Temperature and moisture retrieval from AIRS measurements under cloudy condition over East Asia

Linear regression method was applied for the retrieval of temperature and moisture profiles above cloud from Atmospheric Infrared Sounder (AIRS) measurement over East Asia. To examine the statistical relationship between compressed AIRS radiances and atmospheric states, pairs of ECMWF interim temperature, moisture and cloud water contents data for one year of 2007 and their corresponding simulated radiances are constructed. Regression coefficients varying with cloud top pressure (CTP) and satellite zenith angle were obtained to reduce nonlinearity between temperature and moisture states and AIRS radiance spectrum. Optimizing the algorithm for East Asia, regression was done only with atmospheric status over East Asia. In order to obtain the CTP, a statistical artificial neural network (ANN) approach was taken to retrieve Cloudsat-like CTP from AIRS radiance spectrum. Retrieved results show that spatial distributions of temperature and moisture are in general agreement with ECMWF reanalysis. Error analysis shows that RMS errors of 3-4 K and 1.5 g/kg for temperature and moisture, respectively, in troposphere while mean bias of both of them is close to 0 for the whole troposphere.