An agent-based model for the sustainable management of navigation activities in the Saint Lawrence Estuary

Natural resource managers of protected areas are concerned with the management of human activities potentially harmful to ecosystems’ health and/or integrity. These systems where human interact with natural resources are called social-ecological systems (SES) and possess the characteristics of complex adaptive systems (e.g. co-evolution). The SES of navigation activities and whales interacting within the Saguenay–St. Lawrence Marine Park (SSLMP) and the projected St. Lawrence Estuary Marine Protected Area in Quebec, Canada, has been investigated and modelled using the agent-based modelling (ABM) technology: The resulting Marine Mammal and Maritime Traffic Simulator (3MTSim) is designed to support marine protected area managers in their effort to reduce the frequency and intensity of boat-whale co-occurrences within the St Lawrence Estuary and mitigate the risks of vessel strikes. This dissertation presents the building process of the 3MTSim’s boat ABM. The knowledge extracted from analyses of gathered and collected data relative to all forms of sailing and motorized navigation supported the decision to first focus on the modelling of commercial excursions (including whale-watching trips), cargo ships, and cruise liners. Data analyses allowed, for the first time, to draw a comprehensive portrait of navigation activities throughout the region where whales congregate in great numbers during the summer season. Among others, a quantitative analysis led to an accurate estimate of the total navigation time within each separate ecosystem of the region. This study identified areas intensively used by maritime traffic such as the mouth of the Saguenay River and offshore Les Escoumins. Several field campaigns carried out in the context of this project allowed to link some undesirable collective patterns of whale-watching excursions (regarding both whale conservation and SSLMP visitors’ experience) with contextual factors including whale species’ abundance and distribution, management gaps, and companies and captains’ decisions. The bounded rationality framework was chosen to investigate captains’ decision making and more generally the dynamics of the whole whale-watching SES. A portrait of the decision strategies followed by whale-watching captains has been drawn. The results will lead to a set of recommendations regarding the sustainable management of whale-watching excursions in and around the SSLMP. Results from field investigations and data analyses have fed the model building process, including an explicit representation of the whale-watching captains’ decision making. Data analyses revealed that cargo ships and ocean liners tend to follow predictable routes with low variability. Consequently, a complex behavioural modelling approach was deemed unnecessary in favour of a statistical approach, justified by the large volume of high-quality historical data available for both components. The pattern-oriented modelling approach proved appropriate for selecting a valid model of whale-watching excursions. Model simulations confirmed that whale-watching captains do favour the observation of a few rare rorqual species (e.g. humpback whales), leaving aside the most abundant one, namely the minke whales. Therefore, 3MTSim was run to quantify the impact that whale-watching captains changing their decision strategy could have on both whale exposure to boats (conservation concern) and excursion content (commercial concern). It was found that captains willing to avoid crowded observation sites and/or seeking to increase the diversity of species observed could have statistically significant benefits regarding conservation issues without affecting important features of their excursions. Finally, the convincing performance of the 3MTSim’s boat ABM ensures its safe use as a decision-support tool for management insofar as model limitations are understood and accounted for in the results and discussion.