Résumé
► We developed statistical models to predict the distribution of boats and manatees. ► We derived an index of risk of co-occurrence. ► Accounting for imperfect detection of manatees was important. ► We evaluated the influence of key covariates on manatee and boat distribution. ► Estimating manatee–boat co-occurrence is useful for the optimal design of speed zones.
Collisions between wildlife and vehicles represent a large source of mortality for many species. To implement effective protection zones, it is important to identify areas in which wildlife–vehicle collisions are likely to occur. We used statistical models to derive an index of risk of co-occurrence between manatees and boats. Our statistical models were used to predict the distribution of both manatees and boats, while accounting for observer-specific detection probabilities. Models used aerial survey data and we found that both environmental and temporal covariates influenced manatee and boat distributions. Moreover, the probability of detecting manatees varied substantially with the weather and among observers. To our knowledge, this is the first time that manatee distribution is modeled as a function of key environmental and seasonal covariates, while accounting for imperfect detection of manatees. We computed an index of risk of co-occurrence by multiplying the probability of manatee occupancy by the expected boat density and occupancy to identify areas where manatee–boat collisions are likely to occur. This analytical framework emphasizes the importance of accounting for imperfect detection, and how modeling distribution of both organisms and vehicles as a function of key covariates can help improve predictions of risk of collisions. Risk of collision metrics can then be used in designing protection zones.