C Rainfall RunoffModelling

Rainfall-runoff modelling across the Murray-Darling Basin
A report to the Australian Government from the CSIRO Murray-Darling Basin Sustainable Yields Project
Francis Chiew, Jai Vaze, Neil Viney, Phillip Jordan, Jean-Michel Perraud, LuZang, Jin Teng, Jorge Pena Arancibia, Robert Morden, Andrew Freebairn, Jenet Austin, Peter Hill, Chloe Wiesemfeld and Rachel Murphy
June 2008
Murray-Darling Basin Sustainable Yields Project acknowledgments The Murray-Darling Basin Sustainable Yields project is being undertaken by CSIRO under the Australian Government's Raising National Water Standards Program, administered by the National Water Commission. Important aspects of the work were undertaken by Sinclair Knight Merz; Resource & Environmental Management Pty Ltd; Department of Water and Energy (New South Wales); Department of Natural Resources and Water (Queensland); Murray-Darling Basin Commission; Department of Water, Land and Biodiversity Conservation (South Australia); Bureau of Rural Sciences; Salient Solutions Australia Pty Ltd; eWater Cooperative Research Centre; University of Melbourne; Webb, McKeown and Associates Pty Ltd; and several individual sub-contractors. Murray-Darling Basin Sustainable Yields Project disclaimers Derived from or contains data and/or software provided by the Organisations. The Organisations give no warranty in relation to the data and/or software they provided (including accuracy, reliability, completeness, currency or suitability) and accept no liability (including without limitation, liability in negligence) for any loss, damage or costs (including consequential damage) relating to any use or reliance on that data or software including any material derived from that data and software. Data must not be used for direct marketing or be used in breach of the privacy laws. Organisations include: Department of Water, Land and Biodiversity Conservation (South Australia), Department of Sustainability and Environment (Victoria), Department of Water and Energy (New South Wales), Department of Natural Resources and Water (Queensland), Murray-Darling Basin Commission. CSIRO advises that the information contained in this publication comprises general statements based on scientific research. The reader is advised and needs to be aware that such information may be incomplete or unable to be used in any specific situation. No reliance or actions must therefore be made on that information without seeking prior expert professional, scientific and technical advice. To the extent permitted by law, CSIRO (including its employees and consultants) excludes all liability to any person for any consequences, including but not limited to all losses, damages, costs, expenses and any other compensation, arising directly or indirectly from using this publication (in part or in whole) and any information or material contained in it. Data is assumed to be correct as received from the Organisations. Acknowledgements The authors wish to thank Prof Roger Grayson, etc who provided technical review of the report, and Becky Schmidt, CSIRO, for her copy-editing. Citation Chiew FHS, Vaze J, Viney NR, Jordan PW, Perraud J-M, Zhang L, Teng J, Young WJ, Penaarancibia J, Morden RA, Freebairn A, Austin J, Hill PI, Wiesenfeld CR and Murphy R (2008) Rainfall-runoff modelling across the Murray-Darling Basin. A report to the Australian Government from the CSIRO Murray-Darling Basin Sustainable Yields Project. CSIRO, Australia. 62pp. Publication Details Published by CSIRO 2008 all rights reserved. This work is copyright. Apart from any use as permitted under the Copyright Act 1968, no part may be reproduced by any process without prior written permission from CSIRO. ISSN 1835-095X
Preface
This is a report to the Australian Government from CSIRO. It is an output of the Murray-Darling Basin Sustainable Yields Project which assessed current and potential future water availability in 18 regions across the MurrayDarling Basin (MDB) considering climate change and other risks to water resources. The project was commissioned following the Murray-Darling Basin Water Summit convened by the Prime Minister of Australia in November 2006 to report progressively during the latter half of 2007. The reports for each of the 18 regions and for the entire MDB are supported by a series of technical reports detailing the modelling and assessment methods used in the project. This report is one of the supporting technical reports of the project. Project reports can be accessed at http://www.csiro.au/mdbsy. Project findings are expected to inform the establishment of a new sustainable diversion limit for surface and groundwater in the MDB one of the responsibilities of a new Murray-Darling Basin Authority in formulating a new Murray-Darling Basin Plan, as required under the Commonwealth Water Act 2007. These reforms are a component of the Australian Governments new national water plan Water for our Future. Amongst other objectives, the national water plan seeks to (i) address over-allocation in the MDB, helping to put it back on a sustainable track, significantly improving the health of rivers and wetlands of the MDB and bringing substantial benefits to irrigators and the community; and (ii) facilitate the modernisation of Australian irrigation, helping to put it on a more sustainable footing against the background of declining water resources.
Summary
This report is one in a series of technical reports from the CSIRO Murray-Darling Basin Sustainable Yields o o Project. This report describes the rainfall-runoff modelling for 0.05 x 0.05 grid cells (~ 5 km x 5 km) across the o Murray-Darling Basin (MDB) and presents the runoff estimates for the four modelling scenarios for the 0.05 x o 0.05 grids and for the 18 MDB regions. The analyses of rainfall and other climate variables are described in a companion report (Chiew et al., 2008). The key modelling results for the 18 MDB regions defined in the project are summarised in Appendix A.
Scenario A Historical climate (1895 to 2006) and current development

The mean annual rainfall and modelled runoff, averaged over 1895 to 2006 over the entire MDB, are 457 mm and 27.3 mm respectively. There is a clear eastwest rainfall gradient across the MDB, where rainfall is highest in the south-east (mean annual rainfall of more than 1500 mm) and along the eastern perimeter, and lowest in the west (less than 300 mm). The runoff gradient is much more pronounced than the rainfall gradient, with runoff in the south-east corner (mean annual runoff of more than 200 mm) and eastern perimeter (20 to 80 mm) being much higher than elsewhere in the MDB (less than 10 mm in the western half). In the northern MDB, most of the rainfall and runoff occurs in the summer half of the year, and in the southernmost MDB, most of the rainfall and runoff occurs in the winter half of the year. The runoff estimates in the southern and eastern MDB, where most of the runoff occurs, are relatively good because there are many gauged catchments there from which to estimate the model parameter values. The errors in the mean annual runoff estimated for the southern and eastern MDB are generally less than 50 percent o o for the 0.05 x 0.05 grids, and likely to be less than 10 percent when averaged over the MDB regions. There is less confidence in the runoff estimates in the dry central and western MDB because there are very few or no calibration catchments there from which to estimate the model parameter values.
CSIRO 2008
Scenario B Recent climate (1997 to 2006) and current development

The 1997 to 2006 mean annual runoff averaged over the MDB is 21.7 mm, about 21 percent lower than the 1895 to 2006 long-term mean. The biggest differences are in the southern half of the MDB, where the 1997 to 2006 runoff is more than 30 percent lower than the long-term mean, and up to 50 percent lower in the southernmost parts. Potter et al. (2008) provide a detailed analysis of recent rainfall and runoff characteristics across the MDB.
Scenario C Future climate (~2030) and current development

The future climate is used to assess the range of likely climate conditions around the year 2030. Forty-five future climate variants, each with 112 years of daily climate sequences, are used. The future climate variants come from scaling the 1895 to 2006 climate data to represent the ~2030 climate, based on analyses of 15 global climate models (GCMs) and three global warming scenarios from the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. The majority of the modelling results shows a decrease in mean annual runoff, particularly in the southern MDB where more than two-thirds of the results show a decrease in mean annual runoff. The best estimate or median indicates that the future mean annual runoff in the MDB in ~2030 relative to ~1990 will be lower, by 5 to 10 percent in the north-east and southern half, and by about 15 percent in the southernmost parts. Averaged across the entire MDB, the best estimate or median is a 9 percent decrease in mean annual runoff. There is considerable uncertainty in these estimates, with the extreme dry and extreme wet values in the northern half of the MDB ranging from a 30 percent decrease to a 30 percent increase in mean annual runoff. In the southern half of the MDB, the extreme estimates range from a 40 percent decrease to a 20 percent increase in mean annual runoff, and in the southernmost MDB, the extreme estimates range from a decrease in mean annual runoff of up to 50 percent to little change in mean annual runoff. Averaged over the entire MDB, the extreme estimates range from a 33 percent decrease to a 16 percent increase in mean annual runoff. The biggest uncertainty in Scenario C modelling is in the global warming projections and the GCM modelling of the impact of this global warming on rainfall in the MDB. The uncertainty in the rainfall-runoff modelling of the impact of climate change on runoff is small compared to the uncertainty in the climate change projections. The Scenario C modelling only considers the impact of changes in rainfall and potential evapotranspiration on runoff. The modelling does not take into account the potential effect of global warming and enhanced CO2 concentrations on forest water use. This impact could be significant, but it is difficult to estimate the net effect because of the compensating positive and negative impacts and the complex climate-biosphere-atmosphere interactions and feedbacks.
Scenario D Future climate (~2030) and future development (~2030)

Plantations can significantly affect local runoff, but for the Bureau of Rural Sciences projections of commercial forestry plantations assessed here, the impact on runoff averaged over an entire region is negligible. The impact of the projected increases in farm dams varies from zero to a 1.5 percent reduction in mean annual runoff averaged over 17 of the 18 MDB regions and about 3 percent reduction in mean annual runoff in the Eastern Mount Lofty Ranges region. After the uncertainty in the Scenario C climate change projections, the biggest uncertainty in Scenario D modelling is in the projections of future increases in commercial forestry plantations and farm dam development and the impact of these developments on runoff. The increase in farm dams is estimated by considering trends in historical farm dam growth and current policy controls. There is considerable uncertainty both as to how landholders will respond to development policies and how governments may set policies in the future.
CSIRO 2008
Table of Contents
1 2
2.1 2.2
Introduction................................................................................................................................1 Rainfall-runoff modelling method ...........................................................................................2

Rainfall-runoff modelling...............................................................................................................................................2 Climate scenarios.........................................................................................................................................................3
3
3.1 3.2 3.3 3.4 3.5
Summary of modelling results .................................................................................................5

Reporting regions, subcatchments and calibration catchments ....................................................................................5 Scenario A results (historical climate, recent development)..........................................................................................7 Scenario B results (recent climate, recent development) ............................................................................................ 11 Scenario C results (future climate, recent development)............................................................................................. 13 Scenario D results (future climate, future development) ............................................................................................. 28
4
4.1 4.2
Rainfall-runoff modelling for Scenario A ..............................................................................30

Rainfall-runoff models ................................................................................................................................................ 30 Model calibration and verification ............................................................................................................................... 32
5
5.1 5.2 5.3
Rainfall-runoff modelling for Scenario C ..............................................................................41

Modelling climate change impact on runoff................................................................................................................. 41 Global warming and forest water use ......................................................................................................................... 43 Future bushfire risk and impact on runoff ................................................................................................................... 44
6
6.1 6.2 6.3 6.4
Rainfall-runoff modelling for Scenario D...............................................................................48

Commercial forestry plantations ................................................................................................................................. 48 Farm dams................................................................................................................................................................. 49 Estimation of future farm dam development ............................................................................................................... 53 Results....................................................................................................................................................................... 56
7 8
References ...............................................................................................................................60 Appendix A: Summary of key modelling results .................................................................62
Tables
Table 5-1. Summary of broad assessment of impact of increased future bushfire risk on future runoff............................. 45 Table 6-1. Existing areas of commercial forestry plantations in the Murray-Darling Basin and the projected increases by 2030 ................................................................................................................................................................................. 49 Table 6-2. Existing farm dam storage capacity (GL), listed by data source, region and state ........................................... 53 Table 6-3. Summary of projected increases in farm dam storage capacity (~2030 relative to ~2005) in Murray-Darling Basin regions in New South Wales................................................................................................................................... 55
Figures
Figure 3-1. Map showing 18 reporting regions, subcatchments and calibration catchments............................................... 6 Figure 3-2. Mean annual rainfall, areal potential evapotranspiration and modelled runoff .................................................. 8 Figure 3-3. Mean summer (DJF) rainfall, areal potential evapotranspiration and modelled runoff....................................... 9 Figure 3-4. Mean winter (JJA) rainfall, areal potential evapotranspiration and modelled runoff ........................................ 10 Figure 3-5. Percent difference between 19972006 mean annual runoff and 18952006 long-term mean for 0.05o x 0.05o grid cells (left) and averaged over each of the 18 MDB regions (right) .............................................................................. 12 Figure 3-6. Absolute difference (in mm) between 19972006 mean annual runoff and 18952006 long-term mean for 0.05o x 0.05o grid cells (left) and averaged over each of the 18 MDB regions (right) ......................................................... 12 Figure 3-7. Percent change in mean annual runoff across the Murray-Darling Basin (~2030 relative to ~1990) from 15 global climate models under the medium global warming scenario ................................................................................... 14 Figure 3-8. Percent change in mean summer (DJF) runoff across the Murray-Darling Basin (~2030 relative to ~1990) from 15 global climate models under the medium global warming scenario .............................................................................. 15 Figure 3-9. Percent change in mean winter (JJA) runoff across the Murray-Darling Basin (~2030 relative to ~1990) from 15 global climate models under the medium global warming scenario .............................................................................. 16
CSIRO 2008
Figure 3-10. Number of rainfall-runoff modelling results (using projections from 15 global climate models) showing a decrease (or increase) in future mean annual, summer (DJF), and winter (JJA) runoff ..................................................... 17 Figure 3-11. Percent change in modelled mean annual runoff across the Murray-Darling Basin (~2030 relative to ~1990) for the best estimate or median and the extreme dry and extreme wet scenarios ............................................................. 18 Figure 3-12. Percent change in modelled mean summer (DJF) runoff across the Murray-Darling Basin (~2030 relative to ~1990) for the best estimate or median and the extreme dry and extreme wet scenarios ................................................. 19 Figure 3-13. Percent change in modelled mean winter (JJA) runoff across the Murray-Darling Basin (~2030 relative to ~1990) for the best estimate or median and the extreme dry and extreme wet scenarios ................................................. 20 Figure 3-14. Absolute change (in mm) in modelled annual runoff across the Murray-Darling Basin (~2030 relative to ~1990) for the best estimate or median and the extreme dry and extreme wet scenarios ................................................. 21 Figure 3-15. Absolute change (in mm) in modelled mean summer (DJF) runoff across the Murray-Darling Basin (~2030 relative to ~1990) for the best estimate or median and the extreme dry and extreme wet scenarios ................................. 22 Figure 3-16. Absolute change (in mm) in modelled mean winter (JJA) runoff across the Murray-Darling Basin (~2030 relative to ~1990) for the best estimate or median and the extreme dry and extreme wet scenarios ................................. 23 Figure 3-17. Percent change in modelled mean annual runoff for the 18 Murray-Darling Basin regions (~2030 relative to ~1990) for the best estimate or median and the extreme dry and extreme wet scenarios (results are obtained using climate change projections from a single global climate model run for the entire region, in contrast to Figure 3-11 where the changes are shown for 0.05o x 0.05o grids)....................................................................................................................... 24 Figure 3-18. Mean monthly rainfall and modelled runoff averaged over each of the 18 Murray-Darling Basin regions for the historical climate, with the extreme range for future climate shown in orange ............................................................. 25 Figure 3-19. Projected increases in commercial forestry plantations and farm dam storage capacity in the 18 MurrayDarling Basin regions (~2030 relative to ~2005) ............................................................................................................... 28 Figure 3-20. Percent change in modelled mean annual runoff for the 18 Murray-Darling Basin regions for the best estimate or median and the extreme dry and extreme wet scenarios (the impacts of both climate change and development are included, in contrast to Figure 3-17 where only impacts from climate change are shown)........................................... 29 Figure 4-1. Structure of SIMHYD rainfall-runoff model ..................................................................................................... 31 Figure 4-2. Structure of Sacramento rainfall-runoff model................................................................................................ 32 Figure 4-3. Summary of model calibration and verification results ................................................................................... 34 Figure 4-4. Typical plots comparing modelled and observed monthly runoffs and daily runoff characteristics .................. 35 Figure 4-5. Comparison of mean annual runoff estimated by SIMHYD and Sacramento models for the verification results with the observed runoff ................................................................................................................................................... 38 Figure 4-6. Mean annual runoff estimated by the SIMHYD and Sacramento models across the Murray-Darling Basin .... 38 Figure 4-7. Mean summer (DJF) runoff estimated by the SIMHYD and Sacramento models across the Murray-Darling Basin ................................................................................................................................................................................ 39 Figure 4-8. Mean winter (JJA) runoff estimated by the SIMHYD and Sacramento models across the Murray-Darling Basin ......................................................................................................................................................................................... 39 Figure 4-9. Summary of model verification results across the Murray-Darling Basin ........................................................ 40 Figure 5-1. Comparison of changes in runoff characteristics in ~2030 relative to ~1990 estimated by the SIMHYD and Sacramento models using climate change projections from the IPSL global climate model under the medium global warming scenario ............................................................................................................................................................. 42 Figure 5-2. Comparison of changes in runoff characteristics in ~2030 relative to ~1990 estimated by the SIMHYD and Sacramento models using climate change projections from the INMCM global climate model under the medium global warming scenario ............................................................................................................................................................. 42 Figure 5-3. Comparison of changes in runoff characteristics in ~2030 relative to ~1990 estimated by the SIMHYD and Sacramento models using climate change projections from the CCCMA T47 global climate model under the medium global warming scenario................................................................................................................................................... 43 Figure 5-4. The effect of increased CO2 on catchment water balance processes (adapted from Field et al., 1995) (upward arrow next to a process indicates an increase, downward arrow next to a process indicates a decrease)......................... 44 Figure 5-5. Mean annual cumulative forest fire danger index (FFDI) for the historical climate (19802006) and percent change by ~2030 (second lowest, median and second highest using future climate data informed by six global climate models) ............................................................................................................................................................................ 46 Figure 5-6. Relationship between forest area burnt and mean annual cumulative FFDI based on data from Victoria ....... 47 Figure 5-7. Average change in mean annual streamflow for different species following bushfires .................................... 47 Figure 6-1. Sources of data on existing farm dam storage capacity ................................................................................. 51 Figure 6-2. Spatial density of existing farm dam storage capacity.................................................................................... 52 Figure 6-3. Spatial density of projected increase in farm dam storage capacity (~2030 relative to ~2005) ....................... 55 Figure 6-4. Percent reduction in mean annual runoff due to the projected increase in farm dams (~2030 relative to ~2005) ......................................................................................................................................................................................... 58
CSIRO 2008
Introduction
This report is one in a series of technical reports from the CSIRO Murray-Darling Basin Sustainable Yields Project. The terms of reference for the project are to estimate current and future water availability in each catchment and aquifer in the Murray-Darling Basin (MDB) considering climate change and other risks and surface-groundwater interactions, and compare the estimated current and future water availability to that required to meet the current levels of extractive use. Results from the project were reported progressively for 18 contiguous regions across the entire MDB. The purpose of this report is to describe in more detail the rainfall-runoff modelling undertaken in the project. The main objective of the rainfall-runoff modelling is to use a consistent MDB-wide modelling approach to estimate daily runoff for 0.05o x 0.05o grids (~ 5 km x 5 km) across the MDB for four scenarios. The four scenarios are: Scenario A Historical climate (1895 to 2006) and current development Scenario B Recent climate (1997 to 2006) and current development Scenario C Future climate (~2030) and current development Scenario D Future climate (~2030) and future development (~2030).
The rainfall-runoff modelling method is described in Chapter 2 and the key modelling results are summarised in Chapter 3. The remaining chapters present more details on the modelling: calibration and assessment of the rainfall-runoff models in Chapter 4; application of the models to estimate climate change impact on runoff in Chapter 5; and modelling the impact of development (commercial forestry plantations and farm dams) on future runoff in Chapter 6.
CSIRO 2008
Rainfall-runoff modelling across the Murray-Darling Basin 1
2
2.1
Rainfall-runoff modelling method

Rainfall-runoff modelling
The adopted rainfall-runoff modelling method provides a consistent way of modelling historical runoff across the MurrayDarling Basin (MDB) and assessing the potential impacts of climate change and development on future runoff. The lumped conceptual rainfall-runoff model, SIMHYD, with a Muskingum routing method is used to estimate daily runoff o o o o for 0.05 x 0.05 grid cells (~ 5 km x 5 km) across the entire MDB for the four scenarios. The use of 0.05 x 0.05 grid cells allows a better representation of the spatial patterns and gradients in rainfall. The rainfall-runoff model is calibrated against 1975 to 2006 streamflow data from 183 small and medium size unregulated gauged catchments (50 to 2000 2 km ) across the MDB (referred to as calibration catchments). Although unregulated, streamflow in these catchments may reflect low levels of water diversion and will include the effects of historical land use change. The calibration period is a compromise between a shorter period that would better represent current development and a longer period that would better account for climatic variability. In the model calibration, the six parameters of SIMHYD are optimised to maximise an objective function that incorporates the Nash-Sutcliffe efficiency of monthly runoff and daily flow duration curve, together with a constraint to ensure that the total modelled runoff over the calibration period is within 5 percent of the total recorded runoff (see Chapter 4). The resulting optimised parameter values are therefore identical for all grid cells within a calibration catchment. The runoff for a non-calibration or ungauged subcatchment is modelled using optimised parameter values from the geographically closest calibration catchment, provided there is a calibration catchment within 250 km (subcatchments are defined by the river system models for the 18 MDB regions). Once again, the parameter values for each grid cell within a subcatchment are identical. For subcatchments more than 250 km from a calibration catchment, default parameter values are used. The default parameter values are identical across the entire MDB and are chosen to ensure a realistic runoff gradient across the drier parts of the MDB. The places these default values are used are therefore all areas of very low runoff. There is an exception for the Paroo and Warrego regions in the north-west MDB. In these regions, analyses of local long-term rainfall and runoff data justified the use of optimised parameter values from a single calibration catchment in the Paroo for the entire Paroo region and two calibration catchments in the Warrego for the Warrego region (CSIRO, 2007a; CSIRO 2007b; Young et al., 2006). As the parameter values come from calibration against streamflow from 50 to 2000 km catchments, the runoff defined here is different, and can be much higher than streamflow recorded over very large catchments where there can be significant transmission losses (particularly in the western and northwestern MDB). Almost all the catchments available for model calibration are in the higher runoff areas in the southern and eastern MDB. Runoff estimates are therefore generally good in the southern and eastern MDB and are comparatively poor elsewhere. The same set of parameter values are used to model runoff across the MDB for Scenarios A, B and C using 112 years of daily climate inputs described in Section 2.2. The future climate scenario simulations therefore do not take into account the effect of global warming and enhanced CO2 concentrations on forest water use. This effect could be significant, but it is difficult to estimate the net effect because of the compensating positive and negative impact and the complex climatebiosphere-atmosphere interactions and feedbacks (see Section 5.2 for more discussion). Bushfire risk is also likely to increase under the future climate scenario. In areas where bushfires occur, runoff would reduce significantly as forests regrow. However, the impact on runoff averaged over an entire region is unlikely to be significant (see Section 5.3 for more discussion and broad analysis). The projection of growth in commercial forestry plantations and farm dams and the modelling of the impact of these developments on future runoff are described in Chapter 6. The rainfall-runoff model SIMHYD is used because it is simple and has relatively few parameters and, for the purpose of this project, provides a consistent basis (that is automated and reproducible) for modelling historical runoff across the entire MDB and for assessing the potential impacts of climate change and development on future runoff. It is possible that in data-rich areas, specific calibration of SIMHYD or more complex rainfall-runoff models based on expert judgement and local knowledge, as carried out by some agencies, would lead to better model calibration for the specific modelling
2
2 Rainfall-runoff modelling across the Murray-Darling Basin
CSIRO 2008
objectives of the area. Chapter 4 describes in more detail the model calibration and assessment, and the comparison of SIMHYD with another commonly used rainfall-runoff model (the Sacramento model). The simulations from the two rainfall-runoff models are relatively similar in the context of the application here. The modelled runoff series are modified and used as inputs to drive the river system models to estimate the impact of climate change and development on water availability and water use across the 18 MDB regions. The modelled runoff series are not used directly as subcatchment inflows in the river system models as this would violate the calibrations of the river system models already undertaken by state agencies to different runoff series. Instead, the relative differences between the daily flow duration curves of the historical climate scenario and the other scenarios are used to modify the existing inflows series in the river system models. All the scenario inflow series for the river system modelling therefore have the same daily sequence, but different amounts.
2.2
Climate scenarios
Daily rainfall and potential evapotranspiration (PET) are required to run the rainfall-runoff models. The climate data for the hydrologic scenario modelling across the MDB are described in detail in Chiew et al. (2008). A brief summary is given here. The historical climate scenario (Scenario A) is the baseline against which other scenarios are compared. It is based on observed climate data from 1895 to 2006. The source of the historical climate data is the SILO Data Drill of the Queensland Department of Natural Resources and Water (http://www.nrw.qld.gov.au/silo and Jeffrey et al., 2001). The o o SILO Data Drill provides surfaces of daily rainfall and other climate data for 0.05 x 0.05 grid cells across Australia, interpolated from point measurements made by the Australian Bureau of Meteorology. Areal potential evapotranspiration is calculated from the SILO climate surface using Mortons wet environment evapotranspiration algorithms (http://www.bom.gov.au/climate/averages; Morton, 1983; Chiew and Leahy, 2003). The recent climate scenario (Scenario B) is used to assess future water availability should the climate in the future prove to be similar to that of the last ten years. Climate data for 1997 to 2006 are used to generate stochastic replicates of 112year daily climate sequences. The replicate which produces a mean annual runoff closest to that observed in 1997 to 2006 is selected to define this scenario. The future climate scenario (Scenario C) is used to assess the range of likely climate conditions around the year 2030. Forty-five future climate variants, each with 112 years of daily climate sequences, are used. The future climate variants come from scaling the 1895 to 2006 climate data to represent ~2030 climate, based on analyses of 15 global climate models (GCMs) and three global warming scenarios (the 15 GCMs used are listed in Chiew et al., 2008). The scenario variants are derived from the latest modelling for the Intergovernmental Panel on Climate Change Fourth Assessment Report (IPCC, 2007; CSIRO and Australian Bureau of Meteorology, 2007). The method used here takes into account two types of uncertainties. The first uncertainty is in the global warming projection, due to the uncertainties associated with projecting greenhouse gas emissions and predicting how sensitive the global climate is to greenhouse gas concentrations. The second uncertainty is in GCM modelling of local climate in the MDB. Results from each GCM are analysed separately to estimate the percent change per degree global warming in rainfall and other climate variables required to calculate PET. The percent change per degree global warming is then multiplied by the temperature change projected for high, medium and low global warming. The result is the percent change in the climate variables expected in ~2030 relative to ~1990 under high, medium and low global warming scenarios. As the future climate series (Scenario C) is obtained by scaling the historical daily climate series from 1895 to 2006 (Scenario A), the daily climate series for Scenarios A and C have the same length of data (112 years) and the same sequence of daily climate. Scenario C is therefore not a forecast climate at 2030, but a 112-year daily climate series based on 1895 to 2006 data adjusted to match projected global temperatures at ~2030 relative to ~1990. The method used to obtain the future climate series also takes into account different changes in each of the four seasons as well as changes in the daily rainfall distribution. Considering changes in the daily rainfall distribution is important because many GCMs indicate that future extreme rainfall is likely to be more intense, even in some regions where a decrease in mean seasonal or annual rainfall is projected. As the high rainfall events generate large runoff, the use of
CSIRO 2008
traditional methods that assume the entire rainfall distribution to change in the same way would lead to an underestimation of total runoff.
CSIRO 2008
Summary of modelling results
This chapter summarises the key modelling results, in particular the runoff estimates for the different scenarios. Chiew et al. (2008) provide a similar presentation for rainfall and other climate variables.
3.1
Reporting regions, subcatchments and calibration catchments
Figure 3-1 shows the boundaries of the 18 regions defined for the CSIRO Murray-Darling Basin Sustainable Yields Project, the subcatchments defined for the river system modelling in the 18 regions, and the gauged catchments used to calibrate the rainfall-runoff models. The map also highlights where default model parameter values are used to model areas where the closest calibration catchment is more than 250 km away. Almost all the gauged catchments available for model calibration are in the higher runoff areas in the southern and eastern Murray-Darling Basin (MDB). Runoff estimates are therefore generally good in the higher runoff areas but comparatively poor elsewhere.
CSIRO 2008
Figure 3-1. Map showing 18 reporting regions, subcatchments and calibration catchments
CSIRO 2008
3.2
Scenario A results (historical climate, recent development)
Figure 3-2 shows the mean annual rainfall, areal potential evapotranspiration (APET) and modelled runoff, averaged over 1895 to 2006. Figures 3-3 and 3-4 show this same information for summer (December-January-February) and winter (June-July-August), respectively. The mean annual rainfall, APET and runoff averaged over the entire MDB are 457 mm, 1443 mm and 27.3 mm, respectively. There is a clear eastwest rainfall gradient across the MDB, where rainfall is highest in the south-east (mean annual rainfall of more than 1500 mm) and along the eastern perimeter, and lowest in the west (less than 300 mm). The runoff gradient is much more pronounced than the rainfall gradient, with runoff in the south-east corner (mean annual runoff of more than 200 mm) and eastern perimeter (20 to 80 mm) being much higher than elsewhere in the MDB (less than 10 mm in the western half). In the northern MDB, most of the rainfall and runoff occurs in the summer half of the year, and in the southernmost MDB, most of the rainfall and runoff occurs in the winter half of the year (see Figure 3-14).
CSIRO 2008
Figure 3-2. Mean annual rainfall, areal potential evapotranspiration and modelled runoff
CSIRO 2008
Figure 3-3. Mean summer (DJF) rainfall, areal potential evapotranspiration and modelled runoff
CSIRO 2008
Figure 3-4. Mean winter (JJA) rainfall, areal potential evapotranspiration and modelled runoff
CSIRO 2008
3.3
Scenario B results (recent climate, recent development)
Figures 3-5 and 3-6 show the percent difference, and absolute difference (in mm), respectively, between the modelled mean annual runoff averaged over the past ten years (1997 to 2006) compared to the 1895 to 2006 long-term mean. The 1997 to 2006 mean annual runoff averaged over the MDB is 21.7 mm, about 21 percent lower than the 1895 to 2006 long-term mean. The biggest differences are in the southern half of the MDB, where the 1997 to 2006 runoff is more than 30 percent lower than the long-term mean, and up to 50 percent lower in the southernmost parts. The difference between the 1997 to 2006 runoff and the 1895 to 1996 runoff in the eight southernmost regions (Murray, Murrumbidgee, Eastern Mount Lofty Ranges, Wimmera, Loddon-Avoca, Campaspe, Goulburn-Broken and Ovens) are statistically significant at = 0.2 (with the Student-t and Rank-Sum tests). The difference between the 1997 to 2006 and 1895 to 2006 runoff averaged over the MDB is dominated by the values in south-east MDB where runoff is highest and where the biggest differences occur (see Figure 3-6). Potter et al. (2008) provide a detailed analysis of recent rainfall and runoff and a discussion of annual rainfall and runoff characteristics across the MDB.
CSIRO 2008
Figure 3-5. Percent difference between 19972006 mean annual runoff and 18952006 long-term mean for 0.05o x 0.05o grid cells (left) and averaged over each of the 18 MDB regions (right)
Figure 3-6. Absolute difference (in mm) between 19972006 mean annual runoff and 18952006 long-term mean for 0.05o x 0.05o grid cells (left) and averaged over each of the 18 MDB regions (right)
CSIRO 2008
3.4
Scenario C results (future climate, recent development)
Figures 3-7, 3-8 and 3-9 show the percent change in mean annual, summer, and winter runoff, respectively, for ~2030 relative to ~1990. These results were obtained from the rainfall-runoff modelling using climate change projections from the 15 global climate models (GCMs) under the medium global warming scenario (see Chiew et al. (2008) for description of the GCMs). Figure 3-10 shows the number of GCMs that indicate a decrease (or increase) in mean annual, summer, and winter runoff. The results indicate that the potential changes in runoff as a result of global warming can be very significant. However, there can be considerable differences in the runoff modelling results using climate change projections from the different GCMs. Nevertheless, the majority of the results show a decrease in mean annual runoff, particularly in the southern MDB where more than two-thirds of the results show a decrease in mean annual runoff (Figures 3-7 and 3-10). The majority of the results indicate that the future summer runoff will increase except in the southernmost MDB (Figures 3-8 and 3-10). The results also indicate that future winter runoff is likely to be lower across the MDB, with more than twothirds of the results showing a decrease in winter runoff (Figures 3-9 and 3-10). As most of the runoff in the southernmost MDB occurs in winter, the decrease in winter runoff translates to a significant decrease in mean annual runoff there. The seasonal scaling factors for the 45 future climate variants are obtained by multiplying the percent change per degree global warming (obtained from the 15 GCMs) by the temperature change projected under the low, medium and high global warming scenarios. Thus, the driest and wettest variants will come from the high global warming scenario. The best estimate or median of the change in mean annual, summer, and winter runoff and the extreme range of changes across the MDB are shown as percent changes in Figures 3-11, 3-12 and 3-13, and as changes in runoff amounts (in mm) in Figures 3-14, 3-15 and 3-16. For the best estimate or median, the median result from the medium global warming scenario is used. For the extreme dry estimate, the second driest result from the high global warming scenario is used. For the extreme wet estimate, the second wettest result from the high global warming scenario is used. The second th th driest and second wettest results are used because they represent about the 10 and 90 percentile results. The best estimate or median indicates that the future mean annual runoff in the MDB in ~2030 relative to ~1990 will be lower, by 5 to 10 percent in the north-east and southern half, and by about 15 percent in the southernmost parts. Averaged across the entire MDB, the best estimate or median is a 9 percent decrease in mean annual runoff. The change in future runoff averaged over the entire MDB is dominated by the change in runoff in south-east and eastern perimeter of the MDB where most of the runoff occurs. There is considerable uncertainty in the estimates, with the extreme dry and extreme wet estimates in the northern half of the MDB ranging from a 30 percent decrease to a 30 percent increase in mean annual runoff. In the southern half of the MDB, the extreme estimates range from a 40 percent decrease to a 20 percent increase in mean annual runoff, and in the southernmost MDB, the extreme estimates range from a decrease in mean annual runoff of up to 50 percent to little change in mean annual runoff (Figure 3-11). Averaged across the MDB, the extreme estimates range from a 33 percent decrease to a 16 percent increase in mean annual runoff. Figure 3-17 shows the best estimate or median of the change in mean annual runoff and the extreme range of changes for the 18 MDB regions. Figure 3-17 is similar to Figure 3-11, but unlike Figure 3-11 where the changes are shown for o o 0.05 x 0.05 grid cells across the MDB, the results in Figure 3-17 are obtained using climate change projections from a single GCM run for the entire region (one for each of the extreme dry, median and extreme wet scenarios). These results are used for the whole-of-region modelling presented in the MDBSY regional reports. Figure 3-18 shows the mean monthly rainfall and modelled runoff averaged over the 18 MDB regions, and the extreme range for the ~2030 climate. The extreme range is determined separately for each month from the high global warming scenario, with the second driest and the second wettest monthly result defining the lower and the upper bound, respectively. In the northern regions, most of the rainfall and runoff occurs in the summer half of the year, and in the southernmost regions, most of the rainfall and runoff occurs in the winter half of the year. The plots also highlight the considerable differences between the climate change projections from the GCMs and the modelled runoff results, particularly in the northern regions. In the southernmost regions (Eastern Mount Lofty Ranges, Wimmera, Loddon-Avoca, Campaspe, Goulburn-Broken and Ovens), almost all GCMs predict a decrease in winter rainfall which translates to an even bigger percent decrease in winter runoff when most the runoff in the region occurs.
CSIRO 2008
Figure 3-7. Percent change in mean annual runoff across the Murray-Darling Basin (~2030 relative to ~1990) from 15 global climate models under the medium global warming scenario
CSIRO 2008
Figure 3-8. Percent change in mean summer (DJF) runoff across the Murray-Darling Basin (~2030 relative to ~1990) from 15 global climate models under the medium global warming scenario
CSIRO 2008
Figure 3-9. Percent change in mean winter (JJA) runoff across the Murray-Darling Basin (~2030 relative to ~1990) from 15 global climate models under the medium global warming scenario
CSIRO 2008
Figure 3-10. Number of rainfall-runoff modelling results (using projections from 15 global climate models) showing a decrease (or increase) in future mean annual, summer (DJF), and winter (JJA) runoff
CSIRO 2008
Figure 3-11. Percent change in modelled mean annual runoff across the Murray-Darling Basin (~2030 relative to ~1990) for the best estimate or median and the extreme dry and extreme wet scenarios
CSIRO 2008
Figure 3-12. Percent change in modelled mean summer (DJF) runoff across the Murray-Darling Basin (~2030 relative to ~1990) for the best estimate or median and the extreme dry and extreme wet scenarios
CSIRO 2008
Figure 3-13. Percent change in modelled mean winter (JJA) runoff across the Murray-Darling Basin (~2030 relative to ~1990) for the best estimate or median and the extreme dry and extreme wet scenarios
CSIRO 2008
Figure 3-14. Absolute change (in mm) in modelled annual runoff across the Murray-Darling Basin (~2030 relative to ~1990) for the best estimate or median and the extreme dry and extreme wet scenarios
CSIRO 2008
Figure 3-15. Absolute change (in mm) in modelled mean summer (DJF) runoff across the Murray-Darling Basin (~2030 relative to ~1990) for the best estimate or median and the extreme dry and extreme wet scenarios
CSIRO 2008
Figure 3-16. Absolute change (in mm) in modelled mean winter (JJA) runoff across the Murray-Darling Basin (~2030 relative to ~1990) for the best estimate or median and the extreme dry and extreme wet scenarios
CSIRO 2008
Figure 3-17. Percent change in modelled mean annual runoff for the 18 Murray-Darling Basin regions (~2030 relative to ~1990) for the best estimate or median and the extreme dry and extreme wet scenarios (results are obtained using climate change projections from a single global climate model run for the entire region, in contrast to Figure 3-11 where the changes are shown for 0.05o x 0.05o grids)
CSIRO 2008
Mean monthly rainfall (mm)
60
40
Mean monthly runoff (mm)

J F M A M J J A S O N D
Paroo
20
0 J F M A M J J A S O N D
80 60 40 20 0 J F M A M J J A S O N D
Warrego
100 80 60 40 20 0 J F M A M J J A S O N D
Condamine-Balonne
100 80 60 40 20 0 J F M A M J J A S O N D
Moonie
120 100 80 60 40 20 0 J F M A M J J A S O N D
8 6 4 2 0 J F M A M J J A S O N D
Border
100 80 60 40 20 0 J F M A M J J A S O N D
12 9 6 3 0 J F M A M J J A S O N D
Gwydir
Figure 3-18. Mean monthly rainfall and modelled runoff averaged over each of the 18 Murray-Darling Basin regions for the historical climate, with the extreme range for future climate shown in orange
CSIRO 2008
100 80 60 40 20 0 J F M A M J J A S O N D
8 6 4 2 0 J F M A M J J A S O N D
Namoi
100 80 60 40 20 0 J F M A M J J A S O N D
8 6 4 2 0 J F M A M J J A S O N D
Macquarie-Castlereagh
60

2.0 1.5 1.0 0.5 0.0 J F M A M J J A S O N D
40
Barwon-Darling
20
80 60 40 20 0 J F M A M J J A S O N D
Lachlan
80

10 8 6 4 2 0 J F M A M J J A S O N D
60 40 20 0
Murrumbidgee
60

4 3 2 1 0 J F M A M J J A S O N D
40
Murray
20
Figure 3-18. (continued) Mean monthly rainfall and modelled runoff averaged over each of the 18 Murray-Darling Basin regions for the historical climate, with the extreme range for future climate shown in orange
CSIRO 2008
150
100

60
40
Ovens
50
20
120 100 80 60 40 20 0 J F M A M J J A S O N D
40 30 20 10 0 J F M A M J J A S O N D
Goulburn-Broken
100 80 60 40 20 0 J F M A M J J A S O N D
20 15 10 5 0 J F M A M J J A S O N D
Campaspe
60
40

Loddon-Avoca
20
60

4 3 2 1 0 J F M A M J J A S O N D
40
Wimmera
20
80

10 8 6 4 2 0 J F M A M J J A S O N D
60 40 20 0
Eastern Mount Lofty Ranges
Figure 3-18. (continued) Mean monthly rainfall and modelled runoff averaged over each of the 18 Murray-Darling Basin regions for the historical climate, with the extreme range for future climate shown in orange
CSIRO 2008
3.5
Scenario D results (future climate, future development)
Figure 3-19 shows the projected increases in commercial forestry plantations and farm dam storage capacities by ~2030 relative to ~2005. Commercial forestry plantations are projected to increase significantly in only three of the 18 MDB regions, but they have negligible impact on the mean annual runoff averaged over a region. The projected increase in farm dams in the Eastern Mount Lofty Ranges will reduce mean annual runoff there by about 3 percent. In New South Wales and Victoria, the projected increases in farm dams will reduce mean annual runoff over a region by 0.5 to 1.5 percent, a relatively small impact compared to that from climate change. The projected increase in farm dams has negligible impact on future runoff in the Queensland regions. There is considerable uncertainty in the future projections of commercial forestry plantations and farm dam storage capacity, and the uncertainty in the projections and the resulting impact on runoff are discussed in more detail in Chapter 6. Figure 3-20 shows the impact of both climate change and development on future runoff for the best estimate or median and the extreme dry and extreme wet scenarios. Because the impact of development on runoff is small compared to the impact from climate change, the plots in Figure 3-20 are almost identical to the plots in Figure 3-17 which show only the climate change impact on future runoff.
Figure 3-19. Projected increases in commercial forestry plantations and farm dam storage capacity in the 18 Murray-Darling Basin regions (~2030 relative to ~2005)
CSIRO 2008
Figure 3-20. Percent change in modelled mean annual runoff for the 18 Murray-Darling Basin regions for the best estimate or median and the extreme dry and extreme wet scenarios (the impacts of both climate change and development are included, in contrast to Figure 3-17 where only impacts from climate change are shown)
CSIRO 2008
Rainfall-runoff modelling for Scenario A
The purpose of Chapter 4 and Section 5.1 is to assess and compare the performance of SIMHYD, the rainfall-runoff model used for the CSIRO Murray-Darling Basin Sustainable Yields Project with another commonly used rainfall-runoff model, Sacramento model, and discuss the implications of the choice of rainfall-runoff model for this project.
4.1
Rainfall-runoff models
SIMHYD is a simple lumped conceptual daily rainfall-runoff model with seven parameters (Chiew et al., 2002). Figure 4-1 shows the model structure of SIMHYD and the equations used to model the rainfall-runoff processes. SIMHYD has been used successfully across Australia for various applications, including the estimation of runoff in the National Land and Water Resources Audit (Peel et al., 2002), the estimation of climate change impact on runoff (Chiew and McMahon, 2002), and in various regionalisation studies (Chiew and Siriwardena, 2005). A Muskingum routing algorithm with two parameters (KMUSK and XMUSK), as described in Tan et al. (2005), is used to rout the daily runoff simulated by SIMHYD to the catchment outlet. For the application here, XMUSK is set to 0 (therefore routing with a linear storage), and the two relatively insensitive infiltration capacity parameters, COEFF and SQ (see Figure 4-1), are set to 150 and 2 respectively. There are therefore six parameters in SIMHYD that require optimisation in the application here. The Sacramento model is also a lumped conceptual daily rainfall-runoff model (Burnash et al., 1973), but it is considerably more complex than SIMHYD. Figure 4-2 shows the structure of the Sacramento model. The Sacramento model has been used widely, in particular as part of the river system model implementations in New South Wales and Queensland and for flow forecasting worldwide. The Sacramento model has 17 parameters, but in the application here, only 13 parameters are optimised (ADIMP, LZFPM, LZFSM, LZPK, LZSK, LZTWM, PFREE, REXP, SARVA, IZFWM, UZK, UZTWM, ZPERC) plus one unit hydrograph parameter, with the other four parameters set to default values (PCTIM=0, RSERV=0.3, SIDE=0, SSOUT=0).
CSIRO 2008
Figure 4-1. Structure of SIMHYD rainfall-runoff model
CSIRO 2008
Figure 4-2. Structure of Sacramento rainfall-runoff model
4.2
Model calibration and verification
The daily rainfall and APET data required to run the two rainfall-runoff models have been described in Section 2.2. The models are calibrated and verified using 1975 to 2006 observed streamflow data from 183 small and medium sized unregulated gauged catchments (50 to 2000 km ) (Figure 3-1). The models are run for an extra year before the o o calibration period to allow the model stores to reach the appropriate levels. The modelling is carried out for 0.05 x 0.05 grid cells to allow a better representation of the spatial patterns and gradients in rainfall. The same set of parameter o o values are used for all 0.05 x 0.05 grid cells within a catchment. The objectives of the modelling are to estimate, as reliably as possible: the mean annual runoff, the monthly runoff series, and the daily runoff characteristics. The modelling exercise over the 183 catchments indicated that calibrating the models to reproduce the daily runoff series led to monthly runoff simulations that are almost equally as good as calibrating the models to reproduce the monthly runoff series directly. For this reason, the models are calibrated to optimise the Nash-Sutcliffe efficiency (Nash and Sutcliffe, 1970) of daily runoff, with a volumetric constraint used to ensure that the total modelled runoff is within 5 percent of the total observed runoff. The Nash-Sutcliffe efficiency (E) is defined as follows:
2
(OBS
n
E =
i =1
OBS (MODi OBSi )

2
(OBS
n i =1
i =1
OBS
CSIRO 2008
where MOD is modelled runoff, OBS is observed runoff, observations.
OBS
is mean of all the observed runoff, and n is the number of
To calculate Edaily, all the modelled and observed daily runoff series are compared directly. The Nash-Sutcliffe efficiency (E) expresses the proportion of variance of the observed runoff that can be accounted for by the model and provides a direct measure of the ability of the model to reproduce the observed runoff with Edaily = 1.0 indicating that all the modelled daily runoffs are the same as all the observed daily runoffs. As explained earlier, the model calibration to optimise Edaily led to Emonthly values that are similar to when the models are calibrated to optimise Emonthly directly. As one of the main objectives of the modelling is to estimate as reliably as possible the monthly runoff series, subsequent results show comparisons of the Emonthly values (and the modelled and observed monthly runoffs) rather than the Edaily values. The Edfdc values are also shown in subsequent assessments to reflect the ability of the models to reproduce the observed daily runoff characteristics. To calculate Edfdc, all the modelled daily runoff values and all the observed daily runoff values are sorted from highest to lowest, and the runoff values at the same ranks are compared. Figure 4-3 summarises the Emonthly and Edfdc values from the model calibrations for the 183 catchments. The results indicate that both the SIMHYD and Sacramento models can reproduce reasonably satisfactorily the observed monthly runoff series (Emonthly values greater than 0.6 in more than 90 percent of the catchments and greater than 0.8 in more than 70 percent of the catchments) and the daily runoff characteristics (Edfdc values greater than 0.8 in more than 90 percent of the catchments). The calibration results for the Sacramento model are slightly better than the results for the SIMHYD model, mainly because there are more parameters in the Sacramento model. Figure 4-4 shows typical comparisons of the modelled and observed monthly runoffs and the modelled and observed daily flow duration curves. The calibration to optimise Nash-Sutcliffe E means that more importance is placed on the simulation of high runoff, and therefore the modelling of the medium and high runoffs are considerably better than the simulation of low runoff. Nevertheless, an optimisation to reduce overall error variance will result in some underestimation of high runoff and overestimation of low runoff. This is evident in some of the scatter plots comparing the modelled and observed monthly runoffs and the daily flow duration curves comparing the modelled and observed daily runoff characteristics. The discernible disagreement between the modelled and observed daily runoff characteristics only occurs for runoff that is exceeded less than 0.1 or 1 percent of the time, but is accentuated in the plots because of the linear scale on the y-axis and normal probability scale on the x-axis. In any case, the volumetric constraint used in the model calibration ensures that the total modelled runoff is always within 5 percent of the total observed runoff. Almost all the calibration catchments are in the south-east corner and eastern perimeter of the MDB (Figure 3-1). The calibration catchments cover less than 5 percent of the south-east corner and eastern perimeter of the MDB, and less than 1 percent of the entire MDB. As the runoff for 0.05 x 0.05 grids in subcatchments across the MDB are estimated using optimised parameter values from the geographically closest calibration catchment, the ability of the SIMHYD and Sacramento models to estimate runoff for ungauged catchments is also assessed here. This is done by using the optimised parameter values from the nearest calibration catchment to estimate runoff for each of the catchments and comparing the modelled runoff with the observed runoff. This is called model verification here, to distinguish it from the model calibration discussed above. The model verification results are also summarised in Figure 4-3. As expected, these results are poorer than the model calibration results. However, the Emonthly values in the verification results are generally only less than 0.1 lower than the values in the calibration results in about half the catchments, and less than 0.2 lower in very few catchments (Figure 4-3). The Emonthly values in the verification results for the Sacramento and SIMHYD models are relatively similar. The verification results for the Sacramento model are slightly better than the results for the SIMHYD model in more than half of the catchments (catchments with higher Emonthly values) and slightly poorer than the SIMHYD model in the other catchments. The verification results also show that the errors in the mean annual runoff estimated by the SIMHYD and Sacramento models are less than 20 percent in more than half the catchments and less than 50 percent in more than 85 percent of the catchments, with no bias towards an underestimation or overestimation (Figures 4-3 and 4-5). Figures 4-6, 4-7 and 4o o 8 show the mean annual, summer, and winter runoff estimated by the SIMHYD and Sacramento models for 0.05 x 0.05 grids across the MDB. The plots in Figures 4-5 to 4-8 indicate that the mean runoffs estimated by the SIMHYD and Sacramento models are very similar.
CSIRO 2008 Rainfall-runoff modelling across the Murray-Darling Basin 33
o o
Figure 4-9 shows the spatial Emonthly values and the errors in the mean annual runoff estimates for the verification results for the SIMHYD and Sacramento models across the MDB. As the plots do not show any clear regional differences, the broad conclusions above apply across the MDB. The results therefore indicate that the mean annual runoffs estimated for ungauged subcatchments using optimised parameter values from a nearby calibration catchment have an error of less than 20 percent in more than half the subcatchments and less than 50 percent in most of the subcatchments. As there is little bias towards an underestimation or overestimation, the errors in the mean annual runoff estimates averaged over each of the 18 MDB regions will be considerably smaller because of compensating positive and negative errors. As most of the calibration catchments are in the higher runoff areas in the southern and eastern MDB, the runoff estimates there are generally good. There is less confidence in the runoff estimates in the dry central and western MDB because there are very few or no calibration catchments there from which to estimate the model parameter values. As the simulations from the SIMHYD and Sacramento models are similar, particularly in the mean seasonal and annual runoffs, SIMHYD is used for the modelling in this project because it is simpler and it has been used extensively for climate change impact on runoff studies. It is likely that the use of better regionalisation and parameterisation methods can improve the runoff modelling, particularly the daily runoff estimates (e.g., regionalisation based on catchment similarities, weighted ensemble modelling (of simulations from SIMHYD, Sacramento and other models), and constraining model calibrations with other data types like remotely-sensed evapotranspiration or soil moisture). A system-wide optimisation together with the river system models that also use gauged streamflow data in main river channels can also reduce the uncertainty in the runoff estimates.
Emonthly
1.0
SIMHYD verification
1. 0 0. 8
0.8
0. 6
Sacramento verification SIMHYD calibration

0 20 40 60 80 10 12 14 16 18 0 0 0 0calibration 0 Sacramento
0.6
0. 4
0. 2
0.4
0. 0
0.2
0.0 0 20 40 60 80 100
% catchments where Emonthly is exceeded
Edfdc
1.0 60 40 20 0 0 -20 0.2 -40 0.0 0 20 40 60 80 100 -60
Bias (% difference between modelled and observed m ean annual runoff )
0.8
0.6
0.4
20
40
60
80
100
% catchm ents where Edfdc is exceeded
% catchments where bias is exceeded
Figure 4-3. Summary of model calibration and verification results
CSIRO 2008
Figure 4-4. Typical plots comparing modelled and observed monthly runoffs and daily runoff characteristics
CSIRO 2008
Figure 4-4. (continued) Typical plots comparing modelled and observed monthly runoffs and daily runoff characteristics
CSIRO 2008
Figure 4-4. (continued) Typical plots comparing modelled and observed monthly runoffs and daily runoff characteristics
CSIRO 2008
1000
1000
800
Mean annual Sacramento runoff (mm)

0 200 400 600 800 1000
Mean annual SIMHYD runoff (mm)
800
600
600
400
400
200
200
0 0 200 400 600 800 1000
Mean annual observed runoff (m m)
Mean annual observed runoff (mm )
Figure 4-5. Comparison of mean annual runoff estimated by SIMHYD and Sacramento models for the verification results with the observed runoff
Figure 4-6. Mean annual runoff estimated by the SIMHYD and Sacramento models across the Murray-Darling Basin
CSIRO 2008
Figure 4-7. Mean summer (DJF) runoff estimated by the SIMHYD and Sacramento models across the Murray-Darling Basin
Figure 4-8. Mean winter (JJA) runoff estimated by the SIMHYD and Sacramento models across the Murray-Darling Basin
CSIRO 2008
Figure 4-9. Summary of model verification results across the Murray-Darling Basin
CSIRO 2008
Rainfall-runoff modelling for Scenario C
There are three sections in this chapter. Section 5.1 compares the climate change impact on runoff predicted by the SIMHYD and Sacramento rainfall-runoff models. Section 5.2 discusses the potential impact of global warming and enhanced CO2 concentrations on forest water use, which is not modelled explicitly in the CSIRO Murray-Darling Basin Sustainable Yields Project. Section 5.3 presents a broad analysis of the potential increase in future bushfire risk and its impact on runoff.
5.1
Modelling climate change impact on runoff
Figures 5-1, 5-2 and 5-3 compare the changes in runoff characteristics in ~2030 relative to ~1990 estimated by SIMHYD and Sacramento models using climate change projections from three different GCMs under the medium global warming scenario. The method used to obtain the 112-year daily rainfall and APET series for a ~2030 climate has been described in Section 2.2. The comparisons in Figures 5-1 to 5-3 are for the 183 calibration catchments, where the same optimised parameter values are used to estimate the historical and future runoffs. The three GCMs used here are the IPSL, INMCM and CCCMA T47 GCMs (see Chiew et al. (2008) for description of the GCMs). These three GCMs are chosen because the SIMHYD modelling results using the climate change projections from these GCMs show the second driest, median and second wettest change in future runoff (averaged over the entire MDB) among the results from all 15 GCMs under the medium global warming scenario (see Figure 3-6). The impact of climate change on runoff estimated by the SIMHYD and Sacramento rainfall-runoff models shows the strongest agreement for climate change projections from the wet GCM, CCCMA T47 (Figure 5-3), where the changes in all the runoff variables estimated by the two models differ by less than 5 percent in almost all the 183 catchments. For the other two GCM climate change projections, where reduction in future runoff is estimated, the impact on runoff estimated by SIMHYD is smaller than that estimated by Sacramento (Figures 5-1 and 5-2). The difference in the modelled climate change impact on mean annual runoff and mean winter runoff estimated by the two models is 5 to 10 percent in the majority of the catchments. This difference is significant, but is relatively small compared to the disagreement in the rainfall projections for the MDB from different GCMs. For example, the difference between the median modelled change in the mean annual, summer and winter runoff over the 183 catchments using climate change projections from the dry and wet GCMs (IPSL and CCCMA T47) is about 30 to 40 percent. The biggest disagreement between the SIMHYD and Sacramento models is in the simulation of climate change impact on mean summer runoff. For the medium GCM (INMCM), the median difference from the modelling for the 183 catchments is 7 percent, with differences greater than 17 percent in 10 percent of the catchments (Figure 5-1). For the dry GCM (IPSL), the median difference from the modelling for the 183 catchments is 12 percent, with differences greater than 24 percent in 10 percent of the catchments (Figure 5-2). The difference in the SIMHYD and Sacramento modelling results is most likely due to the different methods used to simulate actual evapotranspiration (from soil wetness and PET), which explains the bigger differences in summer (see Chiew (2006) for more details on estimating hydrologic sensitivity of climate). The disagreements between the modelled climate change impacts on extreme runoff (runoff that is exceeded less than 1 percent of the time) and low runoff (number of days with runoff below 0.1 mm) estimated by SIMHYD and Sacramento are similar to that for mean annual and mean winter runoff.
CSIRO 2008
Mean annual runoff

0 20 0 -20 -20 -40 -60 -60 -60 -80 -80
Mean summer runoff

0
Mean winter runoff
-20
-40
-40
% change SIMHYD
-40
-20
-60
-40
-20
20
-60 -60
-40
-20
Daily runoff that is exceeded less than 1% of the time
100 80
Number of days with daily runoff less than 0.1 mm
-20 60 -40 40 20 -60 -60 0 -40 -20 0 0 20 40 60 80 100
% change Sacramento
Figure 5-1. Comparison of changes in runoff characteristics in ~2030 relative to ~1990 estimated by the SIMHYD and Sacramento models using climate change projections from the IPSL global climate model under the medium global warming scenario
Mean annual runoff

0 -10 -20 -30 10 0
Mean summer runoff

0
Mean winter runoff
-20 -10 -20 -30 -30 -40
% change SIMHYD
-40 -40
-30
-20
-10
-20
-10
10
-60 -60
-40
-20
0 -10 -20 -30
40 30 20 10 0
Num ber of days with daily runoff less than 0.1 mm
-40 -40
-30
-20
-10
10
20
30
40
% change Sacramento
Figure 5-2. Comparison of changes in runoff characteristics in ~2030 relative to ~1990 estimated by the SIMHYD and Sacramento models using climate change projections from the INMCM global climate model under the medium global warming scenario
CSIRO 2008
Mean annual runoff

40 40 20 20 0 0 -20
Mean summer runoff

40
Mean winter runoff
20
% change SIMHYD
-20 -20
20
40
-40 -40
-20
20
40
-20 -20
20
40
60 40 20 0

20 10 0 -10
Number of days with daily runoff less than 0.1 mm
-20 -20
20
40
60
-20 -20
-10
10
20
% change Sacramento
Figure 5-3. Comparison of changes in runoff characteristics in ~2030 relative to ~1990 estimated by the SIMHYD and Sacramento models using climate change projections from the CCCMA T47 global climate model under the medium global warming scenario
5.2
Global warming and forest water use
As the same model parameter values are used to estimate the historical and future runoffs, the modelling does not take into account the potential effect of global warming and enhanced CO2 concentrations on forest water use. There are two main reasons why changes in forest water use can have a significant impact on future runoff. First, runoff is the difference between the two larger water balance terms, rainfall and evapotranspiration, and therefore small changes in forest evapotranspiration can result in large changes in runoff. Second, forests cover a large proportion of the upland areas where most of the runoff occurs. The physiological effects of increased CO2 on the hydrological cycle have been shown to be the major cause of increased runoff for some large river basins around the world (Gedney et al., 2006). However, these effects are generally not considered in predictions of future runoff changes due to global warming, mainly because the interactions and feedbacks between the atmosphere and vegetation under increased CO2 are not well understood. For example, the review of Field et al. (1995) showed that elevated CO2 leads to large decreases in canopy conductance, resulting in increases in water use efficiencies and reduced transpiration. However, elevated CO2 can also lead to increases in leaf area index (LAI). The reduced transpiration per leaf/plant due to higher forest water use efficiency is therefore offset by the higher interception loss and higher transpiration due to higher LAI. Various other processes also interact to modulate the response of evapotranspiration and runoff to increased CO2 (Figure 5-4). The summary results of Ainsworth and Long (2005) from 12 large-scale free-air CO2 enrichment (FACE) experiments found that functional groups differed in their response to elevated CO2 with trees generally more responsive than grass and crops. Trees showed significant increases in LAI, while there was no significant change in LAI in grass and crops. The ecohydrological response to elevated CO2 is likely to vary with hydroclimatic conditions. In catchments where evapotranspiration is limited by energy (e.g. wet catchments in the tropics and northern hemisphere), increased CO2 is likely to result in reduced evapotranspiration and hence increased runoff. However, in water-limited areas, like most parts
CSIRO 2008
of the MDB, the lower transpiration per leaf/plant from CO2 fertilisation is likely to be offset by increased LAI, resulting in little change in evapotranspiration and runoff. Marcar et al. (2006) provide a detailed discussion on the potential impact of global warming on forest water use for the Murray Uplands area. In summary, although the potential changes in forest water use under increased CO2 could significantly impact future runoff, the impact is not estimated in this project. This is because there are compensating positive and negative impacts of global warming on forest water use, and it is difficult to estimate the net effect because of the complex climatebiosphere-atmosphere interactions and feedbacks.
Altered boundary layer
Soil Evaporation Evapotranspiration
Soil Moisture
Runoff CO2 Canopy Conductance LA I
Figure 5-4. The effect of increased CO2 on catchment water balance processes (adapted from Field et al., 1995) (upward arrow next to a process indicates an increase, downward arrow next to a process indicates a decrease)
5.3
Future bushfire risk and impact on runoff
Bushfire risk is likely to increase in the future because of global warming (Hennessy et al., 2005; Lucas et al., 2007). This section presents a broad assessment of the potential impact of increased bushfire risk on future runoff. The Forest Fire Danger Index (FFDI), which has been widely used to assess broad scale daily forest fire risk in southeast Australia, is used here (Luke and McArthur, 1978; Hennessy et al., 2005). The FFDI is calculated using the equation described in Nobel et al. (1980). The daily data used to calculate FFDI are rainfall, maximum temperature, relative humidity at 3 pm, and maximum wind speed. The FFDI is calculated for 0.05 x 0.05 grid cells across the MDB for the historical climate and the future climate. The calculations for the historical climate are based on 1980 to 2006 SILO climate data described in Chiew et al. (2008). The calculations for the future climate are based on scaling the 1980 to 2006 historical climate data with seasonal climate change projections from six GCMs for ~2030: CCCMA_T47, CCCMA_T63, IAP, INMCM, MIROC-M, and MRI (see Chiew et al. (2008) for description of future climate data). These six GCMs are used because they have all the climate data required to calculate FFDI. Figure 5-5 shows the mean annual cumulative FFDI for the historical climate (1980 to 2006) and the percent changes in FFDI by ~2030 for the second lowest, median and second highest FFDI calculated from future climate data informed by the six GCMs. The FFDI is plotted only for grid cells with more than 20 percent woody cover (woody cover data from Furby, 2002). Table 5-1 summarises the results for the 18 MDB regions. The results show increases in FFDI by about 5 percent in the median result, by 10 to 20 percent in the second highest result, and little change in the second lowest result. To estimate the increase in area burnt as a result of higher fire weather risk, a relationship between area burnt and mean annual cumulative FFDI is established and shown in Figure 5-6. The relationship is established using annual data on total area burnt in Victoria from 1921 to 2006 (from Tolhurst, University of Melbourne, pers. comm.) and the FFDI calculated for three sites in the MDB part of Victoria (Bendigo, Mount Beauty and Shepparton). The y-axis shows the percent increase in area burnt, which is estimated as the total area burnt in Victoria divided by the total woody cover in Victoria.
o o
CSIRO 2008
Streamflow increases immediately following the destruction of mature forest by bushfire but then decreases as the forest regenerates (Kuczera, 1987). For this project, streamflow response curves are developed for different groups of species and averaged over 100 years to estimate the long-term change in streamflow resulting from a fire (Figure 5-7). The averaging implies that a small portion of a catchment is burnt each year when in reality fires are likely to be episodic involving large portions of a catchment. However, the average values provide an indication of the impact at the regional scale. The northern mixed eucalypt species is used for Queensland and northern New South Wales and the southern mixed eucalypt species is used for Victoria and southern New South Wales. Small areas of mountain/alpine ash and snowgum are also used for the higher elevation areas in the south-east MDB to reflect the more diverse forest species there. The broad assessment of the potential impact of increased bushfire risk on mean annual runoff in the 18 MDB regions is made by: (i) estimating the increase in mean annual cumulative FFDI (Table 5-1); (ii) estimating the increase in area burnt using the relationship in Figure 5-6; and (iii) estimating the reduction in mean annual runoff as a result of the increase in area burnt using the relationship in Figure 5-7. The estimated reductions in mean annual runoff as a result of increased bushfire risk for the 18 MDB regions, for the second lowest, median and second highest result using climate change projections from the six GCMs, are shown in Table 5-1. Although these results are based on best available current knowledge, there are numerous assumptions involved and therefore there is considerable uncertainty in the estimates. Bushfire risk is likely to increase in the future and in areas where bushfires occur, runoff will reduce significantly as forests regrow. However, the broad assessment here suggests that the impact on runoff from increased bushfires as a result of global warming by ~2030, averaged over a very large region is unlikely to be significant. The impact of increased bushfire risk on mean annual runoff over the MDB varies from zero to a 0.6 percent reduction in the median future scenario and from 0 to a 1.1 percent reduction in the second highest future scenario (Table 5-1).
Table 5-1. Summary of broad assessment of impact of increased future bushfire risk on future runoff Region Current mean annual cumulative FFDI 5583 5158 3814 3522 2454 2322 2466 2620 4216 2952 1897 3037 1834 1609 1988 2406 3058 2174 Percentage change in future mean annual cumulative FFDI Second lowest 1 1 1 0 -1 -1 -1 -2 -1 -1 0 1 2 2 1 1 1 1 Median Second highest 15 16 15 15 13 13 13 13 13 13 16 13 17 17 19 20 15 11 Percentage reduction in mean annual runoff from increased bushfire risk Second lowest 0.0 0.0 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Median Second highest 0.0 0.0 1.0 1.1 0.8 0.4 0.8 0.7 0.3 1.0 0.0 1.0 0.1 0.0 0.1 0.5 1.0 0.1
Paroo Warrego Condamine-Ballonne Moonie Border Rivers Gwydir Namoi Macquarie-Castlereagh Barwon-Darling Lachlan Murrumbidgee Murray Ovens Goulburn-Broken Campaspe Loddon-Avoca Wimmera Eastern Mount Lofty Ranges
8 8 7 7 6 3 3 3 4 4 4 5 6 6 6 6 5 5
0.0 0.0 0.6 0.6 0.3 0.0 0.1 0.1 0.1 0.3 0.0 0.4 0.0 0.0 0.0 0.1 0.4 0.0
CSIRO 2008
Figure 5-5. Mean annual cumulative forest fire danger index (FFDI) for the historical climate (19802006) and percent change by ~2030 (second lowest, median and second highest using future climate data informed by six global climate models)
CSIRO 2008
100
Relative forested area burnt (%)
10
1.0
0.1
0.01 0 1000 2000 3000 4000 5000
Mean annual cumulative FFDI
Figure 5-6. Relationship between forest area burnt and mean annual cumulative FFDI based on data from Victoria
Mean annual rainfall (mm) Average change in mean annual runoff (mm) .
0 0 -50 -100 -150 -200 -250
Mountain Ash
500
1000
1500
2000
2500
-300 -350 -400
Snowgum Mixed Eucalypt - southern Mixed Eucalypt - northern
Figure 5-7. Average change in mean annual streamflow for different species following bushfires
CSIRO 2008
Rainfall-runoff modelling for Scenario D
This chapter describes the methods used to estimate the future growth in commercial forestry plantations and farm dams and the modelling of the impact of these developments on future runoff. There are four steps in Scenario D modelling: (i) estimation of growth in commercial forestry plantations (~2030 relative to ~2005); (ii) estimation of the impact of commercial forestry plantation development on Scenario C subcatchment runoffs to obtain Scenario C+plantations runoff values; (iii) estimation of growth in farm dams (~2030 relative to ~2005); and (iv) estimation of the impact of farm dam development on Scenario C+plantations runoffs to obtain Scenario D subcatchment runoff values. The modelling is carried out to estimate the impact of development on the extreme dry, median and extreme wet Scenario C runoffs to obtain Scenario Ddry, Dmid and Dwet runoff values.
6.1
6.1.1
Commercial forestry plantations

Estimation of future growth in commercial forestry plantations
Plantations here refer only to commercial forestry plantations. Other types of plantations, for example reforestation for carbon sequestration and dispersed small forestry sites, are not considered in this assessment. Commercial forestry plantation is an increasingly important land use in Australia. There are almost two million hectares of plantations in Australia, an increase of 70 percent since 1994 (Parsons et al., 2006). The existing total commercial forestry plantations in the MDB cover about 290,000 ha. Table 6-1 shows the areas of commercial forestry plantations in the MDB and the projected increases by 2030. The development of forestry plantations will depend on opportunities and current and future policies, and there will be considerable uncertainty in any projection of future growth. For this project, the projection of future growth in commercial forestry plantations estimated by the Bureau of Rural Sciences (Parsons, pers. comm., 2007), using data supplied by plantation owners and managers, is used. Based on the projections in Table 6-1, the impact of increased commercial forestry plantations on future runoff is only modelled for three MDB regions: Murrumbidgee (17,000 ha increase in plantations by 2030); Murray (33,000 ha); and Eastern Mount Lofty Ranges (2000 ha). The projected increase in the other 15 MDB regions is small and will have negligible impact on future runoff. The increase in plantations in a MDB region is added based on the 0.05 x 0.05 modelling grid cells. The grid cells are sorted by the mean biomass productivity estimated using the PROMOD model (Battaglia and Sands, 1997), and plantations are added to the non-woody area of successive cells until the projected new plantation area is reached. Plantations are not added to areas where the land use is classified as natural forest. The current vegetation and plantation datasets used are the National Carbon Accounting System 2005 forest dataset and the Bureau of Rural Sciences 2006 plantation dataset (BRS, 2006). For the Eastern Mount Lofty Ranges, additional information from the South Australian Department of Water, Land and Biodiversity Conservation regarding water allocation is used in locating the additional plantations in subcatchments.
o o
6.1.2
Modelling impact of plantation development on runoff
The impact of plantation development on future runoff is modelled using the FCFC (Forest Cover Flow Change) model (Brown et al., 2006; http://www.toolkit.net.au/fcfc). The methodology in FCFC is based on analysis of streamflow data from small experimental catchments that have undergone large changes in forest cover (Brown et al., 2005). The steps used in FCFC are: (i) calculation of the flow duration curve (FDC) for current forest cover; (ii) parameterisation of the FDC based on the method in Best et al. (2003); (iii) estimation of change in mean annual runoff for the new plantation forest cover based on Zhang et al. (2001); (iv) adjustment of the FDC parameters based on change in mean annual runoff for the new plantation forest cover; and (v) adjustment of the current runoff series to reflect the change in the FDC. The data required to run FCFC are the current and future forest cover, and the daily time series of rainfall, PET and current runoff. The areas of current and future commercial forestry plantations have been described in Section 6.1. The
CSIRO 2008
same future daily rainfall and PET series used for the rainfall-runoff modelling are also used to run FCFC. The daily runoff series used is the runoff estimated for Scenarios Cdry, Cmid and Cwet. The FCFC model then adjusts these series to reflect the impact of the increase in commercial forestry plantations to provide daily runoff series for Scenarios Cdry+plantations, Cmid+plantations and Cwet+plantations. The modelling results indicate that commercial forestry plantations can impact significantly on runoff where they are located. However, for the growth projections used here, the impact of commercial forestry plantations averaged over a large region is relatively small. The biggest impact is 0.8 percent reduction in mean annual runoff averaged over the Eastern Mount Lofty Ranges, followed by about 0.3 percent reduction in mean annual runoff averaged over the Murrumbidgee region and the Murray region. The impacts of the small projected growth in commercial forestry plantations in the other 15 MDB regions are negligible. However, as mentioned earlier, there is considerable uncertainty in the projected growth of commercial forestry plantations.
Table 6-1. Existing areas of commercial forestry plantations in the Murray-Darling Basin and the projected increases by 2030 Native forests and woodlands (ha) Avoca River Border Rivers Broken River Campaspe River Condamine-Culgoa River Goulburn River Gwydir River Kiewa River Lachlan River Loddon River Lower Murray River Macquarie-Bogan Rivers Murray-Riverina Murrumbidgee River Namoi River Ovens River Upper Murray River Wimmera-Avon Rivers Total 112,191 1,519,054 120,505 66,329 5,446,152 627,078 764,682 95,639 2,409,542 250,598 1,672,466 2,321,206 217,720 1,503,902 1,456,304 421,942 1,026,565 468,311 17,846,186 Plantation areas, 2005 (ha) Hardwoods 375 0 713 850 145 2,737 0 63 0 971 201 0 0 4 0 838 151 297 7,345 Softwoods 373 2,206 8,771 409 883 14,698 680 3,935 31,050 3,791 1,771 40,637 125 107,119 5,751 19,036 41,661 357 283,253 Total 748 2,206 9,484 1,259 1,028 17,434 680 3,998 31,050 4,762 1,972 40,637 125 107,123 5,751 19,874 41,812 654 290,598 0 17,000 0 0 33,000 0 54,500 Projected increase in plantation areas by 2030 (ha) 0 0 0 1,000 0 500 0 0 0 1,000 2,000
6.2
6.2.1
Farm dams
Estimation of existing farm dam storage capacity
Farm dams here refer only to dams that collect water from their own catchment, not those that store water diverted from a nearby river, as the latter require licences and are usually included within existing river system models. The increases in farm dam storage capacity by ~2030 relative to current (~2005) are estimated for each subcatchment in the MDB, taking into account the policy constraints and the drivers for farm dam development. Except for the Eastern Mount Lofty Ranges, the Statistical Local Areas (SLA), as defined by the Australian Bureau of Statistics in the Australian Standard Geographical Classification 2001 (ASGC2001), are used as the primary units for estimating future farm dam development.
CSIRO 2008
Figure 6-1 shows the sources of data used to estimate existing farm dam storage capacity, Figure 6-2 shows estimates of existing farm dam storage capacity across the MDB, and Table 6-2 summarises the existing farm dam storage capacity in the 18 MDB regions. For New South Wales, the existing farm dam storage capacity is estimated from the satellite imagery captured between 2004 and 2006 and analysed by Geosciences Australia (2007). The farm dam mapping by Geosciences Australia is ongoing, and the data made available to the project at 1 August 2007 are used here. In areas where data are not available, the farm dam storage capacity is estimated based on the assumption that the density of farm dams (ML/ha) is related to existing land use (Figure 6-1). For Victoria, the existing farm dam storage capacity is estimated from Vicmap 1:25,000 scale topographic maps. For South Australia, the total farm dam storage capacity for each management subcatchment is supplied by the South Australian Department of Water, Land and Biodiversity Conservation (DWLBC). The relationship used to convert surface area for each individual farm dam (S in m ) 1.314 to farm dam storage capacity (V in ML) is V = 0.000145 x S (Sinclair Knight Merz, 2004a). A comparison of estimates using this relationship with those from several other studies (Sinclair Knight Merz, 2004a; McMurray, 2004) for farm dams up to 20 ML showed very little differences. The total existing farm dam storage capacity estimated for the MDB is 2164 GL (Table 6-2) with the highest density of farm dams in the eastern MDB. The only previous MDB-wide farm dam study is that undertaken by Agrecon (2005) which digitised farm dams from satellite imagery for 84 square tiles across the MDB. The Agrecon data provides more detailed information including historical farm dam growth rates, but covers only 1 percent of the MDB. The total existing farm dam storage capacity of 2164 GL estimated for this project is remarkably similar to the 2000 GL estimated by Agrecon, based on extrapolating information from a 1 percent sample of the MDB to the entire MDB. In Victoria and Queensland, it is assumed that farm dams smaller than 5 ML are for stock and domestic purposes and farm dams larger than 5 ML are for irrigation purposes. This is based on previous Victorian studies (Sinclair Knight Merz, 2004b) and on the rates provided in Fitzimon (2006) as guidance for Queensland farmers, where (assuming that a typical stock and domestic farm dam would store two years worth of usage) a 5 ML farm dam would supply about four people and 110 cows. By considering only farm dams less than 5 ML from the Geosciences Australia satellite imagery and Vicmap topographic maps, and the population density, the average density of stock and domestic farm dams is 0.58 ML/person in Victoria and 0.32 ML/person in the eastern Queensland part of the MDB. The average density of stock and domestic farm dam in the Queensland part of the MDB is likely to be lower than in Victoria because there are reliable and accessible groundwater aquifers underlying most of southern Queensland that are widely used for stock and domestic water supply.
2
CSIRO 2008
Figure 6-1. Sources of data on existing farm dam storage capacity
CSIRO 2008
Figure 6-2. Spatial density of existing farm dam storage capacity
CSIRO 2008
Table 6-2. Existing farm dam storage capacity (GL), listed by data source, region and state
Region
QLD
NSW and ACT Captured Estimated Total for from GA based on NSW and 2007 data landuse ACT
VIC
SA
TOTAL
Paroo Warrego Condamine-Balonne Moonie Border Rivers Gwydir Namoi Macquarie-Castlereagh Barwon-Darling Lachlan Murrumbidgee Murray Ovens Goulburn-Broken Campaspe Loddon-Avoca Wimmera Eastern Mount Lofty Ranges Total MDB
0 75 263 46 79 -
77 113 98 22 67 197 217 6 -
0 0 46 219 26 63 134 67 -
77 113 145 242 94 261 351 73 -
21 30 105 35 98 34 -
22
0 75 263 46 156 113 145 242 94 261 351 94 30 105 35 98 34 22 2164
6.3
Estimation of future farm dam development
In Victoria and Queensland, current legislation limits future farm dam development in the MDB mainly to those for stock and domestic purposes (Victoria Water Act, 1989; Queensland Water Act, 2000). It is assumed that current requirements for stock and domestic water for existing landholders are met by current surface and groundwater infrastructure. Future subdivisions of rural properties may, however, result in additional requirements for stock and domestic water, which may include the development of new farm dams. To estimate future farm dam development in Victoria and Queensland, it is assumed that the increase in the storage of farm dams is proportional to the projected increase in rural population. The population growth for each Statistical Local Area (SLA) is estimated as the maximum of the projected population between 2006 and 2030 relative to the maximum population between 2001 and 2006, using population data from the Victorian Department of Sustainability and Environment (2004) and Australian Bureau of Statistics (2004). The maximum population (rather than the population at 2030 and at 2006) is used because although farm dams may be developed to meet the maximum population, it is unlikely that they will be removed if the population starts to decline. The farm dam development by 2030 for each SLA is estimated as the projected increase in population in the SLA multiplied by the existing farm dam density in the SLA. In Queensland, where there is insufficient data to estimate the existing farm dam density for an SLA, the Queensland average of 0.32 ML/person is used. In New South Wales, the future development of farm dams is dependent on three factors: current farm dam storage capacity; an upper limit set by the harvestable right defined by the NSW Water Management Act (2000); and a continuation in the observed historical growth rate of farm dams. The estimation of current farm dam storage capacity has been described in Section 6.2.1. The Harvestable Right (HR) policy in New South Wales sets the maximum capacity of farm dams that may be constructed on each individual property without the purchase of a water right. The intent of the policy is that a farm dam capacity equal to the HR will intercept on average 10 percent of the mean annual runoff. To estimate the HR for each parcel of land, the contour map of HR from the NSW Department of Natural Resources
CSIRO 2008
(supplied 7 March 2007) is converted into a grid of HR in ML/ha with a spatial resolution of 1000 m. The HR for each parcel of land in the eastern and central divisions of the New South Wales MDB is calculated by multiplying the area by the average of the HR for the cells that are in the parcel of land. In the estimation of HR, it is assumed that no farm dams will be constructed in the dry western division of New South Wales MDB and in properties on incompatible land use types such as conservation and natural environments, urban residential, roads and water. Prior to the implementation of the HR policy in 1999, landholders could develop farm dams without significant legislative restrictions on the farm dam capacity. To estimate the growth rate of farm dams, data from Agrecon (2005) for 1999 and 2004, which cover less than 1 percent of New South Wales, are used. The limited Agrecon data show a similar growth rate of 0.6 percent per year across New South Wales except for several of the sample squares in the north-east where the apparent growth rate is significantly higher. However, the Agrecon data do not explicitly distinguish between the different types of farm dams, and a large proportion of the data from the Agrecon samples squares in north-east New South Wales are for farm dams that would operate as off-stream storages, collecting water diverted from nearby streams or irrigation channels. As these are not part of the definition of farm dams here, a growth rate of 0.6 percent per year (or 14 percent increase in farm dam between 2006 and 2030) is used to represent farm dam growth across the entire New South Wales. The increase in farm dam storage for each SLA in the eastern and central divisions of New South Wales MDB is estimated by first calculating the available HR as the difference between HR and current farm dam storage capacity, and then estimating the increase in farm dam storage as the lesser of 14 percent multiplied by either the current farm dam storage capacity or the available HR. Table 6-3 summarises the existing farm dam storage capacity, HR, HR taken by existing farm dams, available HR and the estimated increase in future farm dam storage capacity, averaged over each of the eight MDB regions in NSW. In South Australia, future farm dam development for the Eastern Mount Lofty Ranges (EMLR) is estimated for management subcatchments defined by the South Australian Department of Water, Land and Biodiversity Conservation (DWLBC), rather than for SLAs. There is currently a moratorium on future development of farm dams in EMLR while a water management plan is being prepared, so the assumptions on farm dam development policy used here are based on the policy that applied prior to the current moratorium and informal discussions with DWLBC. The future farm dam development is estimated as the lesser of the following two amounts: (i) the increase in farm dam storage capacity based on extrapolation of historical farm dam growth rate; or (ii) the policy limitation for the EMLR which states that future farm dam and commercial forestry plantation development cannot reduce the estimated May to November mean runoff by more than 30 percent. The projected farm dam growth rate estimated from the trend in historical data is about 1.9 percent growth per year, giving a projected increase of 46 percent from existing capacity by 2030. The May to November mean runoff estimates are provided by DWLBC. The increase in farm dam storage capacity for each management subcatchment in the EMLR is therefore estimated as the lesser of: (i) a 46 percent increase in farm dam storage capacity; or (ii) the increase in farm dam storage capacity that will reduce the May to November mean runoff by 30 percent. The estimated total increase in farm dam storage capacity across the EMLR is about 16 percent. Figure 6-3 shows the estimated increase in farm dam storage capacity by 2030 across the MDB. The largest projected increases in farm dam density are in New South Wales, EMLR and central Victoria in the upland areas of the Campaspe and Goulburn-Broken regions. The estimated increase in total farm dam storage capacity by 2030 across the entire MDB is 219 GL, an increase of about 10 percent of current farm dam storage capacity. The increases in farm dam storage capacity for the 18 MDB regions are also summarised in Table 6-4 and Figure 3-19.
CSIRO 2008
Table 6-3. Summary of projected increases in farm dam storage capacity (~2030 relative to ~2005) in Murray-Darling Basin regions in New South Wales Region Estimated existing farm dam capacity (GL) 77 113 145 242 94 261 351 73 1,356 Harvestable Harvestable right (GL) right taken by existing farm dams (GL) 103 141 200 324 131 322 373 80 1,674 31 39 57 103 31 127 134 24 546 Available Projected Projections harvestable additional as a percent right (GL) dams to 2030 of existing (GL) dam capacity 72 102 143 221 100 195 239 56 1,128 11 15 20 38 13 37 48 11 193 14 14 14 16 14 14 14 15 14
Border Rivers* Gwydir Namoi Macquarie-Castlereagh Barwon-Darling Lachlan Murrumbidgee Murray* Total for NSW part of MDB
*Only NSW part of the region
Figure 6-3. Spatial density of projected increase in farm dam storage capacity (~2030 relative to ~2005)
CSIRO 2008
6.3.1
Modelling impact of farm dam development on runoff
The conceptual daily farm dam model, CHEAT (Nathan et al., 2005), is used to model the impact of farm dam development on daily runoff series in subcatchments across the MDB. The time series inputs required to run CHEAT are daily rainfall, PET and runoff. The same future daily rainfall series used for the rainfall-runoff modelling are also used to run CHEAT. The future APET series used for the rainfall-runoff modelling is scaled upwards (by 30 to 70 percent, one scaling factor for each MDB region) to obtain point PET used as input to CHEAT. The daily runoff series used here are the three Scenario C runoff series modified by the plantation impact described in Section 6.1 (Cdry+plantations, Cmid+plantations and Cwet+plantations). The runoff inflows into individual farm dams are assumed to be proportional to the farm dam catchment areas. CHEAT performs a daily water balance for each farm dam in a sample that represents the projected future farm dam development. The water balance for each dam considers inflows from the farm dam catchment area, rainfall falling directly on the dam surface, evaporation losses from the dam, and water demand. New farm dams less than 5 ML are assumed to be for stock and domestic purposes. The annual demand from each stock and domestic farm dam is assumed to be half of the dam storage capacity and distributed uniformly throughout the year. New farm dams larger than 5 ML are assumed to be for irrigation of crops or pasture. For these dams, it is assumed that 83 percent of the storage capacity is extracted for consumption each year and that this demand is concentrated over the months of the year with the highest rainfall deficit, defined as the difference between point PET and rainfall in each month of the year. These demand factors for stock and domestic farm dams and irrigation farm dams are based on surveys performed by Sinclair Knight Merz (2004b). The daily water balance computes the dam storage at the end of the day and the spills. Seepage losses from the dam are assumed to be negligible. The parameters specified for the CHEAT model of each subcatchment are the total increase in farm dam storage capacity, the distribution of farm dam storage size, the upstream catchment area and the storage versus surface area relationship. The estimation of increases in farm dam storage capacity has been described in Section 6.2.2. The increase in farm dam storage capacity in a subcatchment typically comprises several dozen to several hundred new farm dams. The distribution of farm dam size in the projected increase in farm dam storage capacity is assumed to be the same as 2 the current distribution of farm dam size. A two-part relationship is used to estimate the upstream catchment area (km ): upstream catchment area = 0.08 x farm dam storage capacity (in ML), for farm dams smaller than 5 ML; and upstream catchment area = 0.4 + 0.0126 x (farm dam storage capacity - 5), for farm dams larger than 5 ML. This general relationship used across the MDB is based on studies from four catchments in Victoria and South Australia. The farm 2 1.314 is used across the MDB dam storage (V in ML) versus surface area relationship (S in m ) of V = 0.000145 x S (Sinclair Knight Merz, 2004).
6.4
Results
Figure 6-4 shows the modelled reduction in mean annual runoff under Scenario Cmid+plantations caused by the projected increase in farm dams by ~2030 for subcatchments across the MDB. Table 6-4 summarises the farm dam modelling results, averaged over each of the 18 MDB regions. The total projected increase in farm dam storage capacity by ~2030 2 across the entire MDB is 219 GL (or 0.21 ML/km averaged across the MDB). The modelled impact of the future farm dam development is a 0.7 percent reduction in mean annual runoff averaged across the MDB (or 0.17 mm). There is considerable spatial variation in the projected increase in farm dams and the resulting modelled impact on mean annual runoff. The biggest impact of farm dam development is in the Eastern Mount Lofty Ranges, with a 3 percent reduction in mean annual runoff over the region. In Victoria, the projected farm dam development will reduce mean annual runoff by 1.7, 0.8 and 0.4 percent in the Campaspe, Loddon-Avoca and Goulburn-Broken regions, respectively, with negligible impact elsewhere. The impact of projected farm dam development in the eastern and central divisions of the New South Wales MDB will reduce mean annual runoff by 1 to 1.5 percent. The small projected farm dam development in Queensland has negligible impact on future runoff. The estimated impact of farm dam development on future runoff for the projected future farm dam development across the MDB is small compared to the modelled climate change impact on runoff. The impact of farm dams is not spread uniformly within a year. Farm dams capture inflows over the wet season and the water stored in the dams are then lost by evaporation and used for consumptive use, with losses and demands often
CSIRO 2008
concentrated in the dry season. In the southern MDB, the wet season usually commences in May or June and runs through until October or November. Unless low flow bypasses are installed, the farm dams built in these areas capture most of their runoff over May through July and these are the months when farm dams cause the largest reductions in overall catchment runoff. In the Ovens, Goulburn-Broken, Campaspe, Loddon-Avoca, Wimmera and EMLR regions, the farm dams would fill and typically spill during late winter and spring (typically August through November), so relative reductions on runoff in these months are lower. Farm dams will capture any runoff occurring over summer and autumn, although runoff events during these months are infrequent. The seasonal pattern of runoff is reversed in the northern MDB, with the wet season occurring over summer and autumn and the dry season over winter and spring. Therefore in the northern regions the seasonality of the impact is reversed from the southern ones, with the highest impacts occurring over December through February. The Murrumbidgee, Murray and Lachlan regions can produce runoff in both the northern and southern wet seasons, with the seasonal pattern of impacts in these three regions more uniformly distributed across the year. The largest source of uncertainty in estimating farm dam impact on future runoff is the uncertainty in projecting future farm dam growth. The increase in farm dams is estimated by considering trends in historical farm dam growth and current policy controls. There is considerable uncertainty both as to how landholders will respond to development policies and how governments may set policies in the future.
CSIRO 2008
Figure 6-4. Percent reduction in mean annual runoff due to the projected increase in farm dams (~2030 relative to ~2005)
CSIRO 2008
Table 6-4. Summary of farm dam modelling results, averaged over each reporting region (~2030 relative to ~2005)
Region
Projected increase in capacity by 2030 (GL) <1 <1 3 1 13 15 20 38 13 37 48 11 2 9 3 3 <1 3 219
Reduction in mean annual runoff over the region (mm) 0.00 0.00 0.03 0.11 0.32 0.48 0.32 0.40 0.02 0.30 0.46 0.04 0.46 0.56 0.95 0.14 0.01 0.87 0.17
Percent reduction in mean annual runoff
Paroo Warrego Condamine-Balonne Moonie Border Rivers Gwydir Namoi Macquarie-Castlereagh Barwon-Darling Lachlan Murrumbidgee Murray Ovens Goulburn-Broken Campaspe Loddon-Avoca Wimmera Eastern Mt Lofty Ranges Total MDB
0.0 0.0 0.2 0.7 1.1 1.3 1.4 1.3 0.4 1.4 1.0 0.2 0.2 0.4 1.7 0.8 0.1 3.2 0.7
CSIRO 2008
References
Agricultural Reconnaissance Technologies Pty Ltd (Agrecon) (2005) Hillside farm dams investigation. MDBC Project 04/4677DO. Ainsworth EA and Long SP (2005) What have we learned from 15 years of free-air CO2 enrichment (FACE) experiments? A meta-analytic review of the response of photosynthesis, canopy properties and plant production to rising CO2. New Phytologist 165, 351372. Australian Bureau of Statistics (2004) Population projections for statistical local areas 2002 to 2022. Statistical consultancy project by the Australian Bureau of Statistics. Battaglia M and Sands P (1997) Modelling site productivity of Eucalyptus globulus in response to climatic and site factors. Australian Journal of Plant Physiology 24, 831850. Best AE, Zhang L, McMahon TA and Western AW (2003) Development of a model for predicting the changes in flow duration curves due to altered land use conditions. MODSIM 2003, Townsville, July 2003, pp. 861866. Brown AE, McMahon TA, Podger GM and Zhang L (2006) A methodology to predict the impact of change in forest cover on flow duration curves. CSIRO Land and Water, Science Report 8/06. Brown AE, Zhang L, McMahon TA, Western AW and Vertessy RA (2005) A review of paired catchment studies for determining changes in water yield resulting from alterations in vegetation. Journal of Hydrology 310, 2861. Bureau of Rural Sciences (2006) Spatial data for forest industry processing sites in the Murray-Darling Basin. National Inventory Dataset, Bureau of Rural Sciences, Canberra. Burnash RJC, Ferral RL and McGuire RA (1973) A generalised streamflow simulation system conceptual modelling for digital computers. U.S. Department of Commerce, National Weather Service and State of California, Department of Water Resources. Chiew FHS (2006) Estimation of rainfall elasticity of streamflow in Australia. Hydrological Sciences Journal 51, 613525. Chiew FHS and Leahy C (2003) Comparison of evapotranspiration variables in Evapotranspiration Maps of Australia with commonly used evapotranspiration variables. Australian Journal of Water Resources 7, 111. Chiew FHS and McMahon TA (2002) Modelling the impacts of climate change on Australian streamflow. Hydrological Processes 16, 12351245. Chiew FHS, Peel MC and Western AW (2002) Application and testing of the simple rainfall-runoff model SIMHYD. In: Mathematical Models of Small Watershed Hydrology and Applications (Editors: VP Singh and DK Frevert), Water Resources Publication, Littleton, Colorado, pp. 335367. Chiew FHS and Siriwardena L (2005) Estimation of SIMHYD parameter values for application ungauged catchments. Congress on Modelling and Simulation (MODSIM 2005), Melbourne, December 2005, pp. 28832889. Chiew FHS, Teng J, Kirono D, Frost AJ, Bathols JM, Vaze J, Viney NR, Young WJ, Hennessy KJ and Cai W (2008) Climate data for hydrologic scenario modelling across the Murray-Darling Basin. A report to the Australian government from the CSIRO Murray-Darling Basin Sustainable Yields Project. CSIRO, Australia. CSIRO (2007a) Water availability in the Paroo. A report to the Australian Government from the CSIRO Murray-Darling Basin Sustainable Yields Project. CSIRO, Australia. 88pp. CSIRO (2007b) Water availability in the Warrego. A report to the Australian Government from the CSIRO Murray-Darling Basin Sustainable Yields Project. CSIRO, Australia. 89pp. CSIRO and Australian Bureau of Meteorology (2007) Climate change in Australia. Technical report, www.climatechangeinaustralia.gov.au. Field C, Jackson R and Mooney H (1995) Stomatal responses to increased CO2: implications from the plant to the global-scale. Plant Cell Environment 18, 12141225. Fitzimon R (2006) Planning your farm dam. Queensland Government Department of Natural Resources, Fact sheet W24, viewed 18 June 2007, www.drw.qld.gov.au/factsheets/pdf/w24.pdf Furby S (2002) Land cover change: specification for remote sensing analysis. National Carbon Accounting System Technical Report 9, November 2002, www.greenhouse.gov.au/ncas/reports/tech09.html. Gedney N, Cox PM, Betts RA, Boucher O, Huntingford C and Stott PA (2006) Detection of a direct carbon dioxide effect in continental river runoff records. Nature 439, 835838. Geosciences Australia (2007) Man made hydrology spatial data, supplied 30 May 2007. Hennessy K, Lucas C, Nicholls N, Bathols J, Suppiah R and Ricketts J (2005) Climate change impacts on fire-weather in south-east Australia. Consultancy report for New South Wales Greenhouse Office, Victorian Department of Sustainability and Environment, Tasmanian Department of Primary Industries Water and Environment and the Australian Greenhouse Office, CSIRO Marine and Atmospheric Research and Australian Bureau of Meteorology Research Centre, 88 pp. IPCC (2007) Climate Change 2007: The Physical Basis. Contributions of Working Group 1 to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press (www.ipcc.ch). Jeffrey SJ, Carter JO, Moodie KB and Beswick AR (2001) Using spatial interpolation to construct a comprehensive archive of Australian climate data. Environmental Modelling and Software 16, 309330.
CSIRO 2008
Kuczera GA (1987) Prediction of water yield reductions following a bushfire in ash-mixed species eucalypt forest. Journal of Hydrology 94, 215236. Lucas C, Hennessy K, Mills G and Bathols J (2007) Bushfire weather in southeast Australia: recent trends and projected climate change impacts. Consultancy report prepared for the Climate Institute of Australia. Luke RH and McArthur AG (1978) Bushfires in Australia. Australian Government Publishing Service, Canberra, 359 pp. McMurray D (2004) Farm dam volume estimations from simple geometric relationships Mount Lofty Ranges and Clare regions South Australia. Department of Water Land and Biodiversity Conservation, South Australia, Report DWLBC 2004/48. Marcar NE, Benyon RG, Polglase PJ, Paul KI, Theiveyanathan S and Zhang L (2006) Predicting the hydrological impacts of bushfire and climate change in forested catchments of the River Murray Uplands: a review. CSIRO Water for a Healthy Country. Morton FI (1983) Operational estimates of areal evapotranspiration and their significance to the science and practice of hydrology. Journal of Hydrology 66, 176. Nash JE and Sutcliffe JV (1970) River flow forecasting through conceptual models, 1: a discussion of principles. Journal of Hydrology 10, 282290. Nathan RJ. Jordan PW and Morden R (2005) Assessing the impact of farm dam on streamflows, part 1 development of simulation tools. Australian Journal of Water Resources 9, 112. New South Wales Department of Natural Resources (2006) Water Management Act 2000, Order under Section 54, Harvestable Rights Eastern and Central Division Order. New South Wales Government Gazette No. 40, pp. 1628. New South Wales Government (2000) Water Management Act 2000 No 92. Noble IR, Bary GA and Gill AM (1980) McArthurs fire-danger meters expressed as equations. Australian Journal of Ecology 5, 201203. Parsons M, Gavran M and Davidson J (2006) Australias Plantations 2006. Bureau of Rural Sciences, Canberra. Peel MC, Chiew, FHS, Western AW and McMahon TA (2000) Extension of unimpaired monthly streamflow data and regionalisation of parameter values to estimate streamflow in ungauged catchments. National Land and Water Resources Audit. In: http://audit.ea.gov.au/anra/water/docs/national/streamflow/streamflow.pdf. Potter NJ, Chiew FHS, Frost AJ, Srikanthan R, McMahon TA, Peel MC and Austin JM (2008) Characterisation of recent rainfall and runoff across the Murray-Darling Basin. A report to the Australian government from the CSIRO MurrayDarling Basin Sustainable Yields Project. CSIRO, Australia. Queensland Government (2000) Water Act 2000. Sinclair Knight Merz (2004a) Estimating available water in catchments using sustainable diversion limits farm dam surface area and volume relationship (Victoria). Report prepared for the Department of Sustainability and Environment. Sinclair Knight Merz (2004b) Estimating available water in catchments using sustainable diversion limits farm dam demand factors (Victoria). Report prepared for the Department of Sustainability and Environment. South Australian Government (2004) Natural Resources Management Act 2004. Tan KS, Chiew FHS, Grayson RB, Scanlon PJ and Siriwardena L (2005) Calibration of a daily rainfall-runoff model to estimate high daily flows. Congress on Modelling and Simulation (MODSIM 2005), Melbourne, December 2005, pp. 29602966. Victoria Department of Sustainability and Environment (2004) Victoria in future 2004 population projections (online), viewed 16 April 2007, www.dse.vic.gov.au. Victoria Government (1988) Water Act 1989, Act Number 80/1989. Young W, Brandis K and Kingsford R (2006) Modelling monthly streamflows in two Australian dryland rivers: Matching complexity to spatial scale and data availability. Journal of Hydrology 331, 242256. Zhang L, Dawes WR and Walker GR (2001) The response of mean annual evapotranspiration to vegetation changes at catchment scale. Water Resources Research 37, 701708.
CSIRO 2008
Appendix A: Summary of key modelling results

Table A. Summary of key modelling results for the 18 Murray-Darling Basin regions Rainfall Area km2 percent 3.3% 7.2% 12.9% 1.4% 4.1% 2.3% 3.7% 6.9% 13.4% 8.1% 8.2% 19.6% 0.7% 2.1% 0.4% 2.3% 2.9% 0.4% 100.0% Runoff Runoff coefficient percent 15.6 6.6 14.5 16.8 32.2 47.9 28.6 32.9 6.5 17.6 37.2 18.9 171.9 88.6 34.2 9.9 8.1 19.3 21.7 -10% -8% -23% -3% -1% 18% 17% -5% 8% -24% -31% -21% -26% -41% -50% -52% -51% -36% -21% 0.06 0.02 0.04 0.03 0.05 0.06 0.04 0.06 0.02 0.05 0.10 0.07 0.23 0.20 0.12 0.05 0.05 0.07 0.06 2% 2% 9% 1% 5% 3% 3% 8% 3% 7% 16% 16% 6% 11% 1% 2% 2% 1% 96% Contribution to MDB* percent
1895-2006 1997-2006 Difference 1895-2006 1997-2006 Difference mm 311 422 514 528 641 644 633 455 328 461 530 340 1004 764 594 430 403 463 457 310 427 503 541 641 688 663 547 339 425 471 313 895 649 517 381 350 429 440 percent 0% 1% -2% 2% 0% 7% 5% 20% 3% -8% -11% -8% -11% -15% -13% -11% -13% -7% -4% mm 17.4 7.2 18.9 17.4 32.4 40.6 24.4 34.6 6.0 23.1 54.3 24.0 231.3 149.2 68.9 20.7 16.5 30.1 27.3
Paroo Warrego Condamine-Balonne Moonie Border Rivers Gwydir Namoi Macquarie-Castlereagh Barwon-Darling Lachlan Murrumbidgee Murray Ovens Goulburn-Broken Campaspe Loddon-Avoca Wimmera EMLR MDB
35,587 76,615 136,642 14,662 43633 24,947 39780 73453 142173 85532 87331 207723 7813 22337 3961 24918 30640 4693 1,062,440
* Percent runoff contribution is the percentage of total MDB runoff contributed by the region. The Snowy Mountains Hydroelectric Scheme contributes 4% to the overall total.
Scenario C (runoff modelling results) Dry Median Wet percent of percent of percent of percent of results with results results with results with runoff runoff runoff with runoff decrease increase decrease increase >10% >10% Increase in plantations
Scenario D Increase in median runoff farm dams
percent change from Scenario A Paroo Warrego Condamine-Balonne Moonie Border Rivers Gwydir Namoi Macquarie-Castlereagh Barwon-Darling Lachlan Murrumbidgee Murray Ovens Goulburn-Broken Campaspe Loddon-Avoca Wimmera EMLR MDB -16% -25% -20% -24% -28% -28% -31% -25% -22% -34% -31% -37% -44% -44% -46% -43% -47% -44% -33% -2% -6% -9% -10% -9% -9% -6% -6% -2% -10% -9% -10% -13% -13% -16% -16% -17% -15% -9% 40% 46% 26% 20% 20% 31% 37% 30% 50% 17% 13% 7% 1% -2% -4% 0% 1% 0% 16% 53% 60% 60% 60% 67% 67% 67% 60% 60% 67% 67% 73% 80% 100% 100% 100% 87% 93% 47% 40% 40% 40% 33% 33% 33% 40% 40% 33% 33% 27% 20% 0% 0% 0% 13% 7% 47% 40% 35% 60% 53% 53% 47% 40% 33% 60% 60% 53% 60% 60% 67% 67% 60% 67% 40% 33% 33% 27% 20% 27% 27% 33% 40% 27% 13% 7% 0% 0% 0% 0% 0% 0%
ha
GL
percent change from Scenario A
0 0 0 0 0 0 0 0 0 0 17000 33000 0 0 0 0 0 2000 52000
0 0 3 1 13 15 20 38 13 37 48 11 2 9 3 3 0 3 219
NS NS -9% -11% -10% -10% -7% -7% -2% -12% -10% -11% -13% -14% -18% -17% -17% -18% -10%
CSIRO 2008

C Rainfall RunoffModelling

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

C Rainfall RunoffModelling

Transféré par

Droits d'auteur :

Formats disponibles

Rainfall-runoff modelling across the Murray-Darling Basin

Scenario A Historical climate (1895 to 2006) and current development

Rainfall-runoff modelling across the Murray-Darling Basin

Scenario B Recent climate (1997 to 2006) and current development

Scenario C Future climate (~2030) and current development

Scenario D Future climate (~2030) and future development (~2030)

Rainfall-runoff modelling across the Murray-Darling Basin

Introduction................................................................................................................................1 Rainfall-runoff modelling method ...........................................................................................2

Summary of modelling results .................................................................................................5

Rainfall-runoff modelling for Scenario A ..............................................................................30

Rainfall-runoff modelling for Scenario C ..............................................................................41

Rainfall-runoff modelling for Scenario D...............................................................................48

References ...............................................................................................................................60 Appendix A: Summary of key modelling results .................................................................62

Rainfall-runoff modelling across the Murray-Darling Basin

Rainfall-runoff modelling across the Murray-Darling Basin

Rainfall-runoff modelling across the Murray-Darling Basin 1

Rainfall-runoff modelling method

2 Rainfall-runoff modelling across the Murray-Darling Basin

Rainfall-runoff modelling across the Murray-Darling Basin 3

4 Rainfall-runoff modelling across the Murray-Darling Basin

Summary of modelling results

Reporting regions, subcatchments and calibration catchments

Rainfall-runoff modelling across the Murray-Darling Basin 5

6 Rainfall-runoff modelling across the Murray-Darling Basin

Scenario A results (historical climate, recent development)

Rainfall-runoff modelling across the Murray-Darling Basin 7

8 Rainfall-runoff modelling across the Murray-Darling Basin

Rainfall-runoff modelling across the Murray-Darling Basin 9

10 Rainfall-runoff modelling across the Murray-Darling Basin

Scenario B results (recent climate, recent development)

Rainfall-runoff modelling across the Murray-Darling Basin 11

12 Rainfall-runoff modelling across the Murray-Darling Basin

Scenario C results (future climate, recent development)

Rainfall-runoff modelling across the Murray-Darling Basin 13

14 Rainfall-runoff modelling across the Murray-Darling Basin

Rainfall-runoff modelling across the Murray-Darling Basin 15

16 Rainfall-runoff modelling across the Murray-Darling Basin

Rainfall-runoff modelling across the Murray-Darling Basin 17

18 Rainfall-runoff modelling across the Murray-Darling Basin

Rainfall-runoff modelling across the Murray-Darling Basin 19

20 Rainfall-runoff modelling across the Murray-Darling Basin

Rainfall-runoff modelling across the Murray-Darling Basin 21

22 Rainfall-runoff modelling across the Murray-Darling Basin

Rainfall-runoff modelling across the Murray-Darling Basin 23

24 Rainfall-runoff modelling across the Murray-Darling Basin

Mean monthly rainfall (mm)

Mean monthly runoff (mm)

Mean monthly rainfall (mm)

Mean monthly runoff (mm)

Mean monthly rainfall (mm)

Mean monthly runoff (mm)

Mean monthly rainfall (mm)

Mean monthly runoff (mm)

Mean monthly rainfall (mm)

Mean monthly runoff (mm)

Mean monthly rainfall (mm)

Mean monthly runoff (mm)

Rainfall-runoff modelling across the Murray-Darling Basin 25

Mean monthly rainfall (mm)

Mean monthly runoff (mm)

Mean monthly rainfall (mm)

Mean monthly runoff (mm)

Mean monthly rainfall (mm)

Mean monthly runoff (mm)