Submitted to: IEEE Transactions on Geoscience and Remote Sensing
Publication Type: Peer reviewed journal
Publication Acceptance Date: 11/15/2005
Publication Date: 6/30/2006
Citation: Zhan, X., Houser, P.R., Walker, J.P., Crow, W.T. 2006. A method of retrieving high resolution soil moisture from Hydros L-band radiometer and radar observations. IEEE Transactions on Geoscience and Remote Sensing. 44(6):1534-1545. Interpretive Summary: Hydros is an exploratory NASA satellite mission that seeks to deploy the first spaceborne sensor optimally designed to globally measure surface soil moisture. Data products from Hydros have the potential to aid a range of climatic and hydrologic applications (e.g. seasonal weather forecasting, drought monitoring, and flood forecasting) relevant to agricultural management issues. This paper describes results for a numerical simulation of a novel techniques for combining multi-scale observations made by the Hydros sensor. Results will be used to help refine retrieval algorithms and improve the accuracy (and spatial resolution) of Hydros soil moisture products. The Hydros mission is currently in formulation stage within NASA's Earth System Science Pathfinder program and has an expected launch data sometime in 2010/2011. USDA ARS scientists have been involved in the project since its inception and are actively working to develop agricultural applications for Hydros data products.
Technical Abstract: NASA's Earth System Science Pathfinder Hydrospheric States (Hydros) mission will provide the first global scale space borne observations of Earth's soil moisture using both L-band microwave radiometer and radar technologies. In preparation for the Hydros mission, an Observation System Simulation Experiment (OSSE) has been conducted. This involved generating a geophysical fields data set with a land surface model, computing surface microwave emissions and backscatters from the geophysical fields with a microwave emission and backscatter model, simulating the radiometer brightness temperature and radar backscatter observations with the instrument errors, and finally testing various soil moisture retrieval algorithms with the simulated observations. As a part of this OSSE, we explored the potential for retrieving useful surface soil moisture at a spatial resolution of 9km by optimally merging relatively accurate 36km radiometer brightness temperature and noisy 3km radar backscatter cross-section observations using the Extended Kalman Filter (EKF ). This study examined four alternative methods for implementation of the EKF retrieval algorithm and their respective accuracy assessment; one-dimensional fine resolution, one-dimensional medium resolution, two-dimensional fine resolution and two-dimensional medium resolution. Based on the Hydros OSSE data sets, the two-dimensional fine resolution retrieval method performed the best out of the four methods. The root-mean-square error of 9km soil moisture retrievals of the entire OSSE domain across 34 consecutive retrievals was found to be 3.0 % vol/vol, as compared to 3.6 to 4.9 % vol/vol for the other three EKF retrieval methods respectively. This is to be compared with 5.5 % vol /vol and 4.9 % vol/vol for direct iterative inversion of 9km radar data and 36km radiometer data respectively.