A satellite based fog study of the Korean Peninsula

Abstract

Fog has always been a difficult phenomenon to forecast. Its unpredictable nature and propensity to quickly decrease visibilities have had adverse effects on military operations for many years across the Korean peninsula. It is particularly difficult to prepare forecasts or plan operations for remote locations with limited fog detection ability. For detection at night, over large areas, and in remote locations, satellite observations are the best solution. This thesis used NASA MODIS satellite imagery to create an abbreviated climatology data set for remote areas across the Korean peninsula. Imagery from the Terra and Aqua near-polar orbiting satellites was used, providing four images per day: one daytime and one nighttime pass for each satellite. Two decision trees were developed to use as guidelines for fog detection by daytime and nighttime satellite images. It was not always possible to unambiguously determine if fog was in each scene, so various categories were created to supplement a fog or no fog decision. The four midseason months (October 2005, January 2006, April 2006, and July 2006), were analyzed to create a climatology database. The results are tabulated using different variables to make useful comparisons, like day-versus-night or Terra-versus-Aqua. The new totals are compared visually with bar charts and statistically to identify trends that might give insight to planners and forecasters. Seasonal and nocturnal patterns are very evident while differences between the results from two satellites are less obvious. Future work is needed to expand the climatology and increase the statistical results from this study.