Effects of a hurricane on two assemblages of coral reef fishes: Multiple-year analysis reverses a false snapshot' interpretation

The few studies of hurricane impacts on coral reef fishes are generally before vs after snapshots'. Category 2 hurricane Marilyn passed over St. Croix, United States Virgin Islands, in September 1995, amid a multiple-year monitoring project, with stationary point visual fish censuses conducted at monthly intervals. Monitoring had been ongoing for 17 mo prior to, resumed 1 mo after, and continued for 9 mo after the hurricane. Following previous published studies, initial analyses compared pre- and post-hurricane species richness, total fish abundance, and abundance of three selected species. These comparisons indicated a hurricane effect, as all measured variables were lower post-storm. The 'before vs after snapshot' approach provided a false finding of a hurricane effect. Subsequent analyses showed these results were misleading, as linear regression analysis and loess smoothing revealed a decline for all variables over time. Most importantly, this decline was unaffected by the hurricane. Evidently, the impacts of events such as hurricanes can be properly understood only when multiple year trends are included in analyses, which requires continuous monitoring rather than sporadic or single-time snapshots.