| This thesis discusses the developmental processes and issues involved in building a dynamic behavioral model in the NetLogo modeling environment for the Chinese mitten crab (Eriocheir sinensis) within the San Francisco Estuary. Four attributes are studied, which include spatially explicit estimates of temperature, salinity, photosynthetic activity, and sedimentation, to provide insight into the potential behavior of these organisms. Temperature and salinity are updated monthly and cycled over the course of a typical year. Data gathered from sampling stations in the estuary were spatially interpolated to provide continuous gridded representations for the GIS and modeling software. The crabs are individually represented in their development stages. The model was verified for internal consistency, and many simulations were performed under valid conditions, yielding results generally consistent with the potential behavior of these organisms. Observations were made on the basis of sensitivity analysis, investigating a range of environmental and behavioral conditions within the model. Preliminary results collected using the BehaviorSpace tool indicate stable crab populations would survive in the estuary under an array of conditions. BehaviorSpace cannot emulate human pattern recognition, so geographic analysis was performed through dynamic simulations during different times of the year. Locations where simulated crab clusters occur may provide guidance for their detection and eradication. These results indicate that similar agent-based models offer practical geographical guidance for ecosystem-related problems. |
