Decision-support systems: it's the question not the model

Abstract

Decision-support systems (DSSs) do not make decisions, but aid decision making. They encompass a range of techniques from statistics (e.g., Bayesian statistics, classification and regression tree analysis), optimization (e.g., linear programming, simulated annealing), and a variety of simulation techniques and tools. All of them are models because they simplify and abstract reality. Relations between forest dynamics and ungulate herbivory are equally diverse, including influences on tree population dynamics, impacts on other components of biological diversity, and nutrient cycles. The complexity of these relations hinders easy analysis and encourages the use of DSSs to query ungulate-forest systems. Our broad purpose is to suggest ways of matching kinds of DSSs with the variety of questions. We first note the diversity of relations among ungulates and other forest components that stimulates questions about how best to manage forests and forest-dwelling organisms. Then we briefly review DSSs in the broad sense. We recognize that various goals encourage the search for answers and ways of communicating partial answers, and address these under the “purpose” of the DSS. We summarize linkages between questions, purposes, and models, and close by considering use of and errors in DSSs.