Coupling environmental models and geospatial data processing

This research investigated geospatial functions for solving environmental problems from the perspective of the environmental modeler. Its purpose is to better understand the different approaches to coupling complex models and geospatial data processing, plus the implications for the coupled system. To this end, various coupling methodologies were systematically explored using a geographic information system (GIS) and an emissions processor (SMOKE) for air quality models (AQMs).; SMOKE converts an emissions inventory into the format required by an AQM. A GIS creates a file describing the spatial distribution of emissions among the cells in a modeling domain. To demonstrate advantages of a coupled GIS—environmental model system, two methods of spatially distributing on-road mobile emissions to cells were examined. The existing method calculates emissions for each road class, but distributes emissions to the cells using population density. For the new method a GIS builds road density by class and then distributes the emissions using road density. Comparing these methods reveals a significantly different spatial pattern of emissions.; Next, various model-coupling methodologies were analyzed, revealing numerous coupling approaches, some of which were categorized in the literature. Critiquing these categorizations while comparing them with documented implementations led to the development of a new coupling hierarchy. The properties of each hierarchical level are discussed with the advantages and limitations of each design. To successfully couple models, the spatial and temporal scales of all models in the coupled system and the spatiotemporal extents of the data must be reconciled.; Finally, a case study demonstrated methodologies for coupling SMOKE and a GIS. One methodology required a new approach utilizing dynamically linked libraries. Consequently, emissions were processed using SMOKE from a GIS. Also, a new method of converting data from netCDF files into a database was designed and implemented. The case study confirmed that no single coupling methodology is best in all situations. This research contributes new insights into the spatial distribution of vehicle emissions for AQMs. The hierarchy of model coupling methodologies provides design guidance for modeling projects. Finally, recommendations are presented for providing requisite geospatial functionality to multimedia models.