ABSTRACT
Many systems, including those studied in the agricultural sciences and biogeography, are inherently complex and require the use of many different spatial analysis and modeling techniques to understand underlying spatial patterns and processes. In this context, the methods encompassed by the term “geocomputation” represent a rapidly evolving, and increasingly effective, suite of tools to both analyze and visualize problems that cannot be addressed adequately with the use of standard aspatial or spatial statistics. Geocomputation methods include agent-based models, genetic algorithms, cellular automata, and neural networks. Agent-based models, in particular, are increasingly being used to examine a variety of questions involving spatiotemporal dynamics (e.g., changes in landuse/landcover, urban population growth) and the temporal paths of mobile entities (e.g., vehicular trafŽc analysis, pedestrian ¬ows, disease dispersal, animal movements). Presented here is the rationale behind, and instructions for, constructing a simple agentbased model that can be used to form or revise theory regarding the early season infestation of grain storage bins by a small (3-5 mm) beetle called the lesser grain borer (LGB) (Rhyzopertha dominica). LGB is a devastating, long-lived pest of stored cereal grains worldwide and the most damaging pest of stored wheat in the United States. Extending our understanding of LGB populations to include an explicit spatial component via agentbased simulation provides an opportunity to better evaluate the connectivity of grain storage bins on neighboring farms and more fully appreciate the role of landuse, landcover, and landscape conŽguration in shaping insect source-sink dynamics.
