|Zhan, Xiwu - NOAA NESDIS|
Submitted to: BARC Poster Day
Publication Type: Abstract Only
Publication Acceptance Date: April 1, 2007
Publication Date: May 1, 2007
Citation: Ryu, D., Crow, W.T., Zhan, X. 2007. Assimilation of coarse-scale satellite soil satellite soil moisture observations into a fine-scale hydrological model [abstract]. Abs. 9. BARC Poster Day. Technical Abstract: It has been demonstrated that passive microwave observations from space are capable of mapping surface soil moisture over the globe. Currently, the Advanced Microwave Scanning Radiometer (AMSR-E) onboard NASA’s Aqua satellite is providing moisture content in the top 1~2 cm of the soil column, and future L-band Soil Moisture and Ocean Salinity (SMOS) mission by ESA will map global coverage of moisture content at higher accuracies and greater depths. Soil moisture observations from space have been assimilated into basin- or continental-scale land surface models, aiming to improve the models’ predictabilities of surface states or fluxes such as soil moisture, evapotranspiration, and runoff. However, spatial scales of satellite observations are generally coarse (about 50~60 km) compared to those required for watershed-scale hydrologic applications (~1 km). This discrepancy in spatial scales existing between the supports of observation and model can be dealt with either by disaggregating satellite product or by up-scaling the model grid scale. In this work, we apply these two approaches to a synthetic assimilation experiment and compare their differences. The assimilation is performed using the ensemble Kalman filter (EnKF) in a medium-scale basin located in the United States Southern Great Plains. The Noah land surface model in the Land Information System (LIS) is run at 1-km grid scale, and the 1-km surface soil moisture contents are averaged over the basin in order to create a coarse-scale synthetic observation dataset. Up-scaling the model grid or down-scaling the satellite observation is implemented only in the update step of the EnKF. The data assimilation efficiency and benefits of the two approaches are compared against streamflow observations, and the impacts of the up-/down-scaling on the spatial heterogeneity of soil moisture and runoff predictions are discussed.