A Diagnostic Method for Fog Forecasting Using Numerical Weather Prediction (NWP) Model Outputs

Abstract

An attempt has been made in the present study to forecast fog with adiagnostic method using the outputs of global NWP model. The diagnosticmethod is based on the combination of thresholds of meteorologicalvariables involved in fog formation. The thresholds are computed usingthe observations during fog. These thresholds are applied to the output of aglobal NWP model for forecasting fog. The occurrence of fog is a commonphenomenon during winter season over the northern plains of India. Thediagnostic method is used to predict fog occurrences over three stations innorth India. The proposed method is able to predict both occurrences andnon-occurrences of fog at all the three stations. It is found that 94% of thefog events forecasted by the model using the diagnostic method have beenactually observed at the selected stations. The performance of method inpredicting fog is found best over Delhi with the highest accuracy (0.61) andprobability of detection (0.60). The study signifies that diagnostic approachbased on the output of a global model is a useful tool for predicting fogover a single location.