Modeling freezing height level for satellite link communication: Durban, South Africa

In the design process of Earth-space link telecommunication systems at higher frequencies, the precise estimation of attenuation due to rain which depends on rain height is important. This study examines two statistical distribution models for freezing height level over Durban, South Africa, using 5 years data and then studied the monthly, seasonal and annual variations using the parameters estimated by the maximum likelihood of the distributions. Three statistical goodness of fit were also used to determine the performance evaluation of the distributions. It was observed that 3-Parameter Dagum (Dagum) distribution performs better over the next most suitable distribution (3-Parameter Log-logistic) for the modeling of freezing height level over Durban with improvement of 0.94% in Bias error, 0.48% Root Mean Square Error (RMSE) and 7.43% Kolmogorov-Smirnov test (K-S test). Also, the location, shape and scale parameters for the distributions show a wider variation.