Linking economic models to physical substance-flow models. Possibilities and problems