Downscaling of climate scenarios on air temperature using time series of TERRA/MODIS land surface temperature

The characterization of climate within complex landscapes requires knowledge of air temperature at high spatial resolution.In this work a procedure for the estimation of 1 km resolution maximum air temperature has been implemented starting fromLand Surface Temperature (Ts) observed by TERRA/MODIS from 2000 to 2006 and downscaled climate scenario at coarseresolution produced within the project "Agroscenari". The method was tested in the "Beneventano" area (Valle Telesina). Wehave shown that the interannual variability of the ratio of land surface temperature at an arbitrary location Ts (x,y,z) to the landsurface temperature at the reference node Ts (x0,y0,z0) can be neglected. It is exactly this result that makes the method applicableto reconstruct the sub-grid spatial variability of the air temperature obtained by the statistical downscaling of climate scenarioat the 35km x 35 km spatial resolution.