Evaluating land surface parameters in the Biosphere‐Atmosphere Transfer Scheme using remotely sensed data sets

Remotely sensed data sets were used to evaluate the spatial/temporal specification of albedo (α), roughness length ( Z 0 ), and green leaf area index ( L g ) in the Biosphere‐Atmosphere Transfer Scheme (BATS). These parameters were evaluated for 1988 and 1993 in the Upper Mississippi River Basin (UMRB) using data from the International Satellite Land Surface Climatology Project Initiative 1 (ISLSCP 1) and by applying the ISLSCP 1 algorithms to high spatial resolution satellite data (Pathfinder advanced very high resolution radiometer (AVHRR) land (PAL)). ISLSCP 1 L g was more variable spatially than BATS, particularly in midsummer months. ISLSCP 1 α was generally lower than BATS α. BATS and ISLSCP 1 generally showed small spatial variability in α for a given cover type. ISLSCP 1 Z 0 were up to 100% larger than BATS. Spatial/temporal variability in ISLSCP 1 and BATS Z 0 was primarily restricted to the spring. Comparisons between the ISLSCP 1 and PAL data indicated substantial differences in estimates of L g and Z 0 because of differences in the transformation of the vegetation index to L g and grid cell heterogeneity. PAL L g and Z 0 were extremely variable spatially. Large differences between years in the PAL L g were found, particularly in climate‐affected regions. The remotely sensed data indicate that improved parameterizations in BATS are needed to account for spatial variability in L g and Z 0 , as well as interannual variability in L g . However, the comparisons also demonstrate that assumptions inherent in remote sensing algorithms and the impact of these on parameter estimates must be carefully examined.