Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Irrigation Area, Western Australia

Transcription

1 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Irrigation Area, Western Australia Anthony J. Smith and Daniel W. Pollock CSIRO Land and Water Duncan Palmer Department of Environment, Western Australia CSIRO Land and Water Science Report 70/06 February 06 (Rev. A)

2 Copyright and Disclaimer 05 CSIRO 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 Land and Water. 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. Cover Photograph: Ivanhoe Plain, Ord River Irrigation Area Photographer: Anthony Smith 04 CSIRO Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 1

3 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Irrigation Area, Western Australia Anthony J. Smith and Daniel W. Pollock CSIRO Land and Water Duncan Palmer Department of Environment, Western Australia CSIRO Land and Water Science Report 70/06 February 06 (Rev. A) Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 2

4 Acknowledgements This research was undertaken as a partnership between the Ord River Irrigation Cooperative (OIC), CSIRO, Department of Agriculture Western Australia (DAWA), Western Australian Department of Environment (DoE) and Ord Land and Water (OLW). Project funding was provided by the Australian Government and the Government of Western Australia through the National Action Plan for Salinity and Water Quality (NAP) project Improved Water Management in the Stage 1 Ord River Irrigation Area. In-kind contributions were provided by CSIRO, DAWA, DoE, OLW, OIC and growers. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 3

5 Executive Summary Objective and Scope The purpose of this study was to identify a groundwater management strategy to control watertable rise and prevent salinity formation in the Ord Stage 1 Irrigation Area (Ord Stage 1) located at Kununurra in Western Australia. The scope of work for this project included: o a review of the available groundwater information; o construction and calibration of a groundwater flow model of the Ord Stage 1; o o o identification of the factors that contribute to and control groundwater rise within the irrigation district; delineation of groundwater management zones and salinity risk areas; and evaluation of various groundwater management options. The scope of work did not extent to the proposed Ord Stage 2 development. This project was one of fourteen linked sub-projects of the National Action Plan for Salinity and Water Quality (NAP) Project 03: Improved Water Management in the Stage 1 Ord River Irrigation Area. Historical Groundwater Rise beneath the Ord Stage 1 The salinity threat posed by rising groundwater levels was recognised relatively soon after the Ord valley was developed for irrigation in the early 1960s. Nevertheless, up until the mid-1990s there was a relatively small risk of secondary salinity because the depth to watertable was sufficiently large to prevent groundwater evaporation. The groundwater level beneath the irrigation area was monitored during this period in anticipation that the rate of subsurface drainage to the Ord River would increase in response to the mounting hydraulic gradient in the aquifer, and the watertable would re-stabilise at a safe distance below ground surface. More recently it has become clear that the degree of connectivity between the aquifer system beneath the irrigation area and the Ord River is much less than previously thought. Groundwater beneath a large part of Ivanhoe Plain does not drain effectively to the river and the aquifer has filled like a large subsurface reservoir. The watertable beneath Ivanhoe Plain rose steadily by around 0.3 to 0.5 metres per year (15 to metres in total) during the past forty years and had reached to within several metres of ground surface in some areas by the mid-1990s. The watertable beneath Packsaddle Plain rose by a similar amount but the rate of rise was initially more rapid in response to the creation and filling of Lake Kununurra. It is now apparent that the aquifers beneath Packsaddle Plain and Ivanhoe Plain are mostly full. Whereas there was no past need to manage groundwater drainage from the irrigation area because the excess replenishment from rainfall and irrigation was added to the watertable the aquifer storage capacity is now largely exhausted and a deliberate strategy is required to manage groundwater drainage and discharge to the surface environment. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 4

6 Four successive years of above average rainfall during 1998 to 01 raised the watertable beneath northern Ivanhoe Plain to above the irrigation drain inverts. The watertable has since fallen by around 1 to 2 metres in response to below average rainfall and groundwater discharge to the drains. It is anticipated that the rate of watertable fall will slow as the watertable equilibrates with the drain levels. Under the current climatic and irrigation regimes, the watertable in the palaeochannel system beneath the Ord Stage 1 is expected to stabilise around its current level, with fluctuations of 1 to 2 metres in response to variation in rainfall, evapotranspiration and irrigation practice. For example, a future sequence of above average rainfall years may again raise the watertable to the maximum historical level experienced during 01. In some areas where the palaeochannel is absent and the aquifer is poorly drained, the watertable may continue to rise until it reaches almost to the ground surface (e.g., Martins Location and the outer margins of the Ord valley). Salinity Risk and Prevention The depth to watertable below ground surface is an important aspect of salinity risk in arid regions. Continual evaporation of shallow groundwater without subsurface leaching can result in harmful accumulation of salt in the overlying soil and the loss of soil productivity. In the Ord Stage 1, salinity can develop if the depth to watertable is less than approximately 2 metres for an extended period and the subsurface drainage is insufficient to leach the accumulated salts. Other local factors such as groundwater quality and soil structure also are important. The requirement to drain excess groundwater from irrigation areas is an unavoidable element of large-scale irrigated agriculture. A portion of the applied irrigation water (leaching fraction) must be allowed to pass beyond the root zone of the crops to prevent salt accumulation and ensure sustainable growing conditions. If the total amount of water that moves beyond the plant root zone cannot drain naturally from the irrigation area without the formation of areas of shallow watertable, then a deliberate strategy to increase the rate of subsurface drainage by other means (e.g., pumps and drains) is required. In the Ord Stage 1, it is unnecessary and impractical to restore the watertable to a preirrigation state. This would be unfeasible beneath Packsaddle Plain where the raised groundwater level is controlled by the raised water level in Lake Kununurra. On Ivanhoe Plain, it is unnecessary to re-lower the watertable by 15 to metres to prevent salinity development. The available techniques for controlling groundwater drainage in irrigation areas are well known. They include improved on-farm water use efficiency to minimise the leaching fraction beneath crops; improved irrigation water storage and conveyance efficiencies to prevent leakage from the irrigation infrastructure; groundwater pumping to increase the rate of subsurface drainage and balance excess replenishment; and gravity drainage systems including deep open surface drains and subsurface drains. Ord Stage 1 Groundwater Model The Ord Stage 1 groundwater model is a computer-based tool for exploring the affects on groundwater of different land and water management options. It predicts the groundwater elevation and depth to watertable at the valley-scale but does not simulate the salt balance or salt concentration in groundwater. The model was developed and implemented in the subsurface flow and transport simulation system FEFLOW (Diersch, 02). Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 5

7 After matching the model to the historical groundwater conditions (model calibration) it was applied to forecast the future depth to watertable in 25 as an indicator of future salinity risk. A variety of groundwater management scenarios for the period 05 to 25 was then simulated and the results compared. The groundwater modelling encompassed both Packsaddle Plain and Ivanhoe Plain, extending, generally, from the banks of the Ord River to the bedrock outcrops that surround the irrigation district. The Ord valley aquifer system was represented in model as a phreatic, three-layer system consisting of an upper silt and clay aquifer with small transmissivity; an intermediate sand aquifer with moderate transmissivity; and a lower gravel and sand palaeochannel aquifer with large transmissivity. Future Salinity Risk As a basis for identifying the future salinity risk in the irrigation area, the Ord Stage 1 groundwater model was applied to predict the depth to watertable in 25. Three salinity risk classes were defined and mapped: o o o low salinity risk greater than 4 metres depth to watertable; moderate salinity risk between 2 and 4 metres depth to watertable; and high potential salinity risk less than 2 metres depth to watertable. A low salinity risk was identified beneath the central and southern parts of Ivanhoe Plain and the southeast part of Packsaddle Plain. A high potential salinity risk was identified near Packsaddle Creek; in Martins Location and along sections of the M1 supply channel; and in parts of northern Ivanhoe Plain and Caves Spring Gap. The above high risk class is a measure of potential salinity risk only because depth to watertable, alone, does not determine whether salinity will develop. Groundwater quality and the local direction of groundwater drainage also are important factors. If the watertable is less than two metres below ground surface (high potential risk) but there is net replenishment of groundwater and net downward drainage, then subsurface salt accumulation and salinity formation are unlikely. In this case, the potential for waterlogging may be more of a concern. On the other hand, if subsurface drainage is slow and there is net discharge of groundwater from the watertable, then salinity formation is likely. Predicting whether salinity will develop locally in areas of shallow watertable using a valley-scale modelling tool is unrealistic. Modelling at this large-scale is important for understanding the aquifer water balance and valley-scale watertable behaviour but it cannot incorporate the local hydrogeological conditions and soil characteristics that control subsurface drainage at the local-scale. The approach in this study was to classify areas of shallow watertable as having a high potential salinity risk and recommend targeted monitoring of those areas for emergent salinity. Areas with a high potential salinity risk (< 2 metres depth to watertable) and evidence of emergent salinity require immediate management to lower the watertable and leach the accumulated salt from the soil profile. Areas with low to moderate salinity risk (> 2 metres depth to watertable) should be monitored to ensure that the watertable remains at a safe distance below ground surface and there is no emergent sign of salinity formation. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 6

8 Modelling Scenarios The Ord Stage 1 groundwater model was applied to simulate the effect on groundwater levels of various groundwater management scenarios. The output from each model simulation was the predicted depth to watertable for the period 05 to 25. The simulated groundwater management scenarios included pumping at different rates from the palaeochannel aquifers beneath Packsaddle Plain, northern Ivanhoe Plain and Cave Spring Gap; lowering the surface water level in Lake Kununurra; and installation of subsurface drains in selected areas. Specific model simulations to investigate the effect of tree plantations were not conducted in this study because the differences between trees and other irrigated crops, in terms of their leaching fractions and groundwater uptake rates, are not known. Generally speaking, if trees were effective at reducing the amount of groundwater drainage from the irrigation area, then they would have the same effect as reduced replenishment. A model water balance was produced for each simulation to compare the efficiencies of the different management options. For example, in simulations with groundwater pumping, the proportions of the pumped water contributed from aquifer storage, reduced discharge to the surface environment and additional pumping-induced flow were determined. The synthesis of these results is incorporated into the following groundwater management advice. Conclusions and Recommendations Targeted management of groundwater level and salinity in the Ord Stage 1 is recommended based on the following approach: o o o o reduce groundwater replenishment wherever practical, through improved water use efficiency practices; monitor salinity trends in high potential risk areas to trigger management actions; manage the existing infrastructure to exploit natural subsurface drainage from the irrigation area; and construct new drainage systems as required to prevent salinity formation. The first three options are practical if the existing deep drainage processes are adequate to remove the irrigation excess required for root zone leaching of salts. If natural deep through existing groundwater discharge pathways is inadequate to prevent a shallow watertable, then the subsurface drainage needs to be enhanced artificially. Optimisation of water use efficiency practice to minimise the total amount of groundwater drainage from the irrigation area is a proactive measure that can help to reduce the potential extent of salinity formation and the ultimate cost of constructed drainage. Careful monitoring of the watertable elevation and groundwater salinity is recommended in high salinity risk areas to track the affects on groundwater of changes in land and water management practices. Salinity monitoring should be a key component of future groundwater management because the model-predictions of future salinity risk are based only on valley-scale estimates of depth to watertable. The salinity risk maps produced in this study should be used as a basis for targeted salinity monitoring. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 7

9 Where practicable, it is recommended that the existing irrigation infrastructure be managed to improve the efficiency of groundwater drainage from the irrigation area. Examples include the regular maintenance of selected irrigation drains to enhance groundwater discharge, and lowering the level in Lake Kununurra. Irrigation drains that are important for groundwater drainage should be kept relatively clear of sediment buildup and be cut to their maximum practical depth. If the above measures are insufficient to prevent salinity formation, then it is recommended that new drainage systems be constructed to lower the watertable and provide a pathway for salt transport out of the irrigation area. (e.g., groundwater pumps, additional deep open drains, subsurface drains). Packsaddle Plain Moderate to high potential salinity risk is likely to persist in the Packsaddle irrigation area unless action is taken to control the watertable response to future irrigation and rainfall. The degree of watertable control that can be achieved beneath Packsaddle Plain will be affected by the presence of Lake Kununurra and the water level at which it is operated. It is recommended that Lake Kununurra be lowered to the minimum acceptable water level, and that research be undertaken to establish what that level is. Lowering the lake level would enhance the natural subsurface drainage of groundwater from the Packsaddle aquifer into the lake. It is recommended that the lake level not be raised above its present level. If emergent salinity is apparent over the palaeochannel system, then groundwater pumping from the southwest part of Packsaddle Plain is recommended to lower the watertable. The watertable should be lowered to at least 2 metres below ground surface; however, drawdown of the watertable below the level of Lake Kununurra should be minimised to prevent additional lake water being drawn through the pumping scheme. In this context, the pumping would be more effective and efficient if the lake level also were lowered. Pumping from the northeast part of Packsaddle Plain near Lake Kununurra is not recommended unless it is unavoidable or the lake level was lowered and the pumping did not drawdown the watertable below the lake level. If emergent salinity is apparent over parts of the irrigation area where the palaeochannel system is absent and groundwater drainage is poor, then subsurface drains are recommended to lower the watertable in these areas. Ivanhoe South and Ivanhoe Central Shallow watertables are unlikely to develop in Ivanhoe South and Central in response to future rainfall and irrigation due to the natural rate of subsurface drainage and the local topographic controls. It is recommended that Ivanhoe Central be managed to minimise the net rate of groundwater replenishment from irrigation and rainfall. Although the salinity risk is low throughout this part of the palaeochannel system, inattention to groundwater management would be likely to increase the effort required to manage the watertable further north in the palaeochannel system where shallow watertables are expected to develop (i.e., beneath Ivanhoe North). Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 8

10 Ivanhoe North and Cave Spring Gap Areas of high potential salinity risk are likely to develop in parts of Ivanhoe North and Cave Spring Gap in response to future irrigation and rainfall. Salinity formation is most likely in the high potential risk areas where the palaeochannel is absent and groundwater drainage is poor. The irrigation surface drain network appears to be the main control on groundwater drainage in Ivanhoe North. It is evident that groundwater discharges to the drains when the watertable is raised to above the drain inverts; however, areas with a shallow watertable can still develop where the surface drains are shallow, in the areas in between the drains, and where the palaeochannel is absent and subsurface drainage is poor. Groundwater in Cave Spring Gap appears to drain less effectively than it does in Ivanhoe North because the main irrigation drains in Cave Spring Gap run along the flanks of the valley where the palaeochannel pinches out on bedrock and the aquifer is less transmissive. Groundwater pumping is recommended to balance the excess groundwater replenishment if emergent salinity is apparent over the palaeochannel system. The objective of pumping should be to lower the watertable to at least 2 metres below ground surface, while minimising watertable drawdown below the drain levels. Unnecessarily lowering the watertable below the drains would increase the cost of pumping by drawing water through the pumping scheme that would have otherwise discharged effectively to the drains. The disadvantage of pumping from the palaeochannel aquifer to control local salinity formation is that multiple pumps and large pumping rates would be required to drawdown the watertable over a much larger area than required, including areas that are not at risk from salinity formation. It is recommended that the feasibility of deepening selected surface drains and constructing additional deep open drains be investigated. In areas where the palaeochannel system is absent and the aquifer does not drain effectively in response to pumping, subsurface drains are recommended to lower the watertable and reverse emergent salinity. Martins Location and M1 Supply Channel Most of Martins Location and a narrow strip along the M1 supply channel to the south of Martins Location are likely to develop a high potential salinity risk in response to future irrigation and rainfall. A small area within the southeast margin of the irrigation area in Martins Location already has developed secondary salinity and there is evidence of lost productivity near the M1 supply channel. Martins Location is unlikely to be drained effectively using groundwater pumps because the palaeochannel system is absent and natural subsurface drainage is poor. The aquifer has small transmissivity and produces only low pumping yields. Removing excess groundwater from the aquifer is much more difficult in these areas and it is recommended that improved water use efficiency practices and good surface drainage design be implemented to minimise the net rate of groundwater replenishment from irrigation and rainfall. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 9

11 Subsurface drains are recommended to lower the local watertable and remove subsurface salt in areas with emergent salinity. Tree crops also may be effective at reducing the watertable elevation beneath the trees; however, this approach may not be sustainable in the long-term due to subsurface accumulation of salt. It is recommended that a subsurface drainage trial be conducted in Martins location or at a comparable site to determine the feasibility, effectiveness and design parameters for subsurface drains in the Ord Stage 1. The roles of irrigated and dryland trees in managing the aquifer water balance also should be investigated. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 10

12

13 Contents Copyright and Disclaimer...1 Acknowledgements...3 Executive Summary...4 Contents Introduction Objectives Study Area Project Context Aquifer Water Balance Concepts in the Ord Stage Groundwater Replenishment, Discharge and Storage 2.2. Where Does the Irrigation System Leak? Local and Diffuse Groundwater Replenishment Leakage from Lake Kununurra Leaky Area on Ivanhoe North 2.3. Where Does the Aquifer Discharge? Ord Stage 1 Groundwater Flow Model Overview of the Model 3.2. Model Water Balance Previous Groundwater Modelling 4. Groundwater Management Areas Packsaddle Palaeochannel System Influence of Lake Kununurra Packsaddle Plain Ivanhoe South 4.2. Ivanhoe Palaeochannel System Ivanhoe Central, Ivanhoe North and Cave Spring Gap 4.3. Martins Location 4.4. Green Location M1 Supply Channel 5. Assessment of Salinity Prevention Options Techniques for Managing the Watertable Improved Water Use Efficiency Pumping Wells Drains Lowered Water Level in Lake Kununurra Trees Salinity Risk Areas Observed and Simulated Depth to Watertable Future Watertable Scenarios without Management Salinity Management Scenarios Packsaddle Plain Ivanhoe North and Cave Spring Gap Martins Location Conclusions and Recommendations Packsaddle Plain Ivanhoe South and Ivanhoe Central Ivanhoe North and Cave Spring Gap Martins Location and M1 Supply Channel 78 APPENDIX A : Hydrogeology of the Ord Stage APPENDIX B : Hydrological Timeline B1. Rainfall APPENDIX C : Ord Stage 1 Groundwater Flow Model C1. Modelling Environment C2. Model Layers C3. Finite Element Mesh C4. Aquifer Properties Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 12

14 APPENDIX D : Pre-Irrigation Groundwater Conditions D1. Pre-irrigation Groundwater Levels D2. Simulated Watertable D3. Boundary Conditions APPENDIX E : Post-Irrigation Groundwater Conditions E1. Observed Groundwater Rise E2. Simulated Groundwater Rise E3. Simulated Water Balance E4. Initial Conditions E5. Boundary Conditions APPENDIX F : Recent Groundwater Conditions... 1 F1. Increased Responsiveness to Rainfall and Irrigation APPENDIX G : Wet-Season-Dry-Season Hydrograph Separation... 1 APPENDIX H : Calibration Bore Locations... 1 APPENDIX I : Calibration Bore Hydrographs... 1 I1. Packsaddle Plain I2. Packsaddle Creek... 1 I3. Ivanhoe South I4. Ivanhoe Plain Central I5. Ivanhoe North I6. Cave Spring Gap I7. Martins Location I8. M1 Supply Channel I9. Green Location I10. Ivanhoe West References Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 13

15 Figures Figure 1-1: Annual wet season rainfall at Kununurra Airport... Figure 1-2: Location map Figure 2-1: Schematic of the Ord Stage 1 groundwater balance under (a) preirrigation conditions, (b) post-irrigation conditions and (c) managed conditions... Figure 3-1: Ord Stage 1 model domain... Figure 3-2: Simulated aquifer water balance 1975 to 03 from the post-irrigation model calibration... Figure 4-1: Groundwater management areas Figure 4-2: Example hydrographs for monitoring bores P2 and P9 on Packsaddle Plain, and 1F and 1E on Ivanhoe South... Figure 4-3: Secondary salinity outside of the irrigation area adjacent to Packsaddle Creek... Figure 4-4: Example hydrographs from monitoring bores PN9S and 2B on Ivanhoe Central, 4C and 5F on Ivanhoe North, and CS3 and ORD6 in Cave Spring Gap Figure 4-5: A groundwater seep on the river bank directly south of the Frank Wise Research Station... Figure 4-6: Example hydrographs from monitoring bores 11C and 96/06 in Martins Location Figure 4-7: Secondary salinity in Martins Location Figure 4-8: Example hydrographs from monitoring bores 6B and PN12S in Green Location... Figure 4-9: Example hydrographs from monitoring bores 3H, 3K and PN14S near the M1 supply channel... Figure 4-10: Stunted and failed mangoes trees in an area of shallow watertable adjacent to the M1 supply channel Figure 5-1: Estimated depth to watertable in October Figure 5-2: Model calibrated depth to watertable in Figure 5-3: Model calibrated depth to watertable in Figure 5-4: Replenishment scenario W0; model-simulated depth to watertable in Figure 5-5: Replenishment scenario W2; model-simulated depth to watertable in Figure 5-6: Watertable rise above Lake Kununurra in October 02; observed groundwater levels Figure 5-7: Scenario P3; model-simulated depth to watertable in Figure 5-8: Scenario P4; model-simulated depth to watertable in Figure 5-9: Scenario P5; model-simulated depth to watertable in Figure 5-10: Scenario P6; model-simulated depth to watertable in Figure 5-11: Scenario P7; model-simulated depth to watertable in Figure 5-12: Level in Lake Kununurra raised by 1 meter; model-simulated depth to watertable in Figure 5-13: Scenario P8; model-simulated depth to watertable in Figure 5-14: Scenario I1 - simulated depth to watertable in Figure 5-15: Scenario I2 - simulated depth to watertable in Figure 5-: Scenario I3 - simulated depth to watertable in Figure 5-17: Scenario M1; model-simulated depth to watertable in Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 14

16 Tables Table 3-1: Specifications and numerical settings for the Ord Stage 1 groundwater flow model Table 5-1: Summary of groundwater replenishment scenarios Table 5-2: Scenario W0; water balance Table 5-3: Scenario W2: water balance Table 5-4: Summary of groundwater management scenarios; Packsaddle Plain Table 5-5: Scenario P3; water balance Table 5-6: Scenario P4; water balance Table 5-7: Scenario P5; water balance Table 5-8: Scenario P6; water balance Table 5-9: Scenario P7; water balance Table 5-10: Scenario P8; water balance Table 5-11: Summary of groundwater management scenarios; Ivanhoe Plain Table 5-12: Scenario I1; water balance Table 5-13: Scenario I2; water balance Table 5-14: Scenario I3; water balance Table 5-15: Summary of groundwater management scenarios; Martins Location Table 6-1: Summary of groundwater management recommendations Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 15

17 Appendix Figures Appendix Figure B-1: Hydrological timeline for the Ord Stage Appendix Figure C-1: Solid volume of the Ord Stage 1 groundwater flow model Appendix Figure C-2: Finite element mesh in the region of Lake Kununurra Appendix Figure C-3: Detail of the finite element mesh in the region of (a) the M1 offtake and (b) Packsaddle Creek Appendix Figure C-4: Summary of pump test locations and results Appendix Figure C-5: Model layer Appendix Figure C-6: Model layer Appendix Figure C-7: Model layer Appendix Figure D-1: Estimated and simulated groundwater levels in palaeochannel monitoring bores under pre-irrigation conditions Appendix Figure E-1: Groundwater replenishment zones... 1 Appendix Figure E-2: Simulated water level in Lake Kununurra Appendix Figure E-3: Simulated watertable rise above Lake Kununurra; red areas... 1 Appendix Figure E-4: Simulated groundwater discharge to the surface environment; red areas... 1 Appendix Figure F-1: Water level readings beneath northern Ivanhoe Plain from 1984 to Appendix Figure F-2: Wet-season-dry-season hydrograph separation for monitoring bore PB3 (1984 to 02)... 1 Appendix Figure H-1: Calibration bores for the pre-irrigation model Appendix Figure H-2: Calibration bores for the post-irrigation model Ivanhoe Plain... 1 Appendix Figure H-3: Calibration bores for the post-irrigation model Packsaddle Plain Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page

18 Appendix Tables Appendix Table B-1: Average annual wet season rainfall at Kununurra (Station 56) Appendix Table C-1: Specifications and numerical settings for the Ord Stage 1 groundwater flow model Appendix Table C-2: Rules for simulating absent model layers Appendix Table C-3: Summary of pump test results Appendix Table C-4: Aquifer parameter values from the SKM (1998b) model of Ivanhoe Plain Appendix Table C-5: Aquifer parameter values from the current study Appendix Table D-1: Estimated pre-irrigation groundwater level in the palaeochannel aquifer beneath Ivanhoe Plain prior to irrigation development in Appendix Table E-1: Simulated water balance in Appendix Table E-2: Simulated water balance in Appendix Table E-3: Simulated water balance in Appendix Table E-4: Simulated water balance in Appendix Table E-5: Simulated water balance in Appendix Table E-6: Calibrated groundwater replenishment rates Appendix Table E-7: Calibrated transfer rates for the Ord River Appendix Table E-8: Calibrated transfer rates for the M1 and SP1 supply channels Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 17

19 1. Introduction The depth to watertable below ground surface and climate are important aspect of salinity risk and soil salinity formation. Continual evaporation of shallow groundwater without subsurface leaching can result in harmful accumulation of salt in the overlying soil and the loss of soil productivity. In the Ord Irrigation Scheme Stage 1 (Ord Stage 1) annual pan evaporation exceeds annual rainfall by a factor of approximately three and salinity can develop if the depth to watertable is less than approximately 2 metres for an extended period without leaching (Ali and Salama, 03). Other factors such as groundwater quality and soil structure also are important. Figure 1-1: Annual wet season rainfall at Kununurra Airport The Ord region experiences a wet-dry tropical climate that is characterised by well-defined wet and dry seasons. Approximately ninety percent (90%) of the annual precipitation falls during the wet season between mid-november and March. The average wet season rainfall at Kununurra Airport from to was 783 mm/year (Figure 1-1). The annual rainfall amount varies significantly between years. Maximum annual rainfall was 1476 mm in and minimum annual rainfall was 6 mm in Average annual pan evaporation is around 2760 mm/year, which is greater than three times the annual average rainfall. Air temperatures are high to very high throughout the year. July is the coolest month, with mean daily maxima and minima of 31 C and 14 C, respectively. November is the hottest month, with mean daily maxima and minima of 39 C and 25 C.These temperature ranges are suitable for growing crops all year round but the movement of farm vehicles and machinery is restricted by boggy conditions during the wet season. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page

20 Groundwater rise began in the early 1960s due to land clearing and the introduction of irrigated agriculture (APPENDIX E). Infiltration of irrigation water and rainfall during the past forty years has increased the groundwater replenishment rate to above the rate that groundwater can naturally drain from the aquifer. To date, the rising watertable has had only a minor impact on agricultural production because the groundwater has been too deep for widespread watertable evaporation and salinity formation. More recently, the watertable has reached to within several metres of the ground surface and several small areas of salinity have developed around the margins of the irrigation district. To support a sustainable future for agricultural production in the Ord Stage 1, it is critical to establish whether soil salinity is likely to become more widespread in future, which areas are at risk, and what can be done to prevent it Objectives The purpose of this study was to identify a groundwater management strategy to control watertable rise and prevent salinity development in the Ord Stage 1. The scope of work involved the following tasks: a review of the available groundwater information; construction and calibration of a groundwater flow model and subsurface water balance of the Ord Stage 1; identification of the factors that contribute to and control groundwater rise within the irrigation area; delineation of groundwater management zones and salinity risk areas; evaluation of various groundwater management options; and communication of the results, conclusions and recommendations to stakeholders. This report provides a full account of the project methods and outputs. The main body of the report is an overview of the study and includes the key findings and recommendations. Technical detail is contained in the report appendices, which are cross-referenced throughout the main report Study Area The Ord Stage 1 Irrigation Area is located along the banks of the Ord River at Kununurra in the northeast of Western Australia. The Kununurra town site is located approximately thirty kilometres from the State border between Western Australian and Northern Territory. A location map of the irrigation area is presented as Figure Project Context This project was one of fourteen linked sub-projects of the National Action Plan for Salinity and Water Quality (NAP) Project 03: Improved Water Management in the Stage 1 Ord River Irrigation Area. The work described in this report was conducted in association with the linked sub-projects Dewatering Bores (Smith et al., 05) and Whole-System and On-Farm Economic Analysis of Water Use Efficiency Options (Ali et al., in review). Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 19

21 Regional context for Natural Resource Management (NRM) in the Ord catchment is provided in the following documents: Draft for Comment: Kimberley Natural Resource Management Plan, Kimberley Development Commission, 04; A Strategy for Managing the Natural Resources of Western Australia s Rangelands: A Draft for Public Comment and Review, Rangelands NRM Co-ordinating Group, 05; National Action Plan for Salinity and Water Quality Ord Catchment Priority Projects 05, Ord Reference Group, 05. The principal NRM Resource Condition Targets (RCTs) for the Ord catchment include: ORCT1 No new occurrences of salinity in the ORIA by 14. ORCT2 A reduction in the area of salinity affected land in the ORIA by 14. ORCT4 Groundwater levels in the ORIA to be maintained at/or below two metres by 15. ORCT8 Groundwater levels and quality at agreed reference bores meet levels set for the Ord Catchment by 10 with benchmarks set by 06. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page

22 Figure 1-2: Location map Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 21

23 2. Aquifer Water Balance Concepts in the Ord Stage Groundwater Replenishment, Discharge and Storage The aquifer water balance in the Ord Stage 1 can be expressed by the following simple relation: ΔS = R - D where ΔS is change in groundwater storage (i.e., change in groundwater level); R is groundwater replenishment; and D is groundwater discharge. An imbalance between groundwater replenishment and groundwater discharge leads to a change in groundwater level. If the total replenishment is greater than the total discharge (R > D), then the amount of groundwater stored in the aquifer will increase and the watertable will rise. This rise will continue until the rate of groundwater discharge adjusts to re-balance the rate of replenishment. If, on the other hand, the replenishment rate is less than the discharge rate (R < D), then the watertable will fall. This simple water balance concept applied to the Ord Stage 1 is illustrated schematically in Figure 2-1. R = D R > D R = D 15- m Rising Stable >2 m Stable (a) Pre-irrigation (b) Post-irrigation (c) Managed Figure 2-1: Schematic of the Ord Stage 1 groundwater balance under (a) pre-irrigation conditions, (b) post-irrigation conditions and (c) managed conditions In reality, watertables are in constant motion because groundwater replenishment and discharge processes both vary continuously in time. Depth to watertable below ground surface fluctuates according to local-scale and regional-scale hydrological rhythms (e.g., diurnal evaporation, and seasonal and synoptic rainfall patterns). A longer-term trend in the direction of watertable movement is the result of the averaging of those fluctuations. The watertable beneath the Ord Stage 1 has risen in the order of 15 to metres since the early 1960s, which represents around 0.3 to 0.5 metres rise per year. Aquifers in irrigation areas can be replenished through infiltration of rainwater, surface water and irrigation water; and by subsurface inflow from connected aquifers. Groundwater can discharge from the aquifer as seepage to the surface environment (e.g., creeks and ground surface); through uptake by phreatophytic vegetation; and by discharge to drainage systems (e.g., groundwater pumps and open drains). In the Ord Stage 1, changes in rainfall, irrigation management, cropping and land use practices will all affect the aquifer water balance to some degree and cause the groundwater level to vary. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page

24 A relatively stable groundwater level indicates that, on average, the rates of groundwater replenishment and discharge are approximately equal (Figure 2-1). A dynamic balance between groundwater replenishment and discharge is typical in undisturbed aquifer systems that are in dynamic equilibrium with climate and vegetation. Land developments for industry, agriculture and urbanisation alter the local hydrological balance and generally lead to changed groundwater levels Where Does the Irrigation System Leak? Historical groundwater rise beneath the Ord Stage 1 is direct evidence that agricultural development of the valley has increased the rate of groundwater replenishment (APPENDIX E). Vegetation clearing, the formation of Lake Kununurra, leakage of irrigation water from unlined supply channels and drains, and deep drainage beneath flood-irrigated fields have increased the groundwater replenishment rate to above the natural rate of groundwater drainage. It is expected that these effects have compounded over time as the thickness of the unsaturated zone above the watertable has progressively reduced (see APPENDIX F, Section F1) Local and Diffuse Groundwater Replenishment Salama et al. (02a) confirmed that the chemical composition of current day groundwater is a mixture of older native groundwater and the more recent surface water that has filled the aquifer. Nevertheless, the relative contributions of surface waters from local sources (e.g., irrigation channels) and diffuse sources (e.g., flood irrigation and rainfall) have not been determined accurately. There are a number of reasons that make this task difficult: The chemical compositions of surface water in supply channels and drains, and the composition of rainwater are very similar; therefore, natural tracers cannot be used effectively to identify the proportions of surface waters from different sources that have filled the aquifer. In addition, the chemical composition of native groundwater and its variability within the irrigation area are not known accurately because samples were not collected before mixing with surface water had occurred. The Ord Stage 1 is a flow-through scheme wherein surface water moves simultaneously through the irrigation supply channels, crop furrows, tail water drains and the main irrigation drains. It is problematic to investigate the individual effects on groundwater of these local and diffuse replenishment sources because the replenishment from them occurs simultaneously. During the wet season, when the irrigation conveyance system is not used for irrigation purposes, the supply channels and drains can still fill with water following rainfall runoff events, and when the channel system is used for flood relief. Historical records of these flows and levels do not exist. Estimates of surface water infiltration rates from channels and drains based on hydraulic calculations and models (e.g. Barr et al., 03) are inherently uncertain because the parameters that determine the leakage rates in these calculations are known only to an order-of-magnitude accuracy. Estimated leakage rates from supply channels have not been corroborated by field measurement programs and are subject to large uncertainty. It is likely that the infiltration rates from channels have varied significantly over time in response to channel silting and regular channel maintenance, which is carried out to clear weeds and maintain the channel profiles. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 23

25 The primary gravel and sand aquifer beneath the Ord Stage 1 has very large transmissivity that results in small watertable gradients and indistinct local water level responses to groundwater replenishment. For example, recognisable watertable mounds generally do not form beneath the supply channels, except in areas where the primary aquifer is absent and the local subsurface drainage is poor. Without distinctive watertable responses to groundwater replenishment it is difficult to infer the rates of leakage from supply channels and drains based only on hydraulic considerations. Previous work to determine the amount of leakage from the irrigation channel network was motivated mainly by a desire to reduced leakage through the use impermeable liners. It is now apparent that the watertable in some areas has risen to above the base of the irrigation drain network, which has become important for groundwater discharge and drainage. For the above reasons, local and diffuse replenishment are not generally differentiated in the groundwater modelling in this study; however, the role of the irrigation surface drains in removing groundwater from the irrigation area is a critical element Leakage from Lake Kununurra The formation of Lake Kununurra resulted in a large and distinctive watertable rise in the aquifers beneath Packsaddle Plain and Ivanhoe South. Because the aquifer forcing and response are both known in this case, it has enabled more reliable calculations of leakage from Lake Kununurra, and the variation in the leakage over time. Estimates from the groundwater modelling in this study are presented in Section 3.2. and APPENDIX E Section E Leaky Area on Ivanhoe North The existence of a leaky area to the west of the M1 supply channel on Ivanhoe North was apparent during the calibration of the Ord Stage 1 groundwater model (Section 3). The extent of this area, as it was represented in the modelling, is indicated as Zone 6 on Appendix Figure E-1. Hydrographs for the monitoring the bores located within Zone 6 (e.g., 5F, PN2S, PN2D, PN6S, PN5D, 5E and PB1) are presented in APPENDIX I Section I5. Overall, the rate of groundwater rise in the area was noticeably larger than it was in the adjacent replenishment zones on Ivanhoe North. Notable increases in the rate of groundwater rise occurred in Zone 6 from the mid-1970s to the early-1980s. This appears to be a response to rice paddies that were cultivated on Ivanhoe Plain from 1974 to 1982 (see APPENDIX B). Laws and George (1982) reported that rates of groundwater rise under rice crops averaged up to 1.3 metres per year. Banyard (1983) also mentioned that the maintenance of water levels in the rice paddies was difficult due to high surface water infiltration rates. He reported measured infiltration values of 0.43 and 1.0 metre per day in the surface drains D6A and D4F, respectively. Groundwater salinity data compiled by Smith et al. (05) indicated consistently low values of groundwater electrical conductivity (EC) around monitoring bores PB1 and 5E within Zone 6. The existence of fresher groundwater in the area also suggests that the rate of groundwater replenishment from irrigation may be larger than in the adjacent areas where the groundwater EC is larger. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page

26 2.3. Where Does the Aquifer Discharge? Groundwater discharge from the irrigation area occurs through seepage to natural surface drains (e.g., the Ord River and Packsaddle Creek), evaporation from areas of shallow watertable, transpiration by deep-rooted plants (e.g., tree crops) and, more recently, seepage to the irrigation drain network. Significant hydraulic connections between the Ord River and the groundwater systems beneath Packsaddle Plain and Ivanhoe South are apparent. Further north in Ivanhoe Central and North, there appears to be only limited interaction between groundwater and the Ord River. Subsurface groundwater drainage through Cave Spring Gap toward Weaber Plain appears to be relatively small and suggests that the primary gravel and sand aquifer may be discontinuous in this area. There is no geological information to support this assumption; however, a discontinuity in the palaeochannel system is indicated by the steady rate of groundwater rise throughout Cave Spring Gap in comparison with the apparently un-related watertable variation in the monitoring bores on Weaber Plain. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 25

27 3. Ord Stage 1 Groundwater Flow Model A groundwater flow model of the Ord Stage 1 was constructed as a tool to help evaluate the feasibility and effectiveness of various groundwater management options. It predicts the groundwater elevation and depth to watertable at the valley-scale but does not simulate the salt balance or salt concentration in groundwater. The process of constructing and calibrating the groundwater model was valuable because it required the conceptual hydrogeology to be tested within a quantitative and physically based framework. In turn, this provided the opportunity and rationale for improving the conceptual hydrogeology Overview of the Model The geographic extent of the Ord Stage 1 groundwater model is indicated on Figure 3-1. The model domain encompasses both Packsaddle Plain and Ivanhoe Plain and extends, generally, from the banks of the Ord River to the bedrock outcrops that surround the irrigation area (Figure 3-1). The Ord Stage 1 aquifer system was simplified to a phreatic, three-layer model consisting of: Layer 1 an upper silt and clay aquifer with small permeability; Layer 2 an intermediate sand aquifer with moderate permeability; and Layer 3 a lower gravel and sand palaeochannel aquifer with large permeability. The basement bedrock profile was represented by the bottom surface of Layer 3, which was assumed to be impermeable. The model was developed and implemented in the subsurface flow and transport simulation system FEFLOW (Diersch, 02). A summary of the model specifications and numerical settings is presented in Table 3-1. Technical details of the model development and calibration are provided in the following appendices: APPENDIX A overview of the Ord Stage 1 hydrogeology; APPENDIX B timeline of hydrological events in the Ord Stage 1; APPENDIX C detailed description of the Ord Stage 1 groundwater model; APPENDIX D calibration of the Ord Stage 1 groundwater model to pre-irrigation conditions; and APPENDIX E calibration of the Ord Stage 1 groundwater model to post-irrigation conditions. The information in these appendices is critical to the conclusions and recommendations of this study; however, the appendices do not need to be read to understand this report. They are provided as a source of further information and as a complete documentation of the modelling work. Relevant detail is discussed in the main report and referenced to the appendices as required. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page

28 Table 3-1: Specifications and numerical settings for the Ord Stage 1 groundwater flow model Simulation system Type of model Type of aquifer Model layers Type of elements Number of elements Number of nodes Layer 1 Layer 2 Layer 3 Equation solver Time stepping Free surface top Free surface bottom FEFLOW 3-D, saturated flow Unconfined with moveable phreatic surface Three layers, four slices Six-node triangular prisms 1, 835 per layer (4,505 total) 68,2 per 1 slice (272,880 total) Clay and silt (small permeability) Sand aquifer (moderate permeability) Sand and gravel aquifer (large permeability) Iterative Predictor-corrector AB/TR time integration scheme (max. time step size 100 days) Constrained at top slice if touching the top slice Unconstrained (watertable) if fallen dry at the bottom slice 1 The three model layers are defined by four slices, which are the top and bottom surfaces of the layers Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 27

29 Figure 3-1: Ord Stage 1 model domain Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page

30 3.2. Model Water Balance The Ord Stage 1 groundwater model provides a sophisticated water balance tool for the irrigation area. In use, three water balance terms from the model are of primary interest: Groundwater replenishment water that enters the aquifer system from external sources (e.g., channel leakage and percolation of rainfall); Groundwater discharge water that exit the aquifer system to external sinks (e.g., seepage to irrigation drains and outflow through Cave Spring Gap); and Aquifer storage the volume of groundwater accumulated within the aquifer. If, in a given period of time, the total amount of groundwater replenishment is larger than the total amount of groundwater discharge, then the aquifer storage will increase and the average watertable elevation will rise during that period and vice versa. Figure 3-2 depicts the calibrated model water balance for the period 1975 to 03, including the cumulative groundwater storage. This storage term represents the difference between the amount of groundwater stored in the aquifer in 1975 compared to the amount stored at a later date. The following features are evident in Figure 3-2: Cumulative groundwater storage increased steadily up until 01 due to the imbalance between net groundwater replenishment from rainfall and irrigation and net groundwater discharge; Groundwater storage decreased for the first time during 02 and 03 in response to decreased replenishment from rainfall, and increased groundwater discharge to the irrigation drain network; Inflow from Lake Kununurra to the Packsaddle aquifer was reduced to a relatively small component of the water balance by the late 1970s by this time the watertable beneath Packsaddle Plain had risen considerably and the hydraulic gradient between the lake and aquifer had diminished; Inflow from Lake Kununurra to the Ivanhoe aquifer was approximately balanced by groundwater discharge from the aquifer to the Ord River; and Leakage from the SP1 and M1 supply channels, and outflow through Cave Spring Gap were relatively minor components of the model water balance. Overall, the model water balance is consistent with the view that the storage capacities of the Packsaddle and Ivanhoe aquifers within the irrigation area were largely exhausted by the mid-1990s. It is expected that the cumulative groundwater storage and watertable elevation will stabilise around their current values but will fluctuate in response to variations in rainfall, irrigation and future groundwater management. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 29

31 0 Water Flux (GL/yr) Inflow from the SP1 Inflow from the M1 Inflow to Pack. from L.Kun. Inflow to Ivan. from L.Kun. Inflow to Ivan. from the Ord River Inflow from Caves Spring Gap Inflow from creek, drains & ground surface Recharge from irrigation & rainfall Storage (imbalance) Cumulative Storage from 1975 (GL) Figure 3-2: Simulated aquifer water balance 1975 to 03 from the post-irrigation model calibration 3.3. Previous Groundwater Modelling Groundwater flow modelling of Ivanhoe Plain (SKM, 1998b) was carried out in 1997; however, Packsaddle Plain has not previously been modelled. The earlier modelling of Ivanhoe Plain pre-dates the Water and Rivers Commission s review of the Ord Stage 1 hydrogeology, including the production of a new hydrogeology map and hydrogeological cross sections (O Boy et al., 01). Due to the limited information available at the time, the previous groundwater flow model was based on relatively approximate representations of the hydrogeology and aquifer properties, and was considered unsuitable for use in the current study. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page

32

33 4. Groundwater Management Areas The extent and distribution of the gravel and sand palaeochannel aquifer within the Ord Stage 1 is the primary control on lateral groundwater movement and subsurface drainage beneath the irrigation area. Areas that overlie the palaeochannel require coordinated management because they are hydraulically connected through the subsurface by the highlypermeable gravel and sand sediments. These areas can be recognised from the distributions of gravel and sand on the Ord River Irrigation Area Hydrogeology map (Boy et al., 01) and by grouping together the bore locations that exhibit similar hydrograph response characteristics. In contrast, the rate of groundwater drainage from areas where the palaeochannel aquifer is absent is restricted by the reduced transmissivity. The hydrographs of bores in these areas exhibit distinctive local responses. The groundwater management areas identified in Figure 4-1 were delineated based on a review of the existing hydrogeological information and the analysis of approximately one hundred bore hydrographs (APPENDIX I). A more detailed description of each management zone is presented in the following sections. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page

34 Figure 4-1: Groundwater management areas Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 33

35 4.1. Packsaddle Palaeochannel System Influence of Lake Kununurra The water level in Lake Kununurra principally controls the groundwater level beneath Packsaddle Plain and Ivanhoe South. In effect, these areas have become subsurface extensions of Lake Kununurra. There also is an additional effect on the watertable due to groundwater replenishment by irrigation and rainfall. Groundwater rise in these areas began in the early 1960s following the completion of the Kununurra Diversion Dam (APPENDIX B). The timing of groundwater rise in bores on Packsaddle Plain and Ivanhoe South depended on a bore s distance from the lake and the time taken for lake water to flow through the subsurface to that location. By 1974, when irrigated agriculture was commenced on Packsaddle Plain, the aquifers had already substantially filled in response to leakage from the lake. Example hydrographs for monitoring bores P2 and P9 on Packsaddle Plain, and monitoring bores 1F and 1E on Ivanhoe South are shown in Figure 4-2. Additional bore hydrographs and the bore location are presented in APPENDIX I and APPENDIX H, respectively. Water Level (m AHD) Water Level (m AHD) Observed Model calibration L. Kun. Start of Packsaddle irrigation P2 Model basement Model M1 level 1F Observed Model calibrated Water Level (m AHD) Water Level (m AHD) Model DP1 level L. Kun. Start of Packsaddle irrigation P9 Model basement Observed Model calibration Observed Model calibrated Figure 4-2: Example hydrographs for monitoring bores P2 and P9 on Packsaddle Plain, and 1F and 1E on Ivanhoe South L. Kun. 1E Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page

36 Each hydrograph exhibits an initial period of rapid groundwater rise as water from Lake Kununurra reaches the bore location. A subsequent levelling-out period follows as the groundwater level reaches a new equilibrium with the lake level. The equilibrium water levels in bores P2 and 1F were just below the lake level because the bores are located close to the lake shoreline. In comparison, the equilibrium water level in bore 1E on Ivanhoe South was 4 to 5 metres below the lake level because it is mid-way between the lake and the Ord River, which is a groundwater discharge area approximately 10 metres below the lake level. At bore P9 on Packsaddle Plain, there was additional watertable rise of order 3 to 4 metres from around 1974 following the start of irrigation on Packsaddle Plain. The model predicts the watertable to be above the lake level at this location by 00, which would reverse the hydraulic gradient and cause groundwater from irrigation replenishment to drain back into the lake. Prior to 1972 when the Ord River Dam was completed up-stream of Kununurra and Lake Argyle was formed Lake Kununurra was filled periodically by seasonal river flows and was below full capacity for around eight months per year (APPENDIX B). The lake water level was often more than three metres below full capacity during these times and this variability is reflected in the bore hydrographs prior to After the formation of Lake Argyle, there was sufficient up-stream storage to maintain Lake Kununurra at close to full capacity, with only minor operational adjustments to the lake level Packsaddle Plain Packsaddle Plain is bordered to the north and east by Lake Kununurra, to the north by the Dunham River, and to the south and west by the Carr Boyd Ranges. Surface water that flows into the Packsaddle aquifer from Lake Kununurra is transmitted efficiently through the palaeochannel gravels and sands and discharges eventually to Packsaddle Creek, which lies below the lake level. The groundwater level throughout Packsaddle Plain appears to be in approximate equilibrium with the current lake level but fluctuates and develops watertable mounds in response to irrigation and rainfall. The watertable beneath a significant part of Packsaddle Plain is now mounded to above the lake level and groundwater from these areas is draining back into Lake Kununurra. The lake therefore can be both a source and sink for groundwater dependent on the elevation difference between the watertable and lake level. In either case, the lake acts to buffer groundwater level change beneath Packsaddle Plain; re-filling the aquifer if the watertable falls below lake level and emptying it if it rises above. A map of watertable rise above lake level in October 02 is presented later in the report (see Section and Figure 5-6). Assuming that Lake Kununurra will remain at its current level, and that the future watertable elevation will fluctuate somewhere around levels, the areas of Packsaddle Plain with the highest salinity risk are adjacent to Packsaddle Creek and near drain DP3. Secondary salinity outside of the irrigation area near Packsaddle Creek and DP3 (Figure 4-3) is currently monitored by the Department of Agriculture. Shallow watertables occur along the shoreline of Lake Kununurra; however, there are no known reports of salinity development in these areas. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 35

37 05 CSIRO Figure 4-3: Secondary salinity outside of the irrigation area adjacent to Packsaddle Creek Ivanhoe South This area is bordered to the south by Lake Kununurra, to the west by the Ord River and to the east by outcropping bedrock. A relatively steep groundwater gradient exists between Lake Kununurra and the Ord River directly down stream of the Diversion Dam. Lake water that flows into the aquifer beneath Ivanhoe Plain is transmitted efficiently through the palaeochannel gravels and sands and discharges to the Ord River. The uniform watertable gradient in this area is the main evidence of a strong groundwater connection between the lake and river. Seepage faces and groundwater springs are not observed along the river bank in Ivanhoe South (D. Palmer, DoE, 05, pers. comm., 23 June), indicating that groundwater discharges directly to the river across the riverbed. Groundwater replenishment from rainfall and irrigation has a comparably small affect on the groundwater level beneath Ivanhoe South. The exception is where the palaeochannel pinches out along the northeast flank of the valley near the M1 supply channel. In addition, there is indirect evidence that the palaeochannel aquifer is discontinuous between Ivanhoe South and Ivanhoe Central, as indicated by the absence of a distinct groundwater response beneath Ivanhoe Central following the formation of Lake Kununurra Ivanhoe Palaeochannel System Ivanhoe Central, Ivanhoe North and Cave Spring Gap Large parts of Ivanhoe Plain and Cave Spring Gap are connected through relatively continuous palaeochannel sediments. These sediments are highly conductive; however, the palaeochannel system lacks well-developed discharge boundaries and the aquifer has steadily filled since the early 1960s with little evidence that the mounting hydraulic head has activated new major discharge areas. In particular, drainage to the Ord River is apparently restricted. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page

38 Example hydrographs for bores on Ivanhoe Central, Ivanhoe North and Cave Spring Gap are presented in Figure 4-4 (see also APPENDIX I). In general, they demonstrate a remarkably consistent rising trend over a large area and for a long period of time. The similarity between hydrographs indicates that the palaeochannel has behaved like a large subsurface reservoir that filled by vertical infiltration of rainfall and irrigation water but drained only slowly by groundwater outflow through Cave Spring Gap and minor discharge to the Ord River. Water Level (m AHD) Model D2 level PN9S & 2B 2B Observed PN9S Observed PN9S Model calibration Water Level (m AHD) Model D2C level 4C Observed Model calibration Water Level (m AHD) 14 Model D4 level 5F Observed Model calibration Water Level (m AHD) Model D8 level CS3 & ORD6 ORD6 Observed CS3 Observed CS3 Model calibration Figure 4-4: Example hydrographs from monitoring bores PN9S and 2B on Ivanhoe Central, 4C and 5F on Ivanhoe North, and CS3 and ORD6 in Cave Spring Gap Throughout the central and northern parts of Ivanhoe Plain, the palaeochannel aquifer is poorly connected to the Ord River. Steep watertable gradients and elevated groundwater levels next to the river indicate the presence of less permeable sediments that restrict lateral groundwater drainage into the river. A recent survey of the river bank within the irrigation area by the Department of Environment (D. Palmer, DoE, 05, pers. comm., 23 June) established the presence of a limited number of local groundwater springs and seeps located directly south of the Frank Wise Research Station (Figure 4-5). These seeps are consistent with an elevated groundwater level in the river bank and restricted groundwater drainage to the river through local connections. Northward of the Frank Wise Research Station, the palaeochannel alignment deviates northeast through Cave Spring Gap, while the present course of the Ord River continues northward, cutting through outcropping bedrock, and then turns west at the northern end of the irrigation area. Due to this divergence between the palaeochannel alignment and river course, there is no significant groundwater drainage to the river from the aquifer beneath Ivanhoe North. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 37

39 05 Department of Environment Figure 4-5: A groundwater seep on the river bank directly south of the Frank Wise Research Station More recently, the watertable beneath a large part of Ivanhoe North rose to above the base of irrigation drains, which resulted in a subsurface flow gradient toward the drains. The bore hydrographs in this area indicate that the watertable reached the drain inverts during the late 1990s but the aquifer was further topped up by four successive above-average rainfall wet seasons from1998 to 02 (see APPENDIX B Section B1). Recent observations of bank slumping in the drains (A. Price, Ord Irrigation Cooperative, 05, pers. comm.) provide further evidence that groundwater seepage is occurring Martins Location The aquifer beneath Martins Location is bounded to the west by the Ord palaeochannel, to the north by the Cave Spring Range and to the southeast by the Pincombe Range. The sediment in the area is predominantly clay, silt and sand and has relatively small permeability compared to the palaeochannel sand and gravel. Groundwater drainage from Martins Location westward into the palaeochannel system appears to be very slow and groundwater replenishment and discharge in the area are dominated by vertical processes. Increased groundwater replenishment from irrigation has lead to the development of groundwater mounds and shallow watertables. Salinity is present along parts of the southwest margin of Martins Location (Figure 4-7). Example hydrographs for monitoring bores located in Martins Location are presented in Figure 4-6 (see also APPENDIX I). They indicate that the groundwater level has generally risen in response to irrigation; however, there are distinctive differences between individual bore hydrographs that reflect local conditions and responses. For example, groundwater rise in the northeast of Martins Location directly west of drain D4 did not begin until irrigation commenced at the site in 1995, despite watertable rise directly east of the D4 since the mid- 1960s. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page

40 Water Level (m AHD) Model D4 level 11C Observed Model calibration Water Level (m AHD) Model D4 level 96/06 Observed Model calibration Figure 4-6: Example hydrographs from monitoring bores 11C and 96/06 in Martins Location 05 CSIRO Figure 4-7: Secondary salinity in Martins Location 4.4. Green Location Groundwater data for Green Location is limited and the hydrogeology of the area is not well defined or described. The palaeochannel is absent and the watertable response to irrigation appears to be similar to the response in Martins Location. Only a small part of Green Location (approximately 0 hectares) has been developed for irrigation under the Ord Stage 1 scheme. The groundwater level in monitoring bores along the southern margin of Green Location near drains D4 and D7 (e.g., bores 6B and PN12S, Figure 4-8) appear to have stabilised at around the drain inverts. Further north the watertable elevation is lower and the aquifer is still filling. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 39

41 Water Level (m AHD) 14 Model D4 level 6B & PN12S 6B observed PN12S observed 6B model calibrated Model basement Figure 4-8: Example hydrographs from monitoring bores 6B and PN12S in Green Location 4.5. M1 Supply Channel A watertable ramp has developed along the central section of the M1 supply channel where it traverses the eastern flank of the Ord valley. Surface water that leaks from the M1 channel cannot drain through the subsurface effectively because the palaeochannel is absent and the aquifer transmissivity is substantially reduced. Water levels in monitoring bores along the M1 supply channel exhibit local effects due to leakage from the channel, groundwater discharge to local drains, and local variation in the aquifer transmissivity. Bore 3H (Figure 4-9) is located next to drain D2 (Figure 1-2) on the western side of the M1 supply channel. The watertable in that area began rising in the early 1960s and appeared to reach the drain level in the mid-1980s. Stunted crop growth has been observed in the area (Figure 4-10) which might be related to waterlogging and the shallow depth to groundwater, though this has not been confirmed by direct investigation. Bores 3K and PN14S are located on the eastern side of the M1 supply channel where the local watertable appears to be controlled by the channel level. More recently, the groundwater level in bore PN14S has risen to within one metre of the ground surface and above the level of the M1. This appears to be a response to the above average rainfall during 1998 to 01. Water Level (m AHD) Observed Model calibrated Model M1 level Model D2 level 3H Model basement Water Level (m AHD) K observed PN14S observed 3K model calibrated Model M1 level Model basement 3K & PN14S Figure 4-9: Example hydrographs from monitoring bores 3H, 3K and PN14S near the M1 supply channel Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page

42 05 Ord Irrigation Cooperative Figure 4-10: Stunted and failed mangoes trees in an area of shallow watertable adjacent to the M1 supply channel Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 41

43 5. Assessment of Salinity Prevention Options 5.1. Techniques for Managing the Watertable To prevent salinity development in the Ord Stage 1 the groundwater system must be managed to achieve the following states: 1. A stable aquifer water balance in which groundwater replenishment is balanced by groundwater discharge in a way that avoids problematic accumulations of salt; and 2. Watertable elevations at a recommended minimum distance of two metres below ground surface. To accomplish this, it is unnecessary to restore the pre-irrigation groundwater conditions. For example, this would be impractical beneath Packsaddle Plain where the raised groundwater level is a reflection of the raised water level in Lake Kununurra. Attempts to dewater the Packsaddle aquifer by pumping would be overwhelmed by lake water continually refilling the aquifer. Similarly, it is not necessary to re-lower the watertable below Ivanhoe Plain by 15 to metres to prevent salinity formation. The available techniques for managing groundwater levels in irrigation areas are well established. They include: Improved on-farm water use efficiency to minimise groundwater replenishment from deep drainage; Improved irrigation water storage and conveyance efficiencies to prevent leakage from the irrigation infrastructure; Groundwater pumping wells to enhanced discharge; Groundwater gravity drainage systems, including deep open surface drains and subsurface drains; and Tree planting to lower shallow watertables and reduce deep drainage Improved Water Use Efficiency Improved water use efficiency is beneficial because it reduces the rate of groundwater replenishment and avoids the subsequent cost of removing surplus groundwater storage by constructed drainage systems. On the other hand, water use efficiency improvements are unlikely to mitigate completely shallow watertables in the Ord Stage 1, because: The irrigation scheme is a flow-through system that diverts large volumes of river water through many kilometres of unlined supply channels and drains; The impracticality and large cost of lining the conveyance system to eliminate leakage is considered to be prohibitive; It is difficult and undesirable to achieve one-hundred percent on-farm water application efficiency because that would prevent deep drainage and salt leaching beyond the soil root zone; and Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page

44 It is evident from recent groundwater monitoring (APPENDIX F, F1) that wet season rainfall has become an important control on groundwater level that is unrelated to irrigation water use efficiency. The option to leak-proof sections of supply channels and drains using impermeable liners has been considered previously (SKM, 1998a). This option is not investigated in the present study because important areas of leakage from supply channels have not been determined either now or in the past Pumping Wells The effectiveness of groundwater pumping to control the watertable beneath northern Ivanhoe plain was trialled recently. The results are reported by Smith et al. (05), who concluded that groundwater pumping to control watertable rise is feasible, provided that: The implemented pumping strategy targets the entire connected palaeochannel system beneath the central and northern parts of Ivanhoe Plain; Total pumping is balanced against excess groundwater recharge from both irrigation and rainfall; and A suitable number of pumping locations can be established where large pumping rates are possible and the pumping does not induce significant additional recharge from surface water sources. Groundwater pumping wells are expected to be ineffective in areas where the palaeochannel is absent and lateral groundwater transmission is poor. In this situation, low bore yields limit the amount of groundwater that can be pumped and restrict the area of watertable drawdown to a small radius around the bores. If present, local occurrences of more conductive sediments (e.g., discontinuous lenses and stringers of gravel and sand) may provide a suitable target for a local groundwater pumping strategy. Groundwater pumps are not recommended on Packsaddle Plain in locations where there is a strong groundwater connection between the aquifer and Lake Kununurra and pumping would drawdown the watertable below the lake level. Pumping bores that are too close to the lake are likely to draw large amounts of lake water through the pumping scheme and be ineffective at controlling both the watertable and the aquifer water balance Drains When sediment permeability and surface relief are adequate, natural subsurface drainage may be enough to prevent shallow watertables from developing under irrigation; however, these conditions are rare in areas where saline soils occur. Various types of constructed groundwater drainage systems are used successfully throughout the world, including open ditches that intercept the watertable, and subsurface tile drains and perforated plastic pipes. The optimal depth and spacing of drains depends upon the soil characteristics (e.g., layering and permeability) and the local economic considerations. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 43

45 In the Ord Stage 1, open surface drains are now thought to be discharging considerable volumes of groundwater though they were not designed for this purpose. Subsurface drainage systems have not been tried but they may provide a viable option for controlling the watertable in areas where pumping wells are impracticable and the existing surface drains are inadequate (e.g., Martins Location) Lowered Water Level in Lake Kununurra Manipulation of the water level in Lake Kununurra is an additional management option for controlling the watertable beneath Packsaddle Plain. Multiple factors impose limits on the amount of water level variation that is acceptable; however, a minimum water level constraint has not been determined. Influencing factors include, the operational requirements for the M1 off-take and private irrigation pumps; protection of Lake Kununurra as a Ramsar listed wetland; and consideration of local amenity (e.g., foreshore aesthetics and recreation on Lake Kununurra) Trees Recent watertable decline beneath a stand of unirrigated Sandalwood and African Mahogany in Martins Location (see APPENDIX F) has lead to speculation that tree plantations can dramatically lower the watertable through decreasing local groundwater replenishment and increasing transpiration. In this particular case, the watertable drawdown was large but restricted to the immediate area of the plantation. In areas where the aquifer has small transmissivity, like Martins Location and along the flanks of the Ord valley, tree plantations may be effective at lowering the local watertable but this advantage may not be sustainable in the long-term due to salt accumulation in the root zone. In areas underlain by palaeochannel, irrigated tree plantations are unlikely to have a significant effect on the local watertable but may help to reduce subsurface drainage from the palaeochannel system if the leaching fraction beneath the trees is less than for other crop types. An additional advantage may be gained if the trees can derive a portion of their water requirement directly from the watertable (Ayars et al., 05). The results from numerous groundwater pumping tests (e.g., Smith et al., 05 and APPENDIX C Section C4.1) provide clear evidence that small watertable drawdowns would occur beneath trees that are planted over the palaeochannel system. Commercially grown dryland (unirrigated) trees that obtain a significant proportion of their water requirement from the watertable would provide the best benefits with respect to water and salt balance management. Importantly, trees should not be considered as straightforward substitutes for groundwater pumping wells. While mechanical groundwater pumps remove both water and solutes from the aquifer, trees remove only water and concentrate the solutes in the subsurface. If trees are used to manage the aquifer water balance then the subsurface salt balance and sustainability of tree growth also must be considered. To prevent harmful accumulation of salt in the root zone of irrigated crops, total irrigation and rainfall must exceed total evapotranspiration, and that excess must pass through the root zone in a minimum amount called the leaching fraction. This fraction can be minimised through careful management; however, subsurface drainage is an unavoidable element of sustainable irrigated agriculture Salinity Risk Areas The risk of secondary salinity in the Ord Stage 1 was assessed based on the historic rate of groundwater rise, current groundwater conditions and predicted watertable responses to future groundwater replenishment scenarios. Three risk classes were defined: Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 44

46 Low salinity risk greater than 4 metres depth to the watertable; Moderate salinity risk between 2 and 4 metres depth to the watertable; and High potential salinity risk less than 2 metres depth to the watertable The above high risk class is a measure of potential salinity risk only because depth to watertable, alone, does not determine whether salinity will develop. Groundwater quality and the local direction of groundwater drainage also are important factors. If the watertable is less than two metres below ground surface (high potential risk) but there is net replenishment of groundwater and net downward drainage, then subsurface salt accumulation and salinity formation are unlikely. In this case, the potential for waterlogging may be more of a concern. On the other hand, if subsurface drainage is slow and there is net discharge of groundwater from the watertable, then salinity formation is likely. Predicting whether salinity will develop locally in areas of shallow watertable using a valleyscale modelling tool is difficult. Modelling at this large-scale is important for understanding the aquifer water balance and valley-scale watertable behaviour but it cannot incorporate the local hydrogeological conditions and soil characteristics (e.g., local layering) that control subsurface drainage at the local-scale. The approach in this study was to classify areas of shallow watertable as having a high potential salinity risk and recommend targeted monitoring for signs of emergent salinity. Areas with a high potential salinity risk (< 2 metres depth to watertable) and evidence of emergent salinity require immediate management to lower the watertable and leach the accumulated salt from the soil profile. Areas with low to moderate salinity risk (> 2 metres depth to watertable) should be monitored to ensure that the watertable remains at a safe distance below ground surface and there is no emergent sign of salinity formation Observed and Simulated Depth to Watertable Figure 5-1 depicts the interpolated depth to watertable in the Ord Stage 1 in October 02 based on the difference between measured groundwater levels in the surveyed monitoring bores and the digital elevation model constructed for the groundwater model. In comparison, the model-simulated depth to watertable in 02 is presented in Figure 5-2. Together, these maps provide a valley-scale impression of the groundwater conditions in 02 when the watertable throughout much of the irrigation area was at an historical high level (refer to the monitoring bore hydrographs in APPENDIX I). Differences between the maps illustrate the typical size of error between the observed and modelled data. There is a broad-scale agreement between the areas of low, moderate and high potential salinity risk but clear local differences exist. Both the estimates of groundwater level and ground surface elevation contain uncertainty and errors of order 1 to 2 metres. These are significant when trying to predict depth to watertable in the range 2 to 4 metres (i.e., moderate to high potential salinity risk). The maps should be used only to provide an indicative estimate of salinity risk at farm-scale. Low salinity risk exists in the southeast part of Packsaddle Plain, Ivanhoe South, Ivanhoe Central and the western part of Ivanhoe North, where depth to groundwater is generally greater than 4 metres. Moderate to high potential salinity risk is evident beneath most of Packsaddle Plain, Martins Location, Ivanhoe North, Cave Spring Gap and along parts of the M1 supply channel. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 45

47 Figure 5-3 depicts the model-simulated depth to watertable one year later in 03. It indicates that the area with a moderate to high potential salinity risk is less than in 02 because the watertable has fallen slightly in the intervening period. In this context, it is clear that the assessment of long-term salinity risk needs to be based on the likely future groundwater conditions that will develop in response to current irrigation practices and climate. Two possible future watertable scenarios are considered in the following section. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 46

48 Figure 5-1: Estimated depth to watertable in October 02 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 47

49 Figure 5-2: Model calibrated depth to watertable in 02 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 48

50 Figure 5-3: Model calibrated depth to watertable in 03 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 49

51 Future Watertable Scenarios without Management The Ord Stage 1 groundwater model was applied to predict the depth to watertable in 25 for the groundwater replenishment scenarios listed in Table 5-1. These simulations assumed the continuation of current irrigation practices and groundwater management, and were used to estimate the extent of the irrigation area where secondary salinity may ultimately be a threat. Simulated depth-to-watertable maps for scenarios W0 and W2 are presented in Figure 5-4 and Figure 5-5, respectively. Water balance components for selected areas are listed in Table 5-2 and Table 5-3. Table 5-1: Summary of groundwater replenishment scenarios Scenario ID W0 W2 Description Expected replenishment rates Reduced replenishment rates Simulation Specifications Period groundwater replenishment rates No groundwater pumps No subsurface drains Discharge to the irrigation channel network Current irrigation practices groundwater replenishment rates No groundwater pumps No subsurface drains Discharge to the irrigation channel network Discharge to irrigation drains Current irrigation practices Scenario W0 assumed that the model-calibrated groundwater replenishment rates (APPENDIX E Section E5.1) for the period 1995 to 1998 were representative of the contemporary replenishment rates under average rainfall conditions. Rainfall was above average during the following four years (1999 to 01) and the calibrated groundwater replenishment rates were significantly larger in that period. Scenario W2 was based on the model-calibrated groundwater replenishment rates for the period 02 to 03, which were significantly reduced in response to below average rainfall. The net replenishment rates in scenarios W0 and W2 were approximately 45% and 15%, respectively, of the net replenishment rate for the wet period between 1998 and 01. Thus, the reduced replenishment scenario (W2) had a net replenishment rate of around 33% of the expected replenishment rate (W0). A scenario with groundwater replenishment rates was not simulated because it was considered unrealistic to expect above average rainfall for twenty years. In comparison with the estimated depth to watertable in 02 (Figure 5-1 and Figure 5-2) an increased area with medium to high potential salinity risk occurred in 25 for the expected groundwater replenishment rate (scenario W0). The areas most affected were Martins Location and parts of Ivanhoe North where the palaeochannel aquifer is absent. In these areas the aquifer continued to fill because lateral drainage through the subsurface was less than the net replenishment rate. Relatively minor changes in depth to watertable were predicted beneath Packsaddle Plain and the salinity risk remained low in Ivanhoe Central, Ivanhoe South and the southeast of Packsaddle Plain. 2 The groundwater replenishment rate beneath the Sandalwood and Mahogany plantation in Martins Location was set equal to zero for the period of the simulation. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 50

52 Under the reduced replenishment scenario (W2) the depth to watertable and extent of the salinity risk areas were slightly improved compared to current conditions (i.e., compare Figure 5-3 and Figure 5-5). Overall, the results indicated that the salinity risk throughout the irrigation area in 25 is unlikely to be significantly reduced compared to the current situation. It is more likely that there will be an equal or greater area of shallow watertable in 25 if no action is taken to manage groundwater rise. Table 5-2: Scenario W0; water balance 25 Packsaddle Plain - 25 Water Balance Component Flow Rate [GL/yr] Total pumping 0.00 Net inflow from Lake Kununurra 0.10 Net replenishment from irrigation & rainfall 4.23 Discharge to creeks, drains & ground surface Leakage from SP1 0. Aquifer storage / Imbalance 0.0 Ivanhoe North, Green Location & Cave Spring Gap - 25 Water Balance Component Flow Rate [GL/yr] Total pumping 0.00 Net leakage from M Net replenishment from irrigation & rainfall 9.08 Discharge to creeks, drains & ground surface Outflow from Cave Spring Gap Imbalance (storage; inflow from Ivanhoe Central & Martins Location -0.1 Table 5-3: Scenario W2: water balance 25 Packsaddle Plain - 25 Water Balance Component Flow Rate [GL/yr] Total pumping 0.00 Net inflow from Lake Kununurra 1.57 Net replenishment from irrigation & rainfall 1.33 Discharge to creeks, drains & ground surface Leakage from SP Aquifer storage / Imbalance 0.0 Ivanhoe North, Green Location & Cave Spring Gap - 25 Water Balance Component Flow Rate [GL/yr] Total pumping 0.00 Net leakage from M Net replenishment from irrigation & rainfall 3.00 Discharge to creeks, drains & ground surface Outflow from Cave Spring Gap Imbalance (storage; inflow from Ivanhoe Central & Martins Location -0.6 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 51

53 Figure 5-4: Replenishment scenario W0; model-simulated depth to watertable in 25 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 52

54 Figure 5-5: Replenishment scenario W2; model-simulated depth to watertable in 25 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 53

55 5.3. Salinity Management Scenarios This part of the report presents simulated groundwater management scenarios based on the expected groundwater replenishment rates used in scenario W0. The objective of this work was to identify the best management option in each of the salinity risk areas independent of the replenishment rates. For example, if the replenishment rates from scenario W0 are an over-estimate of the real replenishment rates during the next twenty years, then the quantity of groundwater to be pumped or drained will be less, though the same management option will be applicable. Similarly, if the replenishment rates in scenario W0 are an under-estimate, then the quantity of groundwater to be pumped or drained will be larger Packsaddle Plain A summary of the model simulated groundwater management scenarios for Packsaddle Plain are listed in Table 5-4. They include groundwater pumping from the palaeochannel aquifer, lowering the level in Lake Kununurra, and a combination of both options. Simulated maps of depth to watertable in 25 for each scenario are presented in Figure 5-7 to Figure Water balance components are listed in Table 5-5 to Table Table 5-4: Summary of groundwater management scenarios; Packsaddle Plain Scenario ID P3 P4 P5 P6 P7 P8 Description Groundwater pumping from palaeochannel at % replenishment from irrigation and rainfall (valley-scale tactic) Groundwater pumping from palaeochannel at 60% replenishment from irrigation and rainfall (valley-scale tactic) Groundwater pumping from palaeochannel at 100% replenishment from irrigation and rainfall (valley-scale tactic) Lowered level in Lake Kununurra (0.6 m) (valley-scale tactic) Lowered level in Lake Kununurra (1.0 m) (valley-scale tactic) P4 and P6 (valley-scale tactic) Simulation Period Specifications and Figure Number Scenario W0 replenishment rates 2 bores 1,750 KL/day each Total pumping = 3,500 KL/day (1. GL/yr) No subsurface drains Discharge to the irrigation channel network Figure 5-7 and Table Scenario W0 replenishment rates 2 bores 3,500 KL/day each Total pumping = 7,000 KL/day (2.55 GL/yr) No subsurface drains Discharge to the irrigation channel network Figure 5-8 and Table Scenario W0 replenishment rates 4 bores 3,000 KL/day each Total pumping = 12,000 KL/day (4. GL/yr) No subsurface drains Discharge to the irrigation channel network Figure 5-9 and Table Scenario W0 replenishment rates Lake level lowered from 41.6 to 41.0 m AHD No groundwater pumps No subsurface drains Discharge to the irrigation channel network Discharge to irrigation drains Figure 5-10 and Table Scenario W0 replenishment rates Lake level lowered from 41.6 to.6 m AHD No groundwater pumps No subsurface drains Discharge to the irrigation channel network Discharge to irrigation drains Figure 5-11 and Table See above Figure 5-13 and Table 5-10 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 54

56 Natural groundwater drainage from beneath Packsaddle Plain is to Packsaddle Creek and Lake Kununurra in some areas. Figure 5-6 depicts the area of Packsaddle Plain wherein the watertable in October 02 was estimated to be above the water level in Lake Kununurra (refer also to Appendix Figure E-3). From that area, some groundwater drains back toward the lake. The purpose of managed groundwater pumping should be to balance net recharge from irrigation and rainfall. More specifically, the total groundwater pumping volume should be designed to capture the component of groundwater replenishment that does not drain naturally without the watertable rising to within 2 metres of the ground surface. The groundwater management strategy should seek to lower the watertable to at least 2 metres below ground surface but it should minimise the area of watertable that is un-necessarily drawn down to below the surface water level in Lake Kununurra. Lowering the watertable to below the lake level would cause lake water to flow into the aquifer and re-fill it. Scenario P3 - pumping at % replenishment from irrigation and rainfall Two pumping bores screened in the palaeochannel aquifer drained 1. GL per year of groundwater, which was approximately thirty percent of the applied replenishment rate on Packsaddle Plain. The pumping locations (Figure 5-7) were positioned to the southwest of the bedrock high which divides Packsaddle Plain (see Appendix Figure C-7) and away from Lake Kununurra to minimise induced inflow from the lake. This strategy provided only a limited reduction in salinity risk within the irrigation area, indicating that the pumping rate was too small compared to the rate of groundwater replenishment. In 25, approximately % of the pumping was balanced by induced inflow from Lake Kununurra, while 60% was balanced by decreased discharge to the surface environment (i.e., creeks, drains and ground surface). Table 5-5: Scenario P3; water balance 25 Packsaddle Plain - 25 Water Balance Component Flow Rate [GL/yr] Total pumping -1. Net inflow from Lake Kununurra 0.54 Net replenishment from irrigation & rainfall 4.23 Discharge to creeks, drains & ground surface Leakage from SP Aquifer storage -0.1 Increase in inflow from Lake Kununurra as a percentage of pumping % Decrease in drainage to the surface environment as a percentage of pumping 60% Change in aquifer storage as a percentage of pumping 6% 100% Scenario P4 - pumping at 60% replenishment from irrigation and rainfall Two pumping bores screened in the palaeochannel aquifer drained 2.56 GL per year of groundwater, which was approximately sixty percent of the applied replenishment rate on Packsaddle Plain. This strategy reduced the high potential salinity risk along the southeast part of Packsaddle Creek and decreased the area of high salinity risk in the southwest part of the irrigation area. It had only a small effect on salinity risk in the northeast part of the irrigation area i.e., on the northeast side of the bedrock high. In 25, approximately 39% of the pumping was balanced by induced inflow from Lake Kununurra while 59% was balanced by decreased discharge to the surface environment. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 55

57 Table 5-6: Scenario P4; water balance 25 Packsaddle Plain - 25 Water Balance Component Flow Rate [GL/yr] Total pumping Net inflow from Lake Kununurra 1.09 Net replenishment from irrigation & rainfall 4.23 Discharge to creeks, drains & ground surface Leakage from SP Aquifer storage 0.0 Increase in inflow from Lake Kununurra as a percentage of pumping 39% Decrease in drainage to the surface environment as a percentage of pumping 59% Change in aquifer storage as a percentage of pumping 2% 99% Scenario P5 - pumping at 100% replenishment from irrigation and rainfall Four pumping bores screened in the palaeochannel aquifer drained 4. GL per year of groundwater, which was approximately equal to the applied replenishment rate on Packsaddle Plain. Two pumping bores were located to the southwest of the bedrock high (Figure 5-9) as for scenarios P3 and P4 and two additional pumping bores were located in the northwest part of the irrigation area on the northwest side of the bedrock high. This was a successful strategy for controlling the salinity risk throughout most of the Packsaddle irrigation area; however, in 25, approximately % of the pumping was balanced by induced inflow from Lake Kununurra. It is apparent that pumping from both sides of the bedrock high that divides Packsaddle Plain is required in order to manage the watertable across the whole irrigation area. Careful control of pumping in the northeast part of the irrigation area is most critical to minimise the costs associated with drawing lake water through the pumping scheme. Table 5-7: Scenario P5; water balance 25 Packsaddle Plain - 25 Water Balance Component Flow Rate [GL/yr] Total pumping -4. Net inflow from Lake Kununurra 1.92 Net replenishment from irrigation & rainfall 4.23 Discharge to creeks, drains & ground surface Leakage from SP Aquifer storage 0.0 Increase in inflow from Lake Kununurra as a percentage of pumping % Decrease in drainage to the surface environment as a percentage of pumping 56% Change in aquifer storage as a percentage of pumping 1% 99% Scenarios P6 and P7 lowered water level in Lake Kununurra The surface water level in Lake Kununurra was lowered by 0.6 metres to 41 metres AHD, and by 1 metre to.6 metres AHD in scenarios P6 and P7, respectively. Lowering the lake level by up to 1 metre had a significant positive benefit to the salinity risk within the Packsaddle irrigation area, particularly in the northeast part. In 25, % (scenario P5) and 35% (scenario P6) of the total replenishment from irrigation, rainfall and channel leakage was draining back to Lake Kununurra. In comparison, with the lake set at 41.6 metres AHD (scenario W0) there was no net drainage to Lake Kununurra in 25. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 56

58 The results indicate that lowering the level of Lake Kununurra would enhance the natural sub-surface drainage of groundwater into the lake. This would reduce the cost requirements for constructed drainage but introduce other costs associated with the operation of the M1 and SP1 off-takes and private irrigation pumps. Manipulation of the water level in Lake Kununurra is restricted by the existing infrastructure design; operational requirements for irrigation water supply; protection of Lake Kununurra as a Ramsar listed wetland; and local amenity, recreation, property and tourism values. These constraints have not been evaluated. It also is apparent that raising the lake level would significantly increase the salinity risk on Packsaddle Plain and increase the need for constructed drainage. Figure 5-12 shows the simulated increase in salinity risk in 25 if the level in Lake Kununurra were raised by 1 metre to.6 metres AHD. Table 5-8: Scenario P6; water balance 25 Packsaddle Plain - 25 Water Balance Component Flow Rate [GL/yr] Total pumping 0.00 Net inflow from Lake Kununurra Net replenishment from irrigation & rainfall 4.23 Discharge to creeks, drains & ground surface Leakage from SP1 0. Aquifer storage 0.1 Drainage to Lake Kununurra as a percentage of total replenishment % Drainage to the surface environment as a percentage of total replenishment 68% 98% Table 5-9: Scenario P7; water balance 25 Packsaddle Plain - 25 Water Balance Component Flow Rate [GL/yr] Total pumping 0.00 Net inflow from Lake Kununurra Net replenishment from irrigation & rainfall 4.23 Discharge to creeks, drains & ground surface Leakage from SP1 0. Aquifer storage 0.0 Drainage to Lake Kununurra as a percentage of total replenishment 35% Drainage to the surface environment as a percentage of total replenishment 66% 101% Scenario P8 (combined P4 & P6) - pumping at 60% replenishment from irrigation and rainfall and lowered water level in Lake Kununurra Two pumping bores screened in the palaeochannel aquifer drained 2.56 GL per year of groundwater from the southwest part of the irrigation area (scenario P4) and the water level in Lake Kununurra was lowered by 0.6 metres (scenario P6). This was an effective management strategy that significantly reduced the areas of high salinity risk along Packsaddle Creek and within the Packsaddle irrigation area. Lowering the lake level provided a direct benefit in the northeast part of the irrigation area through lowering of the surrounding groundwater level. This alleviated the need to pump groundwater in that area. Pumping was effective at lowering the watertable in the southwest part of the irrigation area on the southwest side of the bedrock high. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 57

59 Most significantly, the pumping did not induce additional inflow from Lake Kununurra due to the lowered lake level. In 25, approximately 94% of the pumping was balanced by reduced discharge to the surface environment. Table 5-10: Scenario P8; water balance 25 Packsaddle Plain - 25 Water Balance Component Flow Rate [GL/yr] Total pumping Net inflow from Lake Kununurra 0.09 Net replenishment from irrigation & rainfall 4.23 Discharge to creeks, drains & ground surface Leakage from SP1 0. Aquifer storage -0.1 Increase in inflow from Lake Kununurra as a percentage of pumping 0% Decrease in drainage to the surface environment as a percentage of pumping 94% Change in aquifer storage as a percentage of pumping 4% 98% Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 58

60 Figure 5-6: Watertable rise above Lake Kununurra in October 02; observed groundwater levels Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 59

61 Figure 5-7: Scenario P3; model-simulated depth to watertable in 25 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 60

62 Figure 5-8: Scenario P4; model-simulated depth to watertable in 25 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 61

63 Figure 5-9: Scenario P5; model-simulated depth to watertable in 25 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 62

64 Figure 5-10: Scenario P6; model-simulated depth to watertable in 25 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 63

65 Figure 5-11: Scenario P7; model-simulated depth to watertable in 25 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 64

66 Figure 5-12: Level in Lake Kununurra raised by 1 meter; model-simulated depth to watertable in 25 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 65

67 Figure 5-13: Scenario P8; model-simulated depth to watertable in 25 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 66

68 Ivanhoe North and Cave Spring Gap Table 5-4 lists the simulated groundwater management scenarios for Ivanhoe North and Cave Spring Gap. They include groundwater pumping from the palaeochannel aquifer and subsurface drains in selected areas. Maps of depth to watertable in 25 for each scenario are presented in Figure 5-14 to Figure 5-17, and the water balance components are listed in Table 5-12 to Table Table 5-11: Summary of groundwater management scenarios; Ivanhoe Plain Scenario ID I1 I2 I3 M1 Description Groundwater pumping from palaeochannel at % replenishment from irrigation and rainfall (valley-scale tactic) Groundwater pumping from palaeochannel at 60% replenishment from irrigation and rainfall (valley-scale tactic) Groundwater pumping from palaeochannel at 100% replenishment from irrigation and rainfall (valley-scale tactic) Subsurface drains in selected areas (local-scale tactic) Simulation Period Specifications and Figure Number Scenario W0 replenishment rates Ivanhoe North o 4 bores 1,850 KL/day each o Total pumping = 7,0 KL/day (2.7 GL/yr) Cave Spring Gap o 1 bore 1,000 KL/day o Total pumping = 1,000 KL/day (0. GL/yr) Discharge to the irrigation channel network Figure 5-14 and Table Scenario W0 replenishment rates Ivanhoe North o 4 bores 3,700 KL/day each o Total pumping = 14,800 KL/day (5.4 GL/yr) Cave Spring Gap o 1 bore 2,000 KL/day o Total pumping = 2,000 KL/day (0.73 GL/yr) Discharge to the irrigation channel network Figure 5-15 and Table Scenario W0 replenishment rates Ivanhoe North o 8 bores 3,000 KL/day each o Total pumping =,800 KL/day (9.1 GL/yr) Cave Spring Gap o 1 bore 3,0 KL/day o Total pumping = 3,0 KL/day (1.2 GL/yr) Discharge to the irrigation channel network Figure 5- and Table Scenario W0 replenishment rates No groundwater pumps Subsurface drains at approximately 2 m below ground Discharge to the irrigation channel network Figure 5-17 Groundwater drainage from Ivanhoe North occurs mainly as discharge to the irrigation channel network and slow drainage through Cave Spring Gap. Discharge to drains D3, D4, D5, D6 and D7 (Figure 1-2) on Ivanhoe North, and D8 on Cave Spring Gap, has been evident since the mid-1990s (refer to Appendix Figure E-4). Shallow watertables with a moderate to high potential salinity risk occur throughout Ivanhoe North and Cave Spring Gap. Based on the replenishment rates simulated in scenario W0, the parts of the irrigation area with moderate to high potential salinity risk were predicted to increase by 25 if no action were taken to manage the watertable. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 67

69 The recent groundwater pumping trial on Ivanhoe North (Smith et al., 05) confirmed that the palaeochannel in this area has large transmissivity and is laterally extensive. Long-term pumping rates of order 5,000 KL/day induced only relatively small watertable drawdowns of order 1 to 2 metres around the pumping locations. In this context, local pumping is not considered to be an effective option for controlling the local watertable because the pumped water can easily be replaced by groundwater flow from other parts of the aquifer. Coordinated pumping from multiple locations is required to manage the water balance at the valley-scale. While some parts of Ivanhoe Plain are unlikely to develop shallow watertables due to topographic controls, attention to groundwater management in these areas is still required to benefit the areas at risk from rising groundwater. Scenario I1 - pumping at % replenishment from irrigation and rainfall Four pumping bores in Ivanhoe North (Figure 5-14) and one pumping bore in Cave Spring Gap drained a total of 3.07 GL per year of groundwater from the palaeochannel aquifer, which was approximately thirty percent of the total applied replenishment rate in these areas. The salinity risk was improved by the pumping although parts of Ivanhoe North and Cave Spring Gap still had a high potential salinity risk in 25. The watertable remained above the drain network and groundwater continued to discharge into the drains. In 25, approximately 85% of the pumping was balanced by reduced discharge to the surface environment, while approximately 14% was balanced by induced inflow from other parts of the aquifer system (i.e., Ivanhoe Central and Martins Location). Of interest, the pumping rate in this scenario was approximately equal to the average pumping rate of 3.3 GL/yr achieved in the aquifer pumping trial on Ivanhoe North. In the analysis of those results (Smith et al., 05) it was concluded that the pumping was effective at removing groundwater from storage but was insufficient compared to replenishment to achieve effective watertable control. It also was concluded that the pumping did not induced significant additional leakage from the M1 supply channel. Table 5-12: Scenario I1; water balance 25 Ivanhoe North, Green Location & Cave Spring Gap - 25 Water Balance Component Flow Rate [GL/yr] Total pumping Net leakage from M Net replenishment from irrigation & rainfall 9.08 Discharge to creeks, drains & ground surface Outflow from Cave Spring Gap Imbalance (storage; inflow from Ivanhoe Central & Martins Location) -0.6 Decrease in drainage to the surface environment as a percentage of pumping 85% Decrease in outflow through Cave Spring Gap as a percentage of pumping 0% Increase in imbalance as a percentage of pumping 14% Increase in leakage from M1 as a percentage of pumping 1% 100% Scenario I2 - pumping at half the rate of replenishment from irrigation and rainfall Four pumping bores in Ivanhoe North and one pumping bore in Cave Spring Gap drained a total of 6.13 GL per year of groundwater from the palaeochannel aquifer, which was approximately sixty percent of the applied replenishment rate. This strategy reduced the salinity risk to a moderate level throughout most of Ivanhoe North and Cave Spring Gap. Several small areas with a high potential salinity risk persisted around the margins of the irrigation area where the palaeochannel aquifer is absent and the groundwater could not drain effectively in response to the pumping. The watertable remained above the drain inverts within a smaller area compared to scenario I1. In 25, approximately 78% of the Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 68

70 pumping was balanced by reduced discharge to the surface environment and approximately 21% (1. GL/yr) was balance by induced inflow from other parts of the aquifer system. Table 5-13: Scenario I2; water balance 25 Ivanhoe North, Green Location & Cave Spring Gap - 25 Water Balance Component Flow Rate [GL/yr] Total pumping Net leakage from M1 0. Net replenishment from irrigation & rainfall 9.08 Discharge to creeks, drains & ground surface Outflow from Cave Spring Gap Imbalance (storage; inflow from Ivanhoe Central & Martins Location) -1.4 Decrease in drainage to the surface environment as a percentage of pumping 78% Decrease in outflow through Cave Spring Gap as a percentage of pumping 0% Increase in imbalance as a percentage of pumping 21% Increase in leakage from M1 as a percentage of pumping 1% 100% Scenario I3 - pumping at double the rate of replenishment from irrigation and rainfall Eight pumping bores in Ivanhoe North and one pumping bore in Cave Spring Gap drained a total of 10.3 GL per year of groundwater from the palaeochannel aquifer. This was approximately equal to the applied replenishment rate. The salinity risk in 25 was low throughout most of Ivanhoe North and Cave Spring Gap, except around the northern and western margins where the palaeochannel aquifer is absent. Discharge to the surface environment was substantially reduced because the watertable was drawn down to below the drain inverts in most areas. Groundwater discharged to small parts of drains D4 and D5 (Figure 1-2) only. In 25, approximately 73% of the pumping was balanced by reduced discharge to the surface environment and approximately 25% (2.55 GL/yr) was balance by induced inflow from other parts of the aquifer system. The pumping from the palaeochannel also resulted in reduced salinity risk along the western margin of Martins Location; however, it was mostly ineffective at draining the groundwater from Martins Location. Table 5-14: Scenario I3; water balance 25 Ivanhoe North, Green Location & Cave Spring Gap - 25 Water Balance Component Flow Rate [GL/yr] Total pumping -10. Net leakage from M1 0. Net replenishment from irrigation & rainfall 9.08 Discharge to creeks, drains & ground surface Outflow from Cave Spring Gap Imbalance (storage; inflow from Ivanhoe Central & Martins Location) -2.7 Decrease in drainage to the surface environment as a percentage of pumping 73% Decrease in outflow through Cave Spring Gap as a percentage of pumping 0% Increase in imbalance as a percentage of pumping 25% Increase in leakage from M1 as a percentage of pumping 2% 100% Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 69

71 Figure 5-14: Scenario I1 - simulated depth to watertable in 25 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 70

72 Figure 5-15: Scenario I2 - simulated depth to watertable in 25 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 71

73 Figure 5-: Scenario I3 - simulated depth to watertable in 25 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 72

74 Martins Location Based on current knowledge, groundwater pumping from Martins Location is considered to be an impracticable management option. The aquifer in that area has poor transmissivity and it is expected that only small pumping rates could be achieved. The exception would be where local disconnected pockets of gravel and sand exist that could be locally dewatered though no such areas have been identified to date. The watertable in parts of Martins Location has been raised to above the drain inverts but the effectiveness of these drains in controlling the watertable is uncertain due to the lack of information about the local drainage characteristics of the soils. It is likely that the drains exert only a local (i.e., tens of metres) influence on the watertable. Table 5-15: Summary of groundwater management scenarios; Martins Location Scenario ID M1 Description Subsurface drains in selected areas (localscale solution) Simulation Period Specifications and Figure Number Scenario W0 replenishment rates No groundwater pumps Subsurface approx. 0 m spacing Discharge to the irrigation channel network Figure 5-17 Scenario M1 subsurface drains in selected locations Scenario M1 (Table 5-15) simulated the effect of subsurface drains installed in selected areas of Martins Location and Ivanhoe North (Figure 5-17). This scenario is presented to illustrate the utility of subsurface drains to control the local watertable in areas where the palaeochannel aquifer is absent and pumping is impracticable. It does not provide evidence that subsurface drainage would work. Field drainage trials, which are planned for 06-08, are required to determine the feasibility and effectiveness of subsurface drains in the Ord Stage 1. Controlling the Watertable using Trees Trees do not provide a simple alternative to groundwater pumps (Section 5.1.5). The concept of in-situ use by crops (Ayars et al., 05) is based upon reducing groundwater drainage and watertable accession by managing crops to extract a percentage of their water requirements from the shallow watertable. Ideally, this could reduce or eliminate the need for pumps and constructed drains but would not eliminate groundwater drainage. The in-situ use approach is best suited to deep rooting crops such as trees. Specific model simulations to investigate the effect of tree plantations were not conducted in this study because the differences between trees and other irrigated crops, in terms of their leaching fractions and groundwater uptake rates, are not known. Generally speaking, if trees were effective at reducing the amount of groundwater drainage from the irrigation area, then they would have the same effect as reduced replenishment. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 73

75 Figure 5-17: Scenario M1; model-simulated depth to watertable in 25 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 74

76

77 6. Conclusions and Recommendations Targeted management of groundwater level and salinity in the Ord Stage 1 is recommended based on the following tactical approach: Reduce groundwater replenishment wherever practical, through improved water use efficiency practices, to minimise the total amount of subsurface drainage through the irrigation area; Monitor salinity trends in high potential risk areas to track the effects of land and water management practices, and to trigger salinity management actions; Manage the existing infrastructure to exploit natural subsurface drainage from the irrigation area (e.g., maintenance and deepening of existing surface drains, and lowering of Lake Kununurra); and Construct new drainage systems as required to prevent salinity formation (e.g., groundwater pumps, deep open drains and subsurface drains). Optimisation of water use efficiency practice to minimise the total amount of groundwater drainage from the irrigation area is a proactive measure that can help to reduce the potential extent of salinity formation and the ultimate cost of constructed drainage. Careful monitoring of the watertable elevation and groundwater salinity is recommended in areas with a high potential salinity risk to track the affects on groundwater of changes in land and water management practices. Salinity monitoring should be a key component of future groundwater management because the model-predictions of future salinity risk are based only on valley-scale estimates of depth to watertable. The salinity risk maps produced in this study should be used as a basis for targeted salinity monitoring. Where practicable, it is recommended that the existing irrigation infrastructure be managed to improve the efficiency of groundwater drainage from the irrigation area. Examples include the regular maintenance of selected irrigation drains to enhance groundwater discharge, and lowering the level in Lake Kununurra. Irrigation drains that are important for groundwater drainage should be kept relatively clear of sediment build-up and be cut to their maximum practical depth. If the above measures are insufficient to prevent salinity formation, then it is recommended that new drainage systems be constructed to lower the watertable and provide a pathway for salt transport out of the irrigation area. (e.g., groundwater pumps, additional deep open drains, subsurface drains). Groundwater management recommendations that are specific to each of the groundwater management areas are discussed below, and a summary of those recommendations is presented in Table Packsaddle Plain Moderate to high potential salinity risk is likely to persist in the Packsaddle irrigation area unless action is taken to control the watertable response to future irrigation and rainfall. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 76

78 The degree of watertable control that can be achieved beneath Packsaddle Plain will be affected by the presence of Lake Kununurra and the water level at which it is operated. It is recommended that Lake Kununurra be lowered to the minimum acceptable water level, and that research be undertaken to establish what that level is. Lowering the lake level would enhance the natural subsurface drainage of groundwater from the Packsaddle aquifer into the lake. It is recommended that the lake level not be raised above its present level. If emergent salinity is apparent over the palaeochannel system, then groundwater pumping from the southwest part of Packsaddle Plain is recommended to lower the watertable. The watertable should be lowered to at least 2 metres below ground surface; however, drawdown of the watertable below the level of Lake Kununurra should be minimised to prevent additional lake water being drawn through the pumping scheme. In this context, the pumping would be more effective and efficient if the lake level also were lowered. Pumping from the northeast part of Packsaddle Plain near Lake Kununurra is not recommended unless it is unavoidable or the lake level was lowered and the pumping did not drawdown the watertable below the lake level. If emergent salinity is apparent over parts of the irrigation area where the palaeochannel system is absent and groundwater drainage is poor, then subsurface drains are recommended to lower the watertable in these areas Ivanhoe South and Ivanhoe Central Shallow watertables are unlikely to develop in Ivanhoe South and Central in response to future rainfall and irrigation due to the natural rate of subsurface drainage and the local topographic controls. It is recommended that the watertable elevation throughout Ivanhoe South and Central be monitored on an ongoing basis to assess long-term trends in response to land use and groundwater management actions. It is recommended that Ivanhoe Central be managed to minimise the net rate of groundwater replenishment from irrigation and rainfall. Although the salinity risk is low throughout this part of the palaeochannel system, inattention to groundwater management would be likely to increase the effort required to manage the watertable further north in the palaeochannel system where shallow watertables are expected to develop (i.e., beneath Ivanhoe North) Ivanhoe North and Cave Spring Gap Areas of high potential salinity risk are likely to develop in parts of Ivanhoe North and Cave Spring Gap in response to future irrigation and rainfall. Salinity formation is most likely in the high potential risk areas where the palaeochannel is absent and groundwater drainage is poor. The irrigation surface drain network appears to be the main control on groundwater drainage in Ivanhoe North. It is evident that groundwater discharges to the drains when the watertable is raised to above the drain inverts; however, areas with a shallow watertable can still develop where the surface drains are shallow, in the areas in between the drains, and where the palaeochannel is absent and subsurface drainage is poor. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 77

79 Groundwater in Cave Spring Gap appears to drain less effectively than it does in Ivanhoe North because the main irrigation drains in Cave Spring Gap run along the flanks of the valley where the palaeochannel pinches out on bedrock and the aquifer is less transmissive. Groundwater pumping is recommended to balance the excess groundwater replenishment if emergent salinity is apparent over the palaeochannel system. The objective of pumping should be to lower the watertable to at least 2 metres below ground surface, while minimising watertable drawdown below the drain levels. Unnecessarily lowering the watertable below the drains would increase the cost of pumping by drawing water through the pumping scheme that would have otherwise discharged effectively to the drains. The disadvantage of pumping from the palaeochannel aquifer to control local salinity formation is that multiple pumps and large pumping rates would be required to drawdown the watertable over a much larger area than required, including areas that are not at risk from salinity formation. It is recommended that the feasibility of deepening selected surface drains and constructing additional deep open drains be investigated. In areas where the palaeochannel system is absent and the aquifer does not drain effectively in response to pumping, subsurface drains are recommended to lower the watertable and reverse emergent salinity Martins Location and M1 Supply Channel Most of Martins Location and a narrow strip along the M1 supply channel to the south of Martins Location are likely to develop a high potential salinity risk in response to future irrigation and rainfall. A small area within the southeast margin of the irrigation area in Martins Location already has developed secondary salinity and there is evidence of lost productivity near the M1 supply channel. Martins Location is unlikely to be drained effectively using groundwater pumps because the palaeochannel system is absent and natural subsurface drainage is poor. The aquifer has small transmissivity and produces only low pumping yields. Removing excess groundwater from the aquifer is much more difficult in these areas and it is recommended that improved water use efficiency practices and good surface drainage design be implemented to minimise the net rate of groundwater replenishment from irrigation and rainfall. Subsurface drains are recommended to lower the local watertable and remove subsurface salt in areas with emergent salinity. Tree crops also may be effective at reducing the watertable elevation beneath the trees; however, this approach may not be sustainable in the long-term due to subsurface accumulation of salt. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 78

80 It is recommended that a subsurface drainage trial be conducted in Martins location or at a comparable site to determine the feasibility, effectiveness and design parameters for subsurface drains in the Ord Stage 1. The roles of irrigated and dryland trees in managing the aquifer water balance also should be investigated. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 79

81 Table 6-1: Summary of groundwater management recommendations Groundwater Management Area Packsaddle Plain Ivanhoe South and Central Ivanhoe North and Cave Spring Gap Martins Location and M1 Supply Channel Salinity Risk Hydrologic Controls Management Options Recommendations Moderate to high Lake Kununurra Palaeochannel Packsaddle Creek Bedrock high Low Palaeochannel Ord River Topography Moderate to high Palaeochannel Irrigation drains Reduced replenishment Lowered level in Lake Kununurra Groundwater pumping Subsurface drains Implementation of water use efficiency improvements Lower Lake Kununurra to the minimum acceptable level Groundwater pumps to balance excess replenishment from irrigation and rainfall with minimal watertable drawdown below Lake Kununurra Subsurface drains to reduce local shallow watertables Not considered Implementation of water use efficiency improvements Groundwater monitoring for rising watertable trends Reduced replenishment Deepening of surface drains Groundwater pumping Sub-surface drains Moderate to high Low transmissivity Reduced replenishment Subsurface drainage Implementation of water use efficiency improvements Groundwater pumps to balance excess replenishment from irrigation and rainfall with minimal watertable drawdown below the base of the irrigation drain network Deepening of selected irrigation drains New surface drains or subsurface drains to reduce local shallow watertables Implementation of water use efficiency improvements Subsurface drains to reduce local shallow watertables Review of surface drainage to maximise surface flows from rainfall and irrigation and minimise infiltration Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 80

82

83 APPENDIX A: Hydrogeology of the Ord Stage 1 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 82

84 3 Water and Rivers Commission (O Boy et al., 01) present the most recent description of the hydrogeology of the Ord valley. The Commission s report draws upon a series of hydrogeological investigations commenced in the 1960s that were undertaken by government agencies from Western Australia and Northern Territory. Important aspects of the conceptual hydrogeology within the Ord Stage 1 are briefly summarised below. The palaeo-ord valley is filled by an aggraded sequence of fluvial sediments that were generally deposited as an upward-fining sequence of gravel, sand, silt and clay. These sediments broadly constitute four hydrostratigraphic units that commonly inter-finger. From deepest to shallowest, the corresponding aquifer units are: Sand and gravel palaeochannel aquifer - moderate to very high permeability with bore yields in the range 5 to >25 L/s; Sand aquifer - moderate permeability with bore yields >5 L/s; Silt, sandy-silt and silty-sand aquifer - low permeability with bore yields in the range 0.5 to 5 L/s; and Clay and silt aquifer - low permeability with bore yields in the range 0.05 to 0.5 L/s. The Palaeozoic and Proterozoic bedrock that underlies the alluvial sediments has small permeability and is believed to have poor hydraulic connection to the overlying aquifers. Interaction between bedrock and the Ord Valley sediments was assumed negligible in this study. Prior to regulation, flow in the Ord River was seasonal and dominated by large floods during the wet season (Ruprecht and Rodgers, 1999). It is likely that the aquifers beneath Ivanhoe Plain and Packsaddle Plains were replenished mostly by infiltration of surface water during seasonal flow events and floods. Groundwater levels were not measured on Ivanhoe Plain until 1964 and on Packsaddle Plain until the early 1970s. Nevertheless, it has been reasoned that pre-development groundwater levels in the immediate vicinity of the Ord River were probably around the same level, or just below the level, of the riverbed. Direct groundwater replenishment from rainfall was considered to have been relatively minor prior to irrigation development (Banyard, 1983). There is no evidence of regional groundwater mounds or groundwater drainage toward the Ord River at that time. The soil unsaturated zone is thought to have rapidly absorbed and stored seasonal rainfall without significant deep drainage and runoff. Average monthly evaporation exceeds average monthly rainfall in all months except February, when average monthly rainfall is around ten percent greater than evaporation. Overall, annual pan evaporation exceeds annual rainfall by a factor of approximately three. Subsurface flow through Cave Spring Gap and toward Weaber Plain is though to have been the only significant route for groundwater drainage from the irrigation area prior to irrigation development. Major discharge boundaries on Packsaddle Plain were not apparent. It is likely that some groundwater drained back to the Ord River as a result of bank storage following seasonal flow events and flooding. Subsurface flows between river meanders may also have occurred when the river was in flow. 3 Now Department of Environment (DoE), Western Australian Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 83

85 Following river regulation and irrigation development in the1960s, the aquifers beneath Ivanhoe Plain and Packsaddle Plain began to fill as a result of increased replenishment from leakage from storage structures and conveyance channels, and deep drainage beneath irrigated fields. The groundwater level beneath Ivanhoe Plain is now above the level of the Ord River, which has become a groundwater discharge boundary for the aquifer. Pump tests of several bores (e.g., McGowan 1983 and O Boy 1997, 1998) have indicated that the upper silts and clays in the irrigation area can be successfully dewatered by pumping from the palaeochannel. Lateral movement of groundwater is mainly through the palaeochannel gravel and sand, with predominantly vertical movement of water in the overlying silts and clays. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 84

86

87 APPENDIX B: Hydrological Timeline Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 86

88 Construction and calibration of the Ord Stage 1 groundwater model was aided by the development of a hydrological timeline for the Ord valley (Appendix Figure B-1). The timeline indicates the sequence of hydrological events that have affected the irrigation area and the timing of hydrogeological investigations in relation to these events. An attempt was made to compile relevant data about the historical development of agriculture in the Ord Stage 1, including a schedule of what the spatial patterns of cropping and irrigation water use have been over the years at farm scale. Nevertheless, the amount of data available in readily useable form was limited and further work is required to establish a detailed, hydrological history of agricultural development. Key hydrological events include: construction of the Ord Diversion Dam at Kununurra and the formation of Lake Kununurra; commencement of irrigated agriculture on Ivanhoe Plain; completion of the Ord River Dam and formation of Lake Argyle; commencement of irrigated agriculture on Packsaddle Plain; and commissioning of the hydropower station below Lake Argyle. Other features include the increase in irrigation during the 1990s in association with an estimated 250 percent increase in the cropped area of Ivanhoe and Packsaddle Plains. Between 1990 and 1999, the cropped area was expanded from approximately 4,0 hectares to 11,000 hectares. This increase coincided with the increase in sugar cane production from less than 2,500 hectares in 1996 to around 7,000 hectares in 00 which was 64 percent of the total cropped area. B1. Rainfall The region has experienced an above-average-rainfall period since , and a particularly wet period from to Average annual rainfall (Appendix Table B-1) during the 10-year period from to was around 2 mm (%) greater than the average for the preceding years ( to ), and 172 mm (%) greater than the long-term average ( to 04 05). Between and 01 02, the average annual rainfall was 1154 mm per year, which was approximately 371 mm (47%) greater than the long-term average. Appendix Table B-1: Average annual wet season rainfall at Kununurra (Station 56) Period Average Annual Rainfall [mm/yr] to (45 years) to ( years) to (10 years) to (4 years) 1154 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 87

89 Appendix Figure B-1: Hydrological timeline for the Ord Stage 1 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 88

90

91 APPENDIX C: Ord Stage 1 Groundwater Flow Model Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 90

92 C1. Modelling Environment The Ord Stage 1 groundwater flow model was developed and implemented in the subsurface flow and transport simulation system FEFLOW (Diersch, 02). 4 ArcGIS software was used to prepare input data and for visualisation of the modelling results. The model specification are summarised below in Appendix Table C-1. Appendix Table C-1: Specifications and numerical settings for the Ord Stage 1 groundwater flow model Simulation system Type of model Type of aquifer Model layers Type of elements Number of elements Number of nodes Layer 1 Layer 2 Layer 3 Equation solver Time stepping Free surface top Free surface bottom FEFLOW 3-D, saturated flow Unconfined with moveable phreatic surface Three layers, four slices Six-node triangular prisms 1, 835 per layer (4,505 total) 68,2 per slice (272,880 total) Clay and silt (low transmissivity) Sand aquifer (moderate transmissivity) Sand and gravel aquifer (high transmissivity) Iterative Predictor-corrector AB/TR time integration scheme (max. time step size 100 days) Constrained at top slice if touching the top slice Unconstrained (watertable) if fallen dry at the bottom slice C2. Model Layers The major hydrostratigraphic units underlying the irrigation area were represented by three model layers (Appendix Figure C-5 to Appendix Figure C-7): Layer 1 - clay and silt with small permeability; Layer 2 - sand aquifer with moderate permeability; and Layer 3 - sand and gravel palaeochannel aquifer with moderate to large permeability. The basement bedrock profile was represented by the bottom surface of the model, which was impermeable. A three-dimensional rendering of the model solid volume is presented in Appendix Figure C-1. 4 ArcGIS is an integrated collection of GIS software produced by Environmental Systems Research (ESRI) see Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 91

93 Appendix Figure C-1: Solid volume of the Ord Stage 1 groundwater flow model Layer extents and thicknesses were estimated based on the Ord River Irrigation Area Hydrogeology map (O Boy et al., 01) and drilling logs supplied by the 5 Water and Rivers Commission. Digital topographic data and survey data of the irrigation supply channels and drains also were used in the construction of the model layers. These data were combined and interpolated to continuous surfaces using ArcGIS software, and then imported into FEFLOW as the basis for assigning the model layer elevations. FEFLOW required the model layers to extend continuously across the model domain; however, the hydrostratigraphic units were discontinuous in areas where the bedrock protruded through them. In these areas, the affected model layers were made absent by assigned a small layer thickness and appropriate aquifer properties to ensure that the correct hydrostratigraphic sequence was maintained. Appendix Table C-2 lists the rules used to assign the aquifer properties in these situations. Appendix Table C-2: Rules for simulating absent model layers Condition Layer 3 absent due to basement subcrop Layers 3 & 2 absent due to basement subcrop Rules for Assigning Aquifer Properties Set Layer 3 to small thickness Set Layer 3 properties to Layer 2 properties Set Layers 3 & 2 to small thickness Set Layer 3 properties to Layer 1 properties Set Layer 2 properties to Layer 1 properties 5 Now Department of Environment (DoE), Western Australian Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 92

94 C3. Finite Element Mesh The model domain was discretised into a mesh of 4,505 triangular elements (1,835 per layer) defined by 272,880 nodes (68,2 per slice). Element boundaries were aligned with key hydrologic features including the Ord River, Dunham River, Packsaddle Creek, Lake Kununurra, the Ord Diversion Dam and the aquifer hydrostratigraphic boundaries (Appendix Figure C-2). The irrigation channels network and channel inverts were built into the mesh top surface (Appendix Figure C-3). Mesh generation was performed using 6 Triangle (Shewchuk, 02). Appendix Figure C-2: Finite element mesh in the region of Lake Kununurra 6 Refer to and FEFLOW documentation Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 93

95 (a) (b) Appendix Figure C-3: Detail of the finite element mesh in the region of (a) the M1 offtake and (b) Packsaddle Creek Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 94

96 C4. Aquifer Properties C4.1. Values Determined from Pumping Tests O Boy et al. (01) reviewed pumping tests carried out to determine hydraulic properties of the ORIA aquifers, including pump testing of three bores by McGowan (1983) and two bores by O Boy (1997, 1998). More recently, Water Corporation (00) pump tested bores 10/00 and 11/00 on Ivanhoe North. A summary of the analyses of these tests is presented in Appendix Table C-3 and Appendix Figure C-4. Pump testing has mainly targeted the high permeability palaeochannel sediments, which have a hydraulic conductivity in the order of hundreds of meters per day. Values of this magnitude are consistent with literature values for gravel and sand-gravel mixes (Bouwer, 1978). Both the short-term and long-term pumping tests by O Boy (1997, 1998) provided reasonably consistent estimates of the palaeochannel transmissivity. Anisotropy of the sand and gravel aquifer system has not been quantified but vertical hydraulic conductivity is expected to be at least order-of-magnitude less than horizontal conductivity in layered sedimentary aquifers with inter-bedding. Estimates of the vertical hydraulic conductivity of the silts and clays are in the order of tensto-hundreds of millimetres per day and are characteristic of fine textured sediments (Bouwer, 1978). These deposits are likely to be highly anisotropic, and it is expected that horizontal conductivity may be at least order-of-magnitude larger than vertical conductivity. Banyard (1983) reported values of saturated vertical hydraulic conductivities measured in Cununurra Clay on Ivanhoe Plain of between 10 and 600 mm per day based on infiltrometer readings. C4.2. Values Used in Previous Modelling Studies The Ivanhoe groundwater system was modelled previously by SKM (1998b) but the Packsaddle Plain aquifer has not been modelled prior to the current study. The calibrated aquifer parameters from the SKM model of Ivanhoe Plain are listed in Appendix Table C-4. Information about vertical hydraulic conductivity and anisotropy in the model was not reported. C4.3. Values Used in This Study Calibrated aquifer parameters from the current study are reported in Appendix Table C-5. Hydraulic conductivity of the silt and clay is consistent with the values adopted by SKM but hydraulic conductivity of the palaeochannel aquifer is an order-of-magnitude larger. This is consistent with both the pump test results in Appendix Table C-3 and the results from the recent groundwater pumping trial on northern Ivanhoe Plain (Smith et al., 05). In particular, a large hydraulic conductivity was required to calibrate the rate of water movement from Lake Kununurra into Packsaddle Plain following the formation of the lake. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 95

97 Appendix Table C-3: Summary of pump test results Bore Location Test Aquifer Aquifer Trans. [m 2 /d] Aquifer Hydraulic Cond. [m/d] Aquitard Vertical Hydraulic Cond. [m/d] Storativity [-] Constant rate: PB1 Ivanhoe Palaeo. 6,000-7, KL/d Constant rate: PB2 Ivanhoe Palaeo KL/d Constant rate: PB3 Ivanhoe Palaeo KL/d PB4 PB4 Ivanhoe Ivanhoe 10/00 Ivanhoe 11/00 Ivanhoe PSPB1 Packsaddle Constant rate: 7.5 1,2 KL/d Long-term: 115 1,2 KL/d Constant rate: 3,000 KL/d Constant rate: 3,000 KL/d Constant rate: 8.5 1,613 KL/d Palaeo. 1,3-3,355 (2,0 mean) (0 mean) (0.01 approx.) 1.8E-5-1.4E-2 (5.0E-5 mean) Source McGowan (1983) McGowan (1983) McGowan (1983) O Boy (1997) Palaeo. 1,460-2, E-4-2.0E-2 O Boy (1998) Palaeo Palaeo. 1, Palaeo. (deep gravel) 1,158-5,179 (2,000 mean) (0 mean) (0.1 approx.) Water Corporation (00) Water Corporation (00) 4.0E-4 (mean) O Boy (1997) PSPB1 Packsaddle Constant rate: 8.5 1,613 KL/d Palaeo. (deep gravel) 1,158-5,179 (1,000 mean) (3 mean) E-2 (mean) O Boy (1997) PSPB1 Packsaddle Long-term: 115 days Palaeo. 2, O Boy (1998) Appendix Table C-4: Aquifer parameter values from the SKM (1998b) model of Ivanhoe Plain Aquifer Property Layer 1 Alluvial Clay & Silt Layer 2 Gravel & Sand Palaeochannel Hydraulic conductivity [m/d] 1-10 Specific yield Appendix Table C-5: Aquifer parameter values from the current study Aquifer Property Layer 1 Clay & Silt Layer 2 Sand Aquifer Layer 3 Palaeochannel Aquifer Horizontal conductivity [m/d] Vertical conductivity [m/d] Effective porosity [1] Specific storativity [1/m] Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 96

98 Dunham River Diversion Dam Ord River PB KL/d Screen: 6-m palaeochannel T : 6,000-7,000 m2/d (McGowan, 1983) Spring Creek PB ,2 KL/d days recovery Screen: 6-m palaeochannel T : 1,460-2,800 m2/d K : 5-4 m/d (O'Boy, 1998) PB ,2 KL/d Screen: 6-m palaeochannel T : 2,0 m2/d K : 0 m/d (O'Boy, 1997) Valentine Creek PB KL/d Screen: 6-m palaeochannel Leaky semi-unconfined T : 0 m2/d K : 14 m/d (McGowan, 1983) Packsaddle Plain Green Location > > > > > > > Lake Kununurra Martins Location Cave Spring Gap PB KL/d Screen: 6-m palaeochannel T : 1,700 m2/d K : 110 m/d (McGowan, 1983) PSPB1 8.5 KL/d Screen: 3-m shallow gravel Screen: 6-m deep gravel T : 2,000 m2/d (deep) T : 1,000 m2/d (shallow) K : 0 m/d (deep) K : 3 m/d (shallow) (O'Boy, 1997) 11/00 3,000 KL/d Screen: 10-m palaeochannel T : 1,100 m2/d K : 105 m/d (Water Corporation, 00) Ivanhoe Plain 10/00 3,000 KL/d Screen: 8-m palaeochannel T : 8 m2/d K : 68 m/d (Water Corporation, 00) > LEGEND Pumping bore > Pump test bore Surface drainage Supply channel Surface drain Irrigation area PSPB1 115 days days recovery Screen: 3-m shallow gravel Screen: 6-m deep gravel T : 2,000 m2/d K : 3 m/d (O'Boy, 1998) Ord River NAP - Improved Water Management in the Stage 1 Ord River Irrigation Area 800 IVANHOE PLAIN NORTH PUMPING TRIAL 800 Transverse Mercator; GDA 1994; MGA Zone Appendix Figure C-4: Summary of pump test locations and results Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 97

99 Appendix Figure C-5: Model layer 1 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 98

100 Appendix Figure C-6: Model layer 2 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 99

101 Appendix Figure C-7: Model layer 3 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 100

102

103 APPENDIX D: Pre-Irrigation Groundwater Conditions Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 102

104 D1. Pre-irrigation Groundwater Levels A pre-irrigation watertable map of the Ord Stage 1 was never produced. Groundwater levels in monitoring bores on Packsaddle Plain were first measured after construction of the Ord Diversion Dam and commissioning of the irrigation scheme in They were un-measured beneath Packsaddle Plain until Analysis of the bore hydrographs to determine the preirrigation groundwater level is problematic because many of the monitoring bores were installed adjacent to supply channels and drains and were affected by channel leakage relatively soon after irrigation began. In the present work, the position of the pre-irrigation watertable beneath Ivanhoe Plain was approximated by backward extrapolation to 1963 of selected hydrographs; this was done visually to the nearest meter. The results are presented in Appendix Table D-1. This method could not be applied to Packsaddle Plain because the watertable rose relatively quickly during the 1960s and reliable backward extrapolation from 1974 was not possible. Appendix Table D-1: Estimated pre-irrigation groundwater level in the palaeochannel aquifer beneath Ivanhoe Plain prior to irrigation development in 1963 Bore Name Easting (m) Northing (m) Estimated Water Level in 1963 (m) 10A C A B C E A C F G A B C C E F A CS CS Note: 1. Source data was provided by the Department of Environment, Map coordinates are in Map Grid of Australia (MGA); GDA Considerable spatial variability is evident in the watertable. The estimates of the pre-irrigation water level elevations varying by up to metres; they range between 12 and metres AHD. In the previous modelling of Ivanhoe Plain by SKM (SKM, 1998b), it was assumed that the pre-irrigation groundwater level was probably around 17 metres AHD throughout the irrigation area. This now seems unlikely. In particular, there is evidence of high pre-irrigation groundwater elevations along the eastern flank of the valley where the palaeochannel aquifer is absent and subsurface drainage is restricted by clay and silt sediments. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 103

105 D2. Simulated Watertable The steady state, pre-irrigation model was constructed primarily to generate appropriate initial conditions for the transient, post-irrigation model. The estimated pre-irrigation groundwater levels in Appendix Table D-1 are compared to the simulated pre-irrigation watertable in Appendix Figure D-1. The general groundwater flow direction was from south to north along the length of the Ord valley. River water replenished the aquifer along up-gradient sections of river meanders and groundwater discharged to the river along the down-gradient sections. On Ivanhoe Plain there was a net replenishment of groundwater from the Ord River, which was transmitted slowly through the palaeodrainage in a northeast direction through Cave Spring Gap. Total outflow through Cave Spring Gap was approximately 7 ML per year. On Packsaddle Plain, aquifer replenishment by river-water was approximately balanced by groundwater discharge back to the river. It is likely that Packsaddle Creek was not a groundwater discharge boundary prior to the filling of Lake Kununurra as the simulated groundwater level was more than 5 meters below the stream bed level. Overall, the aquifers appear to have been relatively poorly developed prior to irrigation and the annual groundwater replenishment was small (see Section D3.1 below). Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 104

106 Appendix Figure D-1: Estimated and simulated groundwater levels in palaeochannel monitoring bores under pre-irrigation conditions Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 105

107 D3. Boundary Conditions D3.1. Rainfall Replenishment Net groundwater replenishment from rainfall percolation and deep drainage was assumed to have been relatively small compared to present day rates (see APPENDIX A). The preirrigation watertable beneath some parts of Ivanhoe Plain may have been above the level of the Ord River, in particular along the flanks of the valley where the palaeochannel pinches out against outcropping bedrock. Nevertheless, small transmissivity in these areas suggests that only a very small amount of rainfall replenishment was required to sustain the watertable at the estimated pre-irrigation level. Without contrary evidence to indicate the existence of active groundwater discharge boundaries or leakage of groundwater into underlying bedrock aquifers, it was assumed that lateral transmission of groundwater to discharge boundaries and net long-term groundwater replenishment from rainfall were relatively minor under pre-irrigation conditions. The model was satisfactorily calibrated against the estimated pre-irrigation groundwater levels with a net rainfall replenishment of zero. This suggests that vertical infiltration following rainfall and river floods was approximately balanced by evapotranspiration from the soil zone. Applying even a very small net replenishment rate resulted in unrealistically high groundwater level along the flanks of the valley where transmissivity is small and groundwater could not drain effectively. Constant head conditions were employed outside of the active model area where the preirrigation watertable was below the estimated basement elevation to establish appropriate initial conditions for the post-irrigation simulations (see APPENDIX E). D3.2. Ord River First-kind ( prescribed head ) boundary conditions were assigned along the portions of the Ord River that were thought to have a significant groundwater connection to the aquifer. Nevertheless, these regions were determined more reliably during the transient model calibration (see APPENDIX E) and, therefore, the pre- and post-irrigation models were calibrated iteratively. The time-constant heads decreased in value linearly along the river from metres AHD at the southern end (Packsaddle Plain) to 10 metres AHD at the northern end (Ivanhoe Plain). This represented a total drop in hydraulic head of metres over an approximately 50-kilometre length of river, and equated to an average bed slope along the Ord River within the Ord Stage 1 of approximately (0.05 %). D3.3. Cave Spring Gap Where the model boundary cut across Cave Spring Gap, a prescribed head boundary condition of 12 metres AHD was assigned to match the estimate pre-irrigation watertable elevation at bore CS6 (see Section D1). Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 106

108

109 APPENDIX E: Post-Irrigation Groundwater Conditions Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 108

110 E1. Observed Groundwater Rise The history of post-irrigation groundwater rise in the Ord Stage 1 is discussed in the main report in Section 4. Bore hydrographs are presented in APPENDIX I and the bore locations are indicated in APPENDIX H. E1.1. Interpretation of Bore Hydrographs The rate at which the watertable rises in an unconfined aquifer depends upon both the imbalance between groundwater replenishment and discharge, and the aquifer storage properties. In general, the rate of watertable rise or fall reflects the volumetric accumulation of groundwater in the aquifer; however, the rate of watertable movement also varies according to the fillable and drainable porosities of the aquifer sediments. For example, the rate of watertable rise in sand and clay can be different even for the same volumetric change in aquifer storage. Thus, spatial heterogeneity in aquifer properties can complicate the interpretation of bore hydrographs. The storage properties of sediments also are nonhysteretic and commonly vary in time according to the antecedent hydrological conditions. Fine-grained sediments, such as silt, clay and fine sand, typically exhibit porosities in the range 0.4 to 0.6 (Bouwer, 1978); however, the fillable and drainable porosities can be appreciably smaller compared to porosity. Medium to coarse grain sand and gravel have porosities in the range 0.2 to 0.4, while sand and gravel mixes have porosities of 0.1 to 0.3. The measured rates of groundwater accession beneath the Ord Stage 1 reflect both the rate of groundwater accumulation and sediment heterogeneity. The separate affects of replenishment and porosity on the rate of watertable rise cannot be separated reliably because sediment heterogeneity and porosity are not known in sufficient detail. The post-irrigation model calibration was based on more than one-hundred bore hydrographs, which had various record lengths between 1963 and 04. Many of these bores were installed adjacent to the irrigation channel network and their hydrographs exhibited both local-scale responses to channel leakage and valley-scale responses to irrigation development. E2. Simulated Groundwater Rise The post-irrigation model was calibrated for the forty-year period from 1963 to 03 inclusive. The primary objective was to match the simulated watertable rise to the observed long-term trends. More than one-hundred hydrographs were used. Time series plots that compare the observed and simulated water levels in each of the calibration bores are presented in APPENDIX I. E3. Simulated Water Balance Model water balances for the years 1965, 1975, 1995, 01 and 03 are presented in Appendix Table E-1 to Appendix Table E-5. In 1965, the aquifers beneath Packsaddle Plain and Ivanhoe South were filling rapidly in response to the formation of Lake Kununurra. As the watertable beneath Packsaddle Plain rose, the hydraulic gradient between the lake and aquifer decreased gradually and the rate of filling slowed. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 109

111 By 1975, the leakage volumes from Lake Kununurra into Packsaddle Plain and Ivanhoe South were approximately equal. On Ivanhoe South, the leakage from Lake Kununurra was approximately matched by groundwater discharge to the Ord River below the Diversion Dam. Head-dependent leakage from the M1 supply channel decreased slowly over time as the aquifer filled beneath the central and northern parts of Ivanhoe Plain. In 01, net groundwater replenishment from irrigation and rainfall was approximately double the rate in 1995, reflecting the increase in cropping and irrigation during the 1990s, and the above average rainfall during 1998 to 01. The increase in groundwater replenishment during this period was balanced by increases in groundwater discharge to the surface environment (e.g., creeks and drains) and aquifer storage. Aquifer storage decreased for the first time during 02 and 03. This was in response to reduced groundwater replenishment by rainfall, and drainage to the irrigation channel network of groundwater mounds that formed during the previous wet years. Appendix Table E-1: Simulated water balance in 1965 Water Balance Component Ivanhoe Plain [GL/yr] Packsaddle Plain [GL/yr] Total [GL/yr] Net replenishment from irrigation & rainfall - Packsaddle Net replenishment from irrigation & rainfall - Ivanhoe south Net replenishment from irrigation & rainfall - Ivanhoe central Net replenishment from irrigation & rainfall - Ivanhoe north & Green Loc Net replenishment from irrigation & rainfall - Caves Spring Gap Net replenishment from irrigation & rainfall - Martins Location Replenishment from Lake Kununurra Replenishment from SP1 supply channel Replenishment from M1 supply channel Outflow through Caves Spring Gap Discharge to the Ord River - Ivanhoe central Discharge to the Ord River - Ivanhoe south Discharge to creeks, drains & ground surface - Packsaddle Discharge to creeks, drains & ground surface - Ivanhoe south Discharge to creeks, drains & ground surface - Ivanhoe central Discharge to creeks, drains & ground surface - Ivanhoe north & Green Loc Discharge to creeks, drains & ground surface - Caves Spring Gap Discharge to creeks, drains & ground surface - Martins Location Aquifer storage Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 110

112 Appendix Table E-2: Simulated water balance in 1975 Water Balance Component Ivanhoe Plain [GL/yr] Packsaddle Plain [GL/yr] Total [GL/yr] Net replenishment from irrigation & rainfall - Packsaddle Net replenishment from irrigation & rainfall - Ivanhoe south Net replenishment from irrigation & rainfall - Ivanhoe central Net replenishment from irrigation & rainfall - Ivanhoe north & Green Loc Net replenishment from irrigation & rainfall - Caves Spring Gap Net replenishment from irrigation & rainfall - Martins Location Replenishment from Lake Kununurra Replenishment from SP1 supply channel Replenishment from M1 supply channel Outflow through Caves Spring Gap Discharge to the Ord River - Ivanhoe central Discharge to the Ord River - Ivanhoe south Discharge to creeks, drains & ground surface - Packsaddle Discharge to creeks, drains & ground surface - Ivanhoe south Discharge to creeks, drains & ground surface - Ivanhoe central Discharge to creeks, drains & ground surface - Ivanhoe north & Green Loc Discharge to creeks, drains & ground surface - Caves Spring Gap Discharge to creeks, drains & ground surface - Martins Location Aquifer storage Appendix Table E-3: Simulated water balance in 1995 Water Balance Component Ivanhoe Plain [GL/yr] Packsaddle Plain [GL/yr] Total [GL/yr] Net replenishment from irrigation & rainfall - Packsaddle Net replenishment from irrigation & rainfall - Ivanhoe south Net replenishment from irrigation & rainfall - Ivanhoe central Net replenishment from irrigation & rainfall - Ivanhoe north & Green Loc Net replenishment from irrigation & rainfall - Caves Spring Gap Net replenishment from irrigation & rainfall - Martins Location Replenishment from Lake Kununurra Replenishment from SP1 supply channel Replenishment from M1 supply channel Outflow through Caves Spring Gap Discharge to the Ord River - Ivanhoe central Discharge to the Ord River - Ivanhoe south Discharge to creeks, drains & ground surface - Packsaddle Discharge to creeks, drains & ground surface - Ivanhoe south Discharge to creeks, drains & ground surface - Ivanhoe central Discharge to creeks, drains & ground surface - Ivanhoe north & Green Loc Discharge to creeks, drains & ground surface - Caves Spring Gap Discharge to creeks, drains & ground surface - Martins Location Aquifer storage Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 111

113 Appendix Table E-4: Simulated water balance in 01 Water Balance Component Ivanhoe Plain [GL/yr] Packsaddle Plain [GL/yr] Total [GL/yr] Net replenishment from irrigation & rainfall - Packsaddle Net replenishment from irrigation & rainfall - Ivanhoe south Net replenishment from irrigation & rainfall - Ivanhoe central Net replenishment from irrigation & rainfall - Ivanhoe north & Green Loc Net replenishment from irrigation & rainfall - Caves Spring Gap Net replenishment from irrigation & rainfall - Martins Location Replenishment from Lake Kununurra Replenishment from SP1 supply channel Replenishment from M1 supply channel Outflow through Caves Spring Gap Discharge to the Ord River - Ivanhoe central Discharge to the Ord River - Ivanhoe south Discharge to creeks, drains & ground surface - Packsaddle Discharge to creeks, drains & ground surface - Ivanhoe south Discharge to creeks, drains & ground surface - Ivanhoe central Discharge to creeks, drains & ground surface - Ivanhoe north & Green Loc Discharge to creeks, drains & ground surface - Caves Spring Gap Discharge to creeks, drains & ground surface - Martins Location Aquifer storage Appendix Table E-5: Simulated water balance in 03 Water Balance Component Ivanhoe Plain [GL/yr] Packsaddle Plain [GL/yr] Total [GL/yr] Net replenishment from irrigation & rainfall - Packsaddle Net replenishment from irrigation & rainfall - Ivanhoe south Net replenishment from irrigation & rainfall - Ivanhoe central Net replenishment from irrigation & rainfall - Ivanhoe north & Green Loc Net replenishment from irrigation & rainfall - Caves Spring Gap Net replenishment from irrigation & rainfall - Martins Location Replenishment from Lake Kununurra Replenishment from SP1 supply channel Replenishment from M1 supply channel Outflow through Caves Spring Gap Discharge to the Ord River - Ivanhoe central Discharge to the Ord River - Ivanhoe south Discharge to creeks, drains & ground surface - Packsaddle Discharge to creeks, drains & ground surface - Ivanhoe south Discharge to creeks, drains & ground surface - Ivanhoe central Discharge to creeks, drains & ground surface - Ivanhoe north & Green Loc Discharge to creeks, drains & ground surface - Caves Spring Gap Discharge to creeks, drains & ground surface - Martins Location Aquifer storage Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 112

114 E4. Initial Conditions Watertable elevations from the pre-irrigation model were imported as the initial conditions in When the post-irrigation model was adjusted during calibration, the same changes were made to the pre-irrigation model and the initial conditions were updated. E5. Boundary Conditions Boundary conditions were assigned to imitate the hydrological conditions associated with construction of the Kununurra Diversion Dam, the formation of Lake Kununurra, and the establishment of irrigated agriculture on Ivanhoe and Packsaddle Plains. A brief chronology of these events is presented in APPENDIX B. E5.1. Net Replenishment from Irrigation and Rainfall Net groundwater replenishment rates beneath the Ord Stage 1 were estimated during the model calibration based on the observed rates of water level rise in the calibration bores (APPENDIX I). Independent estimates of groundwater replenishment were impracticable for several reasons. Firstly, data on irrigation application rates during the past forty years were not available in a useable form and, secondly, if the data were available, it would be nontrivial to estimate the proportion of applied irrigation water that reached the watertable. George (1983) estimated that between 4 and 55 percent of applied irrigation water contributed to groundwater replenishment in Ord Stage 1. SKM (1998b) adopted a net replenishment rate of 6 percent of the applied irrigation water in their modelling study. Such estimates contain large uncertainty because deep drainage, which is only a minor term of the water balance, was estimated from uncertain estimates of water application rates, tail water drainage volumes, evaporation, and crop water requirements. These estimates are subject to large error that is typically the same order of magnitude and larger than the deep drainage term. A number of valley-scale groundwater replenishment zones were identified by grouping together monitoring bores with similar hydrograph characteristics. These zones are depicted in Appendix Figure E-1. Dividing the observed rate of groundwater rise by effective porosity provided an initial estimate of the groundwater replenishment rate. Thus, the assumed value of effective porosity was critical because increasing or decreasing this value implied a proportional increase or decrease in the groundwater replenishment rate. Sediment porosities have not been determined by field-testing but conceivably could vary by a factor 5. The calibrated groundwater replenishment rates are listed in Appendix Table E-6. The general increase in replenishment rates from around 1990 onward was consistent with increased cropping between 1990 and 00 (APPENDIX B); reduction in unsaturated zone thickness during this period (APPENDIX F, Section F1); and above average rainfall during 1998 to 01 (APPENDIX B, Section B1). Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 113

115 Appendix Table E-6: Calibrated groundwater replenishment rates Area ID Description Area (ha.) Period 1 Packsaddle irrigation area 2, , 3 Packsaddle non-irrigated area 3, Ivanhoe South 1, Ivanhoe Central 2, Northern Ivanhoe Plain leaky area 7 Dumas Lookout (basement outcrop) Cave Spring Gap 1, Replenishment Rate ( x10-4 m/d) Replenishment Rate (GL/yr) See Section in the main report Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 114

116 Area ID Description Area (ha.) Period 9 Ivanhoe North 5, Martins Location east Martins Location Sandalwood & Mahogany plantation Martins Location 2, Replenishment Rate ( x10-4 m/d) Replenishment Rate (GL/yr) Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 115

117 Appendix Figure E-1: Groundwater replenishment zones Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 1

118 E5.2. Lake Kununurra Lake Kununurra was simulated using first-kind ( Dirichlet ) boundary conditions around the perimeter of the lake and across the lakebed. The assigned lake stage (Appendix Figure E-2) varied initially because flow up-stream of Lake Kununurra was unregulated until the Ord River Dam was completed in 1972 (refer to Section of the main report) Lake Level (m AHD) Appendix Figure E-2: Simulated water level in Lake Kununurra E5.3. Ord River below Lake Kununurra Interaction between the Ord River and groundwater was simulated using third-kind ( Cauchy ) boundary conditions, which were assigned at boundary nodes where the river and aquifer were believed to be hydraulically connected. This type of boundary condition allows inflow and outflow of water across the riverbed at a rate that is proportional to the head difference between groundwater and surface water and the assigned transfer coefficient. The river stage and transfer coefficient were provided as inputs and the groundwater level and flow rate across the boundary were computed during the model simulation. River stage values and the calibrated transfer coefficients are listed in Appendix Table E-7. Complex interaction between surface water and groundwater is expected in the Ord valley in response to natural variation in riverbed sediments, pinching out of the hydrostratigraphic layers against bedrock subcrops, and incision of these layers by the river channel. In this context, values of the transfer coefficients used in the modelling should be thought of as representing the bulk hydraulic connection between the aquifer and river system at the valley scale. They do not reflect the permeability and extent of the riverbed sediment only. Appendix Table E-7: Calibrated transfer rates for the Ord River Region River Stage [m AHD] In-flowing Transfer Coefficient [10-4 1/d] Out-flowing Transfer Coefficient [10-4 1/d] Ivanhoe South Ivanhoe Central Ivanhoe North N.A. N.A. N.A. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 117

119 1975 (lake level 41 m) 1985 (lake level 41 m) 1994 (lake level 41 m) 1996 (lake level 41.6 m) 00 (lake level 41.6 m) 03 (lake level 41.6 m) Appendix Figure E-3: Simulated watertable rise above Lake Kununurra; red areas Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 1

120 E5.4. M1 and SP1 Supply Channels Surface water leakage from the M1 and SP1 supply channels was simulated using third-kind boundary conditions. The values of channel stage and the calibrated transfer coefficients are listed in Appendix Table E-8. Stage elevation between the channel off-take and relief was decreased by approximately 7 metres and 3 metres, respectively, along the M1 and SP1 based on the channel inverts. Watertable variation along the length of the M1 has been interpreted previously to indicate that the channel leaks at different rates, and that a watertable mound beneath the channel is a sign of high leakage. On the other hand, where the M1 traverses less permeable sediments along the eastern flank of the valley, a local groundwater mound is present. This is consistent with the proposition that lateral transmission of groundwater is poor in this area and may act to limit the amount of channel leakage. In contrast, where the channel traverses large-permeability sediments, the rate of surface water leakage into the aquifer is matched by subsurface transmission and a distinct mound does not form. Thus, the absence of a watertable mound beneath the M1 should not be interpreted, necessarily, to indicate low leakage in that area; and vice versa. Appendix Table E-8: Calibrated transfer rates for the M1 and SP1 supply channels Channel Channel Stage [m AHD] In-flowing Transfer Coefficient [10-4 1/d] Out-flowing Transfer Coefficient [10-4 1/d] M SP ( ) 10 ( ) 0 E5.5. Groundwater Discharge to the Channel Network, Creeks and Ground Surface These drainage components were simulated using boundary constraint conditions that are a feature of the modelling software FEFLOW (Diersch, 02). A detailed digital elevation model of the Ord Stage 1, incorporating the drain network and channel inverts, was developed for this purpose. If, during a simulation, the watertable reached the ground surface or bottom of a drain, then the model nodes at those locations were dynamically switched to first-kind boundary conditions with the head specified according to the ground elevation or drain invert. If the watertable subsequently fell below the constraint elevation, then the boundary condition was de-activated. Thus, the watertable could not rise above ground surface but could fall below it. Discharges to constraint-activated boundary nodes were tracked separately during the model simulations and are report as Discharge to creeks, drain & ground surface in the model water balances presented in Section E3. Appendix Figure E-4 shows in red areas where the watertable was constrained by ground surface and groundwater discharge to the surface environment was simulated. Groundwater drainage to the drain network in Ivanhoe North is evident from around the mid-1990s and corresponds to the exhaustion of the aquifer s storage capacity in that area. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 119

121 Appendix Figure E-4: Simulated groundwater discharge to the surface environment; red areas Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 1

122 E5.6.Packsaddle Creek Groundwater drainage and discharge to Packsaddle Creek was simulated using the technique described in Section E5.5 above. E5.7. Cave Spring Gap Subsurface outflow from Cave Spring Gap was simulated by a third-kind boundary condition. The external stage elevation was set equal to 11 metres AHD and the assigned transfer coefficient was 5 x 10-4 per day. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 121

123 APPENDIX F: Recent Groundwater Conditions Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 1

124 F1. Increased Responsiveness to Rainfall and Irrigation Reduced thickness of the unsaturated zone has accompanied the observed watertable rise beneath the Ord Stage 1. In Ivanhoe North the depth to watertable has decreased from approximately 15- metres in 1963 to less than 4 metres. This change has a number of implications for groundwater replenishment process in the valley, including: Shorter flow paths to the watertable for infiltrated rainfall and irrigation water; Shorter infiltration times to the watertable; and Reduced water storage capacity of the unsaturated zone above the watertable. It is expected that the raised watertable is now more responsive to rainfall and irrigation events than past watertables. Evidence of this effect is presented in Appendix Figure F-1, which depicts 6-hourly water level readings in monitoring bore PB3 on Ivanhoe North (see APPENDIX H for bore locations). Comparing seasonal responses between years, it is apparent that the magnitude of seasonal watertable variation has increased as the unsaturated zone has reduced in thickness. More generally, the water level in bore PB3 is now more variable at shorter time scales in response to rainfall and irrigation events; noting that to were above average rainfall seasons. Annual Wet Season Rainfall Jun - May (mm/year) Annula Rainfall Cumm. Deviation from Mean (mm) Jan-84 Jan-86 Jan-88 Jan-90 Jan-92 Jan-94 Jan-96 Jan-98 Jan-00 Jan-02 Jan-04 Dry period Note: 1. Mean is Dev. = Value - Mean Mean Wet Season Rainfall ( ) Wet period Jan-84 Jan-86 Jan-88 Jan-90 Jan-92 Jan-94 Jan-96 Jan-98 Jan-00 Jan-02 Jan Water Level Elevation (m AHD) Water Level Change (m) Start of dewatering trial Monitoring Bores: PB3 (auto) PB3 (man) 94/01 94/02 94/03 94/04 94/06 94/23 94/ 94/ 94/35 94/ Jan-84 Jan-86 Jan-88 Jan-90 Jan-92 Jan-94 Jan-96 Jan-98 Jan-00 Jan-02 Jan-04 Year Appendix Figure F-1: Water level readings beneath northern Ivanhoe Plain from 1984 to 04 Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 123

125 The change in seasonal watertable responses to rainfall and irrigation is highlighted in Appendix Figure F-2, which show the hydrograph for bore PB3 separated into wet season (November to March) and dry season (April to October) periods. Prior to 1996 the watertable remained relatively steady during the wet season but rose each dry season in response to irrigation. From the late-1990s, the seasonal watertable rise occurred during each wet season and the watertable dropped during the irrigation season as excess groundwater from the wet season replenishment discharged to the irrigation drains. Further examples from other parts of the irrigation area are presented in APPENDIX G. 8 PB3 - Ivanhoe North Water level change (m) Change of season Wet season Dry season Appendix Figure F-2: Wet-season-dry-season hydrograph separation for monitoring bore PB3 (1984 to 02) Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 1

126

127 APPENDIX G : Wet-Season-Dry-Season Hydrograph Separation The following plots show wet-season-dry-season hydrograph separations for selected monitoring bores in the Ord Stage 1. These bores were equipped with automatic water level loggers during the mid-1990s, and are maintained by the Department of Agriculture WA. The loggers provide daily (one per day) water level records. Missing daily records were filled by simple linear interpolation. No other data treatments or corrections were applied. Interpretation: Up-trending wet season (blue) curve aquifer fills during the wet; Down-trending wet season (blue) curve aquifer drains during the wet; Up-trending dry season (orange) curve aquifer fills during the irrigation season; and Down-trending dry season (orange) curve aquifer drains during the irrigation season. Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 1

128 Water level change (m) /01 - Ivanhoe North -2 Change of season Wet season -3 Dry season /02 - Ivanhoe North 4 Water level change (m) Change of season Wet season -3 Dry season /03 - Ivanhoe North 4 Water level change (m) Change of season Wet season -3 Dry season Water level change (m) /04 - Ivanhoe North -1-2 Change of season -3 Wet season Dry season Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 127

129 Water level change (m) /06 - Ivanhoe North -2 Change of season Wet season -3 Dry season / - Ivanhoe North (Scott Tree Farm) 4 Water level change (m) Change of season Wet season -3 Dry season /23 - Ivanhoe North (Scott Tree Farm) 4 Water level change (m) Change of season Wet season -3 Dry season / - Ivanhoe North 4 Water level change (m) Change of season -3 Wet season Dry season Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 1

130 Water level change (m) /25 - Ivanhoe North -2 Change of season Wet season -3 Dry season /27 - Ivanhoe North 4 Water level change (m) Change of season Wet season -3 Dry season / - Ivanhoe North 4 Water level change (m) Change of season Wet season -3 Dry season /37 - Ivanhoe North 4 Water level change (m) Change of season -3 Wet season Dry season Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 129

131 Water level change (m) PB4M1S - Ivanhoe North -2 Change of season Wet season -3 Dry season PB4M2S - Ivanhoe North 4 Water level change (m) Change of season -3 Wet season Dry season PB4M3S - Ivanhoe North 4 Water level change (m) Change of season Wet season -3 Dry season PB4M4S - Ivanhoe North 4 Water level change (m) Change of season -3 Wet season Dry season Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 1

132 94/33 - Martins Location (Mahogany blocks) Water level change (m) Change of season Wet season Dry season Water level change (m) / - Martins Location (Mahogany blocks) -2 Change of season Wet season -3 Dry season /35 - Martins Location (Mahogany blocks) Water level change (m) Change of season Wet season Dry season Water level change (m) Change of season Wet season Dry season 94/ - Martins Location (Mahogany blocks) Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 131

133 Water level change (m) Change of season Wet season Dry season 94/29 - Ivanhoe North (Mahogany blocks) Water level change (m) Change of season Wet season Dry season 94/ - Martins Location (Mahogany blocks) Water level change (m) Change of season Wet season Dry season 94/31 - Martins Location (Mahogany blocks) Water level change (m) Change of season Wet season Dry season 94/ - Martins Location (Mahogany blocks) Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 1

134 Water level change (m) /41 - Ivanhoe South -2 Change of season Wet season -3 Dry season Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 133

135 APPENDIX H : Calibration Bore Locations Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 1

136 Appendix Figure H-1: Calibration bores for the pre-irrigation model Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 135

137 Appendix Figure H-2: Calibration bores for the post-irrigation model Ivanhoe Plain Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 1

138 Appendix Figure H-3: Calibration bores for the post-irrigation model Packsaddle Plain Groundwater Management Options to Control Rising Groundwater Level and Salinity in the Ord Stage 1 Page 137