In this paper we explore the use of high-resolution GOES thermal observations for validation and data assimilation within a land surface hydrological model. The VIC-3L land surface model is run at an hourly time step at 1/4 degree resolution for 1997 for the Red River-Arkansas basin. Previous work has compared surface temperature observations at coarse (1 degree) spatial and temporal (twice daily) scales using remotely-sensed surface temperature from TOVS (1 degree).
We compare surface temperature outputs from the VIC model with remotely-sensed surface temperatures derived from hourly GOES observations using a split-window atmospheric correction algorithm.
The spatial and temporal variability of model and GOES surface temperature fields are analyzed for the year and differences are related to model structure and errors in input variables. This provides an independent means of validating the spatial and temporal
variability of one aspect of the model surface energy balance. Ultimately, GOES observations will be assimilated to improve hydrologic model fluxes.