A Comparison between Evapotranspiration Estimates based on Remotely Sensed Surface Energy Balance and Ground-based Soil Water Balance Analyses

摘要

Remotely sensed and in-situ data were used to investigate dynamics of root zone soil moisture and evapotranspiration (ET) at four Mesonet stations in north-central Oklahoma over an 11-year period (2000-2010). Two moisture deficit indicators based on soil matric potential had spatial and temporal patterns that were consistent with collected precipitation and reference ET data, as well as the state-wide historical trends. Actual ET was estimated by closing the soil water balance using in-situ soil moisture data, as well as by applying a surface energy balance approach to Landsat Thematic Mapper 5 imagery. When averaged over all years, weekly ET estimates with both approaches had a small mean signed difference (0.5 mm·week~(-1)), but a large root mean square deviation (11.6 mm·week~(-1)). This difference may be partly due to errors caused by interpolation of ET estimates between less-frequent satellite overpasses. Averaging over longer periods or modifying the interpolation methods may improve the accuracies for short-term remotely sensed ET estimates.