Assimilation of Hyperspectral Satellite Radiance Observations within Tropical Cyclones

The availability of high resolution temperature and water vapor data is critical for the study of mesoscale scale weather phenomena (e.g., convective initiations, and tropical cyclones). As hyperspectral infrared sounders, the Atmospheric Infrared Sounder (AIRS) and Geosynchronous Imaging Fourier Transform Spectrometer (GIFTS) could provide high resolution atmospheric profiles by measuring radiations in many thousands of different channels. This work focuses on the assessment of the potential values of satellite hyperspectral radiance data on the study of convective initiations (CI) and the assimilation of AIRS radiance observations within tropical storms. First, the potential capability of hyperspectral infrared measurements (GIFTS) to provide convective precipitation forecasts has been studied and assessed. Using both the observed and the model-predicted profiles as input to the GIFTS radiative transfer model (RTM), it is shown that the simulated GIFTS radiance could capture the high vertical and temporal variability of the real and modeled atmosphere prior to a convective initiation, as well as the differences between observations and model forecasts. This study suggests the potential for hyperspectral infrared radiance data to make an important contribution to the improvement of the forecast skill of convective precipitation. Second, as the first step toward applying AIRS data to tropical cyclone (TC) prediction, a set of dropsonde profiles during Hurricane Rita (2005) is used to simulate AIRS radiance data and to assess the ability of AIRS data in capturing the vertical variability within TCs through one-dimensional variational (1D-Var) twin experiments. The AIRS observation errors and background errors are first estimated. Five sets of 1D-Var twin experiments are then performed using different combinations of AIRS channels. Finally, results from these 1D-Var experiments are analyzed. Major findings are: (1) AIRS radiance data contain useful information about the vertical variability of the temperature and water vapor within hurricanes; (2) assimilation of AIRS radiances significantly reduced errors in background temperature in the lower troposphere and relative humidity in the upper troposphere; (3) the near-real time (NRT) channel set provided by NOAA/NESDIS seems sufficient for capturing the vertical variability of the atmosphere in the upper troposphere of TCs, but not in the lower troposphere; and (4) the channels with weighting functions peak within the layer between 500-700 hPa could provide useful information to the atmospheric state below 700 hPa. A channel selection method is proposed to capture most vertical variability of temperature and water vapor within TCs contained in AIRS data. Finally, AIRS radiance data within TCs have been assimilated in the 1D-Var experiments with comparisons of the retrieval temperature and water vapor profiles with co-located Global Positioning System (GPS) radio occultation (RO) soundings and dropsonde profiles. The comparisons of AIRS 1D-Var retrieval profiles with GPS RO sounding show that AIRS data can greatly improve the analysis of temperature and water vapor profiles within TCs. The comparisons of retrieval profiles with dropsonde data during Hurricane Rita, however, showed some discrepancies partly due to the difference of these two measurements and the uncertainties of the AIRS errors.