Predicting mean radiative temperature at millimeter wavelengths in continental climate areas

The aim of this work is to present a parametric model for the prediction of the down-welling mean radiative temperature (Tmr) in the microwave region. For the model setup, the Sky Noise Eddington Model has been exploited to generate a synthetic clouds dataset driven by radiosounding observations. The core of the model, developed in continental climate areas, is the normalization of Tmr for the surface temperature (TS) of the geographic area of interest that makes the Tmr/TS model applicable in each site belonging to the considered climatic area. Retrieval of the predicted Tmr can be done by multiplying the Tmr/TS obtained from the parametric model at the considered attenuation value for the average TS. The model parameters depend on frequency and elevation angle within the range from 23.8 GHz to 49.5 GHz and from 90° to 5°, respectively. A preliminary verification of the model is accomplished, at 23.8 GHz and 31.4 GHz, exploiting Suntracking radiometric measurements available in Rome (NY, USA) with promising results: RMSE never exceeds 0.02 and 0.1 at 23.8 GHz and 31.4 GHz, respectively.