Interpretation of L-band brightness temperatures of differently tilled bare soil plots

The structure of the surface soil layer is strongly influenced by soil tillage practices with important consequences for the soil hydraulic properties and soil moisture dynamics in the top soil layer. In this study, an L-band microwave radiometer and an infrared camera were used to monitor bare soil plots with different structure: tilled, seedbed, and compacted plots. The L-band brightness temperatures were calculated from the raw radiometric data using the radiometer effective transmissivity estimated with the described new algorithm for sky calibration. The new calibration algorithm reduces the bias of brightness temperature estimates. Radiative transfer, dielectric mixing, roughness correction, and soil hydrological models were coupled to determine and disentangle soil hydraulic and surface roughness parameters of the bare soil plots from time series of L-band brightness temperatures using inverse modeling. Two soil hydraulic property models were considered: the uni-modal model of Mualem van Genuchten and the bi-modal model of Durner. Microwave radiative transfer was modeled by two different approaches: the Fresnel equation with depth averaged dielectric permittivity of 2 cm or 5 cm thick surface layers, and a coherent radiative transfer model(CRTM) that accounts for vertical gradients in dielectric permittivity. Two global optimization algorithms (DREAMzs and SCE-UA) were implemented to estimate the optimal solution and the posterior distribution of the soil hydraulic and surface roughness parameters. Brightness temperatures simulated by the CRTM and the 2-cm layer Fresnel model fitted well to the measured ones suggesting that L-band brightness temperatures may be linked to the soil moisture in a 2 cm thick surface layer. Differences in absolute and normalized L-band brightness temperatures between the plots reflect the effect of tillage on the soil structure. The inversely estimated surface roughness parameters compared well with those derived from laser profiler measurements. Both the laboratory derived and the retrieved from L-band brightness temperatures water retention curves were bi-modal. In order to validate the inversely retrieved soil hydraulic functions, simulated water contents were compared with insitu measurements and differences in predicted evaporation rates between the plots were compared with differences in measured IR temperatures. Depth specific calibration relations were found to be essential to derive soil moisture from near-to-surface installed sensors. Furthermore, differences in simulated actual evaporation rates between the plots were confirmed by observed differences in measured IR temperatures. The results, presented in this study, indicate that effects of soil management on soil surface roughness and soil hydraulic properties can be inferred from L-band brightness temperatures