Assessing population-level effects of anthropogenic disturbance on a marine mammal population

Abstract

The Population Consequences of Disturbance (PCoD) model is a conceptual framework used to assess the potential for population-level consequences following exposure of animals to a disturbance activity or stressor. This framework is a four-step process, progressing from changes in individual behavior and/or physiology, to changes in individual health, then vital rates, and finally to population-level effects. Despite its simplicity, there are few complete PCoD models available for any marine mammal species due to a lack of data available to parameterize many of the steps. Here, we present an application of the PCoD framework for migrating humpback whales exposed to a simulated commercial seismic survey scenario. We approached the framework in two ways; first, progressing sequentially forwards through the steps and basing our assessment on lactating females. This cohort was considered to be the most vulnerable in terms of energetic costs of disturbance, and most likely to influence any change in population growth due to future breeding success. Field measurements of behavioral responses of migrating humpback whales to seismic air guns from a previous study were used to parameterize an agent-based model (ABM). This ABM was used to estimate the probability of response, where a response was defined as a change in the migratory movement of female-calf pairs, and the duration of any resulting delay in migration. We then estimated the energetic consequences of any delay in migration for the lactating females and created population growth models with which to assess any population-level effects. The results of the forwards approach suggested a low potential for population consequences of seismic surveys on migrating humpbacks. Working backwards through the framework, we investigated ‘worst case’ scenario(s) that could potentially lead to a population-level effect. Here, we started with increasing calf mortality and assumed that an exposure time greater than 48 hours would increase mortality risk. We determined the most likely context in which this exposure would occur (resting area) and then tested this context within an ABM. This backwards approach illustrates how the PCoD model can be used to make management decisions regarding animal populations and exposure to anthropogenic stressors.