Incorporating agricultural management into an earth system model for the Pacific Northwest region Kiran J. Chinnayakanahalli, Jennifer Adam, Roger Nelson, Claudio Stöckle, Michael Brady, Kirti Rajagopalan, Michael Barber, Shifa T. Dinesh, Keyvan Malek, Georgine Yorgey, Chad Kruger
Outline Background Modeling Approach Data Sources Results Conclusions and Relevance to Earth system modeling
Background
Background: WSU Water Demand Forecasting Every 5 years, The Washington State Department of Ecology’s office of the Columbia River (OCR) is required to submit a long-term water supply and demand forecast to the Legislature The forecast helps improving understanding of where additional water supply is most critically needed, now and in the future Washington State University (WSU) was assigned to develop the 2030s forecast for water supply and out-of-stream demand
Modeling Integration: *Surface Hydrology *Cropping Systems *Water Management *Economics Modeling Approach
Modeling Framework
Economic Modeling
Hydrology and Crop System Models VIC Hydrology Liang et al, 1994 and Elsner et al, 2010 CropSyst Cropping Systems Stockle and Nelson 1994
VIC-CropSyst Model Integration 1. Weather (D) 2. Soil Soil layer depths Soil water content 3. Water flux (D) Infiltrated water 4. Crop type Irrigation water = Crop Water Demand /irrigation efficiency Sow date Crop interception capacity Crop phenology Crop uptake (D) Water stress (D) Current biomass (D) Crop Water demand (D) Harvest day Crop Yield VICCropSyst D – communicated daily
Data Sources
Historical and Future Climate Data
Crop Land Data Layers WSDA – Washington State Department of Agriculture USDA – US Department of Agriculture Non-crop land cover from Elsner et al. 2010
Crops Modeled Winter Wheat Spring Wheat Alfalfa Barley Potato Corn Corn, Sweet Pasture Apple Cherry Lentil Mint Hops Grape, Juice Grape, Wine Pea, Green Pea, Dry Sugarbeet Canola Onions Asparagus Carrots Squash Garlic Spinach Generic Vegetables Grape, Juice Grass hay Bluegrass Hay Rye grass Oats Bean, green Rye Barley Bean, dry Bean, green Other Pastures Lentil/Wheat type Caneberry Blueberry Cranberry Pear Peaches Berries Other Tree fruits Major Crops
Results
Snake River and Columbia River Supplies 2030s Snake River Columbia River
Regulated Supply vs Demand for Columbia River Basin (at Bonneville) 2030 results are for - HADCM_B1 climate scenario - average economic growth and trade Note: Supply is reported prior to accounting for demands Demand: 13.3 million ac-ft Demand: 13.6 million ac-ft 2030s
Regulated Supply and In-Stream Flow Requirements at Key Locations Future (2030s) Historical ( ) Note: Supply is reported prior to accounting for demands
Yakima
Yakima Regulated Supply and Demand Historical Hadcm_B1 (2030s)
Conclusions and Relevance to Earth System Modeling
Relevance to Earth System Modeling VIC-CropSyst will be used in a new Earth system model, BioEarth Overarching Goal: To improve the understanding of regional and decadal- scale C:N:H 2 O interactions in context of global change to better inform decision makers involved in natural and agricultural resource management.
Major Findings A small increase of around 3.0 (±1.2)% in average annual supplies by 2030 compared to historical ( ) Unregulated surface water supply at Bonneville will The irrigation demand under 2030s climate was roughly 2.5% above modeled historic levels under average flow conditions 14.3 (±1.2)% between June and October 17.5 (±1.9)% between November and May
Conclusions Increased irrigation demand, coupled with decreased seasonal supply poses difficult water resources management questions, especially in the context of competing in stream and out of stream users of water supply Some watersheds more impacted than others This new modeling framework to study the coupled dynamics between water supply and water demand in an altered climate will advance our ability to integrate land and water resource decision making into regional-scale Earth system models
Related work Assessing the impact of climate change on Columbia River Basin agriculture through integrated crop systems, hydrologic, and water management modeling Kirti Rajagopalan et al. Tuesday afternoon- Poster
Thank you! Acknowledgements: University of Washington Climate Impacts Group (in particular, Marketa Elsner, Alan Hamlet, and Pablo Carrasco) Peer Reviewers: Alan Hamlet Bob Mahler Ari Michelsen Jeff Peterson
ColSim Reservoir Model (Hamlet et al., 1999) for Columbia Mainstem Model used as is, except for Withdrawals being based on VIC-CropSyst results Curtailment decision is made part of the reservoir model Green triangles show the dam locations
Curtailment Rules (Washington State) Curtailment based on instream flow targets Columbia Mainstem Lower Snake Central Region (Methow, Okanogan, Wenatchee) Eastern Region (Walla Walla, Little Spokane, Colville) Prorated based on a calculation of Total Water Supply Available Yakima
Yakima Reservoir Model Irrigation demand from VIC/CropSyst Curtailment rules Proratable water rights prorated according to Total Water Supply Available (TWSA) calculated each month Monthly Inflows from VIC-CropSyst Total System of Reservoirs (capacity 1MAF approx.) Objectives : Reservoir refill by June 1 st Flood space availability Instream flow targets Gauge at Parker
Model Calibration/Evaluation Calibration: Streamflows (we used calibration from Elsner et al and Maurer et al. 2002) Crop Yields (using USDA NASS values) Irrigation Rules (using reported irrigated extent by watershed) Evaluation: Streamflows (Elsner et al and Maurer et al. 2002) USBR Diversions from Bank’s Lake (for Columbia Basin Project)
Model Scenarios: Low, Middle, High Climate Change Scenarios HADCM_B1, CCSM_B1, CGCM_B1, PCM_A1B, IPSL_A1B Hybrid Delta Downscaling Approach (2030s climate) (UW CIG) GCMs and Emission Scenarios chosen for low/middle/high precipitation and temperature change combinations Water Management Scenarios Additional Storage Capacity Cost Recovery for Newly Developed Water Supply Economic Scenarios International Trade Economic Growth
T – Transpiration I P – Interception capacity I – Infiltration Ir – irrigation Wd- Water demand Q – Runoff Q 01 – Drainage from 0 to 1 Q 02 – Drainage from 0 to 2 Q b – Baseflow W 0 – water content in 0 W 1 – water content in 1 W 2 - water content in 2 Tmin, Tmax – daily minimum and maximum temperature Ws – wind speed RH – Relative humidity SR – Solar radiation QbQb Q 12 T IPIP Redistribute I, W 0, W 1 and W 2 to CropSyst layers Q Q 01 W 0,W 1, W 2 T 0, T 1, T 2, I P, Wd I CropSyst VIC Ir Daily Tmin, Tmax, Ws, RH, SR, I VIC-CropSyst : Coupling Approach
Invoking CropSyst within VIC gridcell Crop 1 VIC grid cell (resolution=1/16° ~ 33 km 2 ) Crop 2 Non-Crop Vegetation CropSyst is invoked Within the cell, VIC does not identify the geographical locations of the crops
Crop type, Soil texture VIC-CropSyst integration - Time view Time 0 At the beginning of each daily time step To CropSyst Weather condition Irrigation- add 20mm if the need ≥ 20 mm Start looking for sow day Crop harvest date Total yield, Biomass etc Crop 1 At the end of each daily time step From CropSyst Irrigation need Transpiration VIC grid cell Crop emerges
Conclusions A biophysical modeling framework was developed and implemented to study the coupled dynamics between water supply and water demand in an altered climate The irrigation demand under 2030s climate was roughly 2.5% above modeled historic levels under average flow conditions
Conclusions Stress due to irrigation demand will be enhanced in summer months due to the shift in the hydrographs (early snow melt) Some crops like corn show decrease in yield while others like alfalfa show an increase for 2030’s climate Bio-earth ?
Alfalfa, average historical yield in tonnes/hectare under full irrigation
Alfalfa, average historical irrigation water demand in mm