4 Spatial analysis and GIS Morton E. O’KelIy 1. Introduction Many of us involved in spatial analysis research are excited to see the explosive growth of interest in GIS. It is clear that the rapid growth of GIS has given a big boost to fundamental research in spatial analysis, and in many ways solidifies the future of the quantitative focus of the discipline. There exist potential linkages between many aspects of spatial analysis and the new information processing, data handling, data storage and data display techniques available through GIS. Furthermore, there are strong emerging links to applications in other disciplines with parallel interests in spatial analysis (including ecology, archaeology, natural resouroes, landscape architecture and geodetic science, etc.) These allied fields have long used spatial analysis tech rüques (see, for example, Upham, 1979; Bartlett, 1975; and Diggle, 1983) but are now increasingly using 018 as a creative tool. Against the background of this growth, and from the perspective of a spatial analyst looking out towards the next stages of development in GIS, two major directions which need attention are apparent: First, the traditional methods for displaying data about spatial situations, and for presenting the results of spatial analyses, need to be overhauled in view of the great advances in information processing technology. This line of attack is straightforward: it suggests that new display techniques be added to the output of existing spatial analysis operations. Ideally, the analyst would use the GIS in query mode (Goodchild, 1987) to develop an improved understanding of the properties of a spatial system. While this is easy to state in principle, the actual implementation will stretch current capabilities and will require a rethinking of the conceptual bases for spatial analysis. One such example might arise if improved display techniques increased the accessibility of multiobjective programming, thereby encouraging analysts to take a multi- objective view, and therefore replacing existing uni-dimensional methods. One obvious area for this to play a role is in locational analysis. Second, spatial analysts need to help GIS end-users and providers to under stand and to improve their sets of tools, and to enhance the appropriate levels of theory and modelling capability in real problem-solving situations. This need is particularly acute if appropriare analytical methods are to lie at the foundation of developments in 015. The capability to adapt existing al gorithms to the data structures in GIS is a critical component of this research. hae: FoM en oi4kqecni, /99tf