Title: Use of inverse modeling tools for improved utilization of earth science data in land surface modeling and data assimilation
Primary Author: Kumar, Sujay
Organization:NASA GSFC
Co-Author(s): Christa Peters-Lidard, Ken Harrison, Soni Yatheendradas, Joe Santanello

The NASA Land Information System (LIS; http://lis.gsfc.nasa.gov; Kumar et al. 2006, Peters-Lidard et al. 2007) is a high-resolution, high-performance, land surface modeling and data assimilation system to support a wide range of land surface research and applications. LIS integrates various community land surface models, ground and satellite-based observations, and ensemble-based data assimilation tools to enable assessment and prediction of hydrologic conditions at various spatial and temporal scales of interest. More recently, the LIS infrastructure has been enhanced with the development of a suite of inverse modeling tools (optimization and uncertainty estimation subsystems) that help in the improved use of Earth science data for hydrologic modeling. The optimization infrastructure in LIS includes a suite of both deterministic and stochastic algorithms and provides a new capability to estimate and improve global soil or vegetation parameters from satellite data. The uncertainty estimation algorithms, which include Bayesian approaches such as Markov Chain Monte Carlo (MCMC), acknowledge the uncertainty in modeling land surface processes and use the observational information to enable probabilistic predictions. These important capabilities represent significant advancements over current practice and further exploit the information content of the Earth Science data. In this presentation, the use of these inverse modeling tools will be demonstrated through a number of case studies for the estimation of soil moisture. Through the use of remotely sensed observational information, the inverse modeling tools are used to estimate land surface parameters and their associated uncertainties. Results from simulations that translate these input uncertainties into uncertainty estimates associated with soil moisture predictions will also be presented.