Impacts of improved land surface representations in a regional climate model with use of MODIS land surface temperature

As much attention has been given to quantify the climatic impacts of human modification of landscape at regional to global scales, realistic parameterization of key biophysical variables in climate models becomes imperative. Remotely sensed biophysical data have recently been employed to improve the land surface representation in climate modeling. Vegetation fractional cover (VFC) is such an important variable which decides the proportions of soil and vegetation in an area. In this paper, newly developed VFC data from the Moderate Resolution Imaging Spectroradiameter (MODIS) are incorporated in the Regional Atmospheric Modeling System (RAMS), replacing the built-in VFC specified by unrealistic mathematical equations. RAMS was run 12 months in 2003 with study area focused on East Africa region, which has one of the most complex landscape in the world. One big issue of climate simulation in this region is its lack of spatially and temporally explicit station observations for model validation. Land surface temperature (LST) products from MODIS Terra and Aqua were used to compare with the RAMS simulated LST. This study found that the built-in VFC in RAMS is too homogeneous spatially to differentiate distinct land surface types across the domain. Temporally, the prescribed VFC varies little with season especially for low-latitude regions such as East Africa. By using MODIS VFC, the simulated LST was greatly improved. Specifically, the bimodal feature of the LST seasonal variation, which is completely missed in the default land surface configurations, is better captured when MODIS VFC is incorporated. Regarding to the spatial characteristics, the Intertropical Convergence Zone (ITCZ) related seasonal migration of LST in the eastern domain has been greatly enhanced. Both MODIS Terra and Aqua LST are used for the first time to evaluate the surface impact on diurnal LST characteristics. This study found that the diurnal LST cycles in the second half of the year are slightly improved due to the new land surface representation. However, the improved land surface has little influence on simulated precipitation over 12 month period.