Fog Forecast for the Kimpo International Airport of Korea

It is difficult to quantitatively predict fog occurrence, duration time, and its intensity despite of improvements in numerical weather prediction for the aviation weather service. The reason is that the fog phenomenon is dependent on complex system such as microphysical and mesoscale processes in the planetary boundary layer that are forced by prevailing synoptic regime. Until now, techniques dependent on the statistical method are only available to operational fog forecast.

In this paper we will present the characteristic features of fog occurrence and introduce to application results of the classification and regression tree(CART) algorithm to forecast the fog occurrence for the Kimpo International Airport of Korea. In the CART algorithm, potential predictors consist of meteorological observation data from a network of surface and upper weather station during recent 10 years from 1990 to 1999. From the Automatic Meteorological Observation System(AMOS) data observed by Kimpo Aviation Weather Office, it was found that the fog occurrence of the Kimpo International Airport is closely related with nocturnal mixing and local circulation, efficient moisture supply from adjacent water reservoirs such as Yellow Sea and Han River, and the visibility of previous day.

MM5 model output which is recently being operated by Korea Meteorological Administration(KMA) is also used to investigate a favorable weather condition for fog occurrence which is a synoptic weather pattern, a moisture transport from adjacent water reservoirs, nocturnal stability and evolution in the planetary boundary layer, and radiative cooling by outgoing longwave radiation during night.