Nighttime multilayered cloud detection using MODIS and ARM data Academic Article uri icon


  • This study reports on recent progress toward the discrimination between pixels containing multilayered clouds, specifically optically thin cirrus overlying lower-level water clouds, and those containing single-layered clouds in nighttime Moderate Resolution Imaging Spectroradiometer (MODIS) data. Cloud heights are determined from analysis of the 15-m CO2 band data (i.e., the CO2-slicing method). Cloud phase is inferred from the MODIS operational bispectral technique using the 8.5- and 11-m IR bands. Clear-sky pixels are identified from application of the MODIS operational cloud-clearing algorithm. The primary assumption invoked is that over a relatively small spatial area, it is likely that two cloud layers exist with some areas that overlap in height. The multilayered cloud pixels are identified through a process of elimination, where pixels from single-layered upper and lower cloud layers are eliminated from the data samples. For two case studies (22 April 2001 and 28 March 2001), ground-based lidar and radar observations are provided by the Atmospheric Radiation Measurement (ARM) Program's Southern Great Plains (SGP) Clouds and Radiation Test Bed (CART) site in Oklahoma. The surface-based cloud observations provide independent information regarding the cloud layering and cloud height statistics in the time period surrounding the MODIS overpass.

published proceedings

  • Journal of Applied Meteorology

author list (cited authors)

  • Baum, B. A., Frey, R. A., Mace, G. G., Harkey, M. K., & Yang, P

publication date

  • January 1, 2003 11:11 AM