River model calibration and scenario analysis for the Roper catchment Australia’s National Science Agency A technical report from the CSIRO Roper River Water Resource Assessment for the National Water Grid Justin Hughes, Ang Yang, Steve Marvanek, Biao Wang, Cuan Petheram and Seonaid Philip ISBN 978-1-4863-1929-9 (print) ISBN 978-1-4863-1930- (online) Citation Hughes J, Yang A, Marvanek S, Wang B, Petheram C, Philip S (2023) River model calibration and scenario analysis for the Roper catchment. A technical report from the CSIRO Roper River Water Resource Assessment for the National Water Grid. CSIRO, Australia. Copyright © Commonwealth Scientific and Industrial Research Organisation 2023. To the extent permitted by law, all rights are reserved and no part of this publication covered by copyright may be reproduced or copied in any form or by any means except with the written permission of CSIRO. Important disclaimer 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. CSIRO is committed to providing web accessible content wherever possible. If you are having difficulties with accessing this document please contact Email CSIRO Enquiries . CSIRO Roper River Water Resource Assessment acknowledgements This report was funded through the National Water Grid’s Science Program, which sits within the Australian Government’s Department of Climate Change, Energy, the Environment and Water. Aspects of the Assessment have been undertaken in conjunction with the Northern Territory Government. The Assessment was guided by two committees: i.The Assessment’s Governance Committee: CRC for Northern Australia/James Cook University; CSIRO; National Water Grid (Departmentof Climate Change, Energy, the Environment and Water); NT Department of Environment, Parks and Water Security; NT Department of Industry, Tourism and Trade; Office of Northern Australia; Qld Department of Agriculture and Fisheries; Qld Department of Regional Development, Manufacturing and Water ii.The Assessment’s joint Roper and Victoria River catchments Steering Committee: Amateur Fishermen’s Association of the NT; Austrade; Centrefarm; CSIRO, National Water Grid (Department of Climate Change, Energy, the Environment and Water); Northern Land Council; NT Cattlemen’s Association; NT Department of Environment, Parks Australia; Parks and Water Security; NT Department of Industry, Tourism and Trade; Regional Development Australia; NT Farmers; NT Seafood Council; Office of Northern Australia; Roper Gulf Regional Council Shire Responsibility for the Assessment’s content lies with CSIRO. The Assessment’s committees did not have an opportunity to review the Assessment results or outputs prior to its release. This report was reviewed by Barry Croke (Associate Professor, Integrated Catchment Assessment and Management Centre (iCAM) and Institute for Water Futures, The Fenner School of Environment and Society and Mathematical Sciences Institute, ANU College of Science) Acknowledgement of Country CSIRO acknowledges the Traditional Owners of the lands, seas and waters, of the area that we live and work on across Australia. We acknowledge their continuing connection to their culture and pay our respects to their Elders past and present. Photo Roper Bar. Source: CSIRO – Nathan Dyer Director’s foreword Sustainable regional development is a priority for the Australian and Northern Territory governments. Across northern Australia, however, there is a scarcity of scientific information on land and water resources to complement local information held by Indigenous owners and landholders. Sustainable regional development requires knowledge of the scale, nature, location and distribution of the likely environmental, social and economic opportunities and the risks of any proposed development. Especially where resource use is contested, this knowledge informs the consultation and planning that underpins the resource security required to unlock investment. In 2019 the Australian Government commissioned CSIRO to complete the Roper River Water Resource Assessment. In response, CSIRO accessed expertise and collaborations from across Australia to provide data and insight to support consideration of the use of land and water resources for development in the Roper catchment. While the Assessment focuses mainly on the potential for agriculture, the detailed information provided on land and water resources, their potential uses and the impacts of those uses are relevant to a wider range of regional-scale planning considerations by Indigenous owners, landholders, citizens, investors, local government, the Northern Territory and federal governments. Importantly the Assessment will not recommend one development over another, nor assume any particular development pathway. It provides a range of possibilities and the information required to interpret them - including risks that may attend any opportunities - consistent with regional values and aspirations. All data and reports produced by the Assessment will be publicly available. Chris Chilcott C:\Users\bru119\AppData\Local\Microsoft\Windows\Temporary Internet Files\Content.Word\C_Chilcott_high.jpg Project Director The Roper River Water Resource Assessment Team Project Director Project Leaders Project Support Communications Activities Agriculture and socio- economics Climate Ecology Groundwater hydrology Indigenous water values, rights, interests and development goals Chris Chilcott Cuan Petheram, Ian Watson Caroline Bruce Chanel Koeleman/Kate Cranney, Siobhan Duffy, Amy Edwards Chris Stokes, Caroline Bruce, Shokhrukh Jalilov, Diane Jarvis1, Adam Liedloff, Yvette Oliver, Alex Peachey2, Allan Peake, Maxine Piggott, Perry Poulton, Di Prestwidge, Thomas Vanderbyl7, Tony Webster, Steve Yeates David McJannet, Lynn Seo Danial Stratford, Laura Blamey, Rik Buckworth, Pascal Castellazzi, Bayley Costin, Roy Aijun Deng, Ruan Gannon, Sophie Gilbey, Rob Kenyon, Darran King, Keller Kopf3, Stacey Kopf3, Simon Linke, Heather McGinness, Linda Merrin, Colton Perna3, Eva Plaganyi, Rocio Ponce Reyes, Jodie Pritchard, Nathan Waltham9 Andrew R. Taylor, Karen Barry, Russell Crosbie, Phil Davies, Alec Deslandes, Katelyn Dooley, Clement Duvert8, Geoff Hodgson, Lindsay Hutley8, Anthony Knapton4, Sebastien Lamontagne, Steven Tickell5, Sarah Marshall, Axel Suckow, Chris Turnadge Pethie Lyons, Marcus Barber, Peta Braedon, Kristina Fisher, Petina Pert Land suitability Ian Watson, Jenet Austin, Elisabeth Bui, Bart Edmeades5, John Gallant, Linda Gregory, Jason Hill5, Seonaid Philip, Ross Searle, Uta Stockmann, Mark Thomas, Francis Wait5, Peter L. Wilson, Peter R. Wilson Surface water hydrology Justin Hughes, Shaun Kim, Steve Marvanek, Catherine Ticehurst, Biao Wang Surface water storage Cuan Petheram, Fred Baynes6, Kevin Devlin7, Arthur Read, Lee Rogers, Ang Yang, Note: Assessment team as at June 15, 2023. All contributors are affiliated with CSIRO unless indicated otherwise. Activity Leaders are underlined. 1James Cook University; 2NT Department of Industry, Tourism and Trade; 3 Research Institute for the Environment and Livelihoods. College of Engineering, IT & Environment. Charles Darwin University; 4CloudGMS; 5NT Department of Environment, Parks and Water Security; 6Baynes Geologic; 7independent consultant; 8Charles Darwin University; 9Centre for Tropical Water and Aquatic Ecosystem Research. James Cook University. ii | River model for the Roper catchment Shortened forms For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Units For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Preface Sustainable regional development is a priority for the Australian and Northern Territory governments. For example, in 2023 the Northern Territory Government committed to the implementation of a new Territory Water Plan. One of the priority actions announced by the government was the acceleration of the existing water science program ‘to support best practice water resource management and sustainable development’. The efficient use of Australia’s natural resources by food producers and processors requires a good understanding of soil, water and energy resources so they can be managed sustainably. Finely tuned strategic planning will be required to ensure that investment and government expenditure on development are soundly targeted and designed. Northern Australia presents a globally unique opportunity (a greenfield development opportunity in a first-world country) to strategically consider and plan development. Northern Australia also contains ecological and cultural assets of high value and decisions about development will need to be made within that context. Good information is critical to these decisions. Most of northern Australia’s land and water resources, however, have not been mapped in sufficient detail to provide for reliable resource allocation, mitigate investment or environmental risks, or build policy settings that can support decisions. Better data are required to inform decisions on private investment and government expenditure, to account for intersections between existing and potential resource users, and to ensure that net development benefits are maximised. In consultation with the Northern Territory Government, the Australian Government prioritised the catchment of the Roper River for investigation (Preface Figure 1-1) and establishment of baseline information on soil, water and the environment. Northern Australia is defined as the part of Australia north of the Tropic of Capricorn. The Murray– Darling Basin and major irrigation areas and major dams (greater than 500 GL capacity) in Australia are shown for context. The Roper River Water Resource Assessment (the Assessment) provides a comprehensive and integrated evaluation of the feasibility, economic viability and sustainability of water and agricultural development. While agricultural developments are the primary focus of the Assessment, it also considers opportunities for and intersections between other types of water-dependent development. For example, the Assessment explores the nature, scale, location and impacts of developments relating to industrial and urban development and aquaculture, in relevant locations. The Assessment was designed to inform consideration of development, not to enable any particular development to occur. As such, the Assessment informs – but does not seek to replace – existing planning, regulatory or approval processes. Importantly, the Assessment does not assume a given policy or regulatory environment. As policy and regulations can change, this enables the results to be applied to the widest range of uses for the longest possible time frame. Preface Figure 1-1 Map of Australia showing Assessment area It was not the intention – and nor was it possible – for the Assessment to generate new information on all topics related to water and irrigation development in northern Australia. Topics not directly examined in the Assessment are discussed with reference to and in the context of the existing literature. Functionally, the Assessment adopted an activities-based approach (reflected in the content and structure of the outputs and products), comprising eight activity groups; each contributes its part to create a cohesive picture of regional development opportunities, costs and benefits. Preface Figure 1-2 illustrates the high-level links between the eight activities and the general flow of information in the Assessment. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Preface Figure 1-2 Schematic diagram of the high-level linkages between the eight activities and the general flow of information in the Assessment. Assessment reporting structure Development opportunities and their impacts are frequently highly interdependent and consequently, so is the research undertaken through this Assessment. While each report may be read as a stand-alone document, the suite of reports most reliably informs discussion and decisions concerning regional development when read as a whole. The Assessment has produced a series of cascading reports and information products: •Technical reports; that present scientific work at a level of detail sufficient for technical andscientific experts to reproduce the work. Each of the eight activities has one or morecorresponding technical report. •A Catchment report; that for the Roper catchment synthesises key material from the technicalreports, providing well-informed (but not necessarily-scientifically trained) readers with theinformation required to make decisions about the opportunities, costs and benefits associatedwith irrigated agriculture and other development options. •A Summary report; that for the Roper catchment provides a summary and narrative for ageneral public audience in plain English. •A Summary factsheet; that for the Roper catchment provides key findings for a general publicaudience in the shortest possible format. The Assessment has also developed online information products to enable the reader to better access information that is not readily available in a static form. All of these reports, information tools and data products are available online at https://www.csiro.au/roperriver. The website provides readers with a communications suite including factsheets, multimedia content, FAQs, reports and links to other related sites, particularly about other research in northern Australia. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Executive summary At present, surface water allocations within the catchment of the Roper River are very low (~0.1 GL/year), particularly as a proportion of the median annual streamflow (~0.002%). The development of the surface water resources of this highly seasonal catchment to enable regional economic development, as has occurred in the south of Australia, would in many instances require rivers to be regulated and water stored/diverted. However, an understanding of the size and nature of surface water resources needs to precede any development, since development provides both opportunity and risks. This report presents information regarding the construction, calibration and scenario assessment of the Roper River model. No river model development has been previously undertaken in the Roper catchment. Hydrological prediction is difficult in northern Australia, and in this instance a lack of streamflow and climate data contributed to the challenges of model construction and calibration. The lack of data, including a poor spatial coverage of suitable streamflow data, required extensive parameter transfer across the catchment, which in turn contributed to model uncertainty. In construction and calibration of the Roper River model, two rainfall-runoff (RR) models were tested in conjunction with various calibration objective functions, calibration shingle configurations and parameter transfer protocols in an iterative fashion. In all, 12 versions of the model were at least partially constructed and calibrated and were used to assess uncertainty arising from model structure and model calibration/parameterisation. The mean annual end-of-system flow for the Roper River at node 90300000 for the reporting period (1910–2019) was simulated to be 5557 GL, while the median annual flow was 4341 GL. Uncertainty estimated using other river model versions suggests a range of plus or minus 12% for these values. Further upstream, where good data were available, streamflow estimates were more certain. For example, mean annual flow at node 90302500 was estimated to be 2414 GL plus or minus 2%. It should be noted that various other sources of uncertainty such as rainfall and streamflow observation uncertainty were not or could not be readily estimated. It is therefore likely that overall uncertainty is higher than those estimated here. Water harvest analyses (i.e. where water is extracted or diverted from a river into an offstream storage) indicated that uncertainty in water harvest estimates (using all suitable model versions) were more sensitive to particular aspects of flow duration, particularly the match of simulated and observed data at flow values around pump start threshold flows, than to total model bias. An analysis at a single node suggested a range of plus or minus 15% variation for water harvest volumes using four different model versions. The availability of soils suitable for irrigation and ringtank construction in the Roper catchment is limited, and when combined with streamflow data the maximum possible annual water harvest volume that could be extracted or diverted into offstream storages at a minimum of 75% annual reliability was 660 GL, although this assumed a very low pump start threshold (200 ML/day) and did not consider any other restrictions due to economic, environmental, cultural or land tenure influences. It is highly likely that potential surface water allocations in the Roper catchment would be accompanied by water licence conditions that seek to mitigate the impact of water extraction on water dependent ecosystems. Such conditions would substantially reduce the volume of water that could potentially be harvested and/or the reliability of extraction/diversion. The behaviour, yield and effects of streamflow at five hypothetical dam sites in the Roper catchment were evaluated for the purpose of understanding the sensitivity of water dependent ecosystems to large instream dams. Analyses considered various dam full supply levels (FSLs) and transparent flow thresholds to determine the most cost-effective FSL in each case. These parameters were then used to simulate the effect of individual dams, two dams and five dams simultaneously. It was found that five dams could release 324 GL in 85% of years, which is modest relative to the potential to regulate water using dams in catchments elsewhere in northern Australia. In considering the likelihood of dam based development in the Roper catchment, it should also be noted that no large dams have been built in northern Australia west of the Great Divide for the purpose of irrigation for over 40 years. Hypothetical groundwater developments from the Cambrian Limestone Aquifer near Larrimah and to the south of Larrimah (i.e. 35 GL/year in addition to the 32.5 GL of existing groundwater entitlements near Mataranka) were modelled to have about a 9% reduction in groundwater discharge to the Roper River in the vicinity of Mataranka and in areas downstream at about the year 2060. However, due to the time lags associated with groundwater flow processes, the majority of the reduction in groundwater discharge was due to the existing groundwater entitlements in the vicinity of Mataranka and not the additional hypothetical 35 GL/year further to the south. Effects on mean annual flow were relatively modest at a catchment scale (i.e. a reduction of mean annual flow of 8 GL at outlet of the Roper River). Future climate analyses in the Roper catchment indicates that there is a large uncertainty in future rainfall. The general circulation model (GCM) future rainfall for 32 GCMs ranges from a reduction of about 12% to an increase of 15%. Five (or 16%) of the GCM projections indicate an increase in mean annual rainfall by more than 5%, nine (or 28%) of the projections indicate a decrease in mean annual rainfall by more than 5%, and 14 (or 56%) of the projections indicate a change in future mean annual rainfall of less than 5% under a 1.6 °C warming scenario. Under a dry future climate (Scenario Cdry) the mean annual rainfall was assumed to be 10% lower than the 1910– 2019 catchment mean annual rainfall and under a wet future climate (Scenario Cwet) it was assumed to be 10% higher. Propagation of these future climates through the Roper river system model resulted in changes in mean annual streamflow of –35% and +41% under scenario Cdry and Cwet respectively. Under Scenario Cdry the entire flow range was impacted but there were notably larger impacts at high flows. Under Scenario Cdry the mean annual discharge from the Roper River was lower than any of the water resource development scenarios examined, including those scenarios that sought to explore the maximum physically plausible water resource development. Contents Director’s foreword ........................................................................................................................ i The Roper River Water Resource Assessment Team ..................................................................... ii Shortened forms .......................................................................................................................... iii Units ............................................................................................................................ iv Preface ............................................................................................................................. v Executive summary .................................................................................................................... viii Part I Main report 1 1 Introduction ...................................................................................................................... 2 1.1 Surface water activity objectives ......................................................................... 2 1.2 Previous surface water modelling studies in the Roper catchment ..................... 3 2 Site characteristics ............................................................................................................ 4 3 Available data ................................................................................................................... 8 3.1 Stream gauge data ............................................................................................. 10 4 Model software and structure ........................................................................................ 13 4.1 River model ........................................................................................................ 13 4.2 Water harvest analyses ...................................................................................... 21 4.3 The effect of groundwater extraction on streamflow in the Roper River .......... 24 4.4 Future climate analyses ..................................................................................... 31 5 River model calibration ................................................................................................... 33 5.1 Calibration objective functions .......................................................................... 37 5.2 Calibration results .............................................................................................. 39 6 Model scenario analyses ................................................................................................. 47 6.1 Scenario A – historical conditions ...................................................................... 48 6.2 Scenario B – development under historical conditions ...................................... 50 6.3 Scenario C – future climate streamflow ............................................................. 66 6.4 Scenario D – future climate and water resource development .......................... 69 7 Discussion and conclusion ............................................................................................... 72 References ........................................................................................................................... 75 Part II Appendices 81 Morton’s Wet Area Potential ET calculation ...................................................... 82 GR7J modification .............................................................................................. 86 Stream gauge data in the Roper catchment ....................................................... 88 River model benchmark plots .......................................................................... 103 Figures Preface Figure 1-1 Map of Australia showing Assessment area ................................................... vi Preface Figure 1-2 Schematic diagram of the high-level linkages between the 8 activities and the general flow of information in the Assessment. ......................................................................... vii Figure 2-1 Roper River Water Resource Assessment area showing the Roper River and tributaries, physiographic provinces after Plumb (1992), significant settlements and roads overlaid on hill-shaded terrain relief ............................................................................................. 5 Figure 2-2 Annual rainfall at Mataranka and Ngukurr under Scenario A: (a) annual rainfall at Mataranka, and (b) annual rainfall at Ngukurr. Scenario A is the historical climate (1910 to 2019). The blue line represents the 10-year running mean .......................................................... 6 Figure 2-3 Soil versatility in the Roper catchment ........................................................................ 6 Figure 2-4 Median annual streamflow (50% exceedance) in the Roper catchment under Scenario A ..................................................................................................................................... 7 Figure 3-1 The distribution of rainfall and stream gauge data in the Roper catchment in 1965 ... 9 Figure 3-2 Stream gauge data for site 9030001 (Elsey Creek at Warlock Ponds) ........................ 11 Figure 3-3 Roper catchment annual rainfall (yellow columns), and stream gauge observation period expressed as mean contributing area rainfall in the respective years of observation ..... 12 Figure 4-1 Floating evaporation pan used to determine water surface evaporation pictured here on the Roper River ...................................................................................................................... 18 Figure 4-2 River model nodes and subcatchment areas ............................................................. 20 Figure 4-3 Simulated annual flow at Ngukurr (90301780) and Red Rocks (90302500) ............... 22 Figure 4-4 The extent of limestone aquifers in relation to the Roper surface water catchment . 26 Figure 4-5 Modelled groundwater discharge to the Roper River upstream of Elsy Homestead (9030013) .................................................................................................................................... 27 Figure 4-6 Minimum modelled October flow across 109 years for all nodes between 90300890 and 90300000 (inclusive) ............................................................................................................ 29 Figure 4-7 Estimates of groundwater discharge at node 90300130 from FEFLOW and via river model emulator .......................................................................................................................... 31 Figure 5-1 Shingle calibration conceptual diagram for a section of five nodes of a river system model. Each calibration shingle has a pre-calibrated input time series (blue), a calibration node (green), and an auxiliary calibration node (red) that is jointly calibrated with green nodes ....... 35 Figure 5-2 Conceptual diagram of a six-node calibration shingle, of which three nodes are ungauged .................................................................................................................................... 37 Figure 5-3 Estimated mean streamflow for various model versions at Red Rocks (n = 10) and Roper end-of-system node 90300000 (n = 7) .............................................................................. 40 Figure 5-4 Median cumulative water harvest volumes for various model versions at node 90300110 using a pump start threshold of 500 ML/day and an unlimited pump capacity ......... 41 Figure 5-5 Flow duration curves for four model versions and observed flow at node 90302500. Horizontal broken lines indicate flow values of 500 and 200 ML/day for reference ................... 41 Figure 5-6 Runoff coefficient vs aridity for four different model versions at the Red Rocks (90302500) model node by decade. Labels indicate the start of each decade for some clusters of data ......................................................................................................................................... 42 Figure 5-7 Simulated mean runoff coefficient by subcatchment for model version 3 ................ 43 Figure 5-8 Simulated mean runoff coefficient by subcatchment for model version 12 .............. 44 Figure 6-1 Annual reliability of irrigation supply for various pump start threshold and irrigation target volumes at node 90300110 for: (a) pump rate 20 days, (b) pump rate 40 days, and (c) pump rate 20 days with an end-of-system requirement of 700 GL/year. Hatched white target volumes are unlikely to be possible due to restricted availability of suitable soils for irrigation in the catchment ............................................................................................................................. 52 Figure 6-2 Annual reliability of supply at node 90301780 using a pump rate of 20 days and an end-of-system requirement of 700 GL/year................................................................................ 52 Figure 6-3 The 80% annual exceedance streamflow relative to Scenario A in the Roper catchment using a pump start threshold of 600 ML/day and a pump capacity of 20 days for various system irrigation targets and end-of-system annual requirement at node 90301780. Hatched white target volumes are unlikely to be possible due to restricted availability of suitable soils for irrigation in the catchment ............................................................................... 53 Figure 6-4 Timing of the first day of annual pumping for end-of-system (node 90301780) requirement across 109 years of model simulation, expressed as a frequency density ............. 55 Figure 6-5 Reservoir surface area related to dams and a selection of river model nodes ........... 57 Figure 6-6 The effect of full supply level and transparent flow threshold on 85% exceedance annual yield, unit capital cost, net mean annual reservoir evaporation and reservoir mean annual spill at site 55 .................................................................................................................. 58 Figure 6-7 The effect of full supply level and transparent flow threshold on 85% exceedance annual yield, unit capital cost, net mean annual reservoir evaporation and reservoir mean annual spill at site 79 .................................................................................................................. 59 Figure 6-8 The effect of full supply level and transparent flow threshold on 85% exceedance annual yield, unit capital cost, net mean annual reservoir evaporation and reservoir mean annual spill at site 108 ................................................................................................................ 60 Figure 6-9 The effect of full supply level and transparent flow threshold on 85% exceedance annual yield, unit capital cost, net mean annual reservoir evaporation and reservoir mean annual spill at site 137 ................................................................................................................ 61 Figure 6-10 The effect of full supply level and transparent flow threshold on 85% exceedance annual yield, unit capital cost, net mean annual reservoir evaporation and reservoir mean annual spill at site 145 ................................................................................................................ 62 Figure 6-11 Percentage change in mean annual rainfall and potential evaporation under Scenario C relative to under Scenario A ...................................................................................... 66 Figure 7-1 A comparison of flow duration curves for various scenarios at (a) and (b) node 90300000 and (c) and (d) node 90302500. ................................................................................. 74 Tables Table 3-1 Stream gauge characteristics in the Roper catchment ................................................ 10 Table 3-2 Runoff coefficients calculated for five Roper catchment gauges ................................. 12 Table 4-1 Loss parameters applied to river model nodes in the mid-Roper ................................ 28 Table 4-2 Scenario effects on groundwater discharge at node 90300130 relative to undeveloped conditions (scenario A) at the year 2060 ..................................................................................... 30 Table 5-1 Rainfall-runoff model and objective function used for various river model calibrations .................................................................................................................................................... 38 Table 5-2 Parameter transfer protocol for model version 12 ...................................................... 45 Table 5-3 Goodness-of-fit statistics for model version 12 ........................................................... 45 Table 6-1 Flow statistics for all model nodes under Scenario A .................................................. 48 Table 6-2 Median annual flow and proportional water harvest allocation for Roper water harvest analyses .......................................................................................................................... 50 Table 6-3 Water harvest parameters and values analysed for Roper water harvest analyses .... 50 Table 6-4 Estimated maximum soil-limited water harvest assuming a pump rate of 20 days and a pump start threshold of 200 ML/day. Values were calculated to ensure a minimum of 75% annual supply reliability .............................................................................................................. 54 Table 6-5 Residual streamflow at selected model nodes for Scenario A and three water harvest scenarios where reliability of supply is maintained to at least 75% in all water harvest nodes. . 55 Table 6-6 Dam full supply levels, yields at 85% annual exceedance and transparent flow thresholds used for combined five dams simulations ................................................................. 63 Table 6-7 Mean annual flow and exceedance flow for dam affected gauges for Scenario A and five dams simulations with and without transparent flow .......................................................... 63 Table 6-8 Mean annual flow and flow rate at 90, 80, 50, 20 and 10% exceedance values for various river model nodes for Scenario A and flow subsequent to hypothetical groundwater development in the Larrimah area .............................................................................................. 65 Table 6-9 Annual flow exceedance values, mean runoff coefficient and mean annual flow for three future climate scenarios .................................................................................................... 67 Table 6-10 Mean annual flow relative to Scenario A for three future climates at two river model node locations ............................................................................................................................. 69 Table 6-11 Mean annual flow and annual flow exceedance values for Scenario D (dry + water harvest) ....................................................................................................................................... 70 Table 6-12 Mean annual flow and annual flow exceedance values at the end-of-system (90300000) for Scenario B (water harvest), Cdry, D (dry climate + water harvest) relative to Scenario A. .................................................................................................................................. 70 Table 6-13 Mean annual flow and annual flow exceedance values for Scenario D (dry + five dams) .......................................................................................................................................... 71 Table 6-14 Mean annual flow and annual flow exceedance values at the end-of-system (90300000) for Scenario B (five dams), Scenario Cdry, and Scenario D (dry climate + five dams) relative to Scenario A .................................................................................................................. 71 Part I Main report 1 Introduction The regulation of surface water resources in southern Australia meets about 70% of Australia’s 25,000 GL mean annual water use (CSIRO, 2011). With the overallocation of water in southern states, the recent millennium drought and projections of a drier future climate in southern Australia, there is interest in developing the water resources of northern Australia. However, the extraction of water from rivers, particularly for high water-using industries such as irrigation, can result in large perturbations to streamflow, which can affect existing industries and users and result in ecological change. To quantify the water resources of a catchment and examine the trade-offs associated with water regulation and extraction, a variety of event-based and continuous hydrological modelling frameworks exist. However, different hydrological modelling frameworks have been developed for different purposes and they have different data requirements and different levels of complexity. At their simplest, hydrological models can be simple statistical relationships, typically with few input data requirements, but can also have a low predictive capacity. At the more complex end are fully distributed physically based models, for which every parameter has physical meaning and can be assigned by measurement. These include soil–vegetation–atmosphere transfer models such as WAVES (Zhang and Dawes, 1988) and TOPOG (O’Loughlin, 1986) and some landscape models. These models can simulate a wide variety of processes and are useful for exploring scenarios that have not been previously observed in the historical record. However, a key challenge in using physically based models is that they have large data requirements, without which many parameters potentially need to be calibrated, which makes them difficult to apply with confidence, particularly across large areas. In between these two extremes are a wide variety of models of intermediate complexity, including those described in this report: • lumped conceptual rainfall-runoff (RR) models (e.g. Sacramento, GR4J, RORB), which are particularly adept at modelling runoff • river system models (e.g. Source, IQQM, AWRA-R), a genre of hydrological model well suited to modelling regulated systems and exploring trade-offs in water use, operation and management rules In selecting an appropriate model or suite of models, it is important to understand the modelling objectives and select a model that is commensurate with the level of data available and then to be cognisant of its predictive capacity and model limitations. 1.1 Surface water activity objectives The Roper River Water Resource Assessment surface water activity seeks to address the following questions: • How do runoff and key water balance terms vary spatially and temporally across the three Assessment areas? • How much water is in the river at different locations and times under current and future climates? • How much water can be extracted in different reaches and with what degree of reliability, and what is the timing of potential extractions? • How may water regulation and extraction perturb downstream flow? This report describes the development of models to assess these items. 1.2 Previous surface water modelling studies in the Roper catchment In 1976 the Australian Water Resources Council (AWRC) oversaw an assessment of Australia’s water resources. In that study, each jurisdiction across Australia provided estimates of mean annual flow and the percentage of mean annual flow that could be diverted in each AWRC river basin in Australia. No other information was provided, and the methods used to make these estimates varied from one jurisdiction to another and were not documented. Thirty years after the AWRC continental assessment of Australia’s water resources the Australian Government commissioned CSIRO to undertake the Northern Australia Sustainable Yields (NASY) project, which was the first hydrological modelling study to examine the water resources of northern Australia (Timor Sea, Gulf of Carpentaria and northern north-east coast drainage divisions), including the assessments of the three study areas, using a consistent set of methods and models (CSIRO, 2009a, 2009b, 2009c). As part of the NASY project, lumped conceptual RR models were calibrated to streamflow data from 125 gauged catchments in northern Australia and then model parameters were transposed to another 500 ungauged catchments using the nearest neighbour (NN) regionalisation method, extensively informed by expert knowledge (Petheram et al., 2009). The lumped conceptual RR models were calibrated to available observed data up to 31 August 2007. Due to time constraints, no new river system models were developed as part of NASY and at the time of the project only four river system models (IQQM models) existed in the northern-draining drainage divisions. Within the catchment of the Roper River, RR models were calibrated to a single, reliable stream gauge. This model was then used to estimate streamflow within the Roper River and simulate streamflow under historical and projected future climates. Subsequent to the NASY study, CSIRO was commissioned to assess the opportunities for water resource development and associated risks in the Flinders and Gilbert (Qld) catchments between 2012-2013 (Petheram et al., 2013a,b) in an assessment known as FGARA, and then in the Fitzroy (WA), Darwin (NT) and Mitchell (Qld) catchments in 2016-2018 in an assessment known as NAWRA. These Assessments employed river models, landscape models and hydrodynamic models to estimate the effects of development scenarios on water resources. Specific information with regard to river and landscape modelling in FGARA is detailed by Lerat et al. (2012) and Holz et al. (2013) and in NAWRA it is detailed in Hughes et al. (2017) and Hughes et al. (2018). The river model and landscape model methods used in this Assessment are largely an evolution of the methods used in the NAWRA. It should also be noted that there is no jurisdictional or any legacy Roper River model that can be referred to for this Assessment. 2 Site characteristics The physical characteristics of the Roper catchment are given in detail in the companion technical report Resource information for assessing potential development opportunities. To assist the reader, an overview of the site characteristics is included here. The Roper catchment is 77,432 km2 in area and features flat, tidally affected coastal plains that extend 20 to 60 km inland and that typically lie at less than 10 mAHD and are prone to seasonal flooding. The Roper River extends approximately 300 km from the east of the river mouth with major tributaries, the Wilton River and the Hodgson River, entering the river mid-catchment from the north and south respectively (Figure 2-1 ). In headwater catchments situated in the north- western part of the catchment, altitudes reach up to 420 mAHD. The Roper catchment is characterised by a distinctive wet and dry season due to its location in the Australian summer monsoon belt. The mean annual rainfall over the Roper catchment for the 109- year historical period (1 September 1910 to 31 August 2019) is 799 mm. Annual rainfall is highest in the northern part of the catchment and lowest in the most southerly part the catchment. This is because the more northerly regions of the catchment receive more wet-season rainfall as a result of active monsoon episodes. The Roper catchment is relatively flat, and consequently there is no noticeable topographic influence on climate parameters such as rainfall or temperature. Approximately 95% of rain falls in the Roper catchment during the wet-season months (1 November to 30 April)(Figure 2-2). The lack of rainfall during the dry season is largely due to the predominance of dry continental south-easterlies and the significant dry air aloft that inhibits shower and thunderstorm formation. The Roper catchment has a mean annual potential evapotranspiration (ET) of 1917 mm (1965–2015) and like rainfall, has a relatively strong north– south gradient across the catchment. The 10-year running mean provides an indication of the sequences of wet or dry years (i.e. variability at decadal timescales, Figure 2-2). For an annual time series, the 10-year running mean is the average of the 5 years of data either side of every annual data point. Under Scenario A, PE exhibits much less inter-annual variability than rainfall (not shown, see companion technical report on climate (McJannet et al., 2023)). The Roper catchment is dominated by soils of a sandy surface texture that are usually associated with Gulf fall and Wilton River physiographic province (Figure 2-3). Soils of higher clay content, usually considered suited to a larger range of irrigated crops, are generally located close to streamlines or in more extensive areas on the Sturt Plateau. Crop versatility, an index that reflects the ability of a soil to support a diverse range of crops and pastures, also reflects this distribution of clay soils (Figure 2-3). For more information regarding soil versatility see the companion technical report (Thomas et al., 2023). In terms of irrigation using surface water, most of the versatile soils are not viable for irrigation due to very low streamflow on the Sturt Plateau (Figure 2-4 and Section 3.1). Figure 2-1 Roper River Water Resource Assessment area showing the Roper River and tributaries, physiographic provinces after Plumb (1992), significant settlements and roads overlaid on hill-shaded terrain relief Map. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Figure 2-2 Annual rainfall at Mataranka and Ngukurr under Scenario A: (a) annual rainfall at Mataranka, and (b) annual rainfall at Ngukurr. Scenario A is the historical climate (1910 to 2019). The blue line represents the 10-year running mean Figure 2-3 Soil versatility in the Roper catchment Chart, histogram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Map. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\3_Roper\1_GIS\1_Map_docs\1_Exports\LL-R-532_Versatility14_v3_10_5.png" Figure 2-4 Median annual streamflow (50% exceedance) in the Roper catchment under Scenario A "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\1_GIS\1_Map_docs\1_Exports\Hy-R-505_Roper_accumulated_AnnualMedian_flow_(E50)_forVideo.png" For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 3 Available data The quality and quantity of the data available to the modeller has a significant influence on the methods used to estimate streamflow, as well as the potential model applications. The coincidence of climate data and streamflow data in space and time are of particular interest since one without the other means model calibration is not straightforward. To that end, the availability of rainfall data, and the availability of stream gauge data was examined for the catchment area on a year-by-year basis to better understand the spatial coincidence of these data (or lack thereof). A series of maps was produced that summarise these data and are provided as supplementary material (companion report). An example of these maps is shown in Figure 3-1 (the first of these figures in the time series and arbitrarily chosen as an example only)Here, rain gauge locations are indicated, along with stream gauge locations and river model subcatchment boundaries that contribute to each gauge. For each calendar year, data completeness is expressed as a proportion of 365 days of data and indicated by circle size (rain) or catchment colour (stream). Rain gauge data was obtained from Queensland Government website as ‘patched point’ data. Only direct observations were utilised for these analyses, while interpolated data were discarded. Stream gauge data were obtained from the Northern Territory Department of Environment, Parks and Water Security (DEPaWS). High-frequency data were aggregated to 9 am for comparison to daily rainfall data. Examination of these data over time shows that distribution of rain gauge data in the southern half of the Roper catchment is generally quite good, while the coverage in the northern portions of the catchment, particularly the Wilton River subcatchment, is poor up until around 2002, after which a single rain gauge in the upper Wilton River has recorded observations (see supplementary material). Unfortunately, all stream gauging had ceased in the Wilton River at this time, meaning that there are no concurrent streamflow and rainfall observations for this major tributary of the Roper River. Obviously, this has consequences primarily for model calibration but also for simulation, given that rainfall estimates are also more uncertain in this area than other parts of the catchment. It should be noted that, for model simulation, ‘patched point’ data are not used, but rather gridded Data Drill climate data from the same provider. This is an interpolated product but relies upon the same rainfall data that has been reviewed here, and therefore is subject to the uncertainties mentioned above. Data Drill data are a daily product supplied at approximately a 5 × 5 km grid resolution. Data from January 1900 until November 2019 were used as model input, although the reporting period begins in 1910. This allows model states to ‘warm-up’, prior to analysis of model outputs. It should also be noted that errors in the rain gauge observations, the availability of gauges at any point in time and the spatial extent of rainfall will all impact the accuracy of interpolated rainfall products. Figure 3-1 The distribution of rainfall and stream gauge data in the Roper catchment in 1965 Size of the green circles indicates the completeness of rain gauge data at that site in the year 1965, while the colour ramp indicates the completeness of stream gauge data for contributing areas in the year 1965. Refer additional maps in supplementary material. Data Drill data were bulk downloaded as spatial layers (netCDF format) and aggregated to daily time series for each grid cell. As a part of this process Morton’s wet area (Mwet) potential ET was calculated using other Data Drill variables. Mwet is an estimate of potential ET over a large area, assuming an unlimited supply of water. The model assumes upwind effects are negligible and local variations are ignored, so the estimate is an areal average (Wang et al., 2001). Chiew and McMahon (1991) found Mwet is similar to Food and Agriculture Organization – Irrigation paper 56 Map. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au (FAO56) (Allen et al., 1998) in a wet climate but lower than FAO56 in a dry climate. Chiew and Leahy (2003) found that Mwet is similar to FAO56 in the coastal areas of south-eastern and eastern Australia. It is possible therefore, that the wet season Morton’s ET used in this study is similar to FAO56, but potentially somewhat lower during the dry season. Morton’s wet area potential ET calculations are detailed in Appendix A. All gridded data was subsequently spatially averaged according to model sub-catchment for use an input in the river models. 3.1 Stream gauge data All stream data, including gaugings and quality codes, were obtained from DEPaWS. These data were examined together to determine the suitability of individual gauges for calibration. Quality code information and plots of all these data are given in Appendix C. An example of these plots is shown in Figure 3-2. These plots assist the hydrologist making judgments of the value of stream gauge data at each site. Gauging period, catchment area and catchment averaged mean annual rainfall are given for Roper catchment stream gauges in Table 3-1. It should be noted that other gauges, usually of a short duration (typically weeks), or with obvious problems (e.g. G9030514), were not assessed further and not subsequently used. Table 3-1 Stream gauge characteristics in the Roper catchment See Figure 4-2 for gauge locations. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au † Gauging has ceased at this site. ‡ Based on daily rainfall data from 1910 to 2019. However, some aspects of gauge quality are not easily quantified, and require more intuition/judgment. For example, a sudden increase in baseflow that is not noted in other gauges at a similar time can increase doubt on the veracity of the observations. Another example is where the peak flows are unusually consistent, or more obviously, where the rating data are poor. Figure 3-2 Stream gauge data for site 9030001 (Elsey Creek at Warlock Ponds) The dashed red line in the top and bottom left panel indicates highest gauged flow. Quality codes are given by yellow points in the top panel, while grey vertical lines indicate missing data A notable feature of the Roper catchment is the very low runoff coefficient recorded at the streamflow gauge on the Sturt Plateau (gauge 9030001, Table 3-2 ). The streamflow record for this site, and the other selected gauges shown in Figure 3-2, is considered good, being of a long duration and with a good range of gauging data to constrain flow estimates (Figure 3-2). Examination of aerial images of the Sturt Plateau also indicate that the area has a very low stream density with numerous sinkholes. This geomorphic property is generally associated with a relatively high regolith transmissivity or infiltration capacity (Carlston and Langbein, 1960; Horton,1945; Strahler, 1956) and is a feature in some basaltic landscapes. Histogram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Table 3-2 Runoff coefficients calculated for five Roper catchment gauges For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au † Runoff coefficient was calculated using observed streamflow between the dates listed in Table 3-1 and missing observations in these periods were patched with modelled data. Conversely, permanent springs and a high river baseflow are observable immediately north of the Sturt Plateau around Mataranka (i.e. in the vicinity of gauge 9030013). See Appendix C (Apx Figure C-3) for more information on gauge 9030013, noting that flow rarely falls below 2 m3/second. Ultimately, stream gauge data are used to derive model parameters and assess the veracity of simulation based on these parameters. Stream gauge data context is an additional consideration here, particularly the climate during observation in relation to the simulation period. These data can be seen in Figure 3-3. The catchment annual rainfall is shown against the climate of the stream gauge contributing area for each gauge during each respective observation period. What is most obvious is that stream gauge data have been collected in a relatively wet period, and there is no data regarding hydrological response in the pre-1960 period. This situation is further exacerbated for gauges 9030003, 9030108, 9030102 and 9030146, where stream gauging is of a relatively short duration and in all cases bridges a relatively wet period in the 1970s. Given the location of these gauges (Table 3-1), data for both the Hodgson and Wilton rivers becomes questionable. Figure 3-3 Roper catchment annual rainfall (yellow columns), and stream gauge observation period expressed as mean contributing area rainfall in the respective years of observation The blue line is the smoothed catchment annual rainfall data Chart, bar chart, histogram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 4 Model software and structure 4.1 River model The AWRA-R model has been selected as a modelling platform for The Assessment. It was selected over ‘Source’ (a commonly used platform by jurisdictions), largely since it is flexible and has very short run times. AWRA-R is not designed to incorporate complicated operations rules as Source might, although for relatively un-developed areas such as northern Australia this presents few, if any, simulation difficulties. Rather, the flexibility and short run times allows for extensive sensitivity analyses of development scenarios, automated model optimisation and an ability to make the model available to users via a website where simulations can be run ‘live’ with development parameters of their choosing and accompanying ecosystem analyses. The AWRA-R model is based upon a series of connected subcatchments which can receive streamflow from upstream nodes, perform various process within each subcatchment, and, using a water balance approach, calculate various fluxes including subcatchment outflow, which may be used as an input to a downstream subcatchment. Outflow points for each subcatchment are generally denoted as a ‘node’. Model parameters and inputs are required for each subcatchment. The AWRA-R model framework used for the Roper River is written in the C language and is used in conjunction with the R language for ease of data processing and access to various functions such as optimisers, goodness-of-fit measures and plotting functions via R packages. A brief summary is provided here, otherwise the reader is directed to the original references of Dutta et al. (2015) and Dutta et al. (2017). Each node in the model requires a configuration vector, a parameter vector, and a time series array as inputs. The standard model output is a time series of model states, including outflow. Where irrigation sub-models are used, these require additional irrigation parameters and configuration vectors, as well as additional time series inputs (e.g. crop coefficient values). 4.1.1 Routing routine Routing represents the transport of water down a river reach from upstream to downstream. A river channel has capacity to store water in varying degrees, which induces a time lag from inflow to outflow, as well some attenuation and dampening of the hydrograph peak across the reach. These effects were simulated using a lagged Muskingum procedure (Koussis, 1980): 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑=𝐼𝐼(𝑡𝑡−𝑙𝑙𝑙𝑙𝑙𝑙)−𝑂𝑂 (1) and 𝑉𝑉(𝑡𝑡)=𝐾𝐾[𝑥𝑥∗𝐼𝐼(𝑡𝑡−𝑙𝑙𝑙𝑙𝑙𝑙)−𝑂𝑂(1−𝑥𝑥)] (2) where V is the routing volume (L3), I and O are the reach inflow and outflow respectively (L3/T), and t is time. K, x and lag are calibrated routing parameters. 4.1.2 Loss model Physically, as water moves along a channel it may experience losses due to exchanges with groundwater or soil water. Transmission losses are very difficult to measure directly, so any loss function that is calibrated jointly with other states (against observed flow) will also implicitly account for gauge error, poor system representation, or overestimates in other model states (e.g. unaccounted runoff). Most model estimates of loss are flow based: 𝐿𝐿(𝑡𝑡)=𝑓𝑓(𝑄𝑄(𝑡𝑡)) (3) where 𝐿𝐿 is the loss and 𝑓𝑓is a function describing the flow-based loss. For the Assessment, the loss estimation method developed by Doble et al. (2012) was used. This method is dependent on physical properties of the river bed material, river stage, river length, river width, depth to groundwater and specific water yield of the aquifer. This method is denoted as ‘Doble loss’ subsequently. The river hydraulic conductivity parameter is calibrated jointly with the RR and routing parameters for each reach. Optimal parameter sets are then used for subsequent simulation. The equations for the Doble loss calculations are given below: 𝑄𝑄𝑔𝑔𝑔𝑔=min(𝐼𝐼𝑟𝑟𝑟𝑟𝑟𝑟,Δ𝑆𝑆𝑟𝑟𝑟𝑟𝑟𝑟+ 𝑄𝑄𝑟𝑟𝑟𝑟𝑟𝑟)∗Λ (4) 𝐼𝐼𝑟𝑟𝑟𝑟𝑟𝑟=𝐾𝐾𝐾𝐾𝑟𝑟𝑟𝑟𝑟𝑟𝑎𝑎𝑟𝑟𝑟𝑟𝑟𝑟􁉀ℎ𝑟𝑟𝑟𝑟𝑟𝑟 𝑑𝑑𝑐𝑐 +1􁉁(5) Δ𝑆𝑆𝑟𝑟𝑟𝑟𝑟𝑟=𝑑𝑑𝑔𝑔𝑔𝑔𝑆𝑆𝑦𝑦𝑥𝑥𝑤𝑤 (6) 𝑄𝑄𝑟𝑟𝑟𝑟𝑟𝑟=𝐾𝐾𝑎𝑎𝑎𝑎𝑑𝑑𝑎𝑎𝑎𝑎 2ℎ𝑟𝑟𝑟𝑟𝑟𝑟 𝑥𝑥𝑤𝑤 (7) where 𝐼𝐼𝑟𝑟𝑟𝑟𝑟𝑟 is potential infiltration rate from river (m2/second), Δ𝑆𝑆𝑟𝑟𝑟𝑟𝑟𝑟 is total storage available (m2/second) within the regolith beneath the streambed, 𝑄𝑄𝑟𝑟𝑟𝑟𝑟𝑟 is maximum volume of water discharging from the aquifer (m2/second), Λ is the reach length (m), 𝐾𝐾𝐾𝐾𝑟𝑟𝑟𝑟𝑟𝑟 is river bed hydraulic conductivity (m/second), 𝑎𝑎𝑟𝑟𝑟𝑟𝑟𝑟 is surface area of the river (m2) obtained from flow-area relationship, ℎ𝑟𝑟𝑟𝑟𝑟𝑟 is depth of river water (m) obtained from flow–depth relationship, 𝑑𝑑𝑔𝑔𝑔𝑔 is depth to groundwater (m), 𝑑𝑑𝑐𝑐 is the thickness of the river bed material, 𝑆𝑆𝑦𝑦 is the aquifer-specific water yield (dimensionless), 𝑥𝑥𝑤𝑤 is the width of the river (m) derived from a flow–width relationship, 𝐾𝐾𝑎𝑎𝑎𝑎 is the aquifer hydraulic conductivity (m/second) and 𝑑𝑑𝑎𝑎𝑎𝑎 is the aquifer thickness. In all cases, depth to groundwater information was not available at appropriate spatial and temporal resolution. Accordingly, depth to watertable was assumed static at 5.0 m. The river bed conductivity was calibrated jointly with runoff and routing parameters. Effectively, the 𝑄𝑄𝑔𝑔𝑔𝑔 calculations are simplified to an estimation of 𝐼𝐼𝑟𝑟𝑟𝑟𝑟𝑟 that is at least partly controlled by the calibrated parameter 𝐾𝐾𝐾𝐾𝑟𝑟𝑟𝑟𝑟𝑟, since 𝑄𝑄𝑞𝑞𝑞𝑞 is taken as the minimum of the two terms 𝐼𝐼𝑟𝑟𝑟𝑟𝑟𝑟 and Δ𝑆𝑆𝑟𝑟𝑟𝑟𝑟𝑟+ 𝑄𝑄𝑟𝑟𝑟𝑟𝑟𝑟, and the final two states Δ𝑆𝑆𝑟𝑟𝑟𝑟𝑟𝑟+𝑄𝑄𝑟𝑟𝑟𝑟𝑟𝑟, are likely to be higher due to depth to groundwater assumptions. This method was favoured partly since it is physically based, but also since it requires only one calibrated parameter and can be applied easily to ungauged locations, requiring an estimate of reach length and assuming the parameters from donor catchments. 4.1.3 Irrigation demand and diversion The irrigation demand modelling is undertaken using the AWRA-R irrigation demand model (Hughes et al., 2013b, 2014b). The AWRA-R irrigation model features a soil water store that represents the water balance for an entire irrigation area within individual reaches. Water is extracted from the virtual soil water store according to demand generated from a crop model. Crop demand is based on the FAO56 method (Allen et al., 1998), using crop factors for sown crops and climate data. As the soil water store becomes depleted, increasing volumes of irrigation demand are triggered. Irrigation demand is zero when the soil store is full. One-dimensional demand is converted to volumetric demands via sown crop area. Sown crop area is determined at a series of crop decision days within the irrigation season. Sown crop area can be adjusted depending upon the volume of available water from each of the three sources: 1. surface water licence (managed irrigation district) 2. on-farm storage 3. groundwater licence. Crop demand for all irrigated crops grown in the reach are determined in the following way: 𝐷𝐷=(𝐾𝐾𝐾𝐾∗𝐸𝐸𝐸𝐸𝑜𝑜∗𝜌𝜌−𝑃𝑃∗𝐴𝐴𝑎𝑎)∗𝐴𝐴𝑐𝑐/𝐿𝐿𝑡𝑡 (8) where 𝐷𝐷 is the total crop demand (m3/second), 𝐾𝐾𝑐𝑐 is the area weighted crop factor (for 1 or more concurrently grown crops), 𝐸𝐸𝐸𝐸𝑜𝑜 is the time step potential ET (m), 𝜌𝜌 is the soil dependant crop water stress (dimensionless, range 0–1), 𝑃𝑃 is the time step rainfall (m), 𝐴𝐴𝑎𝑎 is the proportion of the current irrigation area actively growing crops at the current time step, 𝐴𝐴𝑐𝑐 is the current irrigation area, 𝐸𝐸 is irrigation efficiency (dimensionless) and 𝐿𝐿𝑡𝑡 is the time step length (s). Crop demand is supplied via the soil moisture store, which is in turn supplied via irrigation using the following relationships: 𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼𝐼 = 􁉐 𝐼𝐼𝑚𝑚𝑚𝑚𝑚𝑚 ∗ 𝐴𝐴𝑎𝑎∗𝐴𝐴𝑐𝑐∗𝐸𝐸/𝐿𝐿𝑡𝑡 𝑖𝑖𝑖𝑖 𝜃𝜃𝑡𝑡≤0 𝐼𝐼𝑚𝑚𝑚𝑚𝑚𝑚 ∗ 𝑒𝑒 􁉆 (−1∗𝜃𝜃𝑡𝑡 2) 2𝜎𝜎2􀵘􁉇 ∗ 𝐴𝐴𝑎𝑎∗𝐴𝐴𝑐𝑐∗𝐸𝐸/𝐿𝐿𝑡𝑡 𝑖𝑖𝑖𝑖 𝜃𝜃𝑡𝑡>0 (9) and 𝐼𝐼𝑚𝑚𝑚𝑚𝑚𝑚=𝛾𝛾 𝜎𝜎√2𝜋𝜋 (10) where 𝛾𝛾 and 𝜎𝜎 are user-defined parameters that are adjusted to suit the soil water-holding capacity of the area of interest, 𝐼𝐼𝑚𝑚𝑚𝑚𝑚𝑚 is the highest possible rate of irrigation (m) per time step and 𝜃𝜃𝑡𝑡 is the first estimate of soil water storage for the current time step. When the soil water store is full (say, following rainfall, and 𝜃𝜃𝑡𝑡 = soil capacity), no irrigation is triggered. The model features an on-farm storage module that can extract water from a reach according to user-defined pump parameters, allocation/licence limits and reservoir volumes. Water can then be extracted from the storage as required. Additionally, water can be extracted for irrigation directly from the river, although this feature is more commonly used in larger managed irrigation districts where water is diverted and supplied to irrigators via a channel system. In this Assessment, however, only on-farm storage modules were used (as a part of water harvest analyses). 4.1.4 Rainfall-runoff model RR models take a ‘top-down’ approach to estimating runoff, viz., model parameters are adjusted until the model simulation matches streamflow observations to the satisfaction of the hydrologist. This model can then be used to estimate flow at a time outside of the calibration period (assuming inputs, usually daily precipitation and potential ET, are available). In many situations, no streamflow observations are available at a desired location, and parameters have to be estimated or assumed using other methods. The calculations within the RR models are influenced by observations of hydrological processes, and hence these models are often termed ‘conceptual models’. The ease of use and modest data requirements of these models has seen their widespread application, so much so that these models are applied in a vast array of environments not anticipated (presumably) by the original model authors. Given this, the modeller must take care in their application, especially in environments such as northern Australia. Furthermore, these models are prone to ‘over-fitting’ (i.e. poor predictive performance despite satisfactory representation of observed streamflow during calibration). This is related to the inability of the model to implicitly represent all processes, and fitting to any error in input and streamflow observations. The river model used for the Assessment relies on RR models to generate runoff in each model subcatchment. The Assessment builds upon some previous assessments of RR models in other northern Australian catchments (Hughes et al., 2017). More information regarding these models and choices for this Assessment can be found in Section 5. 4.1.5 Reservoir model Large instream dams have the potential to ‘carry’ water across years and are, therefore, a means of mitigating the impacts of lower rainfall years on irrigation developments. However, disruption to the hydrological characteristics of a stream can also be large, depending upon management, with consequences for ecosystems with a dependency on river flows (Pollino et al., 2018). As an initial assessment, a large range of potential dam sites in the Roper catchment were identified using the DamSite model (Petheram at al., 2023). This model uses a series of algorithms automatically determining favourable locations in the landscape as sites for intermediate to large water storages (Read et al., 2012; Petheram et al., 2017, 2018d). Five dam sites with higher yield per unit cost ratios in distinctly different geographic regions were selected from the output produced by the DamSite model in the Roper catchment (Petheram et al., 2022). At all locations the dam wall dimensions were recalculated using Advanced Land Observing Satellite (ALOS) data to refine the cost of the dam wall and associated infrastructure. Subsequently, relationships were determined for the reservoir stage, reservoir volume and reservoir surface area for each of the analysed dam locations. These relationships formed part of the dam sub-models within the river system model. This allowed assessment of various full supply levels (FSL) for potential yields and reliability of supply. The dam sub-model was configured so that reservoir size and diversion licence volume could be varied to calculate annual reliability of supply for a wide range of reservoir volumes and targets. The reservoir model utilises a water balance equation as follows: 𝑉𝑉𝑡𝑡=𝑉𝑉𝑡𝑡−1+𝑄𝑄𝑙𝑙+𝑄𝑄𝑖𝑖𝑖𝑖−𝐷𝐷𝑡𝑡−𝑆𝑆𝑡𝑡−𝑇𝑇𝑡𝑡+(𝑃𝑃𝑡𝑡−𝐸𝐸𝑡𝑡)∗𝐴𝐴𝑡𝑡 (11) where: 𝑉𝑉𝑡𝑡 is the reservoir volume at time t 𝑉𝑉𝑡𝑡−1 is the reservoir volume at the previous time step 𝑄𝑄𝑙𝑙 is the estimate of local subcatchment streamflow which is, in part, a function of the reservoir surface area 𝑄𝑄𝑖𝑖𝑖𝑖 is the estimate of inflow from all other upstream subcatchments into the reservoir 𝐷𝐷𝑡𝑡 is the diversion out of the reservoir 𝑆𝑆𝑡𝑡 is the dam spill 𝑇𝑇𝑡𝑡 is transparent and/or translucent flow released from the reservoir for environmental purposes 𝑃𝑃𝑡𝑡 is the rainfall on the reservoir surface 𝐸𝐸𝑡𝑡 is the evaporation on the reservoir surface 𝐴𝐴𝑡𝑡 is the surface area of the reservoir at time t. Similar to AWRA-R, the reservoir model was written using C code within an R wrapper. Time series model inputs are: 1. local climate (P and potential ET) 2. reach inflow (from the river model) 3. local runoff (from the river model) 4. daily release pattern for irrigation requirement 5. evaporation correction factor. Additionally, there are various scalar inputs that control such factors as reservoir evaporation adjustment, dam full supply level (FSL) (and height, volume and area relationships for the site), spillway properties, irrigation licence volumes and environmental releases. The evaporation inputs are calculated using the Morton’s wet area algorithm (Morton, 1983). However, these and other typical evaporation estimates have been shown to be different to measured lake evaporation in some instances (Figure 4-1) due to variable fetch conditions. Accordingly, lake evaporation is estimated using dynamic lake area (𝐴𝐴𝑡𝑡) and monthly correction factors. Figure 4-1 Floating evaporation pan used to determine water surface evaporation pictured here on the Roper River The reservoir model features transparent and translucent flow facilities to release water for environmental purposes. Transparent releases are enabled by reservoir inflow thresholds below which 100% of dam inflows are released to the river downstream as if there was no dam present. Importantly, they mimic the range and timing of natural flows. Translucent flows occur at reservoir inflow rates above the transparent thresholds, where a proportion of inflows between the transparent threshold and a higher upper threshold is released. In this Assessment, the potential for transparent flow releases to mitigate impact to water dependent assets was examined and is reported in the companion technical report on ecological modelling, Stratford et al. (2023). This Assessment utilised a release pattern that mimicked demands for a dry-season cotton crop (sown on 1 April). 4.1.6 Node-link structure The number and position of river model nodes across a study area is determined by the study’s modelling objectives. Increasing the number of nodes increases the flexibility of the model in relation to the number and degree of detail possible within model scenarios; however, increasing the number of nodes also increases the computational burden and model run times. For the Assessment, nodes were assigned to: 1. represent stream gauge positions, allowing for model calibration 2. divide the catchment into reasonably evenly distributed subcatchments For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 3.represent communities such as Minyerri and Ngukurr where estimates of water availability may be needed4.potential dam sites (identified using the DamSite model, Petheram et al. 2022) and areas of soil suitable for irrigated agriculture (identified using land suitability data generated by the companion technical report on land suitability, Thomas et al. (2022)). 5.based on the location of ecological assets and where reporting on changes to ecological may be desired. In addition to these influences, nodes were placed in topographic positions that may, at some future time, be suitable for stream gauging, for example, where a stream moves through a narrow gap in a rocky ridge. The final node and subcatchment structure is shown in Figure 4-2 . Figure 4-2 River model nodes and subcatchment areas Note that calibration gauge sites were used as simulation nodes and simulation node ID is the same as calibration gauge ID with the addition of a ‘0’ "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\1_GIS\1_Map_docs\1_Exports\Hy-R-503_Roper_River_Model_v3.png" 4.2 Water harvest analyses Water harvest analyses in this assessment assumes water is pumped from the river into on-farm (or at least offstream) reservoirs without any instream structures. The analysis assumes that river water is extracted by groups or individual irrigators utilising pumps and on-farm storage as a means of supplying and regulating water to irrigated crops. Water harvest scenarios are an attempt to explore trade-offs when water is extracted across a catchment. The goal is to understand how much water can be extracted from the river system at locations where agricultural development is most likely, and how cumulative extractions across the catchment impact on reliability, and how both these factors vary with simple water extraction rules. The analysis is not intended to be prescriptive and is designed to test the extremes of water extraction to understand system dynamics more than evaluate plausible water extraction scenarios. Each water harvest node was given a proportional allocation (i.e. the proportion of the entire river system allocation that could be extracted at that particular node): 𝑎𝑎𝑖𝑖=𝑝𝑝𝑖𝑖∗𝐴𝐴 (12) where: 𝐴𝐴 is the annual system irrigation target (GL) 𝑝𝑝𝑖𝑖 is the proportional irrigation target of node 𝑖𝑖 (dimensionless) 𝑎𝑎𝑖𝑖 is the annual irrigation target for node 𝑖𝑖, such that the sum of proportion node targets in the catchment will sum to 1: Σ𝑝𝑝𝑖𝑖=1𝑛𝑛 𝑖𝑖=1 (13) where 𝑛𝑛 is the number of water harvest nodes in the catchment. The proportional irrigation target was influenced by mean/median annual streamflow for the node and the scale of soils suitable for irrigated agriculture. Opportunities for water harvest at each node will be influenced by node irrigation target, river flow and pump characteristics. In particular, the river flow rate above which the pumping can commence (pump start rate) and the pump capacity (conceptually this could be a physical limit of a pump or it could be a licence condition) will influence the ability to extract water from the river. In order to test the effects of system targets along with pump start rate and pump capacity, combinations for these three variables were tested in each sub-catchment. To test a range of possibilities in terms of water harvest, a range of system irrigation targets were used with a range of pump start thresholds and pump capacities. Rather than use an absolute pump capacity for each permutation of water harvest scenario, relative values were used. More specifically, the pump value was set using the rate by which it would be possible to pump the entire node irrigation target in terms of days. For example, a pump rate of ‘5 days’ would mean that the pump capacity would be high enough to extract the entire node target in 5 days (i.e. if the annual target for a reach was 10 GL, then conceptually a pump rate of 5 days means that all the pumps along that river reach could collectively pump a maximum of 2 GL per day). Minimum pump start thresholds were set at 200 ML/day, a nominal minimum physical threshold at which it was considered that pumping could potentially commence (and assuming suitable waterholes for pump stations). Lastly, we tested the effect of a series of ‘last gauge annual requirements’ for each river. The last gauge annual requirement restricts commencement of water harvesting across the entire Roper catchment until a specified volume has passed the lowermost gauge. The accounting for the last node gauge requirement begins on 1 September each year (the start of the water year). Once the cumulative sum of desired flow has been reached at the lowermost gauge, water harvest can begin for the water year across the river catchment. For the Roper catchment, the ‘lowermost’ gauge for the purposes of this particular analysis was node 90301780 (Roper River at Ngukurr), which is about 42 km below the gauge at Red Rock, the lower most gauge in the Roper catchment. This node was selected because the gauge at Red Rock is located above two major tributaries of the Roper River, the Wilton and Hodgson rivers, and hence only measures about 50% of the streamflow on average. Nonetheless a strong relationship exists between streamflow measured at Red Rock and simulated flow at node 90301780.This relationship may have future utility in relating the results of this analysis to the lower most point in the catchment at which streamflow is measured. Figure 4-3 Simulated annual flow at Ngukurr (90301780) and Red Rocks (90302500) Computationally, each water harvest scenario in each node was enabled by the AWRA-R irrigation model (Hughes et al., 2014b). Typically, the AWRA-R irrigation model represents a crop or series of crops using a specific crop coefficient value (Kc) for each day. However, the aim of the water harvest analysis was to determine if a given annual irrigation target could be extracted given a pump start threshold and pump capacity. Accordingly, for the purpose of this exploratory analysis Kc values were set to 1 for every day in each node. Irrigated area was set to be high enough so that any available water would be used within the year of extraction from the river. On-farm storage was utilised within the model to store extracted water. The size of the on-farm storage was set to equal the node annual irrigation target volume so that on-farm storage would not limit Chart. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au extraction of water. The maximum annual extraction was set to equal the reach irrigation target for each node. All water harvest nodes were simultaneously run within the river model, so that any changes to the flow regime resulting from water extraction is propagated downstream to other nodes. In order to illustrate the opportunity for water harvest, results focus on two aspects of the water harvest simulations, firstly the reliability of supply, and secondly, the resulting effect on streamflow. In terms of supply to irrigation, annual irrigation extraction was compared to the annual allocation across 109 years of simulation. Reliability of supply in each water year was calculated as follows: 𝑧𝑧𝑖𝑖=􁉊 1,Σ𝑒𝑒𝑗𝑗=𝑎𝑎𝑖𝑖 m10,Σ𝑒𝑒𝑗𝑗 m1<𝑎𝑎𝑖𝑖 (14) where: 𝑎𝑎𝑖𝑖 is the annual irrigation target volume (GL) for each river node i 𝑒𝑒𝑗𝑗 is the extraction volume of river water for the node each day j 𝑧𝑧𝑖𝑖 is the extraction success statistic for the water year, with a value of 1 or 0, dependent on the full extraction of the node irrigation target volume 𝑚𝑚 is the number of days in the year. This allows calculation of reliability for each node across the entire time series of simulation: 𝑟𝑟𝑖𝑖=Σ𝑧𝑧𝑖𝑖,𝑡𝑡/𝑡𝑡𝑡𝑡 𝑦𝑦=1 (15) where: 𝑟𝑟𝑖𝑖 is the reliability of irrigation supply in node i t the number of water years of the simulation. The effect of irrigation supply on river flow is judged by comparison to Scenario A (historical river flow with no irrigation development). 4.2.1 Soil-limited water harvest analyses Within the Roper catchment, the availability of soils suitable for irrigation within a reasonable distance of surface water is relatively limited. For this reason, the typical NAWRA-style water harvest analysis (Section 4.2) is less valuable for understanding the physical limitations of irrigation opportunities. For these reasons, an additional water harvest analysis that accounts for both soil and water limitation concurrently was needed. The calculation of the soil-limited water harvest volume calculated both the volume of surface water available to each model node, but also calculated the volume of water required to irrigate a reference crop on all suitable soils within that model subcatchment. The soil-limited water harvest volume (𝑒𝑒𝑠𝑠𝑠𝑠) for the node is taken as the minimum of these three values: 𝑒𝑒𝑠𝑠𝑠𝑠=𝑚𝑚𝑚𝑚𝑚𝑚(𝑒𝑒𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠,𝑒𝑒𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤,𝑒𝑒𝑟𝑟𝑟𝑟) (16) where: 𝑒𝑒𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 is the soil-limited water harvest annual volume (GL) 𝑒𝑒𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 is the annual water-limited diversion volume (see Section 4.2) 𝑒𝑒𝑟𝑟𝑟𝑟 is the volume of water that can be stored on ringtank-suitable soils within 5 km of the river assuming 33% of the arable soil area will need to reserved for water storage. 𝑒𝑒𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠is calculated for the following conditions: 1.Soil area per reach is defined as the area of Class 1 and 2 soils (Thomas et al., 2023) within 5 kmof the main stream within each subcatchment. The available soil area is potentially furtherreduced when accounting for soil spatial continuity (i.e. some areas of Class 1 and 2 soils areisolated by areas of poor soil or topographic features such as larger streamlines) (Thomas et al., 2023). 2.Soil area is further reduced (by 33%) due to the need to build ringtanks, and then a further 20% due to other infrastructure requirements (channels, roads, buildings etc.). 3.Assuming a dry-season crop requires 10 ML/ha on average (including transmission, storage andapplication losses), soil areas are converted to a mean annual water requirement. The soil limitations (𝑒𝑒𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠,𝑒𝑒𝑟𝑟𝑟𝑟) were calculated first and then water harvest analysis was conducted so that irrigation targets did not exceed the calculated soil limit in each reach. Additionally, irrigation volume targets were adjusted so that no reach had an annual reliability of supply less than 75%. 4.3 The effect of groundwater extraction on streamflow in the Roper River Surface water in the form of river flow has, in most cases, some interaction with groundwater, albeit transitory in many situations. In the Roper catchment, surface water–groundwater connections are substantial enough, particularly in the upper Roper River around Mataranka (Figure 2-1), to warrant explicit representation of these processes in the river model. Dry season streamflow in the upper Roper catchment around Mataranka is supplied by aquifers within the Roper River catchment at approximately 4 m3/second of baseflow through the riverbed and springs. The Roper River has continuous inflows along the headwaters from the Cambrian Limestone Aquifer (CLA), as well as more variable discharge from local aquifers. Increased development within the catchment for irrigation or other land use is likely to increase demands on groundwater resources. Additional groundwater abstraction from the high-transmissivity aquifers could lower groundwater levels and thereby reduce the base flow of the river. This is a recognised interaction and models have been constructed for jurisdictional use that attempt to simulate the effects of groundwater development on both groundwater and surface water (Knapton, 2020 and Knapton et al., 2023). The groundwater model consists of a three- dimensional finite element model developed using FEFLOW. The groundwater model was first developed in 2009 and updated in 2018 (Knapton, 2020). The interaction between groundwater and surface water occurs using the FEFLOW transfer boundary conditions. The FEFLOW groundwater model encompasses an area of approximately 159,000 km2 and includes the entire extent of the CLA in the Daly Basin, the northern Wiso Basin, and northern Georgina Basin. These intersect with the Roper surface water catchment but extend far outside it as well (see companion technical report on groundwater modelling, Knapton et al., 2023). The groundwater model was developed with available river geometry and aquifer data. The model was calibrated with available rainfall, river flow, river level, and groundwater level data. The recharge inputs to the FEFLOW model under the eight scenarios were generated using the MIKESHE recharge model. This model was used to examine the effects of three scales of hypothetical groundwater development (extraction) on the Sturt Plateau under current and projected future climates. These scenarios are detailed in Knapton et al. (2023). 4.3.1 Aquifer discharge to the Roper River Most rainfall-runoff models, and river system models based upon rainfall-runoff models, rely on the assumption of mass balance and a “closed catchment”, viz. all fluxes of water within the catchment and any fluxes across the catchment boundary sum to zero. Of note, given the context here, it is often assumed that negligible fluxes of groundwater move across the catchment watershed (other than surface flow at the catchment outlet), since rainfall-runoff models do not generally explicitly recognise groundwater moving across the catchment boundary, and where they do (e.g. GR4J), they are not calibrated against such data, but rather this facility acts as a means of improving streamflow representation. For the Roper catchment, relatively large flows of groundwater, both local and regional occur in the Mataranka region. This cannot be adequately represented by rainfall-runoff models, particularly where groundwater flow across the surface water catchment boundary may be modified by development scenarios. To achieve an adequate representation of streamflow at node location 90300130 (Elsy Homestead), an explicit input of external groundwater flow was required. Estimates of groundwater discharge from the groundwater model in the Mataranka area were included as an input at node 90300130. However, this flow estimate was a combination of modelled local and regional aquifer discharge, which can create some technical problems given that RR models generally represent inputs of local groundwater, since these are usually a component of streamflow, but not regional groundwater fluxes. Since every river model node includes a rainfall- runoff model to estimate ungauged runoff in the node area, local groundwater contribution could be simulated by the rainfall-runoff model, leaving the slower reacting regional groundwater to be represented as an explicit river model input. Figure 4-4 The extent of limestone aquifers in relation to the Roper surface water catchment In terms of river model calibration, a time-filtered product of the modelled groundwater discharge (Figure 4-5) was used to achieve an adequate river model fit at the Elsey Homestead streamflow gauge (i.e. 90300130), noting complications listed above regarding RR models and localised groundwater sources. Practically, this meant iteratively testing various permutations of time filtered, modelled (FEFLOW) groundwater flow. The first iteration of this process was to include the entire estimate of the FEFLOW groundwater discharge from the groundwater model as an "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\11_Groundwater\3_Roper\1_GIS\1_Map_docs\Export\Gr-R-505_hydrogeology_CR_v07.png" input to the river model. However, in that case, this led to an over-estimate of low flow at node 90300130 (due to additional ungauged RR model low flow inputs and inflows to the river model node from upstream). A series of time filtered groundwater flow estimates were iteratively tested until an adequate model fit was obtained. The final “filtered” version of the original GW discharge estimate had a mean flow of 2.10 m3/s, while the original estimate of GW discharge from the groundwater model had a mean flow of 3.19 m3/s. (Figure 4-5). The routine of testing filtered versions of the groundwater discharge as a model input was required largely since RR models generally have problems simulating long-term changes in groundwater storage (Grigg and Hughes, 2018; Hughes and Vaze, 2015) and these problems were further exacerbated by issues associated with regional groundwater discharge. Figure 4-5 Modelled groundwater discharge to the Roper River upstream of Elsy Homestead (9030013) Data sourced from FEFLOW model detailed by Knapton et al. (2023). A filtered timeseries was used as a river model input at node 90300130. Regional groundwater discharging to the Roper River near Mataranka has been measured to have long flow paths and reacts very slowly to changes in recharge (Taylor et al. 2023). At the temporal scale of this assessment (109 years) the discharge of the regional aquifer is almost constant (Knapton et al. 2023). It can be seen in Figure 4-5, that even the filtered version of groundwater input used was still reasonably reactive to short-term climatic changes, and hence is still considered to partly represent localised groundwater. Importantly, this was necessary since the rainfall-runoff model used could not represent flows of similar properties. Examination of gauge data in the Roper River, particularly at Elsy Homestead (9030013) and Red Rock (9030250) indicate substantial dry season losses between those two gauges. Minimum flows during the dry season at gauge 9030013 are generally in the 2 – 4 m3/s range while those at Red Rock are closer to a range of 0.5 – 2 m3/s (Kerle and Cruickshank 2014). As an aside, a prominent feature of streamflow, at least at Red Rock where the record is longer, is that streamflow including low flows has increased across the 20th and into the 21st century. Observed losses during the dry season need to be Chart. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au replicated in the river model reasonably well, since these, in conjunction with reductions in low flow due to groundwater development, are likely to impact on low flow dependant ecosystems. 4.3.2 Simulating streamflow losses in the mid-Roper Theoretically, the Doble loss model (Section 4.1.2) used in previous water resource assessments (Hughes et al., 2017), can be used in a wide range of conditions. However, in this instance, this algorithm (due to a lack of data), was not able to replicate observed losses in the nodes between gauges 9030013 and 9030250. The AWRA-R model does, however, have an anabranch flow facility, that was utilised in the Assessment to generate these losses. In most cases this function is used to estimate the flow of an anabranch after a “split” from the main channel. The anabranch function is as follows; 𝐴𝐴𝑡𝑡=𝛼𝛼𝑄𝑄𝑡𝑡 𝛽𝛽 where 𝐴𝐴𝑡𝑡 is the anabranch flow at time 𝑡𝑡, 𝑄𝑄𝑡𝑡 is flow within the model node (prior to the anabranch), and α and β are model parameters. For most applications, flow in one model anabranch becomes an input to another model node. In this case, the function is used to generate losses, in which case the flow estimates are not returned to other nodes as an input. The calibration process began with by considering no losses while calibrating the reaches from node 9030110 to 90302500, to understand how much excess low flow is apparent at the Red Rock gauge (90302500). This helped to define the magnitude of losses apparent in the Elsy Homestead to Red Rock reach. The desired losses were then distributed along the river model nodes between 90300130 and 90302500 i.e., 90302502, 90302503, 90351440, 90300110, 90302500. Losses were distributed through these model nodes according to the calculated reach length of the main river in each particular node, although these parameters were iteratively adjusted to achieve better calibration goodness of fit at node 90302500. This assumption was necessary since flow data between the gauges 9030130 and 90302500 was sparse. It assumes that the loses are somewhat proportional to the linear distance of the main river channel, although initial losses were assumed to be higher since baseflows will be higher immediately downstream of the regional groundwater discharge area. Final loss parameters are shown in Table 4-1. Table 4-1 Loss parameters applied to river model nodes in the mid-Roper Node ID α β 90302502 1 0.3 90302503 0.007 0.3 90351440 0.2 0.3 90300110 0.085 0.3 90302500 0.6 0.35 The spatial pattern of groundwater discharge and downstream apparent losses need to be represented adequately in the river model. The effects of both modelled groundwater gains and subsequent losses can be seen in Figure 4-6, which depicts the minimum daily flow in October at various river model nodes. October was selected as it is towards the end of the dry season and typically receives very little rainfall (McJannet et al. 2023). It is assumed that, in the majority of years, this is directly related to groundwater flow and loss. Node 90300130 features explicit regional groundwater inputs while all nodes downstream to 90302500 feature losses. Figure 4-6 Minimum modelled October flow across 109 years for all nodes between 90300890 and 90300000 (inclusive) Red dashed line indicates equivalent 200 ML/day flow and dashed blue line indicates equivalent 400 ML/day flow. 4.3.3 Changes in river baseflow with scenario Potential development and changes to the long-term climate can impact the level of baseflow in the Roper River. The most obvious being groundwater extraction in the CLA up-gradient of the regional groundwater discharge to the river around Mataranka. Additionally, for example, a drier future climate will reduce recharge with effects on groundwater levels and baseflow. The effects of various scenarios on groundwater levels and groundwater discharge to the Roper River were quantified using the FEFLOW groundwater model of the CLA (Knapton et al., 2023). Several complications are apparent when trying to understand the effects of groundwater development or long-term changes to climate on groundwater discharge. The most apparent of these is the long time-lags between a change in groundwater management, or groundwater recharge, and a change in groundwater discharge to the Roper River. The timelags are more pronounced with distance from the Roper River. This is related to the long flow paths and residence time of groundwater in the CLA. For example around 32 GL/yr of groundwater extraction is licensed from the CLA in the Chart. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Roper River catchment. The full effects of groundwater abstraction has yet to be realised in groundwater discharge at the Roper River and these effects are modelled to be more apparent at some future time (Knapton et al. 2023). Groundwater modelling indicates that regional groundwater discharge is likely take hundreds of years to achieve equilibrium subsequent to changes in groundwater management or recharge. Additionally, it is thought that increasing rainfall and recharge over the last 40 – 50 years will in part compensate for current licensed groundwater abstractions, although the effects of either on regional aquifer discharge to the river are yet to be fully realised. From a groundwater modelling perspective, the historical input climate sequence of 109 years is inadequate to understand the full effects of any development or future climate scenarios on groundwater discharge. For this reason, an ensemble of climate sequences was run for over 400 years to understand the magnitude of, and time lag of various scenarios. From a practical perspective, the FEFLOW generated changes to groundwater discharge to the river at the year 2060 was used as an index of scenario effects and these were applied within the river model scenario analyses. The change in scenario groundwater discharge relative to observed historical conditions (largely unaffected by groundwater development) is shown in Table 4-2. Table 4-2 Scenario effects on groundwater discharge at node 90300130 relative to undeveloped conditions (scenario A) at the year 2060 Scenario Relative groundwater discharge reduction at 90300130 A 1.00 Acurr† 0.92 B35† 0.91 Cdry† 0.68 Cwet† 0.98 Cmid† 0.77 D35† 0.67 † These scenarios include 32 GL/y of groundwater development already apparent in the study area The method of including time-filtered modelled (FEFLOW) groundwater discharge at node 90300130 satisfied the requirements of an acceptable river model goodness of fit at that node. However, this meant that groundwater discharge in the river model at this location was represented by the time filtered FEFLOW input plus local ungauged runoff estimates from the RR model (which is conceptually representative of localised recharge, which is known to occur around Mataranka based on geochemical analysis of baseflow and hydrographic response at Elsey Homestead streamflow gauge). Modelled groundwater development and climate scenarios result in a proportional change to the FEFLOW modelled. The proportional change (i.e. relative to Scenario A) in the regional groundwater baseflow input was then applied in the river model. Prior to this an estimate of daily groundwater flow in the river model must be calculated. This was undertaken using a Lyne and Hollick digital filter (Lyne and Hollick 1979) of the river model streamflow. Digital filter parameters were adjusted until the mean daily baseflow from the digital filter matched the FEFLOW groundwater discharge at that site: 𝑞𝑞𝑓𝑓(𝑖𝑖)= 𝛼𝛼∗𝑞𝑞𝑓𝑓(𝑖𝑖−1)+(1+𝛼𝛼) 2􀵗∗[𝑞𝑞(𝑖𝑖)− 𝑞𝑞(𝑖𝑖−1)] 17 𝑞𝑞𝑏𝑏(𝑖𝑖)=𝑞𝑞(𝑖𝑖)−𝑞𝑞𝑓𝑓(𝑖𝑖) 18 where 𝑞𝑞𝑓𝑓(𝑖𝑖) is the quickflow response at time 𝑖𝑖, 𝑞𝑞(𝑖𝑖) is the original streamflow at time 𝑖𝑖, 𝑞𝑞𝑏𝑏(𝑖𝑖) is the baseflow response at time 𝑖𝑖, and 𝛼𝛼 is a tuning parameter. The alpha parameter was adjusted until a reasonable fit to the FEFLOW discharge for the 50-year period was achieved. The results can be seen in Figure 4-7. The mean daily groundwater discharge from the FEFLOW model for this period was 3.22 m3/s, while the emulator derived from the river model was 3.19 m3/s. The minimum flow rate across the calibration period from the emulator was 1.64 m3/s, which was assumed to be the magnitude of regional inputs (since these are unlikely to vary across a 50 year period). When the emulator was applied to the river model for a dry climate future the baseflow was reduced by 36 %. This compares well with the estimates from the FEFLOW model of 32 %. Figure 4-7 Estimates of groundwater discharge at node 90300130 from FEFLOW and via river model emulator 4.4 Future climate analyses The Assessment climate activity has produced several variants of 109 years of daily climate series, representative of the five different climate scenarios used by the Intergovernmental Panel on Climate Change (IPCC) in its Sixth Assessment Report (IPCC, 2022). These scenarios are called Shared Socioeconomic Pathways, or SSPs, and are defined as follows: · SSP1-1.9: emissions rapidly decline to net zero by about 2050, and become negative after that · SSP1-2.6: emissions decline to net zero by about 2075, and become negative after that · SSP2-4.5: emissions rise slightly, before declining after 2050, but not reaching net zero by 2100 Chart. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au · SSP3-7.0: emissions rise steadily to become double their current amount by 2100 · SSP5-8.5: emissions rise steadily, doubling by 2050 and more than tripling by the end of the century. GCMs provide information at a resolution that is too coarse to be used directly in catchment-scale hydrological modelling. For example, rainfall occurs too often and at too low intensity (Stephens et al., 2010). This is particularly important in tropical regions, where even high-resolution coupled climate models simulate tropical cyclones as weaker and larger in horizontal extent than those observed (IPCC, 2015). Hence, an intermediate step is generally performed: the broad-scale GCM outputs are transformed to catchment-scale variables. In a comparative assessment of scaling methods in an area dominated by large-scale storm events, Salathé (2003) found simple scaling methods to be as effective in simulating hydrological systems as more complicated downscaling methods (e.g. dynamic downscaling). For these reasons, and due to the scale of the catchments being assessed, which makes it resource intensive to undertake dynamic or statistical downscaling, a simple scaling technique – the pattern scaling method (Chiew et al., 2009b) was adopted. The seasonal pattern scaling (PS) method employed used output from the 32 GCMs to scale the 109-year historical daily rainfall, temperature, radiation and humidity sequences (i.e. SILO climate data), to construct the 32 by 109-year sequences of future daily rainfall, temperature, radiation and humidity. The method is briefly described below. A more detailed description can be found in Chiew et al. (2009b). The method comprised two broad steps. The first step involved estimating the seasonal scaling factors for four 3-month blocks (December to February, March to May, June to August and September to November) for the changes between two time slices centred around 1990 (1975 to 2005) and 2060 (2046 to 2075). This is representative of a 1.6 °C temperature rise under an SSP2- 4.5 emissions scenario (Table 6-1). For each season and over each time slice, the total rainfall was calculated. Seasonal scaling factors were then calculated as the ratio of the total season’s rainfall over the 2060 time slice divided by the total rainfall over the 1990 time slice. The historical climate sequence was then scaled using these seasonal scaling factors. The second step involved rescaling the entire series so that it matches the annual scaling factors, to maintain consistency with annual projected changes in the GCMs (Chiew et al., 2009b; Petheram et al., 2012). This process was repeated for each GCM, for each season and for each GCM grid cell. The results were then expressed per degree warming by dividing the projected change in rainfall by the projected representative global temperature rise of 1.6 °C temperature rise (under the SSP2-4.5 emissions scenario). The method was then repeated for each climate parameter, except for temperature where the difference rather than ratio between the two periods was used to scale the historical sequence. The method of using a pattern scaling method to transform broad-scale GCM outputs to catchment-scale variables is denoted herein as ‘GCM-PS’. 5 River model calibration This section relates to the calibration of the Roper River model. Unlike the NAWRA study, no independent calibration of a landscape model was conducted in the Assessment. Rather, the Sacramento RR model and parameters used in the final river mode version were used to generate gridded runoff data for use in other components of this Assessment. River system models consist of a series of linked nodes or reaches in which the processes of losses and gains are modelled. These nodes are linked in an upstream to downstream chain, enabling representation of the river related states in time and space. In this regard, river system models are semi-distributed, with the outputs of upstream nodes acting as inputs to downstream nodes. RR models are the primary way of estimating runoff/streamflow generated within each river model subcatchment. They utilise subcatchment mean daily precipitation and potential ET time series as inputs, without the need for other inputs other than model parameters. There are many options in terms of RR models and the selection of RR models for this Assessment builds on investigation of RR models used within river models in northern Australia by Hughes et al. (2017). The Hughes et al. (2017) study also tested a range of objective functions, loss models and calibration methods. Hughes et al. (2017) tested the GR4J model (Perrin et al., 2003) and compared it to both Sacramento (Burnash, 1995) and AWRA-L (Viney et al., 2015), noting that AWRA-L is not, strictly speaking, a RR model since it is more ‘physically based’ and has a number of spatially explicit input parameters that are not calibrated. However, AWRA-L can be calibrated to observed streamflow and generate runoff estimates as a RR model might. While results varied with location, they suggested that the Sacramento RR model had better overall performance. This was particularly the case for low-flow representation, and general performance in arid areas. GR4J has only four parameters, which in isolation of performance, makes it attractive for system calibration. Given this, model testing for this Assessment included adaptions of the GR4J model intended to improve low-flow performance. The GR7J model (Hughes et al., 2013; Grigg and Hughes, 2018) is an adaption to GR4J that was designed to cope with long-term changes in storage and climate change better than the GR4J model. GR7J has also been shown to give better performance in drier climates (Hughes et al., 2021). However, low-flow performance of this model, while improved somewhat, is still inadequate for many applications, but most notably ecological assessment. Ideally a model should be able to represent zero-flow days since this is an important ecological metric. To that end adjustments to the GR7J model formulation was attempted and tested. This model was denoted GR7JFM and more information on the model formulation is given in Appendix B. Typically, river system models are calibrated on a ‘reach-by-reach’ basis in which each node is calibrated against an observed gauge in isolation from upstream outputs other than inflows and downstream performance is not considered until that reach is calibrated in turn (e.g. Hughes et al., 2014 Hughes et al., 2017). However, this method suffers from problems since the method for calibration, and simulation for prediction, are quite different (Lerat et al., 2013). An obvious alternative to this is the ‘system-calibration’ method where all parameters for all reaches are calibrated simultaneously, running the model in the same way in calibration as it would be used for predictive modelling. This has benefits (Hughes et al., 2016), but high dimensionality in optimisation becomes a problem depending upon the number of reaches and types of RR models used. The Assessment used a new method which, mechanistically, lies between reach-by-reach and system calibration methods. Shingle calibration divides the catchment, and available stream gauge locations, into overlapping sections or ‘shingles’. At its most simple, each shingle consists of three gauges in serial order in which the uppermost gauge has already been calibrated and a time series of simulations are available. The second gauge and third gauge are then calibrated simultaneously (i.e. the goodness of fit for the second and third gauge are both utilised in the optimising function). Once calibration is complete, the time series of the second gauge is saved and is used as an input to the subsequent shingle in which it is the first gauge. This process is illustrated in Figure 5-1, where a hypothetical river system of five nodes is represented. Blue nodes in each calibration shingle have been calibrated in a previous shingle, while simulation performance of both the green and red nodes is optimised simultaneously in the current shingle. Parameters estimated for the green node are saved and a time series of simulated flow is used as an input in the following calibration shingle. This is a highly simplified example and, in practice, some modifications to this process are required (e.g. for headwater nodes and final nodes), although the same principles apply. The intention of this process was to constrain parameters in each reach by performance at the calibration gauge (green nodes in Figure 5-1) and its effect on at least one downstream node (red nodes in Figure 5-1). This allows many of the benefits of system calibration to be realised, while reducing problems associated with high dimensionality. Figure 5-1 Shingle calibration conceptual diagram for a section of five nodes of a river system model. Each calibration shingle has a pre-calibrated input time series (blue), a calibration node (green), and an auxiliary calibration node (red) that is jointly calibrated with green nodes In most river modelling situations, there will be at least some ungauged nodes (i.e. nodes for which streamflow simulations are required, but there are no available gauge data at that location). In these situations, model parameters must be obtained in another way. Prediction in ungauged areas is an area of ongoing research in hydrology and there have been many publications relating to this situation. Indeed, the International Association of Hydrological Sciences (IAHS) initiated the Prediction in Ungauged Basins program in 2003 (Sivapalan, 2003) to address these needs. During and since this initiative, there have been many publications that aimed to examine various methods of transferring parameters to or estimating parameters for ungauged catchments. This process is termed ‘regionalisation’ and was reviewed by Bloschl et al. (2013). Regionalisation falls into two broad groups, nearest neighbour (NN) and hydrological response unit (HRU) approaches respectively. The HRU approach aims to use various catchment physical descriptors to classify catchments of similar type. The physical descriptors may be related to catchment climate, topography, vegetation and regolith/geological characteristics. Parameters are transferred to the ungauged catchments either as a set or individually (often ignoring parameter correlation), based on similarity with gauged catchments. The NN approach simply takes parameters from the nearest gauged catchment(s). For more information see the review of Parajka et al. (2013). Examination of the relative success of these methods is inconclusive, although the most relevant study in the context of this Assessment is that of Petheram et al. (2012), who concluded that NN methods were the most successful in northern Australia. Diagram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\0_Working\1_Justin\7_Technical_report\shingle.PNG" The shingle calibration method was adapted so that parameters could be ‘shared’ between gauged subcatchments and ungauged subcatchments within a calibration shingle and/or parameters could be ‘borrowed’ from pre-existing calibrations (usually from preceding shingles). Since all gauges within a shingle are jointly calibrated, and any simulation at a gauged location will be affected by upstream and downstream goodness of fit, as well as its own location, a measure of constraint is applied to shared parameters, at least greater than consideration at a single gauge. It should be noted that the protocol in terms of donor and receiver catchments is a decision made by the hydrologist a priori, rather than any automated procedure, and hence there is scope for various parameter sharing possibilities for any single location. For example, if a conceptual calibration shingle has three gauged nodes and three ungauged nodes, parameters must be assigned to the three ungauged nodes (Figure 5-2). Even if we assume that no parameters are ‘borrowed’ from outside the shingle, and only parameters from within the shingle are used, there are still 27 different parameter combinations that could be used across the three ungauged sites. Considering that there will generally be multiple shingles used for any river model, the possibilities for parameter transfer increase exponentially, and as such, the judgment of the hydrologist is required to eliminate most of these permutations. Despite the NN method of parameter transfer being generally favoured, judgment based on physical factors (i.e. a HRU approach) are sometimes warranted. The most obvious example of this in the Roper catchment pertains to the Sturt Plateau (Figure 2-1). The most reliable stream gauge on the Sturt Plateau, 9030001, has a very low runoff coefficient (less than 1%). It is assumed that this is the result of higher permeability of the regolith in the Sturt Plateau. Regardless of the mechanism responsible, the calibrated parameters for gauge 9030001 are considered suitable only for transfer to other Sturt Plateau locations. Figure 5-2 Conceptual diagram of a six-node calibration shingle, of which three nodes are ungauged 5.1 Calibration objective functions Objective functions provide an opportunity for the hydrologist to influence the nature of the simulations that the model will produce. Goodness-of-fit metrics commonly used such as bias and Nash–Sutcliff efficiency (Nash and Sutcliffe, 1970), may not be adequate to produce simulations to the satisfaction of the hydrologist. With some knowledge of how the simulations may be used, the hydrologist may require more emphasis on various components of streamflow. For example, flood studies require more emphasis on accurate peak flow representation, while some ecological studies may require more accurate low-flow representation. In this Assessment a series of model/objective function/parameter transfer combinations were tested in an iterative process with the aim of exploring and producing a more satisfactory collection of simulations (Table 5-1). Broadly, these were all judged on the basis of goodness of fit at gauge locations; the ability to produce an acceptable simulation during the first half of the 20th century, for which there were no stream observations (judged mainly on runoff coefficient); the behaviour of various models in rudimentary development scenarios; and the spatial distribution of runoff coefficient, or more specifically, the spatial coherence of runoff coefficient across subcatchments. Diagram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\0_Working\1_Justin\7_Technical_report\conceptual_parameter_share.png" Table 5-1 Rainfall-runoff model and objective function used for various river model calibrations For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au † This differs from a previous calibration due to a different parameter sharing protocol. Many of these models, particularly earlier iterations, were rejected on model goodness of fit to observations alone, although any of these models may be considered plausible. Objective function terms were formulated so that each term has a value of 1 for a perfect fit. Scores increase as goodness of fit degrades, and the overall objective function score is minimised. It should be noted that all calibration nodes within each calibration shingle are calibrated jointly, and a final aggregated score is computed using an aggregating function: 𝑂𝑂𝑂𝑂𝑠𝑠ℎ𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖=Σ𝑂𝑂𝑂𝑂𝑗𝑗 ∗𝑤𝑤𝑗𝑗 𝑛𝑛 𝑗𝑗 = 1 (19) where 𝑛𝑛 is the number of calibration nodes and 𝑤𝑤𝑗𝑗 is the weighting value for each calibration node. The weighting value is calculated as: 𝑤𝑤𝑗𝑗=Σ𝑞𝑞𝑗𝑗 𝑚𝑚 𝑖𝑖=1ΣΣ𝑞𝑞𝑖𝑖,𝑗𝑗 𝑚𝑚 𝑖𝑖=1 𝑛𝑛 𝑗𝑗=1􀵗 (20) where n is the number of calibration nodes in the shingle and m is the number of days of observation for each calibration node, and q is the observed flow. This weighting vector formulation places more weight of those calibration nodes with relatively longer records and/or higher observed flow rates. Objective function terms are defined as: 1. 𝑁𝑁𝑁𝑁𝑁𝑁0.5 is the Nash–Sutcliffe efficiency for root transformed simulated and observed flows. 2. 𝜀𝜀𝑎𝑎𝑎𝑎 is the mean annual error across all years where the error for each individual year, 𝑎𝑎, is defined as (𝑠𝑠𝑎𝑎 − 𝑜𝑜𝑎𝑎)/𝑜𝑜𝑎𝑎, where 𝑠𝑠𝑎𝑎is the mean simulated flow and 𝑜𝑜𝑎𝑎 is the mean observed flow in year a. 3. 𝜀𝜀ℎ𝑓𝑓 is the mean error between the highest 20 flows for observed and simulated flow in matching time periods, 􀸫𝑠𝑠𝑠𝑠𝑠𝑠𝑘𝑘􀴤􀴤􀴤􀴤􀴤􀴤􀴤− 𝑜𝑜𝑜𝑜𝑜𝑜𝑘𝑘􀴤􀴤􀴤􀴤􀴤􀴤􀸫/𝑜𝑜𝑜𝑜𝑜𝑜𝑘𝑘􀴤􀴤􀴤􀴤􀴤􀴤 where k denotes the highest 20 flows in rank order. 4. 𝐸𝐸𝐸𝐸𝐸𝐸𝑝𝑝90 is the error probability difference between the observed flow and simulated flow for the 90th percentile exceedance observed flow for all non-zero flow observations, defined as 􀸫𝑃𝑃𝑝𝑝90(𝑜𝑜𝑜𝑜𝑜𝑜)−𝑃𝑃𝑝𝑝90(𝑠𝑠𝑠𝑠𝑠𝑠)􀸫, where 𝑃𝑃𝑝𝑝90(𝑜𝑜𝑜𝑜𝑜𝑜) is the quantile probability of the observed flow (90% exceedance non-zero flow), and 𝑃𝑃𝑝𝑝90(𝑠𝑠𝑠𝑠𝑠𝑠) is the quantile probability of the same flow for simulated flow. A similar calculation is used for the 𝐸𝐸𝐸𝐸𝐸𝐸𝑝𝑝70 calculation, with obvious adjustments using exceedance probability flow. 5. 𝐸𝐸𝐸𝐸𝐸𝐸𝑓𝑓200 and 𝐸𝐸𝐸𝐸𝐸𝐸𝑓𝑓500 are similar to the 𝐸𝐸𝐸𝐸𝐸𝐸𝑝𝑝90 calculation except that the flows for calculation of probability difference are 200 ML/day and 500 ML/day, respectively. These flows are chosen since they are close to threshold flows at which water harvest may commence (these are somewhat arbitrary and are not a reflection of anticipated policy). 6. 𝜀𝜀𝑧𝑧𝑧𝑧𝑧𝑧𝑧𝑧 is the error between simulated and observed flows in terms of the number of zero-flow days observed. This is calculated as |𝑠𝑠𝑠𝑠𝑠𝑠𝑧𝑧−𝑜𝑜𝑜𝑜𝑜𝑜𝑧𝑧|/𝑜𝑜𝑜𝑜𝑜𝑜𝑧𝑧, where 𝑠𝑠𝑠𝑠𝑠𝑠𝑧𝑧 is the number of no-flow days in the simulated time series and 𝑜𝑜𝑜𝑜𝑜𝑜𝑧𝑧 is the number of observed no-flow days in the concurrent time series. The denominator is dependent upon the presence of zero-flow days in the observed and or simulated time series (i.e. where no zero-flow days are observed, simulated no-flow days are used in the denominator). 7. 𝛽𝛽 is the simulation bias. In cases where used, simulation bias was set to optimise to –10% in order to conservatively estimate streamflow for diversion. All models were calibrated using the Differential Evolution algorithm of Ardia et al. (2010). This is suitable for river model calibration since it is a heuristic method, well suited to uneven or ‘rough’ optimisation surfaces. Additionally, it was written to allow multi-threading, thereby speeding up the optimisation process. 5.2 Calibration results Despite an array of candidate models, model objective functions and parameter transfer protocols, modelled flow across all versions were relatively stable. This was particularly the case at gauge locations such as 90302500, which was well constrained by the availability of the Red Rocks gauge data (9030250). No gauge data are available on the Roper River downstream of Red Rocks, or beyond the downstream ends of both the Wilton and Hodson rivers (Figure 4-2 ). Therefore, estimates of streamflow downstream of these gauges relies on parameter transfer to estimate end-of-system flow. In some models, gauge data on both the Wilton and Hodgson rivers were deemed to be inadequate, and parameters were transferred to subcatchments. Accordingly, estimates of mean end-of-system flow showed slightly more variation (Figure 5-3). This translates to a variation of approximately plus or minus 12% of the mean end-of-system flow and plus or minus 2% at node 90302500 (Red Rocks). Figure 5-3 Estimated mean streamflow for various model versions at Red Rocks (n = 10) and Roper end-of-system node 90300000 (n = 7) The effect of model version was iteratively tested for its effect on scenario behaviour. More specifically, a simplified ‘water harvest’ scenario was run with four candidate models at a single location (node 90300110). Water harvest pump start thresholds were run at a number of values, while the pump capacity was set to be unlimited. Candidate models (versions 5,6,7 and 11) were tested, and model version was shown to have a significant effect on water harvest volumes and timing (Figure 5-4 ). Using these four model versions, variation in median water harvest cumulative diversion (at the end of March) was approximately plus or minus 15% of the mean. Of these four model versions, only version 11 had the specific constraints placed on matching observed threshold flows of 200 and 500 ML/day (Table 5-1 ). As expected, version 11 had a better fit to observed flow at the specified flow thresholds (Figure 5-5). Chart, box and whisker chart. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\0_Working\1_Justin\7_Technical_report\1_scripts\all_versions_flow_boxplot.R" Figure 5-4 Median cumulative water harvest volumes for various model versions at node 90300110 using a pump start threshold of 500 ML/day and an unlimited pump capacity Figure 5-5 Flow duration curves for four model versions and observed flow at node 90302500. Horizontal broken lines indicate flow values of 500 and 200 ML/day for reference Model version 5, based on GR7JFM, produced slightly lower runoff coefficients in drier decades (Figure 5-6 ), although version 7, based on the same objective function with a differing parameter transfer protocol, did not exhibit the same behaviour. Similarly, version 6, based on Sacramento, showed much higher runoff in the driest decade (1910–1920), while version 11 was more moderate. Chart, line chart. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\0_Working\1_Justin\7_Technical_report\1_scripts\waterHarvest_by_version_plot.R" Chart, line chart. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\0_Working\1_Justin\7_Technical_report\1_scripts\FDC_all models.R" Figure 5-6 Runoff coefficient vs aridity for four different model versions at the Red Rocks (90302500) model node by decade. Labels indicate the start of each decade for some clusters of data Many of the gauges of with shorter duration of observations, or of questionable quality (e.g. 9030003, 9030108 and 9030146) had adverse effects on the spatial distribution of runoff coefficient (Figure 5-7). This was exacerbated by the fact that these observations were made in relatively wetter periods. While some differences in runoff coefficient were expected (e.g. runoff is relatively low from the Sturt Plateau), it is reasonable to expect these differences to be subtle, and reflect the nature of the climate and landscape across the area of interest. A degree of spatial coherence in runoff coefficient was a factor in eliminating model versions from further analyses. It can be seen in Figure 5-8 , that model version 12 had more favourable spatial coherence of runoff coefficient than version 3. For example, for version 3, node 90301080 has an unacceptably low runoff coefficient relative to those surrounding it. While this is possible, it was deemed to be very unlikely. It should be noted that gauge data at 90301080 was used in calibration for this node in version 3. It had data of suspicious properties (Appendix C (Apx Figure C-10)) and was eliminated from subsequent calibrations. Chart, scatter chart. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au //fs1-cbr.nexus.csiro.au/{lw-rowra}/work/2_Hydrology/7_RiverModel/model_compare_Qcoef_v_time_v2.R Figure 5-7 Simulated mean runoff coefficient by subcatchment for model version 3 Map. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\7_RiverModel\3_river_3\4_model_check\model_Qcoef_spatial_check_v8.png" Figure 5-8 Simulated mean runoff coefficient by subcatchment for model version 12 Model version 12 was chosen to use for ongoing scenario analyses since it had the most acceptable qualities in terms of water harvest behaviour, spatial coherence of runoff, low flow characteristics and goodness of fit. Goodness-of-fit plots are shown in Appendix D. The calibration of this model used six gauges, with parameters being transferred from some of these calibration nodes to subcatchments as indicated in Table 5-2. Parameters from gauges 9030250 and 9030089 are used extensively across the model domain, since these gauges have long records, good gauging data and goodness of fit is superior to other gauges (Table 5-3). "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\7_RiverModel\14_river_9\6_modelCheck\model_Qcoef_spatial_check_v9.png" Table 5-2 Parameter transfer protocol for model version 12 For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Table 5-3 Goodness-of-fit statistics for model version 12 For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Using the weighting function described in equations 19 and 20, most of the weight in calibration is placed upon nodes 90300890 and 90302500. This is reflected in the better goodness-of-fit scores for those two nodes (Table 5-3). Very good low flow scores (EPD90, EPD200 and EPD500) were apparent for the nodes 90301760, 90300130 and 90302500. This is related to the need to achieve good calibration scores downstream in the presence of regional groundwater inputs and apparent losses. Gauge 9030250 shows an increase in baseflow across the measurement period (albeit slight), so an apparent change in baseflow and runoff coefficient cannot be dismissed. Given that baseflow around Mataranka is related to interactions with a large regional aquifer and a local aquifer (Taylor et al. 2023), this is a real possibility, although these interactions are difficult to capture and estimate with a river model and can have perverse effects on model performance. Long-term changes in storage/groundwater interactions are not well captured by RR models (Grigg and Hughes, 2018), and even where models are modified to do so, data are required to constrain these models. The final calibrated model, version 12, was the accepted model after rejection of previous versions. The sensitivity of water harvest volumes to model objective function was an important influence in this process (Figure 5-4 ). For this reason, model fit for 𝐸𝐸𝐸𝐸𝐸𝐸𝑓𝑓200 and 𝐸𝐸𝐸𝐸𝐸𝐸𝑓𝑓500 were considered of high importance, particularly for nodes along the Roper River upstream of Red Rocks where the most versatile soils in terms of irrigation are located. 6 Model scenario analyses The Assessment considered four scenarios, reflecting combinations of different levels of development and historical and future climates, much like those used in the Northern Australia Sustainable Yields projects (NASY) (CSIRO, 2009a, 2009b, 2009c), the Flinders and Gilbert Agricultural Resource Assessment (Petheram et al., 2013a, 2013b) and the Northern Australia Water Resource Assessments (Petheram et al., 2018a, 2018b, 2018d): • Scenario A – historical climate and current development • Scenario B – historical climate and future development • Scenario C – future climate and current development • Scenario D – future climate and future development. Scenario A Scenario A and its subset, Scenario A2060, both assume a historical climate. The historical climate series is defined as the observed climate (rainfall, temperature and potential evaporation for water years from 1 September 1910 to 31 August 2019). All results presented in this report are calculated over this period unless specified otherwise. Scenario A assumes no surface water or groundwater development. Because the impacts of licensed groundwater extraction near Mataranka (~24 GL) on baseflow in the Roper River are yet to be fully realised, Scenario A is considered most representative of the hydrological regime in the Roper catchment at 31 August 2019. Scenario A was used as the baseline against which assessments of relative change were made. This will give the most conservative results. Historical tidal data were used to specify downstream boundary conditions for the flood modelling. Scenario A2060 assumes historical climate and current levels of surface water (~0.1 GL) and groundwater development (~24 GL near Mataranka and ~8 GL near Larrimah) assessed around 2060. The difference between Scenario A and Scenario A2060 is that the potential impacts of current groundwater extraction on baseflow in the Roper River are calculated 40 years from 31 August 2019. This corresponds to a period longer than a typical agricultural investment timeframe (~20 30 years), and four times longer than the current groundwater licencing period in the NT. Scenario B Scenario B is historical climate and future development assessed at ~2060. Scenario B used the same historical climate series as Scenario A. River inflow, groundwater recharge and flow, and agricultural productivity were modified to reflect potential future development. Potential development options were devised to assess response of hydrological, ecological and economic systems ranging to small incremental increases in surface water and groundwater extraction through to extraction volumes representative of the likely physical limits of the Roper catchment (i.e. considering the colocation of suitable soil and water). All price and cost information was indexed to June 2021 (i.e. reflective of pre-COVID-19 prices). All water harvesting and dam based development scenarios assume 35 GL of groundwater extraction south of Larrimah in addition to current licensed extractions. It should be noted that the difference in baseflow at 2060 under the three groundwater development scenarios examined in the Assessment, 35, 70 and 105 GL, are negligible (~1%), and the majority of modelled impacts to baseflow at 2060 are due to current licensed extractions near Mataranka. The impacts of changes in flow due to this future development were assessed, including impacts on: • instream, riparian and near-shore ecosystems • Indigenous water values • economic costs and benefits • opportunity costs of expanding irrigation • institutional, economic and social considerations that may impede or enable adoption of irrigated agriculture. Scenario C Scenario C is future climate and current levels of surface water and ground development assessed at ~2060. Future climate impacts on water resources were explored within a sensitivity analysis framework by applying percentage changes in rainfall and potential evaporation to modify the 109-year historical climate series (as in Scenario A). The percentage change values adopted were informed by projected changes in rainfall and potential evaporation under Shared Socio-economic Pathways (SSP) 2-4.5 and 5-8.5. SSP 2-4.5 is broadly considered representative of a likely projection given current global commitments to reducing emissions and SSP 5-8.5 is representative of an (unlikely) upper bound. Scenario D Scenario D is future climate and future development. It used the same future climate series as Scenario C. River inflow, groundwater recharge and flow, and agricultural productivity were modified to reflect potential future development, as in Scenario B. Therefore, in this report, the climate data for scenarios A and B are the same (historical observations from 1 September 1910 to 31 August 2019) and the climate data for scenarios C and D are the same (the above historical data scaled to reflect a plausible range of future climates). 6.1 Scenario A – historical conditions The baseline river model using historical climate was run for the period 1900-01-01 to 2019-11-17, with reporting based upon the period 2010-09-01 to 2019-08-31. The flow statistics for this model run are shown in Table 6-1. Table 6-1 Flow statistics for all model nodes under Scenario A For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au A follow-up simulation was run with full surface water entitlements as they stand at the time of writing. These data were obtained from the Northern Territory Government’s Water Licencing Portal (Northern Territory Government website ). These are limited to a total of 36 ML/ha in node 90301780 and 1649 ML/ha in node 9030110. Additionally, stock and domestic surface water was assumed to take place across the catchment at 1000 ML/ha, although there are little data to confirm this. In all cases, no information regarding any pumping conditions was available. However, given the low volume relative to estimated end- of-system flow, the effects were insignificant. 6.2 Scenario B – development under historical conditions 6.2.1 Water harvest Water harvest analyses were conducted for concurrent diversion at 12 nodes within the river system model. In general, these were targeted to areas where surface water and arable soil was thought to be present, and/or there were communities nearby that may require water for development. Water harvest analysis is partly intended to explore the physical limits of supply across the Roper catchment and how various factors such as pump start thresholds, licence volumes, pump capacity and end-of-system requirements can ‘trade-off’ using annual reliability of supply as in indicator. The analysis was conducted for various system irrigation targets (alternatively denoted licence volumes) across the entire catchment. The irrigation target in each water harvest node was simply a proportion of the system target using the proportions shown in Table 6-2 Table 6-2 Median annual flow and proportional water harvest allocation for Roper water harvest analyses For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au A range of system allocations, pump start thresholds, pump capacities and last node annual requirements were tested with the Roper River model (Table 6-3). Table 6-3 Water harvest parameters and values analysed for Roper water harvest analyses For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au † Pump capacity is the rate at which the pump(s) can operate to extract the reach annual irrigation target (𝑎𝑎𝑖𝑖) in the given number of days. Hence the pump capacity used by the simulation (m3/second) will be a function of 𝑎𝑎𝑖𝑖. It should be noted that these analyses do not account for limits on diversion that may be apparent (e.g. availability of adequate areas of suitable soil with which to irrigate (the assumed primary reason for diversion)), or any other legal/social/economic reasons that may influence diversion of surface water. Subsequent analyses (Section 4.2.1) attempted to incorporate limits due to availability of suitable soil on the diversion of water from the Roper catchment. The number of possible combinations of water parameters across all model locations are beyond the limits of a report to communicate effectively. The sensitivity of the system to a range of water harvest strategies was judged primarily by a series of ‘heat map’ plots that describe the variation in annual reliability of supply for water harvest nodes with changes in annual irrigation target (or licence volume), pump start threshold, pump rate and end-of-system flow requirement. These plots are shown in the supplementary material. An example of annual reliability of diversion is shown for a single node in (Figure 6-1). Additionally, colour ramps and contours are designed to indicate 75% annual reliability of supply as a nominal benchmark. It can be seen from Figure 6-1 that increasing pump start threshold for any scenario will decrease the annual reliability of supply. Decreasing pump rate (increasing the number of ‘pump days’), while all other conditions are identical, will also reduce the reliability of supply for any irrigation target. Interestingly, as pump start thresholds increase, reliability of supply in smaller tributaries (with relatively lower mean flow rates) decreases dramatically. Concurrently, this allows reliability of supply to increase towards the end-of-system, where flows are higher, since less flow can be diverted upstream (Figure 6-2 ). Any diversion will obviously reduce flow in the river itself. In dry years where the river flow is lower, the effects of a given diversion volume will be proportionally higher. In this regard it is useful to examine the effects of various water harvest strategies on the flow relative to Scenario A (historical flow) for drier years. A convenient statistic is the 80% exceedance annual flow relative to Scenario A. Figure 6-3 shows the effect of annual system irrigation target and end-of-system annual requirement at node 90301780 on the relative 80% annual exceedance flow for a selection of nodes. Noting that flow is relative to Scenario A 80% annual exceedance flow and other nodes not shown are included in the whole-of-river analysis (i.e. water harvest is conducted in other nodes not shown in plots). This analysis shows that for system irrigation targets of greater than around 500 GL/year, large reductions in the 80% annual exceedance flow are apparent, although this effect is more marked at downstream nodes relative to headwater nodes since the effects of diversion will accumulate in a downstream direction and diversion in headwater nodes are likely to be relatively restricted by pump start thresholds of 600 ML/day. More notably, it is apparent that end-of-system requirements of at least 700 GL/year are required to significantly reduce the effects of diversion on 80% annual exceedance flow for lower irrigation target volumes (say 800 GL/year). Similarly, for system irrigation targets of less than 250 GL/year, effects on 80% annual exceedance flow are modest with little sensitivity to end-of-system requirements. "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\0_Working\1_Justin\7_Technical_report\WH_mix_v2.png" Figure 6-1 Annual reliability of irrigation supply for various pump start threshold and irrigation target volumes at node 90300110 for: (a) pump rate 20 days, (b) pump rate 40 days, and (c) pump rate 20 days with an end-of-system requirement of 700 GL/year. Hatched white target volumes are unlikely to be possible due to restricted availability of suitable soils for irrigation in the catchment "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\0_Working\1_Justin\7_Technical_report\WH_Ngukurr_20days_EOS700.PNG" Figure 6-2 Annual reliability of supply at node 90301780 using a pump rate of 20 days and an end-of-system requirement of 700 GL/year "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\7_RiverModel\14_river_9\8_waterHarvest\2_output\6_catchReportChap5_plots\catchRep_80annual_residual_flow.png" Figure 6-3 The 80% annual exceedance streamflow relative to Scenario A in the Roper catchment using a pump start threshold of 600 ML/day and a pump capacity of 20 days for various system irrigation targets and end-of-system annual requirement at node 90301780. Hatched white target volumes are unlikely to be possible due to restricted availability of suitable soils for irrigation in the catchment 6.2.2 Soil-limited water harvest Soil-limited water harvest volumes for the Roper catchment were calculated according to the method outlined in Section 4.2.1. Using the Roper River model, estimates were made regarding the volumes that could be supplied at 75% annual reliability for various model nodes run together (ensuring that annual reliability is at least 75% in all nodes). These estimates are shown in Table 6-4 , along with the maximum volume that could be used for irrigation on all suitable soils in each reach. All model nodes in all scenarios used a pump rate of 20 days and a pump start threshold of 200 ML/day. Those analyses were repeated for various end-of-system requirements (100, 400 and 700 GL/year at node 90301780). While the total soil limit in terms of water requirement is higher than the combined soil and water limit for end-of-system (EOS) 0 GL, it is clear that for many nodes where reasonable areas of suitable soil are available, water limits apply, and at other nodes where one may expect larger volumes of water to be available (e.g., 90301780), soil limits to diversion apply. With no end-of- system requirement (EOS 0), a total of 660 GL could be diverted at 75% annual reliability and used for irrigation productively. Using an end-of-system requirement at 90301780 of 700 GL/year limited annual diversion to approximately 65% of this volume (426 GL/year). It is likely that uncertainty in these estimates will be of similar magnitudes to that determined in Section 5.2, where variation was approximately plus or minus 15% relative to the mean water harvest yield. Uncertainty due to model version was not explicitly determined for this particular scenario. Table 6-4 Estimated maximum soil-limited water harvest assuming a pump rate of 20 days and a pump start threshold of 200 ML/day. Values were calculated to ensure a minimum of 75% annual supply reliability For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au The use of end-of-system requirements had the effect of delaying the start of diversion in the wet season, and as such, often reduced the opportunity for pumping a full licence volume. This can be seen in the timing of the first day of pumping across 109 years of simulation. These data are expressed as density plots in Figure 6-4. In some cases, the onset of irrigation diversion was delayed until the commencement of the following wet season, albeit within the same water year. The effects of end-of-system requirement will vary based on conditions, however increasing EOS from 100 GL/year to 400 GL/year delayed the start of irrigation diversion by around 25 days, with a further 10 days’ delay required for an EOS of 700 GL/year. "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\7_RiverModel\14_river_9\8_waterHarvest\2_output\eos_start_day.png" Figure 6-4 Timing of the first day of annual pumping for end-of-system (node 90301780) requirement across 109 years of model simulation, expressed as a frequency density Water harvest operations with a low pump start threshold (200 ML/day), can have substantial effects of streamflow, particularly where no end of system flow condition is invoked. Flow in drier years (as indicated by the 90% annual exceedance flow, Table 6-5), was drastically reduced by water harvest operations. Similar effects were evident for flow rates. However, the use of an end of system flow requirement, while reducing the total available volume for water harvest, substantially reduced impacts of water harvest in dry years. By moving the annual commencement of pumping to the wet season via end of system flow requirements, impacts of water harvest on low to moderate flow (generally during the dry season) is also reduced. Table 6-5 Residual streamflow at selected model nodes for Scenario A and three water harvest scenarios where reliability of supply is maintained to at least 75% in all water harvest nodes. Scenario Node ID Median annual flow (GL) 90% exceedance annual flow (GL) 80% exceedance flow (m3/s) 50% exceedance flow (m3/s) Scenario A 90300110 1729 412 0.981 4.56 90302500 1925 406 0.429 3.726 90300000 4341 1162 0.939 6.6 Water harvest 660 GL/y – EOS 0 GL 90300110 1332 65 0.736 2.197 90302500 1431 32 0.197 1.49 90300000 3695 601 0.357 2.846 Water harvest 579 GL/y – EOS 400 GL 90300110 1399 241 0.752 3.152 90302500 1488 217 0.232 2.345 90300000 3769 774 0.693 5.421 Scenario Node ID Median annual flow (GL) 90% exceedance annual flow (GL) 80% exceedance flow (m3/s) 50% exceedance flow (m3/s) Water harvest 426 GL – EOS 700 GL 90300110 1514 325 0.771 3.577 90302500 1634 316 0.257 2.772 90300000 3918 887 0.725 5.421 6.2.3 Diversion based upon instream dams For five dam sites across the Roper catchment identified as being worthy of further analyses (Figure 6-5), height, volume and surface area relationships were determined (Petheram et al., 2022) and capital costs of construction calculated. Using the reservoir model outlined in Section 4.1.5, 85% annual exceedance supply reliability was determined for all combinations of dam FSL and transparent flow. This allowed further analysis of the unit capital cost of supplied water, along with various components of reservoir water balance (Figure 6-6, Figure 6-7, Figure 6-8, Figure 6-9 and Figure 6-10). These results were then used to determine which FSLs to use in a combined ‘five dams’ simulation. The five dams simulation was intended to test the combined effects of five dams on residual streamflow, since this was important for ecological assessment. A further simulation that included transparent flows was also run, to estimate the effects of such management on residual flows as well as 85% annual supply reliability estimates. Selected transparent flow thresholds were somewhat arbitrary but selected to provide a significant transparent flow that is also comparable to water harvest pump start thresholds and without having major impacts on yields at 85% annual exceedance (Table 6-6). Inclusion of transparent flow at the selected thresholds does decrease yields by 13 to 29%, while using the same FSL. This will mean that the unit cost of supplied water will increase, although in most cases the sensitivity of unit cost of supplied water to transparent flow thresholds at the selected FSLs were relatively low. This can be observed as a ‘bulge’ in the contour plots of unit costs for all dams(Figure 6-6, Figure 6-7, Figure 6-8, Figure 6-9 and Figure 6-10). Map. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\1_GIS\1_Map_docs\1_Exports\Hy-R-511-5_dams_location_v2.png" Figure 6-5 Reservoir surface area related to dams and a selection of river model nodes "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\7_RiverModel\14_river_9\11_dams\2_output\1_dam_55\dam_55_4_panel_plot.png" Figure 6-6 The effect of full supply level and transparent flow threshold on 85% exceedance annual yield, unit capital cost, net mean annual reservoir evaporation and reservoir mean annual spill at site 55 "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\7_RiverModel\14_river_9\11_dams\2_output\2_dam_79\dam_79_4_panel_plot.png" Figure 6-7 The effect of full supply level and transparent flow threshold on 85% exceedance annual yield, unit capital cost, net mean annual reservoir evaporation and reservoir mean annual spill at site 79 "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\7_RiverModel\14_river_9\11_dams\2_output\3_dam_108\dam_108_4_panel_plot.png" Figure 6-8 The effect of full supply level and transparent flow threshold on 85% exceedance annual yield, unit capital cost, net mean annual reservoir evaporation and reservoir mean annual spill at site 108 "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\7_RiverModel\14_river_9\11_dams\2_output\4_dam_137\dam_137_4_panel_plot.png" Figure 6-9 The effect of full supply level and transparent flow threshold on 85% exceedance annual yield, unit capital cost, net mean annual reservoir evaporation and reservoir mean annual spill at site 137 "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\7_RiverModel\14_river_9\11_dams\2_output\5_dam_145\dam_145_4_panel_plot.png" Figure 6-10 The effect of full supply level and transparent flow threshold on 85% exceedance annual yield, unit capital cost, net mean annual reservoir evaporation and reservoir mean annual spill at site 145 Table 6-6 Dam full supply levels, yields at 85% annual exceedance and transparent flow thresholds used for combined five dams simulations For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Reservoir transparent flow thresholds are designed primarily to preserve lower flows where timing of releases is timed to mimic natural flows. The thresholds used in this Assessment were similar to the pump start thresholds used in water harvest, allowing comparisons to be made via ecological analyses. Comparison of flow quantiles between Scenario A, the five dams simulation and the five dams with transparent flow simulation give some indication of the success of transparent flow in restoring low flow (Table 6-7). Inclusion of transparent flow thresholds (Table 6-6) largely restores 90, 80 and 50% exceedance flow to Scenario A levels. This observation is repeated from dams along the river system to the final node at 90300000. Table 6-7 Mean annual flow and exceedance flow for dam affected gauges for Scenario A and five dams simulations with and without transparent flow For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 6.2.4 The effect of groundwater extraction on river flow The effects of a hypothetical groundwater development in the Sturt Plateau (35 GL/year in addition to current licence volumes) were estimated at node 90300130 using the model of Knapton (2020). These changes were then propagated to all downstream nodes using the Roper River model. Results indicate that changes to mean annual flow are modest at 8 GL (at node 90302500). The effects on 90 and 80% flow were also relatively modest, although these may still have substantial impacts on ecosystem assets reliant on dry season flows (Table 6-8). Table 6-8 Mean annual flow and flow rate at 90, 80, 50, 20 and 10% exceedance values for various river model nodes for Scenario A and flow subsequent to hypothetical groundwater development in the Larrimah area For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 6.3 Scenario C – future climate streamflow The relative rainfall and ET of each GCM in the Roper catchment is shown in Figure 6-11 . The selected dry, mid and wet GCMs were GISS–E2–1–H, ACCESS–CM2 and INM–CM4–8 respectively. These approximated the 0.1, 0.5 and 0.9 quantile future climate rainfall. These data were used to ‘pattern scale’ historical SILO-gridded data, which were subsequently aggregated to a single time- series input for each river model subcatchment. For each of these future climate scenarios (i.e. wet, mid and dry), streamflow estimates were simulated for a period corresponding to the historical period (Table 6-9). Relative change to Scenario A flow for two river model nodes is shown in Table 6-10. Interestingly, a rainfall reduction of around 10% (dry scenario) results in a reduction of streamflow of around 35%, while an increase of rainfall of around 10% results in a streamflow increase of around 41%. These responses to climate input are largely a result of the RR models calibrated at nodes 90300890 and 90302500, since these parameters are utilised widely across the model domain. The calibration process and the model structure of conceptual models implicitly captures the streamflow response of the calibration period. To that end, a long calibration period with variable conditions is imperative. However, it must be stated, that for input conditions outside of this range, such as future climate estimates, estimates must be treated with caution. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Figure 6-11 Percentage change in mean annual rainfall and potential evaporation under Scenario C relative to under Scenario A Simple scaling of rainfall and potential evaporation have been applied to global climate model output (GCM-PS) GCM- PSs are ranked by increasing rainfall. Table 6-9 Annual flow exceedance values, mean runoff coefficient and mean annual flow for three future climate scenarios For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Table 6-10 Mean annual flow relative to Scenario A for three future climates at two river model node locations For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 6.4 Scenario D – future climate and water resource development Scenario D simulations are designed to estimate the effects of water resource development in addition to alternative future climates. In terms of water resource availability, the most important climate future is the dry climate since this will reduce streamflow and hence availability to all users, including ecosystem assets. These reductions, in addition to any diversions, will further reduce streamflow with consequences for streamflow-dependant ecosystems and/or diversion supply reliability. 6.4.1 Dry future climate and water harvest The dry climate future in conjunction with soil-limited water harvest analysis was conducted and results in terms of streamflow are tabulated below (Table 6-11). The water harvest parameters for this simulation remained unchanged from those outlined in Section 6.2.2, however in this case, no end-of-system requirement was enforced. Additionally, system and reach irrigation target volumes remained unchanged between Scenario B and this simulation. Given the large reduction in streamflow exhibited for the Cdry scenario relative to Scenario A, reliability of supply will be lower for all water harvest nodes (i.e. the minimum of 75% annual supply reliability was not enforced). In terms of mean flow reductions relative to Scenario A, Scenario D water harvest is the same as the addition of Scenario B water harvest and Scenario Cdry (Table 6-12 ), although Scenario D has large effects on annual flow for higher exceedance probabilities. Table 6-11 Mean annual flow and annual flow exceedance values for Scenario D (dry + water harvest) For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au † Only those nodes effected by water harvest are listed. Table 6-12 Mean annual flow and annual flow exceedance values at the end-of-system (90300000) for Scenario B (water harvest), Cdry, D (dry climate + water harvest) relative to Scenario A. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 6.4.2 Dry climate future and instream dams The dry climate future combined with all five instream dams was combined to assess the effect on streamflow (Table 6-13). For this simulation, dam parameters including FSL and irrigation target volumes were unchanged from Scenario B assessments (Section 6.2.3) and no transparent flow was enforced. For these reasons, reliability of supply was reduced relative to Scenario B simulations (where supply volumes at 85% annual reliability were optimised). Reduction in mean annual flow for Scenario D relative to Scenario A was 37% (Table 6-14). This was of a similar magnitude to the dry climate + water harvest scenario. However, the dry climate + dams scenario exhibited higher reductions in the higher flow years (10 and 20% annual exceedance) and lower low-flow year reductions relative to the dry climate + water harvest scenario. This may seem somewhat counter-intuitive, given that instream dams are usually considered to have greater impacts on surface water flow regimes than diversion into offstream storages via pumps. In this case, the five dam sites were all located in headwaters (Figure 6-5) while the simulated water harvest operations were distributed across the catchment (Table 6-4 ) and had access to a much greater catchment/contributing area than the instream dams. In dry years, the instream dams had a reduced capacity to store water due to their position in headwater areas, while the water harvest operations, particularly those on the Roper River, could access proportionally larger volumes of water at those times. Table 6-13 Mean annual flow and annual flow exceedance values for Scenario D (dry + five dams) For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au † Only those nodes effected by instream dams are listed. Table 6-14 Mean annual flow and annual flow exceedance values at the end-of-system (90300000) for Scenario B (five dams), Scenario Cdry, and Scenario D (dry climate + five dams) relative to Scenario A For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 7 Discussion and conclusion Development of the Roper River model was challenging for various reasons. Hydrological prediction is difficult in monsoonal northern Australia, and in this instance a lack of data, particularly, but not limited to streamflow observations, contributed to the challenges of model construction and calibration. The lack of data, including a poor spatial coverage of suitable streamflow data, required extensive parameter transfer across the catchment, which in turn contributed to model uncertainty. An apparent temporal trend in rainfall across the catchment and sparse rainfall data in the northern portions of the catchment are also acknowledged. In such situations the hydrologist must make various decisions related to calibration routine, calibration objective functions, RR models, gauge suitability and parameter transfer protocols, all of which will influence the river model simulation characteristics. In construction and calibration of the Roper River model, two RR models were tested in conjunction with various calibration objective functions, calibration shingle configurations and parameter transfer protocols in an iterative fashion. In all, 12 versions of the model were at least partially constructed and calibrated. Any of these models were considered adequate; however, one model was deemed to be most appropriate for scenario analyses as judged by various goodness-of-fit metrics, but also by the behaviour of the model in conditions that were outside of the calibration date range (during the early 20th century) and the spatial coherence of runoff estimates across the catchment. The runoff coefficient was very low in the gauged area upstream of node 90300010. The gauge here, Elsey Creek at Warlock Ponds, had good gauging data and a long record, and hence was used for all calibrations. It was also noted that this area coincided with the Sturt Plateau(Figure 2-1), and had very low stream length density suggesting high infiltration of rainfall and low runoff. The final model version chosen (v12) explicitly modelled regional groundwater discharge in the upper Roper and subsequent losses in the mid-Roper that are vital for effective ecosystem analyses. The Mataranka area just up and downstream of the Elsey Creek and Roper River junction features numerous groundwater springs, and gauges in this area (9030176 and 9030013) show high baseflow. Some of this baseflow originates from outside of the model subcatchments, and outside of the entire Roper catchment area, thereby violating the assumption of mass balance by subcatchment that is the foundation of most river models. Water movement across catchment boundaries can be included in models although there is rarely enough data to constrain such models, in which case over-fitting becomes a possibility. In this Assessment the river model used conforms to mass balance by explicitly including regional groundwater discharge (from a jurisdictional groundwater model) as an input at a node representing gauge location 9030013. However, there are also peculiarities with these observations (gauge 9030013, Appendix C (Apx Figure C-5)), in that baseflow (and runoff coefficient) has increased dramatically from the pre-1975 period to the post-2010 period. It is unknown if this is an observational error or there have been genuine and dramatic changes in streamflow characteristics at this site. Such observations suggest a non-stationary streamflow response to increasing rainfall, at least in the Mataranka area. The mean annual end-of-system flow for the Roper River at node 90300000 for the reporting period (1910–2019) was estimated to be 5557 GL. Uncertainty estimated using other river model versions suggests a range of plus or minus 12% for this value. Further upstream, where good data was available, streamflow estimates where more certain. For example, mean annual flow at node 90302500 was estimated to be 2371 GL plus or minus 2%. Water harvest analyses suggested that uncertainty in water harvest estimates (using all suitable model versions) were more sensitive to aspects of flow duration, particularly the match of simulated and observed data at flow values around pump start threshold flows, than to total model bias. An analysis at a single node suggested a range of plus or minus 15% variation for water harvest volumes. It should be noted that this estimate does not consider errors in observed streamflow and rainfall, and is therefore likely to be an underestimate of the true uncertainty. Soils information suggested that the availability of soils suitable for irrigation and ringtank construction was limited, and these combined with streamflow availability meant that the maximum possible water harvest volume that could be supplied at a minimum of 75% annual reliability was 660 GL, although even this assumed a very low pump start threshold (200 ML/day) and did not consider any other restrictions due to economic, environmental, cultural or land tenure influences. This means that a more holistic assessment considering those factors could substantially reduce the availability estimate of water for diversion. Additionally, one must consider the uncertainty of this estimate, and the potential for lower availability than calculated. The behaviour, yield and effects of streamflow at potential dam sites were considered at five locations within the catchment. These sites were selected from a more extensive assessment of water storage locations across the catchment (Petheram et al., 2022). Analyses considered various dam FSLs and transparent flow thresholds to determine the most cost-effective FSL in each case. These parameters were then used to simulate the effect of all five dams concurrently. Using an 85% annual supply reliability benchmark, 324 GL/year could be supplied if no transparent flow is assumed. Groundwater development in the Larrimah area was estimated to have modest impacts on low flows in the Roper River in the vicinity of Mataranka and in areas downstream. Effects on mean annual flow were relatively modest at a catchment scale (a reduction of mean annual flow of 8 GL at node 90300000). The effects of groundwater development (using current development levels plus an additional 35 GL/year) on low flows within the Mataranka to Red Rocks section of the Roper River (or node 903001300 to node 90302500) was also modest all scenarios examined (Figure 7-1). It should be noted that the magnitude of groundwater development effects on river flow used an estimate at the year 2060, since both present and hypothetical development effects will have had relatively small impacts at the present time, i.e. there is a lag between groundwater development and it’s effect on groundwater discharge to the river. These reductions in flow due to groundwater development will be most obvious during the dry season and may affect ecological assets that are related to dry-season flows in this portion of the river. Charts. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Figure 7-1 A comparison of flow duration curves for various scenarios at (a) and (b) node 90300000 and (c) and (d) node 90302500. Future climate analysis in the Roper catchment indicates that there is a large uncertainty in future rainfall. The GCM future rainfall for all 21 selected models ranges from decreases of 27% to increases of 10%. The selected dry future climate used in this Assessment was 10% drier than the 1910–2019 catchment mean annual rainfall, while the selected wet future climate was 10% wetter. Propagation of these future climates resulted in mean flow estimate changes of –35% and +41% for the dry and wet future climates respectively. The dry climate scenario (Cdry) has impacts across the entire flow range, but has notably large impacts at high flows (Figure 7-1).The dry climate scenario has a far greater impact on mean annual flow than water harvest (Table 6-12 ) and the five dams scenario (Table 6-14). Similarly, under the dry future climate, the water harvest and instream dams scenarios have a large, but similar impact in terms of changes in mean annual flow. Notably, under both future dry climate and historical climate, the five dams scenario has a much lower impact in drier/low-flow years than water harvest scenarios as indicated by the 80 and 90% annual exceedance flows (Table 6-12 and Table 6-14). References Allen RG, Pereira LS, Raes D and Smith M (1998) Crop evapotranspiration – Guidelines for computing crop water requirement. FAO Irrigation and Drainage Paper 56, Food and Agriculture Organization of the United Nations. Ardia D, Boudt K, Carl P, Mullen KM and Peterson BG (2010) Differential evolution with ‘DEoptim’: an application to non–convex portfolio optimization. R Journal 3(1), 27–34. DOI:10.32614/RJ–2011–005. Bloschl G, Sivapalan M, Wagener T, Vigliione A and Savenjie H (eds) (2013) Runoff prediction in ungauged basins: synthesis across processes, places and scales. Cambridge University Press, Cambridge. Burnash RJC (1995) The NWS River Forecast System – Catchment Modeling. In: Vijay P Singh (ed) Computer models of watershed hydrology. Revised edition. Water Resources Publications, Highlands Ranch, Colo. Carlston CW and Langbein WB (1960) Rapid approximation of drainage density line intersection method. US Geol. Surv., Water Resource Division Bulletin 11. Chiew FHS and McMahon TA (1991) Applicability of Morton’s and Penman’s evapotranspiration estimates in rainfall–runoff modeling. Water Resources Bulletin 27, 611–620. Chiew FHS and Leahy CP (2003) Comparison of evapotranspiration variables in evapotranspiration maps for Australia with commonly used evapotranspiration variables. Australian Journal of Water Resources 7, 111. Chiew FHS, Teng J, Vaze J, Post DA, Perraud JM, Kirono DGC and Viney NR (2009) Estimating climate change impact on runoff across southeast Australia: method, results, and implications of the modeling method. Water Resources Research 45, W10414. DOI:10. 1023/a:1026035305597. Chiew FHS, Kirono DGC, Kent DM, Frost AJ, Charles SP, Timbal B, Nguyen KC and Fu G (2010) Comparison of runoff modelled using rainfall from different downscaling methods for historical and future climates. Journal of Hydrology 387. DOI:10.1016/j.jhydrol.2010.03.025. CSIRO (2009a) Water in the Fitzroy region. In: CSIRO (ed) Water in the Timor Sea Drainage Division. A report to the Australian Government from the CSIRO Northern Australia Sustainable Yields Project. CSIRO Water for a Healthy Country Flagship, Australia, 61–128. CSIRO (2009b) Water in the Mitchell catchment. In: CSIRO (ed) Water in the Gulf of Carpentaria Drainage Division. A report to the Australian Government from the CSIRO Northern Australia Sustainable Yields Project. CSIRO Water for a Healthy Country Flagship, Australia, 347–416. CSIRO (2009c) Water in the Van Diemen region. In: CSIRO (ed) Water in the Timor Sea Drainage Division. A report to the Australian Government from the CSIRO Northern Australia Sustainable Yields Project. CSIRO Water for a Healthy Country Flagship, Australia, 363–452. CSIRO (2011) Water: science and solutions for Australia. CSIRO Land and Water. DHI (2003) MIKE 11–A Modelling System for Rivers and Channels – User Guide. DHI Water & Environment Pty Ltd. DHI (2012) MIKE 21–2D Modelling of Coast and Sea. DHI Water & Environment Pty Ltd. Doble RC, Crosbie RS, Smerdon BD, Peeters L and Cook FJ (2012) Groundwater recharge from overbank floods. Water Resources Research 48, W09522, DOI:Link to external resource: 10.1029/2011WR011441 . Dutta D, Kim S, Hughes J, Vaze J and Yang A (2015) AWRA–R v5.0 Technical Report. CSIRO Land and Water. Dutta D, Kim S, Hughes J, Vaze J, Yang A, Teng J and Lerat J (2017) Development and application of a large scale river system model for national water accounting in Australia. Journal of Hydrology 547, 124–142. Grigg AH and Hughes JD (2018) Non–stationarity driven by multidecadal change in catchment groundwater storage. A test of modifications to a common rainfall–run–off model.Hydrological Processes 32, 3675–3688. Holz L, Kim S, Petheram C, Podger G, Hughes J, Kehoe M, Aramini D, Podger S, Lerat J, Poulton P, Hornbuckle J, Perraud J (2013) River system modelling for the Flinders and Gilbert Agricultural Resource Assessment case study analysis. A technical report to the Australian Government from the CSIRO Flinders and Gilbert Agricultural Resource Assessment, part of the North Queensland Irrigated Agriculture Strategy. Canberra: CSIRO; 2013. csiro:EP132039. doi.org/10.4225/08/584d97ff6738e Horton RE (1945) Erosional development of streams and their drainage basins: a hydrophysical approach to quantitative morphology. Geological. Society America Bulletin 56(3), 275–370. Hughes JD, Silberstein RP and Grigg AH (2013) Extending rainfall–runoff models for use in environments with long–term catchment storage and forest cover. Proceeding of 20th International Congress on Modelling and Simulation, Adelaide, Australia, 1–6 December 2013. Hughes JD, Mainuddin M, Lerat J, Dutta D, Kim SSH (2013b) An irrigation model for use in river systems modelling, in: 20th International Congress on Modelling and Simulation, Modelling and Simulation Society of Australia and New Zealand, Adelaide. Hughes JD, Dutta D, Vaze J, Kim SSH, Podger G.(2014) An automated multi–step calibration procedure for a river system model. Environmental Modelling and Software 51, 173 – 183. https://doi.org/10.1016/j.envsoft.2013.09.024. Hughes J, Dutta D, Yang A, Marvanek S, Vaze J, Mainuddin M, (2014b) Improvement of AWRA–R Irrigation Model and Its Applications in the Murray–Darling Basin. Technical Report. CSIRO, Land and Water. Hughes JD and Vaze J (2015) Non–stationarity driven by long–term change in catchment storage: possibilities and implications, Proc. IAHS, 371, 7–12, https://doi.org/10.5194/piahs–371–7– 2015. Hughes J, Yang A, Wang B, Marvanek S, Carlin L, Seo L, Petheram C and Vaze J (2017) Calibration of river system and landscape models for the Fitzroy, Darwin and Mitchell catchments. A technical report to the Australian Government from the CSIRO Northern Australia Water Resource Assessment, part of the National Water Infrastructure Development Fund: Water Resource Assessments. CSIRO, Australia. Hughes J, Yang A, Wang B, Marvanek S, Seo L, Petheram C and Vaze J (2018) River model simulation for the Fitzroy, Darwin and Mitchell catchments. A technical report to the Australian Government from the CSIRO Northern Australia Water Resource Assessment, part of the National Water Infrastructure Development Fund: Water Resource Assessments. CSIRO, Australia. Hughes JD, Potter N, Zhang L and Bridgart R (2021) Conceptual model modification and the millennium drought of southeastern Australia. Water 13, 669. Viewed 11 January 2023, Hyperlink to: Conceptual model modification and the millennium drought of southeastern Australia . IPCC (2015) Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team, R.K. Pachauri and L.A. Meyer (eds.)]. IPCC, Geneva, Switzerland, 151 . IPCC (2022) Climate Change 2022: Impacts, Adaptation, and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press. Cambridge University Press, Cambridge, UK and New York, NY, USA. Kerle E, Cruickshank S (2014) End of dry season stream flow measurements, Roper River, 2014. Report Number 13/2015D. Department of Land Resource Management, Palmerston, NT. Knapton A (2020) Upgrade of the coupled model of the Cambrian Limestone Aquifer and Roper River Systems. WRS Technical Report 57/2020. Prepared for the Department of Environment, Parks and Water Security by Cloud GMS, Goodwood SA. Knapton A, Taylor AR, Petheram C and Crosbie RS (2023) An investigation into the effects of climate change and groundwater development scenarios on the water resources of the Roper catchment using two finite element groundwater flow models. A technical report from the CSIRO Roper River Water Resource Assessment for the National Water Grid. CSIRO, Australia. Koussis AD (1980) Comparison of Muskingum method difference scheme. Journal of the Hydraulic Division ASCE 106, 925–929. Lerat J, Dutta D, Kim SSH, Hughes JD, Vaze J (2013). Reducing propagation of uncertainty in river system modelling by optimal use of streamflow data, in: 20th International Congress on Modelling and Simulation, Modelling and Simulation Society of Australia and New Zealand, Adelaide. Lyne V and Hollick M (1979) Stochastic time–variable rainfall–runoff modelling. Institute of Engineers Australia National Conference Publication 79/10, 89–93. McJannet D, Yang A and Seo L (2023) Climate data characterisation for hydrological and agricultural scenario modelling across the Victoria, Roper and Southern Gulf catchments. A technical report from the CSIRO Victoria, Roper and Southern Gulf Water Resource Assessments for the National Water Grid. CSIRO, Australia. Mitchell TD (2003) Pattern scaling: an examination of the accuracy of the technique for describing future climates. Climatic Change 60, 217–242. DOI:10.1029/2008WR007338. Moise A, Wilson L, Grose M, Whetton P, Watterson I, Bhend J, Bathols, J Hanson L, Erwin T, Bedin T, Heady C, Rafter T (2015) Evaluation of CMIP3 and CMIP5 Models over the Australian Region to Inform Confidence in Projections. Australian Meteorological and Oceanographic Journal, 65(1), 19–53, 10.22499/2.6501.004. Morton FI (1983) Operational estimates of areal evapotranspiration and their significance to the science and practice of hydrology. Journal of Hydrology 66, 1–76. Nash JE and Sutcliffe JV (1970) River flow forecasting through conceptual models part I–A discussion of principles. Journal of Hydrology 10(3), 282–290. Viewed 11 January 2023, River flow forecasting through conceptual models part I–A discussion of principles . O'Loughlin EM (1986) Prediction of Surface Saturation Zones in Natural Catchments by Topographic Analysis. Water Resour. Res. 22( 5), 794– 804, DOI:10.1029/WR022i005p00794. Parajka J, Viglione A, Rogger M, Salinas JL, Sivapalan M and Blöschl G (2013) Comparative assessment of predictions in ungauged basins – Part 1: Runoff–hydrograph studies, Hydrology and Earth System Sciences 17, 1783–1795. Viewed 11 January 2023, Hyperlink to: Comparative assessment of predictions in ungauged basins – Part 1: Runoff-hydrograph studies . Petheram C, Rustomji P, Vleeshouwer J (2009). Rainfall–runoff modelling across northern Australia.CSIRO:EP113491. https://doi.org/10.4225/08/584af30de743b Petheram C, Rustomji P, Chiew FHS and Vleeshouwer J (2012) Rainfall–runoff modelling in northern Australia: a guide to modelling strategies in the tropics. Journal of Hydrology 462– 463, 28–41. Viewed 11 January 2023, Hyperlink to: Rainfall–runoff modelling in northern Australia: a guide to modelling strategies in the tropics . Petheram C, Watson I and Stone P (2013a) Agricultural resource assessment for the Flinders catchment. A report to the Australian Government from the CSIRO Flinders and Gilbert Agricultural Resource Assessment, part of the North Queensland Irrigated Agriculture Strategy. CSIRO Water for Healthy Country and Sustainable Agriculture flagships, Australia. Petheram C, Watson I and Stone P (2013b) Agricultural resource assessment for the Gilbert catchment. A report to the Australian Government from the CSIRO Flinders and Gilbert Agricultural Resource Assessment, part of the North Queensland Irrigated Agriculture Strategy. CSIRO Water for Healthy Country and Sustainable Agriculture flagships, Australia. Petheram C, Rogers L, Read A, Gallant J., Moon A, Yang A, Gonzalez D, Seo L, Marvanek S, Hughes J, Poncy–Reyes R, Wilson P, Wang B, Ticehurst C, Barber M (2017). Assessment of surface water storage options in the Fitzroy, Darwin and Mitchell catchments. Technical Report. CSIRO. A technical report to the Australian Government from the CSIRO Northern Australia Water Resource Assessment, part of the National Water Infrastructure Development Fund: Water Resource Assessments. Petheram C, Bruce C, Chilcott C and Watson I (eds) (2018a) Water resource assessment for the Fitzroy catchment. A report to the Australian Government from the CSIRO Northern Australia Water Resource Assessment, part of the National Water Infrastructure Development Fund: Water Resource Assessments. CSIRO, Australia. Petheram C, Chilcott C, Watson I and Bruce C (eds) (2018b) Water resource assessment for the Darwin catchments. A report to the Australian Government from the CSIRO Northern Australia Water Resource Assessment, part of the National Water Infrastructure Development Fund: Water Resource Assessments. CSIRO, Australia. Petheram C, Watson I, Bruce C and Chilcott C (eds) (2018c) Water resource assessment for the Mitchell catchment. A report to the Australian Government from the CSIRO Northern Australia Water Resource Assessment, part of the National Water Infrastructure Development Fund: Water Resource Assessments. CSIRO, Australia. Petheram C, Gallant J, Stone P, Wilson P, Read A (2018d). Rapid assessment of potential for development of large dams and irrigation across continental areas: application to northern Australia. The Rangeland Journal 40, 431–449. Petheram C, Yang A, Seo L, Rogers L, Baynes F, Devlin K, Marvanek S, Hughes J, Ponce Reyes R, Wilson P, Stratford D, Philip S (2022) Assessment of surface water storage options and reticulation infrastructure in the Roper catchment. A technical report from the CSIRO Roper River Water Resource Assessment for the National Water Grid. CSIRO, Australia. Perrin C, Michael C and Andréassian V (2003) Improvement of a parsimonious model for streamflow simulations. Journal of Hydrology 279, 275–289. Plumb KARHG (1992) The geology of Arnhem Land, Northern Territory. Minerals and Land Use Program, Bureau of Mineral Resources, Geology and Geophysics, Canberra City, ACT. Pollino C, Ash A, Ebner B, Grice T, Irvin S, Kenyon R, Merrin L, Nielson D, O’Sullivan J, Paini D, Palmer J, Petheram C, Stratford D, Taylor A, Thomas M (2018). Ecological, biosecurity, offsite and irrigation–induced salinity risks, in: Petheram C, Watson I, Bruce C, Chilcott C. (Eds.), Water resource assessment for the Mitchell catchment. CSIRO. chapter 7, 368 – 435. A report to the Australian Government from the CSIRO Northern Australia Water Resource Assessment, part of the National Water Infrastructure Development Fund: Water Resource Assessments. Prudhomme C and Davies H (2009) Assessing uncertainties in climate change impact analyses on the river flow regimes in the UK. Part 2: Future climate. Climatic Change 93, 197–222. DOI:10.1007/s10584–008–9461–6. Read A, Gallant J, Petheram C (2012). DamSite: An automated method for the regional scale identification of dam wall locations, in: Smith, G. (Ed.), Hydrology and Water Resources Symposium, Engineers Australia: Canberra, ACT. 1 – 10. Salathe EP (2003) Comparison of various precipitation downscaling methods for the simulation of streamflow in a rain shadow river basin. International Journal of Climatology 23, 887–901. DOI:10.1002/joc.922. Sivapalan M (2003) Prediction in ungauged basins: a grand challenge for theoretical hydrology. Hydrological Processes 17, 3163–3170. Viewed 11 January 2023, Hyperlink to: Prediction in ungauged basins: a grand challenge for theoretical hydrology . Stephens GL, L'Ecuyer T, Forbes R, Gettlemen A, Golaz J–C, Bodas–Salcedo A, Suzuki K, Gabriel P and Haynes J (2010) Dreary state of precipitation in global models. J. Geophys. Res. 115(D24), D24211. DOI: 10.1029/2010jd014532. Strahler AN (1956) The nature of induced erosion and aggregation. In: Thomas WL (ed) Man’s role in changing the face of the Earth. Thomas, Chicago–Illinois. Taylor AR, Crosbie RS, Turnadge C, Lamontagne S, Deslandes A, Davies PJ, Barry K, Suckow A, Knapton A, Marshall S, Hodgson G, Tickell S, Duvert C, Hutley L and Dooley K (2023) Hydrogeological assessment of the Cambrian Limestone Aquifer and the Dook Creek Aquifer in the Roper catchment, Northern Territory. A technical report from the CSIRO Roper River Water Resource Assessment for the National Water Grid. CSIRO, Australia. Thomas M, Philip S, Stockman U, Wilson PR, Searle, R, Hill J, Bui E, Gregory, L, Watson, I, Wilson PL and Gallant G (2022) Soils and land suitability for the Roper catchment, Northern Territory. A technical report from the CSIRO Roper River Water Resource Assessment for the National Water Grid. CSIRO, Australia. Viney N, Vaze J, Crosbie R, Wang B, Dawes W and Frost A (2015) AWRA–L v5.0: Technical description of model algorithms and inputs. Wang QJ, McConachy FLN, Chiew FHS, James R, de Hoedt GC and Wright WJ (2001) Climatic Atlas of Australia, Maps of Evapotranspiration. Zhang Z and Dawes W (1998) WAVES – An integrated energy and water balance model. CSIRO Land and Water Technical Report No. 31/98: CSIRO Land and Water, Canberra. Part II Appendices Morton’s Wet Area Potential ET calculation static double CalculateETwp(DateTime today, double Elevation, double Latitude, double tmax, double tmin, double eact, double radin, out double RH) { double height, lat, tavg, esat, pratio, gammap, pdelta, pi, psai, dr, delta, omega, radextra, radinnet, radout, radnet, nNratio, _as, bs, albedo, sigma, term1, term2, term3, fz, ediff, stabfac, Dpfact, vptc, htc, xesat, xtemp, xdelta, tempinc, radnetx, ETpp, ETppx, ETwp, tdiff, esatMax, esatMin, esatMean, ea; int julday; //psea = 101.3; //kPa pi = 3.1415927; _as = 0.25; // Angstorm formula, regression constant bs = 0.50; // Angstorm formula, regression constant albedo = 0.23; sigma = 4.903E–9; // Stefan–Boltzmann constant height = Elevation; lat = Latitude; if (tmax <= –99.0 || tmin <= –99.0 || eact <= –99.0 || radin <= –99.0) // || esat <= –99.0 || radin == 0 ) //|| sunhrs <= –99.0) { ETpp = –99.0; ETwp = –99.0; ea = –99.0; RH = –99.0; tavg = –99.0; } else { tavg = (tmax + tmin) / 2.0; //calculate esat //formula 35 esatMax = 0.6108 * Math.Exp(17.27 * tmax / (tmax + 237.3)); esatMin = 0.6108 * Math.Exp(17.27 * tmin / (tmin + 237.3)); esatMean = 0.6108 * Math.Exp(17.27 * tavg / (tavg + 237.3)); //formula 43 ea = 0.25 * esatMax + 0.5 * esatMean + 0.25 * esatMin; esat = ea; if (esat <= 0.0) esat = 0.0001; if (eact <= 0.0) eact = 0.0001; if (eact > esat) eact = esat; RH = eact / esat; // Calculate ratio of atmospheric at the station to that at the sea level (Pstn/Psea) pratio = Math.Pow((293.0 – 0.0065 * height) / 293.0, 5.26); // Calculate psychrometric constant (kPa/C) if (tavg >= 0.0) gammap = 0.066 * pratio; else gammap = 0.0574 * pratio; // Calculate slope of saturation vapour pressure/temperature curve (kPa/C) if (tavg >= 0.0) pdelta = (4098.0 * esat) / Math.Pow((tavg + 237.3), 2); else pdelta = (5809.0 * esat) / Math.Pow((tavg + 265.5), 2); // Calculate extraterrestrial radiation (Ra=radextra) (MJ/m2/day) psai = (lat / 180.0) * pi; julday = today.DayOfYear; dr = 1.0 + 0.033 * Math.Cos(0.0172 * Convert.ToDouble(julday)); delta = 0.409 * Math.Sin(0.0172 * Convert.ToDouble(julday) – 1.39); omega = Math.Acos(–1.0 * Math.Tan(psai) * Math.Tan(delta)); radextra = (118.1 / pi) * dr * (omega * Math.Sin(psai) * Math.Sin(delta) + Math.Cos(psai) * Math.Cos(delta) *Math.Sin(omega)); // Calculate nNratio based on radin nNratio = (radin / radextra – _as) / bs; // Calculate NET incoming solar radiation (Rns=radinnet) radinnet = (1.0 – albedo) * radin; // Calculate net outgoing longwave radiation (Rnl=radout) term1 = (Math.Pow((tmax + 273.16), 4) + Math.Pow((tmin + 273.16), 4)) / 2.0; term2 = (0.34 – 0.14 * Math.Sqrt(eact)) * (0.10 + 0.9 * nNratio); radout = sigma * term1 * term2; if (radout < 0.0) radout = 0.0; // Calculate net radiation (Rn) : if negative set it to zero radnet = radinnet – radout; if (radnet < 0.0) radnet = 0.0; // Calculate stability factor(stabfac), vapour pressure transfer // coefficient(fa=vptc)and heat transfer coeffieient(lamda=htc) if (tavg >= 0.0) fz = 24.19; else fz = 27.82; ediff = esat – eact; if (ediff <= 0.0) ediff = 0.0001; term3 = gammap * (Math.Sqrt(1.0 / pratio)) * fz * ediff; // term3=gammap*((1.0/pratio)**0.5)*fz*ediff !!!can be wrong because of Jai stabfac = 1.0 / (0.28 * (1.0 + eact / esat) + pdelta * radnet / term3); if (stabfac < 1.0) stabfac = 1.0; //!!!!! MODIFICATION vptc = (Math.Sqrt(1.0 / pratio)) * fz / stabfac; //vptc=((1.0/pratio)**0.5)*fz/stabfac htc = gammap + (1.804E–8 * Math.Pow((tavg + 273.0), 3)) / vptc; //htc=gammap+(1.804E– 8*(tavg+273.0)**3)/vptc; // Carryout iterative procedure to satisfy the energy balance and obtain // equlibrium quantities xesat = esat; xtemp = tavg; xdelta = pdelta; do { tempinc = (radnet / vptc + eact + htc * (tavg – xtemp) – xesat) / (xdelta + htc); tdiff = Math.Abs(tempinc); if (tdiff < 0.01) break; else { xtemp = xtemp + tempinc; if (xtemp >= 0.0) xesat = 0.6108 * Math.Exp((17.27 * xtemp) / (xtemp + 237.3)); else xesat = 0.6108 * Math.Exp((21.88 * xtemp) / (xtemp + 265.5)); if (xtemp >= 0.0) xdelta = (4098.0 * xesat) / Math.Pow((xtemp + 237.3), 2); else xdelta = (5809.0 * xesat) / Math.Pow((xtemp + 265.5), 2); } } while (true); ETppx = radnet – htc * vptc * (xtemp – tavg); ETpp = ETppx * 0.408; if (ETpp < 0.0) ETpp = 0.0; // Calculate Morton Wet Environment Areal Potential Evapotranspiration. ETwp radnetx = ETppx + gammap * vptc * (xtemp – tavg); Dpfact = xdelta / (gammap + xdelta); ETwp = 0.408 * (1.2096 + 1.2 * Dpfact * radnetx); } return ETwp; } GR7J modification GR7J is a modification of the widely used GR4J rainfall–runoff (RR) model (Perrin, 2003). The modifications within GR7J, and its companion GR8J, were intended to improve model performance in situations where there are long–term changes in catchment storage. For example, in situations where extended droughts are apparent, RR models will often under–estimate catchment evapotranspiration (ET) and over–estimate runoff. GR7J/8J was modified to better predict in such situations by modifying the production store. More specifically, the relationship between model store and model ET could be modified to suit catchment conditions more than GR4J and, in particular, runoff sensitivity to model storage can be quite different to ET sensitivity to model storage. These changes are outlined in the conceptual diagram below (Apx Figure B-1). Conveptual diagram of GR4J and GR7J/8J. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\0_Working\1_Justin\7_Technical_report\cartoon_v2.tif" Apx Figure B-1 Schematic of the structure of (a) the GR4J model (Perrin et al., 2003), and (b) a modified structure that better represents long–term changes in storage and runoff (GR7J/8J) Further details on the motivation, structure and application of GR7J and GR8J are given in Hughes et al. (2013), Grigg and Hughes (2018) and Hughes et al. (2021). In previous studies it has been noted that GR4J can perform well in some respects but is generally poorer in more ephemeral river systems where low flow performance is not satisfactory (Hughes et al., 2017). It had been noted by the authors that GR7J had improved low flow, and had very good mid and high flow performance in initial testing using Roper River data. The structure of the GR4J and GR7J routing store is identical. An inconvenient feature of this structure is that total modelled flow can never reach zero (while tending towards zero, i.e. very small values). To further improve model performance, GR7J was further modified to allow losses from the routing store to be better able to represent ‘no–flow days’ in the model, which is a crucial index related to various ecological functions (labelled ‘R’ in Apx Figure B-2). These losses were represented by the following equations: 𝐿𝐿𝑡𝑡=𝑒𝑒􁉀𝑥𝑥𝑥𝑥∗𝑅𝑅𝑡𝑡 𝑥𝑥3􀵗􁉁 (19) where 𝐿𝐿𝑡𝑡 is the loss at time 𝑡𝑡 in mm, 𝑥𝑥𝑥𝑥 is the calibrated loss parameter (of nominal range –140 to –50), 𝑅𝑅𝑡𝑡 is the routing store level at time 𝑡𝑡, and 𝑥𝑥3 is the maximum routing store level. The final model outflow (𝑄𝑄𝑡𝑡) is calculated as: 𝑄𝑄𝑡𝑡=max (0,𝑄𝑄𝑄𝑄𝑡𝑡+𝑄𝑄𝑄𝑄𝑡𝑡−𝐿𝐿𝑡𝑡) (20) When GR7J was modified in this way, the model is denoted GR7JFM. An example of the loss function is shown below in Apx Figure B-2. Chart, line chart. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\0_Working\1_Justin\7_Technical_report\1_scripts\gr7j_fm_loss_plot.R" Apx Figure B-2 An example of the routing loss function used in the GR7J modification (denoted GR7JFM), for an xL parameter value of –80 Stream gauge data in the Roper catchment Apx Table C-1 Stream gauge quality codes (DEPaWS) For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Diagram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Apx Figure C-3 Stream gauge data for site 9030001 (Elsey Creek at Warlock Ponds). The dashed red line in the top and bottom left panel shows highest gauged point. Quality codes are given by yellow points in the top panel, while grey vertical lines indicate missing data Diagram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\4_Data\4_dataCheck\1_data_check_plots\G9030003_dataCheck_plot_v01.jpeg" Apx Figure C-4 Stream gauge data for site 9030003 (Wilton River at Bulman Waterhole). The dashed red line in the top and bottom left panel shows highest gauged point. Quality codes are given by yellow points in the top panel, while grey vertical lines indicate missing data For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Apx Figure C-5 Stream gauge data for site 9030013 (Roper River at Elsey Homestead). The dashed red line in the top and bottom left panel shows highest gauged point. Quality codes are given by yellow points in the top panel, while grey vertical lines indicate missing data Diagram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Description automatically generated with medium confidence Apx Figure C-6 Stream gauge data for site 9030088 (Waterhouse River near Beswick Homestead). The dashed red line in the top and bottom left panel shows highest gauged point. Quality codes are given by yellow points in the top panel, while grey vertical lines indicate missing data Diagram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\4_Data\4_dataCheck\1_data_check_plots\G9030089_dataCheck_plot_v01.jpeg" Apx Figure C-7 Stream gauge data for site 9030089 (Waterhouse River at Road Bridge). The dashed red line in the top and bottom left panel shows highest gauged point. Quality codes are given by yellow points in the top panel, while grey vertical lines indicate missing data Diagram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\4_Data\4_dataCheck\1_data_check_plots\G9030090_dataCheck_plot_v01.jpeg" Apx Figure C-8 Stream gauge data for site 9030090 (Chambers Creek at Wattle Hill). The dashed red line in the top and bottom left panel shows highest gauged point. Quality codes are given by yellow points in the top panel, while grey vertical lines indicate missing data Diagram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\4_Data\4_dataCheck\1_data_check_plots\G9030102_dataCheck_plot_v01.jpeg" Apx Figure C-9 Stream gauge data for site 9030102 (Hodgson River at Wulli Pulli). The dashed red line in the top and bottom left panel shows highest gauged point. Quality codes are given by yellow points in the top panel, while grey vertical lines indicate missing data Diagram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\4_Data\4_dataCheck\1_data_check_plots\G9030108_dataCheck_plot_v01.jpeg" Apx Figure C-10 Stream gauge data for site 9030108 (Flying Fox Creek at Mainoru Road crossing). The dashed red line in the top and bottom left panel shows highest gauged point. Quality codes are given by yellow points in the top panel, while grey vertical lines indicate missing data Diagram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\4_Data\4_dataCheck\1_data_check_plots\G9030124_dataCheck_plot_v01.jpeg" Apx Figure C-11 Stream gauge data for site 9030124 (Daly Waters Creek at Daly Waters). The dashed red line in the top and bottom left panel shows highest gauged point. Quality codes are given by yellow points in the top panel, while grey vertical lines indicate missing data Diagram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\4_Data\4_dataCheck\1_data_check_plots\G9030146_dataCheck_plot_v01.jpeg" Apx Figure C-12 Stream gauge data for site 9030146 (Wilton River at Qualari Waterhole). The dashed red line in the top and bottom left panel shows highest gauged point. Quality codes are given by yellow points in the top panel, while grey vertical lines indicate missing data Diagram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\4_Data\4_dataCheck\1_data_check_plots\G9030176_dataCheck_plot_v01.jpeg" Apx Figure C-13 Stream gauge data for site 9030176 (Roper River downstream of Mataranka Homestead). The dashed red line in the top and bottom left panel shows highest gauged point. Quality codes are given by yellow points in the top panel, while grey vertical lines indicate missing data Diagram. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\4_Data\4_dataCheck\1_data_check_plots\G9030250_dataCheck_plot_v01.jpeg" Apx Figure C-14 Stream gauge data for site 9030250 (Roper River at Red Rocks). The dashed red line in the top and bottom left panel shows highest gauged point. Quality codes are given by yellow points in the top panel, while grey vertical lines indicate missing data River model benchmark plots "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\7_RiverModel\14_river_9\6_modelCheck\1_gof_plots\benchmark_90300010_Version 12.png" Apx Figure D-15 Goodness–of–fit plots for node 90300010, model version 12 "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\7_RiverModel\14_river_9\6_modelCheck\1_gof_plots\benchmark_90300130_Version 12.png" Apx Figure D-16 Goodness–of–fit plots for node 90300130, model version 12 "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\7_RiverModel\14_river_9\6_modelCheck\1_gof_plots\benchmark_90300890_Version 12.png" Apx Figure D-17 Goodness–of–fit plots for node 90300890, model version 12 "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\7_RiverModel\14_river_9\6_modelCheck\1_gof_plots\benchmark_90301240_Version 12.png" Apx Figure D-18 Goodness–of–fit plots for node 90301240, model version 12 "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\7_RiverModel\14_river_9\6_modelCheck\1_gof_plots\benchmark_90301760_Version 12.png" Apx Figure D-19 Goodness–of–fit plots for node 90301760, model version 12 "\\fs1-cbr.nexus.csiro.au\{lw-rowra}\work\2_Hydrology\7_RiverModel\14_river_9\6_modelCheck\1_gof_plots\benchmark_90302500_Version 12.png" Apx Figure D-20 Goodness–of–fit plots for node 90302500, model version 12 As Australia’s national science agency and innovation catalyst, CSIRO is solving the greatest challenges through innovative science and technology. CSIRO. Unlocking a better future for everyone. Contact us 1300 363 400 +61 3 9545 2176 csiroenquiries@csiro.au www.csiro.au For further information Environment Dr Chris Chilcott +61 8 8944 8422 chris.chilcott@csiro.au Environment Dr Cuan Petheram +61 467 816 558 cuan.petheram@csiro.au Agriculture and Food Dr Ian Watson +61 7 4753 8606 Ian.watson@csiro.au