Start Date/Time: Monday, November 04, 2013, 2:30 PM
Ending Date/Time: Monday, November 04, 2013, 3:20 PM
Location: More Hall, Rm 230 (http://uw.edu/maps/?mor)
Roy Rasmussen, Director, Hydrometeorological Applications Program, NCAR Research Applications Laboratory and Martyn Clark, Scientist, Hydrometeorological Applications Program, NCAR Research Applications Laboratory
The NCAR Water System Program has conducted research related to the Regional and Global water cycle since 2001. A key focus of the Water System Program, in collaboration with the US Bureau of Reclamation and the US Army Corps of Engineers, is to understand how changes in climate affect the seasonal evolution of the snowpack, soil moisture, and overall partitioning of precipitation into evapotranspiration and runoff, and to evaluate how well these linkages are expressed in climate change assessment methods that are being used in practice to support water planning studies. Specifically, the program uses a mix of downscaling techniques and hydrologic models to understand how the portrayal of climate impacts on watersheds depends on methodological choices.
For instance, dynamical downscaling using the Weather Research and Forecasting (WRF) model at 4-km grid resolution for the Colorado Headwaters region indicates different climate sensitivities than current statistically-based guidance being provided to water managers. WRF shows wintertime increases in precipitation in the Colorado Headwaters that are consistent with a warmer and moister atmosphere, and occur when topography is adequately resolved by the regional climate model. The statistical downscaling methods are found to have limitations that impact hydrologic responses - e.g., yielding too much drizzle, insufficient extreme events, and improper representation of spatial and temporal scaling characteristics. The choices made in hydrologic modeling also affect projection outcomes, especially those related to selection of hydrologic model structure and methods used to estimate model parameters.
Taken together, the research findings thus far reveal that the current practice of impact assessment unwittingly includes an array of unintended effects - artifacts resulting from the method, data and model choices - that should prompt practitioners to seek a new path. The NCAR-Reclamation-USACE collaboration is therefore working to identify alternatives. On downscaling, for instance, the team is developing advanced hybrid statistical-dynamical downscaling methods to provide a realistic depiction of physical processes at a low computational cost. These methods improve hydrologically relevant metrics such as the spatial representation of extreme precipitation events, and can be applied to a large range of climate scenarios. On hydrologic assessment, the team is developing multi-model, multi-physics and multi-parameter approaches, with advanced multi-objective calibration strategies, to reduce simulation errors and improve characterization of uncertainty in hydrologic models. The improved hydrologic models and more physically realistic downscaling implementation will lead to more dependable projections, and ultimately to improved support for planning and decision-making.