Spatial Seasonal Forecasting of Tropical Cyclone Occurrences Using Sea Surface Temperature and Latent Heat Flux

• Activity Number: 91
• Type: Invited
• Date/Time: Sunday, August 9, 2015 : 9:30 PM to 10:15 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Marcela Alfaro-Córdoba* and Montserrat Fuentes and Joseph Guinness and Lian Xie
• Companies: North Carolina State University and North Carolina State University and North Carolina State University and North Carolina State University
• Keywords: Forecast ; Tropical Cyclone ; Spatial ; Temporal

Abstract

We present a statistical framework to forecast the probability of occurrence and the category of a tropical cyclone at any spatial location in the Atlantic Basin. The different categories of tropical cyclones are modeled simultaneously across space and time as a function of sea surface temperatures and latent heat fluxes. A spatial model with rank reduction techniques was used to model dependence between the proportion of tropical cyclones for each response, time and location and the covariates for the same time and location. We integrate this approach in a hierarchical model to estimate both the total number of storms per year and the probability of occurrence.