Analysis and prediction of hazard risks caused by tropical cyclones in Southern China with fuzzy mathematical and grey models

A hazard-risk assessment model and a grey hazard-year prediction model (GHYPM) are constructed by integrating recent advances in the fuzzy mathematics, grey theory and information spread technique, and then applied to 17-year tropical cyclones (TCs) hazards in Southern China. In constructing the models, a genetic fuzzy mathematical algorithm is first developed to calculate the categorical and ranking weights of TC hazard impact and cause indicators, from which their combined weights are obtained after optimization. The hazard impact and cause index series are then found by coupling the combined weights with their corresponding down-scaled indicators. A two-dimensional normal-spread technique is employed to create a primitive information matrix and a fuzzy relation matrix in order to make fuzzy rough inference of hazard risks with the factorial space theory. An exceeded probability model is developed to assess the possibility of exceeding any given hazard-year category. Results from the GHYPM show that the simulated hazard risk values are more or less consistent with the hazard-impact index series, with more than 60% probability of exceeding a moderate hazard year in Southern China. Results also show small relative errors of the GHYPM, indicating its applicability to the prediction of TC hazard-years up to 20 years.