Comparision of Landsat and MODIS Estimates of Heat Fluxes: Effect of Surface Heterogeniety

Abstract

Using remote sensing to characterize the hydrologic behavior of the land surface on a routine basis is of considerable practical interest. Technique combining information on land and atmospheric properties with remotely observed variables has improved prediction of a number of hydrological variables, such as evapotranspiration rate, which is an important parameter for water resource management especially over agricultural regions. However, spatial extent of typical field scales is not regularly resolved within the pixel resolution of satellite sensors. Therefore landscape heterogeneity will influence the estimate of surface fluxes by satellite sensors with different spatial resolutions. Data from Landsat TM/ETM+ (120/60 m) and MODIS (1 km) satellite platforms are employed to independently estimate evapotranspiration. Evapotranspiration estimates derived with SEBAL at these multiple resolutions were assessed against eddy covariance flux measurements collected at two different areas, Brookings, SD with quite flat landscape and Fort Peck, MT with mountainous terrain. Altogether, these data allow a multiple scale intercomparison of remotely sensed predictions of heat fluxes. A high degree of consistency was observed between the retrievals of the net solar radiation and soil heat flux from Landsat and MODIS, while the comparison for sensible and latent heat fluxes was poor. For comparisons between satellite prediction and flux tower measurement, Landsat performs better than MODIS. The apparent degrade of agreement for both sensible and latent heat estimates with increasing spatial scale suggests that landscape heterogeneity influence the remotely sensed predictions of heat fluxes.