Ciguatera poisoning trends in Florida using a predictive hybrid model

Abstract

The aim of this study is to identify an optimal predictive model for ciguatera poisoning and to determine which variables and time lags best explain the number of cases reported in Florida between 2006 and 2020. Ciguatera poisoning is a debilitating condition caused by consuming coral reef fish contaminated with ciguatoxins, which originate from toxin-producing dinoflagellates and biomagnify through the food chain. In severe cases, the illness can be fatal. Global climate change is expected to increase both the incidence of ciguatera poisoning and its geographic range, extending from tropical and subtropical reefs into temperate regions. This makes understanding and predicting its dynamics particularly urgent, as millions of people worldwide depend on reef fish as a dietary staple. To address this need, we developed an integrated approach combining wavelet cross-coherence analysis with a count modeling framework. Candidate predictors included cumulative monthly landings of Amberjack, Red Snapper, Red Grouper, and Scamp Grouper; the number of tropical cyclones; maximum ocean temperatures; precipitation; season; and Florida’s human population. The optimal model identified was a Zero-Inflated Negative Binomial model. Results showed positive associations between ciguatera cases and (i) maximum ocean temperatures, (ii) storm frequency, (iii) fish landings, and (iv) human population size, alongside a negative relationship with precipitation. By establishing a robust predictive framework, this study advances understanding of the environmental and anthropogenic drivers of ciguatera poisoning. The findings provide a foundation for forecasting outbreaks and offer actionable insights to fisheries and public health agencies aiming to reduce risks for Florida residents and tourists.