Automated Retrieval of ARM Nocturnal Cloud Properties From Multispectral Satellite Data
Smith, W.L., Jr., Analytical Services and Materials, Inc.;
Minnis, P., and Young, D.F., NASA Langley Research Center
Eighth Atmospheric Radiation Measurement (ARM) Science Team Meeting
Current retrievals of cloud properties at night from satellite imager data rely on single-channel infrared techniques to discern cloud amount and a crude estimate of cloud-top altitude. Because many clouds are semi-transparent, the cloud heights are often severely underestimated. To improve the ARM satellite cloud property retrievals at night, a multispectral method is being developed for quasi-operational analysis of 4-km-resolution Geostationary Operational Environmental Satellite (GOES) data. This technique uses the 3.8, 10.8, and 12.0-µm channels of GOES-8 and new surface emissivity data to improve the determination of clear and cloudy scenes. The cloud optical depth, phase, and particle sizes are also estimated for each pixel by matching the radiances in an iterative method to model calculations. The estimation of cloud optical depth allows a correction for semi-transparency facilitating an improved estimate of cloud height. The prototype operational technique and initial results are presented for hourly data taken over several nights to include a variety of cloud types, heights, and mixtures. This solar-infrared, infrared, split-window (SIRS) technique provides much better clear-cloud discrimination than possible with infrared-only methods. Thin cirrus and low-level clouds are detected with greater ease. SIRS also yields realistic cloud microphysical properties and more accurate cloud heights. This method will be incorporated into the ARM satellite cloud data analysis algorithms.
