83rd Annual

Tuesday, 11 February 2003: 12:00 PM
Free Tropospheric Humidity observations from METEOSAT water vapor channel data
Rémy Roca, LMD, Ecole Polytechnique, Palaiseau, France; and H. Brogniez, L. Picon, and M. Desbois
Poster PDF (356.0 kB)
The METEOSAT satellites provide 6.3 microns radiances measurements every 30 minutes at 5km resolution, continuously since around 20 years offering a unique documentation of the water vapor distribution and its variability over Africa and the tropical Atlantic Ocean. In this paper, we present a quantitative analysis of an eleven year part of this long term dataset.

The radiance are first inverted in terms of the weighting-function weighted mean relative humidity of the atmospheric column. Owing to the variability of the weighing function this parameter will be designed as the Free Tropospheric Humidity (FTH). The retrieval algorithm was inspired by the EUMETSAT operational UTH algorithm and includes a temperature vertical distribution normalizing factor and a wider vertical layer for the look-up table computations. The algorithm has been validated using METEOSAT-5 measurements over the Indian Ocean and an ensemble of 240 radiosondes. The METEOSAT-retrieved FTH agrees well with the sonde derived FTH with a small bias of 2.7% and a standard deviation of 6%.

This retrieval technique is then applied onto the 1983-1994 METEOSAT archive extracted from the ISCCP-B3 dataset at a resolution of 30km and 3 hours over Africa and the tropical Atlantic. Previous to the retrieval two processing steps are need. First, owing to the change of radiometer along the time and to the calibration technique updates, the homogenization of the long term series is performed. A dedicated method using the relative departure of the observed radiances from the ERA-15 analysis simulated ones is used yielding to strong correction of the original radiances which insure a stable long term time series. Second, only the clear sky scenes are used in building the FTH climatology. The cloud clearing only concerns the mid-to-upper cloudiness which are detected and characterized using the ISCCP-DX dataset. The results of the METEOSAT archive homogeneization will be shown as well as the outputs of the cloud clearing approach. Finally, the inter annual variability of FTH during this period will also be presented at the conference.

Supplementary URL: