Atmospheric water-vapour profiling from passive microwave sounders over ocean and land. Part I: Methodology for the Megha-Tropiques mission

International audienceA water-vapour retrieval algorithm has been developed that uses satellite observations in the microwave region. It is based on neural-network modelling and includes a dedicated calibration scheme for the satellite observations. The water vapour is retrieved for clear and cloudy scenes, over both ocean and land surfaces. Precipitation cases are excluded. The atmospheric relative humidity profile is retrieved on six atmospheric layers, together with the total column water vapour. By-products are also retrieved by the algorithm, including surface temperature and microwave emissivities over the continents and surface wind speed over the ocean. A first version of a retrieval chain has been produced for the French-Indian Megha-Tropiques mission launched on 12 October 2011. The algorithm has been further developed for the instruments AMSR-E/HSB (resp. AMSU-A/MHS) on board the AQUA (resp. MetOp) platform, in order to test it on existing satellite observations. In this article, the principles of the inversion method are presented and the theoretical retrieval uncertainties are assessed using direct tests on simulated data as well as estimations using the traditional information-content analysis. Results of the retrieval algorithm will be evaluated in a companion article for AQUA and MetOp observations using comparisons with European Centre for Medium-Range Weather Forecasts (ECMWF) analysis and radiosonde measurements. © 2012 Royal Meteorological Society