A rule-based method for diagnosing radiation fog in an arid region from NWP forecasts

Abstract

A reliable fog forecast can assist in water resource management and mitigate against hazardous effects on transport networks. However, diagnosing fog from numerical weather prediction models remains a challenge and various methods have been developed to achieve this. One approach is to use a multi rule-based method, where threshold values of multiple variables are employed to define fog conditions in the forecast. Such a method is presented here for next day forecasts of radiation fog at Abu Dhabi in the United Arab Emirates. Variables tested include the dew point depression, wind speed, modified Richardson index, inversion depth, relative humidity and liquid water content. Of these, the best performing rule included relative humidity (≥94%), wind speed (≤3ms−1) and inversion depth (>250 m) with a probability of detection (POD) (false alarm ratio (FAR)) of 0.62 (0.37). The worst performing rule was liquid water content with a POD (FAR) of 0.37 (0.31). Model performance was more sensitive to inversion depth than wind speed, and rules that included inversion depth performed better than rules that included the modified Richardson number. A qualitative spatial analysis showed that rules produced realistic fog patches when compared to satellite images. This suggests that the geophysical parameters chosen to define rules are well suited for defining fog as long as the model bias is sufficiently small to fall within the threshold values. This method worked well for the study region and did not demonstrate any seasonal bias.