Coupling Geographic Information Systems and Urban Dispersion Model: An Example using ArcGIS and QUIC

The paper demonstrates a case of coupling between Geographic Information Systems (GIS) and Urban Dispersion Models (UDM). There are two main strategies to bridge GIS and environmental models: 1) integration and 2) coupling. Integration means implementing both GIS and environmental models based on a common data model and method (Bernard and Kruger, 2000). Coupling is defined as linking two separate systems together with data transfer; it can be loosely or tightly coupled (Nyerges, 1992). In the coupling, GIS technology is used to acquire, process, and analyze spatial data from different resources and scales and display geospatial information as maps, while environmental models offer prediction and simulation for complex environmental conditions. This paper shows a loose coupling between GIS and UDM using ArcGIS 9 and the Quick Urban and Industrial Complex dispersion modeling system (QUIC) as an example. ArcGIS 9 is a popular commercial GIS software developed by the Environmental Systems and Research Institute (ESRI, Redlands California). QUIC is a new dispersion model currentlydeveloped by Los Alamos National Laboratory (Pardyjak and Brown, 2002 and Williams et.al. 2002). It uses a diagnostic wind field model (QUIC-URB) that has been coupled with a Lagrangian dispersion model (QUIC-PLUME) and provides fast predictions of atmospheric dispersion in urban areas. It requires detailed building information within the modeling domain as input data. At present, this information is typically manually generated within the so-called city builder of QUIC, in which the user specifies the location and dimensions of all buildings resolved by the model. For operational use, an automatic generation of the building database is desirable.

The presented loose coupling starts with an algorithm developed in Visual Basic (VB) which converts a building shape file form ArcGIS 9 to QUIC's input data format. The building shape file is derived from LiDAR data and validated with aerial photographs and ground observations. Within the VB script, the user can specify the horizontal and vertical resolutions of the modeling domain. After the conversion, a QUIC input.dat file is automatically generated and ready for use in the urban dispersion model. The loose coupling provides a fast and convenient way to feed GIS data into QUIC model runs. Expanding upon the linking of ArcGIS and QUIC, the paper discusses the advantages and limitations of loose coupling and concludes that even though the loose coupling approach may seem ad-hoc or model specific, it could be beneficial for future improvement of UDM and provide a stepping stone to a true integration of GIS and UDM.