Impact assessment of additional predictors to the retrieval of atmospheric profiles from infrared radiances

Abstract

Impacts of additional predictors on inverting atmospheric infrared radiance for temperature and humidity profiles are investigated using Atmospheric Infrared Sounder (AIRS) real measurements and empirical orthogonal function expansion method (EOF). These predictors are microwave channels, latitude, topography, surface altitude, surface temperature, and surface air pressure. The results suggest that microwave channels can remarkably help the improvement of the accuracy of retrieved profiles at lower troposphere (below 800hPa) and have little effect on that above 800hPa. With dataset classified by latitude, better retrievals are obtained. The root mean square errors (RMSE) of retrieved temperature at complicated terrain are significantly greater than that at plat area. For humidity retrievals it was found that RMSE exhibit weak sensitivity to topography. By combined use of infrared measurements and additional predictors, great improvements have achieved in the retrieval of atmospheric temperature and humidity profiles at lower troposphere.