Climatological Tools for Low Visibility Forecasting

Abstract

Forecasters need climatological forecasting tools because of limitations of numerical weather prediction models. In this article, using Finnish SYNOP observations and ERA-40 model reanalysis data, low visibility cases are studied using subjective and objective analysis techniques. For the objective analysis, we used an AutoClass clustering algorithm, concentrating on three Finnish airports, namely, the Rovaniemi in northern Finland, Kauhava in western Finland, and Maarianhamina in southwest Finland. These airports represent different climatological conditions. Results suggested that combining of subjective analysis with an objective analysis, e.g., clustering algorithms such as the AutoClass method, can be used to construct climatological guides for forecasters. Some higher level subjective “meta-clustering” was used to make the results physically more reasonable and easier to interpret by the forecasters.