Agent-based modelling to improve management strategy evaluation in complex problems of biodiversity conservation

Abstract

Amid the Earth’s 6th mass extinction, conservation of the remaining biodiversity is more urgent than ever. But conservation faces the many challenges associated with the management of ecosystems, including the different sources of uncertainty around their joint social and ecological dynamics. Moreover, mismanagement can have detrimental consequences for both conservation and people’s livelihood. To help managers design efficient and equitable policies in such complex situations, the use of simulation models as virtual laboratories to evaluate management scenarios have become widespread. Yet, most of these models are mathematical models, which can be limited in conservation problems. Indeed, the latter are rather small-scale, often spatially explicit, and unexpected responses often stem from local interactions between the parts of the system. These features are key assets of agent-based modelling (ABM), which should therefore be more appropriate but is still scarcely applied in management strategy evaluation. In this thesis, I further develop the use of ABM in conservation by designing agent-based tools to evaluate alternative management strategies in two complex conservation problems: the management of conservation conflicts between species protection and agriculture, and the management of species endangered by apparent competition. First, I address the timing of management intervention in a conservation conflict between a manager aiming to conserve an animal population and farmers aiming to maximize yield by protecting their crop from consumption by the animal population. Building upon the agent-based software GMSE, which simulates the budget-constrained adaptive management of conservation conflicts, I propose a novel management strategy that dynamically alternates between intervention and waiting based on the monitoring of the distance between the population density and manager’s target. The evaluation shows that my strategy can produce at least as efficient and equitable results as unconditional intervention while allowing critical budget savings. This strategy is now available in the GMSE package to be evaluated in other cases of conservation conflicts and my method proposes a way to do so in a more equitable way. Then, I introduce an ABM of trophic interactions between several species in several trophic layers in which apparent competition can emerge. I validate it in its essential version according to the ABM framework and discuss its adequacy with apparent competition theory. Overall, my model behaves as theory and empirical cases predict, with some interesting contradictions challenging predictions of previous mathematical models. The model is now fit for the addition of more complex features needing further understanding in apparent competition theory and conservation. Finally, I demonstrate how my model can be used to evaluate removal strategies to conserve a prey species endangered by apparent competition with an alternate prey species mediated by a shared predator. Predator removal only, alternate prey removal only, and simultaneous removal of the predator and alternate prey all successfully prevented the extinction of the endangered population, but the simultaneous removal strategy was the most efficient in conserving the endangered population while also ensuring more stable dynamics and higher densities of the other species. Any combinations of these removal strategies are now available in my model for evaluation by researchers or managers in other cases, and my method proposes to conserve the endangered population while also ensuring the persistence of the other species in the system. Across these chapters, I demonstrate how the agent-based approach can efficiently integrate the complexity of conservation problems to produce informative tools for biodiversity management strategy evaluation. Notably through human individual decision-making, complex individual behaviour, and spatially explicit modelling along with the simulation of different sources of uncertainty. I have shown, with general examples, how to use the modelling tools I provide, with methods enlarging the scope of conservation objectives: a better consideration of the equity of management measures between conservation and land-users’ livelihood in conservation conflicts, and a better inclusion of management consequences on other species in interaction with the population of conservation interest. Through my models’ development and validation, I have questioned and enhanced theory of conservation conflicts' adaptive management and and of apparent competition’s underlying mechanisms and management.