ABSTRACT
Geographical information systems and environmental simulation models started to be used together round about the end of the 1980s. From my point of view, there was no particular application, no landmark paper that monuments the beginning of this fortuitous working together of the different technologies. It was more the case that the benefits of the idea independently sprung up on different types of projects in different parts of the world. It was, however, not inevitable that GIS and environmental modelling would come together (Parks, 1993). They are rather different technologies. GIS focus on representations of location, the spatial distri bution of phenomena and their relationships to one another in space. These are usually static representations. Environmental simulation models on the other hand are principally concerned with system states, mass balance and conservation of energy, that is, focusing on quantities (popu lations, chemicals, water) in time. Whilst, the distribution of ‘actors’ (Fedra, 1993) within environmental simulation models are affected in their interactions and dynamics by their spatial distribution, many of the early models of the late 1960s, 1970s and early 1980s did not treat the spatial dimension explicitly. If we treat environmental modelling in its broadest sense and include the logical models of land use suitability then, yes, there has been a considerable tradition of treating space as explicit and it was these types of applications that propelled GIS forward in its earliest stages (see Chapter 2). Nevertheless, even with the advent of com puter-based numerical simulation models in the atmospheric sciences, hydrology, biology and ecology, the GIS link was not a foregone conclu sion. To be sure, with the aid of GIS, hydrological models, for example, could more easily move from a 1-D treatment of the drainage basin to a distributed-parameter approach. GIS could be used to calculate gradient and aspect for a distributed model, discretise soil and land use coverages, interpolate sampled parameters and thus handle spatial variability expli citly. This is not to say that GIS cannot have a role in 1-D modelling (e.g. in producing spatially averaged parameters), it is just that in this mode, the two technologies can, if necessary, be used quite independently of each other and there are not nearly the same level of synergies to be achieved
by getting GIS and environmental simulation models to work together. Nevertheless, in either case, it is absolutely critical to remember that GIS are not sources of spatial data, they are a technology for handling, manipu lating and displaying the spatial data provided to them (Fedra, 1993; Maidment, 1993b).
