P4.4
Near real time snow data assimilation for streamflow forecasting using MODIS snow data products
Qiuhong Tang, University of Washington, Seattle, WA; and A. W. Wood and D. P. Lettenmaier
Snow plays a major role in streamflow generation in the western United States, hence estimates of snow water storage during winter are a key predictor of water availability in the spring and summer. In an effort to improve the estimation of snow cover, snow water storage, and consequently forecasts of seasonal streamflow, we have implemented an experimental approach for using MODIS snow cover imagery to reduce errors in snow states simulated by the Variable Infiltration Capacity (VIC) macroscale hydrology model which is used for ensemble streamflow prediction (ESP) in our Westwide Seasonal Forecast System. We have implemented the approach in both retrospective and real-time contexts. Our approach involves a composite insertion of MODIS/Terra Snow Cover Data into the VIC model. The assimilation of satellite snow cover imagery for streamflow forecast improvement requires a snow cover depletion curve which relates snow cover fraction to snow water equivalent (SWE). This relationship is formed by using NRCS SNOTEL point observations of SWE concurrently with MODIS snow cover data. The derived estimates of SWE (based on MODIS snow covered area) are then assimilated into the VIC model. For two study basins, the upper Klamath River of Oregon and the Feather River of California, our findings to date show mixed results for streamflow prediction.
Supplementary URL: http://www.hydro.washington.edu/forecast/rsda/
Poster Session 4, Advances in Remote Sensing and Data Assimilation in Hydrology
Thursday, 24 January 2008, 9:45 AM-11:00 AM, Exhibit Hall B
Previous paper Next paper