Meteo-France has been working on AAPP (ATOVS & AVHRR Processing Package) & ICI (Inversion Coupled with Imageur) software development for several years. These softwares are used to treat ATOVS (HIRS + AMSU) sounding data and AVHRR images on NOAA satellite.
MAIA (Mask AVHRR for Inversion ATOVS) software is included in AAPP and is operational at CMS for several years. MAIA process the AVHRR pixels mapped inside HIRS F.O.V and determines the mean clear pourcentage cover on the HIRS spot and when some clear AVHRR pixels are detected in the fov, a skin surface temperature from AVHRR spilt-window, with two different formula over sea and land. The used forecast air surface temperature (over land) or the sea surface climatology (over sea) is also kept. Also when a black-body cloud is found, a cloud top temperature is given.These pieces of information are necessary to obtain good ICI inversion results and have been validated in the past over several climatological areas.
The MAIA software is a multispectral threshold technique for which values are adapted to the observation conditions by specific computations. In the first version of the software (MAIA-1), the test thresholds were calculated for mid-latitude regions (some were determined by experience over Europe). The MAIA-2 version have been developped for dynamically adapting the thresholds to variable climatological conditions, mainly the total water vapor content and the skin surface temperature. This was required for implementing the AAPP/ICI package in other HRPT centers everywhere in the world. The developments were first undertaken within the SAF in support to Nowcasting and Very Short Range Forecasting (SAF NWC) by using GOES-8 imagery daily recieved at CMS to simulate extra mid-latitude conditions. This study gave good results and the method has been adapted to the AVHRR imager.
The new version of MAIA is now available in the AAPPv3 package and was implemented in the CMS operational scheme since June1999. The use of MAIA-2 should improve the cloud mask, particularly over wet tropical regions. MAIA-2 software is installed in LaReunion, China and Brazil.
The new version of the mask can also be easely used for full AVHRR resolution applications (SST, NDVI ...) by removing the HIRS ellipse shape in the processing and by keeping the cloud cover and surface temperature outputs at full resolution. However, a validation of the full resolution products should be undertaken. MAIA-2 also should be considered as a basic software for imaging users as it allows to compute the cloud cover (and by now the surface temperature) at the AVHRR full resolution.
Ardouin L., G. Monnier, L. Lavanant, 1999: Adjustment, validation and implementation of MAIA2 in AAPP software. Internal report.
Derrien M, Farki B., Harang L., Le Gléau H., Noyalet A., Pochic D., A.Sairouni, 1993 : Automatic cloud detection applied to NOAA-11 / AVHRR imagery, Remote Sensing of Environment, 46, pp 246-267.
Derrien D, H. LeGléau, 1999: Cloud classification extracted from AVHRR and GOES imagery, Proceedings of Eumetsat Meteorological satellite data conference- 1999
Drezen C., P. Bryere, 2001: General specifications for the AAPP preprocessing package. scientic part. software architecture
Lavanant L., H. LeGléau, M. Derrien, S. Levasseur, G. Monnier, L. Ardouin, P. Brunel, B. Bellec, 1999: AVHRR Cloud Mask for Sounding Applications. ITSC-10 proceedings.
Lavanant L., 2000: Retrieval of cloud and surface parameters with AVHRR for IASI processing. Validation part. MASSIF2 report.
LeGleau H., M. Derrien, 2000: Prototype scientific description for Meteo-France/CMS.
SAF/NWC/MFCMS/MTR/PSD, isuue 1
(you can get a copy at the web page : www.meteorologie.eu.fr/safnwc)