ENHANCING SATELLITE-BASED PRECIPITATION PRODUCTS USING REMOTELY-SENSED SOIL MOISTURE RETRIEVALS AND DATA ASSIMILATION TECHNIQUE
Hydrology and Remote Sensing Laboratory
2011 Annual Report
1a.Objectives (from AD-416)
This proposal seeks to develop, test and implement a robust data assimilation system capable of correcting remotely-sensed precipitation inputs via the assimilation of remotely-sensed soil moisture retrievals into a land surface water balance model.
1b.Approach (from AD-416)
This project objective will be addressed in three phases:
1) Starting with a simple existing baseline system - based on a simple linear surface model modern and Kalman filtering – define and construct alternative systems based on more complex data assimilation and/or land surface modeling techniques. Potential data assimilation alternatives include the use of an Ensemble Kalman filter or Particle filtering techniques. Alternative land surface model techniques include the range of complex land surface models currently implemented within the NASA Land Information System.
2) Test the modified system using existing remote sensing datasets and evaluate the degree to which various systems are able to correct satellite-base precipitation to match aggregated rain gauge observations of short-term precipitation accumulations.
3) Prototype the identified system for the eventual availability of NASA Soil Moisture Active/Passive mission soil moisture measurements by running the system using L-band based soil moisture retrievals obtained from the European Space Agency (ESA) Soil Moisture/Ocean Salinity mission.
Completed the development of a new data assimilation to enhance satellite-based rainfall accumulations via the integration of remotely-sensed surface soil moisture retrievals into a land surface model and reported on research in a peer-reviewed manuscript (currently in review at Water Resources Research). The manuscript describes the adaptation of our original algorithm to include cases of more complex data assimilation methodologies and land surface modeling strategies. A number of these changes lead to substantial enhancements in performance. However, not all changes were associated with improved performance. In particular, changes which reflected the impact of more complex land surface modeling approaches were associated with little or no improvement in results. All changes that resulted in improved algorithm performance (i.e. better rainfall accumulation estimates when compared with rain gauge data) were collected and used to construct the Soil Moisture Analysis Rainfall Tool (SMART). Once defined, SMART was then successfully applied to the quasi-global (land area between 50 S and 50 N) correction of 3-day precipitation accumulation products derived from the TRMM Multi-Satellite Precipitation Analysis (TMPA) 3B40RT rainfall product.