An experimental approach to assess the combined effects of multiple stressors on a large vertebrate species.

The design of experiments to investigate the combined effects of multiple stressors requires exposing target organisms to multiple combinations of stressor doses. Concurrent manipulation of stressors is often infeasible with wildlife, but long-lasting health effects allow individual health to be used as an integrator of prior stressor exposure. The population of bottlenose dolphins in Barataria Bay, Louisiana, experienced long-lasting health effects after exposure to oil from the Deepwater Horizon spill. We investigated whether compromised health status affects dolphins' ability to respond to other stressors, specifically vessel traffic, potentially leading to increased collision risk. We used this case study to develop a hypothesis-driven experimental approach to assess the combined effects of multiple stressors in a large, long-lived vertebrate species. We conducted controlled vessel approaches to test whether the health status of targeted dolphins affected their behavioral responses. Our results highlighted some effects of health, suggesting that oil spills may exacerbate the effects of other stressors in coastal populations. For example, lung disease was associated with a delayed dive response, which could affect the ability to vertically avoid vessels and lead to increased collision risk. However, health effects on response probability were overall smaller than anticipated, and other contextual variables (e.g., sex, age, calf presence, prior experience, and exposure context) contributed to response variability. Our work demonstrated the value of formalizing stressor interactions as multidimensional dose-response functions and showed the feasibility of an experimental, multiple-stressor study in a wild system in which individual health status can be used as an integrator of prior stressor exposure. This approach has broad implications for other species that are difficult to handle experimentally. The quantification and management of the cumulative risk from multiple stressors on wildlife will require a combination of empirical and mechanistic approaches to inform long-term, population consequences.