Impacts of the satellite-derived leaf area index on GCM simulation of near-surface climate

The leaf area index (LAI), as define as leaf surface area per unit ground area, is an important parameter to determine radiative transfer and partitioning surface available energy on the vegetated surface of biophysical land surface models. Recent efforts to produce LAI estimates from AVHRR measurements encourage us to assess two satellite-based LAI datasets, which are called as FASIR and GIMMSBU datasets, in order to investigate the role of vegetation in climate variations. Three sets of numerical experiments using the NCEP General Circulation Model (GCM) coupled with the Simplified Simple Biosphere Model (SSiB) were conducted for this purpose, particularly, by focusing on interannual variability. As a surface boundary condition, prescribed LAI values from the look-up table which have no interannual variation was used in control experiment. The other two experiments employed the LAI values provided by FASIR and GIMMSBU datasets, respectively. Each experiment consists of 20-month long simulation of the NCEP GCM/SSiB from January to August of next year for 6 ENSO years (three El-nino and three La-Nina years) during 17 years from 1982 to 1998.

FASIR and GIMMSBU LAI datasets show qualitative agreements to each other, but contain substantial discrepancies in terms of magnitude and amplitude of interannual variability especially in high-latitudes of northern hemisphere and East Asia. From the results of three experiments, the authors analyzed interannual variability defined as difference between warm and cold ENSO years. Both the FASIR and GIMMS LAI datasets played a role in reducing biases of surface air temperature and precipitation simulated by control experiment during summer in terms of horizontal distribution. For the surface air temperature, the ENSO signals in mid- and high latitudes of North America and Eurasia continents which were not captured in control experiment were reproduced apparently in FASIR and GIMMS experiments. In the meanwhile, the most sensitive region of the simulated precipitation to satellite-LAI forcings was the East Asian summer monsoon area. FASIR experiments produced slightly better results than GIMMS. Further analysis to find the mechanism of the land-atmosphere interaction associated with the vegetation forcing in improvements of the interannual variability in simulating surface air temperature and precipitation will be given in the presentation.