Abstract:
Land Surface Temperature (LST) is an important parameter to assess the energy state of a surface. Synoptic satellite observations of LST must be used when attempting to estimate fluxes over large spatial scales. Due to the close coupling between LST, root level water availability, and mass and energy fluxes at the surface, LST is particularly useful over agricultural areas to help determine crop water demands and facilitate water management decisions (e.g., irrigation). Further, LST can be assimilated into land surface models to help improve estimates of latent and sensible heat fluxes. However, the accuracy of LST products and its impact on surface flux estimation is not well known. In this study, we quantify the uncertainty limits in LST products for accurately estimating latent heat fluxes over agricultural fields in the Rio Grande River basin of central New Mexico. We use the Community Land Model (CLM) within the Land Information Systems (LIS), and adopt an Ensemble Kalman Filter approach to assimilate the LST fields into the model. We evaluate the LST and assimilation performance against field measurements of evapotranspiration collected at two eddy-covariance towers in semi-arid cropland areas. Our results will help clarify sensor and LST product requirements for future remote sensing systems.