Abstract
A method is proposed for testing microwave measurements from spaceborne sensors by computing collocated simulated brightness temperatures from the ECMWF numerical weather meteorological model using an atmospheric radiative transfer model and the Stogryn model for the surface emissivity. In this paper this method is tested on Scanning Multichannel Microwave Radiometer (SMMR) and Special Sensor Microwave/Imager (SSM/I) data. The comparison with observed brightness temperature underlines a rather good agreement for SSM/I but a significant discrepancy for SMMR in the vertical polarization, confirming that the SMMR was biased (up to 10 K). Correcting functions are proposed for SMMR 18- and 21-GHz channels, which are validated by comparing Wilheit and Chang's water vapor algorithm results and radiosonde measurements. Small biases in the SSM/I 19- and 22-GHz brightness temperatures comparison can be similarly eliminated. This calibration is validated in the same way by applying water vapor and surface wind retrieval algorithms and comparing the results with corresponding local measurements. The SSM/I algorithms were developed using the same radiative transfer and surface emissivity model, applied to a dataset based on five months of ECMWF analyses. This study establishes the validity of the model, and the reliability of ECMWF analyses, both for validating microwave brightness temperatures and for developing retrieval algorithms in view of future sensors.
