No abstract

No abstract

Remote sensing of the earth in the shodmidwave infrared (SMWIR) spectral region can be greatly obscured by the presence of clouds. In fact, data collections by space-based infrared sensors can be deemed inappropriate due to cloud cover. The purpose of this paper is to analyze data acquired with the Airborne Remote Earth Sensing (ARES) instrument in the 2.2-6.4 pm spectral regime, in order to quantify cloud coverage and cloud altitude to evaluate the potential of similar algorithms contemplated for future space-based systems.The Airborne Remote Earth Sensing (ARES) Program hosted a Lockheed-Martin Palo Alto Research Laboratory (LMPARL) dual mode radiometedimaging spectrometer aboard a NASA WB-57F to collect spatially and spectrally resolved images in the 2.2-6.4 pm region. Data collected over Hanscom AFB, MA on September 16, 1995 is used in this analysis. Cloud coverage and altitudes were estimated from spectra taken at nadir and 70" Look Zenith Angles over the ground site where radar measurements of cloud altitude were made.The method to determine cloud altitude and coverage is based on a linear least squares (LLSQ) algorithm which uses a matrix of spectral templates to estimate the fractional contribution of each to the input spectrum. Cirrus and cumulus cloud templates (predicted by the HYPEX software program) and terrain templates extracted from the data were utilized in the algorithm.The algorithm correctly identified the cloud conditions and yielded low residuals (4%-11%) and high confidence for the comparisons with radar during the nadir observations. Results for the off-nadir observations yielded higher residuals (139640%), possibly due to difficulties extracting clear terrain s8pectra and insufficient coverage of the actual conditions with the selected templates. Although these offnadir residuals were large, the LLSQ algorithm successfully selected templates for cloud types and altitudes consistent with the radar data measured for these tracks.