Events / Media

Data Assimilation Seminar Series

Ms. Chang Yaping and Dr. Shunji Kotsuki (Mar. 2, 2016, 15:30-)

Affiliation Chinese Academy of Science and RIKEN-AICS
Title Ensemble data assimilation of MODIS surface temperature into land surface model
Abstract Land surface temperature (LST) is an important variable for energy and water balances in land surface models. In this study, the LST measured by Moderate Resolution Imaging Spectroradiometer (MODIS) is assimilated into the Simple Biosphere Urban Canopy (SiBUC) model using the local ensemble transform Kalman filter. Ensemble data assimilation for land surface models needs perturbed forcing or observation data (PF or PO methods) to maintain the ensemble spread. This study proposes a new PF method with a parameter estimation algorithm in the data assimilation framework. Unlike the existing methods, the new scheme estimates flow-dependent amplitudes of the forcing perturbations for each ensemble member. We perform data assimilation experiments at two sites from the Asia-Flux project: the Habei Grassland and Yucheng Sites. Using the flux observations as the ground truth, we validate whether LST and other land surface variables such as sensible and latent heat fluxes are improved by assimilating LST from MODIS. First, we investigate the sensitivity of the SiBUC LST simulation to the forcing data. The results indicate that the forcing data can cause large errors in LST. Next, we perform a data assimilation experiment without the PF scheme and compare with the experiment without data assimilation. The results show that assimilating the MODIS LST observations improved the LST estimates remarkably in winter. For further improvement, we are now running another experiment with the new PF scheme, and the results will be presented at the seminar.
PDF Link to PDF file