Retrieval of atmospheric and surface properties from high-resolution FTIR measured radiances

As more and more high-resolution FTIR instruments planned for future satellite remote sensing, robust and fast physical retrieval algorithms are needed to invert the measured radiances to geophysical parameters. One of the important components of a physical retrieval algorithm is the fast forward model. For interferometer-based sounders, the Sensor Response function (SRF) may not be localized. A fast and accurate forward model, Optimal Spectral Sampling (OSS), is used to calculate upwelling radiances in our physical retrieval algorithm. It's capable of modeling radiance spectra with either a localized or a non-localized SRF with negative side lobes. Derivatives with respect to atmospheric and surface properties can be calculated analytically and efficiently. The OSS fast RT model is idea for atmospheric sounding or atmospheric compensation applications. The inversion method is based on an optimal estimation algorithm. Empirical Orthogonal Functions (EOF) are used to transform the atmospheric profiles into more compact form for fast and stable inversions. The non-linearity of the radiative transfer function is taken into account in the algorithm so the inversion is very robust. The algorithm has been used to simulate the Environmental Data Record (EDR) retrieval performance for the Cross-track Infrared Sounder (CrIS). We extended the method to model cloud directly. The cloud parameters are retrieved simultaneously with the atmospheric and surface parameters. This algorithm has been successfully applied to the NPOESS Aircraft Sounder Testbed (NAST-I) measured radiances.