Abstract: Statistical equations for sea-fog forecast were established using the Yangshan automatic weather station data, sea surface temperature of reanalysis data, and buoy temperature data over Yangshan Sea area in recent years. Many elements, including wind direction, wind speed, air temperature, temperature difference between air and sea surface, and relative humidity were selected as factors for sea-fog forecast. For establishing the forecast equations, firstly each of the forecast factors was divided into several ranges, which were then crossly combined to form a number of groups. Each group corresponds to a kind of conditions for generating and sustaining sea fog. Secondly, based on their historical sample groups, visibility forecast equations were established. This method can overcome the weakness that traditional linear regression equations can’t describe the nonlinear relationship between forecast factors and visibility, and the weakness that critical thresholds must be set in traditional regression equations. In this study, the sea fog in different conditions can be forecasted by more applicable equations, which have physical meanings. The results show that using this method good forecast can be acquired for sea-fog events with sufficient historical samples. With accumulation of sea-fog samples in different conditions, sea-fog forecast can be further improved.

Key words: sea fog, similarity analysis, objective forecast