Using Multi-Model Ensemble Methods to Assess Climate Change Impacts on Water Management Throughout the State of Washington, USA
Abstract
Reservoir systems throughout the western U.S. depend on mountain snowpack for springtime storage and refill. In the state of Washington, almost all reservoirs are small relative to their mean annual inflows, so that in almost all years they fill. However, they also can be quite susceptible to changes in the seasonality of inflows. Climate change is expected to result in earlier snowmelt runoff, and reduced summer flows, which may make water supply systems that rely on reservoir storage for metropolitan water supply and irrigation increasingly vulnerable to inability to meet their nominal demands. Seasonal and long-term water management planning typically uses reservoir models to assess how reservoir releases will respond to variations in streamflow. For instance, for municipal water supply, the goal is often 98% system reliability, or failures occurring on average in only one year in 50. For future planning, planning methods that use historical streamflows, which implies an underlying assumption of climate stationarity, may need to be modified to incorporate transient change and uncertainty in climate predictions. We use an approach that generates future streamflow simulations using regional hydrological models forced by statistically downscaled climate from IPCC A1B and B1 scenarios as simulated by 20 GCMs, for which output was archived as part of the IPCC 2007 report for the period 2000 to 2099. This ensemble of streamflow simulations provides information that can then be used to evaluate future reservoir system performance in a probabilistic manner. We explore how to best synthesize this information and use it within established reservoir system models, which are managed for diverse needs such as irrigation, municipal water supply, environmental flows, hydroelectric power, and recreation. To highlight advantages and methodological challenges, we provide an overview of this multi-model process for four unique water supply systems: three which serve municipal water supply needs in the Puget Sound Basin (Seattle, Tacoma, Everett) and one which serves irrigation needs in the Yakima River Basin. Based on our analyses, we provide insights into a framework that uses the ensemble method to provide metrics for water managers that (1) capture future trajectories and (2) adequately represent uncertainties in climate change impacts through time.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2008
- Bibcode:
- 2008AGUFMGC21B..05V
- Keywords:
-
- 1630 Impacts of global change (1225);
- 1637 Regional climate change;
- 1816 Estimation and forecasting;
- 1880 Water management (6334);
- 1884 Water supply