Abstract
A numerical model of radiation fog was developed in order to test the sensitivity of variables comprising the model, and evaluate its capability for forecasting the onset of fog from standard radiosonde weather data. Four case studies were considered that included both fog and no-fog occurrences. The variables examined–initial surface temperature and moisture conditions, eddy exchange profiles, radiative flux divergence, and dew formation–were all found to influence critically the model's performance. Prediction of fog occurrence and temperature were reasonably encouraging provided a judicious (though somewhat arbitrary) choice of eddy mixing values was made.
