Yarra River Application Project Source Catchments Hydrology Calibration Report

Transcription

1 Yarra River Application Project Source Catchments Hydrology Calibration Report Richard Carty and Simon Pierotti ewater Technical Report

2

3 Yarra River Application Project Source Catchments Hydrology Calibration Report Progress Report Richard Carty and Simon Pierotti 1 1 Melbourne Water November 2010 ewater Cooperative Research Centre Technical Report

4 ewater Cooperative Research Centre 2009 All rights reserved. No parts of this work may be reproduced in any form or by any means - graphic, electronic, or mechanical, including photocopying, recording, taping, or information storage and retrieval systems - without the written permission of the publisher. Products that are referred to in this document may be either trademarks and/or registered trademarks of the respective owners. The publisher and the author make no claim to these trademarks. While every precaution has been taken in the preparation of this document, the publisher and the author assume no responsibility for errors or omissions, or for damages resulting from the use of information contained in this document or from the use of programs and source code that may accompany it. 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. Published online November 2010 ewater CRC Innovation Centre, University of Canberra ACT 2601, Australia Phone (02) Fax (02) info@ewatercrc.com.au Web ewater CRC is a cooperative joint venture whose work supports the ecologically and economically sustainable use of Australia s water and river systems. ewater CRC was established in 2005 as a successor to the CRCs for Freshwater Ecology and Catchment Hydrology, under the Australian Government s Cooperative Research Centres Program.

5 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 Water ewater CRC. For all his assistance with regards to modelling and data requirements. 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 Hydrstra 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. Yarra River Application Project Team, ewater CRC and Melbourne Water generally for all the assistance provided along the way. i

6 Executive Summary The Yarra River catchment encompasses an area of some 4,000km 2 and provides habitat for various aquatic species including Macquarie Perch, Murray Cod, various frogs, birds and eels. The catchment is also highly regulated with a number of irrigation diversions and is used extensively for water harvesting to supply the city of Melbourne. The state of Victoria has been experiencing a severe drought for over 10 years. The resultant effects of the drought have lead 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. Recently the Victorian government has been exploring ways to reduce the environmental flows to the river by 10 GL/annum. However concerns have been raised regarding what impact the lower flows are having 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 on going drought has also caused a gradual decline in both the overall total flow and the number of high flow events within the river. Through the ewater CRC s Applications projects, the Source Catchments 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 will then be used by the ecological team to test the ecological impacts of various current and future flow management scenarios. The target area of the application project was the mid to lower Yarra (ie. Warrandyte to Dights Falls). Source Catchments has been successfully calibrated up to Coldstream on the Yarra River upstream of which an observed flow time series was used due to a lack of available data on extractions. Four different pumping scenarios were tested to achieve the minimum flow requirements in the Yarra River downstream of Yering Gorge. This report describes the hydrology calibration and validation process. Key findings from the hydrology calibration and validation process include: The runoff generated by Source Catchments generally represented the system well, with predicted (modelled) monthly flow within ±10% of the observed runoff. The Nash Sutcliffe objective function statistics ranged from for calibrated subcatchments, with the main reach of the Yarra having a monthly E value ranging from Influence of the lower tributaries on baseflows in the lower Yarra River was found to be minimal, with most being <5% of the total baseflow. Poor gauging data quality especially in low flow periods for many of the tributaries meant that calibration in RRL proved difficult. Manual adjustments were used to match up peak flows. Four pumping scenarios were tested and compared to observed data. Further work is required to assess the range of available node models in Source Catchments better represent the current pumping activities The Source Catchments model provided a relatively flexible tool for catchment modelling. Unfortunately the extent of regulation and complexities of the Yarra catchment, made it difficult to calibrate the entire catchment. ii

7 The development of this base model will provide a simulation tool which enables stakeholders of the Yarra catchment to easily assess a range of current and future scenarios of interest Over the next 12 months a more detailed assessment of the range of node models available in Source Catchments will be undertaken to best represent current pumping operations. A new plug-in may be developed to improve the output of these scenarios. The intent for any future work is to look at modelling the remainder of the catchment, where observed flow is currently used including adding storage models to account for the larger reservoirs in the headwaters of the Yarra River. If time permits modelling of constituents and climate change scenarios will also be looked at. iii

8 Disclaimer The following disclaimer relates to all Melbourne Water maps and data used in the preparation of the Source Catchments 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 anyway 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 carry out its own investigations if appropriate. iv

9 Table of Contents List of Figures...vi List of Tables...vii 1.0 Introduction Catchment Description Victoria in Drought Implications for environmental flows Dissolved Oxygen and River Flow Source Catchments model ewater Application Project Objectives Methodology Model Inputs Model Sub-catchment and Stream Generation Rainfall and Potential Evapotranspiration (PET) Land use and Pervious Fraction Determination Determination of hydrological parameters Observed inflows Yering Gorge Pumping Station Hydrology calibration and validation Scenario Testing Yering Gorge Pumping Scenarios Results and Discussion Hydrology Calibration Results Influence of the Observed Inflow time series Quality of Observed and Predicted Data Scenario Testing Yering Gorge Pumping Scenarios Conclusions References Appendix 1 SIMHYD Parameter List Appendix 2: Calibration Results v

10 List of Figures Figure 1: Yarra River Catchment indicating monitoring sites, reservoirs, streams and broad land use classes Figure 2: Decreasing trend in minimum Dissolved Oxygen concentration over a 10 year period. Plotted against monthly flow at Kew (Chandler HWY Gauge)... 3 Figure 3: Yarra River Application Project Area (yellow line) and the locations of the gauging stations used during calibration (Landsat Imagery - Commonwealth of Australia 2006). Sites with blue labels are gauges on the Yarra River, sites with yellow labels are gauges on Yarra River tributaries and sites with red labels are gauges where observed inflows were used (see Table 3)... 4 Figure 4: Yarra River Catchment Model generated by the 50m DEM with the node link network overlayed. The red box indicates the target area of the application project... 5 Figure 5: Land Use Classification Map... 7 Figure 6: Diversion Graph for 200 ML/d environmental flow requirement Figure 7: Predicted and observed daily streamflow scatter plots for selected gauges Figure 8 - Daily Hydrograph for a poorly calibrated gauge, showing poor observed data quality in low flows Figure 9: Typical cross verification daily hydrograph on the Yarra River Figure 10: Annual Flow Volumes at Yarra River, Yarra Glen Figure 11 - Annual flow volumes at Yarra River, Chandler Highway Figure 12: Daily extraction from Yering Gorge Pumping Station indicating that the flood harvesting tool did not accurately match extractions from the River Figure 13: Annual diversion volumes for all 4 scenarios indicating that the maximum annual diversion volume was met in most years vi

11 List of Tables Table 1: Land Use Classification Based upon BMT WBM Report (2008)... 6 Table 2: Gauging Stations within the Target Area of the Model... 8 Table 3 : Gauging Stations used as observed inflows... 9 Table 4: Calibration results for the thirteen gauging stations Table 5: Estimated Influence of Major Tributaries on Yarra River Flow Table 6: SIMHYD Parameter Values Used for Calibration vii

12 1.0 Introduction 1.1. Catchment Description The Yarra River catchment is located to the north east of Melbourne and encompasses an area of some 4000 km 2 (Figure 1). The average annual rainfall across the catchment is variable, but typically ranges from 680mm in Burnley (near Melbourne) to 1080mm in the river headwaters around the Upper Yarra Reservoir (above Warburton) (MWC and PPWPCMA 2004a). Melbourne Figure 1: Yarra River Catchment indicating monitoring sites, reservoirs, streams and broad land use classes. 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 and has 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, approximately 40% of the catchment retains natural vegetation, 40% is used for agriculture and 20% is urbanised (though urbanisation is rapidly expanding) (MWC and PPWCMA 2004a, MWC and PPWCMA 2004b). In addition to direct water harvesting through the placement of water supply reservoirs at various locations throughout the catchment, water is extracted from the various rivers and creeks via weirs and a large pumping station in the mid Yarra (Yering Gorge ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 1

13 Pumping Station), which can pump to a maximum of 1000 ML/d. There are also numerous farm dams throughout the catchment making the Yarra River catchment a highly regulated system (MWC and PPWCMA 2004a) Victoria in Drought Implications for environmental flows 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). 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 1500 ML/d during April-May, or 2000 ML/d June-September 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 has now adopted these reduced environmental flow targets with the requirement of a yearly review on the ecological impact on the river. Once the level of water restrictions in Melbourne is eased, the original 245ML/d flow entitlements will be re-established (Victorian Government 2007 and MWC 2007). The Yering Gorge Scenario tests will attempt to simulate the impacts of these environmental flow constraints on the Yarra River 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 (Figure 2). Low DO concentrations in the river have serious ecological implications as fish and other aquatic life rely on oxygen to breathe; if there is insufficient oxygen in the water column they may not survive (ewater 2009). ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 2

14 Chandler HWY Dissolved Oxygen (Minimum Values) Time Series DO Concentration (mg/l) Monthly Flow (ML/Month) 0 0 1/08/ /12/ /04/2001 9/09/ /01/2004 5/06/ /10/2006 1/03/ /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) Figure 2 shows a decreasing trend of DO minimum concentrations over the past 10 years. Additionally, there are also indications of a significant increase in the variability of DO concentrations over the past 5 years coupled with some very low spikes (eg mg/l during March 2007). There has also been a consistent decrease in the monthly flow at this site over the past 10 years and a noticeable drop in the number of high flow events during the past 4 to 5 years. The concerns around the low dissolved oxygen concentrations has lead to the need to assess the impacts different flow regimes have upon the level of DO in the river. This project will model a range of flow scenarios to enable ecological modellers to assess the impacts of various flow regimes on DO 1.4. Source Catchments model Source Catchments is a whole of catchment modelling tool, capable of predicting the hydrologic behaviour of catchments. The model structure operates as a node-link network. Runoff and constituent loads that are generated within sub-catchments are fed into nodes and then routed throughout the model via the links (ewater 2008). The Yarra River project is one of a number of projects across South Eastern Australia testing the applicability of new ewater software to real world projects. The Source Catchments modelling software is one such tool being tested as part of this application project ewater Application Project Objectives The Yarra River ewater application project is a partnership project between Melbourne Water, SKM, Monash University, EPA Victoria, Department of Sustainability and Environment (DSE), Southern Rural Water and ewater. Specifically the project objectives are to: Test the suitability of the Source Catchments modelling software (under development by ewater) to a real world project, in this case generating a hydrological time series for DO modelling in the Yarra River. ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 3

15 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 in instigate certain management actions, such as refreshing water in poor DO zones by augmenting river flows (ewater 2009). Given the assumption that low DO concentrations are at least in part caused by low flows, this model will be focussing on low river flows and small to medium runoff events. Source Catchments was used to create a predicted runoff time series for the Yarra River catchment at specific gauging stations between Warrandyte (mid Yarra) and Dights Falls, Kew (Figure 3) to represent current conditions. This included calibration of subcatchments between the Coldstream (229653) and the Chandler Hwy (229143) gauging stations. Warrandyte Figure 3: Yarra River Application Project Area (yellow line) and the locations of the gauging stations used during calibration (Landsat Imagery - Commonwealth of Australia 2006). Sites with blue labels are gauges on the Yarra River, sites with yellow labels are gauges on Yarra River tributaries and sites with red labels are gauges where observed inflows were used (see Table 3). ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 4

16 2.0 Methodology The most recent available data was used to construct the Source Catchments model. A modelling period of 32 years was chosen (June ) 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 Model Inputs Model Sub-catchment and Stream Generation The Source Catchments sub-catchments and links were generated from a 20m digital elevation model (DEM) sourced from the Department of Sustainability and Environment Vicmap elevation state-wide DEM as at The DEM has a vertical accuracy of 5-10m on moderate to steep slopes and 1-5m on undulating or flat plain areas (DSE 2005). The 20m DEM was then re-sampled to 50m and pit filled. A minimum subcatchment area of 50 km 2 was chosen for the model and the locations of thirty seven Melbourne Water gauging stations added to the subcatchment configuration. Overall eighty two subcatchments were generated. This included the nine sub-catchments and six gauging stations in the estuarine reach of the river which have no calibration data available.. The nodes located throughout the model represent the catchment outlets and in cases where a gauging station is present the location of this station (Figure 4). Figure 4: Yarra River Catchment Model generated by the 50m DEM with the node link network overlayed. The red box indicates the target area of the application project. ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 5

17 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 Catchments 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 subcatchment was then extracted and used for hydrologic calibration external to Source Catchments Land use and Pervious Fraction Determination Although the focus of this Source Catchments model was the hydrology, land use configuration was considered given the impact land use can play on runoff in particular the imperviousness of urban areas. Secondly, by incorporating relevant land use categories, future users have the option to model constituents. The PortsE2 (Argent 2006) land use map and codes (Table 1) were used for this project for consistency. In total, 16 land use categories were created. These are summarised in Table 1 and Figure 5 Table 1: Land Use Classification Based upon BMT WBM Report (2008) 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 ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 6

18 Figure 5: 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 impervious surfaces. A pervious fraction of 0.7 for urban areas was applied in the model as per PortsE2 report, Argent (2006). This was then used as a basis for the pervious fraction parameter in the model calibration Determination of hydrological parameters Source catchment offers several rainfall runoff models for calibration. The available models were reviewed and SIMHYD was selected for rainfall runoff modelling. 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 the aforementioned simplicity and because it had been used within the modelling region in other projects. A set of SIMHYD hydrologic parameters were derived upstream of each gauging station. Nine parameter sets were derived for the catchment (Appendix 1). For ungauged catchments, one of the nine parameter sets were assigned to each subcatchment based on a comparison of spatial attributes between the gauged and ungauged catchments. Attributes considered was pervious fractions, land use types, and rainfall. 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 th June 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 (229200B) in the mid Yarra with <1% of the data missing, in which case missing data was ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 7

19 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 MW officers 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 about 70 gauging stations of variable quality. Source Catchments is not designed to model heavily regulated catchments. Therefore with the upper Yarra being heavily regulated, observed gauged timeseries were used in a number of the upper catchments as discussed in section For the modelled area, 29 gauging stations were assessed. Streamflow data from 13 gauges were deemed suitable for calibrating the hydrological model (Table 2), with nine of these being used in the SIMHYD parameterisation. Marked station locations are shown in Figure 3. Table 2: Details of the 13 Gauging Stations used for Model calibration Stream Gauge ID Location Catchment Area (km 2 ) Date Commenced Years Record % Data Missing Resolution (mm) Darebin Creek Bundoora Diamond Creek Koonung Creek Mullum Mullum Creek Watsons Creek Hurstbridge ^ Eltham * Bulleen * Doncaster East Kangaroo Ground Sth Olinda Creek * Mt Evelyn Stringybark Creek * Mt Evelyn ^ Yarra Glen Yarra River # Warrandyte # Fitzsimons La, Templestowe # Banksia St, Heidelberg # Chandler HWY, Fairfield * Gauge at a midpoint of a sub-catchment but still used in calibration ^Residual Calibration undertaken at this gauge with , and with # Not used for parameterisation ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 8

20 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 Catchments model and calibrated against the observed flow for 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 1. Four of the selected gauges locations did not correspond to a Source Catchment model node however they were still deemed viable to be used in calibrating that sub-catchment. Gauges and 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 Observed inflows Due to the complexity and impact of the Melbourne Water Headworks system on flow in the Yarra River catchment, it was decided that the full Yarra catchment would not be modelled. Suitable Calibrations were unable to be obtained for gauging stations in the Upper Yarra reaches of the model above Yarra Grange (Coldstream) as well as on the Plenty River, where Yan Yean and Toorourrong Reservoir have a major impact on streamflow. To alleviate these problems, observed flows taken from suitable gauges (Table 3) were used as inputs into the Source Catchments model as Replacement Flow node models. The gauges used were selected as they had appropriate length of record and data was reliable. Any missing data from these sites was in-filled to obtain a complete data set to correspond with the model run period from June 1978 to June Table 3 : Details of Gauging Stations used as observed inflows (Sites with red labels, Figure 3) Stream Gauge ID Location Plenty River Yarra River Catchment Area (km 2 ) Date Commenced Years Record Resolution (mm) Lower Plenty Yarra Grange (Coldstream) Yering Gorge Pumping Station One of the major diversions in the Yarra River Catchment is the Yering Gorge pumping station. During calibration an observed extraction node model was used in Source Catchments to extract this water from the main reach of the Yarra River (see Figure 3 for location). For scenario testing, the River Harvesting Diversion Node Model was used to attempt to mimic various pumping rules at Yering Gorge 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 ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 9

21 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, , , and ), were then used to validate the model within Source Catchments. This enabled the ungauged sub-catchments to be validated when gauged sub-catchment parameters were assigned to them. The results can be seen in Appendix Scenario Testing Yering Gorge Pumping Scenarios Four Scenarios were tested for the Yering gorge pumping station in this analysis. The flood harvest diversion tool in Source Catchments was used to test these scenarios. The flood harvest diversion tool allows user input for; Flow to be diverted over a certain upstream flow threshold A yearly total diversion amount A yearly diversion period The scenarios in this analysis revolved around the minimum flow requirements in the Yarra River downstream of Yering Gorge. Each Scenario had a diversion maximum of 130,000 ML/year which was historically the maximum amount of water taken over a year. The diversion period was set from May through to February as minimal pumping is done during the summer months. Currently a minimum flow of 200 ML/d must be in the river downstream of the pumps. Operators therefore observe upstream flows and limit pumping when flow gets to a critical level. Four 250 ML/d pumps are able to be used on the Yarra River to extract water to Sugarloaf Reservoir for Water Supply purposes. Each of these pumps can extract between ML/d. For modelling purposes only 3 pumps will be modelled as the fourth pump is normally used in the Maroondah Aqueduct and not the Yarra River (it has been used 3 times in the Yarra over 30 years). A maximum pumping rate of 750ML/d can be achieved. A flow diversion vs. upstream flow rate table was created for each of the four scenarios. An example of this for the 200 ML/d minimum flow requirement scenario can be seen in Figure 6. ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 10

22 Diversion Volume (ML/d) Upstream Flow Rate (ML/d) Figure 6: Diversion Graph for 200 ML/d environmental flow requirement Four scenarios were tested in the model using the flood harvest diversion tool; Scenario ML/d minimum flow Scenario ML/d minimum flow Scenario ML/d minimum flow Scenario ML/d minimum flow ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 11

23 3.0 Results and Discussion 3.1. Hydrology Calibration Results Table 4 summarises the hydrology calibration results at each of the suitable gauging stations. All predicted flows come from the Source Catchment model outputs, except for the four areas calibrated as residual catchments (229229, , , ). The calibration results listed in table 4 for these four catchments were derived from RRL. Table 4: Calibration results for the thirteen gauging stations Stream Gauge ID Location Darebin Creek Diamond Creek Koonung Creek Mullum Mullum Creek Watsons Creek Nash Sutcliffe Criteria (E daily ) Nash Sutcliffe Criteria (E monthly ) Predicted/ Observed Bundoora Hurstbridge Eltham * Bulleen * Doncaster East Kangaroo Ground Sth Olinda Creek * Mt Evelyn Stringybark Creek * Mt Evelyn ^ Yarra Glen B^ Warrandyte Yarra River ^ ^ Fitzsimons La, Templestowe Banksia St, Heidelberg Chandler ^ HWY, Fairfield *Midstream Calibration results taken from RRL ^Indicates that calibration is downstream on observed time series which is likely influencing the result. Total predicted streamflow volumes on the main stretch of the Yarra River and in the focus area were all within 10% of observed flow volumes. For the Yarra Tributary calibrations this was not the case, however this was deemed reasonable as gauge resolution and quality at these sites especially for baseflows is quite poor. Therefore the calibration of these sites focused on improving event flow predictions. Coefficient of efficiency values (E), which are a measure of the correlation between measured and modelled data range from for daily (E daily ) predicted and observed values. The monthly correlation (E monthly ) was generally higher then the daily data. For sites where an observed time series has no influence on downstream flow, the E daily ranges from Once again these sites had poor gauged data especially around baseflow estimates, therefore maximising the E value during calibration was not the main focus. Runoff estimates in the focus area of the Catchment can be considered good. However it must be noted that using observed time series upstream of the calibration area is ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 12

24 having a major influence on these results. For the purpose of this project this is a reasonable assumption. 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 in the model. A one day lag was applied upstream of the Yering Gorge pumping station. This proved to be extremely reliable, as it significantly improved the daily predicted and observed flow comparisons. The location of the lag was selected 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 four of the calibrated gauges are presented in Figure 7. For the Yarra River gauges, the plots indicate a very good agreement between predicted and observed flows across the full range of daily flow volumes. Some poor fits are seen on the Yarra tributary gauges, however as calibration of these sites looked at matching up event flows due to unreliable data at low flows the calibrations were deemed acceptable for the application. Figure 8 shows a verification of daily observed and predicted runoff for a poorly calibrated gauge. The poor data quality low flow periods of the observed time series is evident. Figure 9Error! Reference source not found. shows a typical verification of daily observed and predicted runoff for a Yarra River Gauge. The plot indicates that once calibrated the hydrologic model has estimated the daily streamflow satisfactorily Yarra River, Chandler HWY Yarra River, Warrandyte 30,000 25,000 Predicted daily flow (ML) 25,000 20,000 15,000 10,000 5,000 Predicted daily flow (ML) 20,000 15,000 10,000 5, ,000 10,000 15,000 20,000 25,000 30, ,000 10,000 15,000 20,000 25,000 Observed daily flow (ML) Observed daily flow (ML) Darebin Creek, Bundoora Diamond Creek, Eltham 2,500 6,000 Predicted daily flow (ML) 2,000 1,500 1, Predicted daily flow (ML) 5,000 4,000 3,000 2,000 1, ,000 1,500 2,000 2, ,000 2,000 3,000 4,000 5,000 6,000 Observed daily flow (ML) Observed daily flow (ML) Figure 7: Predicted and observed daily streamflow scatter plots for selected gauges ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 13

25 Watsons Creek, Kangaroo Ground Sth 1200 Daily Streamflow (ML) observed predicted /06/90 1/08/90 1/10/90 1/12/90 1/02/91 1/04/91 1/06/91 1/08/91 1/10/91 1/12/91 1/02/92 1/04/92 1/06/92 1/08/92 1/10/92 1/12/92 1/02/93 1/04/93 1/06/93 Figure 8 - Daily Hydrograph for a poorly calibrated gauge, showing poor observed data quality in low flows 20, Yarra River, Chandler Highway 17,500 15,000 Daily Streamflow (ML) 12,500 10,000 7,500 observed predicted 5,000 2, /06/90 1/08/90 1/10/90 1/12/90 1/02/91 1/04/91 1/06/91 1/08/91 1/10/91 1/12/91 1/02/92 1/04/92 1/06/92 1/08/92 1/10/92 1/12/92 1/02/93 1/04/93 1/06/93 1/08/93 1/10/93 1/12/93 1/02/94 1/04/94 1/06/94 Figure 9: Typical cross verification daily hydrograph on the Yarra River 3.2. Influence of the Observed Inflow time series The calibration stations listed in Table 4 from the Yarra River are downstream of an observed time series. This had a major impact upon the calibration results for gauges on the Yarra River, due to this flow being a driving factor in the amount of flow passing these gauges. Figure 9 shows a very good calibrated hydrograph, however this is a bit misleading due to the influence that the observed time series has. On a catchment where there is no observed flow series input into the model upstream of a gauge, the ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 14

26 calibration quality is diminished. Figure 8 shows one such catchment that has an average calibration result. The majority of small tributaries which join the Yarra River between Warrandyte and Dights Falls generate minimal runoff when compared to flow in the Yarra River. This is especially the case for baseflows. The impact of event flows is slightly higher due to the lower sections of these tributary s having higher urban impervious zones then further up the catchment. Table 5 shows that the two main tributaries of the lower Yarra Rivers have a small influence on flow in the Yarra River. Table 5: Estimated Influence of Major Tributaries on Yarra River Flow Tributary % of Yarra Annual Flow Diamond Creek 4.8 Darebin Creek 3.9 The combination of the upstream observed time series and the limited influence of the tributaries on the Yarra flow, have reduced the sensitivity of this model. Unfortunately it currently has not been possible to calibrate upstream of any of the observed time series sites due to significant irrigation/water supply diversions and minimal gauged data. Further model builds may attempt to incorporate this, however for this study the lower Yarra catchment was the area of most interest. It is expected that calibration results for the Yarra River gauges will not be as good as with the observed time series, however to properly run climate change scenarios in the future the Upper Yarra section of the model will need to be incorporated 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 particularly 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 baseflows already being unreliable in that data, 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 in the baseflow and low to medium flow events. Predicted flows were generally higher than observe. At the majority of downstream gauges, predicted flows were approximately 10% greater then observed flows. Figure 9 and Figure 10 show predicted and observed annual flow volumes for 2 gauges on the Yarra River. Gauge is located above Yering Gorge, while is the most downstream gauge in the model. Annual volumes can vary greatly in the Yarra River, from over 1000 GL on a wet year to less then 200 GL on a dry year. The relationship between these gauges is difficult to see due to the Yering Gorge extraction from the Yarra River that is situated between these gauges. ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 15

27 Yarra River, Yarra Glen 1,100 1, Total Annual Flow (GL) Observed Predicted Figure 10: Annual Flow Volumes at Yarra River, Yarra Glen Yarra River Chandler Highway Total Annual Flow (GL) 1,400 1,300 1,200 1,100 1, Figure 11 - Annual flow volumes at Yarra River, Chandler Highway Observed Predicted The cause of the predicted flows being greater then 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 Yarra River catchment contains approximately 1480 diversion licences and 560 registered farm dams. Most of these diversions are not gauged and only total annual diversion license volumes are available. Therefore they were not included in the model. ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 16

28 Some sites have recently been put on telemetry; however data is only available for the past few years and was not used in this analysis. Farm dams and loses need to be further explored before being implemented. Future model builds may attempt to incorporate these three processes Scenario Testing Yering Gorge Pumping Scenarios The limited functionality of the flood harvesting tool meant that the extractions at the Yering Gorge pumping station did not match up with the observed extractions (Figure 12). This is due to the way the pumps are operated in practice. The flood harvesting tool cannot take into consideration the reservoir levels in sugarloaf which is an integral part of the operation of the Yering Gorge pumps. All the scenario does is extract water when it is above a certain stream threshold. As can be seen in Figure 13, all 4 scenarios pump the maximum yearly amount for most of the modelled period. The only years it does not reach the 130,000 ML/year maximum is during extremely dry years where river flows were extremely low. As expected the 150 ML/d flow requirement will extract more water then the 300 ML/d scenario observed extraction Scenario ML/d Daily Diversion Volume (ML/d) Figure 12: Daily extraction from Yering Gorge Pumping Station indicating that the flood harvesting tool did not accurately match extractions from the River. ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 17

29 130, ,000 Annual Diversion Volume (ML) 110, ,000 90,000 80,000 70,000 60, scenario 4 scenario 3 scenario 2 scenario 1 Figure 13: Annual diversion volumes for all 4 scenarios indicating that the maximum annual diversion volume was met in most years. The scenarios may still be useful when investigating the sensitivity of DO during low flows in the lower Yarra River. Further work may be done to improve the quality of these scenarios. This may involve a new Source Catchment plug-in to attempt to mimic the pump operations a bit more realistically. 4.0 Conclusions The objective of this project was to calibrate a whole of catchment model for the Yarra River. The 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 Development of the Yarra River catchment model was successful within the target area. Acceptable calibration results were obtained on the suitable gauges within the catchment focus area. Predicted runoff volumes in the focus area of the Yarra River were all within 10% of the observed flow. Improvements in gauged data from many of the tributaries are needed to improve the calibrations at low flows, however the use the observed time series in the Yarra River at Coldstream has minimised the impacts of these poorer calibrations especially in regards to baseflow. The Yering Gorge minimum flow requirement scenarios will be useful for testing the sensitivity of DO at low flows. However, they do not mimic what is done in the actual operation of the pumping station. Over the next 12 months a more detailed assessment of the range of node models available in Source Catchments will be undertaken to best represent current pumping operations. A new plug-in may be developed to improve the output of these scenarios. The intent for any future work is to look at modelling the rest of the catchment, including adding storage models to account for the larger reservoirs in the headwaters of the Yarra River, and applying some level of constituent modelling. Climate change scenarios will also be looked at in the future. ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 18

30 5.0 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. 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 Connell Wagner (2006) Toorourrong Reservoir Capacity Survey Report Reservoir Remedial Works. Connell Wagner. South Melbourne. 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 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: 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. 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 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: 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 Victorian Government. Melbourne. ewater CRC Yarra River Application Project Source Catchments Hydrology Calibration Report 19