Soil moisture and temperature determination by recursive assimilation of multifrequency observations using simplified models of soil heat and moisture flow and emission

Summary form only given, as follows. Although the sensitivity of radiobrightness to soil moisture content has been well established, direct inversion of remote sensing measurements is impeded by variable surface cover (vegetation and snow), surface roughness, and soil temperature and moisture profiles. During interstorm periods, soil moisture and temperature conditions change dynamically under atmospheric and radiative forcing. A model of these dynamics may be used to encapsulate a priori soil information and constrain the inversion solution. A soil and land-atmosphere model driven by local meteorology or forecasts can propagate soil conditions between times at which remote sensing measurements are available. The predictions of the model soil would facilitate incorporation of imperfect measurements from multiple sources through comparisons of the modeled and observed states. This paper presents the results of an assimilation study using simple models of the coupled heat and moisture transport in soil and microwave emission. Simple models enable the full utilization of available microwave and thermal infrared measurements through recursive application of Kalman filtering. Comparisons are made between soil emission models based on coherent wave radiative transfer, classical radiative transfer (CRT), a first order approximation to CRT, and a gray body. Data from the BARC (Beltsville Agricultural Research Center) 1994 experiment-including micrometeorology, soil conditions, and L-band (1.4 GHz) radiobrightness-drive the model. The authors test the sensitivity of the model to temporal radiometric measurement availability ranging from hourly to daily and show that a simplified emission parameterization is adequate.