Title: Bayesian modeling of the frequency of tropical storms
Authors: Joyee Ghosh - The University of Iowa (United States) [presenting]
Abstract: Knowing the frequency of tropical storms in advance can help in improved preparedness before a hurricane season. It is known from the existing literature that sea surface temperatures during the peak hurricane season are good predictors of tropical storm activity. However, sea surface temperatures are available after the end of the hurricane season, and thus cannot be directly used for prediction. Instead, forecasts of sea surface temperatures are available from multiple climate models. A Bayesian model averaging approach is developed that combines forecasts from multiple climate models. Based on simulation studies and North Atlantic tropical cyclone activity data, it is illustrated that this model can provide improved predictive performance compared to some existing models in the literature.