recommended that it be
improved upon in future model builds.
1,250
Extracted Volume (ML/day)
1,000
750
Loss Curve
500
250
0
0
250
500
750
1,000
1,250
1,500
1,750
2,000
Upstream Flow (ML/day)
Figure 5 - Yering Gorge extracted flow analysis and model loss curve
Source hydrology calibration and dissolved oxygen modelling
12
Determination of Hydrological Parameters
Source offers several rainfall runoff models for calibration. The SIMHYD rainfall runoff model
was used as per Argent (2006). SIMHYD uses daily rainfall and aerial PET data to estimate
daily stream flow (CRC for Catchment Hydrology 2004). SIMHYD was selected for the Yarra
River Application project due to its simplicity and because it was used for similar modelling
projects in the region.
Initially the Yarra Application project model concentrated on calibrating the lower reaches of
the Yarra River using an observed upstream inflow node, however this was unsuitable to run
scenarios. Due to time constraints the upper reaches (above Coldstream 229653) of the
Yarra have been parameterised using existing SIMHYD parameters from the Yarra
Catchment in the Ports E2 model (WBM 2008). This allows scenarios to be tested in the
model while improving on the Ports E2 calibration with updated parameters for the more
urbanised regions in the lower Yarra River catchment.
For the lower Yarra reaches, a set of SIMHYD hydrologic parameters were derived upstream
of each gauging station. Nine parameter sets were derived for these catchments (Appendix
A). For ungauged catchments, either one of the nine parameter sets, or the Ports E2
parameter sets were assigned based on a comparison of attributes between the gauged and
ungauged catchments. Attributes considered were location, pervious fraction, land use type,
and rainfall. Figure 6 shows the sub-catchments that incorporated the Ports E2 parameter
set.
To derive these parameters each sub-catchment must be calibrated against observed data.
Gauging station data was used to do this. Observed flow data was extracted from Melbourne
Water’s Hydstra database from 1st January 1970 – 30th June 2009. The actual length of
historical records available and the quality of the data was highly variable depending upon
the station. The site with the longest record was located at Warrandyte (229200) in the mid
Yarra with <1% of the data missing. Missing data was then infilled via correlation to an
upstream or downstream gauge to create a full record. While some stations had less than 10
years data, most stations had data beginning in the 1970s through to the present day. After
extraction from the database, the data was assessed to determine data quality. The criteria
used to determine data suitability was;
• Consultation with Melbourne Water hydrographers to assess data quality, data gaps
and the quality of the rating curves for the gauging stations
• At least 10 years of streamflow data
• Visual assessment of rainfall and runoff data to identify any obvious problems
The entire Yarra Catchment contains over 60 gauges, with 34 flow gauging stations
assessed. Streamflow data from 13 gauges below Coldstream were suitable for calibrating
the hydrological model (Table 4), with nine of these being used in the SIMHYD
parameterisation. Marked station locations are shown in Figure 6.
13
Source hydrology calibration and dissolved oxygen modelling
Table 4 – Details of the 13 Gauging Stations used for model calibration and validation
Gauge
ID
Location
Catchment
Area (km2)
Date
Commenced
Years
Record
% Data
Missing
Resolution
(mm)
Darebin
Creek
229612
Bundoora
93.7
1977
32
9.22
10
Diamond
Creek
229619
Hurstbridge
256.74
1978
31
3.84
5
Diamond
Creek
229618
Eltham
321
1977
32
6.05
10
Koonung
Creek
229229
Bulleen
31.7
1997
12
5.2
10
Mullum
Mullum
Creek
229648
Doncaster
East
34.6
1980
29
2.4
10
Watsons
Creek
229608
Kangaroo
Ground Sth
73.2
1980
29
7.1
10
Olinda
Creek
229690
Mt Evelyn
27.2
1987
22
2.01
5
Stringybark
Creek
229401
Mt Evelyn
9
1999
10
6.4
5
Yarra River
229206
Yarra Glen
1963
1991
18
0
10
Stream
#
Yarra River
229200
Warrandyte
2447
1970
39
0
5
Yarra River
229142#
Fitzsimons
Ln
2780
1975
34
2.6
10
Yarra River
229135#
Banksia St
3240
1975
34
6.9
10
Yarra River
229143#
Chandler
Hwy
3420
1975
34
1.6
10
# Validation Gauges
For the selected gauging stations the Rainfall Runoff Library (RRL) software was used to
optimise the parameter values for the SIMHYD model. RRL is a calibration tool containing
several different rainfall runoff models, calibration optimisers and display tools to assist with
model calibration (CRC for Catchment Hydrology 2004). Rainfall and PET files were
extracted from the Source model and calibrated against the observed flow for each gauging
station. The calibration period depended on the amount and quality of data from each
gauging station. Initially, the optimisation methods included in the RRL package were used
to obtain a preliminary best fit. Some parameters such as pervious fraction were then fixed at
a pre-determined value or altered manually and fixed before running the optimisation. The
most reliable method during the optimisation process was the Rosenbrock Single Start using
starting parameters taken from the PortsE2 report (Argent 2006). Some manual fitting of
parameters was also implemented to obtain suitable parameters for use in the model. The
final adopted parameter sets are included in Appendix A.
Gauges 229618 and 229206 were downstream of headwater gauges and therefore a
residual calibration was undertaken whereby a set of hydrologic parameters was derived for
the area between the two gauges. An observed flow series for residual calibration was
estimated by subtracting the upstream flow from the downstream gauge flow
Source hydrology calibration and dissolved oxygen modelling
14
Figure 6 - Location of Gauges used in Calibration and parameter sets used
15
Source hydrology calibration and dissolved oxygen modelling
3.2
Hydrology Calibration and Validation
The primary objective function used to optimise the SIMHYD model parameters was the
coefficient of efficiency (E) or Nash-Sutcliffe Criterion (Nash and Sutcliffe 1970). It provides a
measure of the ability of the model to reproduce recorded flows, with a value of E=1.0 to
indicate that all the estimated flows are the same as the recorded flows (Chiew and Scanlon,
2002). The secondary objective functions used were either runoff difference, percentage and
flow duration curve depending upon which gave the best fit to the data. The best visual fit did
not always seem to be the best calibration for a number of gauges due to poor resolution in
the low flows of many of the gauges, therefore some manual changes were used to match
up the low to medium sized peak events.
The four gauges in the lower Yarra River in the project target area (229200, 229142,
229135, and 229143), were then used to validate the model within Source. The validation
period was between 1981 and 2009. This enabled the ungauged sub-catchments to be
validated when gauged sub-catchment parameters were assigned to them and also
indicated how well the Ports E2 parameter sets used in the upper Yarra Reaches worked in
the model. Two of these validation gauges were used as the sites for the hydrological time
series outputs for the DO modelling and scenario testing. Given the assumption that low DO
concentrations are at least in part caused by low flows, the calibration of this model will be
focussing on low river flows and small to medium runoff events. The results can be seen in
Appendix B.
3.3
Dissolved Oxygen Modelling in QUASAR
The Quasar software was applied to model Dissolved Oxygen (DO) within two reaches of the
Yarra River. These were;
• Chandler Reach (from Chandler Highway, Kew to Dights Falls, Abbotsford); and
• Banksia Reach (from Fitzsimons Lane, Templestowe to Banksia Street, Heidelberg)
This section outlines the inputs applied and calibration approach for the Quasar modelling.
Model Inputs
Reach Parameters
The channel is defined by a set of spatial parameters that are entered into the model. These
define the channel length, width and most upstream location of the reach. These spatial
parameters were obtained from Google Earth. The channel width information was taken from
numerous (>30) width measurements that included runs and bends that were taken where
the edge of the banks were relatively obvious for measurement. The mean channel depth
was obtained from a longitudinal depth survey undertaken for this project.
The flow-velocity relationship was based on cross-section data supplied by Melbourne
Water.
Monthly algae values are entered as chlorophyll-a concentration in micrograms/litre for a
month. This information was obtained from monitored data from 1985-1990 and was the only
available dataset for this parameter.
Source hydrology calibration and dissolved oxygen modelling
16
Rate coefficients were entered as an initial set of values based on previously established set
in Quasar; these are discussed further in Section 2.4.2.
Determinands
Flow and water quality data were entered into Quasar as reach determinand data.
Flow data was obtained from the Source modelling for Chandler Highway and Fitzsimons
Lane. Base case conditions were provided in addition to five scenarios (see Table 5).
Water quality inputs were obtained from monitored water quality data for Chandler Highway
and Warrandyte for input into the Chandler and Banksia reaches respectively. This
contained readings of frequencies varying between weekly and monthly. For data and years
where only one spot reading was taken for the month, that value was applied. When more
than one spot reading was taken during the month, the average of the readings was taken.
Quasar Calibration
To calibrate the Chandler model, the rate coefficient parameters were adjusted to match the
modelled data to daily measured data. Suitable values of these reaction rate coefficients are
not widely published, however two research papers were obtained (Eatherall, et al 1998 and
Lewis, et al 1997). These studies stated that the process of the calibration is important but
that if site specific values for various rate coefficients cannot be obtained (which is the case
for the Yarra River due to lack of suitable data) then calibration is best achieved by adjusting
rate coefficients so that modelled data is matched to existing data. There was no measured
daily data available over the modelling periods for the Banksia reach, therefore the rate
coefficients applied for the Chandler reach were considered appropriate for the Banksia
reach. This is consistent with approaches applied by Eatherall, et al (1998).
3.4
Scenario Testing
One of the aims of the Yarra River application project was to test various flow scenarios in a
DO model. Two types of scenarios were modelled to assess the impact on DO in the lower
Yarra River, these were altered flow regimes, and climate change. A description of each
scenario are shown in Table 5 below.
Table 5 - List of scenarios tested
17
Scenario
Description
1
Base Case (current conditions)
Calibrated Model
2a
No Diversions and No Yering Pumping
No Yering Pumping, No diversions
2b
Diversion and No Yering Pumping
No Yering Pumping, but with full diversions
2c
Increase min Enviro Flow
Minimum flow upstream of Yering before pumping
starts set to 500 ML/d
3a
Climate Change 2030
Rainfall and PET from CSIRO GCM model
3b
Climate Change 2060
Rainfall and PET from CSIRO GCM model
Source hydrology calibration and dissolved oxygen modelling
For each scenario listed above, daily time series outputs from Source were provided for an
average (2004), dry (2006), and wet (1996) year to be used as inputs into QUASAR. The
Chandler reach applied the modelled streamflow output from 229143, Chandler Hwy) and
the Banksia reach output from 229142, Fitzsimons Ln).
Altered Flow Regimes
Three scenarios were modelled to simulate an altered flow regime in the Yarra River
catchment. Scenario 2a and 2b were tested by removing the appropriate Loss Node models
which simulate irrigation diversions, water harvesting diversions, and the extraction of water
from the Yarra River at the Yering Gorge Pumping station. Scenario 2a turns off all
diversions in the Catchment including the Yering Gorge pumps. Scenario 2b only turns off
the Yering Pumps, while still allowing the tributary diversions for water harvesting and
irrigation diversions in the system. The impacts of these changes are then assessed against
the base case calibration model reflecting current conditions. Scenario 2c looks at increasing
the minimum flow requirement upstream of Yering before pumping can occur (Figure 7). In
the calibration run it is set at 250ML/day, and for Scenario 2c it is set to 500ML/day.
1,000
Extracted Volume (ML/day)
Base Case (1)
Scenario 2c
750
500
250
0
0
250
500
750
1,000
1,250
1,500
1,750
2,000
Upstream Flow (ML/day)
Figure 7 - Yering pumping loss curves for the Base Case and increased environmental flow scenario (2c)
Climate Change
Climate change scenarios were modelled by altering the input rainfall and PET grids in
Source. The input data came from the CSIRO/SEACI derived climate change projection
data. The projections used are for 1 degree of global warming (2030 projection) and 2
degrees of global warming (2060 projection). 15 Global Climate Models (GCMs) are included
in these projections. To simplify the modelling a GCM of “median” runoff was used. To
identify which GCM rainfall data to use, the mean annual runoff output for each Global
Warming (GW) scenario was ranked. The selected GCM model producing median runoff (i.e.
ranked 8th out of the 15 GCMs) of which the CSIRO/SEACI derived climate change
projection data (rainfall and APET) was the CCCMA-t47 model (Canadian Centre for Climate
Modelling and Analysis) - the same GCM model ranked median for both 2030 (1.0 deg C
warming) and 2060 (2.0 deg C) under the MEDIUM global warming scenario.
Source hydrology calibration and dissolved oxygen modelling
18
4
Results and Discussion
4.1
Hydrology Calibration Results
Table 6 summarises the hydrology calibration results at each of the suitable gauging stations
calibrated in RRL, as well as the validation results for the four sites in the Lower Yarra River.
All predicted flows come from the Source Catchment model outputs.
Table 6 - Calibration and Validation Results
Stream
Gauge ID
Location
Nash Sutcliffe
Criteria
(Edaily)
Nash Sutcliffe
Criteria
(Emonthly)
Predicted/
Observed
Darebin Creek
229612
Bundoora
0.6
0.72
0.85
Diamond Creek
229619
Hurstbridge
0.57
0.65
0.93
Diamond Creek
229618
Eltham
0.61
0.7
1.06
Koonung Creek
229229
Bulleen
0.51
0.51
0.63
Mullum Mullum
Creek
229648
Doncaster East
0.45
0.87
0.96
Watsons Creek
229608
Kangaroo Ground
Sth
0.4
0.62
0.73
Olinda Creek
229690
Mt Evelyn
0.56
0.62
0.92
Stringybark
Creek
229401
Mt Evelyn
0.69
0.75
0.84
Yarra River
229206
Yarra River
Yarra Glen
0.48
0.89
1.09
#
Warrandyte
0.56
0.88
0.97
#
229200
Yarra River
229142
Fitzsimons Ln
0.54
0.87
0.95
Yarra River
229135#
Banksia St
0.65
0.89
0.95
Yarra River
229143#
Chandler Hwy
0.69
0.89
0.99
# Validation Gauges
Total predicted streamflow volumes were within 10% of observed flow volumes for 5 of the 9
calibration gauges, and within 5% for the sites used in the DO modelling. Two of the
calibration gauges were within 20% and two within 40% of observed volumes, this was
deemed reasonable as gauge resolution and data quality at these four sites was poor. The
calibration of these sites focused on improving medium event flow predictions and not on
maximising the E value. Coefficient of efficiency values (E), ranged from 0.4 – 0.69 for daily
(Edaily) predicted and observed values. The monthly correlation (Emonthly) was acceptable and
ranged from 0.51 - 0.89. Runoff estimates in the focus area of the catchment were generally
considered within an acceptable range for the modelling objective looking at relative change.
19
Source hydrology calibration and dissolved oxygen modelling
An initial comparison of daily predicted and observed flows highlighted that there was a flow
lag to be accounted for in the runoff routing through main stretch of the Yarra River. A one
day lag was applied upstream of the Yering Gorge pumping station. This significantly
improved the daily predicted and observed flow comparisons. The lag was applied upstream
of the pumping station as it was noted that flows after the pumping station were dropping
below the environmental flow requirement. When the lag was applied most of these
instances disappeared and the timing of runoff peaks were better aligned.
Calibration results can also be represented via scatter plots. Scatter plots showing predicted
and observed daily streamflow for the two validation gauges used as input into Quasar are
presented in Figure 8. For the Yarra River validation gauges, the plots indicate a good
agreement between predicted and observed flows in the low to mid flow volume range. The
model generally over predicted high flow events. The scatter plots for the other calibration
gauges can be seen in Appendix B. Figure 9 shows a cross-verification of daily observed
and predicted runoff for the poorest calibrated gauge on Watsons Creek. The poor data
quality low flow periods of the observed time series is evident with missing observed data in
many of the low flow periods. These cross-verification graphs proved useful in trying to
match up small event flows for these sites, which fit reasonably well and can be considered
good calibrations for the purpose of this modelling. Figure 10 shows a typical verification of
daily observed and predicted runoff for the downstream gauge at Chandler Highway on the
Yarra River. The flow duration curve in Figure 11 is a good match. Predicted flows tend to be
less than observed flows for flows less than 1000 ML/d, however the medium size flow
events match up very well.
The base case calibration scenario was then used to investigate the effects of the various
scenarios tested in the modelling. The base case scenario will be used to compare the
effects of DO with the altered flow regimes and climate change scenarios in the lower
reaches of the Yarra River.
229143 - Yarra River, Chandler Hwy (E=0.69)
25,000
25,000
20,000
20,000
Predicted daily flow (ML)
Predicted daily flow (ML)
229142 - Yarra River, Fitzsimons Ln (E=0.54)
15,000
10,000
5,000
15,000
10,000
5,000
-
-
5,000
10,000
15,000
20,000
Observed daily flow (ML)
25,000
-
5,000
10,000
15,000
20,000
25,000
Observed daily flow (ML)
Figure 8 - Predicted and observed daily streamflow scatter plots for the gauges to be used as Quasar inputs
Source hydrology calibration and dissolved oxygen modelling
20
229608 - Watsons Creek, Kangaroo Ground Sth
200
180
160
Daily Streamflow (ML)
140
120
observed
predicted
100
80
60
40
20
0
01/1998
02/1998
03/1998
04/1998
05/1998
06/1998
07/1998
08/1998
09/1998
10/1998
11/1998
12/1998
01/1999
Figure 9 - Daily hydrograph showing poor (deleted flat lining data) observed data quality in low flows
229143 - Yarra River, Chandler Highway
8,000
7,000
Daily Streamflow (ML)
6,000
5,000
4,000
observed
predicted
3,000
2,000
1,000
0
1/01/98
1/02/98
1/03/98
1/04/98
1/05/98
1/06/98
1/07/98
1/08/98
1/09/98
1/10/98
1/11/98
Figure 10 - Typical cross verification daily hydrograph on the Yarra River
21
Source hydrology calibration and dissolved oxygen modelling
1/12/98
1/01/99
Flow Duration Curve - 229143 Chandler Highway
100000
10000
Runoff (ML)
1000
100
10
1
0%
20%
40%
60%
80%
100%
0.1
0.01
Frequency
Observed
Predicted
Figure 11 - Flow duration curve for Chandler Hwy showing a good calibration for daily flows
4.2
Quality of Observed and Predicted Data
A recurring problem in attempting to calibrate many of tributary gauges was the availability of
reliable gauge data. Many Melbourne Water gauging stations are used for Flood Warning
systems only, therefore are not of high resolution at low flows. This creates problems when
trying to calibrate these gauges especially the baseflow. As RRL tries to optimise the
coefficient of efficiency (E) over the entire data set it tends to underestimate baseflow and
overestimate peak events. With gauged baseflows being unreliable, good calibrations proved
difficult when looking at the total predicted versus observed flow volumes as well as the E
values. Manual adjustments were then used to match up the hydrographs a lot better in the
small to medium sized event flows. This resulted in E values decreasing. However this was
still a reasonable result as an aim of this project is to try and have reasonable results for
baseflow and low to medium flow events as low flows are seen as a driver for low DO levels.
Annual volumes can vary greatly in the Yarra River, from over 1000 GL on a wet year to less
than 200 GL on a dry year. Looking at annual total flow at Yarra Glen (Figure 12) which sits
upstream of the Yering Gorge pumping stations, it can be noted that predicted flow is
generally greater than the observed flows. The cause of the predicted flows being greater
than the observed is possibly due to the fact that a number of potential losses from the
system were not represented in the model, these include;
• Irrigation/Other Diversions
• Farm Dams
• Channel Losses
The chandler highway gauge downstream of Yering tends to have a higher annual observed
flow then predicted flow (Figure 13). This is due the way the Yering pumps are actually run in
the system compared to how they were modelled. Too much water is being extracted from
the river in the model. Figure 14 shows the observed extractions from Yering Gorge and the
extracted volume from the base case over time. It can be seen that the modelled extractions
are significantly greater pre 1997, however match up a lot better in the drier years post 1997.
Source hydrology calibration and dissolved oxygen modelling
22
This is most likely due to Sugarloaf Reservoir being full pre 1997. To improve the extractions
in the model, Sugarloaf Reservoir levels would need to be modelled. For the modelling
purposes of this project this was still deemed suitable enough to test the effects of the Yering
pumps on river flow and DO.
229206 - Yarra River, Yarra Glen
1100
1000
900
Total Annual Flow (GL)
800
700
600
Observed
Predicted
500
400
300
200
100
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
0
Figure 12 – Annual Flow Volumes at Yarra Glen, upstream of the Yering Gorge pumping station
229143 - Yarra River Chandler Highway
1,200
1,100
1,000
Total Annual Flow (GL)
900
800
700
Observed
Predicted
600
500
400
300
200
100
0
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
Figure 13 – Annual flow volumes at Chandler Hwy, downstream of the Yering Gorge pumping station
23
Source hydrology calibration and dissolved oxygen modelling
1200
1100
1000
Extracted Volume (ML/day)
900
800
700
600
500
400
300
200
100
Predicted
1/01/09
1/07/08
1/01/08
1/07/07
1/01/07
1/07/06
1/01/06
1/07/05
1/01/05
1/07/04
1/01/04
1/07/03
1/01/03
1/07/02
1/01/02
1/07/01
1/01/01
1/07/00
1/01/00
1/07/99
1/01/99
1/07/98
1/01/98
1/07/97
1/01/97
1/07/96
1/01/96
1/07/95
1/01/95
1/07/94
1/01/94
1/07/93
1/01/93
1/07/92
1/01/92
1/07/91
1/01/91
1/07/90
1/01/90
0
Observed
Figure 14 - Modelled and predicted Yering Gorge extractions
4.3
Scenario Testing
The scenarios run in Source to produce a flow time-series for input into the DO model and
compared against the base case were;
• Scenario 2a – No diversions and No Yering Pumping. Compared to the base scenario
the total average annual runoff increased by 33% for the dry (2006) year, 34% for the
average (2004) year at Chandler Hwy
• Scenario 2b – No Yering Pumping. Total average annual runoff is only marginally less
than that of Scenario a. with an increase of 29% for the dry year and 25% greater for
the average year compared to the base case at Chandler Hwy
• Scenario 2c – Increased Environmental Flow. Increasing the minimum environmental
flow is notably greater in dry years with an increase in total average annual runoff of
17% compared to a 4% increase for the average year at Chandler Hwy
• Scenario 3a,b– Climate Change 2030 & 2060. Total average annual runoff was
reduced significantly for the two climate change scenarios due to a decrease in rainfall
inputs. The dry year yielded a decrease of 25% and 35% for the 2030 and 2060
climate change, while the average year showed a decrease of 35% and 46%
respectively at Chandler Hwy.
Figure 15 suggests that scenario (2a) (no diversion and no pumping) produced more runoff
then that of scenario 2b (no pumping). However, including diversions (scenario 2b) only
results in a 3% reduction in annual flow compared to scenario 2a for the representative dry
year and a 7% reduction in the representative average year. This means that diversions in
the system have less of an impact than pumping on flows in the lower reaches of the Yarra
River. The reduction of flow under average conditions seems to be more significant than
under dry as more water is available to be diverted. It is not viable to completely stop these
diversions and pumping however reducing the amount appears to have a significant increase
in flow in the lower reaches of the Yarra River in both dry and average years. Scenario 2c
Source hydrology calibration and dissolved oxygen modelling
24
shows that by increasing the minimum environmental flow passing Yering, an increase in
flow can be achieved. This increase is more notable during dry conditions where DO is also
expected to be very low in the lower reaches. Further work may be required to examine
these results more closely and to investigate whether the minimum environmental flow target
at Yering needs to be increased. Both climate change scenarios 3a, b show a decrease in
total annual flow with the average year 10% more than the dry year.
600
Total Flow (GL)
500
400
300
200
100
0
Base
2a
2b
2c
3a
3b
Scenario
Dry (2006)
Average (2004)
Figure 15 – Total annual flow comparisons for the selected representative dry and average year at Chandler Hwy
DO modelling results for Chandler and Banksia reaches are presented as time series,
exceedance plots and spells plots during a typical wet year, dry year, and average rainfall
year. The results can be seen in Appendix C (Figures 18-31).
The DO modelling suggests that flow has a strong influence on DO and is the major driver.
All years modelled show a variation of DO, being lower during the warmer months and
higher during the cooler months. Figure 16 shows that there is not a large variation between
the scenarios in the distribution of spells of DO below 6mg/L for the representative average
year. This is also the case for the dry year (Figure 17). As expected the dry year has the
longest durations of DO being below the threshold. Each of the altered flow regime scenarios
showed higher DO than that of the base case as the input flow was nominally higher. The
climate change scenario flows typically show lower DO than the base case and altered flow
regime scenarios. Climate change is predicted to have the greatest impact on DO via
reductions in flow that translate to increased periods / frequency of low DO occurrences,
however further work is required to parameterise QUASAR with locally relevant rate constant
data for the Yarra to improve estimates.
25
Source hydrology calibration and dissolved oxygen modelling
Distribution of Spells
2004
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Distribution of Spells
2006
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
6.0 m g/L B ase
6.0 m g/L N o D i v
6.0 m g/L N o Y er D i v
6.0 m g/L I nc Env Flow s
6.0 m g/L CC 2030
6.0 m g/L CC 2060
Dec
6.0 m g/L M easured
Figure 16 - Chandler Hwy Do spell durations of less than 6mg/L for an average (2004) and dry (2006) year
Source hydrology calibration and dissolved oxygen modelling
26
5
Conclusions
The Source model was built to assist in the establishment of appropriate trigger levels, to
determine when and how to augment flow in the river to minimise low dissolved oxygen
events and minimise ecological impacts. Hydrological time series of various scenarios from
the Source model were able to provide input into the QUASAR DO model to help assess the
subsequent impacts on river DO levels.
Acceptable calibration results were obtained within the catchment focus area. Predicted
runoff volumes in the focus area of the lower Yarra River were within 5% of the observed
flow, however improvements in gauged data from many of the tributaries are needed to
improve the calibrations at low flows. The modelling of the Yering Gorge pumping station
must also be improved to gain better calibration results.
The results suggest that flow is the major driver for DO in the river. Scenario modelling has
shown that climate (dry versus average and wet years) has more influence on DO in the
Yarra rather than specific altered flow scenarios. Climate change is predicted to reduce flow
in the river, which results in more frequent and longer periods of low DO. Under climate
change conditions various management options may need to be considered to combat a
decline in DO in the Yarra. Further work is needed to establish what the critical limits are and
how management options can keep DO at sustainable levels.
The work demonstrates that Source is a very useful and flexible modelling package that
allows a variety of scenarios to be tested easily.
27
Source hydrology calibration and dissolved oxygen modelling
6
References
Argent R. M. (2006) PortsE2: A Decision Support System for Water Quality Improvement in
Port Phillip and Western Port. CEAH Report 01/06.
BMT WBM (2008) PortsE2 Model Calibration and Extension Project: Draft Report. BMT
WBM. Brisbane.
Commonwealth of Australia (2006) LandSat Imagery. Commonwealth of Australia.
Cox, B.A. (2003). A review of currently available in-stream water-quality models and their
applicability for simulating dissolved oxygen in lowland rivers. The Science of
the Total Environment 314 –316: 335–377.
CRC for Catchment Hydrology (2004) Rainfall Runoff Library – User Guide. CRC for
Catchment Hydrology.
Chiew, F., and Scanlon, P. (2002). Estimation of Pollutant Concentrations for E2Modelling of
the South-East Queensland Region. – Technical Report. Cooperative Research
Centre for Catchment Hydrology. Available from
http://www.catchment.crc.org.au/pdfs/technical200202.pdf
Connell Wagner (2006) Toorourrong Reservoir Capacity Survey Report – Reservoir
Remedial Works. Connell Wagner. South Melbourne.
Delgado, P., Kelley, P., Murray, N. & Satheesh, A. (2011), Source User Guide, eWater
Cooperative Research Centre, Canberra, Australia.
DSE (2005) Vicmap Elevation. Victorian Government – Department of Sustainability and
Environment. Melbourne.
DSE (2007) Our Water Our Future – Next Stage of the Governments Water Plan. Victorian
Government – Department of Sustainability and Environment. Melbourne
Eatherall, A., Boorman, D.B., Williams R.J., and Kowe, R. (1998). Modelling in-stream water
quality in LOIS. The Science of the Total Environment 210/211 (1998) 499-517.
eWater (2008) Watercast: The Water and Contaminant Analysis and Simulation Tool for
Modelling – User Guide. eWater CRC. Canberra.
eWater (2009) Yarra River Focus Catchment. eWater CRC. Canberra. Available through:
http://www.ewatercrc.com.au/reports/Yarra%20focus%20catchment%20aug091.pdf
Lewis, D.R., Williams, R.J., Whitehead, P.G. (1997). Quality simulation along rivers
(QUASAR): an application to the Yorkshire Ouse. The Science of the Total
Environment 194/195 (1997) 399-418.
MMBW (1985) Toorourrong Reservoir Survey – Unpublished Records. Melbourne and
Metropolitan Board of Works. Melbourne.
MWC (2003) Yan Yean Reservoir Data Sheet – Internal Document, Melbourne Water
Corporation. Melbourne.
Source hydrology calibration and dissolved oxygen modelling
28
MWC and PPWCMA (2004a) Port Phillip and Westernport Regional River Health Strategy,
Melbourne Water Corporation, Melbourne.
MWC and PPWCMA (2004b) Melbourne’s Rivers and Creeks 2004, Melbourne Water
Corporation, Melbourne.
MWC (2007) Yarra River Extreme Drought Environmental Flows - Recommendations Report
September 2007. Melbourne Water Corporation, Melbourne.
Nash, J.E., and Sutcliffe, J.V. (1970). River Forecasting Using Conceptual Models, 1.a A
Discussion of Principles. Journal of Hydrology, 10: 282-290.
http://www.elsevier.com/wps/find/journaldescription.cws_home/503343/descripti
on#
SILO – Queensland DERM (2009). The Silo Data Drill. Department of Environment and
Natural Resource Management (DERM).
Victorian Government (2007) Water Act 1989 – Temporary Qualification of Rights in the
Melbourne Water Supply System – Yarra October 2007. Victorian Government.
Melbourne.
Waters D. (2008) Water quality monitoring and E2 modelling in the South West NRM
Region, Queensland –Technical Report for South West NRM. Department of
Natural Resources and Water. Coorparoo. ISBN 978-1741723175
29
Source hydrology calibration and dissolved oxygen modelling
Acknowledgements
The author would like to acknowledge the contributions and assistance of the following
people, without whose assistance the model would not have been completed.
• David Waters – eWater CRC. For all his assistance with regards to modelling and data
requirements as well as reviewing and editing the report.
• Shane Haydon – Melbourne Water. For assisting with the modelling process.
• Phillip Jordan – SKM. Also for assisting with the modelling process.
• Ian S Watson – Melbourne Water. For providing information and data regarding water
supply reservoirs and diversions.
• Steve Hosking – Melbourne Water. For providing assistance and data relating to
Irrigation Diversions.
• Sholto Maud and Peter Waugh – Melbourne Water. For providing and assisting with
Melbourne Water’s Hydstra database and gauging station information.
• Tam Hoang – Melbourne Water. For providing assistance with catchment area
information.
• Christine Hughes – Melbourne Water. For providing information on Yarra
environmental flow entitlements.
• Trevor Buckingham – Yarra Valley Water. For providing information and data on
catchment STPs.
• Jacinta Burns and Ted Chylinski – Melbourne Water. For providing GIS data and
assistance.
• Graham Rooney, Rhys Coleman and Melbourne Water’s Research and Technology
Team for reviewing the paper.
• Yarra River Application Project Team, eWater CRC and Melbourne Water generally
for all the assistance provided along the way.
• CSIRO for climate change data
• SKM for QUASAR water quality modelling
Source hydrology calibration and dissolved oxygen modelling
30
Appendix A
Catchment
Gauge
ID
Baseflow
coeff.
Impervious
Threshold
Infiltratio
n coeff.
Infiltratio
n shape
Interflo
w coeff.
Pervious
fraction
RISC
recharg
e coeff.
SMSC
Darebin Creek
229612
0.186
5
400
4
0.05
0.88
2
0.422
297
Diamond
Creek
229619
0.397
5
214
2.6
0.03
1
4.8
0.184
384
229618
0.2
5
88
1.7
0
0.91
5
0.309
400
229229
0.02
5
53
4
0
0.75
5
0.5
300
229648
0.114
3
362
5
0.3
0.8
0.3
0.4
180
229608
0.28
5
270
5
0.03
1
5
0.21
500
229690
0.065
5
397
1
0.08
1
5
0.4
500
229401
0.08
5
398
2.9
0.3
1
0.4
0.6
350
229206
0.158
5
120
0.8
0
1
2
0.8
190
Koonung
Creek
Mullum
Mullum Creek
Watsons
Creek
Olinda Creek
Stringybark
Creek
Yarra River
31
Source hydrology calibration and dissolved oxygen modelling
Appendix B
229143 - Yarra River, Chandler Hwy (E=0.69)
Flow Duration Curve - 229143 Chandler Highway
25,000
10000
1000
Runoff (ML)
15,000
10,000
100
10
1
5,000
0%
20%
40%
60%
80%
100%
0.1
-
5,000
10,000
15,000
20,000
25,000
0.01
Observed daily flow (ML)
Frequency
Observed
Predicted
229143 - Yarra River, Chandler Hwy
45,000
42,500
40,000
37,500
35,000
32,500
30,000
27,500
observed
predicted
25,000
22,500
20,000
17,500
15,000
12,500
10,000
7,500
5,000
2,500
Source hydrology calibration and dissolved oxygen modelling
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
0
1978
Daily Streamflow (ML)
Predicted daily flow (ML)
100000
20,000
32
Flow Duration Curve
229135 - Yarra River, Heidelberg (E=0.65)
100000
10000
20,000
1000
Runoff (ML)
15,000
10,000
100
10
1
5,000
0.1
-
5,000
10,000
15,000
20,000
25,000
0%
20%
40%
60%
80%
0.01
Observed daily flow (ML)
Frequency
Observed
Predicted
229135 - Yarra River, Heidelberg
47,500
45,000
42,500
40,000
37,500
35,000
Daily Streamflow (ML)
32,500
30,000
27,500
observed
predicted
25,000
22,500
20,000
17,500
15,000
12,500
10,000
7,500
5,000
2,500
33
Source hydrology calibration and dissolved oxygen modelling
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
0
1978
Predicted daily flow (ML)
25,000
100%
Flow Duration Curve
25,000
100000
20,000
10000
1000
Runoff (ML)
15,000
10,000
5,000
100
10
1
-
5,000
10,000
15,000
20,000
25,000
0.1
0%
20%
40%
60%
80%
100%
Observed daily flow (ML)
0.01
Frequency
Observed
Predicted
229142 - Yarra River, Templestowe
47,500
45,000
42,500
40,000
37,500
35,000
32,500
30,000
27,500
observed
predicted
25,000
22,500
20,000
17,500
15,000
12,500
10,000
7,500
5,000
2,500
Source hydrology calibration and dissolved oxygen modelling
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
0
1978
Daily Streamflow (ML)
Predicted daily flow (ML)
229142 - Yarra River, Tem plestow e (E=0.54)
34
Flow Duration Curve
229200 - Yarra River, Warrandyte (E=0.56)
100000
10000
20,000
1000
Runoff (ML)
15,000
10,000
5,000
100
10
1
-
5,000
10,000
15,000
20,000
0.1
25,000
Observed daily flow (ML)
0%
20%
40%
60%
80%
100%
0.01
Frequency
Observed
Predicted
229200B - Yarra River, Warrandyte
45,000
42,500
40,000
37,500
35,000
Daily Streamflow (ML)
32,500
30,000
27,500
observed
predicted
25,000
22,500
20,000
17,500
15,000
12,500
10,000
7,500
5,000
2,500
35
Source hydrology calibration and dissolved oxygen modelling
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
0
1978
Predicted daily flow (ML)
25,000
Flow Duration Curve
229206 - Yarra River, Yarra Glen (E=0.48)
100000
10000
20,000
1000
Runoff (ML)
15,000
10,000
5,000
100
10
1
-
5,000
10,000
15,000
20,000
0.1
25,000
0%
20%
40%
60%
80%
100%
0.01
Observed daily flow (ML)
Frequency
Observed
Predicted
229206 - Yarra River, Yarra Glen
47,500
45,000
42,500
40,000
37,500
35,000
Daily Streamflow (ML)
Predicted daily flow (ML)
25,000
32,500
30,000
27,500
observed
predicted
25,000
22,500
20,000
17,500
15,000
12,500
10,000
7,500
5,000
2,500
0
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
Source hydrology calibration and dissolved oxygen modelling
2008
2009
36
Flow Duration Curve
229229 - Koonung Creek (E=0.51)
10000
600
1000
500
Runoff (ML)
Predicted daily flow (ML)
700
400
300
200
100
10
1
100
0%
20%
40%
60%
80%
0.1
-
100
200
300
400
500
600
700
0.01
Observed daily flow (ML)
Frequency
Observed
Predicted
229229- Koonung Creek
4,500
4,000
Daily Streamflow (ML)
3,500
3,000
2,500
observed
predicted
2,000
1,500
1,000
500
37
Source hydrology calibration and dissolved oxygen modelling
2009
2008
2008
2007
2007
2006
2006
2005
2005
2004
2004
2003
2003
2002
2002
2001
2001
2000
2000
1999
1999
1998
1998
1997
1997
1996
1996
0
100%
Flow Duration Curve
229401 - Stringybark Creek (E=0.69)
100
100
Predicted daily flow (ML)
80
10
Runoff (ML)
60
40
20
1
0%
20%
40%
60%
80%
100%
0.1
-
20
40
60
80
100
Observed daily flow (ML)
0.01
Frequency
Observed
Predicted
229401 - Stringybark Creek
100
Daily Streamflow (ML)
80
60
observed
predicted
40
20
0
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
Source hydrology calibration and dissolved oxygen modelling
2009
38
Flow Duration Curve
229608 - Watsons Creek (E = 0.37)
10000
1000
100
Runoff (ML))
1,500
1,000
500
10
1
0%
20%
40%
60%
80%
0.1
-
500
1,000
1,500
2,000
0.01
Observed daily flow (ML)
Frequency
Observed
Predicted
229608 - Watsons Creek, Kangaroo Ground Sth
1600
1400
1200
Daily Streamflow (ML)
1000
observed
predicted
800
600
400
200
39
Source hydrology calibration and dissolved oxygen modelling
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
0
1990
Predicted daily flow (ML)
2,000
100%
Flow Duration Curve
229612 - Darebin Creek, Bundoora (E=0.6)
10000
1000
1,500
100
Runoff (ML)
1,000
500
10
1
0%
20%
40%
60%
80%
100%
0.1
-
500
1,000
1,500
2,000
Observed daily flow (ML)
0.01
Frequency
Observed
Predicted
229612 - Darebin Creek, Bundoora
2600
2400
2200
2000
1800
1600
observed
predicted
1400
1200
1000
800
600
400
200
Source hydrology calibration and dissolved oxygen modelling
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
0
1978
Daily Streamflow (ML)
Predicted daily flow (ML)
2,000
40
Flow Duration Curve
6,000
10000
5,000
1000
4,000
Runoff (ML)
Predicted daily flow (ML)
229618 - Diam ond Creek, Eltham (E=0.61)
3,000
2,000
100
10
1
1,000
0%
20%
40%
60%
80%
0.1
-
1,000
2,000
3,000
4,000
5,000
6,000
0.01
Observed daily flow (ML)
Frequency
Observed
Predicted
229618 - Diamond Creek, Eltham
9000
8000
Daily Streamflow (ML)
7000
6000
5000
observed
predicted
4000
3000
2000
1000
41
Source hydrology calibration and dissolved oxygen modelling
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
1978
0
100%
Flow Duration Curve
229619 - Diam ond Creek, Hurstbridge (E=0.57)
10000
1000
5,000
Runoff (ML)
4,000
3,000
2,000
100
10
1
1,000
0%
-
1,000
2,000
3,000
4,000
5,000
20%
40%
60%
80%
100%
0.1
6,000
Observed daily flow (ML)
0.01
Frequency
Observed
Predicted
229619 - Diamond Creek, Hurstbridge
6000
5500
5000
4500
4000
3500
observed
predicted
3000
2500
2000
1500
1000
500
Source hydrology calibration and dissolved oxygen modelling
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
1979
0
1978
Daily Streamflow (ML)
Predicted daily flow (ML)
6,000
42
Flow Duration Curve
1,500
10000
1,250
1000
1,000
Runoff (ML)
Predicted daily flow (ML)
229648 - Mullum Mullum Creek, Doncaster East (E=0.45)
750
500
100
10
1
0%
250
20%
40%
60%
80%
0.1
-
250
500
750
1,000
1,250
0.01
1,500
Frequency
Observed daily flow (ML)
Observed
Predicted
229648 - Mullum Mullum Creek, Doncaster East
2000
1800
Daily Streamflow (ML)
1600
1400
1200
observed
predicted
1000
800
600
400
200
43
Source hydrology calibration and dissolved oxygen modelling
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
1980
0
100%
Flow Duration Curve
229690 - Olinda Creek, Mt Evelyn (E=0.52)
1000
200
100
Runoff (ML)
150
100
50
10
1
0%
20%
40%
60%
80%
100%
0.1
-
50
100
150
200
250
Observed daily flow (ML)
0.01
Frequency
Observed
Predicted
229690 - Olinda Creek, Mt Evelyn
500
400
observed
predicted
300
200
100
Source hydrology calibration and dissolved oxygen modelling
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
0
1987
Daily Streamflow (ML)
Predicted daily flow (ML)
250
44
Appendix C
Chandler Site
Time series of DO
T ime Series Plot
12
B ase
N o D iv
11
N o Y er D iv
10
I nc Env Flow s
CC 2030
9
CC 2060
D O (m g/L )
8
Measured
7
6
5
4
3
2
1
0
Jan96
Mar96
May96
Jul96
Sep96
N ov96
Figure 17 Chandler DO time series for a wet year, 1996
45
Source hydrology calibration and dissolved oxygen modelling
T ime Series Plot
12
B ase
N o D iv
11
N o Y er D iv
10
I nc Env Flow s
CC 2030
9
CC 2060
D O (m g/L )
8
Measured
7
6
5
4
3
2
1
0
Jan04
Mar04
May04
Jul04
Sep04
N ov04
Figure 18 Chandler DO time series for an average year, 2004
T ime Series Plot
12
B ase
N o D iv
11
N o Y er D i v
10
I nc Env Flow s
CC 2030
9
CC 2060
D O (mg/L )
8
M easured
7
6
5
4
3
2
1
0
Jan06 Feb06 Mar06 A pr06 M ay 06 Jun06 Jul06 A ug06 Sep06 Oct06 N ov06 D ec06
Figure 19 Chandler DO time series for a dry year, 2006
Source hydrology calibration and dissolved oxygen modelling
46
DO exceedance curves
Exceedance Frequency Plot
12
B ase
11
N o D iv
D O (mg/L )
10
N o Y er D iv
9
I nc Env Flow s
CC 2030
8
CC 2060
7
M easured
6
0
10
20
30
40
50
60
70
80
90
100
T ime Exceeded (% )
Figure 20 Chandler DO exceedance plot for a wet year, 1996
47
Source hydrology calibration and dissolved oxygen modelling
Exceedance Frequency Plot
12
11
B ase
10
N o D iv
D O (mg/L )
9
8
N o Y er D iv
7
I nc Env Flow s
6
5
CC 2030
4
CC 2060
3
M easured
2
1
0
10
20
30
40
50
60
70
80
90
100
T ime Exceeded (% )
Figure 21 Chandler DO exceedance plot for an average year, 2004
Exceedance Frequency Plot
12
B ase
11
N o D iv
D O (m g/L )
10
9
N o Y er D iv
8
I nc Env Flow s
7
CC 2030
6
CC 2060
5
M easured
4
0
10
20
30
40
50
60
70
80
90
100
T ime Exceeded (% )
Figure 22 Chandler DO exceedance plot for a dry year, 2006
Source hydrology calibration and dissolved oxygen modelling
48
DO spells plots
D istribution of Spells
1996
Jan
Feb
Mar
A pr
M ay
6.0 m g/L CC 2030
Jun
Jul
A ug
Sep
Oct
N ov
D ec
6.0 m g/L CC 2060
Figure 23 Chandler DO spell durations of less than 6mg/L for a wet year, 1996
D istribution of Spells
2004
Jan
Feb
M ar
A pr
M ay
Jun
Jul
A ug
Sep
Oct
N ov
6.0 mg/L B ase
6.0 m g/L N o D iv
6.0 mg/L N o Y er D iv
6.0 mg/L I nc Env Flow s
6.0 m g/L CC 2030
6.0 mg/L CC 2060
D ec
6.0 mg/L M easured
Figure 24 Chandler DO spell durations of less than 6mg/L for an average year, 2004
D istribution of Spells
2006
Jan
Feb
Mar
A pr
M ay
Jun
Jul
A ug
Sep
Oct
N ov
6.0 mg/L B ase
6.0 m g/L N o D i v
6.0 mg/L N o Y er D i v
6.0 mg/L I nc Env Flow s
6.0 m g/L CC 2030
6.0 mg/L CC 2060
D ec
6.0 mg/L M easured
Figure 25 Chandler DO spell durations of less than 6mg/L for a dry year, 2006
49
Source hydrology calibration and dissolved oxygen modelling
Banksia Site
Time series of DO
T ime Series Plot
12
10
D O (mg/L )
8
6
4
2
0
Jan96
Mar96
M ay 96
Jul96
Sep96
N ov96
Figure 26 Banksia DO time series for a wet year, 1996
Source hydrology calibration and dissolved oxygen modelling
50
T ime Series Plot
B ase
11
N o D iv
N o Y er D i v
10
I nc Env Flow
9
CC 2030
CC 2060
D O (mg/L )
8
M easured
7
6
5
4
3
2
1
0
Jan04
M ar04
May04
Jul04
Sep04
N ov04
Figure 27 Banksia DO time series for an average year, 2004
T ime Series Plot
B ase
10
N o D iv
N o Y er D i v
9
I nc Env Flow s
8
CC 2030
CC 2060
D O (mg/L )
7
M easured
6
5
4
3
2
1
0
Jan06
Mar06
M ay 06
Jul06
Sep06
N ov06
Jan07
Figure 28 Banksia DO time series for a dry year, 2006
51
Source hydrology calibration and dissolved oxygen modelling
DO exceedance curves
Exceedance Frequency Plot
13
12.5
B ase
12
N o D iv
D O (mg/L )
11.5
11
N o Y er D iv
10.5
I nc Env Flow s
10
CC 2030
9.5
CC 2060
9
8.5
Measured
8
0
10
20
30
40
50
60
70
80
90
100
T ime Exceeded (% )
Figure 29 Banksia DO exceedance plot for a wet year, 1996
Exceedance Frequency Plot
11.5
B ase
11
10.5
N o D iv
D O (mg/L )
10
N o Y er D i v
9.5
9
I nc Env Fl ow
8.5
8
CC 2030
7.5
CC 2060
7
6.5
M easured
6
0
10
20
30
40
50
60
70
80
90
100
T ime Exceeded (% )
Figure 30 Banksia DO exceedance plot for an average year, 2004
Source hydrology calibration and dissolved oxygen modelling
52
Exceedance Frequency Plot
10.5
B ase
10
N o D iv
D O (mg/L )
9.5
9
N o Y er D iv
8.5
I nc Env Flow s
8
CC 2030
7.5
CC 2060
7
Measured
6.5
6
0
10
20
30
40
50
60
70
80
90
100
T ime Exceeded (% )
Figure 31 Banksia DO exceedance plot for a dry year, 2006
DO spells plots
Note that no spells plots were compiled for the Banksia reach as DO was typically above the
6mg/L threshold.
53
Source hydrology calibration and dissolved oxygen modelling
Source hydrology calibration and dissolved oxygen modelling
54
eWater Cooperative Research Centre
eWater Limited ABN 47 115 422 903
UC Innovation Centre
University Drive South
Bruce ACT 2617
Phone: +61 2 6201 5168
[email protected]
www.ewater.com.au
55
Source hydrology calibration and dissolved oxygen modelling
© 2012 eWater Ltd