Using Economic Information to Anticipate Collapses in Social-Ecological Systems

Social-ecological systems typically undergo gradual transformations, but may also suddenly change their qualitative behavior. Ecological examples include fish stocks switching to a less productive regime or sudden losses in biodiversity due to habitat degradation. Social examples include small communities switching from a cooperative sustainable harvesting regime to an unsustainable one, where each community member exploits the resource to grab as many resource rents as fast as possible. We develop a complex social-ecological model, in which individual agents face the temptation to overexploit the resource, while a cooperative harvesting norm spreads through the community via interpersonal relations. A catastrophic transitions from relatively high levels of cooperation to widespread norm violation can be induced by biological or climatic forces, such as loss carrying of capacity and decrease in intrinsic growth rate, but also by socioeconomic driving forces, such as an increase in the number of harvesting agents, or higher harvesting efficiency. While social collapse results in wasteful overexploitation, it may also have irreversible consequences and trigger also a collapse of the ecological system. Anticipating such collapse is still a major challenge, especially with incomplete biological information, such as biomass estimates from stock assessments. Here we show that collapse can be anticipated by investigating only socio-economic response variables, even if biological information is absent. Even with full social-ecological complexity, early warning signals derived from economic information, such as profits, can foretell a resource collapse in our model. However, we also find that any early warning signs that are decreasing in severity may simple document the attempts undertaken by some individuals to save the system just before it collapses, rather than a restoration of the ecological system itself. By overlooking these social complexities, one may falsely conclude that the system is recovering, while a resource collapse just lures around the corner.