A two-phase study of the potential impact of Geosynchronous Imaging Fourier Transfer Spectrometer (GIFTS) radiance data to the prediction of strong convective events was developed. In the first phase of the project, a statistical analysis of six runs of the Fifth Generation Pennsylvania State University/National Center for Atmospheric Research Mesoscale Model (MM5), version 3, was performed. These runs incorporate different size domains, numbers of vertical levels, numbers of nesting domains, and physical schemes. Using high-resolution National Center for Environmental Prediction (NCEP) Stage IV precipitation estimates, mesonet data, and radar reflectivity, it was determined that of all runs, one was chosen as being most appropriate for simulating GIFTS radiance. This run incorporates the simple ice microphysical scheme, the Grell cumulus scheme, the Blackadar planetary boundary layer scheme, and a simple atmospheric radiation scheme. Furthermore, this run was nested, with the mother domain (12-km resolution) of size 163 x 127 x 54 and the nested domain (4-km resolution) of size 103 x 127 x 54. In the second phase of the project, two sensitivity studies were carried out. In the first sensitivity study, the sensitivity of simulated GIFTS radiance to temperature and water vapor were examined. The 14 most sensitive channels within the GIFTS spectral range, out of 3,073, were chosen for further analysis. Through an analysis of an MM5 grid point that had relatively minimal cloud cover, it was determined that the most sensitive atmospheric layers at eight channels are in the lower troposphere (temperature) and lower to mid-troposphere (water vapor). At the other six, the most sensitive region is in the mid- to upper troposphere. The layers of maximum sensitivity are consistent with peaks of the weighting functions of these channels. The second sensitivity study examined the sensitivity of convective precipitation forecasts to the initial conditions of temperature and water vapor. The purpose of this study was to "bridge" the results of the first sensitivity study to the MM5 quantitative precipitation forecast (QPF) results. It was found that the most sensitive region is over the Central Plains of the United States and that the convective QPF is most sensitive to both water vapor content and temperature in the low-levels of the troposphere. Furthermore, temperature is deemed more sensitive to convective QPFs than water vapor. The results from these sensitivity tests, when linked together, demonstrate that GIFTS radiance at the eight wavenumbers most sensitive in the lower troposphere may be more effective to improve QPF than higher wavenumber radiance and that temperature in the Central Plains is the key meteorological variable to which the convective QPF is most sensitive. In a future four-dimensional variational data assimilation (4D-Var) study, simulated and real atmospheric observations from various sources will be assimilated into the MM5, with the GIFTS model representing the observation operator. Through this current study, a better sense of the utility of data from GIFTS to the forecasting of convective precipitation is ascertained, which would help streamline the 4D-Var study.
