Visibility Over Land from Contrast Analysis of Multi-Spectral Satellite Observations
Abstract
The objective of this thesis is to investigate the viability of using contrast reduction in multi-spectral satellite observations to characterize surface visibility reduction due to heavy aerosol loading. Two methods are explored. First, the spectral distribution of standard deviation of surface reflectance over a homogeneous background (urban, agriculture, or forested) is plotted for three aerosol conditions (dust, smoke, and low aerosol loading). Second, the same cases are analyzed using a pixel-to-pixel differencing of surface reflectance. The spectral distributions of the means for the resulting difference fields are constructed. Each aerosol type was found to exhibit a relatively unique spectral distribution for both methods. Each background was found to exhibit a characteristic amount of contrast in the absence of heavy aerosol loading. The unique spectral characteristics for each aerosol- background combination may be correlated to aerosol optical depths or surface visibilities with corrections for sensor view angle variations, Rayleigh scattering, and masking of clouds and surface water. The spectral distribution-aerosol optical depth correlation can be used to build an empirical model for aerosol optical depth and surface visibility retrievals from satellite observations. This method may be applied to multi-spectral or panchromatic imagery, unlike current aerosol optical depth retrievals over land.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 2003
- Accession Number
- ADA418306
Entities
People
- Dominick A. Vincent
Organizations
- Naval Postgraduate School