Estimating root-zone moisture and evapotranspiration with AVHRR Data

A model framework for parameterized subgrid-scale surface fluxes (PASS) has been applied to use satellite data, models, and limited surface observations to infer root-zone available moisture (RAM) content and evapotranspiration rate with moderate spatial resolution over large terrestrial areas. Data from Advanced Very High Resolution Radiometers on the NOAA-14 satellite were collected and adjusted for atmospheric effects using LOWTRAN7 and radiosonde data. The normalized difference vegetation index (NDVI) and surface radiant temperature were evaluated using the AVHRR data. Local surface observations provided data on downwelling solar irradiance, air temperature, relative humidity, and wind speed. Surface parameters including roughness length, albedo, surface water conductance, and the ratio of soil heat flux to net radiation were estimated; pixel-specific near-surface meteorological conditions such as air temperature, vapor pressure, and wind speed were adjusted according to local surface forcing; and estimates of pixel-scale RAM content were made by using surface energy balance and aerodynamic methods. After initial values of RAM content and NDVI from clear-day satellite overpasses were obtained, they were combined with the standard surface meteorological data and used in PASS to estimate evapotranspiration rates and the associated changes in RAM content for subsequent days. Estimates of RAM content were made for four soil layers using a linear relationship between an evaporation rate factor and the relative available soil moisture content. This modeling system makes effective use of satellite data and can be run for large areas for which flux data do not exist and surface meteorological data are available from only a limited number of ground stations. The PASS approach was applied to, and assessed with, observations from the 1997 Cooperative Atmosphere-Surface Exchange Study at the Atmosphere Boundary Layer Experiments site in the Walnut River Watershed in Kansas. Comparisons with radar cumulative precipitation observations, in situ soil moisture measurements, and eddy correlation moisture fluxes from aircraft indicated that the spatial and temporal variations of RAM were simulated reasonably well by PASS.