Problems, issues, potential solutions (continued)
     A provisional radiance calibrated nighttime lights product was produced using DMSP-OLS data acquired at three gain settings acquired during the winter months of 1999-2000 under low lunar illumination. We noticed areas with anomalously bright and extensive lighting in this initial product. Upon investigation we determined that when the gain setting is high, the OLS lights from the outer quarter panels are slightly brighter and more extensive than the lights from the inner half of each swath. This is due to automatic switching of the gain in the outer quarter panels built into the OLS for the purpose of "constant contrast" cloud imaging. Another factor which appears to result in brighter and more extensive lights is the presence of snow cover.

     Our solution was to construct a new radiance calibrated cloud-free composite of the nighttime lights the USA using data acquired during September, October, and November of 2001. These months wer selected to avoid solar contamination present in OLS low light imaging data acquired mid-summer at mid-to-high latitudes and snow cover which may be present during winter months. The 2001 lights were processed only using the inner half of the swaths when the OLS gain was set high. This eliminated the edge effects found in the lights present in the outer quarter panels, provided for improved geolocation accuracy and sharper lights.
Back  |  Title Page  |  Next