Human activities are increasing in intensity, duration, and spatial extent. As a result, humans play an increasingly influential role in nearly all of the planet’s ecosystems. In order to maintain ecosystem function alongside growing human impacts, we must first understand the nature and drivers of these impacts, and, just as importantly, understand how to manage them. Behavior is a critical consideration in addressing this two-part challenge. Animal behaviors that underlie ecosystem function can be affected by human activities, and human behaviors determine human impacts and management outcomes. This thesis integrates ecological and social approaches to improve our understanding of both animal and human behavior in the context of ecosystem management, with a particular focus on Caribbean coral reef systems. I begin with a comprehensive assessment of the potential pathways in which human activities may influence ecosystem outcomes through impacts on animal behavior across aquatic and terrestrial ecosystems. While such impacts have been increasingly documented, the implications of these behavior changes for ecosystem function remain understudied. Through synthesis of empirical evidence and ecological theory, I find that human-induced animal behavior change has the potential to alter ecosystems through numerous pathways. Whether or not these pathways are realized will depend on several factors, primarily the nature and distribution of human activities and the ecological role of an altered species’ behavior. To better understand potential variations in ecologically critical animal behaviors, in my second chapter I examine the feeding behaviors of herbivorous fish in coral reef ecosystems experiencing different human disturbance. Herbivory on coral reefs is a keystone driver of ecosystem health through facilitation of coral survival, growth, and recruitment by suppressing otherwise dominant algal competitors. Across thirteen reef sites in three Caribbean islands with varying levels of coral reef health and anthropogenic impacts, I find that the grazing behavior of herbivorous fish varies significantly. Feeding rates, or the number of bites taken per minute, varied as much as nine-fold across sites, while the grazing intensity, or number of consecutive bites taken while feeding, varied as much as six-fold. I present several hypotheses for these trends and discuss theoretical links to human impacts such as fishing that can guide future studies. In my third chapter, I evaluate the outcomes of efforts to shift human fishing behavior off of coral reefs and onto less exploited pelagic resources through a case study of the Guadeloupe fishery, a mixed-gear small-scale fishery in the Caribbean. I find that these efforts, namely the introduction of moored fish aggregating devices (MFADs), have allowed more vessels to specialize in exclusively offshore fishing and likely facilitated an increasing proportion of landings coming from pelagic species, but that they have also made offshore fishing less accessible to certain portions of the fleet. These findings suggest the importance of understanding human behavior in modifying human impacts on threatened ecosystems.