Abstract. Single-footprint Atmospheric Infrared Sounder spectra are used in an optimal estimation-based algorithm (AIRS-OE) for simultaneous retrieval of atmospheric temperature, water vapor, surface temperature, cloud-top temperature, effective cloud optical depth and effective cloud particle radius. In a departure from currently operational AIRS retrievals (AIRS-V6), cloud 10 scattering and absorption are in the radiative transfer forward model, and Level 1b AIRS thermal infrared data are used directly rather than Level 2 cloud-cleared spectra. Coincident MODIS Level 2 cloud data are used for cloud a priori. Using Level 1b spectra improves the horizontal resolution of the AIRS retrieval from ~45 km to ~13.5 km at nadir, but as microwave data are not used, retrieval is not made at altitudes below thick clouds. An outline of the AIRS-OE retrieval procedure and information content analysis is presented. Initial comparison of AIRS-OE to AIRS-V6 results show increased horizontal detail in the water 15 vapor and relative humidity fields in the free troposphere above clouds. Comparisions of temperature, water vapor and relative humidity profiles against coincident radiosondes show good agreement. Future improvements to the retrieval algorithm, and to the forward model in particular, are discussed.