Streamflow Estimation from Hydrologic Model Assimilation of Remotely Sensed Snow Information in Snowmelt Dominated Basins
|Title||Streamflow Estimation from Hydrologic Model Assimilation of Remotely Sensed Snow Information in Snowmelt Dominated Basins|
|Publication Type||Conference Proceedings|
|Year of Conference||2004|
|Authors||Dressier, K, Leavesley, G, Bales, R, Fassnacht, S|
|Conference Name||American Geophysical Union, Fall Meeting 2004|
The USGS Precipitation Runoff Modeling System (PRMS) hydrological model was used to evaluate the utility of experimental, gridded, 1-km2 snow covered area (SCA) and snow water equivalent (SWE) products in improving the modeling of snowmelt runoff from three headwater basins in the Southwestern United Sates. The SCA product was the fraction of each 1-km2 pixel covered by snow and was derived from NOAA Advanced Very High Resolution Radiometer imagery. The SWE product was developed by combining the SCA product with SWE estimates interpolated from National Resources Conservation Service Snow Telemetry (SNOTEL) point measurements. An eight-year period (1995-2002) was used to compare PRMS simulated streamflow generated with and without the use of the SCA and SWE products. The test basins were the Upper Rio Grande (3,397 km2) in Colorado, and the White (1,634 km2) and Black (1,441 km2) which are tributaries to the Salt River in Arizona. In model runs using the SCA and SWE products, PRMS simulated SCA and SWE values were replaced with the SCA and SWE product values each time step the products were available. The simulated energy and mass balance states of PRMS were also adjusted based on the difference between the current model state and the assimilated estimate. The largest differences between PRMS simulations of SCA and SWE, and those estimated in the SCA and SWE products, occurred in the complex, higher elevation terrain. Simulated streamflow using the assimilated products were as much as 50% less than observed streamflow over the eight year period. The largest differences between observed streamflow and that simulated using the assimilated products occurred in the topographically complex Upper Rio Grande. Differences were smaller in the White and Black basins. Use of an averaging filter to smooth the SCA and SWE products prior to assimilation improved simulated streamflow volume, especially for the Upper Rio Grande basin.