ASSIMILATION OF CYCLOPES LEAF AREA INDEX INTO THE ISBA-A-GS LAND SURFACE MODEL OVER FRANCE

摘要

In our study the SURFEX modeling platform is used inrnoff-line mode. It contains the land surface model ISBAA-rngs that simulates photosynthesis and plant growth. Arnversion of the Extended Kalman Filter scheme is used forrnthe assimilation in situ and CYCLOPES LAI product derivedrnfrom the SPOT4/VEGETATION sensor within thernISBA-A-gs model. The LAI data is incorporated into thernmodel every 10 days to analyse its impact on vegetationrnbiomass. The experiment is performed over France at arnspatial resolution of 8 km. A model grid box is divided inrna number of patches each having its own set of prognosticrnvariables. The filter algorithm is designed to providernthe analysis for each patches independently by using onernobservation per grid box. When needed, the updated valuesrnare aggregated by computing a weighted average. Inrnthis study we demonstrate that the assimilation schemernworks effectively within the multi-patch version of ourrnmodel. An important reduction of the LAI bias betweenrnthe model and measurements is achieved.