Combined use of active and passive microwave satellite data to constrain a discrete scattering model

Emissivity and backscattering have different sensitivities to soil and vegetation parameters and provide complementary information on the observed scene. For this reason, the main objective of this study was to investigate the potential of the combined use of active and passive microwave signatures to improve the understanding of emission and scattering mechanisms over land. The discrete scattering model developed by the Tor Vergata University, driven by a single set of input parameters, was used to simulate both emission and backscattering. The model accuracy was assessed by comparing modelled emissivity, multiplied by the effective temperature, with C-band AQUA AMSR-E observations and modelled backscattering coefficient with Metop ASCAT observations. The investigation was carried out using datasets collected over the Maqu area, which is located in the north-eastern part of the Tibetan Plateau and is covered by homogeneous grassland with a thin litter layer. In situ soil moisture measurements and MODIS Terra/Aqua 8-days LAI products served as model inputs, whereas the most sensitive other parameters, such as surface roughness and litter and vegetation parameters, were calibrated to obtain the best match between modelled outputs and both ASCAT and AMSR-E 2009 observations.

With the calibrated parameters, measured soil moisture and MODIS LAI as input, the model correctly simulated the temporal variations of ASCAT backscattering coefficient, with R2 of 0.9 and rmse of 0.5 dB. The results obtained in the passive case strongly depended on the adopted effective temperature. The best match with satellite observations was found when the surface temperature was obtained from Ka-band AMSR-E data, leading to R2 of 0.8 and rmse of 6 K for H polarisation data and R2 of 0.5 and rmse of 6 K for V polarisation data.

The validation carried out using the 2010 data, without any further refinement of the calibrated parameters, showed a slightly worse agreement between modelled and observed signatures.

The results showed that the combined use of active and passive microwave signatures contributed to an unambiguous interpretation of emission and scattering processes over the study area, it allowed us to impose constraints to non-available inputs of the Tor Vergata model and it led to correct simulations of the satellite signatures. These findings are important to better understand the potential of the synergistic use of radiometer and radar data, like those to be provided by NASA's SMAP.