UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Changing with the flow : an analysis of water supply and demand in a subwatershed of the Okanagan Basin,.. Harma, Kirsten Joy 2010

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
24-ubc_2010_fall_harma_kirsten.pdf [ 2.05MB ]
Metadata
JSON: 24-1.0071203.json
JSON-LD: 24-1.0071203-ld.json
RDF/XML (Pretty): 24-1.0071203-rdf.xml
RDF/JSON: 24-1.0071203-rdf.json
Turtle: 24-1.0071203-turtle.txt
N-Triples: 24-1.0071203-rdf-ntriples.txt
Original Record: 24-1.0071203-source.json
Full Text
24-1.0071203-fulltext.txt
Citation
24-1.0071203.ris

Full Text

CHANGING WITH THE FLOW: AN ANALYSIS OF WATER SUPPLY AND DEMAND IN A SUBWATERSHED OF THE OKANAGAN BASIN, BRITISH COLUMBIA by Kirsten Joy Harma B.S. Western Washington University, 2001 A thesis submitted in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE in The Faculty of Graduate Studies (Resource Management and Environmental Studies) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  August 2010 © Kirsten Joy Harma, 2010  Abstract Surface water is critical for meeting water needs in British Columbia’s Okanagan Basin, but the timing and magnitude of its availability is being altered through climate and land use changes and growing water demand. WEAP, an integrated water management model, was used to consider future scenarios for water supply and demand in an unregulated and a reservoir-supported stream that supply the District of Peachland. Potential changes to the magnitude and timing of streamflow were evaluated in response to the following scenarios: (i) climate change (derived from the HadCM3 and CGCM2 GCMs for the 2020s and 2050s), (ii) a simulated prolonged drought, (iii) land cover change resulting from a Mountain Pine Beetle (MPB) outbreak, and (iv) combinations of these conditions. These changes, in combination with likely demand increases and reservoir operating rules were evaluated in terms of stress on water availability for human use and aquatic life. Results demonstrate that anticipated future climate conditions will critically reduce streamflow relative to demand (societal and ecological) in at least a few months of “normal” and “dry” years. On the unregulated creek, an earlier recession of peak spring snowmelt, accompanied by higher demands at the beginning of the summer outdoor watering season as early as the 2020s, reduced the ability to meet downstream needs. On the regulated stream, two “very dry” years under a climate change scenario resulted in deficits for municipal water users under a higher reservoir release scenario. In all scenarios, even with the higher flows expected under the MPB scenario, some combinations of demand, reservoir operations and climate variability resulted in less than optimal conditions for instream ecological flow needs. Beyond the implications for the District of Peachland, this work demonstrates a method of using an accessible modeling tool for integrating knowledge from the fields of climate science, forest hydrology, water systems management and stream ecology to aid in water and land management decision-making.  ii  Table of Contents Abstract........................................................................................................................... ii Table of Contents............................................................................................................iii List of Tables .................................................................................................................. v List of Figures ................................................................................................................vii Glossary and Acronyms ................................................................................................viii Acknowledgements......................................................................................................... x Chapter 1: Introduction.................................................................................................... 1 1.1) Problem Definition and Study Region................................................................... 1 1.2) Modeling for Water Resource Management ......................................................... 2 1.3) Research Objectives ............................................................................................ 4 1.4) Thesis Organization ............................................................................................. 4 Chapter 2: Background .................................................................................................. 5 2.1) Climate Change and Hydrology........................................................................... 5 2.1.1) Climate Change Projections for the Okanagan Basin .................................... 5 2.1.2) Climate Change Impacts on Snow Accumulation and Timing of Melt ............ 6 2.1.3) Climate Change Impacts on Evaporative Water Demand .............................. 6 2.1.4) Subsequent Impacts on Stream Flow ............................................................ 7 2.1.5) Drought, Implications Separate from “Climate Change” Alone....................... 8 2.1.6) Implications for Management ........................................................................ 9 2.2) Hydrologic Impacts of Land Cover Change ....................................................... 10 2.2.1) Links between Loss of Forest Cover and Hydrology.................................... 10 2.3) Human Demand................................................................................................. 14 2.4) Impacts of Hydrologic Change on Aquatic Ecosystems...................................... 15 Chapter 3: Methods ...................................................................................................... 17 3.1) Study Site Selection and Description of Study Area ........................................... 17 3.1.1) Existing Land Use/ Land Cover ................................................................... 18 3.1.2) Changes to Land Use/Land Cover .............................................................. 18 3.2) Model Description .............................................................................................. 19 3.3) Constructing Model to Represent Study Area..................................................... 21 3.3.1) Hydrological Response Unit Delineation...................................................... 21 3.3.2) Choice of Time Step.................................................................................... 25 3.3.3) Infrastructure ............................................................................................... 25 3.3.4) Demand Sites.............................................................................................. 26 3.4) Representing Climate Variables ......................................................................... 27 3.5) Model Calibration ............................................................................................... 29 3.5.1) Choice of Calibration and Validation Period................................................. 29 3.5.2) Calibration of Snow Parameters .................................................................. 32 3.5.3) Energy Parameter Specifications ................................................................ 32 3.5.4) Landscape Parameter Specifications .......................................................... 35 3.5.5) Calibration Adjustments in Peachland Creek Watershed............................. 37 3.5.6) Calibration and Validation Performance Metrics .......................................... 39 3.6) Representing Human Water Demand................................................................ 40 3.6.1) Data Availability and Limitations .................................................................. 40 3.6.2) Developing the Model.................................................................................. 40 3.6.3) Location of Water Use................................................................................. 41 3.6.4) Values Used in Activity-based Water Use Estimates ................................... 41 3.6.5) Values Used in Extraction-Based Water Use Estimates .............................. 44 3.6.6) Connecting Supply and Demand ................................................................. 45 3.7) Representing Instream Flow Targets.................................................................. 45  iii  3.8) Developing Scenarios ........................................................................................ 46 3.8.1) Climate Change .......................................................................................... 46 3.8.2) Drought and Climate Change ...................................................................... 48 3.8.3) Land Cover Change: Mountain Pine Beetle Attack ...................................... 49 3.8.4) Management: Water Demand..................................................................... 52 3.8.5) Management: Reservoir Operations........................................................... 54 3.9) Metrics of Change .............................................................................................. 55 Chapter 4: Results ....................................................................................................... 57 4.1) Calibration and Validation Results...................................................................... 57 4.2) Scenario Results................................................................................................ 62 4.2.1) Hydrologic Changes .................................................................................... 63 4.3) Interactions between Supply and Demand ......................................................... 76 4.3.1) Trepanier Creek .......................................................................................... 76 4.3.2) Peachland Creek......................................................................................... 83 Chapter 5: Analysis....................................................................................................... 95 5.1) Ranking of Impacts ............................................................................................ 95 5.1.1) Change in Annual Streamflow ..................................................................... 95 5.1.2) Volume of Deficit in District of Peachland .................................................... 96 5.1.3) Reduction in Reservoir Storage................................................................... 98 5.2) Sensitivity to Changes...................................................................................... 102 5.2.1) Timing of Sensitivities................................................................................ 102 5.2.2) Peachland’s Sensitivity to Peak Flow Changes ......................................... 103 5.2.3) Peachland’s Sensitivity to Changes in Demand vs. Supply ....................... 103 5.2.4) Instream Flow Target Sensitivity to Water Use .......................................... 104 Chapter 6: Discussion and Conclusion........................................................................ 106 6.1) Key Findings .................................................................................................... 106 6.2) Comparison to Other Studies ........................................................................... 112 6.2.1) Climate Change ....................................................................................... 112 6.2.2) Land Cover Change .................................................................................. 113 6.2.3) Drought ..................................................................................................... 114 6.3) Water Supply and Demand Estimates in Context ............................................. 115 6.3.1) Reservoir Operations ................................................................................ 115 6.3.2) Demand Estimates .................................................................................... 115 6.3.3) Instream Flows.......................................................................................... 116 6.3.4) Lessons from Model Calibration ................................................................ 117 6.4) Hydrologic Modeling: Limitations and Advantages of the Methods ................... 118 6.5) Future Work ..................................................................................................... 120 6.6) Conclusion ....................................................................................................... 121 References ................................................................................................................. 122 Appendices ................................................................................................................. 130 Appendix A) Formulas............................................................................................. 130 Appendix B) Data Used in Water Use Estimates, Baseline ..................................... 131 Appendix C) Increases in Demand in Future Scenarios .......................................... 133 Appendix D) Streamflow in Relation to Instream Flow Targets ................................ 135  iv  List of Tables Table 1: Range of impacts to snow from land cover change relative to a mature forest ............12 Table 2: Range of changes in stream hydrology expected from land cover changes ................13 Table 3: HRU specifications ......................................................................................................24 Table 4: Data availability by year ..............................................................................................30 Table 5: Precipitation regime for the calibration and validation years ........................................31 Table 6: Initial increase in Rnet_other to account for HRU-specific conditions ..........................34 Table 7: The two types of water use estimates used in this study .............................................40 Table 8: Water use activity by demand site ...............................................................................41 Table 9: Example of how the amount of lodgepole pine in an HRU translates to additional precipitation under the MPB scenario (example assumes a medium-elevation HRU) ...............50 Table 10: Reduction in Kc and corresponding reduction in AET ................................................51 Table 11: Calibration and Validation: Performance Metrics .......................................................57 Table 12: Measures of model performance in other studies using WEAP .................................61 Table 13: Scenario names and descriptions..............................................................................62 Table 14: Snowpack summary statistics for the 2020s climate change scenarios, compared to baseline ....................................................................................................................................63 Table 15: Annual streamflow summary statistics: Trepanier Creek, measured above the water withdrawals, for the 2020s climate change scenarios compared to baseline .............................64 Table 16: Monthly flow statistics: Trepanier Creek, measured above the water withdrawals, for the 2020s climate scenarios compared to baseline ...................................................................65 Table 17: Snowpack summary statistics for the 2050s climate change scenarios compared to baseline ....................................................................................................................................67 Table 18: Annual streamflow summary statistics: Trepanier Creek, measured above the water withdrawals, for the 2050s climate change scenarios compared to baseline .............................67 Table 19: Monthly flow statistics: Trepanier Creek, measured above the water withdrawals, for the 2050s climate change scenarios compared to baseline ......................................................68 Table 20: Total annual streamflow over the course of the drought ............................................70 Table 21: Snowpack summary statistics for the Mountain Pine Beetle scenarios compared to baseline ....................................................................................................................................71 Table 22: Annual streamflow summary statistics: Trepanier Creek, measured above the water withdrawals, for MPB scenarios compared to baseline..............................................................71 Table 23: Monthly flow statistics: MPB Only and MPB + 2020s scenarios.................................72 Table 24: Comparison between maximum snowpack and streamflow data, MPB scenarios, climate scenario, and baseline ..................................................................................................74 Table 25: PET and AET for a low elevation, mid elevation and high elevation HRU affected by pine beetle damage compared between baseline, MPB only and MPB + 2020s scenarios .......75 Table 26: Unmet demand in Peachland System 1 under the 2020s HadCM3 scenario, extraction-based water use estimate.........................................................................................76 Table 27: Unmet demand in Peachland System 1 under the 2050s CGCM2 scenario ..............78 Table 28: Unmet demand in Peachland System 1 under the 2050s HadCM3 scenario .............78 Table 29: Unmet demand in Peachland System 1 under the Drought + 2020s scenario ...........80 Table 30: Changes in carry-over storage each year for all scenarios, using the lower initial conditions and higher extraction rates.......................................................................................83 Table 31: Unmet demand in Peachland System 3 under the 2020s HadCM3, higher reservoir extraction scenario....................................................................................................................84 Table 32: Unmet demand in Peachland System 3 under the 2050s CGCM2, higher reservoir extraction scenario....................................................................................................................87 Table 33: Unmet demand in Peachland System 3 under the 2050s HadCM3 higher reservoir extraction scenario....................................................................................................................88 Table 34: Number of days when zero flow in Peachland Creek during a “normal”, “very wet” and “very dry” year under the 2050s CGCM2 scenario, both water use and reservoir releases .......90  v  Table 35: Number of days when zero flow in Peachland Creek during a “normal”, “very wet” and “very dry” year under the 2050s HadCM3 scenario, both water use and reservoir releases ......90 Table 36: Unmet demand in Peachland System 3 under the Drought + 2020s, high reservoirextraction scenario....................................................................................................................92 Table 37: Number of days when zero flow in Peachland Creek. Drought under climate change (2020s CGCM2), both water use and reservoir release scenarios.............................................92 Table 38: Change in average annual streamflow ranked by scenario........................................95 Table 39: Ranking of total volume deficit (thousand m3) in Peachland System 1 across the 9year scenario run ......................................................................................................................96 Table 40: Ranking of total volume deficit (m3) in Peachland System 3, high reservoir release, across the 9-year scenario run..................................................................................................97 Table 41: Proportion of the 9 scenario years with reduction in carry-over storage, ranked by total carry-over storage under all management conditions. ...............................................................98 Table 42: Ranking of reduction in number of days instream flow targets met under scenario conditions compared to baseline ...............................................................................................99 Table 43: Ranking of reduction in number of days instream flow targets met under scenario conditions compared to baseline. ............................................................................................100 Table 44: Number of days where the instream flow targets on Trepanier Creek were met under baseline conditions, but not under scenario conditions............................................................101 Table 45: Change in number of days instream flow target were met each month....................102 Table 46: Difference in number of days meeting instream flow targets between extraction-based and activity-based water use estimate ....................................................................................105 Table 47: MPB Only and MPB + 2020s: Difference in meeting instream flow targets between extraction-based and activity-based water use estimate. ........................................................105 Table 48. Summary of hydrologic response to two global climate models downscaled to the study area for the 2020s and 2050s periods. ..........................................................................106 Table 49: Metrics of change in stream hydrology in the Drought + 2020s and and Mountain Pine Beetle attack scenarios compared to baseline conditions .......................................................107 Table 50: Summary of impact of higher water demand on deficits in Peachland .....................109 Table 51: Cases where targets were fully met November to March, but met less fully May to September. .............................................................................................................................111  vi  List of Figures Figure 1: Study area location in British Columbia ......................................................................17 Figure 2. WEAP functional schematic ......................................................................................20 Figure 3: Trepanier and Peachland Watersheds Map. HRU boundaries overlain on biogeoclimatic zones.................................................................................................................23 Figure 4: Representation of Study Area in WEAP .....................................................................27 Figure 5: Temperature and wind speed in 1984 and likely additional energy input ....................33 Figure 6: Energy available for snowmelt as represented in WEAP............................................34 Figure 7: Water use by irrigation as a percentage of total use by agriculture (annual)...............42 Figure 8: Monthly apportionment of annual water use summed for all water users: activity-based water use estimate....................................................................................................................43 Figure 9: Monthly apportionment of annual water use from Peachland Creek: extraction-based estimate. ...................................................................................................................................44 Figure 10: Conservation flow targets (MAD) for Trepanier and Peachland Creeks, and BCIFN target for Trepanier Creek.........................................................................................................45 Figure 11: Temperature change data (°C) and precipitation change data (% change from baseline) from the 2050s time slice as downscaled to the study area. ......................................48 Figure 12: Calibration results from Trepanier Creek (1984 – 1986) ...........................................58 Figure 13: Validation results from Trepanier Creek (1991 – 1993) ............................................58 Figure 14: Validation Results, Trepanier Creek (1983 – 1993) ..................................................59 Figure 15: Calibration Results, Greata Creek (1973-1976)........................................................59 Figure 16: Validation Results, Greata Creek (1983-1986) .........................................................59 Figure 17: Comparison between Trepanier Creek flow and demand in the first through fourth year of the simulation for the scenario 2050s HadCM3 activity-based and extraction-based water use estimates ..................................................................................................................79 Figure 18: Comparison between flow and demand during three drought years and two subsequent years under the drought under climate change (2020s CGCM2) scenario, Trepanier Creek: activity and extraction-based water use estimate...........................................................81 Figure 19: Peachland Reservoir: Inflow, storage volume and pre-set reservoir releases under the 2020s HadCM3 higher reservoir extraction scenario, activity-based water use estimate.....84 Figure 20: Reservoir volume, inflow to reservoir and set reservoir release, 2050s CGCM2 lower initial conditions, higher reservoir release scenario, extraction-based water use estimate.........86 Figure 21: Reservoir volume, inflow to reservoir and release for Peachland, 2050s CGCM2 higher initial conditions, lower release scenario, extraction-based demand estimate ................86 Figure 22: Reservoir volume, inflow to reservoir and release for Peachland, 2050s HadCM3 lower initial conditions, higher release scenario, extraction-based demand estimate ................87 Figure 23: Reservoir volume, inflow to reservoir and release for Peachland, 2050s HadCM3 higher initial conditions, lower release scenario, extraction-based demand estimate ................88 Figure 24: Streamflow above and below withdrawal relative to demand. Scenario: 2050s CGCM2, lower reservoir releases and extraction-based water use ...........................................89 Figure 25: Reservoir volume, inflow to reservoir and set reservoir release during Drought + 2020s, high reservoir release scenario, extraction-based water use estimate ...........................91 Figure 26: Reservoir volume, inflow to reservoir and set reservoir release during Drought + 2020s, low reservoir release scenario, extraction-based water use estimate ............................91 Figure 27: Monthly precipitation change, as percent, for the 2020s HadCM3 scenario and the 2050s CGCM2 scenario............................................................................................................97  vii  Glossary and Acronyms 7-Day Minimum: For a given year, the lowest seven day average of a given value occurring during that year. Used to describe extreme conditions in streamflow that may indicate environmental stress. 7-Day Maximum: For a given year, the highest seven day average of a given value occurring during that year. Used to describe extreme conditions in streamflow that may indicate environmental stress. Ablation: Processes leading to loss of snow. Includes evaporation, sublimation and melt. Activity-based water use estimate: Water use estimate based on discrete values for water use by activity: indoor domestic, outdoor domestic, agricultural, commercial/industrial, and parks. Albedo: A measure of how strongly an object reflects a light source. Is applied to the reflectivity of snow in the context of this report. Dimensionless. BCIFN: British Columbia Instream Flow Needs. Developed by Hatfield et al. (2003) and represents flows needed to support all stream ecosystem functions. Calibration: The process by which the match between a measurement of a known quantity and a modeled approximation of that quantity is maximized. CGCM2: Coupled Global Climate Model version 2. A global climate model designed by the Canadian Centre for Climate Modelling and Analysis. cms: Cubic Meters per Second. Used as a measure of stream discharge. Demand: Volume of water requested by a user or activity: includes consumptive demand for agriculture or urban water supply and non-consumptive demand for ecosystem production. Distributed Hydrologic Model: A grid-based hydrologic model where climate input and land-based processes are calculated for each grid cell before being routed to a stream or groundwater. ECA: Equivalent Clearcut Area. The land area in a watershed that is “like a clearcut” in terms of lack of forest canopy. Used in studies of the hydrologic consequences of forest canopy loss. Extraction-based water use estimate: Water use estimate based on measured use from each water source (Peachland Creek, Trepanier Creek, Okanagan Lake) from 1999 to 2002, or a water license.  viii  Freshet: Flood resulting from spring snowmelt. Usually synonymous with peak flow for areas that receive significant snowfall. GCM: Global Climate Model (also refers to the similar General Circulation Models). HadCM3: The Hadley Centre for Climate Prediction and Research Global Circulation Model, version 3. HRU: Hydrologic Response Unit. A representation of a land area that is likely to have similar hydrologic properties throughout. IHA: Indicators of Hydrologic Alteration. Software developed by The Nature Conservancy to assess the impact of hydrologic alteration on a set of indicators relevant to biological health in streams. Used in this study to statistically compare baseline streamflow conditions to conditions under the different scenarios. Instream Flow: Also known as “environmental flow,” refers to the amount of flow, both maximums and minimums, necessary to meet the life-stage requirements of fish and aquatic macroinvertebrates in streams, as well as maintain channel structure. MAD: Mean Annual Discharge. The mean of monthly streamflow discharges, over the course of a year. Used to determine flow thresholds under which less than optimal conditions for fish will result. MPB: Mountain Pine Beetle (Dendroctonus ponderosae) is a species of bark beetle that bores through pine bark, resulting in eventual death of the tree. The range and severity of Mountain Pine Beetle attacks has increased in British Columbia in recent years. Peak Flow: The term “peak flow” can refer to the single largest flow event recorded in a year, the largest flow over a 7-day period, and the total volume of water delivered over the high-flow season. Rnet_other: The “other energy” variable available for use in WEAP. Represents energy available for snowmelt besides direct solar radiation: advective, latent and sensible heat. Semi-Distributed Hydrologic Model: A model that uses areas of similar hydrologic response, or HRUs, to represent watershed characteristics. WEAP: Water Evaluation And Planning system. The water resource management model used to integrate hydrological data and test different scenarios in this study.  ix  Acknowledgements Primarily, I would like to thank my advisor Mark Johnson for bravely taking me on as his first student and subsequently providing me with unfailing encouragement and technical support, always with patience and a smile. I would also like to thank Stewart Cohen for serving on my committee and the great community of professors at UBC who were happy to meet with me to answer questions about hydrology, forestry, climate modeling, soil science, or whatever subject I needed guidance on. Additionally, I would like to thank the many professionals throughout B.C., too numerous to list here, who graciously gave me their time to answer my questions or provide me with data. I hope to be able to pass on their goodwill when I develop enough expertise that students begin to ask for my help. I would like to thank the University of British Columbia and the Center for Interdisciplinary Studies for much appreciated financial support, and acknowledge the contribution of a CFI grant in support of GIS related analyses. Additionally, I would like to express gratitude for the many opportunities I have had as a graduate student at IRES, the opportunity to live in beautiful Vancouver, and for being part of a great RMES cohort who challenged me, inspired me, and encourage me to get out and have fun. A big thanks to Don, for helping me to see my “story.” Through the writing process I was also thankful for my Mom’s help editing all my papers in high school, as it makes writing that much easier at present. Finally, I would like to thank my parents for supporting me on whatever endeavor I’ve chosen to pursue, and, who although they would really just like me to get a job and settle down, conceded that taking on the Master’s was a task worthy of forgoing settling for awhile.  x  Chapter 1: Introduction 1.1) Problem Definition and Study Region Communities in Central British Columbia’s Okanagan Basin face multiple challenges for sustainably meeting their water needs. Changes to the quantity and timing of the streamflow used for water supply are occurring as a result of land use changes and may be exacerbated by global climate change. Simultaneously, population and economic activity are growing in these areas, creating greater demand for water resources overall, with demand peaks currently occurring during the late summer outdoor watering and crop irrigation season (Cohen, Neilsen & Welbourn, 2004). In light of these changes, trade-offs will be necessary between meeting municipal and irrigation demands without depleting a water supply that is also needed to support aquatic ecosystems. Annually replenished surface water sources are critical for meeting the Okanagan’s water 2  needs. All water used in the 8000 km Okanagan Basin (Figure 1), whether extracted from streams, reservoirs, groundwater or Okanagan Lake, originates as snowfall or rain falling within the basin. Sixty-seven percent of the water used in the Okanagan comes from surface water sources with the majority extracted from high-elevation storage reservoirs or directly from streamflow (Okanagan Basin Water Board (OBWB), 2010). Most communities do not have other options for developing new sources to meet their primary water needs (R. Hrasko, Water Supply Association of B.C., personal communication 6/8/2009). A limited amount of water is extracted from Okanagan Lake, since taking more than what is renewed annually by streamflow would result in “mining” the lake water. For these reasons, supporting future water use needs will require appropriately managing surface waters and preparing to adapt to change. st  By the mid 21 century global climate change is predicted to cause hydrological changes in the Okanagan Basin, including earlier snowmelt, more precipitation in the form of rainfall, and reduction in the volume of annual and spring flows (Merritt et al., 2006). Demand for water is highest in the late summer and early fall, precisely when less water is available; this low-water period is predicted to become even more stressed under climate change (Merritt et al., 2006). The fact that the watersheds provisioning community water supplies are also mixed-use watersheds means that changes to the landscape could alter the quantity and timing of water availability as well. Loss of forest cover, accelerating due to attack by the Mountain Pine Beetle (MPB) and subsequent salvage logging, are predicted to increase stream flows and lead to greater discharge peaks earlier in the spring (Grainger, 2009; Huggard & Lewis, 2007). The interactions of climate change, beetle attack and logging could also increase the risk of wildfire, thus further changing the landscape and hydrology. While there are many predictions for what types of impacts to hydrology these changes could bring, there is inherent uncertainty about what change they will bring. No amount of study will provide the answers for precisely what will happen. However, it is important that resource managers and  1  stakeholders be prepared for the types of changes that are considered likely, as well as be aware of changes that are considered possible. This study was undertaken in order to contribute to the dialogue on adaptation to changing climate and land use as they impact water resources.  1.2) Modeling for Water Resource Management Hydrologic models are used to answer questions when environmental, physical, or social variables cannot be manipulated directly. They are used for understanding why a process may be occurring, and to answer “what if” questions about potential future conditions (Loucks, Stedinger, & Haith, 1981). The fields of climate science, forest hydrology, stream hydrology and water management each have models tailored to their own fields of investigation, each with their own strengths and limitations. Coupled global climate models (GCMs) rely on scenarios about potential atmospheric CO2 levels and resulting climatic impacts, and are currently designed for making climate change projections at the scale of hundreds of kilometers (Le Treut et al., 2007). Forest hydrology modeling is based on linking data collected from stand-level measurements with an understanding of physical processes and can be used to predict the impacts of forest management at a watershed scale (Beckers, Smerdon, & Wilson, 2009). Water resource modeling typically considers what happens in a watershed as “boundary conditions” and focuses on manipulation and allocation of water once it is in a form that can be channeled to use for hydropower, municipal demand or for irrigation (Loucks, Stedinger & Haith, 1981). Yet land, climate, water and the human agent are all linked and interacting simultaneously. In order to model how changes in each variable affect the system as a whole it is necessary to link these processes together. What are referred to as “watershed” models and “Decision Support Systems for integrated water resource management” models have the ability to do just that (Georgakakos, 2007; Singh & Woolhiser, 2002). Like all models, these are designed to meet specific objectives, with watershed models focusing more on hydrological processes, and water management models focusing more on decision-making. The model chosen for use in this study is the Water Evaluation And Planning system (WEAP), created by the Stockholm Environment Institute (1997). WEAP integrates both hydrologic modeling and a decision support system, allowing for exploration of questions related to climate, hydrology and water management. WEAP simulates water’s movement through bio-physical systems in a semi-distributed hydrologic model and then accounts for how stream and ground water is stored, allocated and delivered for human consumptive purposes and subsequent impacts on aquatic ecosystems (Yates Sieber, Purkey, & Huber-Lee, 2005). WEAP also accounts for water use patterns by including economic, technological and demographic variables. Importantly, the WEAP code is implemented in a graphical user interface, so its use is not limited to computer programmers, allowing resource managers to model and evaluate scenario analyses specific to their needs. The potential utility of WEAP in application in this study is demonstrated by the fact that it has been applied in other areas facing challenges similar to the Okanagan. It has been applied extensively in California’s Sierra Nevada region to examine the impacts of climate change on  2  downstream water users (Young et al., 2009), tradeoffs between agriculture and instream flow needs (Yates, Purkey, Sieber, Huber-Lee, & Galbraith, 2005) and to explore options for how water purveyors can meet water needs over the next century (Huber-Lee et al., 2006). Recently, WEAP was applied on the U.S. side of the Okanagan Valley (where it is called the Okanogan Valley) to examine the trade-offs between agriculture and instream flow needs (Donley, 2010). Other modeling tools and platforms have been used to study water systems in the B.C. Okanagan Basin, all with unique sets of objectives and constraints. The UBC Watershed Model, originally developed by Quick and Pipes (1972) and later parameterized for the Okanagan Basin by Merritt et al. (2006) is designed to generate streamflow forecasts in snow-dominated watersheds. It focuses on the supply side of water resource questions (e.g. hydrological processes) and, as in the Merritt (2006) study, is applicable to studies of the influence of climate change on hydrology. Recently, a study supported by the Okanagan’s water governance body, the Okanagan Basin Water Board, has modeled water supply using a physically-based runoff generation model, MIKE-She (Summit Environmental, Inc. & DHI Water and Environment, 2009) and linked it with the Okanagan Water Budget Model (DHI Water and Environment, 2009) to determine the relationships between supply and demand projections. Output data include tributary flows into Okanagan Lake but do not include details on conditions within the subwatersheds. Work by Langsdale, Beall, Carmichael, Cohen, & Forster, (2007) used the “Stella” system dynamics model to link supply and demand information with the primary purpose that it be a decision-making tool, and, again, applied broadly to the Okanagan as a whole. WEAP differs from the models described above in that it includes a model for generating hydrological predictions and links supply with demand in a single modeling tool. Like the work of Langsdale et al. (2007), it can be used to help support decision-making, but with the added advantage of being specifically designed for water resources. This application of WEAP differs from the projects mentioned above in that it focuses on a single subwatershed in the Okanagan, includes a hydrological modeling component capturing the topographic variability of the study watershed at a finer spatial scale, runs on a daily time step rather than monthly or weekly, and incorporates study area-specific water demand estimates rather than generalized estimates from the basin as a whole. With the advantages of this modeling effort in mind, it should be considered that, while integrating many types of information within a single tool is a powerful way to understand interconnections, complexity comes at the cost of specificity. In general, when models have a dual purpose of facilitating decision-making and advancing scientific research, tradeoffs on each end must be made. Ivanovic & Freer (2009) summarize these tradeoffs as those between (i) political design, which supports the use of the simplest model (typically an empirical model), and (ii) a scientific design, which aims to attain the right answers for the right reasons – usually best achieved through a realistic approach (typically a physically-based model). Where there are losses from using one of these types of models, there are potentially gains from using the other. This thesis project is an attempt to find a balance between accuracy and pragmatism. Current scientific literature has been reviewed with the aim of accurately representing climatic,  3  hydrologic and landscape processes while also keeping representation of those processes simple enough to provide output that is useful for answering management questions. This effort aims to address questions that can only be answered by using a multi-purpose modeling approach with the intent of advancing understanding of the interactions between human and natural systems.  1.3) Research Objectives The aim of this study is to examine various scenarios for changes to water supply and demand in a subwatershed of the Okanagan Basin. The first study objective is to identify potential changes to the magnitude and timing of stream flow in response to global climate change, a prolonged climatic drought, land cover change resulting from a Mountain Pine Beetle outbreak, and interactions resulting from combinations of these conditions. The second objective is to determine which of these conditions lead to the most stress on human water needs and aquatic life. The objectives will be met by application of an integrated water resource management model and involve the following steps: •  Develop the model to represent the specifications of the study area  •  Calibrate the model to measured streamflow and snowpack records  •  Downscale global climate models to the study area  •  Represent hydrologic changes resulting from forest canopy loss  •  Determine current water demand and potential future demand given population growth and climate-related water use increases  •  Quantify and compare changes to streamflow resulting from the supply change scenarios and implications for water storage systems, aquatic life and human water use needs  1.4) Thesis Organization Chapter 2 provides background on how climate change and land cover change affect surface water hydrology, with specific reference to British Columbia and the Okanagan Basin. The chapter also introduces the factors which may influence future water demands and contribute to or reduce water stress, and then describes current methods in place for assessing the impacts of hydrologic changes on aquatic ecosystems. Chapter 3 provides details on the methods used in this study, including: model description, model specification to the study area, calibration procedures, scenario development and metrics used to assess change. Chapter 4 presents results of hydrologic changes from the scenarios compared to baseline and impacts on downstream water users. The scenarios are examined in aggregate in Chapter 5. The different scenarios are ranked by degree of impact on human and aquatic water users and a summary is provided of patterns present in multiple scenarios. The thesis concludes in Chapter 6 with a summary and discussion of research findings, and suggestions for future work.  4  Chapter 2: Background 2.1) Climate Change and Hydrology The Intergovernmental Panel on Climate Change (IPCC) projects that water quantity in North America will be particularly sensitive to a changing climate (Field et al., 2007). Changes in climate in North America are likely to impact total flows and flow seasonality with implications for ecosystems and how water managers can meet multiple objectives for water (Field et al., 2007). These changes will not happen in isolation, but will occur in combination with changes to landscapes, vegetation composition, and water use. The following review of literature looks at how climate change may affect water supply in the Okanagan Basin with a view towards understanding how climate change might interact with other changes occurring in the region.  2.1.1) Climate Change Projections for the Okanagan Basin A changing climate is not just an expectation for the future of the Okanagan Basin, but has already been occurring over the past century. An analysis of trends in temperature at a valley climate station showed an increase in daily maximum temperature during the winter of 2.4°C over the past century, with increases in daily maximums and minimums between 1 and 2°C in the other seasons (Taylor & Barton, 2004). An analysis of precipitation showed statistically significant increases in spring and summer, with slight decreases in the fall but no significant change in winter (Taylor & Barton, 2004). The trend of warmer temperatures is predicted to continue in winter and summer (Cohen &Kulkarni, 2001) with the trend of increasing precipitation continuing in the winter but not in the summer (Taylor and Barton, 2004). Determining precisely when and by how much the climate is expected to change in the Okanagan depends on climate model used. Global Climate Models (GCMs) mathematically link information about atmosphere and ocean for use at a global scale, but all use different techniques and assumptions and produce differing results when applied at a regional scale. For example, when applied in a study by Taylor and Barton (2004), the Hadley Center’s HadCM3 A2 climate model predicts that winter precipitation will increase 5% by the 2050s, and the CSIRO Atmospheric Research Mark 2b A1 model predicts that it will increase by 22%. The Canadian Global Coupled Model CGCM2 B2 model predicts that summer temperature will increase 1.8°C by the 2050s, and the HadCM3 A2 predicts that it will increase by 4.1°C. What all models do tend to agree on is the general direction of the change. All models generally suggest warmer, wetter winters and hotter, drier summers (Taylor & Barton, 2004). Thus, while the actual magnitude cannot be precisely determined, since, according to the IPCC all models and scenarios have to be regarded as equally likely (Kundzewicz et al., 2007), when examined together they present a range of expected conditions.  5  2.1.2) Climate Change Impacts on Snow Accumulation and Timing of Melt The Okanagan Basin has snow-dominated hydrology (Merritt et al., 2006). Hence, the primary influence of the changing climate on hydrology will be through the mechanism of formation and melt of snow. Air temperature has a strong influence on whether precipitation becomes rain or snow. If temperatures are consistently warmer, more precipitation will fall as rain and the snowpack will consequently be lower. Past trends have already shown that less precipitation has been falling as snow in the lower elevations, but higher elevations have been buffered from change given the temperatures are generally well below freezing all winter (Taylor & Barton, 2004). Future temperature projections by many climate models, however, indicate warmer temperatures at the beginning and end of the winter (CCIS Project, 2010), so less precipitation will be falling as snow, leading to a diminished snowpack at even higher elevations. Another factor affecting the snowpack under warmer conditions is the increased frequency of rain-on-snow events. When there are more precipitation days when the temperature is above 0°C at the end of winter, the resulting rain will hasten the ripening of the snowpack, and therefore advance the timing of snowmelt (Merritt & Alila, 2004).  2.1.3) Climate Change Impacts on Evaporative Water Demand A secondary impact of climate change on basin hydrology will be through interactions with vegetation. In general, higher temperatures lead to higher evaporative water demand (higher potential evapotranspiration). Higher evaporation related to warming tends to offset the effects of increases in precipitation, while magnifying the effects of reduced precipitation (Stonefelt, Fontaine, & Hotchkiss, 2000; Fontaine, Klassen, Cruickshank & Hotchiss, 2001). If more rainfall occurs during the summer and fall, the evaporative water demand could be such that that water is readily used by plants rather than contributing to an increase in streamflow. Changes in evapotranspiration and thus streamflow may also occur in the winter. Plants transpire less or not at all when air temperatures are below freezing (Kaufmann, 1983), but with the addition of more early spring days with temperatures above zero, greater transpiration can occur. Simultaneously, transpiration may decrease under climate change. When there is more CO2 in the atmosphere as would be the case under any climate change scenario, plant stomata can respond by closing and thus decreasing transpiration (Stonefelt et. al., 2000). It is unknown whether these couplings of changes from climate change will offset each other in terms of changes in water yield, or whether one will have a greater magnitude of change than the other. A modeling exercise by Nash and Gleick (1991) showed that increasing temperature by 4°C will increase plant water use such that it will offset a 10% increase in precipitation in terms of water available for streamflow. A modeling exercise by Stonefelt et al. (2000) which added an increase in net radiation to temperature and precipitation predictions found that annual water yield decreased due  6  to increased evaporation and plant water use. However, when relative humidity was also considered, it offset these changes so that only a slight deviation from baseline was seen.  2.1.4) Subsequent Impacts on Stream Flow Changes to streamflow are not a direct consequence of increased levels of CO2 in the atmosphere, but will occur as a result of projected changes to temperature, precipitation and other climate factors that affect snow accumulation, melt, water use by plants, interception and evaporation. Although all changes expected from climate change will likely affect streamflow to some degree, temperature is generally regarded as having the largest influence in interior snow-dominated watersheds like the Okanagan (Cohen & Kulkarni, 2001; Merritt & Alila, 2004; Loukas, Vasiliades, & Dalezios, 2004). At least five studies have used Global Climate Models (GCMs) downscaled to regional scale hydrologic models to predict impacts of climate change on hydrology in British Columbia (Cohen & Kulkarni, 2001; DHI, 2010; Loukas et al., 2004; Merritt et al., 2006; Morrison, Quick, & Foreman, 2002; Taylor & Barton, 2004). Although results differed between application, even when the same climate model was used as a starting point, most studies showed a shift to a more rainfall-driven streamflow regime and a decrease in the magnitude of flood events. Three studies looked specifically at impacts in the Okanagan Basin. The earliest, Cohen and Kulkarni (2001) looked at six unregulated streams in the Okanagan Basin by applying the Coupled Global Climate Model, version 1 (CGCM1), the Hadley Centre Climate Model, version 2 (HadCM2) and the German ECHAM4 GCMs as their driving climate models. Their results showed that spring peak flows would occur earlier than presently, by as much as 6 weeks in some catchments by the 2080s (less for the 2050s), that winter flow would increase, peak flow volume would decrease, and summer flow would decrease. There was no consensus between GCMs on changes to total annual flow. Merritt et al. (2006) also examined unregulated streams in the Okanagan, using three GCMs and two emissions scenarios. The threshold of system sensitivity to change in temperature was found to be at the boundaries of the period of winter snowfall: October and April. Considerably more precipitation was allocated to rainfall during those months under all GCM and emissions scenarios. In their model, this led to earlier peak flows, with the onset of peak flow from 1 to 34 days earlier in the 2020s, depending on GCM and study area, and 13 to 37 days earlier in the 2050s. Peak volumes were either lower or higher in the 2020s, and lower in the 2050s (with the exception of the results generated from the HadCM3 model). Streamflow response to change under all GCMs was a significant decrease in annual volumes by the 2050s and 2080s. For all GCMs and emissions scenarios, in the 2050s, monthly flow volumes were found to decrease significantly from April to June. Most models also showed decreases July-August, and increases January to March. The DHI (2010) study only used one GCM and one emissions scenario (CGCM2 A2) but looked at all streams flowing into Okanagan Lake. Their results showed little change in net annual streamflow volume, but higher winter flows (January – March) in the 2020s and 2050s and a decline  7  in July – September flows in the 2020s and 2050s, and lower, earlier peak flows, which occur up to 2 weeks earlier by the 2050s. In a study in a snow-dominated watershed to the northeast of the Okanagan Basin, Loukas, Vasiliades, & Dalezios, (2004), applied just one climate model and emissions scenario (CGCM A2) during one time period (2080s). The authors projected winter snowpack to decrease by 21%, leading to a reduction in mean annual runoff by 21%. Their model showed peak snow accumulation 1 month earlier than baseline conditions (from March to February) and snow depletion at lower and middle elevations up to 2 months earlier (from August to June). The resulting impact of the changes to snow hydrology were shown to result in an annual maximum flow and mean annual flood peak that decreased by about 12%, and a maximum flood that occurred 14 days earlier than baseline. These findings were similar to those of Merritt et al. (2006) but differed from DHI (2010) in terms of mean annual runoff. Loukas, Vasiliades, & Dalezios (2004) used a hydrological model (UBC Watershed Model) that included processes to represent cloud cover and vegetation distribution, biomass production and plant physiology impacts, energy effects of north and south aspect on calculation, modifications in albedo and consequence short wave energy changes. Yet in their results they attributed streamflow changes to the form of precipitation changing from snowfall to rainfall and consequently decreasing snowpack, indicating the lesser importance of changes in the other variables.  2.1.5) Drought, Implications Separate from “Climate Change” Alone Not only are annual and monthly mean temperatures and precipitation expected to change under climate change, but precipitation extremes are expected to increase in the mid to highlatitudes, with periods of greater intensity and longer periods between rainfall events (Field et al., 2007). Climate simulations in the Okanagan Basin have predicted an increased frequency of drought relative to historic conditions (Nielsen et al., 2006). Actual year-to-year variability cannot be predicted, but if that variability is outside the historic range then there will be consequences for human institutions used to operating within a fixed set of rules. In terms of precipitation, 1929, 1967 and 2003 were the driest in the period of record in the Okanagan (Taylor & Barton, 2004). In terms of flows measured in the unregulated Camp Creek, April to September flows in 2003 were the lowest on record, with flows between 24% and 49% of the 19652003 mean (Taylor & Barton, 2004). Therefore, 2003 is both an extreme drought and one within the current experience of water managers. Examining the impact of consecutive years with climate conditions like 2003 would be a reasonable way of assessing the impact of droughts occurring more frequently in the future.  8  2.1.6) Implications for Management The climate changes discussed above all have implications for water management in the Okanagan Valley. Hotter, drier summers will lead to decreased plant-water availability, longer growing seasons and therefore the need for increased irrigation (Neilsen et al., 2006). These predicted increases in demand coincide with predicted lower low flows (Merritt et al., 2006; Neilsen et al., 2006). High elevation reservoirs, which typically help to alleviate some of the summer shortages, will likely have less water entering them in the later time-steps. When the timing of the bulk of the flow entering the reservoirs shifts to earlier under climate changes, releases to prevent flooding will need to occur earlier. Thus, the time that the reservoir is full will be earlier in the year, leaving a longer period of time where the reservoir is drawn down and subsequently less water available will be available later in the summer for downstream users. Additionally, as the precipitation shifts to be more rain dominated, the snowpack will become a less reliable reservoir, making it more difficult to predict how much water will be available to fill the reservoirs and thus more difficult to plan for periods of shortages.  9  2.2) Hydrologic Impacts of Land Cover Change In addition to the array of impacts to watershed hydrology caused by climate, there are many processes that happen on a landscape that affect the water cycle. Forest canopy loss due to timber harvest, insect infestation and wildfire are the most widespread land cover changes in British Columbia in terms of area affected. Link and Marks (1999) postulated that these types of changes in land cover may result in more significant changes to hydrology than changes in climate in areas with snow-dominated hydrology due to the major role of the canopy in snow accumulation and ablation. The Mountain Pine Beetle (MPB) infestation, hitting the Okanagan Basin beginning 2004 and projected to intensify at least through 2012 (Walton, 2010), is of particular concern to water resource managers in the area (Uunila, Pike & Guy, 2006) and will receive most focus in this section.  2.2.1) Links between Loss of Forest Cover and Hydrology Logging, insect-damage and fire can be described jointly as loss of mature forest canopy. Removing the influence of trees in snow-dominated watersheds affects hydrology in two major ways: 1) less water will be removed from the watershed via transpiration and interception (Spittlehouse, 2006) and 2) more snow will accumulate on the ground in the winter and it will melt faster in the spring (Boon, 2009). Researchers throughout British Columbia are looking at these processes at a stand-level scale (Boon, 2007; Boon, 2009; Winkler, Spittlehouse, & Golding, 2005; Winkler et al., 2008) but few field-based studies provide conclusive evidence about what the impacts will be on streamflow at watershed outlets. Most of the theories about how forest loss affects streamflow come from paired-watershed studies or hydrological modeling exercises (Cheng, 1989; Jost, Weiler, Gluns, & Alila, 2007; Moore & Scott, 2005; Moore & Wondzell, 2005; Schnorbus, Winkler, & Alila; Whitaker, Alila, & Beckers, 2002). The research that most directly addresses what changes might be expected in the Okanagan is discussed in the next sections.  Snow Accumulation The primary effect of loss of mature canopy on streamflow in snow-dominated watersheds is by means of the amount of snow that reaches the ground in the winter, and then forms the bulk of streamflow during spring melt. Living trees intercept snowfall and a portion of the snow sublimates before reaching the ground and becoming part of the snowpack. In cleared areas all precipitation that falls as snow reaches the ground, so clearcuts are likely to see the greatest increase in snow water equivalent (SWE) (Boon, 2009; Winkler et al., 2009). Areas affected by stand-replacing fires, dead or “grey” MPB-attacked stands and stands experiencing needle-loss or “red” attacked stands (Redding et al., 2008) will also have increased SWE. When some structure of the canopy remains, either from trees killed by wildfire that have not yet fallen down, or beetle-killed trees that have not yet lost their  10  needles and branches (Huggard & Lewis, 2007; Winkler et al., 2009), snowfall is intercepted, evaporated and sublimated, processes that cumulatively prevent up to 25% of precipitation from reaching the forest floor (Spittlehouse, 2006). Canopy changes observed at the stand level result in different impacts on snow accumulation depending on location and climate. The results of a few studies are provided here to give a range of potential impacts on snowpacks. In a study where snowpack depth was measured after clearcut logging in the Penticton Creek watershed, the portion of the watershed previously dominated by lodgepole pine showed an increase in maximum SWE by 12%, while in an area previously dominated by mixed forest, the clearcut had 27% more SWE than the surrounding forest (Winkler 2001 in Winkler et al. 2009). In a study in Colorado, 50% canopy removal led to a 21% increase in SWE (Troendle & King, 1987). In a study in Montana, 50% removal of lodgepole pine, in the form of a thinning rather than a clearcut, led to SWE increase by between 17 and 33% (Woods, Ahi, Sappington, & McCaughey, 2006). Winkler et al. (2009) found that reductions in snow interception and increases in net precipitation are generally proportional to reduction in canopy cover, so can therefore be modeled as a linear relationship. For every 1% loss of forest cover, there was about a 0.4% increase in SWE (or about a 20% increase in SWE when 50% of the forest cover is lost). For forests attacked by insects such as the Mountain Pine Beetle, the relationship between affected trees and SWE is more complicated. Whereas the greatest impacts from clear-cutting occur immediately after the cut, the greatest impacts from beetle-kill occur several years after. According to Boon (2009) needles fall off 3-5 years post attack, resulting in only minimal impact on SWE, and after 10-15 years branches fall, resulting in greater impact. Despite these nuances, a few commonalities in the impacts are apparent. Stands with more structure present intercept more snow and allow less solar and advective energy to melt snow. If there are more needles and branches on the trees, the snow accumulation is similar to that in a forest, but if the branches have fallen or the trees have fallen, it is more similar to a clearcut (Winkler et al., 2009). These impacts will be stronger or weaker depending on other factors like slope, aspect, subcanopy structure, and whether it is a wet year or a dry year (Jost et al., 2007; Winkler et al., 2008).  Ablation Rates Change in the rate that snow melts is important in terms of when the water in the snowpack will be available for use by vegetation and for streamflow. In cleared areas, snow can be expected to melt faster because of higher wind and associated turbulent heat transfer, and because of higher net shortwave radiation providing energy to melt the snowpack. In a multi-year study comparing clearcut and Mountain Pine Beetle attacked stands, the clearcut consistently showed faster ablation rates (Boon, 2009). Table 1 summarizes the snow accumulation and ablation impacts of forest canopy loss.  11  Table 1: Range of impacts to snow from land cover change relative to a mature forest All values are means, and except for snow depletion date, all values are increases relative to baseline (mature forest). Based on data found in: Winkler et al. 2009; Winkler & Boon, 2009; Boon, 2009; Boon, 2007. Land Cover Max SWE % Ablation mm/day Snow depletion date Clearcut 12 to 27 1 to 9 < 9 days earlier MPB Green/Red Attacked 9 to 23 4.2 to 8.5 Later than clearcut MPB Grey Attacked 10 to 24 2.1 to 9.7 Later than clearcut Burned 9 to 50 While Table 1 shows some of the few quantitative values available to describe relationships between snow and forest cover, it cannot be used to extrapolate what might happen in other stands at other times given all the many factors of topography and climate at play. However, given that these are the only values available, they will be used as a rough guideline in this study for what might be expected to occur.  Evaporation Rates Evapotranspiration (ET) from forests with dry canopy in British Columbia is estimated as 2 4.5 mm/day on a sunny day (Winkler et al., 2009) whereas bare soil evaporation during intermittent wet and dry periods has been found to average 1- 2 mm/day, and during extended dry periods, less than 0.5 mm/day (Novak & Black, 1982). A modeling study in BC spruce forests found ET to average 1.66 mm/day (Maayar & Chen, 2006). Reduction in mature forest canopy leads to a significant reduction in evapotranspiration in a watershed. While the understory continues to transpire, it does not intercept and transpire as much water as a mature forest canopy (Winkler et al. 2009). Reduced evapotranspiration results in increased soil moisture, which in turn increases the percent soil saturation and can lead to more runoff during snow melt or heavy rainfall events (VanShaar, Haddeland, & Lettenmaier, 2002). In a study of forests in the southern interior of British Columbia, Spittlehouse (2006) found evaporation during the summer from a high-elevation clearcut to be 30% less than from forest, with the greatest difference between forest and clearcut when soils were moist. For similar reasons, VanShaar et al. (2002) concluded that the greatest change to streamflow after forest loss will occur during the spring melt season, when soil moisture is not limited. Overall, the effect of reduced ET due to loss of forest cover augments the effect of increased net precipitation by increasing soil water content and thereby streamflow.  Stream Response Streams integrate all of the changes occurring in the uplands, from evaporation changes in the forests, to the timing and quantity of water generated by snowmelt. In general, mechanisms that lead to more snow accumulation result in more water available for streamflow in the spring. Mechanisms that lead to earlier melt result in peak streamflow occurring earlier.  12  In a compilation of studies, Winkler et al. (2009) found that in snow-dominated watersheds, post-logging water yields increased from 0.25 mm to more than 3 mm for each percentage of watershed area harvested. In a modeling exercise by Schnorbus et al. (2004), forest removal of 60% increased magnitude of the snowmelt-generated streamflow by 15%. In a watershed in Colorado, streamflow was 45% higher after logging for the first 5 years post logging compared to preharvest (Troendle & King, 1987). On the high end of these findings, a field-based study by Spittlehouse (2006) found up to an 87% increase in surface water drainage in one cleared catchment in one year, though average increases were closer to 50%. Peak flow is defined as the flow derived primarily from snowmelt in the spring measured at the day of maximum flow, the highest 7-days of flow, or flow over the entire melt period. Peak flow generally increases after forest canopy loss, though studies of the extent of that increase show great variability. Van Haveren (1988) found that 100% clearcutting produced a 50% increase in mean peak flow (Winkler et al., 2009). In a study in Colorado, watersheds from 20 to 40% harvested had from 20 to 90% increases in peak flow (Troendle & King, 1987). Timing of these flows is also extremely variable. In a study of an extreme amount of canopy removal after salvage logging following MPB attack, the peak flow was over 2 weeks earlier following harvest. Depending on where the harvest is in the watershed, however, flows can “desynchronize” and leave the timing of the melt unchanged (Winkler et al., 2009). Salvage harvesting is expected to have a greater hydrologic impact than MPBrelated mortality alone because more canopy is removed (Redding et al., 2009). Table 2 summarizes the results of the studies discussed above. These can be used, broadly, as a comparison to model results, but are not comprehensive enough to permit generalization.  Table 2: Range of changes in stream hydrology expected from land cover changes Watershed size, area affected, location, and season all vary between studies (data from a compilation of research by Winkler et al 2009; Redding et al., 2008; Spittlehouse, 2006) Land Cover Clearcut Green/Red Attacked Grey Attacked Burned  Total Volume Increase (%) 15 to 87 less than clearcut less than clearcut  Peak Flow (%) 20 to 90 less than clearcut less than clearcut 100  Onset of Freshet 3 to 21 days earlier not as early as clearcut not as early as clearcut  Low Flow (%) 0 to 17 0 to 10  2 weeks earlier  In terms of management considerations, hydrological changes resulting from forest removal due to a pine beetle attack, fire, or logging of most concern to water managers are the increases in flood risk, damage to drainage structures, and increased sedimentation (Winkler, Rex, Teti, Maloney & Redding, 2008). Where a high elevation reservoir is present in a watershed, impacts on the timing of peak flow are also of concern (Dobson, 2009). When peak flows occur earlier in the spring, as with climate change, the reservoir must be filled earlier and thus less water may be available to meet later summer/early fall demands (Merritt et al., 2006).  13  2.3) Human Demand Humans are both the agent of climate and land use changes and are also affected by the impacts of these changes. The conditions leading to MPB outbreaks are also likely due at least in part to human management of forests, as well as to climate change (Taylor & Carroll, 2003; Williams & Liebhold, 2002). In an area where humans depend on a snow-fed, landscape-mediated water supply, these changes come to bear on water available to meet human needs, and people become recipients of the afore-mentioned changes. In turn, the resulting water made available is used by humans for a variety of purposes, and the choices made about how to use that water impact how much is available for each of many specified needs. In addition to the changes to water supply described in Sections 2.1 and 2.2, there will be a variety of choices that take place once the water is made available to human use that affects the ease with which different needs are met. Existing land use in the Okanagan locks the region in to a certain amount of agricultural water use. With climate change, the irrigation season is expected to begin sooner and last longer, and could be as much as 30% longer by the 2080s. With increased efficiency, agricultural consumption could be reduced (Neilsen et al.,2006). Across all water use sectors, climate change alone could increase future use basin-wide between 3% and 10% by the 2020s, and 7% to 19% by the 2050s (Neale, Carmicheal & Cohen, 2005). When combined with population growth, the increase could be up to 21% by the 2020s and 86% by the 2050s (Neale et al., 2005). Increased efficiency, taking place largely in the outdoor watering sector, could substantially reduce water use (Neale et al., 2005; Maurer, 2010, DHI, 2010). In one study area in the Okanagan, Neale et al. (2005) found that by the 2050s efficient water use, low population growth and the “best case scenario” for climate change, could reduce water use to below 2001 levels. Results from the Okanagan-wide DHI (2010) study did not project quite as substantial savings, but still suggest that if efficiency measures are put in place in all water use sectors, water consumption could be 35% less than with no efficiency by the 2020s. The type of development is another factor affecting future water demand. In a study of four communities in the Okanagan, Maurer (2010) found that if increased development occurs in the form of ground-based dwelling, outdoor water use may increase 55% by 2026, whereas if densification take places, the increase in outdoor use will only be 12%. These demand changes have implications for improving streamflow and thus meeting a variety of ecological needs. The study by DHI (2010) predicted that a 14% increase in the average annual flow in Mission Creek could be achieved through various planning, policy, and water use efficiency measures. In summary, water “stress” will likely increase with future changes in demand, though there are options for reducing that stress through water conservation measures.  14  2.4) Impacts of Hydrologic Change on Aquatic Ecosystems Modifications to streamflow through changes in climate and land use, in combination with changes in amount and timing of water extracted from streams for human use, have significant implications for aquatic life. There is a great amount of uncertainty surrounding the connections between water quantity and aquatic ecosystem integrity, with only limited field research having been conducted on flow-habitat relationships in Okanagan streams (Summit Environmental, 2004). Recent hydrological research suggests that the safest way to ensure that biodiversity and ecosystem integrity are maintained is through maintaining the full range of natural intra- and inter-annual flow variability (Richter, Baumgartner, Powell, & Braun, 1996). However, under climate change, “natural” flow regimes are likely to vary outside of their historic range (Field et. al.,2007), and with the added influence of land use change and water extraction to meet increasing demands, future streamflow patterns are likely to be well outside the range of historic variability. British Columbia water law is currently under revision and methods for determining how to most efficiently keep enough water in streams for aquatic life are being assessed (Ministry of Environment, 2010). Future regulations will likely include a recommendation that local assessments take place before major projects are undertaken that alter stream flow, but for “low-impact” withdrawals such as for irrigation, standardized “rules-based” approaches will be used (Ministry of Environment, 2010). Currently, two forms of rules-based estimates are being employed. While neither has any regulatory backing (both are “discretionary” only), both are useful for assessing whether current flows are likely sufficient for maintaining critical ecological functions for fish and invertebrates, and thus may be used for assessing whether flows under future scenarios may also protect aquatic life. The first method is applicable to determining if fish habitat needs are being met by stream flow at a regional scale, known as the B.C. Modified Tennant method. It requires only data about annual flow and determines monthly requirements based on percent Mean Annual Discharge (MAD) (Ptolemy & Lewis, 2002). This method was used to determine “Conservation Flows” for Kokanee and Rainbow Trout in tributaries to Okanagan Lake for the agency formerly known as B.C. Fisheries, now Ministry of Environment (Northwest Hydraulic Consultants, 2001). A more recently developed method for ensuring that stream flow mimics historic flow regimes, thus meeting all necessary ecological requirements, was developed specifically for British Columbia (known as British Columbia Instream Flow Needs, BCIFN) (Hatfield, Lewis, Ohlson, & Bradford, 2003). At least 20 years of daily flow data are required to determine historic variability and thereby set a threshold above which the goal of maintaining key features of the stream’s hydrograph and minimizing risk to fish and other biota can be met (Hatfield et. al., 2003). As one proceeds below these thresholds, the likelihood of flow-related constraints on aquatic productivity increases. The true level at which harm occurs, however, is unknown without detailed stream-specific studies.  15  Both MAD and BCIFN methods for setting flow targets use a monthly timestep. However, many critical fish life stages occur on a shorter timescale. The B.C. Modified Tennant recommends establishing the duration per annum that each flow level should be met so that no harmful disruption will occur. As a general guideline, Ptolmey and Lewis (2002) suggested that the low flows set for fish rearing and incubation be met for months, flows to connect side channels last for weeks. They suggest that higher flows, around 400% MAD, occur 1 to 2 days annually to provide sufficient force to flush sediment from the stream substrate. The NHC conservation flows report also recognizes the limitations of a monthly target, stating that “a more refined analysis would use daily flows in both a ‘wet’ and a ‘dry’ year. Since some life history events occur over days to weeks, flows over these shorter time spans would determine the likelihood of success for passage or spawning better than a mean monthly value.” (NHC, 2001, p. 31). Thus it might be considered that the duration of violation of a monthly target is one way to examine potential harm in more detail: the impact of streamflow dipping below a target for a few weeks is probably more detrimental to fish health than dipping below for a few days. *** In this chapter, the types of changes likely to impact water resources in the Okanagan in the near and long-term as well as interactions between supply and demand have been reviewed. The next chapter will describe the methods used to integrate this information in WEAP, an integrated water resource management model, and thereby address the study question of how potential changes to the magnitude and timing of stream flow in response to global climate change, a prolonged climatic drought, land cover change resulting from a Mountain Pine Beetle outbreak may impact human water needs and aquatic life.  16  Chapter 3: Methods 3.1) Study Site Selection and Description of Study Area The characteristics considered for selection of a study watershed were that: a) there be a hydrometric station present, and b) surface water provide the primary source of water for a community. The District of Peachland is served primarily by water from Trepanier and Peachland Creeks. Trepanier Creek has an active hydrometric station with 30 years of streamflow data, while Peachland Creek has a gauged tributary as well as historic streamflow data on the main channel. With climate stations located near the base headwaters of these streams, and with a small but 2  growing municipality at their outlets, the 380 km watershed area was ideal for use as the study area. These two watersheds, the District of Peachland, and demand sites outside the District that extract water from these watersheds will be included in the bounds of the study area (Figure 1).  Figure 1: Study area location in British Columbia  17  3.1.1) Existing Land Use/ Land Cover The land cover in the Trepanier Creek watershed is predominantly forested. The basin is in a relatively natural state, with only some areas which were logged in the 1990s. Logging occurred in mostly the higher elevations in the northeast portion of the basin, with some logging also occurring in the south and northwest. The Peachland Creek watershed is mostly forested. Beginning in the early 1990s there has been extensive logging in the uplands, especially around Peachland Lake. Some recently logged areas are also present in the northern edge of the Greata Creek watershed. The low to mid elevations (600 to 1400m) in both watersheds are characterized as being in the Interior Douglas-Fir Biogeoclimatic zone (Ministry of Environment, 2008a). Above 1400m is the Montane Spruce, with some of the highest elevations, e.g., above 1800 m, classified as EnglemannSpruce Subalpine fir. Within those zones, dominant tree species are ponderosa pine (P. ponderosa), lodgepole pine (P. contorta), Douglas-fir (Pseudotsuga menziesii) and Englemann Spruce (Picea engelmannii) (Ministry of Environment, 2008a).  3.1.2) Changes to Land Use/Land Cover Forest Practices Likely changes to the land cover in these watersheds will result from planned Forest Practices activities and activities permitted for private land owners. About 84% of the Peachland Creek watershed is designated as “Short Term Retention” for forest practices, with additional areas in the watershed marked as either “Planned Logging” or “Long Term Retention,” which exists mostly around lakes and rivers, indicating their protected status (Tolko Industries, Ltd, 2009). In the Trepanier Creek basin there are 2,884 hectares of protected area (Ministry of Environment, 2000) equivalent to about 10% of the basin, with the rest open to potential logging in the future.  Insect Outbreaks Mature lodgepole pine is the target species Mountain Pine Beetle (Huggard & Lewis, 2007). Forty-five percent of the study area is mature lodgepole pine, with 40% mature pine in the Trepanier basin and 50% in the Peachland basin (Ministry of Forests and Range, 2009b). An outbreak of Mountain Pine Beetle has already affected the area, and as of 2008, in Trepanier: about 0.5% was 1  severely affected; 5% moderately affected; 40% lightly affected. In the Peachland watershed, 30% has been lightly affected. The current infestation in the Okanagan is projected to intensify through 2012 (Walton, 2010), resulting in increasing mortality. With the combination of potential increased beetle-kill and resulting salvage logging and the possibility of a severe wildfire, there is a potential for dramatic land cover change in the study area. 1  MPB attack severity categories by % mortality: Lightly = <11%; Moderately = 11-30%; Severely = 31-50%. (Ministry of Environment 2008c)  18  3.2) Model Description The model used in this study, the Water Evaluation And Planning system (WEAP), integrates a hydrologic model with representation of how water is allocated and delivered for different human purposes and then allows for examination of subsequent impacts on aquatic ecosystems (Yates et al., 2005b). A conceptual diagram of model structure is provided in Figure 2. WEAP represents hydrologic processes using a semi-distributed approach where the water balance is calculated for user-defined area of similar hydrologic properties. In this way, the user can group areas with similar land cover, soil, elevation, slope, aspect etc. expected to have the same hydrologic response throughout. These areas are referred to as hydrologic response units or HRUs. The hydrological model component of WEAP represents soil as two layers; an upper layer and a lower layer, each with a user-designated water holding capacity and depth. Water flows into and out of the soil layers based on hydrologic conductivity and a defined ratio for partitioning soil moisture between layers and into interflow and baseflow. The upper layer provides water for transpiration by plants which is depleted according to evapotranspiration calculated using the Penman-Monteith equation. Streamflow results from calculations of interflow, baseflow and surface runoff. Groundwater recharge can be modeled by specifying that the lower layer be an aquifer rather than a soil layer. WEAP uses a temperature-index snow melt model to estimate snow water equivalent and snow melt in each HRU (Yates et al., 2005b). Precipitation is partitioned into rain or snow depending on melting and freezing temperature thresholds. Melt is triggered by melting temperature thresholds, albedo, solar radiation, and user-defined “other energy” available for melt (Young et al. 2009). The two main variables that describe vegetation function include the crop coefficient, Kc, which is applied with the Penman-Monteith equation to describe the amount of water used by a plant in the watershed relative to a “reference crop,” and a Runoff Resistance Factor, described as integrating Leaf Area Index and slope of the land surface in terms of how they affect surface runoff (Stockholm Environment Institute, 2007). Water demands in WEAP are user-defined, but can include municipal and industrial demand, irrigation demand, and instream flow requirements. Each site of demand for water is assigned a priority (1-99) and linked to its available supply sources. Each supply source may have one or more demands that it needs to satisfy, so the model allocates available supply to all demands based on priority. Future municipal demand can be simulated by using projected population growth rates and projections for water reduction strategies. Reservoirs capture water generated within the HRUs. A priority can be assigned to the reservoir so that at a given time, streamflow will either go to filling the reservoir or meeting downstream demands.  19  Figure 2. WEAP functional schematic * = Agricultural demand can also be simulated by temperature and solar radiation data as is shown in the upper left hand corner of the diagram. Many additional components are included in WEAP but are not shown here since they do not relate to the study. The diagram presented here represents WEAP as applied by the author in this study.  20  3.3) Constructing Model to Represent Study Area Hydrologic processes are represented in WEAP using a semi-distributed approach where the water balance is calculated for user-defined “catchments” which can represent different subwatersheds or elevation bands within the watershed. Within each catchment, discrete units that represent different landscape characteristics, or hydrologic response units (HRUs), can be defined. In this way, one can treat areas with similar land cover, elevation, slope and aspect etc. as having the same hydrologic response. The hydrological model component of WEAP represents soil as two buckets; an upper layer and a lower layer, each with a defined depth. Water flows into and out of the buckets based on hydrologic conductivity and a defined ratio for partitioning soil moisture between layers and into interflow and baseflow. The upper layer provides water for transpiration by plants which is depleted according to evapotranspiration calculated using the Penman-Monteith equation (Yates et al., 2005b). Streamflow results from calculations of interflow, baseflow and surface runoff (Yates et al., 2005b). The specification of the model to the study area will be described here.  3.3.1) Hydrological Response Unit Delineation The Trepanier and Peachland Creek watersheds encompass elevations from 600 masl (meters above sea level) to 1880 masl, include four biogeoclimatic zones and are characterized by topography ranging from steep slopes to rolling plateaus. Due to the diverse terrain of these watersheds, hydrologic processes vary throughout. In this study, the delineation of areas of similar hydrologic properties, known as hydrologic response units (HRUs), was based on the objective of best representing the dominant hydrologic processes appropriate to the region given the capabilities of WEAP. Given that these are snow-dominated watersheds, representation of processes that affect snow accumulation and melt were the first priority for delineation of HRUs. Therefore, the first step was to create HRUs that represent areas of similar elevation (to define a common temperature throughout) followed by slope and aspect (to capture subtler accumulation and melt processes). In order to represent changes to land cover, areas of similar land cover were aggregated in the same HRUs. Details on the choice of delineation of these HRUs is described here.  Subwatersheds and Elevation Bands Subwatersheds were delineated for major tributaries. Where there were many small tributaries contributing to the main channel, the subwatershed outlet was defined as a length of the main channel (Figure 3). The subwatersheds were next divided into elevation bands. The first criterion for choosing the limits of the elevation bands was that they align with snow surveys associating an elevation of snowline with a date, identified through aerial surveys that were conducted during two years (Dobson, 2003). In this way, the snow accumulation in these bands could be calibrated to observed snowmelt timing. The next step in delineation of elevation bands  21  involved placing limits at approximately 250m intervals (based on TRIM contour lines (Ministry of Agriculture and Lands, 2003)). The plateaus in the upper reaches were placed into HRUs with less elevation range over a larger area to represent their flat topography and based on the assumption that snow melts throughout the plateaus at a similar time (Grainger, 2009).  Biogeoclimatic Zone The next criterion for choosing elevation band divides was that they be at the edges of mapped biogeoclimatic zones. These zones represent the spatial extent of major vegetation groups (Englemann-Spruce-Subalpine-Fir; Montane Spruce; Interior Douglas Fir; Ponderosa Pine) and also capture differences in soil and climate. The biogeoclimatic zones also happen to line up with the elevations chosen to represent the snow line from the snow survey. (Source: Biogeoclimatic Zone Map, Ministry of Environment, 2008a).  Slope Areas with steep, moderate or shallow slopes were generally grouped into separate HRUs. In areas where no slope type dominated, parameters in the model were set to reflect the mixture of these conditions. Thus, an area comprised of both “shallow” and “steep” slopes would be classified as “moderate”. (Source: Digital Terrain and Soils Map (Ministry of Energy, Mines and Petroleum Resources, 2009) ; TRIM Contour Lines 1:20,000 (Ministry of Agriculture and Lands, 2003))  Aspect Since aspect is an important driver of snow melt due to spring irradiation, separate HRUs were delineated for south, north, east and west-facing slopes. (Source: TRIM Contour lines 1:20,000, Ministry of Agriculture and Lands (2003), acquired through UBC Map Librarian).  Vegetation Vegetation data were acquired from a GIS layer created by the Ministry of Forest Vegetation Resource Inventory (VRI) vegetation survey (Ministry of Forests and Range, 2008b). Where possible, areas of like vegetation were grouped into a single HRU. Classification of vegetation type within each HRU was based on the fields: BC Land Cover Level 4 (Coniferous, Broadleaf, Shrub-Tall, Shrub-Low, Herb, Graminoid, Forb, Exposed Land) and BC Land Cover Level 5 (Dense (60-100% cover), Open (26-60%) and Sparse (10-25%)). These fields were combined in order to create a single category that represented both vegetation type and density. Land cover classes in the study area are Coniferous-Open, and Coniferous-Dense and Exposed Land. HRUs were assigned the dominant vegetation-density type present. The different aspects generally agreed with vegetation type. In general, south facing slopes tended to have open coniferous stands, and north-facing slopes to have dense coniferous stands.  22  VRI datasets were available for 2008 and 1994. The 1994 data were used to identify vegetation type during the calibration period, and the 2008 data were used to assess current conditions. Both datasets had information on “stand age” and could be used to backtrack the timing of clearcut harvesting. This information was used to determine when to classify an area as shrub/herb rather than mature coniferous forest. For the 1983-1986 calibration period, approximately 95% of the Trepanier watershed, and 98% of the Peachland Creek watershed were forested.  Soils Where possible, HRU boundaries were chosen that also lined up closely with transitions in soil type. Given that the soils in this area are heterogeneous and do not always follow topographic divides, most elevation bands represent a heterogeneity of soil classes. (Source: Ministry of Energy, Mines and Petroleum Resources, 2009). Major characteristics of the HRUs are shown in Table 3 and they are presented in relation to biogeoclimatic zone in Figure 3.  Figure 3: Trepanier and Peachland Watersheds Map. HRU boundaries overlain on biogeoclimatic zones  23  Table 3: HRU specifications Nomenclature: T= Trepanier; P = Peachland; 1,2,3, etc. = subwatershed; L, M, H = low, mid, or high elevation; subclasses 1 or 2 to separate HRUs in subwatersheds with similar general elevation; S, N, E, W = south, north, east, west-facing aspect, P= plateau. Median HRU Area Elevation Elevation Slope Vegetation Type and Name (hectares) Limits (m) (m) Spacing Steep Conifer - Open T 1L1 1646 560 - 1060 810 Moderate Conifer– Open/Dense T 1L2S 1516 820-1180 1000 Shallow Conifer - Open T 1MP 2063 1180 - 1560 1370 Steep Conifer - Open T 2L1S 1216 600 - 1040 820 Moderate Conifer- Open/Dense T 2L2S 1300 1040 - 1300 1170 Shallow Conifer - Dense T 2MP 1411 1300 - 1500 1430 Steep Conifer - Dense T 3L1N 1072 600 - 1040 820 Shallow Conifer - Dense T 3L2P 917 1040 - 1200 1120 Moderate Conifer- Open/Dense T 3L3N 355 1200 - 1380 1290 Moderate Conifer - Dense T 4LN 442 780 - 1220 1000 Shallow Exposed/ Conifer D. T 4M1P 671 1220 - 1380 1300 Shallow Exposed/Conifer D. T 4M2P 1161 1380 - 1520 1450 Moderate Conifer - Dense T 4H1P 558 1520 - 1620 1570 Shallow Conifer-Open T 4H2P 634 1620 - 1880 1750 Steep Conifer - Dense T 5LS 513 800 - 1400 1100 Steep Conifer - Dense T 5LN 671 800 - 1400 1100 Steep Conifer-Open T 5MS 882 1400 - 1640 1520 Moderate Conifer-Open T 5MN 645 1400 - 1640 1520 Shallow Conifer-Open T 5H1P 339 1640 - 1860 1750 Moderate Conifer-Open T 5H2P 553 1640 - 1900 1770 Moderate Conifer-Open T 6L1E 202 780 - 1040 910 Moderate Conifer-Dense T 6L1W 298 780 - 1040 910 Steep Conifer-Open T 6L2E 570 1040 - 1340 1190 Moderate Conifer-Dense T 6L2W 577 1040 - 1340 1190 Moderate Conifer-Open T 6MP 2426 1340 - 1580 1460 Moderate Conifer-Open T 6HP 887 1580 - 1880 1730 Moderate Conifer-Open P 1L 1492 580-1080 830 Moderate Conifer- Open/Dense P 2LS 1734 920-1400 1160 Steep Conifer-Dense P 2LN 1228 920-1400 1160 Shallow Conifer- Open P 2MP 422 1400-1520 1460 Moderate Conifer– Dense/Open P 2H1P 507 1400-1640 1520 Moderate Conifer-Open P 2H2P 565 1400-1700 1550 Moderate Conifer-Open P 3LS 1108 920-1240 1080 Moderate Conifer-Dense P 3LN 680 920-1240 1080 Moderate Conifer-Dense P 3M1N 812 1240-1420 1330 Shallow Conifer-Dense P 3M2S 1614 1240-1480 1360 Shallow Conifer-Open/Dense P 4MP 835 1260-1400 1330 Moderate Conifer- Dense P 4H1E 849 1400-1600 1500 Moderate Conifer-Open P 4H2P 621 1600-1720 1660  24  3.3.2) Choice of Time Step WEAP can run with either a monthly, weekly, or daily time step. Choice of time step should be based on how long it takes a drop of water to reach the watershed outlet (Loucks et al., 1981), which can be approximately determined through calculation of the “time of concentration”. The “time of concentration” was determined by using a formula generated by the BC Ministry of Forests (1988), tc = a +b(x - 0.2), where x is the square root of the watershed area, and a and b are coefficients based on slope. The study watersheds have a time of concentration of 0.8 (Peachland) and 1.7 (Trepanier) days, so a daily time step is appropriate for modeling basin hydrology. Additional reasons to use this time step are that climate input data are available as daily averages and expected responses to some of the changes in climate and land use variables examined in this study are on the scale of days rather than weeks or months.  3.3.3) Infrastructure Infrastructure related to water storage and delivery was modeled and included water retention reservoirs (both active and inactive reservoirs), wells, and inter-basin diversion pipes.  Reservoirs Peachland Lake is used by the District of Peachland to moderate stream flows and provide water to meet demand for the District of Peachland. The reservoir was modeled using a reservoir node with a set total volume of 11,770 ML (W. Grundy, District of Peachland Public Works, personal communication, 4/1/2010). Reservoir releases vary dynamically with available water each year. Release volumes are based on a combination of a water license for fish held by the Ministry of Environment and expected needs by the District of Peachland. Normally, releases are on the order of 0.71 cubic meters per second (cms) in May and June during the spring freshet, and 0.14 cms during the rest of the year (Ministry of Water, Land and Air Protection (1999); confirmed by District staff as meeting current operating conditions). WEAP will either fill the reservoir or release water depending on a user-defined priority. The District of Peachland was given a demand priority of 1 and the reservoir a priority of 2 so that the reservoir would only fill once municipal needs were met. An “instream flow” node was modeled just downstream of the reservoir which pulled water from the reservoir (if available) at the appropriate times.  Interbasin Diversions An interbasin diversion is present between the headwaters of Trepanier and Peachland Creeks. The diversion exists because Brenda Mines, a molybdenum extraction facility, was in  25  operation from 1970 through 1990 and required more water for the mining process than available from the headwaters of Trepanier alone (Patterson, n.d.). Prior to 1990 water was diverted from MacDonald Creek (tributary to Trepanier) to Peachland Reservoir and then diverted back into Trepanier via the Brenda Mines site. Environment Canada’s hydrometric gauge (8NM218) recorded flow through the diversion May-August in 1973-1979 (Environment Canada, 2007), but beyond that period the exact timing and volume of diversions is unknown. The mine was closed in 1990 and District of Peachland claimed the previously diverted water for use in filling its reservoir at Peachland Reservoir. After 1990, no more water was officially diverted from MacDonald to Peachland (Patterson, n.d.), though a broken control valve continues to allow some of Trepanier’s water to flow to Peachland Reservoir (D. Dobson, Dobson Engineering, personal communication, 11/1/2009). In the present study, the period from 1984-1986 was used as the calibration period, so the diversion was set as active during calibration. The amount of water sent through the diversion during this period was assumed to be the same as in the gauged period in the 1970s. The volume of water 3  diverted (approximately 1.7 million m annually) was compared to the approximate volume supplied by the different catchments (P4H1, P4H2, P4M, T4H1 and T4H2) and a combination of catchments was chosen that generated the appropriate volume to be delivered to the diversion (from Peachland to Brenda) (P4H1, P4H2, T4H2), with the others being delivered to Peachland Reservoir to later be released downstream. Future scenarios were based on post-1990 conditions where most of the flow from the headwaters of MacDonald Creek supplied Trepanier Creek and Peachland Reservoir was fed by its own headwaters as well as portion of water entering from the MacDonald Creek Diversion (represented as catchment T4H2, approximately 20% of total inflow into the reservoir).  3.3.4) Demand Sites Water demand sites are defined in WEAP such that they may represent demand within the watershed itself (evaporative demand by vegetation) or demand by a human activity. Demand by vegetation in the HRUs is incorporated within the hydrologic component of the model. Additional demand sites were defined where human activity required more water than available from precipitation alone. Demand sites were created to appropriately represent discrete units of the water management system. Individual licenses in the Trepanier watershed were given a demand node separate from the District of Peachland since rules governing their extraction and water use are different than in the District of Peachland. Within the District of Peachland, separate demand nodes were created based on how the District manages water distribution. Details on demand sites are provided in Section 3.6. Figure 4 shows the compilation of model components described in this section.  26  Figure 4: Representation of Study Area in WEAP  3.4) Representing Climate Variables Climate Data In developing a semi-distributed hydrologic models it is important to represent temperature and precipitation in as spatially-accurate a manner as possible for calibration and later for representing climate change. Climate stations near the study watersheds are in the towns of Peachland (345 masl; latitude: 49.78, longitude: -119.72), Kelowna (429 masl; latitude: 49.96, longitude: -119.38), and at the high elevation mine site of Brenda Mines (1520 masl; latitude: 49.87, longitude: -120.00). Since a low and a high elevation station were available, lapse rates were calculated to determine climate values for elevations between these stations and therefore provided climate data specified to 39 HRUs within the watersheds.  27  Temperature Daily temperature data available from Brenda Mines and Peachland climate stations (Environment Canada, 2009a) were used to determine the lapse rate each day of the period of record using the equation:  LRT = Tu – TL Eu - EL  (1)  Where Tu = temperature at upper station; TL = temperature at lower station; Eu = elevation at upper station and EL = elevation at lower station. For comparison, - 0.0065 °C/m is the global mean lapse rate (Dodson & Marks, 1997). The average lapse rate calculated using the method above is -0.0064 °C/m. This rate was used to calculate temperature values at the mean elevation of each elevation band.  TE = TL + LR *( E – EL)  (2)  Where TE is temperature at the desired elevation, TL = lower station temperature value; LR = lapse rate; E = median elevation of HRU; EL = lower elevation station elevation. There were a few months for which temperature data at Peachland were not available. To determine temperature at that time the inverse of the lapse rate was applied to the Brenda Mines station. The default was derived from Peachland in the years where data were available.  Precipitation The low and high elevation climate stations were used to calculate a lapse rate for precipitation, but since precipitation did not occur every day, and since precipitation did not occur on the same days at high and low elevations, instead of using a daily lapse, an average lapse rate over each month was calculated and then applied the mean lapse rate to daily values (G. Jost, Research Associate, University of British Columbia, personal communication, 11/5/2009).  LRP = PU – PL PU/EU-EL  (3)  Where LR= lapse rate; PU = monthly mean precipitation at the upper station; PL = monthly mean precipitation at the lower elevation station; EU = upper station elevation; EL = lower elevation station elevation. Since it often rained at the high elevation station but not at the low elevation station, adding the lapse to the lower station data put too much of the weight of the increase onto the days that it rained, so instead the inverse of the precipitation lapse rate was applied to the higher station to  28  determine values for the lower elevations. This method was deemed appropriate since the mountainous conditions of the upper station were representative of larger areas of the study area than low elevation, lakeside station. The equation for determining daily precipitation in each HRU is:  PED = PUD * (1-LR*(E – EU))  (4)  Where PED = daily precipitation at median elevation of HRU and PUD = daily precipitation at the upper climate station; LR = lapse rate; EU = Upper station elevation; E = median elevation of HRU.  Relative Humidity and Wind Speed Data on relative humidity and wind speed were only available on an hourly time step from the Kelowna A weather station. Hourly data were aggregated to determine daily averages. Where no data were available for any hour in a day, the average of the preceding day was used to fill in the missing values.  3.5) Model Calibration Calibration of the hydrologic model began with selecting a calibration and validation period and then adjusting the climate data series to compensate for gaps. Next, modeled snowfall was calibrated to match observed records. The following step was acquisition of landscape data to calibrate the remaining parameters in WEAP so that modeled streamflow most closely approximated measured streamflow data. Measured streamflow data was available for one site in Trepanier Creek, above the withdrawals for the District of Peachland. Peachland Creek did not have a long enough flow record near its outlet, so flow for one of Peachland Creek’s tributaries, Greata Creek, was used for calibration in this watershed. The calibration methods will be described here and the outcomes of the calibration, including performance statistics, are included in Section 4.1.  3.5.1) Choice of Calibration and Validation Period The calibration and validation period were chosen for time periods where complete daily data were available for temperature and precipitation at both the Peachland and Brenda Mines climate stations, stream flow (Gauge 041 in Trepanier Creek, and Gauge 173 in Greata Creek in the Peachland Creek basin (Environment Canada, 2007)), and snow water equivalent (SWE) (River Forecast Center at Brenda Mine (Station 2F18, 1520masl) and MacDonald Lake (Station 2F23, 1780masl)(Ministry of Environment, 2009b)) (Table 4).  29  Stream flow: Peachland  Stream flow Greata  Streamflow - Trepanier  SWE Brenda  Temp Brenda  Temp Peachland  Precipitation Brenda  Precipitation Peachland  Table 4: Data availability by year  Water Year 1973* 1974* 1975* 1976 1977 1978 1979 1980 1981 1982 1983 1984• 1985• 1986• 1987 1988 1989 1990 1991° 1992° 1993° 1994 1995 1996 1997 1998 1999 * = Years chosen for Greata Creek calibration; • = Years chosen for Trepanier Creek calibration and Greata Creek validation; ° = years chosen for Trepanier Creek validation. Water years begin in October of the previous calendar year (ie, Water Year 1974 = October 1973 – September 1974) Calibration and validation period were selected from the 3-year blocks where data were most complete (Table 4). For Trepanier Creek this was Oct 1983 – Sept 1986 (calibration) and Oct 1990 – Sept 1993 (validation). For Greata Creek this was Oct 1973 – Sept 76 (calibration) and Oct 1983 – Sept 1986 (validation). The mid 1970s and ‘80s were appropriate as a calibration period given land use conditions: less than 2% of the watersheds were harvested (Ministry of Forests and Range, 2008), meaning the hydrology reflected mostly “natural” land cover conditions. A few of the landscape parameters were adjusted in order to reflect the addition of some clearcuts present during the validation period. Additionally, the validation period was adjusted relative to the calibration period to reflect changes in interbasin diversions at the Brenda Mines site.  30  As a secondary validation step, the calibration parameters were used in the longest period of record for which climate data were available at the high and low elevation stations: Oct 1983 – Sept 1993. These time periods captured precipitation variability seen throughout the longest period of record available (1973-1993) for the Brenda Mine climate station. Since these are snow dominated watersheds, this station represents precipitation at the elevations most likely to affect streamflow. Total annual precipitation was calculated for each of the 20 years of data, which were then divided into quintiles, with the top quintile assigned as “very wet” then as “wet” “normal” “dry” and “very dry.” Precipitation was also described by how many standard deviations from the mean of the 20 years of data each annual precipitation value represents (Table 5). Thus, the calibration and validation periods represent a range of historic precipitation regimes. Table 5: Precipitation regime for the calibration and validation years Water Relative Number of standard deviations Year Precipitation from the mean (annual Quintile precipitation) 1984 Dry -0.46 1985 Dry -0.07 1986 Normal 0.40 1987 Very Dry -1.50 1988 Very Dry -1.40 1989 Very Wet 1.04 1990 Normal 0.31 1991 Wet 0.48 1992 Dry -0.89 1993 Very Wet 1.75 In order to complete the second phase of model validation which examined the 10-year period (1983-1993), it was necessary to fill in the missing data based on information from years with data at both the Peachland and Brenda Mines site. Gaps were filled with temperature and precipitation lapse rates data from years that were similar to the year with missing data in terms of their climate record. In the 1983 through 1993 period, data were missing for the Peachland station, but present for the Brenda Mines station. Thus, the amount of precipitation likely to be present at the lower elevations was calculated by applying a lapse rate to the Brenda Mines data. For 1986, lapse rates came from 1992; for 1987, used lapse came from 1985; for 1989, lapse rates came from 1983; for 1990, lapse rates came from 1983. For temperature, a record of historic annual mean temperatures (Environment Canada, 2009b) for the closest climate station for which these data were available (Summerland) was used to find which years were similar to the years with missing data in terms of annual mean temperature as well as with similar mean temperatures in winter and spring. For 1986, this was 1983, for winter and spring 1987, this was winter and spring 1983, for 1989, 1991 and for winter and spring 1990, this was winter and spring 1988. The lapse rate from those years was applied to the Brenda Mines temperature data.  31  3.5.2) Calibration of Snow Parameters The variables pertaining to energy in WEAP were calibrated so that they produced the best fit of snow accumulation depth and timing of melt with the SWE records from Brenda Mines and the timing and melt records from the aerial survey by Dobson (2003). Priority was given to generating an accurate portrayal of the timing of snowmelt, followed by an attempt to match the depth of SWE to the monthly snow surveys throughout the snow season.  3.5.3) Energy Parameter Specifications Freezing point Within WEAP, the parameter used to specify the temperature at which precipitation forms as snow was set at the default value of 0 °C. Melting point A melting point value of 4 or 5 °C has been used in other projects using WEAP to show that snow does not start to melt as soon as temperature reach 0.01 °C (Young et al., 2009). Setting the melting point above 0 °C may account for cold content: the fact that the temperature of the entire snowpack has to be brought to the melting point before melt can occur. A value of 4 °C was chosen for use here. Albedo An additional physical representation of the snowmelt process is provided through a decaying albedo parameter that represents “ripening” of the snow surface of old snow. A script was made available by model designers to model albedo decay over a weekly time step (Sieber, 2008). In order for the model to dynamically decay albedo on the daily time step in this model, the code needed to be rewritten. Given that it is not possible for the user to rewrite the model code within WEAP, an alternative option was used: creating “user-designed variables” in version 2.3053 of WEAP (code in Appendix A). The rate of decay used, 0.019/day, is the approximate rate of decay determined based on research by the U.S. Army Corps of Engineers (1956) and approximately the same rate as found from measurements in B.C. forests and clearcuts (albedo decay from 0.8 on April 1 to 0.2 by May 1 = 0.0194/day) (Boon, 2009). To account for the fact that sometimes snow fell after the albedo already began to decay, the formula was modified so that additional snow less than 10 mm did not bring the albedo value up to the high value (0.85) but kept the albedo at the value of the previous day before the snowfall. Final calibration of the melting rates required beginning albedo decay 5 days earlier than generated by the model.  32  Rnet_Other This parameter approximates “other energy” available to a snowpack for melt besides shortwave solar radiation and its use in this model is based on the observation that there is much more energy available to melt snow in the spring than provided by shortwave radiation alone (Young et. al. 2009). The model designers qualify Rnet_other as a lumped variable that represents sensible, latent, and advective energies (Young et al., 2009). The sum of net solar radiation and Rnet_other is then converted to a depth of melt and used in equations to determine accumulation and melt rate of the snowpack. The variable can also be used to represent an increase in solar radiation (C. Young, Senior Scientist at Stockholm Environment Institute, personal communication, 11/8/2008). It should be emphasized here that the purpose of this variable is to stimulate snowmelt rather than physically represent energy properties throughout the year. In order to calibrate the model with Rnet_other values that approximate physical realities during the time of snowmelt, aerodynamic resistance and vapour densities were calculated using curves from Oke (1987), based on 1983-1986 temperature and wind speed data from Kelowna Station A to determine what the average energy inputs from sensible and latent heat would be each 2  month. Average temperatures and wind speeds generated energy increases up to 20 MJ/m /day in 2  April and 80 MJ/m /day in July (Figure 5). These are averages and do not take conditions on individual days into account. On warm days (above 22°C) with high winds (above 10 m/s) this 2  advective energy can represent up to 250 MJ/m /day at the snow surface.  100 80  Average Wind (m/s)  60  MJ/m2/day  Temp (C)  40 20 0 -20  Oct  Nov Dec  Jan Feb Mar Apr May June  Jul  Aug Sept  Figure 5: Temperature and wind speed in 1984 and likely additional energy input Monthly averages. Data from the Kelowna A weather station. Rnet_other is a multiplication factor: the values put into WEAP by the user for Rnet_other are multiplied by the net solar radiation determined to be arriving at the snowpack during each time step (Figure 6). As seen in Figure 5, these other energies do appear to increase in a pattern similar to increases in solar radiation. Rnet_other only comes into effect once the net solar radiation is greater than zero, as can be seen in Figure 6. Net solar radiation is largely a function of albedo, so initial  33  snow decay in response to the albedo function, driven by the fact that new snow is no longer falling, is what drives the onset of melt in this model. 100  6  90  MJ/m2/day  70  Net Solar Radiation (T4L)  60  Actual Other Energy  50  Rnet Multiplier  4 3  40  2  30 20  Rnet multiplier  5  80  1  10 0 9/1/1985  8/1/1985  7/1/1985  6/1/1985  5/1/1985  4/1/1985  3/1/1985  2/1/1985  1/1/1985  12/1/1984  11/1/1984  10/1/1984  0  Date  Figure 6: Energy available for snowmelt as represented in WEAP Calculated in WEAP based on latitude, air temperature and relative humidity), Rnet_other multiplier (user-defined) and resulting “other” radiation at a sample HRU (T4L).  Adjustment by HRU Increasing Rnet_other decreases the SWE, with the most impact in April, May and June when most energy is available. Therefore, Rnet_other should vary depending on expectations for how fast snow will melt in the spring in each HRU. Rnet_other values were chosen to approximate the relative magnitude of energy likely to be available under different conditions. Since a greater amount of advective and other energy was assumed to be available in plateaus than north-facing slopes, HRUs characterized by plateaus were assigned the highest Rnet_other increase (Table 6). Table 6: Initial increase in Rnet_other to account for HRU-specific conditions Baseline Rnet multiplier is shown in Figure 6 (above). “*” Denotes a factor not present in the calibration period, but applied where applicable in the test period, and was used during the scenarios (see Section 3.7.3) Factor Impact Increase in Rnet South-facing slope Greater solar radiation during the winter + 0.5 and late spring East or WestSlightly greater solar radiation during +0.25 facings slope the winter and late spring Plateau Greater sensible/advective energy from +1 wind. Slightly greater solar radiation during the winter and late spring “Open” canopy (26- Greater sensible/advective energy from +0.5 60% cover) wind and greater exposure of ground surface to solar radiation Clearcut* (0% Greater sensible/advective energy from +2 cover) wind and greater exposure of ground surface to solar radiation  34  Since the choice of increase in Rnet_other was arbitrary and represented expected ranking of conditions rather than actual measured values, after these values were applied to each catchment they were adjusted so that model output for snow accumulation and depletion date approximately matched snow depth surveys (River Forecast Center) for Brenda Mines (1430m) and McDonald Lake (1720m). Additional adjustments were made for elevation ranges below 1400, 1400-1600 and above 1600, to reflect melt date values from 2002 aerial surveys (Dobson, 2003).  3.5.4) Landscape Parameter Specifications Once the model was calibrated to match snow data, the next step in calibration was to determine physically-realistic ranges of values for the landscape parameters, then select different combinations of those values which maximized the agreement between modeled and measured streamflow. The Trepanier Creek basin was calibrated first since a stream gauge provided data for the entire time period of interest. The calibrated values were then applied to the Peachland watershed and adjusted to optimize model performance in the Greata Creek tributary.  Root Zone Capacity This parameter represents soil pore space, measured as a depth, available for holding water. Several datasets were examined to determine soil characteristics in this area (British Columbia Ministry of Environment Land and Parks, 2009; Ministry of Energy, Mines and Petroleum Resources, 2009). None of the data sources gave “rooting depth” per se. In a “windshield” survey in the study watersheds, soil profiles were easily visible in cutout areas along the highways. The root zone appeared to be shallow with root masses up to a meter in depth. The Root Zone Capacity parameter should be set at less than that depth since only a fraction of a 1 meter deep rooting zone accounts for pore spaces available for holding water. This model optimized for values between 280 and 380mm of 3  3  potential water holding capacity (e.g. 0.28 and 0.38 cm of water for each cm of soil). Higher values were assigned to lowlands, and lower values assigned to HRUs with steep slopes.  Root Zone Conductivity This parameter represents saturated hydraulic conductivity. The dominant soil types in this area are sandy loams (Ministry of Energy, Mines and Petroleum Resources, 2009). Even for soils of a single classification, there is a wide range of possible hydraulic conductivities (0.36 mm/day for silty loams (Hanks & Ashcroft, 1980) to 8640 mm/day for silty sand (Hornberger, Raffensperger, Wiberg, & Eshleman, 1998)). Since it is not possible to find a physically-realistic value to represent soil across an HRU hundreds of hectares in size, selection of values for this parameter can only be chosen based on what produces the expected stream response in combination with the chosen root depth. Model outputs for baseflow and peakflow were found to be sensitive to changes in root zone  35  conductivity. A value of 0.75 mm/day worked best to optimize the magnitude of the peak flows and baseflow and was applied to all HRUs.  Deep Soil Capacity and Conductivity Little data exist on depth of soil below the root layer and associated conductivity. Hence, these parameters were tuned within a range of physically meaningful bounds. A soil depth (1700mm) and conductivity (1.5 mm/day) that generated the appropriate amount of baseflow were selected as baseline. Depth was then increased slightly in lowlands and decreased on steep slopes.  Preferred Flow Direction Preferred flow direction is a parameter specific to WEAP that is used to represent the general slope of landscape and therefore how much water flows as interflow or is transferred to the lower soil bucket. Here it was set higher in HRUs with steep slopes and lower on plateaus.  Vegetation Parameters The two main variables that describe vegetation function in WEAP include the crop coefficient, Kc, which is applied with the Penman-Monteith equation to describe the amount of water used by a plant in the watershed relative to a “reference crop,” and a Runoff Resistance Factor, described as integrating Leaf Area Index and slope in terms of how they affect surface runoff (Stockholm Environment Institute, 2007).  Crop Coefficient (Kc) Kc modifies the reference crop evaporation in the Food and Agriculture Organization (FAO) version of the Penman-Monteith equation, which represents the total demand for water given unlimited soil water availability (Food and Agricultural Organization, 1998a). WEAP is constructed to calculate the actual amount of evaporation (actual evapotranspiration, or AET) that will occur given the amount of water available in the soil at a given time step. Increases in Kc will increase the AET and decreases will decrease AET. Thus, Kc can be used to adjust how much water is used by vegetation in a watershed, and as a time variant. The FAO set the value for Kc in coniferous forests at 1, with the explanation that coniferous trees do not exhibit a distinct growing season like herbaceous crops (FAO, 1998b). However, the FAO suggests setting the value of Kc lower in the late or early growing season, when less transpiration is occurring. This reflects the fact that less AET occurs with less plant activity. In coniferous forests in temperate climates, such as are the conditions in the study watershed, the amount of transpiration does vary throughout the year. A study of coniferous forests in Western Montana found that when air temperatures are below freezing (0 °C), stomata no longer open and transpiration occurs through cuticles only, reducing the amount of water transpired (Running et al., 1989). This would suggest that Kc be set lower in the winter when both daytime and nighttime temperatures are below freezing. According to FAO, additional modifications to Kc should be made  36  based on the frequency of soil wetting. A study in lodgepole pine in Wyoming found that up to 40% of annual transpiration occurred during the spring drainage period (Knight et al., 1985), and similarly, in the study area, the soil can be expected to be the moistest in the spring during snow melt. Finally, FAO suggest that additional modifications be made to Kc to represent stressed conditions. In the summer, drought can lead to stomatal closure, (Chaves et al., 2002; Tognetti et al., 1999) meaning that transpiration is reduced as a function of plant processes as well as water availability. Thus, the best Kc curve for this area should include lower values in the winter, highest values in the spring, and lower values in the summer. A Kc curve was generated based on these considerations and calibrated to achieve the expected values of ET in the study watersheds.  Runoff Resistance Factor (RRF) WEAP does not explicitly account for plant processes, but does include a lumped parameter called the “Runoff Resistance Factor” (RRF) which, according to the WEAP User’s Guide (SEI, 2007), “reduces surface runoff by accounting for Leaf Area Index and slope”. This variable is designed to adjust modeled streamflow once all other parameters have been set so that it more closely replicates hydrometric data. It was assumed here that “resistance” to runoff is low when the ground is covered in snow and that it starts increasing after snowmelt due to growth of trees and understory which take up more water and slow the movement of water across the soil surface. Thus, the value was set lower in the winter and higher in the spring. The annual RRF curve was shifted up in HRUs with a “dense” vegetation category, and down with an “open” designation. Since the RRF variable also incorporates the effect of slope, HRUs with steeper slopes were given a lower RRF, and shallow slopes a higher one.  Z1 and Z2  Z1 and Z2 are parameters representing initial soil moisture content in upper and lower layers.  These values for were set for each HRU based on the average October 1 value which resulted after the model initialization period.  3.5.5) Calibration Adjustments in Peachland Creek Watershed It was not possible to calibrate the Peachland watershed for the 1983-1986 time period because no hydrometric gauge data at the watershed outlet are available from that time. An additional complication in the Peachland Creek watershed is the fact that the upstream Reservoir has been regulating flows for the entire period for which streamflow gauge data are available (19651982). Greata Creek, a tributary to Peachland Creek, does have a hydrometric station that has been recording flow data for the period for which climate data are available, and it was this subwatershed that was used to verify that calibrated model parameters done for Trepanier Creek were appropriate for Peachland Creek.  37  The Peachland model was set up with parameters from the Trepanier calibration and run with the same climate input data for the period from 1983-1986. When modeled Greata Creek flows were first compared to gauge data, flow was consistently over-estimated. Examination of the period from 1983-1993 revealed the same trend. Since the Peachland watershed is located to the south of the Trepanier watershed, it has been declared as in the “rain-shadow” relative to the location of the Brenda Mines climate station (Canada-British Columbia Okanagan Basin Agreement,1974). To compensate for this, the precipitation data were reduced by 0.3 mm per day of precipitation, representing about 10% of annual flow. This change resulted in improved model performance. Additional modifications to input parameters were made that resulted in an improved match: Root Zone Capacity: Increased to 450 mm Deep Soil Capacity: Decreased to 1400 mm Deep Soil Conductivity: Decreased to 0.75 mm/s Rnet_other: Increased in selected catchments to improve match with snow records and improve alignment of peak discharge. These values were applied to all HRUs in the Peachland Creek watershed.  38  3.5.6) Calibration and Validation Performance Metrics The success of a calibration can be measured using any of a number of performance metrics, either visual or statistical. A statistical measure designed specifically for analyzing goodness of fit in river discharge is the Nash Sutcliffe Efficiency Index (Nash & Sutcliffe, 1970). This index is based on the sum of square of errors between model and observations. Other measures often used in hydrology are Root Mean Square Error (RMSE), coefficient of correlation, coefficient of determination and Index of Agreement (Jain & Sudheer, 2008; Krause, Boyle, & Base, 2005). Each measure may put emphasis on different types of model behavior. The Nash-Sutcliffe index and other indices based on squared differences tend to focus on the goodness of fit during peak flows, and poorly represent improvements in baseflows (Krause et al., 2005). Performing the Nash-Sutcliffe Efficiency analysis on the natural log transformed stream data and model predictions improves the assessment of baseflows while detecting systematic over or under-prediction (Krause et al., 2005). According to Krause et al. (2005), “for scientific sound model calibration and validation a combination of different efficiency criteria complemented by the assessment of the absolute or relative volume error is recommended” (p.97). To assess the model accuracy here, calculations were done for Total Volume Error, NashSutcliffe Efficiency (NSE) and Nash-Sutcliffe performed on the natural log transformations (NSEln). Goodness of fit between modeled and observed SWE was assessed based on a visual assessment and on application of the Pearson’s correlation coefficient. In hydrology, NSE values over 0.65 are considered relatively high (Krause et al., 2005). Values between 0.5 and 0.6 are considered acceptable in a daily rainfall-runoff model (Young, 2006). According to Ingol-Blanco & McKinney (2009), in hydrology studies used for management purposes, total volume errors below 10% are very good, error between 10 and 20 is good, and between 20 and 30 is fair. Calibration and validation results are provided in Section 4.1.  39  3.6) Representing Human Water Demand 3.6.1) Data Availability and Limitations Water use in the District of Peachland and surrounding communities is not currently metered. There are no current measurements of either total amount of water used from sources, nor data on the proportion of water used by each type of activity throughout the year. Various consultants have attempted to come at these values indirectly, but there are discrepancies between estimates for total annual water use, and even less information indicating how much water might be used on a daily basis. WEAP is classified as being “demand-driven” in how it accounts for water use, but in the District of Peachland, the type of water use activity and water used by each activity is unknown. The greatest uncertainty exists regarding the amount of agricultural land actively being irrigated and amount of water being used for irrigation, and the amount of water being used for commercial and industrial purposes. Indirect methods have been used to determine amount of water used for domestic purposes in general, but there are no data on how much is used indoors vs. outdoors.  3.6.2) Developing the Model Data to populate the demand side of this model came largely from consultant reports specific to the study area. In some cases values came from studies done in surrounding areas. Since the objective driving each study was different, and since the methods for gathering and analyzing data were all different, there were significant discrepancies between most of values in these reports. It is beyond the scope of the present study to verify which of the reports provided the most “correct” estimate or to physically collect water use data, so the values in the consultant reports were used to provide a range of possible water use values. Two type of estimates were utilized: those driven by population and water use rate per “activity” (referred to hereafter as activity-based estimates), and those based on measurements of the amount of water extracted from each source during metering from 1999-2002, without information as to where it was then used (referred to hereafter as extractionbased estimates)(Table 7). Activity-based estimates, if data were completely available, would provide the most complete view of water use since use rates can be adjusted for household use rates, outdoor use, agricultural use, etc. Since these data were not consistently available, extraction-based estimates generate values for use by the study area in aggregate. Details on the sources of information used in generating the activity-based and extraction-based water use estimates are provided in Appendix B. Table 7: The two types of water use estimates used in this study “*” denotes multiplication Activity-based Extraction-based Data Population* use per capita indoor Metered withdrawal from each water sources source (1999-2002) Population* use per capita outdoor Agricultural land area * use per hectare Water licenses Estimate 3 3 3.24 million m per year 5.31 million m per year total  40  3.6.3) Location of Water Use Activity-based demand for water was divided among six different sites which have unique water sources and water uses (Table 8). The District of Peachland was divided into three sites based on system operations. Peachland System 1 is currently supplied by a combination of water from Trepanier Creek and Okanagan Lake, Peachland System 2 is entirely supplied by groundwater, and Peachland System 3 is supplied entirely by Peachland Creek (Summit Environmental Consultants, Inc., 2004). There are two nodes representing areas not serviced by the District of Peachland. These include higher priority license holders for water on Trepanier Creek, and the small communities of Star Place and Dietrich, also served by Trepanier Creek. Table 8: Water use activity by demand site Demand Site Water Use Activity Trepanier Licenses Domestic and Agriculture Star Place & Dietrich Domestic Peachland System 1 Domestic, Commercial/Institutional/Parks, Agriculture Peachland System 2  Domestic and Golf  Peachland System 3  Domestic, Commercial/Institutional/Parks, Agriculture  3.6.4) Values Used in Activity-based Water Use Estimates Population Peachland population data from 2008 (Economic Development Council, 2009) were used to represent “baseline conditions” (pop= 5,232). The approximate division of the population by water service area was obtained from District of Peachland Public Works staff (W. Grundy, personal communication, 3/31/2010). Approximately 2,616 (50%) are served by System 1, 1,046 on System 2 (20%), and 1,570 on System 3 (30%). Population in the area serviced by private licenses to Trepanier Creek was based on an estimate by Summit (2004). Population estimates for the Star Place and Dietrich system were based on Summit’s assessment that there are 8 residential lots on Star Place. The Okanagan Economic Development states that that 2.2 people/household is the average in Peachland (ECD, 2009), which would mean about 18 people are on this system. No estimates were given for Dietrich, so given that the license for the area was similar to that of Star Place, the same population was assumed.  Use per Capita Water metering took place on the inlet valves from all of the sources to the District of Peachland system from 1999 to 2002 (Urban Systems, 2005). Only total amount of water provided by each source was measured. Urban Systems (2005) used these values to estimate amount of 3  water used domestically per person in the District. They determined a rate of 0.900 m /person/day for 3  Peachland, which equates to about 329 m per person per year.  41  A report released in 2010 by the Okanagan Basin Water Board (2010) determined the annual 3  water use in municipalities in the Okanagan to be 0.675 m /person/day, which equates to 246 3  m3/cap/year. The OBWB split the amount of domestic use into indoor, 0.15 m /person/day and 0.525 3  3  3  m /person/day outdoor. This amounts to 55 m /person/year indoor and 192 m /person/year outdoor. Since the OBWB report’s data came from many sites throughout the Okanagan Basin, and since it is unlikely that domestic indoor use varies much between municipalities, the OBWB estimate of annual indoor use was used here. The remainder of the estimate of total annual water use from the Urban System’s study was assumed to be for outdoor use.  Agriculture  The Summit (2004) and Urban Systems (2005) reports employed methods for determining  agricultural water use which differed greatly, as did their results. The Summit (2004) report used Agriculture/Agrifoods Canada GIS data and determined that 28 hectares (ha) of irrigated agriculture exist in Peachland. The 2005 Urban Systems report estimated 161 irrigated ha of agriculture, which represented about 60% of total agricultural area. Current District staff confirmed that the higher of the two estimates is closer to reality since 122 ha are currently listed as having “farm” status, with an additional 35 ha in Agricultural Land Reserve that may or may not be actively farmed. (W. Grundy personal communication, 4/14/2010). The estimate by the current district staff was used here as the activity-based estimate since it represents expert opinion and is between these two values. In order to split agricultural land between System 1 & 3, the areas were proportioned according to address (provided by W. Grundy, 4/14/2010). Urban Systems estimated water use per acre to be slightly below the District’s allowed rate of 7 gallons/minute. In this study, Summit’s (2004) estimate was used instead since it is based on crop use determined by Agriculture-Agrifoods Canada and thus better represents an “activity-driven” value. 3  The average use rate across all crop types was 7,193 m /ha/year. This value was multiplied by the estimated agricultural land area per System to determine total demand for water for agriculture. Information on seasonal apportionment of water for agriculture came from Dobson (2006) who cited average values from throughout the region (Figure 7). These values were applied for all agriculture activities and to parks and outdoor domestic use. Oct 6% Sept 15%  Aug 26%  Apr 2%  May 1%  Jun 18%  Jul 32%  Figure 7: Water use by irrigation as a percentage of total use by agriculture (annual). Based on estimate for water use from average values from throughout the Okanagan (Dobson, 2006)  42  Commercial/Industrial and Parks Water use by the commercial and industrial sector is even less understood than water use by agriculture. Both Summit (2004) and Urban Systems (2005) used indirect methods of calculation only. Urban Systems (2005) used GIS data to determine which lands were neither residential nor agricultural and assumed those represented commercial or park areas. Use values were only determined for the month of maximum use. Summit (2004) used the domestic rate for water use (which includes indoor and outdoor use) and applied it to number of employees in each business sector in Peachland. Since this method appeared to be double-counting water use, and produced an estimate larger than municipalities twice the size of Peachland (10-times higher than Westbank, population 24,600), it was not included here. Urban Systems’ estimate of commercial/industrial use was one third the amount of domestic indoor on the day of maximum use. This implies that commercial use is likely mostly indoor use since: 1) The 2001 census recorded 2,400 employees in Peachland, which is about half the population contributing to the domestic indoor estimate, 2) if the commercial estimate included outdoor watering the maximum day estimate would likely be more than the estimate for outdoor use. Thus, it was assumed that water use by the commercial/industrial sector was the same all year as it was on the day of maximum use. 3  For water use in parks, Urban Systems (2005) estimated that 6,820 m are used in July. Here it was assumed that most of this was for outdoor watering, so, given the same monthly apportionment as for agricultural irrigation, an annual total was derived from the assumption that the July value represented one third of use. Half of the total value was applied to Peachland System 1 and half to Peachland System 3 since there are no parks in the Peachland System 2 area. 30%  Monthly apportionment of water use: activity-based estimate  25% 20% 15% 10% 5% 0% Oct  Nov  Dec  Jan  Feb  Mar  Apr  May  Jun  Jul  Aug  Sept  Figure 8: Monthly apportionment of annual water use summed for all water users: activitybased water use estimate  43  3.6.5) Values Used in Extraction-Based Water Use Estimates Summit (2004) and Urban Systems (2005) did not provide complete activity-driven data for calculating water use. Instead they used the data they had available to estimate water used, which included water license information and calculations of water use by source. In this study, these values were used as another type of estimate for water use. For all demand sites, the extractionbased estimate was based on the highest estimate from each of the consultant reports. On Peachland System 1, this was based on the Summit (2004) estimate of water used for “waterworks” from Trepanier Creek, the Dobson (2006) estimate of use of Okanagan Lake use for “waterworks,” and the Dobson (2006) estimate of use of Trepanier Creek for agriculture. For Peachland System 2, it was based on the Summit (2004) estimate of total extraction capacity from the wells. For Peachland System 3, it was based on the Dobson (2006) estimate for total use of Peachland Creek. For the Trepanier Licenses and the Star Place and Deitrich systems, not even an extraction value has been measured, so the total licenses for those systems were used as the upper bound (see Appendix B for summary of activity-based and extraction-based water use estimates and sources of information). The monthly distribution of the extraction-based estimate was formed based on the average distribution of water extractions measured during the monitoring period (1999-2002) for Peachland Creek (Dobson, 2006) (Figure 9). Note that values in May are higher than expected based on irrigation data. These may partially represent the fact that Peachland Creek supplemented the system normally fed by Trepanier Creek at that time, but also more closely represent system-wide demand at the time monitoring took place (Urban Systems, 2005). 20% 18%  Monthly apportionment of water use: extraction-based estimate  16% 14% 12% 10% 8% 6% 4% 2% 0% Oct  Nov  Dec  Jan  Feb  Mar  Apr  May  Jun  Jul  Aug  Sep  Figure 9: Monthly apportionment of annual water use from Peachland Creek: extraction-based estimate. Based on water extraction metering on Peachland Creek, 1999-2002 (Urban Systems, 2005).  44  3.6.6) Connecting Supply and Demand Trepanier Creek was set as the sole source for Trepanier Licenses and Star Place-Dietrich. Trepanier Creek was also assigned top priority as the source for Peachland System 1, with the exception of during April through July. System 1 uses Okanagan Lake to supplement its requirements during the peak flow period when the water is too sediment laden to be adequately treated by the current treatment system (D. Allin, Public Works Director, District of Peachland, personal communication, October 26, 2009). Therefore, Okanagan Lake was given top priority during peak flow months so that the licensed volume from the lake was used before the stream. Groundwater was established as the only supply for Peachland 2. Since no data are available on recharge, it was assumed to meet all current and future demands. Peachland Creek comprises the only source for Peachland System 3.  3.7) Representing Instream Flow Targets The ability of streamflow to meet instream flow targets on Trepanier and Peachland Creeks under baseline conditions was used as a reference when analyzing scenarios. Instream flow nodes were set at the furthest downstream point on each creek, below the District of Peachland intakes, in order to see if the requirements are met in the areas likely to be the most stressed. Two types of targets were assigned to Trepanier Creek, given the two main ways B.C. agencies currently quantify flows for aquatic life needs: conservation targets established by NHC (2001) based on Mean Annual Discharge (MAD), and BCIFN targets based on mean daily flows as per Hatfield et al. (2003) (Figure 10). Since unregulated flow data were not available to calculate BCIFN targets on Peachland Creek, only the MAD target was used (Figure 10).  2.5 Peachland Creek Conservation Flow Target  2  Trepanier Creek Conservation Flow Target  1.5  cms  Trepanier BCIFN Threshold  1 0.5  9/1  8/1  7/1  6/1  5/1  4/1  3/1  2/1  1/1  12/1  11/1  10/1  0  Figure 10: Conservation flow targets (MAD) for Trepanier and Peachland Creeks, and BCIFN target for Trepanier Creek  45  3.8) Developing Scenarios Once this model was developed to represent “current” conditions with the greatest degree of accuracy possible, scenarios for changes to those conditions could be developed. Scenarios for changes in climate, land use, population growth and water use are described here. All supply change scenarios were perturbations of the 9-year time series (1984-1993) used in calibration and validation given that it represents a range of climatic conditions ensuring that perturbations will also represent a variety of climatic conditions.  3.8.1) Climate Change Since there has been such a wide range in the results produced by each global climate model and the likelihood of each occurring is deemed equally valid, the IPCC (Kundzewicz et al., 2007) recommends that the results from more than one model be considered in conducting climate change impacts work. Here, two GCMs and one SRES emissions scenario were used. Data are available for the grid cell above the Okanagan Basin from the Canadian Climate Impacts Scenarios Group for 15 GCMs (CCIS Project, 2003). The Canadian Global Coupled Model, version 2 (CGCM2) provided the most conservative estimates of impacts of climate change in a modeling study by Merritt & Alila (2004) and the HadCM3 showed greatest summer precipitation changes and greatest increase in summer mean temperatures. Therefore, these two models were deemed appropriate show a wide range of possible impacts. Selection of an emission scenario was based on a desire to show the most extreme impacts likely under future climate change. The A2 SRES scenario represents mid- to high-range projection for CO2 emissions (Nakicenovic et al. 2001) and thus was chosen to examine what may happen to water resources under more extreme conditions. Two time slices; the 2020s (2010-2039) and the 2050s (2040-2069) were used. In this study, the only GCM output variables used were temperature and precipitation. No other modifications were made to vegetation parameters to account for climate change. For example, potential reductions in evapotranspiration due to CO2 fertilization (Gedney et al., 2006) were beyond the scope of this study. The dominant role of temperature and precipitation in snow-dominated watersheds like those of Trepanier and Peachland Creek, described in Section 2.1.2, combined with the fact that precipitation and temperature are the simplest to represent in WEAP, were the drivers behind this choice. In order to downscale the regional GCMs to the study area, the “delta” method was used. Each application of the delta method begins with development of a Normal time series to use as a baseline. Many climate studies use average monthly values from the time period 1961 to 1990 as the “normal” baseline time period (CCIS Project, 2003; Wang et al., 2006). Since this study uses a daily time step, use of monthly averages was not appropriate, particularly for precipitation data. Instead, daily data from 1983 to 1993 was used as the baseline which was then perturbed.  46  Accounting for Temperature Change A grid of temperature change data for British Columbia was developed through the ClimateBC project that is more representative of local conditions than the raw GCM grid cell data (Wang, Spittlehouse, & Aitken, 2006). The data include climate features such as rain shadows, temperature inversions, slope and aspect effects at a scale of several kilometers, and lapse-rate driven temperature differences as a function of elevation represented at the scale of hundreds of meters (Mbogga, Hamaan, & Wang, 2009). Temperature data for the Normal period (1961-1990) along with values from the GCMs and scenarios were downloaded from the ClimateBC web interface for the central latitude, longitude and elevation of each of the subwatersheds (T 1-6 and P 1-4). Since differences between subwatersheds were found to be insignificant, the average change was applied to all HRUs. The process was repeated for each of the GCMs and time slices to find the difference between Normal and future for each month. The expected temperature change for each month was then applied to the daily timeseries. The lapse rates between the low elevation and high elevation station were not modified. To save on computation time, instead of starting with a baseline year and running the model forward in time, scenarios were examined by perturbing the climate inputs for the 9-year period (1983-1993) and applying the expected changes to model forcings that would represent either 2020s or 2050s climate conditions.  Accounting for Precipitation Change Since precipitation data were not available from the ClimateBC project, precipitation predictions at the Okanagan-wide scale were used to perturb climate data in this study. Okanagan GCM grid data are available from the Canadian Climate Impacts Scenarios website (CCIS Project, 2003). Values for potential changes in precipitation are given in percent change, thus, the method for applying GCM values to regional conditions was to perturb the local daily precipitation values for days with precipitation at each HRU with the average value given for each month. This method creates a scenario where only total volume of precipitation will change, not number of days with precipitation, implying that historical precipitation characteristics will be preserved over time. However, since historical precipitation characteristics are not likely to be preserved under future climate change (Field et al., 2007), a drought sequence was also developed as one way to explore the impacts in the change in mean as well as total volume of precipitation. Projections for changes in temperature and precipitation for the study area in the 2050s time slice are shown in Figure 11.  47  60  4  40  2  20  0  0  -2  -20  -4  -40  -6  Change in precipitation (%)  Temperature (C)  6  Temperature: CGCM2 A2 Temperature: HadCM3 A2 Precip: CGCM2 A2 Precip: HadCM3 A2  -60 Oct  Nov  Dec  Jan  Feb March Apr  May  Jun  Jul  Aug  Sept  Month  Figure 11: Temperature change data (°C) and precipitation change data (% change from baseline) from the 2050s time slice as downscaled to the study area. Includes data from the CGCM2 and HadCM3 climate models, A2 emissions scenario. Temperature data for coordinates Lat. = 49 52 58, Long. = 119 48 35, available from the ClimateBC project (Wang et al. 2003). Precipitation data for the Okanagan GCM grid (CCIS Project, 2003).  3.8.2) Drought and Climate Change Conditions that are most difficult for water managers and water users occur when there is a sequence of many dry years (for example, there will not be enough water to fill the reservoir and hold over to the next year; fruit trees become more stressed after a sequence of droughts, etc.). Yet it is this subtlety of determining what future precipitation patterns could be and how they will vary year to year that is the most difficult to determine at even a regional level. Therefore, in order to look at extreme, rather than “expected” conditions, a scenario that looks at the system under a sequence of drought years was developed. To construct this scenario, first, 2003 precipitation, wind speed and humidity data were obtained from the Environment Canada HYDAT database (2007). Since there were not data available for Brenda Mines in 2003, a precipitation lapse rate derived from actual data could not be used to calculate precipitation at each HRU. Instead, a lapse rate was chosen from a year where data were available at Peachland and Brenda Mines stations and where both were considered “dry” years: the 1985 water year. Only the days for which there was recorded precipitation at Peachland received precipitation at Brenda Mines and all elevation bands in between. Although this is not consistent with general patterns (there was precipitation on 14% more days at Brenda Mines over the period 1983-1993), it did capture the fact that a greater volume of water falls at the upper station (45% more at upper station 1983-1993). To account for the drought occurring in the context of climate change, temperature data during the drought were taken from the series of data for each HRU developed using the CGCM2 CGM (2020s).  48  th  th  th  Next, a sequence of “years like 2003” was placed in the 4 , 5 , and 6 year of the 9-year scenario. That is, the precipitation, wind speed, and humidity data from 2003 were entered for those years. For the remaining years, precipitation and temperature data from the climate change scenario for the 2020s, using the CGCM2 GCM was used.  3.8.3) Land Cover Change: Mountain Pine Beetle Attack The Mountain Pine Beetle (MPB) infestation currently underway in British Columbia is projected to continue to affect areas of live pine in areas on the periphery of the outbreak, including the Okanagan, through at least 2012 (Walton, 2010). As such, hydrologic impacts will be increasingly seen in the short term. However, since the greatest hydrologic impacts are not expected until 15 to 20 years post-attack (Huggard & Lewis, 2007), a more reasonable scenario is a combination of pine beetle attack with projected climate changes in the 2020s time period. Here, a scenario representing the maximum hydrologic impact of the Mountain Pine Beetle attack was developed for baseline climate conditions in order to examine the impacts of the attack alone, and a second scenario was developed for Mountain Pine Beetle attack in addition to climate change projections for the 2020s. In both scenarios, an MPB attack was set to occur in all mature pine (as designated by the Ministry of Forests and Range 2007 GIS shapefile) in all HRUs. The first year of the simulation was set as 15 years after an attack. The attack was represented by a 50% increase in Equivalent Clearcut Area (ECA) for all pine stands, as projected by Huggard & Lewis (2007), for an area 100% affected in the first year of the attack. All nine years of the scenario were equally representative of th  this 15 year post-attack so that they could represent what conditions would be like under different baseline precipitation and temperature conditions. Since WEAP does not physically represent vegetation canopy, it cannot represent precipitation interception, understory vs. overstory water use and other meteorological and biophysical variables that would result from loss of forest canopy, it therefore does not allow for directly representing stand-level vegetation changes. Instead, broad, processed-based changes were represented based on generalized assumptions about impacts. While not predictive, this approach gives an estimate of the types of changes that might occur due to land cover change in the study watershed. Given that this scenario is set up based on ECA, it could also represent logging taking place in an equivalent area in the watersheds. Such a representation is appropriate given that salvage logging typically follows a pine beetle outbreak.  Snow Accumulation Impacts In this study, the ECA of each HRU was determined by reducing the area mature pine in the HRU by 50% to represent the canopy loss 15 years post-attack. Reduction in interception due to loss of mature canopy was simulated in this study through an increase in precipitation. The studies reviewed by Winkler et al. (2009) suggest that there will be about a 0.4% increase in snow water  49  equivalent (SWE) for every 1% forest canopy lost. Therefore, the percentage of the HRU without canopy was multiplied by 0.4 to determine the increase in SWE. The first step in determining the quantity of precipitation that would lead to that expected increase in maximum SWE was selection of a high, medium and low elevation HRU (T 4L; T 2M; T 5H2). The average precipitation value (over 3 years) that would represent a 10% increase in their maximum SWE was determined. For T 4L, this was about 30 mm, which, if applied across the 70 days of precipitation over the 157 days up to the date of maximum accumulation, would be 0.43 mm per day of precipitation. In the mid-elevation HRUs, the value was 0.47 mm/day, and in the high elevations, 0.54 mm/day. The proportion of each HRU that was “clearcut-like” was multiplied by the appropriate increase in precipitation to get the desired increase in maximum SWE (for example, if a 5% increase in SWE was expected in a high elevation HRU, the precipitation data series in that HRU would be increased by half of what was expected for 10% increase, 0.54 mm, leading to an increase by 0.27 mm for each day of precipitation) (Table 9). Table 9: Example of how the amount of lodgepole pine in an HRU translates to additional precipitation under the MPB scenario (example assumes a medium-elevation HRU) Percent Percent forest Percent Expected Increase in precipitation Mature Pine remaining 15 years “Clearcut increase in for each day of (baseline) post-attack like” SWE (%) precipitation (mm/day) 15 93 7 3 0.14 50 75 25 10 0.47 80 60 40 16 0.71  Snow Melt Impacts The rate of snowmelt during the melt season is predicted to be from 2 – 10 mm/day faster in grey-attacked stands than forested stands due to increases in radiative and advective energies (Winkler et al., 2009). In the present study, the conditions driving increased snowmelt, or increase in the energy available for melt, was simulated using the Rnet_other variable. In model set-up, the additional energy for a clearcut was represented by an increase in Rnet_other by 2 between April 15 and August 27 (days of equivalent solar radiation), with gradual increase starting February 25. To determine how much to increase Rnet_other in the HRUs, the percentage of each HRU that was determined to be “clearcut-like” in the Mountain Pine Beetle scenario was multiplied by 2. For each HRU, this value was applied in addition to the adjustments of topography and the relative openness of the rest of the canopy as determined during the calibration. Additional energy (+0.5 Rnet_other) was applied to south-facing catchments to account for the strong impact expected in these catchments due to reduced canopy.  Evapotranspiration Impacts Spittlehouse (2006) found that 30% less actual evapotranspiration (AET) occurs in a clearcut than in a forest over the course of a year. Here, this value was multiplied by %ECA to determine total reduction in evapotranspiration for each HRU. For example, in an HRU that was 50% pine, a total of  50  25% of the catchment would be clearcut-like, and therefore 25% of the catchment would have 30% less evapotranspiration, resulting in a 7.5% reduction in AET for the whole HRU. Through tests of the relationship between Kc and AET throughout a 9-year period, the relationship between Kc and AET reduction was established (  Table 10). These values  were used to decrease the entire Kc curve for each HRU, depending on affected area, to create the expected decreases in AET. The specified Kc was applied to create expected AET reductions in these ranges. For example, an expected reduction of 4% was created by a 0.05 reduction in Kc. Table 10: Reduction in Kc and corresponding reduction in AET (Based on test of catchments T 4L, 2M and 5H2, 1984-1993) Kc reduction AET reduction (annually) -0.05 3 – 5% -0.1 8 – 11% -0.2 14 – 23% -0.3 24 – 37% -0.4 39 – 52%  Wind Speed Increases Wind speed affects the rate of evaporation through the Penman-Monteith equation. In a sensitivity test performed by increasing wind speed +/- 10% over the course of three years, wind speed was found to have a minor, though noticeable impact on streamflow. When canopy is lost due to insect-kill or logging, wind speeds increase in the affected area. A study in Central B.C. by Boon (2007) found the mean wind speed in a clearcut stand to be about 0.36 m/s greater than in a live stand. In this study, an increase in wind speed of 0.5 m/s was assumed for an HRU entirely killed by pine beetle. The value for each HRU was determined at 0.5 times the ECA.  Mountain Pine Beetle and Climate Change For the scenario where a Mountain Pine Beetle attack occurred in the context of climate change, precipitation values were determined by first altering precipitation based on projected climate change influence, then applying the increase derived from the ECA. Temperature values from the CGCM2 scenario were used here. All other input values were the same as in MPB Only.  51  3.8.4) Management: Water Demand Where possible, the scenarios for future demand were based on existing information. Developing new demand scenarios should involve considerable participation from water managers and water users, which was is outside the scope of this project. Future demand was estimated for two time periods; the 2020s and the 2050s. Demand values were applied uniformly across the modeled scenarios involving different climatic conditions.  Assumptions In all scenarios, it was assumed that all demand for Systems 1 and 3 is met by surface water sources rather than groundwater, and that no increased extraction would occur from Okanagan Lake. This assumption is based on the fact that licensing decisions are the subject of Basin-wide negotiations which are outside the consideration of this study. It was assumed that only the amount of Okanagan Lake water licensed to the District of Peachland is used, rather than the use values reported by Urban Systems, 1999-2002. It was assumed that no new licenses would be added to meet demand, given the stipulation that all reasonable alternatives be pursued from existing licensed sources before permission is granted for new licenses (Summit, 2004). Since a maximum use based on licenses was not specified in the model, all final demand scenarios were checked against the licensed value to assure that they did not exceed that license. No hydrologically-significant change was assumed to occur in land use in the District of Peachland. The land currently in agriculture was assumed to remain in agriculture given the objectives of maintaining rural character and directing growth to existing urban areas as outlined in the Official Community Plan (District of Peachland, 2003). The current plan by the District of Peachland is that by 2025, more water is going to be supplied to the District by Peachland Creek and less by Trepanier Creek (Urban Systems, 2007; D. Allin, pers comm.. 2009)). Therefore, a future change scenario was modeled for the study area where Peachland Creek provides supply for both Systems 1 and 3.  Population Growth Population growth rate was determined for the 2020s and 2050s. Population growth estimates are based on the community build-out projections that are in Official Community Plans prepared by local governments (Summit, 2004). These plans projected a 35% increase (relative to 2001) in Peachland’s population by 2020 (to 6,277) and a 90% increase by 2050 (to 8,840). The population served by private water licenses on Trepanier Creek is projected to decrease to 100 by 2020 and 87 by 2050. Changes in population were applied proportionally to its current distribution in Systems 1, 2 and 3. Since population could not be directly accounted for in the extraction-based estimate, population growth was incorporated into the total demand estimate (see “Extraction-Based Increases”).  52  Use Rates Rate of water use was changed in all future scenarios in order to reflect predicted increases in demand that come along with population growth and the warmer temperatures and increased evaporative demand associated with climate change. Climate-related increases were applied to agricultural and domestic outdoor use, and population-related increases were applied to the domestic uses. A combined use increase value was applied in the extraction-based scenarios as described below. Since the main goal of the Mountain Pine Beetle Only scenario was to assess the hydrologic impacts of the loss of canopy without the influence of climate change, it was only paired against population growth related increases, not climate-related increases.  Domestic The domestic use rate for Peachland estimated from the Urban Systems (2005) report and used in the model base scenario is 33% higher than per capita use rates estimated for the Okanagan region as a whole (DHI, 2010). The domestic use rate for Peachland was therefore not increased in the future scenarios. The indoor use rate per capita was not changed, as indoor use is not projected to change under climate change (Maurer, 2010). For the activity-driven estimates at Trepanier Licenses and Star Place/Dietrich, projected increases in domestic outdoor use were applied, based on estimates by Summit (2004), who projected that climate change will increase outdoor use 16% by 2020 and 30% by 2050. The distribution of monthly use as a percentage of annual total was set to be the same for outdoor use and agriculture as under baseline conditions, so that a proportionally higher amount of that increase occurred during the months of most outdoor use.  Agriculture Agriculture/Agrifood Canada estimates that total water consumption by agriculture will increase 16% by 2020 and 30% by 2050s due to climate-related increases to evaporative demands alone (Nielsen et al., 2006). To account for these changes, the activity-based use rate was increased by the respective percentages. The monthly distribution was the same as under baseline conditions.  Commercial/Industrial and Parks Summit (2004) estimated that future use in Peachland will increase by 50% by the 2020s and 132% by 2050s due to growth in this sector. The estimates used in the baseline scenario for commercial/institutional were increased by these percentages. Projections for increases in use by golf courses and parks were based on increased evaporative demand due to climate change only: 16% by 2020, 30% by 2050 (Summit, 2004).  53  Extraction-Based Increases For the extraction-based water use increases, it was not possible to differentiate increases by sector. Thus, for the extraction-based estimate, modifications were made to account for the effects of climate change and population growth indirectly. Demand in November and December was assumed to account for domestic and commercial indoor use. Therefore, baseline November and December values were increased by the projected population increase, 35% for the 2020s and 90% for the 2050s, and applied all year as indoor use values. The assumed baseline indoor use value was subtracted from baseline extraction-based estimate with the remainder assumed to be for outdoor use. That amount was increased by the projected climate-related increase of 16% for the 2020s and 30% for the 2050s. Projected future indoor and outdoor use values were then summed. These calculations resulted in a 23% increase in the total water demand in the 2020s and a 47% increase in the 2050s. The resulting future projections for extraction-based and activity-based water use estimates differed due to the differing base data and methods for projecting increases. Baseline and future estimates are shown in Appendix C. Most notably, as with the baseline data, activity-based estimates exceed extraction-based estimates June through September on Peachland System 1, but are lower on System 3. This may reflect the fact that Peachland Creek was supplementing water needs on the system usually fed by Trepanier Creek during the years where the metering took place (1999-2002). The activity-based estimates reflect how much System 1 should require based on water use activities only.  3.8.5) Management: Reservoir Operations The role of the Peachland Reservoir was found to be critical in how downstream water users on Peachland Creek were affected by changes in supply. Therefore, two reservoir management scenarios were developed to examine the impacts on human and instream users. The first reservoir management scenario assumes “typical” release rates under current District of Peachland water management operating procedures. These rates are set to support instream flows in the winter and release freshet waters in the spring. Since the first simulation using these rates showed lower reservoir volumes at the end of the simulation run compared to the beginning, this scenario run starts with a low reservoir volume based on the assumption that it had been slowly depleted by this level of release prior to the beginning of the simulation. The next reservoir management scenario assumes that water managers would want to keep as much water in the reservoir as possible to meet Peachland’s future water use needs. Therefore, the release to meet instream flow needs was reduced to approximately half of the release in the scenario described above, and consequently initial reservoir volumes were assumed to be higher.  54  3.9) Metrics of Change The choice of metrics by which to evaluate model output was based on the following objectives: a) assessing if the hydrology varied from baseline conditions; b) assessing hydrologic response to climate change and land use change; c) providing results applicable to addressing water management questions. The metrics used to evaluate change are: •  Maximum SWE in each HRU, measured each water year. This metric is helpful in understanding the role of snow in hydrologic response to the change variables. This metric is often used by forest hydrologists in assessing the effects of land use change and thus is applicable for comparing model output to other studies.  •  Date of snow depletion in each HRU, measured each water year. This metric reveals influence of energy parameters on snow processes.  •  Mean annual stream discharge. This metric shows the integrated watershed response to change parameters. In order to determine the effects of the hydrologic change without human alteration, results were summarized for the unregulated Trepanier Creek, above the points of withdrawal, and for the tributaries contributing to Peachland Reservoir.  •  Difference in cumulative monthly flow volume. Difference between scenario and baseline. Provides information on water available for municipal use. Analyzed in Trepanier Creek above withdrawals and for the tributaries contributing to Peachland Reservoir.  •  Median daily discharge, by month. Central tendency of streamflow (cms). Relevant to water management operations and ecological functions. Analyzed in Trepanier Creek above withdrawals and for the tributaries contributing to Peachland Reservoir.  •  3  Peak flow volume (total). Total discharge (m ) over the week of maximum flow (calculated using a moving-average). Analyzed in Trepanier Creek and the tributaries contributing to Peachland Reservoir. It should be noted that this is just one means of assessing when the majority of the spring freshet takes place. In most years the highest week of flow represents the center of the peak flow, but in some years several large peaks occur.  •  Peak flow timing. Start and end date of the week of maximum flow for each water year. Determined for Trepanier and the tributaries contributing to Peachland Reservoir.  •  Low flow discharge. Average cms of low flows (defined as lower than 30% of flows from the period of record). Determine using IHA software (The Nature Conservancy, 2009).  •  Dates unmet. Dates where there was less water delivered than demanded (deficit) in the District of Peachland. 3  •  Amount of deficit. Volume (m ) of the water deficit.  •  Total demand required at time of deficit. Total volume of water demanded during the period when there was a deficit.  55  •  Percent of demand met. Total amount of water demanded minus amount of deficit divided by total water demanded. This metric shows how severe the deficit was.  •  Increase in demand due to population growth and climate change during the period of deficit.  •  Demand increase ÷ deficit. Volume of increase in demand due to climate change and population growth divided by volume water deficit. This metric is used to represent the amount of the deficit that could be attributed to demand-related increases alone. Where this value is below 100%, the change cannot be attributed to demand change alone but also was influenced by a decrease in supply.  •  3  Change in carry-over storage. Storage (m ) in Peachland Reservoir at the end of the water year minus the beginning of the water year. Used to determine if storage increased or decreased over the course of the year.  •  Number of days instream flow targets met. Calculated for each month in a “very wet,” “normal,” and “dry” year (based on conditions in baseline).  •  Difference between flow delivered and target (% met): An assessment how far below the target the streamflow is each month in a “very wet,” “normal,” and “dry” year.  •  Number of days with no flow in stream. Calculated as annual total.  •  Number of years in simulation where flows reach 400% MAD. Assessment of whether the recommendation that flows above 400% of the MAD be met once every 2 years (NHC, 2001) can be met.  Difference in mean maximum snow accumulation (SWE) was assessed using one-tailed paired ttests in order to determine if maximum SWE increased or decreased significantly between baseline and scenario. Total annual stream discharge in Trepanier Creek above withdrawals, and in Peachland Creek above the reservoir was also tested using one-tailed paired t-tests to determine if there was an increase or decrease in flow in each scenario. Assessment of hydrologic variability in daily discharge was conducted using non-parametric statistics to avoid problems associated with assumed normality of data. These tests were performed using the Indicators of Hydrologic Alteration (IHA) software, version 7.1 (The Nature Conservancy, 2009). IHA allows comparisons of pre-impact and post-impact hydrologic data. It can be used to compare flows from two different stream gauges or from two different model runs or scenarios (The Nature Conservancy, 2009). One way IHA quantifies differences between pre-and post-impact datasets is by generating significance counts, which can be considered as somewhat analogous to pvalue. These counts are generated by randomly shuffling pre and post-impact median values for each metric (e.g., October low flows) 1000 times. The significance count is calculated as the percent of all simulations (out of 1000) that result in a deviation factor (percent change between pre-impact and post-impact) greater than the observed data. Low values are most significant (Zimmerman, 2006). Significance counts in the lowest quartile (e.g. counts < 0.25) will be presented as indicating the highest confidence of change in this study.  56  Chapter 4: Results 4.1) Calibration and Validation Results The model performance was evaluated by comparing measured and modeled values for streamflow and for snow-water equivalent in the snowpack. The modeled SWE match measured SWE well for all calibration and validation. It should be noted that snow accumulation and melt were not only calibrated to the Brenda Mines station (for which the index below was calculated), but also to “timing of melt” data from aerial fly-over surveys. The whole model was calibrated so that best fit for timing of melt was found for all 39 HRUs. Modeled streamflow was optimized to the measured streamflow for the entire range of flow conditions. The NSE tests performed on the log-transformed data (NSEln) resulted in values generally above 0.7 (Table 11) suggesting that there is a good match between modeled and measured baseflow and across the hydrograph during all calibration and validation periods. Full hydrologic model calibration and validation performance metrics are provided in Table 11. Measured vs. modeled calibration and validation hydrographs are shown in Figure 12 through Figure 16. Table 11: Calibration and Validation: Performance Metrics Trepanier – Trepanier Trepanier Greata – Greata – Calibration Validation Validation Calibration Validation (`84-`86) (`91-`93) (`84-`93) (`74-`76) (`84-`86) SWE: Brenda Mines and Similar-Elevation 0.94 0.93* 0.88 0.90 0.95 2 HRU (R ) NSE 0.55 0.51 0.42 0.61 0.42 NSEln 0.76 0.71 0.72 0.72 0.66 Streamflow: Total Volume Error (%): +3 - 11 -1 +13 +7 Average Streamflow: Total Volume Error (%): -9 to +1 -30 to +5 -31 to +30 +7 to +22 -32 to +37 Range Baseflow (Oct. 1 to March 1): Total -1 -11 -3 + 28 +51 Volume Error (%): Average Baseflow (Oct. 1 to +21 to March 1): Total -44 to +42 -33 to +19 - 38 to +56 -13 to + 55 +105 Volume Error: Range *SWE data were not available for 1993, so the correlation was performed between 90-92 only  57  Figure 12: Calibration results from Trepanier Creek (1984 – 1986)  Figure 13: Validation results from Trepanier Creek (1991 – 1993)  58  Figure 14: Validation Results, Trepanier Creek (1983 – 1993)  Figure 15: Calibration Results, Greata Creek (1973 - 1976)  Figure 16: Validation Results, Greata Creek (1983 - 1986)  59  The Nash-Sutcliffe index indicates that the model matches the observed data acceptably over the calibration and 3-year validation period for Trepanier Creek (NSE = 0.42 to 0.61 and NSEln = 0.66 to 0.76). These results can be interpreted when accompanied by a visual assessment of measured and modeled hydrographs ( Figure 12 to Figure 16). The index is likely responding to the fact that although the peaks occur at about the right time and baseflow occurs at about the right time, correspondence between flows on a given day during the peak or baseflow period is never exact. The NSE is lower when applied across the longer test period. This fits with the analysis of Young (2006) who asserted that since NSE was created for event-based assessments it tends to decrease as the length of the simulation increases. In this simulation, there was greater flow variability over the longer period that was not captured as well by the model. An additional consideration for model performance over this longer time period is the fact that some climate data were missing for the longer test period and thus the hydrograph was not reflecting data from the entire time period. Sufficient length of flow record was not available to permit calibration and testing on separate 9 year records, which would have facilitated a calibration to a wider range of flow characteristics. On Trepanier Creek, the total volume error averaged over the 10-year test period is only -1%. The modeled flow is greater than observed during six of the 10 years and less in 4 years. When seen as an aggregate, the model neither under nor over-predicts the total flow volumes. Total volume error is “very good” (below 10%) during four years, “good” (between 10 and 20%) during two years, and “fair” (20 to 30%) during the remaining four years. The model tends to under-predict peak flow, with seven years showing less flow in the model than was observed, and slightly over-predicts baseflow, with six of ten years over-estimating baseflow. Overall, the range of match between measured and modeled total volumes across the hydrograph and during baseflow periods is less than optimal, but the average very is good. Given that many metrics of change in this study relate to flow volumes, this match was important to obtain. Visual assessments of the match between modeled and measured hydrographs shows that the model’s peak flows occur at the right time and are generally of the same magnitude as those in the observed data but the days of the highest peak usually differ between modeled and measured streamflow ( Figure 12 to Figure 16). This indicates that the model performs well for capturing the general timing of peak flow, but that the exact days of the highest peak flow should not be expected to be accurate. On Greata Creek, Nash-Sutcliffe Efficiency indices were higher for the calibration than they were in the Trepanier calibration, likely due to the lower amount of “flashiness” in the flow on Greata Creek at the time of calibration. The NSEln reveals a good correspondence of base flows. Average flow volume was higher during two of the test years and lower during the third, suggesting no systematic over- or under-prediction. Baseflows, however, were always slightly over-predicted. The visual assessment reveals that the peak flows occur at the right time and base flows stay generally consistent throughout the calibration period ( Figure 15 and Figure 16). Underestimation of peak flow during one of the validation years ( Figure 16) likely contributed to the lower validation NSE.  60  A test of the influence of time-step on resulting NSE values was performed by aggregating the daily Trepanier validation (1984-1993) data on a weekly time step. This resulted in an improved NSE (from 0.42 on a daily time step to 0.57 on a weekly time step). This suggests that calibration for a weekly or monthly time step would likely have improved performance in this study, as well. Further, attention to time steps is an essential factor for comparing model performance for disparate studies. Table 12 gives performance results from some other studies that used WEAP compared to the present study. The table reveals that measures of model performance in this study fall within the range of that in other studies. Important to note is that the other studies used monthly or weekly timesteps which inherently have less variability than the daily time step used here, and therefore should be expected to result in better correlation. Table 12: Measures of model performance in other studies using WEAP Values compared to Trepanier and Greata Creek calibration and validation period, with the exception 2 of the R , which compares snow accumulation values Metric Young et al. Young et al. Ingol-Blanco & This study (2009) (2009) McKinney (daily time step) (monthly (weekly time (2009) (monthly time step) step) time step) Total Volume Error (%) - 33 to 2 - 32 to + 37 NSE 0.65 to 0.84 0.42 to 0.61 NSEln - 0.12 to 0.87 0.66 to 0.76 RMSE (%) 38 to 65 70 to 203 86 to 136 BIAS (%) - 6 to 0 - 22 to 31 -4 2 Correlation coefficient (R ) 0.71 to 0.94 0.78 to 0.95 0.88 to 0.95 Overall, since the calibration resulted in adequate match between all aspects of the hydrograph (peak flows, base flows, and flow volume); it was determined to be adequate for use in scenario analysis.  61  4.2) Scenario Results A total of seven water supply scenarios were implemented in WEAP (Table 13). All scenarios were paired with water use estimates (activity-based and extraction-based) with projected increases based on population and climate change up to the appropriate future scenario. Two scenarios for reservoir operations were examined with each supply change scenario (Table 13).  Scenario Name 2020s CGCM2  2020s HadCM3  2050s CGCM2  2050s HadCM3  Drought + 2020s  MPB Only  MPB + 2020s  High Reservoir Release Low Reservoir Release  Table 13: Scenario names and descriptions Description Climate input data provided for the 2020s by the CGCM2 GCM, A2 emission scenario downscaled to the study area. Demand data increased based on population growth projections for the 2020s and expected increases in demand for outdoor watering due to climate change. Climate input data provided for the 2020s by the HadCM3 GCM, A2 emission scenario downscaled to the study area. Demand data increased based on population growth projections for the 2020s and expected increases in demand for outdoor watering due to climate change. Climate input data provided for the 2050s by the CGCM2 GCM, A2 emission scenario downscaled to the study area. Demand data increased based on population growth projections for the 2050s and expected increases in demand for outdoor watering due to climate change. Climate input data provided for the 2050s by the HadCM3 GCM, A2 emission scenario downscaled to the study area. Demand data increased based on population growth projections for the 2050s and expected increases in demand for outdoor watering due to climate change. Climate input data for the first and last three years of the scenario and temperature data for the middle three years (drought) provided for the 2020s by the CGCM2 GCM, A2 emission scenario downscaled to the study area. Three year drought sequence precipitation, humidity and wind speed from recorded 2003 data. Demand data increased based on population growth projections for the 2020s and expected increases in demand for outdoor watering due to climate change. Climate input data from the 1984-1993 baseline period. Alterations to precipitation, energy and evapotranspiration parameters to simulate a time period 15 years after a Mountain Pine Beetle attack affects all mature lodgepole pine in the watersheds. Demand data based on increased population projections but not climate-related increases. Climate input data provided for the 2020s by the CGCM2 GCM, A2 emission scenario downscaled to the study area. Alterations to precipitation, energy and evapotranspiration parameters to simulate at time period 15 years after a Mountain Pine Beetle attack affects all mature lodgepole pine in the watersheds. Demand data increased based on population growth projections for the 2020s and expected increases in demand for outdoor watering due to climate change. Low initial reservoir storage volume and current release rates used in combination with all of the above water supply change scenarios. High initial reservoir storage volume and lower release rates used in combination with all of the above water supply change scenarios.  The following chapter first presents changes to snowpack and streamflow, followed by impact on downstream water users. The latter section is divided between supply and demand effects on the unregulated stream (in order to show the direct interactions between hydrologic change and demand  62  increase), and effects on the regulated stream (in order to show conditions under several reservoir management scenarios).  4.2.1) Hydrologic Changes 4.2.1.1) Climate Change Scenarios: 2020s Both CGCM2 and HadCM3 climate models used in this study resulted in less snow accumulation and earlier snowmelt for the 2020s time slice compared to baseline; but the similarity ends there. The CGCM2 model produced higher annual streamflow volumes relative to baseline with cumulative monthly flows also mostly higher than baseline. Relative to baseline, the HadCM3 model resulted in less annual flow, increases in winter flows, but decreases in summer flows. Peak flow occurred earlier with generally higher but variable peak flow volumes under the CGCM2 scenario relative to baseline, and differed little from baseline in terms of timing. Lower median flows relative to baseline were found under the HadCM3 scenario. The details on the hydrologic variation resulting from these two scenarios will be described here.  Snowpack The average maximum SWE across all HRUs was lower than the baseline (p=0.03) in the 2020s CGCM2 scenario (Table 14). A lower SWE was present all years in HRUs below 1100 m, and most years at higher elevations, leading to this watershed-wide effect of a lower SWE. The CGCM2 A2 2020s climate model predicted increased monthly precipitation, which led to a net annual increase in precipitation when downscaled to this basin (average annual increase 8%; range from +6% to +12%). Thus, the decrease in maximum SWE under the future climate scenario was not a result of the change in precipitation so was likely due to increased temperatures reducing snow accumulation early in the season. The average date of depletion of snow pack in the 2020s CGCM2 scenario was 9 days earlier than baseline (Table 14). When comparing date of depletion for each HRU between baseline and the scenario, melt occurred earlier in 315 of 324 instances (39 HRUs over 9 scenario years) showing that the general effect of the 2020s CGCM2 scenario was earlier snow melt. Table 14: Snowpack summary statistics for the 2020s climate change scenarios, compared to baseline 2020s 2020s Baseline CGCM2 HadCM3 245 238 205 Mean annual maximum SWE (mm) (n=9) 60 65 64 Standard deviation (S.D.) 20 22 21 Standard error (S.E.) Difference in date of depletion (n=351) -8.9 -4.4 Mean (days) 7.7 5.3 S.D. 2.6 1.8 S.E.  63  Average maximum snow accumulation was also lower under the HadCM3 scenario than baseline (p< 0.001). Mean SWE was lower than determined using the CGCM2 scenario, likely due to the fact that there was a reduction in monthly precipitation predicted by the HadCM3 model, leading to lower flow volumes when downscaled to this basin (average annual change -2%; range -6% to 0%). Thus, lower precipitation, in addition to warmer temperatures, led to a lower maximum snow accumulation relative to the CGCM2 scenario and baseline. Snow melted an average of 4.4 days earlier than baseline in the 2020s HadCM3 scenario, with snow in individual HRUs melting from 26 days earlier to 8 days later. Overall, melt occurred earlier than baseline in 295 instances out of 324 (39 HRUs examined over 9 years).  Streamflow Annual streamflow, measured above the water withdrawals on Trepanier Creek, was significantly higher (p< 0.001) under the 2020s CGCM2 scenario than baseline during all years (Table 15). However, the lower snowpack under the 2020s HadCM3 scenario led to less annual streamflow (p=0.03). Table 15: Annual streamflow summary statistics: Trepanier Creek, measured above the water withdrawals, for the 2020s climate change scenarios compared to baseline 2020s 2020s Baseline CGCM2 HadCM3 Annual Flow Volume 3 (million m /year) 28.4 31.1 26.0 S.D. 10.1 10.5 9.21 S.E. 3.39 3.50 3.07 Mean % volume change 11% -9% There were large differences between outputs from using these different GCMs in terms of monthly distribution of flow. On a monthly basis, there was significantly higher discharge in March to April (Table 16). Median discharge did not change during June, July and August, although cumulative flow volume was 23% lower in June (p=0.03). Though the CGCM2 climate model predicts more precipitation November to June in the 2020s, only the early spring months show significantly higher discharge (Table 16). The results suggest that the greatest change in streamflow under the 2020s CGCM2 scenario was during the spring freshet, with any increase in precipitation in June offset by an earlier recession of peak snowmelt. In the HadCM3 scenario there was less cumulative monthly flow June (p=0.01) to September (p<0.001), and significantly higher discharge December to April (Table 16). This pattern partially parallels the change in precipitation, which the HadCM3 A2 scenario predicts to be lower for June, July and September, and higher for December, March, and April. The consistently lower mean cumulative flows June to September, despite the increase in precipitation in May and August, likely result from the earlier timing of snowmelt and increased evaporative demand in the summer due to warmer temperatures. The consistently higher winter flows likely result from increased baseflows due to greater overall precipitation during the winter months.  64  Table 16: Monthly flow statistics: Trepanier Creek, measured above the water withdrawals, for the 2020s climate scenarios compared to baseline (blue = increase and red = decrease relative to baseline. Significance counts in the lowest quartile are in bold, indicating that these results as generated in IHA are highly significant) Baseline 2020s CGCM2 2020s HadCM3 Month October November December January February March April May June July August September  Median cms 0.231 0.292 0.279 0.264 0.255 0.254 0.606 5.144 1.102 0.449 0.321 0.275  Median cms 0.240 0.341 0.326 0.301 0.303 0.337 1.487 5.658 0.744 0.417 0.310 0.300  Sig. count 0.73 0.46 0.37 0.36 0.27 0.09 0.00 0.82 0.56 0.59 0.76 0.73  % Volume 3 (m ) diff. 7 13 13 14 18 35 145 10 -23 -15 0 6  Median cms 0.242 0.321 0.327 0.307 0.308 0.323 0.695 3.937 0.672 0.373 0.267 0.225  Sig. count 0.59 0.47 0.07 0.06 0.04 0.06 0.22 0.52 0.49 0.40 0.37 0.17  % Volume 3 (m ) diff. -9 12 20 20 24 25 25 -4 -27 -25 -15 -15  The metrics of change for the spring freshet also show this change in timing and volume. In Trepanier Creek, the week of maximum flow occurred an average of 8.75 days earlier in the 2020s CGCM2 scenario than baseline though with a large amount of inter-annual variability: occurring from 30 to 0 days earlier. The onset of the freshet was an average of 9.7 days earlier and ended an average of 5.3 days earlier, demonstrating that the shift in 7-day peak is representative of the freshet shift as a whole. The mean difference in annual discharge was an increase in peak flow 0.731 million 3  m . There was great inter-annual variability in difference between baseline and the 2020s climate 3  scenario, which resulted in as low as 0.099 million m less flow during a dry year. Peak discharge was higher (significance count in lowest quartile) under the 2020s CGCM2 climate scenario relative to baseline for 7-day maximum flow as calculated in IHA but not over the 30-day maximum discharge. The timing and volume of maximum flow occurred both earlier and later than from baseline conditions in the 2020s HadCM3 scenario. The 7-day peak flow occurred an average of 1.6 days earlier and a maximum of 4 days earlier under this scenario. When the whole freshet was examined, the onset occurred as early as 11 days early in one year, but more often did not occur earlier. Total volume of the peak either increased or decreased, depending on the year, with an overall average decrease in peak flow. IHA results confirm the result that peak flows differ little from baseline, showing no significant difference in 3-, 7- and 30-day maximums.  Inflow to Reservoir The catchments contributing to the Peachland Reservoir generally followed the same streamflow patterns as those found near the Trepanier Creek outlet and reflect precipitation patterns generated by each climate model. Total annual inflow to the reservoir was 8% greater under the 2020s CGCM2 scenario conditions, and 11% lower under the 2020s HadCM3 scenario.  65  The timing of the greatest contribution of water to the reservoir was slightly earlier in the catchments contributing to the reservoir under the CGCM2 scenario than for Trepanier Creek as a whole, with the 7-day peak flow occurring an average of 12 days earlier than baseline and ranging from 33 days earlier to no change. Though earlier, this peak in the freshet brought more flow to the reservoir, paralleling a significant increase in median flows in March and April. Under the 2020s HadCM3 scenario, the 7-day peak flow occurred an average of 7 days earlier than baseline. The timing of the freshet similarly showed little change from baseline, with onset occurring 4 to 0 days earlier, and recession 14 to 1 day earlier. There was an average decrease in flow volume during the 7-day peak flow period, though median peak flow discharges, measured as 3-, 7- and 30-day maximum flows, were not significantly different from baseline.  4.2.1.2) Climate Change: 2050s Hydrologic changes in the 2050s time slice were more extreme than those observed in the 2020s. Application of both CGCM2 and HadCM3 models resulted in lower maximum snowpack, earlier snowmelt, and reduced annual streamflow. Results for whether streamflow increased or decreased during a specific month varied depending on which climate model was used. Though the timing of the spring freshet shifted more in this time slice than in the 2020s, on average, the freshet occurred less than a month earlier. Details on the hydrologic changes for this time slice are provided here.  Snowpack The average annual maximum SWE was lower than the baseline in all years in the 2050s CGCM2 scenario (Table 17) (p<0.001). Though the 2050s CGCM2 climate model downscaled to the study area resulted in a higher net annual precipitation (average annual change +4%; range -1% to +9%), the warmer temperatures explain the diminished accumulation of the snowpack, especially at the lower elevations. Whereas there were both increases and decreases in SWE in the 2020s CGCM2 scenario, in this time slice SWE decreased in all but a few higher elevation HRUs during the th  th  6 and 8 scenario year. These lower snow packs contributed to the earlier snowmelt: melt occurred earlier in all HRUs in all years under this scenario, with an average melt date over 2 weeks earlier. Application of the HadCM3 2050s model resulted in even lower average annual maximum SWE relative to CGCM2 (p<0.001) though a similar difference in date of depletion. Similar to the CGCM2 2050s scenario, maximum accumulation values decreased in all but a few HRUs in a few years under the 2050s HadCM3 climate change scenario. The lower snowpack compared to the 2050s CGCM2 scenario is explained by the lower monthly precipitation predicted by this climate model, which led to lower total annual precipitation when downscaled to the study area (average annual change -11%; range: -16% to -7%). Lower precipitation, combined with the effect of the higher temperatures converting less of the precipitation to snow led to this lower maximum SWE in the future climate scenario.  66  Table 17: Snowpack summary statistics for the 2050s climate change scenarios compared to baseline 2050s 2050s CGCM2 HadCM3 Baseline 245 209 200 Mean maximum SWE (mm) 60 60 56 S.D. 20 20 19 S.E. Difference in date of depletion (n=351) -18 -17 Mean (number of days) 9.6 8.3 S.D. 3.2 2.8 S.E.  Streamflow In Trepanier Creek measured above the water withdrawals, total annual stream discharge was lower in the 2050s CGCM2 scenario than baseline by an average of 6% (p=0.09) (Table 18). These results contrast the net increase in annual precipitation over each year of the scenario run. The timing of the increases in precipitation, in combination with higher potential evapotranspiration, explains these results. Increases in precipitation generally occur in the fall months, with decreases in the early summer precipitation accompanied by higher evaporation further reducing soil moisture at that time. The fact that net annual streamflow increases in the first three years of the scenario run, but decreases thereafter, support the conclusion that any initial effects of increased precipitation are later cancelled out by the effects of increased evapotranspiration. Table 18: Annual streamflow summary statistics: Trepanier Creek, measured above the water withdrawals, for the 2050s climate change scenarios compared to baseline 2050s 2050s Baseline CGCM2 HadCM3 Annual Flow Volume (million 3 m /year) 28.4 26.7 20.7 S.D. 10.2 9.35 6.57 S.E. 3.39 3.12 2.19 Mean % volume change -6% -27% In terms of monthly flow, in the 2050s CGCM2 scenario there was less cumulative flow volume at p < 0.05 May to September, and higher discharge and cumulative flow (p < 0.05) from December to April (Table 19). Changes in flow volume show a definite shift whereby the bulk of the 3  flow that previously occurred in May (average reduction = 2.31 million m ) occurs in April in the future climate (percent change in Table 19). The CGCM2 A2 2050s climate model predicted less precipitation in May, June, and July, and more August and September, so the lower flows in August and September are likely the result of an earlier peak in spring runoff and subsequently lower baseflow due to the net decrease in snowpack and increase in potential evapotranspiration, with precipitation increases during those months not enough to out weigh those factors.  67  Table 19: Monthly flow statistics: Trepanier Creek, measured above the water withdrawals, for the 2050s climate change scenarios compared to baseline (blue = increase and red = decrease relative to baseline. Significance counts in the lowest quartile are in bold, indicating that these results as generated in IHA are considered highly significant) Baseline 2050s CGCM2 2050s HadCM3 Month October November December January February March April May June July August September  Median cms 0.231 0.292 0.279 0.264 0.255 0.254 0.606 5.144 1.102 0.449 0.321 0.275  Median cms 0.242 0.354 0.343 0.326 0.354 0.417 1.69 2.99 0.504 0.331 0.267 0.257  Sig. count 0.84 0.28 0.15 0.09 0.02 0 0 0.52 0.14 0.29 0.43 0.68  %Volume 3 (m ) diff. 0 17 22 29 40 68 230 -22 -57 -32 -18 -8  Median cms 0.242 0.355 0.347 0.335 0.334 0.361 0.695 3.653 0.475 0.379 0.271 0.248  Sig. count 0.69 0.31 0.06 0.04 0.03 0.01 0.24 0.46 0.41 0.35 0.38 0.48  %Volume 3 (m ) diff. -34 -18 -6 -5 -2 9 124 -28 -59 -48 -45 -44  Metrics of peak flow for the 2050s CGCM2 scenario show similar results to the monthly change data. The week of maximum streamflow in Trepanier Creek occurred an average of 16 days earlier, with an inter-annual variability of between 32 and 2 days earlier. Flow volumes changed 3  substantially during that peak week, with an average increase of 1.66 million m but with a range of 3  3  differences between a decrease of 3.57 million m and increase of 2.07 million m . The entire freshet occurred earlier than baseline under this scenario, rising from 8 to 28 days earlier (mean = 15) and receding 0 to 32 days earlier (mean = 15). IHA comparison of median flows showed a significantly earlier day of annual maximum flow under the scenario conditions (Julian day 123 instead of 150). Under the 2050s HadCM3 scenario, total annual flows in Trepanier Creek above the withdrawals were significantly lower than baseline in all scenario years (p=0.001). With an average reduction by 27%, these were even lower than under the 2050s CGCM2 scenario (Table 18). The two types of monthly flow metrics show two different phenomena occurring under the 2050s HadCM3 scenario. Cumulative monthly flow volumes were lower than baseline May through November (p<0.05), with increases only occurring during the onset of the spring freshet in April (p<0.01). Median flows, however, were significantly higher than baseline December through April, with no significant change May through September. These differences reflect the flashiness of the stream hydrograph, whereby there were more frequent high flow events (reflected in the change in daily medians), though less flow volume overall (reflected in the cumulative volume). The timing of the freshet shifts more in the 2050s HadCM3 time slice than it did during the 2020s time slice. About two-thirds of the decrease in May flows in Trepanier Creek 0.103 million m  3  3  can be accounted for in the increase in April flows (71,000 m ). The maximum 7-day flow was an average of 14 days earlier under this scenario, ranging from 33 to 1 day earlier. Peak flow volumes 3  3  were also consistently lower, ranging from 2.57 m to 0.293 million m lower.  68  Inflow to Reservoir Hydrologic changes in the catchments contributing to the Peachland Reservoir were similar to those observed near the mouth of Trepanier Creek under both 2050s climate change scenarios. In both scenarios, annual inflow into the reservoir decreased, though more so under the HadCM3 scenario (range: - 44% to - 6% compared to - 21% to + 4% for the 2050s CGCM2 scenario). Under the 2050s CGCM2 climate scenario, the total volume of water flowing into the reservoir was lower in all but the first two years of the scenario run. This resulted in a reduction by 8% across 3  the simulation. Flow volumes were reduced by an average of 18,000 and 12,000 m in May and 3  June, and increased by 18,000 m in April, showing that a substantial amount of the freshet was shifted to the earlier month, with an impact on the relationship between inflow and reservoir release timing. When examined in terms of median flow, the median flow in May shifted from 0.760 cms (baseline) to 0.520 cms (scenario). When the operating rules were set to release 0.570 cms during May, more water was released from the reservoir at that time than was coming in. The date of the 7day peak flow was found to occur between 0 and 20 days earlier than baseline over the scenario run. Under the 2050s HadCM3 scenario, there was less annual inflow to the reservoir relative to baseline during all scenario years. Similar to the results for Trepanier Creek, patterns differed between cumulative monthly flow and median daily flow metrics: cumulative inflow into the reservoir was lower than under baseline conditions during all months (p<0.05) but July, though median flows were only significantly lower in September and November, indicating flashy flows during the other months. The 7-day peak flow occurred an average of 8 days and a maximum of 2 weeks earlier under the HadCM3 scenario than under baseline conditions. Although the onset in peak flow occurred only about a week earlier than baseline, the peak receded more than a week earlier in most years. The volume of peak flow was lower than baseline in seven of the nine scenario years, reflecting both lower peaks and the earlier recession.  4.2.1.3) Drought During Climate Change Using the CGCM2 2020s Model The three-year drought scenario cannot be compared to the baseline in the same way as the climate change scenarios. In the drought scenario, baseline data were not perturbed by a uniform set of conditions, but rather baseline data was replaced with three years of drought data in the middle of a climate change sequence. Whereas under the climate scenarios this study uses multiple years to show the effects of inter-annual variability, in this drought scenario it is the sequence of events that is of interest.  69  Snowpack  Average maximum SWE during the drought years is 142 mm, compared to the 245 mm in the  baseline. This value reflects the results of both the precipitation data from the drought year (222 mm at Peachland in the 2003 water year) and the temperature data from the 2020s CGCM2 scenario. The average date of snowpack depletion is 23 days earlier than the baseline average in the scenario.  Streamflow The total annual streamflow during the drought years, as measured in Trepanier Creek above the withdrawal, was between 70% and 74% lower than the mean volume under baseline conditions. However, streamflow volumes decreased over the course of the drought (Table 20). Since the input data was the same for all three of the drought years (precipitation, temperature, wind speed and humidity), the reduction in streamflow reflects likely the result of depletion of soil moisture and ground water levels in the watershed over the course of the drought. Table 20: Total annual streamflow over the course of the drought Drought Year st  1 nd 2 rd 3  Total annual streamflow, Trepanier 3 Creek (million m ) 8.55 7.81 7.42  Total annual inflow to Peachland Reservoir, 3 (million m ) 1.48 1.38 1.33  Inflow to Reservoir The total volume of water flowing into the reservoir was an average of 72% lower during the drought years compared to the average from the baseline years (annual totals in Table 20). This reduction is greater than the 65% reduction projected to be the 1:100 year drought conditions of inflow to Peachland Lake as calculated in a study from 1977 (“Annual Runoff Estimates for West Side Okanagan Valley”, reported in Dobson, 2006).  4.2.1.4) Land Cover Change: Mountain Pine Beetle Attack Maximum snow accumulation and annual streamflows were higher than baseline under the Mountain Pine Beetle attack scenarios. The snow melted earlier, resulting in an earlier and higher spring freshet in both scenarios. The addition of the climate change variables resulted in significant hydrologic changes relative to the Mountain Pine Beetle Only scenario. The results of each scenario, and differences between them, will be presented here.  70  MPB Under Present Climate Conditions Snowpack The mean maximum SWE was higher on average over the simulation (Table 21) and in each year of the MPB scenario (p<0.001), which is to be expected given the increase in precipitation representing the decrease in canopy interception. When examined as a percent increase averaged across all HRUs, regardless of percent canopy loss, maximum SWE was between 5% and 10% higher. The mean date of snowpack depletion was an average of 4 days earlier in the MPB scenario than baseline. Since loss of forest canopy is expected to lead to a faster rate of snowmelt (Boon, 2009; Winkler, 2009), this result matches expectations. Given that the maximum SWE was higher to begin with under MPB conditions compared to baseline, the faster snowmelt rate does not significantly impact the date of depletion metric. Among HRUs, snowpack generally depleted between 0 and 10 days earlier than baseline, 2  although this sometimes occurred later. There is no significant linear correlation (R = 0.003, p=0.75) between difference in melt date and elevation, showing that the MPB Only scenario did not exacerbate elevational trends. Table 21: Snowpack summary statistics for the Mountain Pine Beetle scenarios compared to baseline MPB + Baseline MPB Only 2020s 245 263 255 Mean annual maximum SWE (mm) (n=9) 60 62 68 S.D. 20 21 23 S.E. Difference in date of depletion (n=351) -4 -11 Mean (days) 6.2 8.6 S.D. 2.1 2.9 S.E.  Streamflow There was higher annual streamflow in the MPB scenario than baseline (p<0.001)(Table 22). Annual percent volume increase ranged from 43% to 84% higher in the scenario than under baseline conditions, with an average of a 57% increase (Table 22). Table 22: Annual streamflow summary statistics: Trepanier Creek, measured above the water withdrawals, for MPB scenarios compared to baseline MPB + Baseline MPB Only 2020s Annual Flow Volume (million 3 m /year) 28.6 44.9 48.5 S.D. 10.7 15.5 15.9 S.E. 3.55 5.16 5.31 Mean % volume change 57% 73%  71  Increase in median streamflow was significant in all months but June (Table 23), with the largest monthly volume increase occurring in April (p<0.01). This indicates that snow was melting earlier so some of the freshet was transferred to earlier months, resulting in flows that were comparable to baseline in June. Monthly volume increases were the lowest in December through February. Streamflow may not have increased as much in the winter relative to the other months because the increased precipitation was forming as snow. Low flows, calculated as the lowest 30% of flows in the scenario, increased from 18% to 31% October through January, and 26% in August, though did not increase significantly in June or July (Table 23). Table 23: Monthly flow statistics: MPB Only and MPB + 2020s scenarios (blue = increase and red = decrease relative to baseline. Significance counts in the lowest quartile are in bold, indicating that these results as generated in IHA are considered highly significant) Baseline MPB Only MPB + 2020s Month October November December January February March April May June  Baseline Median 0.231 0.292 0.279 0.264 0.255 0.254 0.606 5.144 1.102  Median cms 0.389 0.452 0.417 0.377 0.359 0.395 1.75 7.57 1.34  Sig. count 0.01 0.02 0.03 0.03 0.04 0.04 0 0.11 0.47  Volume 3 (m ) Diff % 55 45 37 35 34 44 215 63 24  Median cms 0.386 0.466 0.440 0.398 0.412 0.562 3.58 6.44 0.902  Sig. count 0.01 0.01 0.02 0.03 0.02 0 0 0.24 0.59  Volume 3 (m ) Diff % 65 66 53 53 61 112 451 53 17  July August September  0.449 0.321 0.275  0.618 0.483 0.410  0.06 0.07 0.01  70 63 56  0.620 0.476 0.438  0.03 0.06 0.02  48 68 66  In Trepanier Creek measured above the withdrawal, the 7-day maximum flow was an average of one week earlier under this scenario (range: 20 days to 0 days earlier). When examining the entire freshet, the onset was earlier than shown in the 7-day maximum data (from 10 to 20 days earlier than baseline). However, the earlier onset of the peak was not followed by an earlier recession of the peak to the same degree, beginning 0 to 9 days earlier. 7-day peak flow volumes 3  were from 1.06 to 4.87 million m higher, representing a volume increase between 22% and 146% (average = 59%).  Inflow to Reservoir Inflow to Peachland Reservoir was significantly higher in the MPB Only scenario: by volume, from 52% to 91% higher. The timing of the highest 7-days of flow into the reservoir was an average of only 1 day earlier under this scenario, and ranged from 5 to 0 days earlier among the scenario years. When the entire freshet was examined, the onset was from 15 to 7 days earlier, and recession was  72  from 6 days earlier to 2 days later. Thus, although the freshet began earlier, it lasted longer, contributing a greater amount of water to the reservoir overall.  MPB Under Climate Change, 2020s Snowpack The effect of the Mountain Pine Beetle attack, in combination with climate change simulated based on data from the 2020s CGCM2 climate model, was to create the greatest increase in annual effective precipitation arriving at the soil surface of any of the scenarios examined here. This results from the combination of increased precipitation from the GCM, as well an increase in effective precipitation arriving at the soil surface as simulated for the MPB scenario. The precipitation under the MPB + 2020s scenario is the arithmetic combination of the increase from the 2020s CGCM2 and MPB Only scenario. Thus, from precipitation alone, one would expect a maximum SWE greater than both of these scenarios. However, while greater than baseline (p=0.03) and greater than the 2020s CGCM2 scenario (p<0.001), maximum SWE were lower than the MPB Only scenario (p=0.02), suggesting that increased temperatures due to climate change offset the effect of increased precipitation from the MPB scenario in terms of maximum SWE. The average snowpack measured across all HRUs and all years (n=351) depleted 11 days earlier than baseline under the MPB + 2020s scenario (Table 21). Since the maximum snow accumulation was slightly higher under this scenario than under baseline conditions (and occurred on about the same day between baseline and scenario), this can be attributed to a higher melt rate under the scenario, likely a combination of increased temperatures and increased radiation. The snowpack tended to melt between 2 and 20 days earlier than baseline during all scenario 2  years. The most extreme differences in melt timing occurred at the lowest elevations. The R of the regression line each year ranged from 0.003 to 0.51 (p<0.001 to p=0.74), showing that in terms of variability from baseline, the MPB + 2020 scenario may have exacerbate elevational trends in some years but not others (i.e., melt low elevations earlier or high elevations later). The slight trend here in an earlier melt at the lowest elevations was not present in the MPB Only scenario, suggesting that this results from the climate influences.  Streamflow As in the MPB Only scenario, average streamflow increased on average over the simulation and during all years of the MPB + 2020s scenario (Table 22). On a monthly basis, the median flow values and cumulative monthly flow volume were all significantly higher than baseline except for June. These results align with a visual examination of the hydrograph, where earlier and higher peak flows are followed by lower flows in June of most years. The Trepanier Creek 7-day maximum flow was on average 10 days earlier in this scenario, with an inter-annual variability between 22 and 0 days earlier. The shift in peak flow timing was not  73  as extreme as the shift in timing of snowmelt, likely reflecting the greater influence of the higher elevations (which did not melt as early) on peak flow. The onset of the freshet was earlier (from 4 to 18 days) than baseline conditions during all scenario years and recession begins from 23 to 0 days 3  earlier. The volume of the 7-day peak was consistently higher than baseline (mean = 3.07 million m ; 3  S.D. = 1.74 million m ), confirmed by significantly higher 7-, 30- and 90-day maximum flow values.  Inflow to Reservoir Similar to Trepanier Creek as a whole, there was more inflow to the reservoir under the MPB + 2020s scenario during all months of all years except for June. The timing of the week of maximum volume flowing into the reservoir was generally earlier in the scenario (mean = 3 days) though there was a large amount of inter-annual variability with 7-day peak discharge from 20 days earlier to 4 days later. When examining the freshet as a whole, onset was 20 to 10 days earlier, and recession from 13 days earlier to 8 days later. In some cases the entire freshet shifted to earlier in the year, and in some cases it began earlier and lasted for longer, increasing the total volume of water entering the reservoir.  MPB Only and MPB + 2020s Compared Both mean annual maximum SWE (p=0.05) and mean annual streamflow (p<0.001) are significantly different between the MPB Only and MPB + 2020s scenarios (Table 24). Table 24: Comparison between maximum snowpack and streamflow data, MPB scenarios, climate scenario, and baseline Mean Annual Mean Annual Scenario 3 Max SWE (mm) Streamflow (million m ) Baseline 245 28.4 2020s Only 238 31.1 MPB Only  263  44.9  MPB + 2020  255  48.5  The fact that the maximum SWE was lower under the MPB + 2020s scenario than in the MPB Only scenario (despite the increase in precipitation) suggests that the decrease in SWE from the temperature increase partially offsets the increase from the canopy loss. Since the MPB Only scenario resulted in a greater increase in maximum SWE than climate change caused a decrease, the effects of pine beetle damage were slightly stronger. The fact that the MPB + 2020s scenario resulted in lower maximum SWE but higher streamflow suggests that more precipitation falls as rain than snow with the addition of the climate change variables. Comparing the relative magnitude of impacts on hydrology, temperature change had a stronger effect on ablation, and therefore dates of depletion and peak streamflow than the radiation and advective energy changes due to canopy loss alone. The average date of depletion was 11 days earlier under the MPB + 2020s scenario than baseline. Considering that melt occurred an average of  74  9 days earlier under the scenario of climate change alone, these results might indicate that the addition of the conditions present during the pine beetle scenario advanced melt by an average of only 2 days. Without the effects present in the climate change scenario, the changes in net radiation in the MPB scenario advanced melt by 5 days. Following the fact that the snow melted earlier in the MPB + 2020 than the MPB Only scenario is the fact that, compared to the MPB only scenario, the MPB + 2020 resulted in a greater overall shift in the entire peak flow period (more of the peak occurred earlier and less occurred later). Evapotranspiration reductions from the MPB attack offset increases in potential evapotranspiration from climate change only, leading to an overall decrease in actual evapotranspiration in the MPB + 2020s scenario. The increased temperatures under the MPB + 2020 scenario had the effect of increasing the potential evapotranspiration (PET) relative to the MPB Only scenario, and the decreased crop coefficient (Kc) present in both MPB scenarios had the effect of decreasing the actual evapotranspiration (AET) relative to baseline. The resulting changes in AET and PET for sample HRUs are shown in Table 25. Table 25: PET and AET for a low elevation, mid elevation and high elevation HRU affected by pine beetle damage compared between baseline, MPB only and MPB + 2020s scenarios Low Elevation HRU Mid Elevation HRU High Elevation HRU (P 3LN; ECA = 30%) (P 2M; ECA = 25%) (P 4H2; ECA = 30%)  Average Annual PET Average Annual AET  Baseline  MPB Only  MPB 2020  Baseline  MPB Only  MPB + 2020  Baseline  MPB Only  MPB + 2020  653  475  506  727  567  598  571  417  439  426  378  405  498  452  481  428  353  371  Table 25 shows that the MPB Only scenario had the effect of decreasing the average annual PET (defined as reference crop potential evapotranspiration * Kc in WEAP) and AET relative to baseline. The potential evapotranspiration was reduced in the MPB scenario due to modification of the potential evapotranspiration with the Kc reduction, with a resulting impact on the amount of water actually transpired (AET). Potential evapotranspiration was increased in the climate change scenario due to the warmer temperatures, verified by the fact that the MPB + 2020 scenario had a higher PET than the MPB Only scenario. However, since the MPB + 2020 scenario resulted in a lower AET than the baseline scenario, the effects of the pine beetle attack on evapotranspiration were stronger than those of climate change. Combined, changes in effective precipitation due to lack of interception and change in evapotranspiration brought about by the beetle attack had a greater impact on volume of streamflow. Overall, the effects of the MPB + 2020 conditions were to increase precipitation available for use by plants and streamflow relative to the climate change only scenario, and increase temperature for all months (based on 2020 CGCM2 data) relative to the MPB Only scenario.  75  4.3) Interactions between Supply and Demand 4.3.1) Trepanier Creek The interactions between the supply changes described above and the changes in water demand resulting from population increases and climate change can be seen most directly on the unregulated Trepanier Creek. The timing of the decreases in supply as they overlap with the increases in Peachland’s demands points out the critical periods during which to anticipate deficit on this system. Changes in the ability of the creek to meet instream flow needs were seen both from supply changes alone and in combination with the greater demands.  4.3.1.1) Climate Change: 2020s The increases in population and climate-related water demand were the same for all scenarios set to occur in the 2020s (Appendix C). While the evaporative demands of forest lands within the watersheds themselves varied depending on whether the CGCM2 or the HadCM3 model was used, increases in water use by agriculture and outdoor watering due to the warmer temperatures and longer growing seasons expected to accompany climate change were approximated using the activity-based and extraction-based water use estimates described in Section 3.7.4 and did not vary with climate model used. The results described here reflect the changes in streamflow as they interact with the demands of the water users.  Meeting Peachland’s Demands No water deficit was present for the users drawing water from Trepanier Creek under the 2020s CGCM2 climate scenario. Although annual demands were higher, and June and July flows were lower than baseline, these changes were not enough to cause a deficit for Peachland. Trepanier Creek streamflow resulting from the changes produced under the HadCM3 scenario, which projected lower summer precipitation, was low enough in the summer that there was a deficit on Trepanier Creek during the lowest water years under the extraction-based water use estimate (Table 26). That deficit was not considerable: 98 to 99% of Peachland’s demands could be met. Table 26: Unmet demand in Peachland System 1 under the 2020s HadCM3 scenario, extraction-based water use estimate Year Dates Unmet Amount Total % of Increase in Demand of Deficit required at Demand Demand Increase ÷ (thousand that time Met (thousand Deficit 3 3 m) (thousand m) 3 m) st 1 8/18 - 8/31 2.3 170 99% 34 >100% rd 3 8/28 - 8/31 1.2 48 98% 8 >100%  76  The volume of water required by Peachland from Trepanier Creek was higher under the activity-based water use estimate than the extraction-based water use estimate, though no deficit was present using the former. The difference can be explained by the role of the Trepanier Licenses, which are set as the total volume of the license under the extraction-based estimate, which is greater than demand generated by the activity-based estimate. The fact that the reduced flows resulting from the higher extraction-based estimate resulted in a deficit downstream in Peachland reveals that streamflow was at a critical threshold where the amount of water use upstream determined whether or not there would be a deficit downstream.  Meeting Instream Flow Targets Under the 2020s CGCM2 climate scenario, the Mean Annual Discharge (MAD) and B.C. Instream Flow Needs (BCIFN) instream flow targets were met for fewer days than baseline May through September of the “normal” and “very dry” years even though there was more total flow during August and September above the withdrawal, likely owing to the increased withdrawals. Under the 2020s HadCM3 scenario, there was virtually no difference in meeting targets between scenario and baseline in the “very wet” year. In the “normal” and “very dry” years there were consistent reductions May through September for meeting both targets based on MAD and BCIFN targets, showing that the system was more sensitive to the lower flows predicted by the HadCM3 scenario only in the drier years. The >400% MAD flushing target was met for more than 2 weeks in seven of the nine CGCM2 scenario years, exceeding the NHC (2001) recommendation. Though peak flows were lower under the HadCM3 scenario, the recommended 400% MAD flushing flow target was met for more than a day in seven of the nine scenario years, also surpassing NHC recommendations.  4.3.1.2) Climate Change: 2050s As with the 2020s scenarios, the increases in population and climate-related water demand were the same for all scenarios set to occur in the 2050s (Appendix C). Differences between results under the two climate models used here result from the timing of the change in flow and interactions with the increases in demand.  Meeting Peachland’s Demands Under the 2050s CGCM2 scenario, there was unmet demand in the area of Peachland receiving water from Trepanier Creek under both water use estimates during both “very dry” years. Though these deficits lasted up to a month and a half, between 90% and 99% of demand at that time was met. The activity-based water use estimate was much higher here, resulting in a greater volume deficit than the extraction-based water use estimate.  77  Table 27: Unmet demand in Peachland System 1 under the 2050s CGCM2 scenario Year Dates Unmet Amount of Total required % of Increase in Demand Deficit at that time Demand Demand Increase 3 (thousand (thousand m ) Met (thousand ÷ Deficit 3 3 m) m) Activity-based water use estimate st 1 7/16 - 8/31 81 992 92 372 >100% rd 3 7/13 - 8/31 15 1,132 99 433 >100% Extraction-based water use estimate st 1 7/28 - 8/31 47 477 90 74 >100% rd 3 8/11 - 8/31 5 287 98 44 >100% Lower flows May to September in Trepanier Creek under the HadCM3 scenario than the CGCM2 scenario contributed to more days where demands could not fully be met. There were deficits for 5 of the 9 scenario years (all but the “wet” and “very wet” baseline years) under the extraction-based water use estimate, and 4 years under the activity-based water use estimate (Table 28 and Figure 17). Despite the flow changes, Trepanier Creek could still meet over 75% of Peachland’s demand in the times of deficit. Table 28: Unmet demand in Peachland System 1 under the 2050s HadCM3 scenario Year Dates Unmet Amount Total required % of Increase in Demand of Deficit at that time Deman Demand Increase 3 3 (thousand (thousand m ) d Met (thousand m ) ÷ Deficit 3 m) Activity-based water use estimate st 1 7/13 – 8/31 184 931 80 372 >100% nd 2 7/28 – 8/31 34 617 95 244 >100% rd 3 7/1 – 8/31 248 1,182 79 483 >100% th 4 7/21 – 8/31 80 763 89 304 >100% Extraction-based water use estimate st 1 7/23 – 8/31 132 639 77 187 >100% nd 2 8/7 - 8/31 37 442 90 114 >100% rd 3 7/15 – 8/31 149 826 78 222 >100% th 4 7/25 – 8/31 60 540 89 187 >100% th 8 8/21 – 8/31 12 271 92 48 >100% Under both climate scenarios, in all scenario years, the increase in demand due to climate change and population growth was greater than the amount of unmet demand. This shows that deficit would not likely have occurred without the increase in demand.  78  m3/day  25,000  20,000  Trepanier Flow, above Withdrawal (activity-based)  15,000  Trepanier Flow, above Withdrawal (extraction-based) Demand, Activitybased estimate  10,000  Demand, Extractionbased Estimate  5,000  10/1  7/1  4/1  1/1  10/1  7/1  4/1  1/1  10/1  7/1  4/1  1/1  10/1  7/1  4/1  1/1  10/1  0  Figure 17: Comparison between Trepanier Creek flow and demand in the first through fourth year of the simulation for the scenario 2050s HadCM3 activity-based and extraction-based water use estimates This graph displays Trepanier’s streamflows at the scale of water demands to highlight deficits in the 3 low flow periods. Actual flows were up to 1.2 million m /day during spring freshets during the time period covered.  Meeting Instream Flow Targets Under the 2050s CGCM2 climate scenario, Trepanier Creek MAD and BCIFN targets were met for fewer days and to a lesser extent in “very wet,” “normal” and “very dry” years in May through July. Targets were also met for fewer days and to a lesser extent August to September of the “normal” and “very dry” year. These changes parallel the effect of the climate scenario above the withdrawal of less cumulative flow volume during May through September. Under the 2050s HadCM3 scenario, the MAD instream flow targets continued to be met during December to March of the “normal” and “very wet” year, likely owing to the higher streamflows during those months counteracting any demand increases. Targets were met for fewer days in the summer and fall however, and in even the “normal” year the target was met at less than 30% from June through September. The Trepanier BCIFN target was even less fully met each month. Although there were no days with no flow in Trepanier Creek under this scenario, flow was low enough so that an average of less than 0.1% of the August MAD and BCIFN targets were met in the “normal” and “very dry” year. Thus, despite the fact that most of Peachland’s demands were met with the lower flows of this scenario, instream flows were significantly compromised during the “normal” and “very dry” years. The 400% MAD flushing flow target was met for more than one week in 7 of the 9 scenario years on Trepanier Creek under the 2050s CGCM2 scenario, and was met for more than one week in 5 of the 9 scenario years under the 2050s HadCM3 scenario. Thus, even under this most extreme of the climate change scenarios, the NHC suggested target of reaching 400% MAD for 1 or 2 days once every two years can be met on Trepanier Creek.  79  4.3.1.3) Drought and Climate Change Meeting Peachland’s Demands There was a water deficit in Peachland during all three drought years, with less demand met each year of the drought (i.e., there was more deficit in the third year of the drought than the first year)(Table 29). Since demand was set to be the same each year of the drought, the difference was solely due to less streamflow, which likely resulted from sequentially less soil water in the watershed over the course of the three-year drought. Table 29: Unmet demand in Peachland System 1 under the Drought + 2020s scenario Year Dates Amount of Total % of Increase Demand Unmet Deficit required at Demand in Demand Increase (thousand that time met (thousand ÷ Deficit 3 3 m) (thousand m) 3 m) Activity-based water use estimate st 1 Drought year 7/26 – 7/31 4 91 96 16 >100% nd 2 Drought year 7/19 – 8/31 28 585 95 120 >100% rd 3 Drought year 7/15 – 8/31 59 646 91 132 >100% Extraction-based water use estimate st 1 Drought year 7/25 – 8/31 91 446 80 155 >100% nd 2 Drought year 7/18 – 8/31 138 528 74 184 >100% rd 3 Drought year 7/15 – 8/31 171 563 70 197 >100% Much less of Peachland’s total demand can be met under the extraction-based estimate, despite the lower total demand estimate under that scenario. This shows the effects of lower streamflow volumes resulting from the Trepanier Licenses withdrawals (Figure 18). Regardless of the cause, the difference in deficits between demand estimates shows that flows were at the threshold whereby the difference between demand estimates made a difference: in the third year of the drought, 91% of demand was met during the period of deficit under the activity-based water use estimate, but only 70% was met under the extraction-based water use estimate for that same time period.  80  20,000 18,000  Trepanier Flow, above withdrawal, activity-based  16,000  m3/day  14,000  Trepanier Flow, above withdrawal, extraction-based  12,000 10,000  Demand, Activitybased estimate  8,000 6,000  Demand, Extractionbased Estimate  4,000 2,000 6/1  2/1  10/1  6/1  2/1  10/1  6/1  2/1  10/1  6/1  2/1  10/1  6/1  2/1  10/1  0  Figure 18: Comparison between flow and demand during three drought years and two subsequent years under the drought under climate change (2020s CGCM2) scenario, Trepanier Creek: activity and extraction-based water use estimate. This graph displays Trepanier’s streamflows at the scale of water demands to highlight deficits in the 3 low flow periods. Actual flows were up to 1.2 million m /day during spring freshets during the time period covered.  Meeting Instream Flow Targets As can be expected for a drought, instream flow targets were less fully met during the drought years than under average baseline conditions. The highest flow targets in May were only met for a few days in each drought year. The Trepanier Creek MAD and BCIFN targets were only met for a few days in each of the drought years. The percent of the target that was met, however, decreased over the period of the drought, likely due to flow reductions. On Trepanier Creek, the recommended 400% MAD flushing target was only met one day during the drought. If there were normal water years on either end of the drought, then the NHC recommendation of reaching this target every other year may be able to be met, otherwise, it is unlikely that this target would be met.  4.3.1.4) Land Cover Change: Mountain Pine Beetle Attack Meeting Peachland’s Demands There was no unmet demand in Peachland under the MPB scenario set to current climate conditions. The shift in the timing of the Trepanier Creek freshet in the MPB + 2020s scenario resulted in flows that were comparable to baseline in June. However, given that flows generally recovered during summer and fall precipitation events, there was still and increase in flow during the months of increase in peak demand.  81  Meeting Instream Flows With only the influence of the Mountain Pine Beetle attack, the Trepanier Creek instream flow targets were more fully met during all months but June of the “normal” year, where the MAD target was met for fewer days than baseline, and the BCIFN target was met half as many days as baseline in the “normal” year, both likely owing to the shift in timing of the freshet. With the influence of climate change and MPB, both the Trepanier MAD and BCIFN, targets were met for fewer days and to a lesser extent in May and June of the “very wet” and “normal” year compared to baseline. Under the higher water use estimate in the “very dry” year, July MAD targets were also met to a lesser extent. Under the MPB + 2020s scenario, flows in Trepanier Creek below the withdrawal for Peachland were lower than baseline in June and no different from baseline July and August during many of the scenario years. These flows occurred despite the higher flows above the withdrawal over the course of the MPB + 2020s scenario. This indicates that the increase in flow from canopy loss was offset by the increase in water demand under the population growth and climate change projections in this scenario. As a result, conservation targets were not fully met.  82  4.3.2) Peachland Creek The ability of Peachland Creek to meet human and instream needs differed from Trepanier because of its high-elevation reservoir. Although this study developed supply perturbation scenarios over a range of baseline climate conditions with the intent of showing the range of variability that they could bring, the sequence in which these climate conditions occurred was found to be important on Peachland Creek. When the initial reservoir levels were set low and the extraction-rates higher, deficits in the District of Peachland were found to occur after the sequence of “very dry” years (#3 and #4). The change in volume water in the reservoir “carried over” from one water year to the next is shown in Table 30. These data show the impact of each water year independent of each other, demonstrating that scenario results could vary based on the initial conditions and order of wet and dry years. Table 30: Changes in carry-over storage each year for all scenarios, using the lower initial conditions and higher extraction rates A = activity-based water use estimate; E = extraction-based water use estimate. Decreases in red. 3 Values in million m . Year: 1 2 3 4 5 6 7 8 9 2020s – CGCM2 A 0.426 1.11 E 0.124 0.950 2020s – HadCM3 A -0.129 0.126 E -0.442 -0.070 2050s – CGCM2 -0.171 0.783 A -0.638 0.388 E 2050s – HadCM3 -0.795 -0.691 A -1.33 -1.24 E Drought + 2020s 0.426 1.11 A 0.124 0.950 E MPB + 2020s 3.07 0.841 A 2.96 0.949 E  -1.17 -1.49  -2.97 -3.23  1.39 1.48  1.28 1.23  1.75 1.67  -1.80 -1.93  2.80 2.80  -1.88 -2.22  -2.51 -1.66  1.68 1.62  0.049 -0.065  0.662 0.524  -2.36 -2.01  2.05 5.45  -1.62 -2.14  -3.31 -2.01  1.11 0.938  -0.169 -0.466  0.753 0.439  -1.75 -0.898  2.06 2.03  -2.83 -1.83  -0.069 0  0.299 0  -0.274 0  0.372 0  -0.379 0  0.299 0  0.067 -0.220  -2.25 -2.60  -2.40 -2.79  -2.41 -0.954  1.50 1.36  -1.50 -1.35  3.51 3.51  -0.708 -0.779  -0.732 -0.735  1.51 1.58  -0.019 -0.019  0.010 0.010  -0.025 -0.025  0.040 0.040  83  4.3.2.1) Climate Change Scenarios: 2020s Meeting Peachland’s Demands Under the 2020s CGCM2 scenario, even when the main freshet contribution to the reservoir occurred in the first weeks of May rather than the last weeks of May, the total volume contributed (and subsequently released) during that month was similar between baseline and scenario. The total volume of water entering the reservoir was greater than baseline in six of the scenario years, and therefore no downstream deficits were present in either reservoir release scenario. The reservoir-filling freshet under the 2020s HadCM3 scenario had less total volume than under both 2020s CGCM2 and baseline conditions. When initial reservoir volumes were set low and rd  extraction rates high, reservoir volume decreased after the 3 year (dry year), and did not increase in th  th  the 5 and 6 year to the same extent as under baseline conditions, such that it was drawn down th  completely by the 8 year (another dry year) (Figure 19). These conditions led to a short period of deficit in two of the scenario years under the extraction-based estimate (Table 31). Under the activity-based water use estimate, the District of Peachland only experienced deficit during the 4th “very dry” year.  Table 31: Unmet demand in Peachland System 3 under the 2020s HadCM3, higher reservoir extraction scenario Year Dates Unmet Amount Total required % of Increase Demand of Deficit at that time Demand in Demand Increase ÷ 3 (thousand (thousand m ) Met (thousand Deficit 3 3 m) m) Activity-based water use estimate th 4 7/1 to 8/31 128 700 82% 179 >100% Extraction-based water use estimate th 4 6/4 to 9/30 779 1,907 59% 425 55% th 8 7/8 to 9/30 439 1,388 68% 307 70% 2  6,000,000  1.8 5,000,000  Reservoir Volume (m3)  1.6 1.4  4,000,000  1  3,000,000  0.8 2,000,000  0.6  cms  m3  1.2  Inflow to Reservoir (cms) Reservoir Release (cms)  0.4  1,000,000  0.2 0 4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  0  Figure 19: Peachland Reservoir: Inflow, storage volume and pre-set reservoir releases under the 2020s HadCM3 higher reservoir extraction scenario, activity-based water use estimate  84  With the higher initial reservoir volume and lower rate of extraction for instream flows, reservoir levels were maintained consistently high and there was no unmet demand in Peachland.  Meeting Instream Flow Targets Increased withdrawals due to population growth and climate change led to an impact on aquatic life as measured by the instream flow targets. Under the higher reservoir release scenario, differences between baseline and the 2020s CGCM2 scenario can be seen even in the “wet” year. In that year, the conservation flow target was not met during October, whereas under baseline conditions it was met for the whole month. Under all precipitation regimes, targets continued to be fully met in the winter, supported by reservoir releases, but were met for fewer days in May through September in the “very wet” and “normal” years compared to baseline. The same situation was the case with the 2020s HadCM3 scenario: winter flows were mostly supported by the reservoir releases while the number of days the target was met and the percent of target met was lower June to September all years. There was no difference between the higher and lower water use estimates during the “very wet” year of the CGCM2 scenario, but during the normal year, the target was met for fewer days in October, May and June, and less fully met July though October. In the 2020s CGCM scenario, where there were low flows in July and August under the lower water use estimate, there were times of zero flow under the higher use estimate. In the HadCM3 scenario, flows were low enough that this situation was exacerbated even more: with the both demand estimates there was no flow during July and August of the “very dry” year, and in the higher extraction-based demand scenario, there was no flow July – September and half of June. In the 2020s CGCM2 scenario, compared to higher reservoir releases, the lower reservoir release scenario reduced ability to meet targets in November to January “very wet” year under high extraction estimates, and during October to June of the “very dry” year. Increases in percent of target met occurred in April and May of the “normal” year, and in June and July of the “very wet” year. When reservoir releases were lower under the 2020s HadCM3 scenario, winter targets were only fully met in the “normal” year under the low water use estimate. In the “very dry” year of the 2020s HadCM3 scenario, the instream flow targets were not met a single day under either water use estimate. The number of days that target was met only increased in February of the “very wet” year. In the 2020s CGCM2 scenario, the >400% MAD flushing flow targets were met for one or two days during at least six of the scenario years using both reservoir release estimates, meeting NHC recommendations. When the HadCM3 model generated flow projections, the >400% flushing target was met for only one day during two years under both the high and low demand estimate of the higher reservoir release and for one day in 6 of the scenario years under the lower reservoir release. Flows generated under the HadCM3 scenario were therefore insufficient to meet the NHC recommendations with the first set of operating rules, but not the second.  85  4.3.2.2) Climate Change: 2050s The initially low reservoir volumes and higher releases rates when combined with the inflows to the reservoir projected under the 2050s CGCM2 scenario resulted in two years where the reservoir was completely drawn down (Figure 20) and therefore in subsequent deficit for the District of Peachland. 7,000,000  2 1.8  6,000,000 1.6 5,000,000  Reservoir Volume (m3)  1.4  m3  1 3,000,000  cms  1.2  4,000,000  0.8  Reservoir Release (cms)  0.6  2,000,000  Inflow to Reservoir (cms)  0.4 1,000,000 0.2  4/1  4/1  10/1  10/1  4/1  10/1  4/1  4/1  10/1  10/1  4/1  4/1  10/1  10/1  4/1  10/1  4/1  0 10/1  0  Figure 20: Reservoir volume, inflow to reservoir and set reservoir release, 2050s CGCM2 lower initial conditions, higher reservoir release scenario, extraction-based water use estimate When the higher initial reservoir levels were used, accompanied by the lower release rates, the reservoir was consistently full enough to meet District of Peachland demands (Figure 21). 14,000,000  2 1.8  12,000,000 1.6 10,000,000  Reservoir Volume (m3)  1.4  m3  1 6,000,000  0.8  cms  1.2  8,000,000  Inflow to Reservoir (cms) Reservoir Release (cms)  0.6  4,000,000  0.4 2,000,000 0.2  4/1  4/1  10/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  0 10/1  0  Figure 21: Reservoir volume, inflow to reservoir and release for Peachland, 2050s CGCM2 higher initial conditions, lower release scenario, extraction-based demand estimate Deficits in Peachland vary based on management practices for the reservoir operations, as shown in Figure 20 and Figure 21 above. In the scenario where a deficit was present, there was a shorter and lower period of deficit under the activity-based demand estimate than the extraction-  86  based estimate, corresponding with higher demand under the extraction-based estimate on Peachland Creek (Appendix C). Table 32: Unmet demand in Peachland System 3 under the 2050s CGCM2, higher reservoir extraction scenario Year Dates Unmet Amount of Total required % of Increase Demand Deficit at that time Demand in Demand Increase ÷ 3 (thousand (thousand m ) Met (thousand Deficit 3 3 m) m) Activity-based water use estimate th 8 7/1 - 8/31 107 873 88 312 >100% Extraction-based water use estimate th 4 5/23 – 9/30 1,130 2,360 52 2,050 >100% th 8 5/24 - 9/30 780 2,350 67 2,060 >100% The increase in demand due to population and climate change-related increases in could entirely explain the deficit under both demand estimates (Table 32). The lower cumulative volume of flows under the 2050s HadCM3 scenario resulted in even greater deficits when the higher reservoir extraction rates were used. Similar to the 2050s CGCM2 scenario, under the high reservoir release scenario, the reservoir dropped down to the lowest level  2.0  4,500,000  1.8  4,000,000  1.6  3,500,000  1.4  3,000,000  1.2  2,500,000  1.0  2,000,000  0.8  1,500,000  0.6  1,000,000  0.4  500,000  0.2  0  0.0  Reservoir Volume (m3) Inflow to Reservoir (cms) Reservoir Release (cms)  4/1  10/1  4/1  10/1  4/1  4/1  10/1  10/1  4/1  10/1  4/1  10/1  4/1  4/1  10/1  10/1  4/1  cms  5,000,000  10/1  m3  after a sequence of “dry” and “very dry” years (Figure 22).  Figure 22: Reservoir volume, inflow to reservoir and release for Peachland, 2050s HadCM3 lower initial conditions, higher release scenario, extraction-based demand estimate Under the lower release scenario, reservoir volumes were maintained (Figure 23). Though, unlike the CGCM2 scenario, volume decreased over the scenario run, indicating that reservoir depletion may be seen over a longer simulation with similar conditions.  87  14,000,000  2.0 1.8  12,000,000  1.6 10,000,000  1.4  Reservoir Volume (m3)  1.2  m3  cms  8,000,000  1.0 6,000,000  0.8 0.6  4,000,000  Inflow to Reservoir (cms) Reservoir Release (cms)  0.4 2,000,000  0.2  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  0.0 10/1  0  Figure 23: Reservoir volume, inflow to reservoir and release for Peachland, 2050s HadCM3 higher initial conditions, lower release scenario, extraction-based demand estimate  In the reservoir release scenario that contributed to deficits, demand could not be fully met from Peachland Creek in six scenario years under the higher extraction-based estimate and only 2 years on the lower activity-based water use estimate (Table 33). When there was a deficit on this system, the proportion of demand met at that time was low, between 38 and 71% of total demand.  Table 33: Unmet demand in Peachland System 3 under the 2050s HadCM3 higher reservoir extraction scenario Year Dates Unmet Amount Total required % of Increase in Demand of Deficit at that time Demand Demand Increase 3 (thousand (thousand m ) Met (thousand ÷ Deficit 3 3 m) m) Activity-based water use estimate th 4 6/16 – 9/30 580 1,312 56 504 87 th 8 6/7 – 9/30 461 1,400 67 541 >100 Extraction-based water use estimate rd 3 6/5 – 10/31 1,540 2,490 38 882 57 th 4 5/15 – 10/31 1,490 2,810 47 1,000 67 th 5 9/1 – 10/23 170 592 71 181 >100 th 6 8/14 – 10/25 468 1,010 54 382 82 th 7 9/15 – 10/31 133 452 71 175 >100 th 8 5/13 – 10/31 1,360 2,840 52 1,010 75 Notably, there was unmet demand in October under the extraction-based water use estimate but not under the activity-based estimate, although the total demand in this month was much less than during the summer months. This is the result of less water available for reservoir releases and lower baseflows in the rest of the watershed; this pattern was not present on Trepanier Creek. In fact, all available water was used by Peachland in October of the deficit years.  88  Meeting Instream Flow Targets Despite the fact that Peachland’s municipal and agricultural demands were entirely met under the lower reservoir release scenario, there were consequences for meeting instream flow targets. Figure 24 below shows two example years where reservoir releases met Peachland’s demands, but left no water in the stream from July through September.  200,000  Peachland Creek, above withdrawal  m3/ day  150,000  Peachland Creek, Below withdrawal  100,000  Demand: extraction-based  50,000  9/1  8/1  7/1  6/1  5/1  4/1  3/1  2/1  1/1  12/1  11/1  9/1  10/1  8/1  7/1  6/1  5/1  4/1  3/1  2/1  1/1  12/1  11/1  10/1  0  Figure 24: Streamflow above and below withdrawal relative to demand. Scenario: 2050s CGCM2, lower reservoir releases and extraction-based water use These are the first and second year of the scenario run, but in all scenario years there were periods of no flow left in the creek after Peachland’s demands were met. Under the higher reservoir release scenario in the 2050s CGCM2 scenario, targets were met November to March among all water years, supported by reservoir releases. Targets were met for fewer days and to a lesser extent June through September. In the HadCM3 scenario, instream flow targets were similarly supported by reservoir storage November through March of the “normal” year, though, likely because it was near the beginning of the simulation and the reservoir had not been depleted yet, as winter flows were not fully supported in the “very wet” year. The HadCM3, high reservoir release scenario had the effect of making it so that water in the reservoir was insufficient to meet late summer targets, and there was no flow in the stream July and August under all but the very wet precipitation regime under the lower water use estimate (Table 34). Where reservoir release rates were lower, the number of days the target was met, and the extent to which they were met decreased during all water years, with the exception of April of the “very wet” and “normal” year in the 2050s CGCM2 scenario. Of note, instream flows were supported during October through December of the “normal” year due to baseflows that were higher than Peachland’s needs. In the HadCM3 lower reservoir scenario, unlike with the 2050s CGCM2 scenario, winter targets could not be fully met under any of the water-year conditions due to the fact that Peachland took all of the set reservoir releases, and baseflows from the rest of the watershed were not high enough to meet those targets on their own. In both climate scenarios and both reservoir release scenarios, the extraction-based estimate resulted in a more severe condition than the activity-based water use estimate (Table 34,Table 35).  89  Table 34: Number of days when zero flow in Peachland Creek during a “normal”, “very wet” and “very dry” year under the 2050s CGCM2 scenario, both water use and reservoir releases Activity-Based Estimate Extraction-Based Estimate High release Low release High release Low release “Very wet” year 0 0 8 19 “Normal” year  0  62  62  109  “Very dry” year  0  70  104  111  Of note, for the 2050s HadCM3 scenario, there were a similar number of days with no flow in Peachland Creek between the two reservoir release scenarios (Table 35). Under the higher reservoir release scenario, this was because the reservoir was drawn down to empty and thus no water was available to meet the set release for instream flow requirement. Under the lower reservoir release scenario, Peachland took all of the water released from the reservoir, leaving none for instream flows. No additional water was released to meet instream flow needs since the reservoir was given a higher priority for filling in order to maintain water in the reservoir for subsequent years. In the first scenario, meeting winter flows perhaps came at the cost of meeting early fall flows, and in the second scenario, maintaining water in the reservoir for subsequent years came at the cost of water year-round for instream flows, with no additional late fall flows. Table 35: Number of days when zero flow in Peachland Creek during a “normal”, “very wet” and “very dry” year under the 2050s HadCM3 scenario, both water use and reservoir releases Activity-Based Estimate Extraction-Based Estimate High release Low release High release Low release “Very wet” year 0 31 100 106 “Normal” year  47  80  107  111  “Very dry” year  98  108  165  113  Under the 2050s CGCM2 scenario the 400% MAD flushing flow target was unmet on Peachland Creek under the higher reservoir extraction estimate (high and low demand estimate). Interestingly, it was met under the lower reservoir release scenario for one or two days in 5 of the scenario years. The fact that the higher reservoir levels resulting from the lower releases satisfied all of Peachland’s demands must have left enough of the residual flow to flow downstream rather than being used by Peachland. Under the 2050s HadCM3 scenario, where natural peak flows were lower, the >400% MAD flushing target was never met under either demand or reservoir release scenario.  90  4.3.2.3) Drought and Climate Change (2020s) Meeting Peachland Demands Where reservoir levels were kept higher from the beginning of the scenario and less water released, reservoir levels decreased dramatically during the drought, not recovering until the “very wet” year two years after the drought (Figure 25). When initial reservoir levels were high and releases reduced, there was still enough water in the reservoir to meet Peachland’s demands at the end of the drought (Figure 26). None the less, with the lower release rates, there is a steady decline 3  in reservoir level over the period of the drought, with final volumes 4.17 million m lower than initial conditions. As can be seen in Figure 25 and Figure 26, the initial reservoir conditions prior to the drought will be the determinant of whether there will be a deficit for meeting downstream needs during and after the drought. 2  7,000,000  1.8  6,000,000  1.6 5,000,000  Reservoir volume (m3)  1.4 1.2 1  3,000,000  Inflow to Reservoir (cms)  cms  m3  4,000,000  0.8 Reservoir Release (cms)  0.6  2,000,000  0.4 1,000,000  0.2 0 4/1  4/1  10/1  10/1  4/1  4/1  10/1  4/1  10/1  10/1  4/1  4/1  10/1  4/1  10/1  10/1  4/1  10/1  0  Figure 25: Reservoir volume, inflow to reservoir and set reservoir release during Drought + 2020s, high reservoir release scenario, extraction-based water use estimate 2  12,000,000  1.8 10,000,000  1.6  Reservoir Volume (m3)  1.4  8,000,000  1  cms  m3  1.2 6,000,000  Inflow to Reservoir (cms)  0.8 4,000,000  0.6  Reservoir Release (cms)  0.4  2,000,000  0.2 0 4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  4/1  10/1  0  Figure 26: Reservoir volume, inflow to reservoir and set reservoir release during Drought + 2020s, low reservoir release scenario, extraction-based water use estimate  91  Table 36: Unmet demand in Peachland System 3 under the Drought + 2020s, high reservoirextraction scenario Year Dates Amount of Total % of Increase Demand Unmet Deficit required at Deman in Demand Increase (thousand that time d Met (thousand ÷ Deficit 3 3 m) (thousand m) 3 m) Activity-based water use estimate st 1 Drought year NA 0 NA NA NA NA nd 2 Drought year 7/1 – 9/30 742 866 86% 177 24% rd 3 Drought year 7/1 – 9/30 363 866 42% 177 49% nd 2 year after 8/28-8/31 736 40 2% 6 >100% Extraction-based water use estimate st 1 Drought year 10/24–10/31 3,650 44 8% 40 >100% nd 2 Drought year 5/26 -10/31 1,120 2,150 52% 2,118 >100% rd 3 Drought year 5/18 – 10/31 1,130 2,246 50% 2,228 >100% nd 2 year after 8/3 – 9/30 271 879 31% 938 >100%  Meeting Instream Flow Targets Under the higher reservoir release rates, instream flow targets were fully met November through March of the first two years of the drought under the activity-based water use estimate, and during the first year under the extraction-based water use estimate. This came at the cost of meeting late season demands, however, which were not met for a single day May to September all years of the drought, and met at less than 40% during the final year of the drought under the extraction-based demand estimate. Where reservoir releases were lower, those winter demands were never fully met under the activity-based water use estimates, and still not met to a higher degree during the summer months. There were more days with no flow in Peachland Creek under the higher reservoir release scenario than the lower (Table 37). This is because, despite the fact that winter flows were more fully met under the former, the reservoir became more quickly depleted and could not support both Peachland’s demands and instream flow targets during the drought. Table 37: Number of days when zero flow in Peachland Creek. Drought under climate change (2020s CGCM2), both water use and reservoir release scenarios Activity-Based Estimate Extraction-Based Estimate High release Low release High release Low release st 1 year drought 0 0 67 67 nd  2  year drought  92  62  134  78  3 year drought  92  62  160  129  rd  The target recommendation of 400% MAD flushing target was never met during the drought.  92  4.3.2.4) Land Cover Change: Mountain Pine Beetle Attack Meeting Peachland Demands Although reservoir volume decreased in the relatively drier years during summer reservoir releases, it always filled up enough during the next winter and spring to meet release requirements. The reservoir was never drawn all the way down in either the MPB Only or the MPB + 2020s scenario. There was no unmet demand under either water use estimate in this scenario.  Meeting Instream Flow Targets Instream flow targets were met more fully than baseline during the “wet,” “normal,” and “dry” year under the MPB Only scenario combined with higher reservoir extraction scenario. Even with these increases, during the “very dry” year, the targets were less than 75% met in April through August under the lower activity-based water use estimate and April and May of the extraction-based water use estimate. Under the MPB + 2020s scenario combined with higher reservoir releases, the MAD instream flow target was met more fully than baseline under all precipitation regimes and extraction estimates. None the less, the target was still less than 75% met during April through August of the “very dry” year. Under the lower reservoir release rate there was a reduction in the number of days the target was met in the winter of the “very dry” year relative to baseline, and a reduction in the amount of target that was met during the summer of the normal year. In the MPB + 2020s low reservoir release scenario, water was being withdrawn to meet Peachland needs, but not enough (as specified by the set reservoir releases) to meet aquatic life needs since reservoir filling was still prioritized first over meeting those needs. Since the reservoir was close to full throughout the scenario run, there could likely have been reservoir releases to meet instream flow needs to compensate for these deficiencies.  4.3.2.5) Meeting all of Peachland System 1 and 3 Water Needs with Reservoir Storage The supply change scenario with the greatest reduction in total annual streamflow was in the 2050s with the HadCM3 climate model. Since it had the potential to lead to the most stress for water users, this scenario was the best choice for testing the ability of Peachland Reservoir to meet the demands of both District of Peachland System 1 and 3 under more highly stressed conditions relative to current conditions. Given that there were deficits when System 3 alone was supplied under the lower initial volume, higher reservoir release combination, this simulation was run starting out with a higher reservoir volume and lower release rate. This simulation resulted in no deficit on either Peachland System 1 or 3 during the 9-year run. However, since the priority set for reservoir operations was that the reservoir be filled to meet the next  93  year’s demands before releasing water for instream flow needs, the scenario resulted in severe consequences for streamflow volumes below the District intake. There were a total of only 9 days where the target was met: in April of the “normal” year. The target was only met for 8 days of the “very wet” year. Targets were less than 50% met throughout the “very dry” year. There was no flow in Peachland Creek for 113 days in June to September of the “very dry” year, and 106 days of the “very wet” year under the extraction-based water use estimate. Of note, the number of zero flow days in June through September was no different than the result of meeting only Peachland System 3’s needs under the same climate scenario and reservoir operating rules. This is due to the fact that in both cases, Peachland used all available streamflow during the low-flow months, and the only additional water released from the reservoir went to meet the Peachland System 1’s demands. Therefore, the greatest change to instream flow needs generated from the addition of Peachland System 1 was on meeting targets October to May. The additional reservoir releases to meet Peachland 1’s demands had the consequence of drawing down the reservoir further than when just Peachland System 3 was served. The reservoir still 3  had 6.00 million m of water during the driest year of the scenario, suggesting that there may have been potential to reduce reservoir levels further to at least partially support instream flows. However, 3  it should be noted that the reservoir volume was 2.76 million m lower at the end of the scenario run than the beginning of the scenario run. This would suggest that if this simulation was run continuously it would be depleted within 27 years. Reallocating additional reservoir storage to meet conservation targets would hasten the draining of the reservoir.  94  Chapter 5: Analysis 5.1) Ranking of Impacts A principle objective of this study was to determine how the different supply scenarios impact water use needs for humans and ecosystems. An important component of that objective was to determine how the scenarios relate to each other in terms of their impact. Those comparisons are summarized here.  5.1.1) Change in Annual Streamflow The two global climate models used in this study do not indicate the same direction of change in annual total streamflow in the near term, with the CGCM2 model showing an increase in the 2020s and the HadCM3 model showing a decrease (Table 38). In the longer term (2050s), both climate models indicate a decrease in annual streamflow. Notably, the average decrease resulting from the HadCM3 model in the 2020s is similar to the decrease projected by the CGCM2 model in the 2050s. The climate change scenario for the 2050s using the HadCM3 model resulted in the greatest annual decrease in flow and the average across the 9 scenarios years and the Mountain Pine Beetle scenario in the 2020s resulted in the greatest increase in flow (Table 38). The Drought + 2020s scenario across the full 9 years is considered here and in the subsequent rankings. Important to note 3  is that during the 3 years of drought, the average annual streamflow was 20.1 million m lower than 3  3  the baseline average (7.93 million m vs. 28.4 million m ). Table 38: Change in average annual streamflow ranked by scenario Values ranked from highest to lowest (for decreases, the highest ranking is for the greatest decrease). Ranks based on mean annual flow over the 9-year scenario, Trepanier Creek, above the withdrawals. Values for drought under climate change (2020s CGCM2) compare the entire scenario, with the 3 year drought situated within the 9 year scenario, to baseline. Increase Decrease Rank Value Value Scenario Scenario 3 3 (million m ) (million m ) 1 MPB + 2020s 20.1 2050s HadCM3 -7.68 2  MPB Only  16.6  Drought + 2020s  -5.64  3  2020s CGCM2  2.76  2020s HadCM3  -2.34  2050s CGCM2  -1.71  4  95  5.1.2) Volume of Deficit in District of Peachland Peachland System 1 The climate and drought scenarios resulted in deficits in Peachland System 1, Peachland’s water system drawing from Trepanier Creek, for all but the 2020s CGCM2 scenario. The deficits from the combination of those scenarios and the two water use estimates are displayed in rank order from greatest to least deficit in Table 39. 3  Table 39: Ranking of total volume deficit (thousand m ) in Peachland System 1 across the 9year scenario run Scenarios include climate conditions and water use estimate (Section 3.5). Highest ranking denotes the greatest cumulative deficit (summed across all years) Deficit Rank Scenario 3 (thousand m ) 1 2050s HadCM3, extraction-based 758 2  2050s HadCM3, activity-based  546  3  Drought + 2020s, extraction-based  400  4  2050s CGCM2, activity-based  95  5  Drought + 2020s, activity-based  90  6  2050s CGCM2, extraction-based  52  7  2020s HadCM3, extraction-based  4  8  2020s HadCM3, activity-based  0  The much higher deficit present during the drought with the extraction-based estimate in place than the activity-based estimate can be explained by the influence of Trepanier Licenses, upstream from Peachland. The greater water demand by Trepanier Licenses during the drought led to a greater draw down of the stream, and hence a greater deficit for Peachland System 1. Also of note is that although the 2020s scenario using the HadCM3 model had a greater decrease in average annual flow due to the 2050s CGCM2 scenario, it resulted in less of a deficit than that. This can be explained by the similar reductions in summer precipitation between climate models as applied to the study area (Figure 27), and the earlier timing of peak flow under the CGCM2 scenario, which resulted in lower streamflows during the time of peak demand.  96  Percent change in monthly precipitation  30  20  10  0  -10  Precip: 2020s HadCM3 -20  Precip: 2050s CGCM2  -30 Oct  Nov  Dec  Jan  Feb  March  Apr  May  Jun  Jul  Aug  Sept  Figure 27: Monthly precipitation change, as percent, for the 2020s HadCM3 scenario and the 2050s CGCM2 scenario. Precipitation change data from the Okanagan GCM grid (CCIS Project, 2003).  Peachland System 3, Higher Reservoir Releases Considering that the drought years only reduced flows three years out of the nine, it led to a deficit similar to the 2050s CGCM2 scenario (Table 40). This shows that the drought in the near term has similar impact to climate change in the 2050s due to the significant impact of drawing down the reservoir over even a short time period. Also of note is that even though there was greater reduction in total annual flow under the 2020s HadCM3 scenario than 2050s CGCM2 scenario, the greater demand in the 2050s, extraction-based scenario pulled the reservoir down further, leading to a greater deficit (Table 40). 3  Table 40: Ranking of total volume deficit (m ) in Peachland System 3, high reservoir release, across the 9-year scenario run Highest ranking denotes the greatest cumulative deficit (summed across all years) 3 Rank Scenario Deficit (thousand m ) 1 2050s HadCM3, extraction-based 5,153 2  2050s CGCM2, extraction-based  1,961  3  Drought + 2020s, extraction-based  1,418  4  2020s HadCM3; extraction-based  1,218  5  2050s HadCM3; activity-based  1,041  6  Drought + 2020s, activity-based  363  7  2020s HadCM3, activity-based  128  8  2050s CGCM2, activity-based  107  97  5.1.3) Reduction in Reservoir Storage The metric of “carry-over storage,” defined as remaining reservoir storage at the end of the water year, helps assess the impact of a scenario on reservoir storage regardless of initial conditions. The results in Table 41 show that there were several years with a reduction in carry-over storage, even in scenarios where there was no deficit in the District of Peachland. Table 41: Proportion of the 9 scenario years with reduction in carry-over storage, ranked by total carry-over storage under all management conditions. Highest ranking denotes the greatest number of years with a reduction in carry-over storage (summed across reservoir release scenarios and water use estimates). 3 ( ) values in parentheses denote proportion of years with a reduction over 1 million m ° denotes presence of years where no change because reservoir over-tops denotes presence of years where no change because the reservoir is completely drawn down over the course of the water year, so the value here is likely an underestimate of potential for reduction High Reservoir Release Low Reservoir Release Rank  Scenario  1  2050s CGCM2  0.56 (0.33)  0.56 (0.22)  0.56 (0.11)  0.56 (0.11)  2  2020s HadCM3  0.44 (0.33)  0.67 (0.33)  0.44 (0)  0.44 (0)  3  2050s HadCM3  0.67 (0.11)  0.33 (0.33)  0.44 (0.22)  0.44 (0.22)  4  Drought + 2020s  0.44 (0.44)  0.44 (0.33)  0.44 (0.22)  0.44 (0.33)  5  2020s CGCM2  0.33 (0.33)  0.33 (0.33)  0.22 (0)°  0.22 (0.11)  6  MPB + 2020s  0.44 (0)  0.44 (0)  0.11 (0)°  0.11 (0)°  Activity  Extraction  Activity  Extraction  There was the greatest reduction in reservoir storage under the 2050s CGCM2 scenario, despite the higher annual flows for the that scenario compared to the 2050s HadCM3 and Drought + 2020s scenario. This surprising result occurs in response to the fact that demands and reservoir releases were higher than inflows to the reservoir in 5 of the 2050s CGCM2 scenario years, even under the low reservoir-release scenario. Another factor influencing this result is the fact that in the Drought + 2020s and 2050s HadCM3 scenarios the reservoir was drawn down fully for several of the water years with the higher reservoir release conditions, thus under-representing the full potential reductions. In the 2050s CGCM2 scenario, the reservoir filled up the year after it is emptied, allowing for complete representation of the supply/demand interactions. The fact that there are more years with reservoir storage reduction under the high release rate in the 2020s HadCM3 scenario than the 2050s scenario using the same climate model stems from the underestimate of decreases due to complete reservoir draw-down in the later time slice. When difference between reservoir release scenario is examined, some important patterns are revealed. The lower release did not consistently lead to more years with increased carry-over storage. There was no difference between reservoir management scenarios under the 2050s CGCM2 climate scenario, though the severity of the reduction was less under the lower reservoir release (as demonstrated by the values in parentheses). During the drought there was also no difference, though the lower release also decreases the volume of the reduction. The apparently  98  lower proportion of days with a reduction in carry over storage in the 2050s HadCM3 model (extraction-based estimate) results from the underestimate of reductions in the higher reservoir release scenario because of the complete draw-down of the reservoir. The lower reservoir release led to improvement in terms of increasing carry-over storage with both demand values only in the scenarios with an increase in annual streamflow: the 2020s climate change (CGCM2 model) and Mountain Pine Beetle scenarios. Seen in aggregate, three of the supply change scenarios resulted in a reduction in carry-over storage in over 50% of the scenario years with some combination of reservoir management and water use estimate. These results may suggest a potential for a net decrease in water available for impoundment in reservoir storage over the longer term. Reservoir management impacted how frequently carry-over storage was reduced in four of the scenarios. Water demand impacted how frequently carry-over storage was reduced in only two of the scenarios, though had a significant impact on amount of reduction in an additional two scenarios.  5.1.4) Reduction in Number of Days Instream Flow Targets Met To determine which scenario had the greatest impact on instream flows, the net reduction in number of days for which the instream targets were met relative to the baseline scenario was calculated. Days targets were met were summed across Trepanier MAD and BCINF and Peachland MAD conservation targets during the “very dry”, “normal” and “very wet” year for both water use estimates and reservoir management scenarios (representing a total of 24 years). Net change accounts for increase and decrease in days targets were met (Table 42). “Decreases Only” accounts for number of days where there was a reduction. Higher rank for the “Decreases Only” column of Table 42 denotes more days where the target was met under baseline conditions but not under the scenario. Table 42: Ranking of reduction in number of days instream flow targets met under scenario conditions compared to baseline Value includes number of days out of 24 years (“wet”, “dry” and “normal” years for all targets, water use estimates and reservoir management scenarios) when there was a net or total decrease. Highest ranking denotes greatest reduction. Net Change Decreases Only Rank Scenario Value Scenario Value 1 Drought + 2020s -3851 Drought + 2020s -3880 2  2050s HadCM3  -2675  2050s HadCM3  -2909  3  2050s CGCM2  -947  2050s CGCM2  -1486  4  2020s HadCM3  -789  2020s HadCM3  -1110  5  2020s CGCM2  -624  2020s CGCM2  -1101  6  MPB + 2020s  -274  7  MPB Only  -31  99  The reduction in days where the instream flow targets were met parallel difference in annual flow volumes only for the two driest scenarios. Despite the higher annual flows, the 2050s CGCM2 has a greater reduction than the 2020s HadCM3 scenario, which shows the greater influence of timing of flow reductions and higher withdrawals for Peachland under the later scenario. The 2020s CGCM2 and MPB + 2020s scenarios show decreases despite annually higher flows because of flow reductions in June and July combined with higher water extraction upstream those months. Changes in days meeting instream flow targets separated for the regulated and unregulated streams are presented in Table 43. Table 43: Ranking of reduction in number of days instream flow targets met under scenario conditions compared to baseline. Trepanier values include reduction in meeting BCIFN and MAD targets in “wet”, “dry” and “normal” years (12 years total). Peachland values only reflect reduction in meeting MAD target during “wet”, “dry”, and “normal” years (6 years for each reservoir release scenario). Highest ranking denotes greatest reduction. Trepanier Creek Peachland Creek (high Peachland Creek (lower Rank reservoir release) reservoir release) Scenario Value Scenario Value Scenario Value 1 Drought + 2020s -1961 2050s HadCM3 -925 2050s HadCM3 -1128 2  2050s HadCM3  -856  Drought + 2020s  -830  Drought + 2020s  -1089  3  2050s CGCM2  -396  2050s CGCM2  -350  2050s CGCM2  -740  4  2020s CGCM2  -331  2020s HadCM3  -252  2020s HadCM3  -589  5  2020s HadCM3  -269  2020s CGCM2  -222  2020s CGCM2  -548  6  MPB + 2020s  -107  MPB + 2020s  -167  On Trepanier Creek, the CGCM2 model resulted in a greater reduction than the HadCM3 model in the 2020s even though annual flows increased. This is mostly explained by the similar reduction from lower flows and higher upstream water withdrawals during June through September under both scenarios. Despite the fact that under the 2020s CGCM2 scenario there was annually higher streamflow volume and more water filling the reservoir, there was a net reduction (-171 under the high reservoir release) in the number of days the instream flow targets could be met. This likely results from the reservoir attenuating some of the higher flow events that would have contributed to meeting targets, as well as the increased water withdrawals for the District of Peachland under this scenario. The rankings are the same between high and low reservoir extraction scenarios, with the exception that the MPB + 2020s scenario resulted in a reduction in meeting the targets under low reservoir conditions. This likely stems from the fact that the operating rules incorporated into the model prioritized filling the reservoir, and the lower release volumes were not enough to meet targets even when the inflow to the reservoir was high. Inter-annual variability in precipitation has important implications for whether or not instream flow targets are met in the future, just as in the present. Under all future scenarios, best illustrated on the unregulated Trepanier Creek, there was the least change when the “very dry” years were used as  100  baseline compared to when the “very wet” and “normal” years were used (Table 44). This stems from the fact that under baseline conditions, targets were not often met in “very dry” years. Table 44: Number of days where the instream flow targets on Trepanier Creek were met under baseline conditions, but not under scenario conditions. Summarized across all months and both water use estimates. Includes only decreases and not increases in meeting instream flow targets. 2020s 2050s MPB CGCM2 HadCM3 CCGM2 HadCM3 Only 2020s Sum Trepanier Creek (BCIFN Target) -567 "Very Wet" -150 -77 -94 -228 -1 -17 -398 "Normal" -48 -58 -66 -174 -16 -36 -19 "Very Dry" 0 -2 -9 -8 0 0 Trepanier Creek (MAD Target) -333 "Very Wet" -9 -25 -58 -225 -4 -12 -510 "Normal" -88 -82 -132 -158 -10 -40 -82 "Very Dry" -16 -10 -28 -26 0 -2 An analysis of the patterns of when instream flow targets were and were not met on a monthly basis reveals other similarities between scenarios. Summarized across the “very wet,” “normal,” and “very dry” years for all three targets, there was the greatest reduction in number of days meeting targets in October, followed by November through January (Table 45). These reductions were driven by the substantial decrease in Trepanier Creek flows under the 2050s HadCM3 and Drought + 2020s scenario, and decreases under the Peachland Creek low reservoir release scenario for all supply change scenarios. There was a net increase in number of days meeting targets in April, and the second highest increase during March, both indicating the effect of the shift in the timing of the freshet in all scenarios. Although reduction in days meeting May targets also reflects some of that shift in the timing of the freshet, there was the greatest reduction in ability to meet May targets due to the drought. In June through September there were reductions (but no increases) in meeting the instream flow targets in all but the MPB scenarios. These changes were influenced by reservoir reductions under the more extreme low supply conditions, the earlier freshet recessions on Trepanier Creek, and higher demands during summer months.  101  Table 45: Change in number of days instream flow target were met each month Summarized across the “very wet” “normal” and “very dry” years for all three flow targets (24 years total). Includes high and low water use scenarios and high and low reservoir release scenarios. “+” denotes increase in days meeting target, “-” denotes decrease. Scenario CGCM 2020s HadCM3 2020s CCGM 2050s HadCM3 2050s Drought + 2020s MPB Only MPB + 2020s Sum  Oct + + + + + + +  -251 50 -300 7 -285 7 -417 0 -452 0 0 104 -39 68 1570  Nov -94 20 -130 18 -174 52 -303 0 -389 0 0 152 -45 103 -790  Dec -84 1 -63 28 -111 21 -393 4 -486 0 0 72 -31 42 1000  Jan  Feb  Mar  Apr  May  Jun  July  Aug  Sept  -62 0 -71 0 -117 8 -287 1 -495 0 0 62 -31 31  -56 5 -56 17 -71 43 -240 13 -431 0 0 56 -22 28  -70 129 -62 113 -70 145 -271 60 -350 0 0 140 0 133  -9 235 -6 95 -4 261 -3 156 -59 29 0 202 0 381  -71 37 -2 41 -171 2 -184 0 -344 0 -5 113 -40 184  -131 0 -116 0 -173 0 -187 0 -229 0 -26 51 -59 38  -87 0 -91 2 -99 0 -191 0 -210 0 0 172 -3 107  -97 0 -101 0 -103 0 -207 0 -209 0 0 152 -3 78  -107 0 -112 0 -108 0 -226 0 -226 0 0 206 -1 132  -961  -714  -103  1296  -440  -832  -400  -490  -442  5.2) Sensitivity to Changes This study revealed changes to stream hydrology occurring during all scenarios and seasons. However, all hydrologic changes did not result in “harm” to human or ecosystem users in all cases. There were certain conditions and certain time periods where the changes were felt by human and aquatic water users. Those users can be said to be the most “sensitive” to hydrologic change at those times. The following section demonstrates the condition under which water users are sensitive to change.  5.2.1) Timing of Sensitivities The most deficits were identified for July and August for the Peachland System 1, which draws from the unregulated Trepanier Creek. Thus, Peachland System 1 is the most sensitive to changes in supply and demand, particularly during July and August when streamflows were lower and demands were higher. On the regulated Peachland Creek, when the higher extraction rule was used, there were also deficits in July and August and in the fall when the reservoir was depleted earlier in the season. By the 2050s, although there was less of a demand increase in October relative to other months, this month experienced the most severe water shortages because there was insufficient water in the reservoir and the rest of the Peachland drainage to meet those increased demands. As noted in the individual scenario results in Section 4.3, the different water use estimates provided a good indication of when human and ecosystem users might be the most sensitive to supply or demand change. For example, on Peachland Creek under the 2050s HadCM3, high  102  reservoir extraction scenario, in a year where there were deficits in June to September under the lower water use estimate, deficits extended May through October under the higher use estimate because more water was withdrawn from the reservoir to meet higher demands earlier in the year. Timing of instream flow sensitivity to demand estimate is explored in section 5.2.4.  5.2.2) Peachland’s Sensitivity to Peak Flow Changes Given that these watersheds have a snow-dominated hydrology with most of the water flowing downstream in the spring, the timing of these peak flows is an important consideration for water management. When deficits were present on the Peachland water system drawing from Trepanier Creek, they occurred at the tail end of the spring freshet and resulted from both a reduction in flows and the earlier freshet recession, as shown in Figure 17 and Figure 18. The timing and quantity of filling the reservoir are important parameters for water management. Whatever volume of water is in the reservoir at the end of the freshet period is essentially the amount available for use during the summer and fall. Thus, a freshet that recedes two weeks earlier than normal (assuming no volume change) represents two additional weeks for which the reservoir will have to be drawn upon to meet water needs. If the reservoir is drawn down to the maximum withdrawal limit, the time in which there is no water in the stream begins earlier. In this study, even though the bulk of the freshet never consistently occurred as much as a month early, the earlier reservoir filling (from 0 days to 20 days earlier, see Section 4.2.1) could have contributed to some of the earlier draw downs. The impact of the timing of the freshets exclusive of volume changes could not be conclusively quantified since there was no “control” for which there was a change in timing but not a change in volume. The volume of water entering the reservoir had a noticeable impact on the ability of the reservoir to meet downstream water demands. In every scenario where there were two consecutive years where the volume of water contributed to the reservoir during the spring freshet was lower than the volume being released during May and winter flows were also lower than baseline, there was a deficit the following year in the District of Peachland.  5.2.3) Peachland’s Sensitivity to Changes in Demand vs. Supply The metric of demand-related increase compared to deficit in water available to meet Peachland’s demands showed the relative contribution of supply and demand changes to deficits. The interactions between changes in water supply and water demand can be most clearly seen on the unregulated Trepanier Creek. On Trepanier Creek the deficit was less than the increase in demand from climate change and population growth. This was found for all scenarios except the Drought + 2020s scenario with the extraction-based estimate. This finding would suggest that if population and use-related demand increases do not occur, future supply may be sufficient to meet all demands through the 2050s except under drought conditions. However, such may not be the case on the regulated stream. On this stream, especially during the later time slice, the change in demand  103  due to population growth and climate change was less than the deficit, indicating that supply-related reductions had a greater share of contribution to water stress. Some of this supply-related deficit was due to the management of the reservoir itself, which is discussed in Section 6.3.  5.2.4) Instream Flow Target Sensitivity to Water Use The two demand estimates used in this study to represent a range of possible total water use were applied to an assessment of ecosystem sensitivity to variation in water demand. The number of days where there was a difference in ability to meet instream flow targets between extraction-based and activity-based water use estimates was summed across all scenarios. On average, 10% of days of each month exhibited differences. In all scenarios under all precipitation conditions, there were at least three months where more than 30% of days resulted in a differential ability to meet instream flow targets, indicating a high degree of sensitivity of stream life to Peachland’s demands under climate change, even in the near-term. Even when the District of Peachland’s municipal demands were fully met, differences in water use estimates had an impact on meeting instream flow needs. For example, under the 2020s HadCM3 scenario, even during years where there was no deficit in the District of Peachland, the higher demand estimate led to a month and a half longer with no stream flow for Peachland Creek relative to the same scenario using the lower demand estimate. There was the greatest difference between water use estimates in October (Table 46). This difference was mostly present on Peachland Creek and strongest during the “very dry” years, where, under the high reservoir release scenario, there was not enough reservoir storage left over after meeting the higher estimate for summer demands to support any streamflow for any days in October. Since the reservoir was empty in both situations, this was because there was enough baseflow remaining after meeting the lower demands, but not the higher ones. Differences were also great in August and November (Table 46). The August difference was due to conditions in the “very wet” year. On Trepanier Creek, this was indicating that the streamflow above the withdrawal node was close to the threshold, whereby the higher demand drew the creek down below that threshold. November differences were present only on Peachland Creek, and result from the mix of flow and reservoir releases being close to the target threshold under multiple scenarios. There was the least difference between demand estimates in June through September of the “normal” and “very dry” years (Table 46). This reflects the fact that during these months the flows were usually far enough below the target that differential demand estimates did not make a difference on streamflow.  104  Table 46: Difference in number of days meeting instream flow targets between extractionbased and activity-based water use estimate Summarized across the “very wet” (W) “normal” (N) and “very dry”(D) years for MAD and BCIFN flow targets, Trepanier Creek, and MAD target on Peachland Creek, high and low reservoir release scenario (12 years total). 2020s 2020s 2050s 2050s Drought + CGCM2 HADCM3 CGCM2 HADCM3 2020s Sum W N D W N D W N D W N D W N D Oct 45 19 62 31 24 62 27 27 62 27 11 62 27 0 62 548 Nov 4 4 2 29 10 21 33 31 30 0 4 1 30 31 0 230 Dec 21 0 0 10 0 17 34 17 9 7 8 2 41 31 0 197 Jan 0 0 0 0 9 0 25 11 21 7 16 1 31 32 0 153 Feb 0 0 5 0 0 17 15 0 5 18 35 4 30 43 0 172 Mar 11 0 16 4 0 1 9 0 6 11 25 3 31 49 18 184 Apr 8 15 21 6 10 3 4 12 3 6 6 3 5 2 3 107 May 27 18 17 12 7 2 9 11 1 4 11 1 4 15 1 140 June 13 3 9 3 3 4 4 1 4 5 0 4 1 0 4 58 July 29 9 1 9 9 1 13 9 1 15 9 1 12 9 1 128 Aug 76 7 0 46 7 0 44 7 0 50 7 0 48 7 0 299 Sept 64 3 0 16 2 0 34 2 0 34 2 0 34 2 0 193  The MPB + 2020 scenario had the greatest sensitivity to demand during the “very dry” year (Table 47) due to conditions in Trepanier Creek and in the lower reservoir release scenario on Peachland Creek. The MPB Only scenario resulted in the greatest sensitivity in July through October on Peachland Creek (Table 47). The fact that differences occur here indicates that even with general increases in flow, the combination of demand, reservoir operations, and underlying climate conditions can result in less than optimal conditions for aquatic life. Table 47: MPB Only and MPB + 2020s: Difference in meeting instream flow targets between extraction-based and activity-based water use estimate. Summarized across the “very wet” (W) “normal” (N) and “very dry” (D) years for MAD and BCIFN flow targets, Trepanier Creek, and MAD target on Peachland Creek, high and low reservoir release scenario (12 years total). Month Oct Nov Dec Jan Feb Mar Apr May June July Aug Sept  MPB Only W N 28 0 0 1 10 0 0 0 0 0 1 0 1 1 2 4 3 2 9 28 24 33 17 4  D 31 1 0 0 0 5 2 2 4 19 35 25  MPB + 2020s W N D 27 8 0 0 1 35 10 0 61 0 0 62 0 0 50 1 0 18 1 5 7 3 9 6 2 2 11 10 7 16 40 4 21 18 23 28  105  Chapter 6: Discussion and Conclusion 6.1) Key Findings Use of the hydrologic modeling components of WEAP in this study allowed information on global and regional scale processes to be integrated and translated into an assessment of potential changes at a local scale. Use of the scenario development component of WEAP allowed for further insight into how those changes might interact with local scale conditions. The scenarios developed in this study are not predictions for the future, but indications of what might happen given a set of conditions. None the less, in the case of the downscaled climate models, the conditions under which many scenario runs lead to similar outputs points to a greater likelihood that those changes will take place. Where human intervention is present, such as through land use choices, demand, or reservoir management, opportunities are revealed whereby some of the negative implications of the climate scenarios may be reduced. Instances where multiple scenario runs led to similar outcomes will be presented in this section, followed by identification of patterns seen in interactions between supply and demand and indicators of possibilities to reduce impacts on municipal users and aquatic life.  Climate Many similarities are present in the hydrologic responses for both time slices and both global climate models. Maximum annual snowpack was reduced relative to baseline and the snowpack melted earlier (Table 48). These conditions resulted in earlier peak runoff. January and February had higher stream discharge, due to a greater proportion of precipitation occurring as snow. June and July had less cumulative flow volume due to lower monthly precipitation and higher evaporative demand. Similarity between models stems in part from similarity between GCMs downscaled to the study area during these time periods (Figure 11). Although present in all scenarios, these changes were generally more significant in the later time periods (e.g. 2050s) (Table 48). Table 48. Summary of hydrologic response to two global climate models downscaled to the study area for the 2020s and 2050s periods. * denotes significant difference (p <0.1); ** denotes cumulative monthly flows significantly lower in August and September; ° denotes significantly earlier date of maximum flow 2020s 2050s Metric CGCM2 HadCM3 CGCM2 HadCM3 Annual streamflow higher* lower* lower* lower* Winter streamflow higher higher* higher* higher* Summer streamflow lower lower** lower** lower** Spring freshet earlier (9 days) earlier (2 days) earlier° (16 days) earlier° (14 days) Freshet volume higher no difference higher lower Despite some convergence for GCM results for the 2050s, there is also variability between models and time slices (Table 48). This is seen in annual flow statistics, which are higher in the near term using the CGCM2 GCM, but not the HadCM3 GCM. The spring freshet occurred earlier under  106  all scenarios, with the 2050s climate change scenarios bringing the freshet up to two weeks earlier. Climate scenarios do not agree on whether the spring freshet will be accompanied by higher or lower volumes.  Drought All hydrologic indicators during the drought scenarios revealed signficantly lower flow volumes (Table 49). Since it was shown in Section 4.3 that the sequence of dry years had more of a negative impact than one dry year occurring on its own (Table 28 and Table 36), both due to depletion of soil moisture and depletion of the reservoir, it is a sequence of dry years that should be prepared for in the context of climate change. The deficits, reservoir draw-downs, and reductions in meeting instream flow targets resulting from the three-year drought in the context of climate change were comparable to impacts from climate change over 9 years in the 2050s time slice (Table 39, Table 40, Table 42, Table 43). This suggests that generally lower flows by mid-century (even with inter-annual variability) could lead to conditions like a drought from the 2020s. Table 49: Metrics of change in stream hydrology in the Drought + 2020s and and Mountain Pine Beetle attack scenarios compared to baseline conditions * denotes significance at p<0.05 for mean annual data, and significance counts in the lowest quartile for monthly median data. Significance data not provided with the drought data given that the drought data presented here is only from the 3 years of drought within the 9-year scenario. Metric Annual streamflow Winter streamflow Spring freshet Freshet volume Summer streamflow June streamflow  Drought 2020s CGCM2 lower lower earlier (24 days) lower lower lower  Mountain Pine Beetle Present Climate higher* higher* earlier (5 days) higher higher* higher  Attack 2020s CGCM2 higher* higher* earlier (11 days) higher higher* lower  Land Cover Change All indicators from the Mountain Pine Beetle scenario point to significant increases in annual mean and monthly median streamflows (Table 49). These results strongly suggest that increased flow will result from a MPB attack during the 15 year post-MPB stage modeled in the present study. When MPB conditions were combined with climate change in the 2020s, there were was, on average, less streamflow in June. However, this result was not statistically significant. An important finding of this research is the significant differences between when MPB was considered on its own versus in the context of climate change. Table 24 shows the differences between mean annual maximum SWE and mean annual streamflow, both of which were significantly different between MPB Only and MPB + 2020s scenarios. Climate change was found to offset some of the increase in maximum snow accumulation due canopy loss in the MPB scenario, and led to a greater impact on date of snow depletion than MPB conditions alone. The MPB + 2020s scenario led  107  to a greater shift in timing of freshet than either climate change or Mountain Pine Beetle conditions on their own. In terms of meeting downstream water needs, the climate related changes had more of an impact than the Mountain Pine Beetle related ones. In the MPB + 2020s scenario, the notable impact on meeting downstream needs was the lower flows in June (and no change in flow July and August) below the withdrawal for Peachland. Here, the reduced flow for June was found to be statistically significant. Changes occurred as a result of the earlier snowmelt, increased evaporative demand accompanying warmer summer temperatures, and increased demand due to climate change. In this study, the effect of the 2020s HadCM3 scenario was to reduce snowpack relative to baseline to a greater degree than in the 2020s CGCM2 scenario (Table 14). Therefore it is possible that the effects of increased maximum snow pack due to forest canopy loss under MPB conditions and decreased snow due to climate change as projected by the HadCM3 model may have offset each other even more. With snowpack volume similar to baseline conditions, the radiation-related impacts on timing of snowmelt from MPB conditions would further advance snowmelt timing and thereby increase the impacts due to timing of recession of peak flow. A further assumption may be that if a MPB scenario were to occur under a climate similar to the HadCM3 GCM projections (even warmer summer temperatures and lower summer precipitation (Figure 11)), the result would be further reductions in early summer streamflows. A definitive result of this research is that some of the negative climate-related impacts on water supplies may take place even when combined the generally higher flows that are anticipated to accompany a Mountain Pine Beetle attack. Given the likelihood of a future insect outbreak occurring in the context of climate change, the combination of these perturbations on hydrologic change are important to consider. The results of this study represent an initial step in that direction.  Interactions Between Supply and Demand and Impacts on Downstream Users Peachland Some patterns were found to emerge across several of the supply change scenarios when combined with the demand change scenarios. On Trepanier Creek, deficits occur in both scenarios set in the 2050s. These deficits occurred when “dry” and “very dry” years from baseline were perturbed. Under these conditions, both high and low demand estimates led to deficits, showing that level of demand did not make a difference in whether a deficit was present or not; just the severity of the deficit. The deficit always was present in half of July and all of August. On Peachland Creek, deficits were always present in the 2050s under the higher reservoir release rate after two consecutive “very dry” years and after the first drought year in the 2020s. Under these conditions, deficits occurred under both demand estimates but were usually present during more years under the higher demand estimate. On Peachland Creek, deficits occurred, at minimum, during all of July and August. In conclusion, given greater likelihood of summer low flows in the 2050s as shown from the climate results, deficits on both systems are likely to occur if paired with the demand increases under business-as-usual.  108  Table 50 provides a summary how the water demand affects presence/absence, volume, and duration of deficits in Peachland under the different supply change scenarios. Here, a deficit is defined as the volume of water demanded that was unmet. Table 50: Summary of impact of higher water demand on deficits in Peachland Scenario  Stream  Impact of higher water demand  2020s (HadCM3)  Trepanier Peachland Trepanier Peachland Trepanier  Deficit present with higher demand, but not with lower Adds another year with a deficit; increases volume of deficit Increases volume of deficit Adds another year with a deficit; increases volume of deficit Increases volume of deficit Adds 4 years with a deficit, increases length of deficit, increases volume of deficit Increases volume of deficit Adds two years with deficit; increases volume and duration of deficits  2050s (CGCM2) 2050s (HadCM3)  Peachland Drought + 2020s  Trepanier Peachland  Aquatic Ecosystems All scenarios resulted in a reduction in meeting instream flow targets (Table 42), implying that there is an increased likelihood that there will not be fully productive fish populations or resilient aquatic ecosystems in the future. Importantly, instream flow targets were met for fewer days even in months where there were no deficits in Peachland. For example, under the 2020s CGCM2 scenario, even though flows increased in August and September and there were no deficits in Peachland, flows below the municipality decreased, owing to the increased withdrawals in the future scenario. This shows that an indication of “no harm” to the municipality does not mean there will also be no harm downstream. For all scenarios, the ability to meet instream flow targets was reduced during very few days in the “very dry” years (Table 44). This is due to the fact that even under baseline conditions, targets are not fully met in “very dry” years. Thus, when future conditions further reduce flows, there cannot be improvement in meeting those targets. A key finding of this study, however, is that in all climate change scenarios and the drought scenario there was a net reduction in meeting targets across all precipitation regimes (Table 42). Greater reductions in meeting targets even in the wetter years should be a cause for concern. On a seasonal basis, the greatest likelihood of a reduction in meeting instream flow targets taking place is in the summer months. All climate change, reservoir management and water use scenarios led to reductions, but no increases in meeting instream flow targets in June through September (Table 45). The fact that all scenarios show the same pattern during these months gives higher weight to the likelihood of these impacts occurring. These changes were influenced by reservoir reductions under the more extreme low supply conditions, the earlier freshet recessions on Trepanier Creek, and higher demands during summer months. There may be opportunities for meeting those targets through reduction in municipal demand or alternative reservoir management strategies.  109  Opportunities for Improvement in Meeting Instream Flow Targets Water Demanded by Peachland The two demand estimates used in this study represent high and low use volumes. They can therefore be used to identify response to variation in demand and point to opportunities for improvement if demand reduction is implemented. Improvements in meeting targets through demand reduction are possible when the streamflow is close to the target. On the unregulated Trepanier Creek, this was the case under the most scenarios in the “very wet” year for all months but January and February (where targets were always met). Opportunities for improvement also may be present in July and August of “normal” years. On the reservoir-regulated Peachland Creek, there was the greatest difference between demand estimates, and thus may be a greater likelihood of improvement in meeting targets, in October of the “very dry” year, and August and September of the “very wet” year (Peachland Creek is the main contributor to the differences during those months as shown in Table 46). This response was consistent across all climate, drought, and pine beetle scenarios, indicating that the combination of flow and reservoir releases put streamflow close to the target threshold in all scenarios. The difference between demand estimates in terms of ability to meet instream flow target was found to be least in June through September of the “normal” and “very dry” years on both creeks under all scenarios (Table 46). This reflects the fact that during these months the flows were usually far enough below the target that differential demand estimates do not make a difference on meeting those targets. Therefore, there are few opportunities for improvement under those conditions with the range of demand estimates presented here.  Reservoir Operations Further opportunities for improvement in meeting instream flow targets can be seen from the following consideration of reservoir management. In all of the climate change scenarios and both reservoir operating scenarios, Peachland Creek’s instream flow targets in the “very wet” year were less frequently met relative to baseline from May to September. Thus, if reservoir release strategies modeled in this study are applied in the future, there will likely be reductions in meeting targets in these months. However, some opportunities to meet those targets (and targets during “normal” years) through a change in reservoir release strategies may be present. In all but the 2050s HadCM3 climate change scenarios (low reservoir release), there were at least some precipitation conditions where May to September targets were not met, but November through March targets were fully met (Table 51). This suggests that targets during the summer months could possibly be met by less fully meeting winter targets. Table 51 also shows that under the driest conditions (“dry” years within the 2050s, HadCM3 model, and by the third year of the drought) summer targets could not be supported by reduction in meeting winter targets by either reservoir release or demand scenario. Support for  110  instream flows during those times would have to result from reducing reservoir levels or even less fully meeting Peachland’s demands. Table 51: Cases where targets were fully met November to March, but met less fully May to September. Y = cases present; N = case not present. A “Y” indicates an opportunity to support summer flows through reduction in meeting winter targets. NA = targets are not less fully met May to September, so opportunity not present. *Indicates targets fully met December to March, but less fully met May to September Low Reservoir Release High Reservoir Release Scenario Water Year Low High Low High Demand Demand Demand Demand Y* Y* Y Y 2020s Wet CGCM2 Y Y Y Y Normal Y Y* Dry N N Y Y* Y Y Wet Y Y Y Y HadCM3 Normal Y Y* Dry N N 2050s  CGCM2 HadCM3  Drought + 2020s  MPB + 2020s  Wet Normal Dry Wet Normal Dry Year 1 Year 2 Year 3 Wet Normal Dry  Y Y N N N N N N N NA Y Y  N Y N N N N Y* N N NA Y N  Y Y Y N Y N Y Y N NA Y Y  Y Y Y N Y N Y N N NA Y Y  The lower water demand (activity-based estimate) only had more opportunities than the higher demand (extraction-based estimate) to support summer flows by reducing attainment of winter targets in a few instances shown in Table 51. It should be noted that Table 51 looks at monthly differences, whereas improvements due to differential demand is seen more clearly on the scale of days (such as seen in Table 46 and Table 47). This section has shown patterns among scenarios that indicate a greater degree of confidence in given outcomes occurring. Yet the confidence in these results needs to be qualified by consideration of how representative the model is of potential future conditions and other factors that may have led to an under- or over-estimate of outcomes. These will be discussed in the following sections.  111  6.2) Comparison to Other Studies 6.2.1) Climate Change Of the climate change modeling work conducted in southern B.C., the Merritt et al. (2006) study is the closest in geographic area and methodology to that of the present study. Merritt et al. (2006) projected climate change impacts on eight subwatersheds throughout the Okanagan Basin, although not the subbasins that are the focus of the present study. Merritt et al. (2006) did employ a semi-distributed model using the same climate models and emissions scenarios as those employed in this study. The changes in mean annual flow volumes found in the present study showed the same direction of change as those of the Merritt et al. (2006) study. In the present study and in seven of the 8 watersheds in the Merritt et al. (2006) study, mean annual volume increased using the CGCM2 climate model and A2 emissions scenario in the 2020s, and decreased in all watersheds in the 2050s. When the HadCM3 model was applied using the A2 scenario, total volume decreased in both the 2020s and 2050s, in the present study as in the Merritt et al. (2006) study. In the present study, when using the CGCM2 climate model and the A2 emissions scenario for the 2020s, the peak flow occurred earlier on average and with flow volumes either greater than or less than baseline conditions. Using the same GCM and time slice, Merritt et al. (2006) similarly projected earlier peaks and little mean difference in peak flow volume compared to current conditions, though with considerable inter-annual variability. Using the HadCM3 model, the present study found that the timing of peak flow changed very little, though with lower volumes on average. Merritt et al. (2006) predicted higher peak flow volumes with lower inter-annual variability with the HadCM3 GCM. These results likely differ due to the fact that timing of peak snow melt is sensitive to elevation, land cover, and baseline climate conditions, all of which differed between studies. What is clear from both studies is that under both climate models, snow is likely to melt earlier and recede earlier, though temporal distribution of spring flows may vary. In the 2050s, application of the CGCM2 model resulted in a freshet that occurred between 8 and 28 days earlier. Merritt et al. (2006) projected peak flows that overlapped with the latter end of that range, from 13 to 41 days earlier than baseline. Using the HadCM3 model, peak flows were significantly lower than baseline in this study. In the Merritt et al. (2006) study, peak flows were higher, though to a lesser degree than in the 2020s. In both studies, use of the HadCM3 model resulted in reductions in flow volumes in July and August. In this study volumes decreased 17% to 36% in the 2020s, and 50% to 65% reductions by the 2050s; in Merritt et al.’s (2006) application peak volumes reduced 50% to 80% in July to August by the 2080s. Thus, despite a different projected timing of peak, there was a projected reduction in flow volumes in both studies at the time of highest water demand, therefore leading to similar results during critical periods.  112  6.2.2) Land Cover Change The hydrologic response resulting from the Mountain Pine Beetle attack scenario in the present study corresponds with general predictions from other research efforts to date (Redding et al., 2008; Spittlehouse, 2006; Winkler et al., 2009). In the MPB Only scenario of the present study, the average maximum SWE increase (from 5% to 10%) among the study watersheds’ HRUs was lower than expected for a stand affected by just MPB (10% to 24%) (Winkler et al., 2009; Winkler & Boon, 2009). However, since few of the HRUs in the present study were entirely affected by Mountain Pine Beetle due to being less than 100% forested with lodgepole pine, having a lesser impact than a fully beetle-attacked stand is to be expected. The date of snowpack depletion in the MPB Only scenario in the present study was an average of 5 days earlier than baseline, which is consistent with expectations for more rapid snow melt following MPB attack (Winkler & Boon, 2009). The proportional increase in streamflow in this study was on the high end of the range of expectations based on results from other studies (Schnorbus et al., 2004; Spittlehouse, 2006; Troendle & King, 1987). Total annual streamflow in the present study was from 43% to 84% higher in the MPB Only scenario, whereas the highest value found in the literature was an increase by 87% in a single catchment that was clearcut (Spittlehouse, 2006). Though differences in results are to be expected due to study area, time period, field vs. model-based study, etc., these higher results may indicate that the assumptions used here regarding reduced interception and evapotranspiration may be overestimates. The freshet timing results in the present study are consistent with the range of the findings in the literature, which suggest that the timing of the freshet may not change, or it may occur up to three weeks earlier if beetle-damaged trees are subjected to clearcut salvage-logging. However, timing of the freshet is expected to be less if trees affected by beetle kill are not harvested (Redding et al., 2008; Spittlehouse, 2006; Winkler et al. 2009). In the present study, the onset of the freshet was firmly within that range of expectations, beginning 10 to 20 days earlier, with the peak of the freshet a week earlier than baseline. In terms of freshet volume, though high, the results are within the range of those reported by others. Previous studies have shown increases below 90% (Winkler et al. 2009; Redding et al., 2008, Spittlehouse, 2006). The results from this study showed that, on an annual basis, 7-day peak flow volumes increased from 22% to 146%; the average, however, was within the range of expectations at 59%. Low flows, here defined as flows with less than 30% of the daily discharge for the study period, were between 18% and 31% higher October to January in the scenario. Few studies have provided quantitative values for low flow changes, but two studies showed an increase in low flows: one by up to 10% and another by up to 31% (Winkler et al., 2009 Redding et al., 2008). Although the present study’s results were within the range of expectations, there are limited studies for which changes in low flow conditions following a pine beetle attack have been considered. The reason for the lack of research has been cited as freezing conditions inhibiting flow measurements when flows  113  are lowest in the winter (Winkler et al., 2008). While not the lowest annual flows, the results of the present study indicate that June flows may reduce under canopy loss conditions, warranting further attention to low flows during these time periods. There is nothing in the literature at present on the likely changes to stream hydrology under the combination of a pine beetle attack and climate change, thus, the results from the MPB + 2020s scenario could not be compared.  6.2.3) Drought In this study, the drought conditions, when placed in the context of climate change in the 2020s, resulted in a reduction in streamflow of between 72% and 74% in Trepanier Creek. These results were more extreme than those projected by the DHI (2010) study, where, though summarized across all tributaries to Okanagan Lake, a future drought reduced annual flow by only 50%. The DHI (2010) report, however, admitted that its drought projections were conservative, thus, the estimates here might not be outside the range of possible outcomes. In a 1977 study of flows in Trepanier Creek (as cited in Dobson, 2006), the 1:100 year drought was forecast to be a 72% reduction over baseline conditions, the same as was represented here.  114  6.3) Water Supply and Demand Estimates in Context 6.3.1) Reservoir Operations The results of the present study showed as little as 38% of Peachland’s demand being met June through October in a low water year under the most extreme climate change scenario in the 2050s (Table 33). These results came from the higher reservoir release scenario and may be an over-estimate of the downstream water deficits. For both reservoir release scenarios, the model was set to release a specified amount of water, even with less water coming into the reservoir. Generally, water managers will adjust releases to vary with amount of water predicted to flow into the reservoir, releasing less in drier years. Thus, some of the stress on downstream users in this study can be attributed to releasing too much water from the reservoir relative to inflow based on simplified reservoir management rules which may have resulted in an overestimate of deficit. The present study may have also over-estimated the amount of water left in the reservoir at the end of the spring freshet, thereby under-estimating the deficit. A certain amount of unpredictability is inherent in reservoir operations. In some cases more water will be released prior to spring than flows into the reservoir during the freshet in anticipation of flooding (DHI, 2010). This management component was not modeled in this study. This means that the model may have overestimated the amount of water available at the end of the spring freshet, with a resulting overestimate of water available for meeting summer requirements. In such a case, even less of Peachland’s demands and instream flow targets would be met.  6.3.2) Demand Estimates The two demand estimates used in this study reflect the uncertainty about actual water use in the study area, even at present. The two estimates were derived from two different studies with different methodologies and results. The estimates roughly represent high and low values of possible water use, with the extraction-based estimate assuming 60% more use than the activity-based estimate across the study area (Table 7). It might be assumed that the “true” values for water use are somewhere in the middle, with each estimate representing error bars around that value. Unfortunately, the uncertainty about the accuracy of the baseline estimates also limits the accuracy of the future demand estimates and thus the projected future water deficits. Of the two estimates, the activity-based estimate appears to be the best suited to representing changes in system demand, since each activity can be increased or decreased separately according to demand increase estimations. The extraction-based estimate is based on measurements of withdrawal from the source or licensed values, and thus modifications to reflect future uses cannot differentiate use by sector, but rather increase estimates need to be arrived at indirectly. As with baseline data, data on future water use cannot be assumed to be more correct with one estimate or the other, but rather provides a range of possible increased future use.  115  Another consideration of the future estimates is that they are conservative, business-as-usual estimates, as they do not account for potential adoption of water use efficiency requirements and water conservation measures. However, since the future estimates are based on a high and a low estimate for current use, the future values could be used to represent a “business as usual” and “efficiency” scenario for future use. DHI’s (2010) modeling work projected that future water use in the Okanagan as a whole would be 35% less by the 2020s with efficiency than with no efficiency. In the present study, the total water use in Peachland was 48% higher in the 2020s using the extractionbased estimate than using the activity-based estimate. In the 2050s it was 32% higher. Thus, the activity-based estimate could be seen as the estimate for efficient water use, and the extractionbased estimate could be seen as the “business-as-usual” estimate. While precise determinations of the impacts of future efficiency require more accurate data about current water use, the present study at least indicates the likely consequences of using more vs. less water.  6.3.3) Instream Flows The reductions in meeting instream flow targets may over-estimate the severity of future conditions. For consideration, some of the apparent reductions in meeting the flow targets were not actual reductions in meeting that level of flow. Rather, the discharge required to meet the target was met at an earlier time. This was found to be the case in Trepanier Creek under the 2050s HadCM3 scenario where the target for May was met in April during the “normal” and “very wet” years. Given that the timing of the cut-off for monthly targets occurs arbitrarily at the end of the calendar month, the precise timing of flow may not correspond with the required habitat function. According to a regional fisheries biologist, if the targets for one month are met on either end of that month, the habitat function of that level of flow will likely still be met (P. Epp, personal communication, 4/6/2010). None the less, as the timing of meeting the established targets begins to shift further due to natural and anthropogenic hydrologic changes, the impact on stream biota will need to be re-evaluated. In this study, the targets still may not be supporting the full range of ecosystem functions, even under the scenarios for which the streamflows targets are met. After this study was initiated, a report was released that outlines new methods for determining instream flow targets for all streams in the Okanagan Basin (ESSA Technologies Ltd. and Solander Ecological Research, 2010). These proposed targets are based on the same methodology as described by Hatfield et al. (2003), with the exception that they shift the monthly targets to align with a stream’s hydrograph rather than the calendar month, and are based on only 10 years of data rather than the recommended 20 (ESSA, 2010). As an additional method for assessing instream flow needs, they also include targets based on use of the %MAD with the specific intention of providing flow for salmonid species. Given that this information was only recently released, it was not incorporated into this study’s methodology. For comparison, however, the newly developed targets based on the BCIFN criteria represent higher flows (e.g. less conservative) than those used in the present study for all months but May. Thus, for every day but in May that the target was not met in this study, it would not be met under the new criteria, either.  116  6.3.4) Lessons from Model Calibration The model calibration showed that the degree of match between modeled and measured values varied between years, even when all parameters were optimized to give the best fit for the whole hydrograph during the calibration years. For this reason, the results given here for the various scenarios should be interpreted as changes relative to the baseline conditions within the model itself, rather than relative to the measured values. None the less, a few systematic differences between observed flows and modeled flows will be discussed here. Although the week of maximum flow matched well between measured and modeled hydrograph, the day of peak flow often differed. Thus, the model cannot be assumed to have the ability to precisely represent the day of peak flow. Where the day of peak flow dominated the resulting value for week of peak flow, it should be considered that this may be a misrepresentation of likely actual conditions. This problem might be avoided by use of a weekly time step rather than the daily time step employed in this study. On Greata Creek, the model tended to underestimate peak flows and overestimate baseflows compared to the baseline. Therefore, it is possible that flows generated from Greata Creek and therefore Peachland Creek in the model produced lower peak flows and higher baseflows that would be expected under actual conditions. The apparent systematic difference in timing of peak flow recession on Trepanier Creek is particularly important to note here. On Trepanier Creek, over the 10-year validation period, the recession occurred earlier than observed in seven of the ten years (mean=8 days earlier, S.D.= 4). This meant that there was less volume of water over the period of recession from early June to late August in those years. The scenarios revealed that on Trepanier creek, the system was most sensitive to changes in flow during June through September when the demand increases are the greatest. If the model decreases peak flow earlier than in reality, then it is showing a greater deficit than may actually occur. In conclusion, this model was optimized to match the entire hydrograph (e.g. high and low flows) over many years. While performance measurements showed a generally successful calibration, there were differences between measured and modeled hydrograph as is typical of watershed models in general. For future studies, it may be useful to optimize the model to match a specific part of the hydrograph of interest, such as the peak flow recession in June through September, in order to address specific water management questions. This approach was not feasible in the present study however, which was interested in the full range of flow conditions.  117  6.4) Hydrologic Modeling: Limitations and Advantages of the Methods Other considerations about the confidence in the results presented in this study stem from the functions of the model itself. As mentioned in Section 1.3, there are trade-offs between accurate physical representation and practical application of a model. WEAP has the advantage of only requiring data that are available from most climate stations and of being useable by those without advanced quantitative modeling training. The ability of the WEAP model to represent watershed hydrology, reservoir storage, lake storage, and downstream water use in an integrated platform was critical to addressing the full range of water balance questions. There are many limitations to a semi-distributed model being used to represent changes to the hydrology of a watershed. According to a review of hydrologic models for forest management and climate change applications in British Columbia by Beckers et al., (2009), WEAP would classify as having useful features as well as structural limitations. WEAP would rank as having “high” functionality for its use of the Penman-Monteith equation in calculating evapotranspiration, but “low” functionality for its use of temperature-based snowmelt calculations, conceptual separation of soil layers, and lack of representation of vegetation canopy. In terms of representing climate change, WEAP is not able to represent “rain-on-snow” events that will likely be important in characterizing spring snowmelt processes. In terms of examining forest management scenarios, lack of representation of a canopy layer makes it difficult to project the effect of removing a canopy layer, with only empirical methods and user-defined adjustments available to account for this type of modification. Thus, there are many nuances in land cover change and climate change that were likely not captured here. Fully distributed, physically-based models would appear to be a more useful choice of model structure. However, a more data-intensive, distributed model would have required more data than are available for the study area, resulting in omissions and inaccuracies in the model output. Additionally, improved performance in hydrological predictions alone would not have helped to address the breadth of questions explored in this study. WEAP as applied to the study area provided an appropriate level of function to meet the study objectives. There are advantages and disadvantages to the choices made in model set up. In this study, discussion is merited regarding the choice of time step. The use of the daily time step in this study provided the advantage of more closely representing the scale of the watershed given the time of concentration of approximately one day. This allowed for modeling the rapid stream response to storm events and more precise modeling of snow melt timing. In terms of impacts on downstream users, many thresholds of change were apparent with the daily time step that would not have been seen on a monthly or even weekly time step. For instance, often deficits in Peachland occurred for only part of a month. For water users, a day is a more appropriate scale than a month. Similarly, many changes were detected in the ability to meet instream flow targets on the scale of days that would not have been detected over a longer time period. Given that many ecological processes occur over the span of days or weeks (Ptolmey & Lewis, 2002), the use of the daily time step  118  provides additional ability to see if those functions may be met under future scenarios. Drawbacks of the use of the daily time step are the increased imprecision in reproducing measured flows. The Nash-Sutecliffe Efficiency index over the 10-year validation period shows the degree to which WEAP as parameterized and calibrated in this study was able to consistently reproduce daily streamflows over a long time period. This may indicate that the assumption that the model is adequately capturing the impact of alterations to hydrology on a daily basis is an overestimate of confidence in model results. Finally, a limitation of the model is present in its representation of “demand-side” processes. This study showed the importance of reservoir operation on meeting downstream water needs. However, the model did not include script to represent the full range of operating rules that water managers in this area are likely to use. Although WEAP allows for development of “conservation buffers” and evaporation from the reservoir surface (functions which were not used in this study) it does not include a function for making decisions about reservoir operations based on winter snowpack levels. Application of such a function would have likely led to reduced deficits downstream in the cases where the reservoir was drawn down so far early in the season that it was not able to meet late summer demands.  119  6.5) Future Work Many more scenarios can be run with the model in its current state, and more questions can be asked of the model with slight modifications to model structure. As currently designed, the model could be extended to: •  test the impacts of other pine beetle and partial logging scenarios;  •  simulate varying severities of wildfires with varying recovery times;  •  determine hydrological changes with different assumptions about radiation impacts;  •  incorporate changes in wind speed and humidity into climate scenarios;  •  simulate increased use of Okanagan Lake;  •  activate the other reservoirs in the watersheds;  •  add interbasin diversions to or from the study watersheds;  •  test the impacts of high-efficiency water use scenarios;  •  test the impacts of increased or decreased development in the District of Peachland;  •  re-prioritize meeting fish conservation targets vs. Peachland’s water needs during different months or even weeks;  •  test possible outcomes of new instream flow targets under a changing climate.  With only a few modifications to the model, it could be used to: •  model water quality;  •  model stream and reservoir temperature to test impacts of changes in atmospheric climate on fish;  •  represent ground water as an interaction with the lake or streams.  There are many data needs for any future modeling exercise, regardless of platform used. Foremost, there is a large gap in information about water use in the District of Peachland. Metering the amount of water being used on a daily basis in all of the different sectors would provide valuable information to address questions about an array of water issues in the District. The most pressing needs are to: determine how much agricultural land is irrigated; determine how much water is used by the commercial and industrial sector; determine how much water is used indoors vs. outdoors on a monthly, if not daily, basis.  120  6.6) Conclusion This thesis explored future scenarios for water supply and demand in a watershed in British Columbia’s Okanagan Basin by modeling the spatial and temporal interdependencies between natural processes and human and ecological water users. The interaction between climatic and watershed processes on the timing of the recession of the peak spring snowmelt on the unregulated stream was of major importance to meeting downstream needs by the 2050s at the beginning of the summer outdoor watering season in “normal” through “very dry” years. On the regulated stream, three scenarios resulted in a reduction in storage in over 50% of the scenario years, which, combined with the timing of the sequence of “wet” and “dry” years was found to be important for whether or not water needs could be met. Since the timing of lowest streamflow corresponds with that of highest demand, there were reductions in ability to meet instream flow targets June to September on the regulated stream downstream of the municipal withdrawal nodes under all future scenarios. Where summer instream flow needs were given lower priority relative to consumptive uses and reservoir filling, there were always severe consequences for instream flow levels, suggesting that these consequences will occur unless tradeoffs are considered for when and how long the optimal ecological condition will be maintained. Beyond its implications for assessing future water supply and demand in the Trepanier and Peachland Creek watersheds, this study represents a conceptual effort of integrating knowledge from the fields of climate science, forest hydrology, water systems management and stream ecology. Although WEAP has only limited means of representing changes in climate and forest cover, these means were employed to the fullest extent so as to best represent different physical characteristics in the study area. Given that this project was built on limited data compiled for use in a widely accessible platform, the methods developed here could apply to other studies with similar objectives and constraints. A potential outcome of this work is application to water related decision-making. The study showed a range of possible supply changes and potential human and ecosystem sensitivities to water use and reservoir management choices, allowing decision-makers to determine how to focus adaptation efforts. To be fully employed as a decision-making tool, the next step would need to involve stakeholders and decision-makers in further scenario development. In this way, trade-offs can be examined between stakeholder-defined social and economic priorities for land and water use as flow regimes become further outside the range of experience of water managers.  121  References BC Stats. (2009). Environmental Statistics. April, 2009. Beckers, J., Smerdon, B., & Wilson, M. (2009). Review of hydrologic models for forest management and climate change applications in British Columbia and Alberta No. 25). Kamploops, B.C.: Forrex Forum for Research and Extension in Natural Resources. Black, T.A., Spittlehouse, D.L.; Novak, M. (1989). Estimation of aerial evapotranspiration. In: Spittlehouse, D.L. (Ed.), Estimating evapotranspiration from land surfaces in B.C. IAHS Publication # 177. pp. 245-256. IAHS Press, Oxfordshire. Boon, S. (2007). Snow accumulation and ablation in a beetle-killed pine stand in Northern Interior British Columbia. B.C. Journal of Ecosystems Management, 8(3) 1-13. Boon, S. (2009). Snow ablation energy balance in a dead forest stand. Hydrological Processes, 23, 2600-2610. British Columbia Ministry of Environment, Land and Parks. (2009) Example Terrain Map and Legend Retrieved 10/2/2009 from http://www.ilmb.gov.bc.ca/risc/pubs/teecolo/terclass/sm.htm Canada-British Columbia Okanagan Basin Agreement. (1974). Technical supplement 1: Water quantity in the Okanagan Basin. Penticton, B.C.: Office of the Study Director. Canadian Climate Impacts Scenarios (CCIS). (2003). Retrieved 11/1/2009, 2009, from http://www.cics.uvic.ca/scenarios/data/select.cgi Chan, D. (2006). Assessing the water balance and future consumption scenarios for demand management of the Aldergrove aquifer in B.C. Master of Science Thesis, University of British Columbia, Vancouver, B.C. Chaves, M.M., Pereira, J.S., Maroco, J., Rodrigues, M.L., Ricardo, C.P., Osorio, M.L., Carvalho, I., Faria, T. and C. Pinheiro. (2002) How plants cope with water stress in the field? Photosynthesis and growth. Annals of Botany. 89: 907-916. Cohen, S., & Kulkarni, T. (Eds.) (2001). Water management and climate change in the Okanagan Basin. Environment Canada. Cohen, S., Neilsen, D., & Welbourne, R. (Eds.). (2004). Expanding the dialogue on climate change and water management in the Okanagan Basin, British Columbia. Final report. Environment Canada. District of Peachland (2003). Summary: Official Community Plan (OCP). Retrieved 8/19/2010 from http://www.peachland.ca/services/planning/ocp/ocp_maps/OCP2002.pdf DHI Water and Environment (DHI). (2009) Okanagan Water Accounting Model. Draft. DHI Water and Environment (DHI). (2010) Okanagan Basin Water Accounting Model. Final Report. May, 2010. Dobson, D. (2003). Extent of snow cover during the 2002 spring freshet for the Peachland Creek watershed (second year of snowline data) No. File: 544-011 Project: 22043). Dobson Engineering, Ltd. Dobson, D. (2006). District of Peachland water availability analysis No. 557-002/25068.  122  Dobson, D. (2009). Impacts of the Mountain Pine Beetle on community water supplies. Mountain Pine Beetle and Water Management Workshop Proceedings. Kelowna, B.C.: FORREX Forum on Research and Extension in Natural Resources Society. Dodson, R., & Marks, D. (1997). Daily air temperature interpolated at high spatial resolution over a large mountainous region. Climate Research, 8. 1-20. Donley, E. (2010). Strategic planning for water right acquisitions in the Central Columbia Basin: an assessment of regional streamflow response to climate change. Masters Thesis, University of Washington, Seattle. Economic Development Commission (EDC) (2009). District of Peachland 2009 Population Projections. Regional District of the Central Okanagan. Kelowna, B.C. Retrieved 2/15/2010, from: http://www.regionaldistrict.com/departments/edc/default.aspx Environment Canada (Producer) (2007). HYDAT for Windows version 2.04. [CD-ROM] Water Survey of Canada. Environment Canada (2009a). National Climate Data and Information Archive. Retrieved 10/1/2009 from www.climate.weatheroffice.gc.ca Environment Canada (2009b) Adjusted Historical Canadian Climate Data. Retrieved 10/10/2009 from http://www.cccma.bc.ec.gc.ca/hccd/data/access_data_e.shtml ESSA Technologies Ltd. and Solander Ecological Research. (2009). Instream flow needs analysis for the Okanagan Water Supply & Demand Project. Report prepared for the Okanagan Basin Water Board (OBWB), Kelowna, BC. 152 p. Food and Agriculture Organization. (FAO). (1998a). Chapter 2: FAO Penman-Monteith Equation. FAO Irrigation and Drainage Papers. Retrieved 3/21/2010 from: http://www.fao.org/docrep/X0490E/x0490e06.htm Food and Agriculture Organization. (FAO). (1998b). Chapter 6: Etc Single Crop Coefficient Kc. FAO Irrigation and Drainage Papers. Retrieved 3/21/2010 from http://www.fao.org/docrep/X0490E/x0490e0b.htm Field, C. B., Mortsch, L. D., Brklacich, M., Forbes, D. L., Kovacs, P., Patz, J. A., et al. (2007). North America. In M. L. Parry, O. F. Canziani, J. P. Palutikof, P. J. van der Linden & C. E. Hanson (Eds.), Climate change 2007: Impacts, adaptation and vulnerability. Contribution of working group II to the fourth assessment report of the Intergovernmental Panel on Climate Change. Cambridge, U.K.: Intergovernmental Panel on Climate Change: Cambridge University Press. Fontaine, T.A., Klassen, J.F., Cruickshank, T.S., Hotchiss, R.H. (2001). Hydrological response to climate change in the Black Hills of South Dakota, USA. Hydrological Sciences, 46(1), Gedney, N., Cox P.M., Betts, R.A., Boucher O., Huntingford C., Stott P.A. (2006). Detection of a direct carbon dioxide effect in continental river runoff records. Nature, 439: 835-838. Georgakakos, A. P. (2007). Decisions support systems for integrated water resource management with application to the Nile Basin. In A. Castelletti, & R. Soncini-Sessa (Eds.), Topics on system analysis and integrated water resource management (1st ed.,). Amsterdam, The Netherlands: Elsevier. Grainger, B. (2009). Trout Creek hydrologic risk assessment No. 08-012. B.C. Ministry of Environment. Hanks, R., & Ashcroft, G. (1980). Applied soil physics. Berlin-Heidelberg-New York: Springer-Verlag.  123  Hatfield, T., Lewis, A., Ohlson, D., & Bradford, M. (2003). Development of instream flow thresholds as guidelines for reviewing proposed water uses. British Columbia. Hornberger, G. M., Raffensperger, J. P., Wiberg, P. L., & Eshleman, K. N. (1998). Elements of physical hydrology. Baltimore and London: The John Hopkins University Press. Huber-Lee, A., Swartz, C., Sieber, J., Goldstein, J., Purkey, D., Young, C., et al. (2006). Decision support system for sustainable water supply planning. Water Research Foundation. Huggard, D., & Lewis, D. (2007). Summary of: ECA effects of options for mountain pine beetle salvage - stand and watershed level reports. Ingol-Blanco, E., & McKinney, D. (2009). Hydrologic model for the Rio Conchos Basin: calibration and validation No. 08-09. Austin, Texas: Center for Research in Water Resources. Ivanovic, R., & Freer, J. (2009). Science versus politics: truth and uncertainty in predictive modeling. Hydrological Processes, 23, 2549-2554. Jain, S., & Sudheer, K. (2008). Fitting of hydrologic models: A close look at the Nash-Sutcliffe Index. Journal of Hydrologic Engineering, 13(10) Jost, G., Weiler, M., Gluns, D., & Alila, Y. (2007). The influence of forest and topography on snow accumulation and melt at the watershed scale. Journal of Hydrology, 347, 101-115. Kaufmann, M. (1983). A canopy model (RM-CWU) for determining transpiration of subalpine forests. I: Model development. Canadian Journal of Forest Resources, 14, 218-226. Knight, D. Fahey, T., & S. Running. (1985). Water and nutrient flow from contrasting lodgepole pine forests in Wyoming. Ecological Monographs, 55(1). 18-29. Krause, P., Boyle, D., & Base, F. (2005). Comparison of different efficiency criteria for hydrological model assessment. Advances in Geosciences, 5, 89-97. Kundzewicz, Z. W., Mata, L. J., Arnell, N. W., Döll, P., Kabat, P., Jiménez, B., et al. (2007). Freshwater resources and their management. In M. L. Parry, O. F. Canziani, J. P. Palutikof, P. J. van der Linden & C. E. Hanson (Eds.), Climate change 2007: Impacts, adaptation and vulnerability. Contribution of Working Group II to the fourth assessment report of the Intergovernmental Panel on Climate Change Langsdale, S., Beall, A., Carmichael, J., Cohen, S., & Forster, C. (2007). An exploration of water resource futures under climate change using system dynamics modeling. The Integrated Assessment Journal, 7(1), 51-79. Le Treut, H., Somerville, R., Cubasch, U., Ding, Y., Mauritzen, C., Mokssit, A., et al. (2007). Chapter 1: Historical overview of climate change science. In S. Solomon, et al. (Eds.), Climate change 2007: The physical science basis. Contribution of Working Group I to the fourth assessment report of the Intergovernmental Panel on Climate Change, 2007). Cambridge, U.K. and New York, N.Y.: Cambridge University Press. Link, T., & Marks, D. (1999). Point simulation of seasonal snow cover dynamics beneath boreal forest canopies. Journal of Geophysical Research, 104(D22), 841-857. Loucks, D. P., Stedinger, J. R., & Haith, D. A. (1981). Water resource systems planning and analysis. Englewood Cliffs N.J.: Prentice-Hall. Loukas, A., Vasiliades, L., & Dalezios, N. (2004). Climate change implications of flood response of a mountainous watershed. Water, Air, and Soil Pollution: Focus, 4, 331-347.  124  Maayar, M., & Chen, J. (2006). Spatial scaling of evapotranspiration as affected by heterogeneities in vegetation, topography and soil texture. Remote Sensing of Environment, 102(1-2), 33-51. Maurer, N. (2010). Modeling urban development trends and outdoor residential water demand in the Okanagan Basin, British Columbia. Master’s Thesis, University of British Columbia, Vancouver, B.C. Mbogga, M., Hamaan, A., & Wang, T. Historical and projected climate data for natural resource management in Western Canada. Agriculture and Forest Meteorology, 149(5), 881-890. Merritt, W., Alila. Y. (2004). Chapter 7: Hydrology. In Cohen, Neilsen and Welbourn (Eds.), Expanding the dialog on climate change in the Okanagan Basin. 75-80. Merritt, W., Alila, Y., Barton, M., Taylor, B., Cohen, S., & Neilsen, D. (2006). Hydrologic response to scenarios of climate change in sub watersheds of the Okanagan Basin, British Columbia. Journal of Hydrology, 326, 79-108. Ministry of Agriculture and Lands (2003). TRIM Contour Lines 1:20,000. [computer file] Victoria, B.C. 3/11/2003. Ministry of Energy, Mines and Petroleum Resources (2009). Digital Terrain & Soils Map Library. Retrieved 10/1/2009, 2009, from http://www.empr.gov.bc.ca/Mining/Geoscience/TerrainandSoilMaps/Pages/default.aspx Ministry of Environment (2010). British Columbia’s Water Act modernization: Technical background report. Water Stewardship Division. Retrieved 4/30/2010 from: http://www.livingwatersmart.ca/water-act/ Ministry of Environment (2009a). Community Watershed Boundaries [computer file]. Victoria, B.C.: Environmental Science and Information Branch. 7/15/2009. Retrieved 10/1/2009 from Geographic Data Discovery Service: www.geobc.gov.bc.ca. Ministry of Environment (2009b). Historic Snow Survey Data. River Forecast Centre. Retrieved, 9/15/2009, from http://www.agf.gov.bc.ca/rfc/ Ministry of Environment (2008a). Biogeoclimatic Ecosystem Classification Map [computer file]. Victoria, B.C.: Forest Research Branch. 5/15/2008. Retrieved 10/1/2009 from Geographic Data Discovery Service: www.geobc.gov.bc.ca. Ministry of Environment (2008c). Mountain pine beetle attack map [computer file]: Remote Sensing and Geospatial Applications. Ministry of Environment (2000). Protected Areas of British Columbia Act. SBC 2000, Chapter 17. Victoria, B.C. Retrieved, 5/18/2010, from http://www.bclaws.ca/EPLibraries/bclaws_new/document/ID/freeside/00017_06 Ministry of Forests and Range. (2009) Mountain Pine Beetle GIS Data [ArcGIS Personal Geodatabase v 9.1]. Victoria, B.C.: Remote Sensing and Geospatial Applications. 6/23/2008. Ministry of Forests and Range (2008). Vegetation Resource Inventory [computer file]. Victoria, B.C.: Retrieved 10/1/2009 from Geographic Data Discovery Service: www.geobc.gov.bc.ca. Ministry of Forests and Range (1990). Vegetation Resource Inventory [computer file]. Victoria, B.C. Ministry of Water, Land and Air Protection. (1999). Peachland lake dam operation agreement between the corporation of the District of Peachland and Ministry of Water, Land and Air Protection.  125  Moore, D., & Scott, D. F. (2005). Camp Creek revisited: Streamflow changes following salvage harvesting in a medium-sized, snowmelt-dominated catchment. Canadian Water Resources Journal, 30(4), 331-344. Moore, D., & Wondzell, S. (2005). Physical hydrology and the effects of forest harvesting in the pacific northwest: A review. Journal of the American Water Resources Association, 41(4) 763784. Morrison, J., Quick, M. C., & Foreman, M. G. (2002). Climate change in the Fraser River watershed: flow and temperature projections. Journal of Hydrology, 263(1-4), 230-244. Nakicenovic, N., Alcamo, J., Davis, G., de Vries, B., Fenhann, J., Gaffin, S., Gregory, K. et al. (2001). Special report on emissions scenarios. Intergovernmental Panel on Climate Change. GRIDArendal. Retrieved 8/19/2010 from http://www.grida.no/publications/other/ipcc_sr/ Nash, L.L., Gleick, P.H. (1991). Sensitivity of streamflow in the Colorado Basin to climatic changes. Journal of Hydrology, 125(3-4) 221-241. Nash, J. E. and J. V. Sutcliffe (1970), River flow forecasting through conceptual models part I — A discussion of principles, Journal of Hydrology, 10 (3), 282–290. Neale, T., Carmicheal, J. & Cohen, S.(2007). Urban Water Futures: A multivariate analysis of population growth and climate change impacts on urban water demand in the Okanagan Basin, B.C. Canadian Water Resources Journal, 32(4), 315-330. Neilsen, D., Smith, S., Frank, G., Koch, W., Alila, Y., Merritt, W.S., Taylor, W.G., Barton, M.,Hall, J.W., and Cohen, S.J. (2006). Potential impacts of climate change on water availability for crops in the Okanagan Basin, British Columbia. Canadian Journal of Soil Science. 86: 921-936. Northwest Hydraulic Consultants (NHC) (2001). Hydrology, water use and conservation flows for Kokanee Salmon and Rainbow Trout in the Okanagan Lake Basin, B.C. Prepared for: B.C. Fisheries, Victoria, B.C. Northwest Hydraulic Consultants (NHC). (2003a). Trepanier Creek Stream Summary (draft). Prepared for B.C. Fisheries (Fisheries Management Branch). Northwest Hydraulic Consultants (NHC). (2003b). Peachland Creek Stream Summary (draft). Prepared for B.C. Fisheries (Fisheries Management Branch). Novak, M. D., & Black, T. A. (1982). Test of an equation for evaporation from bare soil. Water Resources Research, 18, 1735-1737. Okanagan Basin Water Board (OBWB) (2010) Key Findings: Okanagan Water Supply and Demand Project (2010). Retrieved March 26, 2010 from http://www.obwb.ca/fileadmin/docs/100326_key_findings.pdf Oke, T. R. (1987). Boundary layer climates (Second Ed.). Methuen, London: University Press. Quick, M.C., & Pipes, A. (1972). Daily and seasonal forecasting with a water budget model. Proceedings of the UNESCO/WMO/IAHS Symposium. 106, 1017-1034. Banff, B.C. Patterson, M. (n.d.) Water management and molybdenum treatment at the closed Noranda Inc. Brenda Mines site, Peachland, B.C. Westbank, B.C.: Noranda, Inc. Ptolmey, R., & Lewis, A. (2002). Rationale for multiple British Columbia instream flow standards to maintain ecosystem function and biodiversity. Ministry of Water, Land and Air Protection and Ministry of Sustainable Resource Management. Victoria, B.C. Redding, T., Winkler, R., Spittlehouse, D., Moore, R., Wei, A., & Teti, P. (2009) Mountain Pine  126  Beetle and watershed hydrology: A synthesis focused on the Okanagan Basin. One Watershed One Water Conference Proceedings. Redding, T. R., Winkler, R., Teti, P., Spittlehouse, D., Boon, S., Rex, J., et al. (2008). Mountain pine beetle and watershed hydrology. Paper presented at the Mountain Pine Beetle: From Lessons Learned to Community-Based Solutions, 9(3) 33-50. Richter, B., Baumgartner, J., Powell, J., & Braun, D. (1996). Method for assessing hydrologic alteration within ecosystems. Conservation Biology, 10(4), 1163-1174. Running, S., Nemani, R., Peterson, D., Band, L., Potts, D., Pierce, L., et al. (1989). Mapping regional forest evapotranspiration and photosynthesis by coupling satellite data with ecosystem simulation. Ecology, 70(4), 1091-1101. Schnorbus, M., Winkler, R., & Alila, Y. Forest harvesting impacts on the peak flow regime in the Columbia mountains of southeastern British Columbia: An investigation using long-term numerical modeling. Water Resources Research, 40. Singh, V. P., & Woolhiser, D. A. (2002). Mathematical modeling of watershed hydrology. Journal of Hydrologic Engineering, 7(4), 270-291. Spittlehouse, D. L. (2006). Annual water balance of high elevation forest and clearcut sites. Paper presented at the 27th Conference on Agricultural and Forest Meteorology. Stockholm Environment Institute. (1997). Water Evaluation and Planning System, Version 2.3053 [computer program]. Stockholm Environment Institute. (2007). Water Evaluation and Planning System: User’s Guide. Stonefelt, M. D., Fontaine, T. A., & Hotchkiss, R. H. (2000). Impacts of climate change on water yield in the Upper Wind River Basin. Journal of the American Water Resources Association, 36, 321336. Summit Environmental Consultants Ltd. (Summit) (2004), Trepanier landscape unit water management plan. Vernon, B.C. Summit Environmental, Inc. & DHI Water and Environment (2009). Okanagan Hydrology Model (MIKE-She). Summit Environmental, Inc. & Polar Geoscience Ltd. (2009). Surface water hydrology and hydrologic modeling study, Part 1: “State of the Basin” report. Draft, March 2009. Vernon, B.C. The Nature Conservancy. (2009). Indicators of Hydrologic Alteration software, version 7.1 [computer program] Taylor, B., Barton, M. (2004). Chapter 4. Climate. In Cohen, S., Neilsen, D., and Welbourn, R. (Eds.) Expanding the dialog on climate change and water management in the Okanagan Basin, British Columbia, Final Report, 24-45. Taylor, S.W., Carroll, A.L. (2003). Disturbance, forest age, and Mountain Pine Beetle outbreaks in B.C.: A historical perspective. Mountain Pine Beetle Symposium: Challenges and Solutions. October 30-31, 2003. Victoria, B.C.: Natural Resources Canada. Tognetti, R., Longbucco, A.M, Miglietta, F. and Raschi, A.(1999). Water relations, stomatal response and transpiration of Quercus pubescens trees during summer in a Mediterranean carbon dioxide spring. Tree Physiology, 19(4-5): 261-270. Tolko Industries, Ltd. “Peachland/Trout Retention Plan Overview” [map] 1:30,000. Vernon, B.C. 2009.  127  Troendle, C. A., & King, R. M. (1987). The effect of partial and clearcutting on streamflow at Deadhorse Creek, Colorado. Journal of Hydrology, 90(1-2), 145-157. Urban Systems (2005). The District of Peachland water conservation drought management study. Kelowna, B.C. Urban Systems (2007). District of Peachland: Water Master Plan. Kelowna, B.C. U. S. Army Corps of Engineers. 1956. Snow Hydrology, Summary Report of the Snow Investigations, North Pacific Division, Portland Oregon, 437 pp. Uunila, L., Guy, B., & Pike, R. Hydrologic effects of mountain pine beetle in the interior pine forests of British Columbia: Key questions and current knowledge. Streamline Water Management Bulletin, 9. VanShaar, J., Haddeland, I., & Lettenmaier, D. (2002). Effects of land-cover changes on the hydrological response of interior Columbia River Basin forested catchments. Hydrological Processes, 16, 2499-2520. Walton, A. (2010). Provincial-level projection of the current mountain pine beetle outbreak: Update of the infestation projection based on the 2009 provincial aerial overview of forest health and the BCMPB model (year 7). B.C. Forest Service, Research Branch. Wang, T. H., A., Spittlehouse, D., & Aitken, S. (2006). Development of scale-free climate data for western canada for use in resource management. International Journal of Climatology, 26, 383397. Whitaker, A., Alila, Y., & Beckers, J. T., D. (2002). Evaluating peak flow sensitivity to clear-cutting in different elevation bands of a snowmelt-dominated mountainous catchment. Water Resources Research, 38(9), 1172. Williams, D. W., Liebhold, A.M. (2002). Climate change and the outbreak ranges of two North American bark beetles. Agricultural and Forest Entomology, 4 (2): 87-99. Winkler, R., Moore, R., Redding, T., Spittlehouse, D., Carlye-Moses, D., & B. Smerdon. (2009). Chapter 6 - Hydrological Processes and Watershed Response. In: Compendium of Forest Hydrology and Geomorphology in British Columbia, Victoria, B.C.: B.C. Ministry of Forests and Range, Research Branch. Winkler, R., Rex, J., Teti, P., Maloney, D., Redding, I. (2008). Mountain Pine Beetle, forest practices and watershed management. Extension Note 88, Victoria, B.C.: Ministry of Forests and Range Science Program. Winkler, R., Spittlehouse, D., & Golding, D. (2005). Measured differences in snow accumulation and melt among clearcut, juvenile, and mature forests in southern British Columbia. Hydrological Processes, 51-62. Winkler, R., Spittlehouse, D., Allen, D., Redding, T., Giles, T., Hope, G. (2008). The Upper Penticton Creek watershed experiment: integrated water resource research on the Okanagan plateau. One Watershed - One Water Conference Proceedings, Kelowna, B.C.: Canadian Water Resources Association. Woods, S., Ahi, R., Sappington, J., & McCaughey, W. (2006). Snow accumulation in thinned lodgepole pine stands, Montana, USA. Forest Ecology and Management, 235(1-3), 202-211.  128  Yates, D., Purkey, D., Sieber, J., Huber-Lee, A., & Galbraith, H. (2005a). WEAP21 - A demand, priority, and preference-driven water planning model. part 2: Aiding freshwater ecosystem service evaluation. International Water Resources Association, 30(4), 501-512. Yates, D., Sieber, J., Purkey, D., & Huber-Lee, A. (2005b). WEAP21 - A demand, priority, and preference-driven water planning model. part 1: Model characteristics. Water International, 30(4), 487-500. Young, A. R. (2006). Stream flow simulation within UK ungauged catchments using a daily rainfallrunoff model. Journal of Hydrology, 320, 155-172. Young, C., Escobar-Arias, M., Fernandes, M., Joyce, B., K., M., Mount, J., Mehta, V., et al. (2009). Modeling the hydrology of climate change in California's sierra Nevada for subwatershed scale adaptation. Journal of the American Water Resources Association,1-15. Zimmerman, J. (2006). Hydrologic effects of flood control dams in the Ashuelot River, New Hampshire, and West River, Vermont. The Nature Conservancy. Retrieved 7/6/2010 from http://conserveonline.org/library/Hydrologic%20analysis%20Ashuelot%20and%20West.R1.do c  129  Appendices Appendix A) Formulas Albedo formula developed for the daily time series used in this study: If(ts=1,0.2, (Snow Accumulation[mm])=0, Albedo Lower Bound, (Snow Accumulation[mm])>(PrevTSValue(Snow Accumulation[mm])), (If(Or((Snow Accumulation[mm])-PrevTSValue(Snow Accumulation[mm])>=10,PrevTSValue(Albedo Data)=Albedo Upper Bound),Albedo Upper Bound,PrevTSValue(Albedo Data))), (Snow Accumulation[mm])=(PrevTSValue(Snow Accumulation[mm])), (PrevTSValue(Albedo Data)), Max(Albedo Lower Bound, PrevTSValue(Albedo Data)-0.019)) *Ts = time step; PrevTSValue = Previous time step value; Max = maximum of the referenced; Albedo Lower Bound = albedo of a snow-free surface; Albedo Upper Bound = albedo of new snow  130  Appendix B) Data Used in Water Use Estimates, Baseline Demand Site Trepanier Licenses  Star Place  Dietrich  Activity Agriculture Domestic Indoor  Annual 3 Total (m ) ExtractionBased  158,246  20,928  Total  185,169  Domestic Indoor  990  Domestic Outdoor  3,456  Total  4,356  Domestic Indoor  990  Domestic Outdoor  3,456  Total  4,356  814,000  11,000  OBWB use per person * population (Summit) Remainder per capita use (Urban Systems estimate - indoor use estimate from OBWB * population (Summit) Extraction-based based on License; Activitybased is sum of above OBWB use per person * households (Summit)* average persons per household (OED) OBWB use per person * households (Summit)* average persons per household (OED) Extraction-based based on License; Activitybased is sum of above OBWB use per person * Est. population (Summit)  10,000  OBWB use per person* Est. population (Summit) Extraction-based based on License; Activitybased is sum of above OBWB use per person * population (Allin, pers comm. 10/26/2009)  137,500  Domestic Outdoor  685,000  Commercial/ Institutional  127,458  Remainder per capita use (Urban Systems estimate - indoor use estimate from OBWB * population (Allin, pers comm.) Half of Urban System's estimated based on land in commercial/industrial (assuming yearround rate was same a maximum day)  Parks  10,230  Half of the estimate from Urban Systems for Parks.  55,000  Total from "Waterworks" from Trepanier creek (Summit) + Dobson's estimate for lake use (higher than Summit's) District of Peachland's est. for hectares [57] * Summit's average annual water use/ha Estimate of use of water by agriculture based on extraction data (Dobson) Sum of Above OBWB use per person * population (Allin, pers comm)  274,000  Remainder per capita use (Urban Systems estimate - indoor use estimate from OBWB * population (Allin, pers comm.)  880,250 Agriculture  Peachland System 2  Data Source Area and demand/ha estimate (Summit)  5,995  Domestic Outdoor  Domestic Indoor Peachland System 1  Annual 3 Total (m ): ActivityBased  Total Domestic Indoor Domestic Outdoor  410,001  1,370,189  488,000 1,368,250  131  Demand Site  Activity  Commercial/ Institutional  Peachland System 3  Total Domestic Indoor Domestic Outdoor  Annual 3 Total (m ): ActivityBased  Annual 3 Total (m ) ExtractionBased  Data Source  767,000  82,500  Golf Course ha * estimated use/ha (Summit) Activity-based: Summit's estimate based on total capacity of withdrawal from wells; extraction-based: Sum of above OBWB use per person * population (Allin, pers comm)  288,000  Remainder per capita use (Urban Systems estimate - indoor use estimate from OBWB * population (Allin, pers comm.)  250,000 329,000  Commercial/ Institutional  127,458  Parks  10,230 1,611,000 467,545  Agriculture Total Grand Total  975,733 2,868,803  631,000 2,242,000 5,212,250  Half of Urban System's estimated based on land in commercial/industrial (assuming yearround rate was same a maximum day) Half of the estimate from Urban Systems for Parks. Dobson Waterworks estimate DoP's est. for hectares [65] * Summit's average annual water use/ha Estimate of use of water by agriculture based on extraction data (Dobson) Sum of above  132  Appendix C) Increases in Demand in Future Scenarios Due to population growth and climate change, 2020s and 2050s, activity-based and extraction-based water use estimates C.1) Peachland System 1: Climate and population-related increases in demand (activitybased estimate and extraction-based estimate), 2020s  Water Demand in Baseline and 2020s: Peachland System 1 Actitybased and Extraction-based Demand requirements 600,000 Baseline Activity-Based Est  500,000  2020s Activity-Based Est 400,000  Baseline Extraction-Based Est  300,000  2020s Extraction-Based Est  200,000 100,000 0 Oct  Nov  Dec  Jan  Feb  Mar  Apr  May  June  July  Aug  Sept  C.2) Peachland System 3: Climate and population-related increases in demand (activitybased estimate and extraction-based estimate), 2020s  Water Demand in Baseline and 2020s: Peachland System 3: Actity-based and Extraction-based Demand requirements 600,000 Baseline Activity-Based Est  500,000  2020s Activity-Based Est  400,000  Baseline Extraction-Based Est 2020s Extraction-Based Est  300,000 200,000 100,000 0 Oct  Nov  Dec  Jan  Feb  Mar  Apr  May  June  July  Aug  Sept  133  C.3) Peachland System 1: Climate and population-related increases in demand (activitybased estimate and extraction-based estimate), 2050s  Water Demand in Baseline and 2050s: Peachland System 1 Activitybased and Extraction-based Demand requirements  700,000 600,000  Baseline Activity-Based Est  500,000  2050s Activity-Based Est  400,000  Baseline Extraction-Based Est 2050s Extraction-Based Est  300,000 200,000 100,000 0 Oct  Nov  Dec  Jan  Feb  Mar  Apr  May June July  Aug Sept  C.4) Peachland System 3: Climate and population-related increases in demand (activitybased estimate and extraction-based estimate), 2050s  700,000  Water Demand in Baseline and 2050s: Peachland System 3: Actitybased and Extraction-based Demand requirements Baseline Activity-Based Est  600,000  2050s Activity-Based Est  500,000  Baseline Extraction-Based Est 2050s Extraction-Based Est  400,000 300,000 200,000 100,000 0 Oct  Nov  Dec  Jan  Feb  Mar  Apr  May  June  July  Aug  Sept  134  Appendix D) Streamflow in Relation to Instream Flow Targets Sample Scenarios, Years and Instream Flow Targets E.1) Trepanier Creek MAD, CGCM2 2020s, Normal Year Trepanier Creek, 2050s CGCM2 "normal" year  2.5  Trepanier Creek Conservation Target (MAD)  2  cms  Delivered, Activity-based est. 1.5 Delivered, Extraction-based est. 1  0.5  0 10/31  12/1  1/1  2/1  3/4  4/4  5/5  6/5  7/6  8/6  9/6  E.2) Trepanier Creek MAD, CGCM2 2020s, Very Dry Year Trepanier Creek, 2050s CGCM2 "very dry" year 2.5 Trepanier Creek Conservation Target (MAD)  2 1.5  Delivered, Extraction-based est.  1 0.5  9/29  9/1  9/15  8/4  8/18  7/7  7/21  6/9  6/23  5/26  4/28  5/12  3/31  4/14  3/3  3/17  2/4  2/18  1/7  1/21  12/24  11/26  12/10  10/29  11/12  10/1  0 10/15  cms  Delivered, Activity-based est.  135  E.3) Trepanier Creek MAD, 2050s CGCM2, Very Wet Year Trepanier Creek, 2050s CGCM2 "very wet" year  2.5  Trepanier Creek Conservation Target (MAD)  cms  2  Delivered, Activity-based est.  1.5  Delivered, Extraction-based est.  1 0.5  9/1  8/1  7/1  6/1  5/1  4/1  3/1  2/1  1/1  12/1  11/1  10/1  0  E.4) Peachland Creek, CGCM2 2050s, Low Reservoir Release, Normal Year Peachland Creek, 2050s CGCM2, Low Reservoir Release "normal" year 1.2 Peachland Creek Conservartion Target (MAD)  1  Delivered, Activity-based est.  cms  0.8 Delivered - Extraction-based est.  0.6 0.4 0.2 0 10/31  12/1  1/1  2/1  3/4  4/4  5/5  6/5  7/6  8/6  9/6  E.5) Peachland Creek, CGCM2 2050s, High Reservoir Release, Normal Year Peachland Creek, 2050s CGCM2, High Reservoir Release "normal" year 1.2 1  cms  0.8 0.6  Peachland Creek Conservartion Target (MAD) Delivered, Activity-based est. Delivered - Extraction-based est.  0.4 0.2 0 10/31  12/1  1/1  2/1  3/4  4/4  5/5  6/5  7/6  8/6  9/6  136  E.6) Peachland Creek, CGCM2 2050s, Low Reservoir Release, Very Dry Year Peachland Creek, 2050s CGCM2, Low Reservoir Release "very dry" year 1.2 1  Peachland Creek Conservartion Target (MAD) Delivered, Activity-based est.  cms  0.8 0.6  Delivered - Extraction-based est.  0.4 0.2 0 10/1  11/1  12/1  1/1  2/1  3/1  4/1  5/1  6/1  7/1  8/1  9/1  E.7) Peachland Creek, CGCM2 2050s, High Reservoir Release, Very Dry Year Peachland Creek, 2050s CGCM2, High Reservoir Release "very dry" year 1.2 1  cms  0.8 0.6  Peachland Creek Conservartion Target (MAD) Delivered, Activity-based est. Delivered - Extraction-based est.  0.4 0.2 0 10/1  11/1  12/1  1/1  2/1  3/1  4/1  5/1  6/1  7/1  8/1  9/1  E.8) Peachland Creek, CGCM2 2050s, Low Reservoir Release, Very Wet Year Peachland Creek, 2050s CGCM2, Low Reservoir Release "very wet" year 1.2 1  cms  0.8 0.6  Peachland Creek Conservartion Target (MAD) Delivered, Activity-based est. Delivered - Extraction-based est.  0.4 0.2 0 10/1  11/1  12/1  1/1  2/1  3/1  4/1  5/1  6/1  7/1  8/1  9/1  137  E.9) Peachland Creek, CGCM2 2050s, High Reservoir Release, Very Wet Year Peachland Creek, 2050s CGCM2, High Reservoir Release "very wet" year 1.2 Peachland Creek Conservartion Target (MAD)  1  Delivered, Activity-based est.  cms  0.8 Delivered - Extraction-based est.  0.6 0.4 0.2 0 10/1  11/1  12/1  1/1  2/1  3/1  4/1  5/1  6/1  7/1  8/1  9/1  E.10) Peachland Creek, MPB + 2020s, Low Reservoir Release, Normal Year Peachland Creek ~ MPB + 2020s, Low reservoir release scenario "normal" year 1.2  Peachland Creek Conservartion Target (MAD) Delivered, Activity-based est.  1  Delivered - Extraction-based est.  cms  0.8 0.6 0.4 0.2  9/1  8/1  7/1  6/1  5/1  4/1  3/1  2/1  1/1  12/1  11/1  10/1  0  E.11) Peachland Creek, MPB + 2020s, Low Reservoir Release, Very Dry Year Peachland Creek ~ MPB + 2020s, Low reservoir release scenario "very dry" year 1.2 Peachland Creek Conservartion Target (MAD) Delivered, Activity-based est.  1  Delivered - Extraction-based est.  0.6 0.4 0.2  9/1  8/1  7/1  6/1  5/1  4/1  3/1  2/1  1/1  12/1  11/1  0 10/1  cms  0.8  138  E.12) Trepanier Creek, MPB + 2020s, Normal Year Trepanier Creek ~ MPB + 2020s scenario "normal" year 2.5 Trepanier Creek Conservation Target (MAD) Delivered, Activity-based est.  cms  2  Delivered, Extraction-based est.  1.5 1 0.5  9/1  8/1  7/1  6/1  5/1  4/1  3/1  2/1  1/1  12/1  11/1  10/1  0  E.13) Trepanier Creek, MPB + 2020s, Very Dry Year Trepanier Creek ~ MPB + 2020s scenario "very dry" year 2.5 Trepanier Creek Conservation Target (MAD) Delivered, Activity-based est. Delivered, Extraction-based est.  1.5 1 0.5  9/1  8/1  7/1  6/1  5/1  4/1  3/1  2/1  1/1  12/1  11/1  0 10/1  cms  2  139  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0071203/manifest

Comment

Related Items