Use of Climate Forecasts in Hydrologic Prediction Applications

Use of Climate Forecasts in Hydrologic Prediction Applications Alan F. Hamlet Dennis P. Lettenmaier JISAO/CSES Climate Impacts Group Dept. of Civil and Environmental Engineering University of Washington

Weather Forecasts Flood Control and Hydropower Management Flood Forecasts 1 hr - 1 week Seasonal to Interannual Climate Forecasts Seasonal Streamflow Volumes Water Resources Management 1– 24 months Forecast Lead Time Climate Change Scenarios Long-Range Streamflow Forecasts Water Resources Planning 10 -100 years

Example of a Seasonal Forecasting System Based on Regression Models NRCS SNOTEL Network NRCS/NWRFC Water Supply Forecasts Hydrologic Index Regression Equation Streamflow Volume

Example of a Seasonal Forecasting System Using a Physically-Based Hydrologic Model Temperature and Precipitation Forecast Estimated Hydrologic State UW West-Wide Seasonal Hydrologic Forecast System VIC Hydrologic Model http: //www. hydro. washington. edu/forecast/westwide/ Hydrologic Forecast

Schematic Diagram of a Hydrologic Forecasting System Initial Hydrologic State • Soil Moisture • Snowpack Hydrologic Forecast: Hydrologic Model Future Temperature and Precipitation Forecast • Streamflow • Soil Moisture • Snowpack • Evaporation

Simulated Water Balance for the Pacific Northwest

Relative Roles of Future Precipitation and Initial Hydrologic State at Different Forecast Dates In April inputs to the water balance from future precipitation and storage are comparable. 46% 99% In October future precipitation dominates the inputs to the water balance. Simulated Long-Term Water Balance for the Pacific Northwest

Seasonal to Interannual Time Scales

Hydroclimate Signals in the Columbia River Basin Associated with ENSO and Pacific Decadal Variability Cool PDO Warm PDO Red = warm ENSO Green = neutral Blue = cool ENSO

Evaluation of Sept 1 ECMWF Nino 3. 4 Forecast for WY 2003 -2005 WY 2003 WY 2004 WY 2005

In the PNW Climate forecasts based on ENSO predictions can provide useful information about future streamflows with lead times up to 12 months. Retrospective tests in the Columbia River basin have shown that during cool or warm events, ENSO-based streamflow forecasts are superior to assumptions of “normal” conditions about 65 % of the time on Oct 1 Natural Streamflow (cfs) Natural Streamflow Columbia River at The Dalles, OR Red = Unconditional mean Blue = Ensemble mean Black = 2005 Observed

Can we predict decadal precip. variability? -0. 845 -0. 264 (Regional to PDO -0. 438 Correlation R 2 ) -0. 053 PNW Trend CRB Trend

7 -14 Day Time Scales

NOAA’s reforecast data set Model: T 62 L 28 NCEP GFS, circa 1998 Initial Conditions: NCEP-NCAR Reanalysis II plus 7 +/- bred modes. Duration: 15 days runs every day at 00 Z from 19781101 to now. (http: //www. cdc. noaa. gov/people/jeffrey. s. whitaker/refcst/week 2). Data: Selected fields (winds, hgt, temp on 5 press levels, precip, t 2 m, u 10 m, v 10 m, pwat, prmsl, rh 700, heating). NCEP/NCAR reanalysis verifying fields included (Web form to download at http: //www. cdc. noaa. gov/reforecast). Real-time probabilistic precipitation forecasts: http: //www. cdc. noaa. gov/reforecast/narr Hamill, T. M. , J. S. Whitaker, and X. Wei, 2004: Ensemble re-forecasting: improving medium-range forecast skill using retrospective forecasts. Mon. Wea. Rev. , 132, 14341447. http: //www. cdc. noaa. gov/people/tom. hamill/reforecast_mwr. pdf

Precipitation Skill Hamill, T. M. , J. S. Whitaker, and X. Wei, 2004: Ensemble re-forecasting: improving medium-range forecast skill using retrospective forecasts. Mon. Wea. Rev. , 132, 14341447. http: //www. cdc. noaa. gov/people/tom. hamill/reforecast_mwr. pdf

Compositing Approaches Using the Pacific North America Pattern Gridded Temperature Precipitation Data compositing Medium Range Forecast Model PNA Ensemble 7 -14 Day Temperature Precipitation Forecast

Ensemble PNA Forecast http: //www. cpc. ncep. noaa. gov/products/precip/CWlink/pna/new. pna_index_ensm. html

Implications of Climate Change

Observed 20 th century variability °C warming over the next 50 years. 2) Mitigation of greenhouse gas emissions will play an le in determining the scope of late 21 st century impacts. +3. 2°C +0. 7°C +1. 7°C 0. 9 -2. 4°C 0. 4 -1. 0°C Pacific Northwest 1. 2 -5. 5°C

Simulated Changes in Natural Runoff Timing in the Naches River Basin Associated with 2 C Warming • Increased winter flow • Earlier and reduced peak flows • Reduced summer flow volume • Reduced late summer low flow

Climate change assessments using scenarios show significant hydrologic changes due to temperature in basins with substantial snow accumulation in winter. 20 th Century Change + 2. 25 °C Naches River Basin on the East Slopes of the Cascades.

Observed Temperature Trends in the Western U. S. (1916 -2003) Tmax Canada PNW CA USA GB CRB Tmin

Trends in April 1 SWE 1950 -1997 Mote P. W. , Hamlet A. F. , Clark M. P. , Lettenmaier D. P. , 2005, Declining mountain snowpack in western North America, BAMS, 86 (1): 39 -49

Regionally Averaged Cool Season Precipitation Anomalies PRECIP

Climate Models vs. Statistical Approaches

Dynamic Approaches Based on Climate Models Potential Advantages: Ideally accounts for extraordinary changes in the climate system in ways that statistical procedures may ignore or discount (e. g. loss of sea ice in the arctic, altered radiative forcing due to current greenhouse forcing, effects due to soil moisture anomalies, etc. ) Reforecast approaches suggest useful precipitation skill in some areas at 6 -10 days, but linkage to hydrologic models is needed for test purposes. Limitations: Model bias must almost always be removed (so not a purely dynamic system) With some few exceptions, there has been very limited demonstrated precipitation skill in the U. S. at seasonal to interannual time scales in real-time forecast applications. (Role of downscaling procedures? ) Relatively expensive to implement.

Statistical Approaches Potential Advantages: Exploits strengths of models (e. g. large scale index forecasts) while avoiding problems (precipitation simulation). Relatively low implementation costs. Relatively high skill. Limitations: Cannot account easily for unique features of individual years (e. g. different “flavors” of ENSO ) Requires a training data set—problems with parameter stationarity.

Conclusions Useful hydrologic forecasts based on weather or climate forecasts are available with lead times ranging from a few hours (flood forecasts) to 50 years or more (climate change scenarios). Many operational hydrologic forecasting systems are currently based on statistical models, however dynamic, physicallybased tools are increasingly being used in both academic and operational forecasting systems. Dynamic forecasting systems based on weather or climate models directly linked to physically-based hydrologic models have important advantages in a rapidly evolving climate system. Short-term forecasts based on weather models have already reached a useful state of development, but many challenges remain at seasonal or longer time scales.

Extras

Example of a Short Time Scale Flood Forecasting System MM 5 mesoscale atmospheric model Estimated Hydrologic State DHSVM distributed hydrologic model Streamflow Forecast River Stage Forecast

Background: Forecast System Schematic local scale (1/8 degree) weather inputs Hydrologic model spin up NCDC met. station obs. up to 2 -4 months from current 1 -2 years back LDAS/other real-time met. forcings for spin-up gap soil moisture snowpack INITIAL STATE streamflow, soil moisture, snow water equivalent, runoff Hydrologic forecast simulation ensemble forecasts SNOTEL Update 25 th Day, Month 0 ESP traces (40) CPC-based outlook (13) NCEP GSM ensemble (20) NSIPP-1 ensemble (9) Month 6 - 12