mountainous 
ain 
  
  
  
  
  
  
  
us terrain 
  
  
COUPLING GIS AND ENVIRONMENTAL MODELLING: 
THE IMPLICATIONS FOR SPATIO-TEMPORAL DATA MODELLING 
Cheng Tao, Wolfgang Kainz, Robert A. van Zuidam 
International Institute for Aerospace Survey and Earth Sciences (ITC) 
350 Boulevard 1945, P.O.Box 6, 7500 AA Enschede, The Netherlands 
e-mail: (cheng, kainz }@itc.nl 
Commission Ill, Working Group 4 
KEY WORDS: Environmental Modelling, Object-oriented Data Model, Spatio- Temporal GIS 
ABSTRACT 
This paper discusses spatio-temporal data model design for coupling environmental models with GlSs. It is 
organized along three levels of data modelling, i.e., the spatial modelling level; the conceptual level and the logical 
level. Issues to be discussed include: whether an application-oriented or a system design-oriented perspective 
should be taken at the spatial modelling level; how the objects should be constructed at the conceptual level 
(geometric-based, temporal-based or attribute-based), if an object-oriented approach is selected; should it be the 
loosely-coupled or tightly-coupled method chosen for the logical implementation; and finally how environmental 
modelling can be implemented in objects. It is hoped that these discussions will contribute to the data model 
design for the integration of environmental models with GISs, and to the development of object-oriented concepts in 
GIS data modelling. 
EE 
   
  
   
  
  
  
  
  
  
  
  
  
  
  
  
  
1. INTRODUCTION 
In recent years, the integration of environmental 
models with GISs has attracted much research 
attention. The past experiences demonstrate that 
three levels of such an integration can be achieved, 
depending on the circumstances of the GlSs and 
the environmental models: 
a. A basic level of separated but interfaced system. 
At this level, environmental models communicate 
with GISs via data files. The GIS data are converted 
to the formats (or structures) required by the 
environmental models, or the way around, so that 
the data stored in the GISs can be used by the 
models, and the data generated by models can be 
represented in the GISs. Models and GISs are 
independent of each other. 
b. An intermediate level of "seamlessly" interwoven 
systems sharing a common user interface. At this 
level, macro-languages provided by GISs are used 
to build a user interface, under which the GISs and 
environmental models are linked. Sometimes, the 
macro-language can also be used to develop part of 
the environmental models. The user interface 
provides menus to invoke the modelling process 
module and GISs functionality. So the gap between 
GISs and modelling processes may not be noticed 
by the users. 
c. An advanced level of a fully integrated system. At 
this level, the environmental modelling is fully 
embedded in a GIS. The environmental models can 
directly read GIS data and the data produced by 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B3. Vienna 1996 
models can be directly managed and represented by 
the GIS. 
Currently, most environmental models are 
integrated with GISs at the first level. Such GISs are 
used as a tool for preparing data input for 
environmental models, and for post-processing data 
output from environmental models. However, even 
this basic level of integration is normally made ad 
hoc, and a standardized procedure is yet to be 
created (Cremers et al., 1995). Given the fact that 
many commercial GISs provide macro-languages or 
other tools to facilitate applications in different 
fields, there are ample opportunities to achieve the 
second level of integration. However, the static and 
2-D or 2.5-D features of current GlSs may only 
allow the simple and static models to be easily 
integrated. Complex, 3-D and dynamic 
environmental models, which are important 
analytical tools in environmental studies, may be 
difficult. Most benefits of using GISs for 
environmental modelling can only be achieved by 
the fully coupled environmental modelling processes 
within a GIS framework. Whether or not such an 
objective can be achieved largely depends on what 
is offered by the GIS and what is required by the 
modelling process. One of the most important 
factors in this supply and demand relationship is 
whether or not the data structure of the GIS can 
support the data organization requirements of the 
modelling process concerned. That justifies the 
importance and the necessity of discussing the 
issues of the spatio-temporal GIS data model 
design for coupling GIS and environmental 
modelling in this paper.