In this study, fogs are classified based on the spatial and temporal characteristics over South Korea using the visibility data and the empirical orthogonal function (EOF) and wavelet analyses. With fog defined in terms of visibility (<1 km), the EOF analysis is performed to extract spatial distribution characteristics via dimension reduction, whereas the space-time wavelet expansion is applied to the EOF time series to specify the fog characteristics in the space of time versus scale (i.e., period in this study). The first EOF mode occupies 48.9% of total variance and shows the fog distribution covering almost entire areas of South Korea with one sign (+), except at the eastern coast and western part of the southern coast. The wavelet analysis reveals that this fog occurs based on meteorological conditions of various scales from daily to seasonal, thus classified as mixed fog. The second EOF mode, which occupies 19.5% of total variance, shows distinct separation of spatial distribution of fog, with a negative (-) sign in winter over northwestern coastal/inland, western coastal, and south central mountain areas of South Korea and a positive (+) sign in other seasons elsewhere. With cycles of 1-2 weeks and 1-2 months being dominant in the wavelet analysis, this fog is considered to be strongly affected by synoptic scale weather systems and monsoon. Fog over the positive area is mostly affected by monsoon and/or cyclonic frontal systems, thus classified as frontal fog, whereas that over the negative area is affected by the cold-core anticyclones moving over warm sea surface in winter or by radiative cooling, thus classified as steam fog (coastal/sea) or radiation fog (inland), respectively. The mountain area may have upslope fog because of orographic lifting. The third EOF mode, occupying 6.7% of total variance, depicts distinct spatial separation of fog distribution around the coastal areas with a negative (-) sign and in the inland areas with a positive (+) sign. The former, with a dominant 1-2 week cycle, is classified as sea fog affected by migratory anticyclones and monsoon in late spring and summer, while the latter, with a dominant diurnal variation, represents radiation fog under clear sky in autumn. It turns out that the combined EOF and wavelet analyses are useful to assess the detailed spatial and temporal characteristics of various types of fog occurrence in South Korea.