4 — original pdf

Backup

Update on WF24 Climate and Hydrology Analysis September 20, 2022 Planning for Uncertainty ▪ Develop range of futures ▪ Find common near-term strategies that work for a broad range of futures ▪ Develop adaptive plan with key decision points ▪ Re-evaluate at key decision points Range of demands Possible climate futures DWDRs Regional supply trends Uncertainty in water availability Water Forward 2024 Decision points A C D B E 2125 Goals of Climate & Hydrology Analysis Update  Look at a range of possible future climate scenarios  Identify high-level climate trends in the basin  Generate climate change- adjusted streamflow data to test in the Water Forward Water Availability Model (WF WAM) Differences from 2018 WF Plan  Partnership with UT Austin  Climate technical advisory group scenarios  Looking at multiple climate  New hydrologic models Climate and Hydrology Analysis Update – Tasks 2022 2023 Task 1: Project management and external communication (WFTF, climate TAG, etc.) Task 2: Select GCMs representative of the region to use for update Task 3: Perform GCM downscaling and trend analysis Task 4A/B: Develop hydrologic models to predict streamflow from downscaled GCM outputs Task 4C: Generate time series of naturalized flows Task 5: Package flow data for use in the WF Water Availability Model Task 6: Develop stochastic drought sequences using historical and climate-adjusted hydrology Task 7: Continue coordination with WF update process (communication, presentations, reports, etc.) Through 2024 We are here Selection of GCMs  What: choose global climate models (GCMs) that best represent climate over the Colorado River Basin  Why: want to use GCMs that can best project possible climate futures for the Colorado River Basin  How: evaluate how well GCMs simulate historical climate over the Colorado River Basin and select the best performing set of models Evaluation of GCMs  Historical simulations of 35 global climate models (GCMs) from the Coupled Model Intercomparison Project Phase 6 (CMIP6) are evaluated on their ability to represent the following observed characteristics: Top-scoring GCMs Top 10 best-scoring GCMs based on model performance over the Colorado River Basin (CRB), as measured by skills scores (S) Model 𝑺𝒔𝒑𝒂𝒕𝒊𝒂𝒍,𝑻 𝑺𝒔𝒑𝒂𝒕𝒊𝒂𝒍,𝑷 𝑺𝒔𝒑𝒂𝒕𝒊𝒂𝒍,𝑵𝑫𝑫 𝑺𝒕𝒆𝒎𝒑𝒐𝒓𝒂𝒍,𝑻 𝑺𝒕𝒆𝒎𝒑𝒐𝒓𝒂𝒍,𝑷 𝑺𝒕𝒆𝒎𝒑𝒐𝒓𝒂𝒍,𝑵𝑫𝑫 𝑺𝒐𝒗𝒆𝒓𝒂𝒍𝒍 Ranking CNRM-CM6-1-HR HadGEM3-GC31-MM UKESM1-0-LL HadGEM3-GC31-LL CNRM-CM6-1 CNRM-ESM2-1 KACE-1-0-G GFDL-ESM4 ACCESS-CM2 EC-Earth3 0.90 0.93 0.92 0.91 0.90 0.91 0.91 0.91 0.89 0.91 0.94 1.00 0.88 0.91 0.79 0.82 0.94 0.78 0.95 0.95 0.97 0.96 0.70 0.70 0.72 0.75 0.79 0.81 0.77 0.95 0.98 0.94 0.94 0.95 0.96 0.95 0.94 0.94 0.95 0.92 0.63 0.21 0.75 0.54 0.72 0.65 0.08 0.54 0.31 0.13 0.55 0.68 0.68 0.75 0.70 0.63 0.72 0.49 0.33 0.04 0.864 0.845 0.818 0.809 0.801 0.798 0.779 0.775 0.758 0.747 1 2 3 4 5 6 7 8 9 10 Selected GCMs Climate Scenario # CMIP6 SSP emission scenarios Average End-of- Century Warming for CRB (°F) Selected GCMs 1 2 3 1-2.6 2-4.5 5-8.5 4.4 6.6 11.6 ACCESS-CM2 CNRM-CM6-1 EC-Earth3 KACE-1-0-G UKESM1-0-LL • Different Shared Socioeconomic Pathways (SSPs) for different greenhouse gas emission scenarios according to different climate policies • Number of GCMs and emission scenarios selected covers a wide range of possible futures • Wide range of possibilities will support robust decision-making approach (image source: IPCC, 2021) Downscaling and bias correction of GCM data  What: downscale and bias correct data outputs from the global climate models into data that can be used over the Colorado River Basin  Why: improves data resolution in the area of interest and removes biases in GCM data to allow for more in-depth analysis  How: use statistical downscaling over the Colorado River Basin and bias correct downscaled GCM data based on statistical relationship with observed data Downscaling process Effectively bias-corrects modelled evaporation to match observation (TWDB monthly total lake evaporation) Outputs of Global Climate Model (GCM) 27 1-degree grids bordering the Colorado River Basin Regrid Bilinear Interpolation Nearest Neighbor Interpolation Daily Maximum and Minimum Temperature Daily Precipitation Hargreaves Evapotranspiration equation Monthly Total Lake Evaporation (Inch/Month) Trend analysis of downscaled GCM data  What: examine GCM projections of future climate and identify relevant trends in the data  Why: trends in the data help us determine if temperature is generally increasing across all scenarios, if rainfall is generally decreasing, etc.  How: Compare GCM future projections to historical data and identify differences Trend analysis process  Trend analysis performed on the variables below calculated using bias-corrected daily temperature and precipitation : • Annual average precipitation and temperature • Number of days in a year with precipitation below 0.01 inch (dry days) days) 90°F (hot days) 100°F (hot days) (cold days/nights) • Number of days in a year with precipitation above 2 inches (wet • Number of days in a year with maximum temperature above • Number of days in a year with maximum temperature above • Number of days in a year with minimum temperature below 32°F • Annual maximum 5-day total precipitation  Trend analysis over Austin and CRB 709 710 809 810 93W107W37N25NTemperature • Annual mean temperature is projected to increase • Number of hot days with temperatures above 100°F are projected to increase • Rainfall distribution is projected to • Less frequent and more intense rainfall events are projected Rainfall change Dry Days • Number of dry days with precipitation below 0.01” are projected to increase Projected high-level climate trends in the basin Based on initial results Projection of streamflows using GCM data  What: use precipitation and temperature data across the Colorado River Basin to develop streamflow projections  Why: streamflow projections will be used in the water availability model to evaluate portfolios of strategies across future time horizons  How: considered 2 modeling methods: runoff projections from GCM datasets and multivariate model similar to the method used in WF18 Colorado River Basin  31,000 square miles of drainage area  45 control points for flow inputs to the WAM  27 weather locations (quadrangles) Austin Precipitation and Temperature Features  Features created using measures of precipitation and temperature time series from nearby quadrangles Pedernales River near Johnson City Precipitation Exp. Avg., n = 6 Colorado River at Austin Precipitation Exp. Avg., n = 4 Maximum Temperature Arith. Avg., 17 months Minimum Temperature Arith. Avg., 11 months Dry Days Exp. Avg., n = 11 Dry Days Exp. Avg., n = 24 Precipitation > 0.25” per day Arith. Avg., 13 months Dry Days Current month only  10 unique features selected for Dry Days Exp. Avg., n = 2 Dry Days Arith. Avg., 4 months each flow control point  Features selected for relevancy to flow at each location, and minimum redundancy between features Precipitation-Evaporation Arith. Avg., 8 months Precipitation > 2” per day Current month only Hottest 7 days per month Arith. Avg., 13 months Precipitation-Evaporation Arith. Avg., 12 months Max of daily Precip.-Evap. Exp. Avg., n = 8 Precipitation > 1” per day Current month only Precip.-Evap. > 2” per day Exp. Avg., n = 2 Precip.-Evap. > 2” per day Arith. Avg., 3 months Precip.-Evap. > 1” per day Arith. Avg., 2 months Precipitation Arith. Avg., 7 months Flow Model Framework, Training, and Projections  Neural network flow models  Flow models connect to downstream control point models Precipitation and Temperature Features using Historical Data Precipitation and Temperature Features using GCM Projections Flow Model Training Process Evaluate Outputs Versus Historical Flows Established Flow Model Projections of Future Flows ΣΣUpstream Flow ModelsFlowsWeather FeaturesFlow Projection Overview  Projections show greater frequency of lower flows as the climate warms over time.  Extreme high flow events can increase in magnitude. 2041-2060 2081-2100 Climate & Hydrology Analysis Next Steps  Create ensembles of streamflow data for testing in the WF WAM  Develop stochastic streamflow series for both historical and climate-adjusted data  Test water management strategies against all possible streamflow series developed • Determine which strategies perform best over the most scenarios Questions?