Evaluating the semi-empirical H − Q model, used to calculate the emissivity of a rough bare soil, with a numerical modeling approach
Résumé
In this paper, a numerical modeling approach was used to evaluate the semiempirical H-Q model used in the Soil Moisture and Ocean Salinity (SMOS) retrieval algorithm to account for roughness effects over bare soil. The H-Q model uses four parameters, HR, QR, NRH, and NRV 10 , which are usually calibrated at the ground scale for different surface types. The aim of this paper is to investigate whether these empirical parameters could be linked to the physical roughness parameters of standard deviation of surface heights σ and autocorrelation length Lc . First, a numerical modeling approach was used to calculate rough soil emissivities for different roughness and soil moisture conditions. Second, H-Q model parameters were retrieved by minimizing a cost function between these emissivities and those calculated by theH-Q model. It was found that the retrievedHR could be related directly to Zs = σ2/Lc and thatQR,NRV , and NRH were dependent onHR.HR was found to have a negligible dependence on soil moisture. Based on these results, a new model was proposed where the four H-Q model parameters were calibrated to Zs. This model was tested on the PORTOS 1993 data set and found to yield a root-mean-square difference between the retrieved and measured soil moisture values of ∼ 0.03 m3/m3, which was within the desired 0.04-m3/m3 error margin for the SMOS mission.