Dependence of retrievals of cloud and atmospheric properties on the spectral resolution of infrared measurements

Abstract:

This work investigates the extent to which atmospheric properties (including cloud properties) can be retrieved from downwelling infrared spectra for different spectral ranges and resolutions. For this purpose, hundreds of spectra have been simulated that represent the range of clear and cloudy conditions experienced in the Arctic, at resolutions ranging from perfect (monochromatic) to less than 8 cm^-1. This dataset allows retrievals to be performed on many different realizations of the Arctic atmosphere to develop statistical knowledge of how well the retrieval system performs. Since simulated spectra are used, the atmospheric and cloud conditions are a priori and can thus be compared with what is retrieved. Retrieved properties include cloud height and cloud microphysical properties. For cloud height, three methods are compared: CO2-sorting, the Minimum Local Emissivity Variance, and optimal estimation. Retrievals of cloud microphysical properties are found to be relatively insensitive to the resolution of the infrared spectra, because cloud properties vary slowly across the infrared spectrum.