Recently, more biophysical parameters (e.g., vegetation cover, leaf area index; LAI) retrieved from the measurement of Advanced Very High Resolution Radiometer (AVHHR) on board of the polar orbiting NOAA satellites are available from 1982 to recent years. In spite of the same source of measurements, however, the retrieved data can be largely different to each other according to its retrieval algorithms and data processing methodology. It is important to evaluate these vegetation properties for climate studies, because it is generally agreed that presence and/or change of vegetation play significant roles in modulating climate variability. Therefore, in preparation for the assessment of LAI datasets using a GCM, we analyzed spatial distribution and temporal behavior of the two available monthly-LAI datasets for 17 years (1982-1998): one is FASIR obtained from the International Satellite Land Surface Climatology Project (ISLSCP) Initiative II, and the other one is derived from Global Inventory Monitoring and Modeling System (GIMMS), which is also the ISLSCP holding.

In the sense of 17-year climatology, the most significant differences between LAI/FASIR and LAI/GIMMS were shown in tropical regions of Amazon basin and central Africa not only in the magnitude but also the interannual variability. In boreal summer season, there also exist large differences in mid- and high latitudinal North American and Eurasian continents. The seasonal and interannual variations of the data sets were examined using EOF analysis for different regions to find anomalous variations. The singular value decomposition (SVD) analysis was applied to examine the precipitation-LAI coupled variability. The most-dominant coupled mode explains about 30% of the total variability and correlation coefficients between principle components of precipitation and LAI are greater than 0.6 over Africa and South America. But, it was hard to recognize significant differences of coupled variability between LAI/FASIR and LAI/GIMMS datasets. Although it is not easy to identify merits and limitations of the currently available LAI dataset due to the lack of ground measurements, it is necessary for the climate simulation to take account for interannual variability of vegetation intensity, which can be provided by only satellite-based LAI datasets. In order to investigate feasibility to use these datasets as a surface boundary forcing for climate simulation, we are conducting 17-year climate simulation with the NCEP GSM coupled with the SSiB. In the preliminary results, two LAI datasets induced significant differences in the simulated climate especially for the monsoon evolution over some areas, such as East Asia. Further detail analysis of LAI datasets and simulation results will be discussed in the presentation.