Acceptance Criteria for Using Individual‐Based Models to Make Management Decisions

Abstract

Individual‐based population models are beginning to have substantial influence in conservation biology. One of the dangers of such models is that their reliability may be poorly understood, and generally over‐estimated. To help avoid this problem a set of guidelines is suggested here for evaluating individual‐based models. Major components of the evaluation include a description of the model and estimates of the reliability of its predictions. The description should include detailed explanations of the model's purpose(s), its structure, and the assumptions it makes. Analyses of reliability at four levels may be useful: structural assumptions, parameter values, secondary predictions of the model, and primary predictions of the model. The guidelines also recommend that "best‐" and "worst‐case" scenarios, intended to span the range of plausible outcomes, be prepared using the model, rather than presenting single predictions or conclusions. I argue that results from such an evaluation should be prepared and peer reviewed before the model is used to make or defend management decisions.