discrete and static data about spatial structures (distribution and spatial 
relationships) and the relevant contents. Therefore, the spatial reality to be 
analyzed (or modeled) should be mapped or viewed as a set of spatial realities 
or spatial properties of these entities and their spatial relationships can 
then be used by spatial models for analysis and modelling. 
There are two general approaches to map the spatial reality into spatial 
objects in GIS. The first one is to classify the spatial reality into spatial 
objects in terms of points, lines, and polygons, each having its spatial 
properties or contents as associate attributes. The second is to divide the 
continous space. In either case, the spatial data contained in GIS is used to 
define the spatial objects. : 
On the basis of the above mapping, it is therefore possible to formulate a 
GIS's spatial analysis model by integrating the analytic capabilities or 
process modeling functionalities of the spatial model with the spatial data 
management and map-based analysis capabilities of GIS. The map-based analysis 
and spatial data may be used in developing model parameters and deriving input 
data during runs of models [kemp, 1992], and GIS supports the interactive graphic 
display and efficient data management during the spatial analysis. 
3). Choosing a coupling environment 
As stated by Nyerges (1992), coupling spatial model with GIS means the software 
and hardware connection that implements the integration of the analytic 
funtions (or spatial process modeling) with GIS capabilities. The most coupling 
occurs through the DBMS of GIS which requires the internal modeling within 
GIS,i.e, the spatial models are implemented by using GIS provided language or 
related funtions [Chou and Ding, 1992]. The compatibility of the spatial 
objects defined above with the data model of GIS determine how easy or 
difficult it is to couple the two. 
3. Some examples of urban GIS's spatial analysis models. 
1). Urban traffic volume analysis model 
An urban traffic volume analysis model was developed by coupling the 'four step 
model^ of traffic planning with GIS [Chen, 1992]. The spatial analytic model was 
used for traffic analysis and simulation, including trip generation, mode 
choice and trip assignment. As shown in Fig.1, the spatial entities to be 
analyzed by the ‘four-steps’ model (such as traffic zones, road network, and 
traffic facilities) are mapped into spatial objects in terms of point, line and 
polygons with the associate attributes in GIS database. GIS is used not only 
for supporting data storage, display and interactive modeling, but also for. 
166 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996