Inversion of cloudy satellite sounding radiances by nonlinear optimal estimation. II: Application to TOVS data

A new approach to inverting satellite sounding data for use in numerical weather prediction (NWP) was proposed in a previous paper, where it was applied to simulated data from the TIROS Operational Vertical Sounder (TOVS). In this paper the method has been applied successfully to real TOVS data from the NOAA‐9 satellite, in a case study taken from the period of the Generation of Atlantic Lows Experiment (GALE). Temperature and humidity profiles, surface temperature and microwave emissivity, and cloud‐top pressure and fractional coverage have been retrieved simultaneously. The retrieved profiles have been compared statistically with co‐located radiosonde measurements. Also, fields of retrieved values have been compared qualitatively with corresponding NWP background fields (which provided the ‘first guess’ for the inversion) and with image data from TOVS itself and from the Advanced Very High Resolution Radiometer on the same satellite. Impact on the temperature field was found to be small, as expected over an area for which the conventional observing network is relatively dense. However, changes to the background field were generally found to be beneficial. At mid to upper tropospheric levels, the retrievals modified the background humidity field considerably. Retrieved values were both horizontally coherent and consistent with corresponding images and the expected dynamics. Retrieved cloud parameters have not been studied in great detail but have been found to be reasonable when compared with image data.