Deconvolution of Soil Moisture and Vegetation Emissions from Passive Microwave Brightness Temperatures Using Visible/Infrared Observations

Abstract:

Passive microwave (PM) observations of soil moisture (SM), like those produced from data observed by the AMSR-E, WindSat, AMSR2, and SMOS instruments, provide global soil moisture data sets with moderate resolution (~25km), reasonable accuracy (±10%), and short revisit times (2-3 days). A principal source of the current error in these SM data sets is due to heterogeneous topography below the native resolution of the PM instrument. A single PM instantaneous field of view (IFOV) may encompass surface water, dense and/or sparse vegetation, and bare soil. We show that by using satellite-based, high resolution (~250m) visible/infrared (VIS/IR) observations to estimate the fractions of water, vegetation, and bare soil in each PM IFOV, we can deconvolve the brightness temperatures from each individual component. This allows for greatly increased accuracy in the estimated soil moisture content. We will present our results in applying this technique to the WindSat soil moisture algorithm using WindSat PM data and vegetation and water fraction estimates derived from MODIS VIS/IR data.