Integrated land use and environmental models: a survey of current applications

We are all witnesses to the second wave of socioeconomic modeling. The frustration aroused by thècomprehensive' black-box models of the 1980s is over; cellular automata and multiagent systems have come to the fore. After the retreat of the 1980s and early 1990s, the geographic community is again in favor of modeling, with the number of conferences and journal publications growing constantly. The emerging approaches and tools of environmental modeling are the focus of this book, a compendium of papers presented at a symposium convened at Arizona State University in 2000 to survey the progress made in integrated land-use and environmental modeling. The book's four sections consider specific aspects of this trend. The first section, ``Evolving definitions öchanging practices'', begins with``Advances in urban and environmental modeling: surveying the terrain and demarcating frontiers'', contributed by the book's editor, Subhrajit Guhathakurta. The paper offers a general evaluation whilst also claiming that land-use and environmental models are still overly biased towards economic phenomena, a trend constraining incorporation of ecological components that are often weakly if ever related to economics. Aggregation, another sinful tradition of economic modeling, continues to endure despite the demise of the black-box era. Guhathakurta claims that the future lies in high-resolution representation and calls for the development of agent-based`computer interfaces' to integrate better our knowledge of urban and environmental processes. I certainly agree with his argument that agent-based or, more generally, object-based modeling presents an improved way of articulating the nature of environmental systems and the principles of general systems theory in its modern guise. However, I would prefer to wait for additional successful examples of this school of modeling before I am totally convinced of its supremacy. The author is aware that new tools do not necessarily obliterate old problems. He therefore discusses the uncertainty still attached to environmental knowledge and social rules in addition to the inability to construct models with the rigidity of natural science or that overcome problems of temporal and spatial scaling. Guhathakurta also calls for proper definition of a model hierarchy and criteria for judging model performance. His response to these dilemmas is captured in his proposal to view models as tools for constructing a narrative öa meaningful story about the processöan idea that I like very much. The still popular format for communicating a modelöits description accompanied by illustrative simulation runsöis clearly inadequate for this purpose; to construct a narrative you need a thorough …