Many industrial ecology models may be classified as single-use products-constructed to address a specific problem and of little use beyond their original context. But a knowledge ecosystem based around single-use representations of reality wastes effort and resources and limits the impact of industrial ecology as a field. In recent years, tangible progress has been made in areas such as integrated modeling and multisimulation. However, this work tends to discount the evolutionary nature of models and their embeddedness within a changing sociotechnical environment-aspects which, we argue, are central to enabling the sustained usefulness of models in industrial ecology.We define a multimodel ecology as an interacting group of models coevolving with one another in a dynamic sociotechnical environment. A multimodel ecology perspective can facilitate model integration and reuse and highlight different ways in which models-mental, conceptual, and computational-may interact and evolve. To demonstrate the use of this perspective, we introduce and analyze an existing multimodel ecology-the Energy Modeling Laboratory. We conclude with a set of guidelines for facilitating model reuse and integration. Among others, these include use open standards, build simple components, leverage the Web, borrow proudly, and enforce sharing.