The Utility of Remotely-Sensed Land Surface Temperature from Multiple Platforms For Testing Distributed Hydrologic Models over Complex Terrain

Land surface temperature (LST) is a key parameter in watershed energy and water budgets that is relatively unexplored as a validation metric for distributed hydrologic models. Ground-based or remotely-sensed LST datasets can provide insights into a model's ability in reproducing water and energy fluxes across a large range of terrain, vegetation, soil and meteorological conditions. As a result, spatiotemporal LST observations can serve as a strong constraint for distributed simulations and can augment other available in-situ data. LST fields are particular useful in mountainous areas where temperature varies with terrain properties and time-variable surface conditions. In this study, we collect and process remotely-sensed fields from several satellite platforms - Landsat 5/7, MODIS and ASTER - to capture spatiotemporal LST dynamics at multiple resolutions and with frequent repeat visits. We focus our analysis of these fields over the Sierra Los Locos basin (~100 km2) in Sonora, Mexico, for a period encompassing the Soil Moisture Experiment in 2004 and the North American Monsoon Experiment (SMEX04-NAME). Satellite observations are verified using a limited set of ground data from manual sampling at 30 locations and continuous measurements at 2 sites. First, we utilize the remotely-sensed fields to understand the summer seasonal evolution of LST in the basin in response to the arrival of summer storms and the vigorous ecosystem greening organized along elevation bands. Then, we utilize the ground and remote-sensing datasets to test the distributed predictions of the TIN-based Real-time Integrated Basin Simulator (tRIBS) under conditions accounting static and dynamic vegetation patterns. Basin-averaged and distributed comparisons are carried out for two different terrain products (INEGI aerial photogrammetry and ASTER stereo processing) used to derive the distributed model domain. Results from the comparisons are discussed in light of the utility of remotely-sensed LST for testing distributed hydrologic models in regions of complex terrain with seasonal changes in land surface conditions.