Evaluating the semi-empirical H-Q model used in the SMOS retrieval algorithm. Can calibrated parameters HR, QR, NRV and NRH be linked to physical roughness parameters ?
Résumé
The Soil Moisture and Ocean Salinity (SMOS) mission was successfully launched in November 2009 and is the first mission dedicated to providing global surface soil moisture fields at a high temporal resolution of three days. The satellite carries a 1.4 GHz (L-band) interferometric radiometer. A forward inversion algorithm, using the so-called Lband Microwave Emission of the Biosphere (L-MEB) model, is used to retrieve land soil moisture from brightness temperature measurements. The L-band emissivity of bare soil is dependent on soil moisture, but is also significantly affected by surface roughness. Surface roughness effects are accounted for in L-MEB by a semi-empirical model known as the H − Q model. This model was chosen for use in the SMOS forward algorithm due to its simplicity and ease of computation: In the H−Q model the emissivity is calculated from the Fresnel equations, modified by an exponential factor. The model includes the parameters HR, QR and NRp, for polarization p, which account for surface roughness effects, and must be calibrated using measurements. In this paper we present a study which aimed to evaluate potential relationships between the parameters HR, QR, and NRp and physical roughness parameters (standard deviation of surface heights σ and autocorrelation length Lc), using a numerical modeling approach. The modeling approach used in this paper was presented comprises the use of Ansoft's High Frequency Structure Simulator (HFSS) software which solves Maxwell's equations using the numerical Finite Element Method (FEM)