Surfaces for GIScience

The articles included in this special issue were presented as part of a meeting `Surfaces in Geography' sponsored by the Quantitative Methods Research Group of the Royal Geographical Society-Institute of British Geographers (RGS-IBG) and the Modelling and Advanced Techniques Special Interest Group of the Remote Sensing Society. The original meeting comprised 16 papers and eight posters and took place on 5 January 2000 as part of the annual meeting of the RGS-IBG held this year at the University of Sussex, England. The original meeting was convened jointly by Nick Tate (University of Leicester) with Peter Fisher (University of Leicester), Jo Wood (City University) and Peter Atkinson (University of Southampton). I am especially grateful to these individuals for their assistance with the original meeting. The five articles in this issue are all concerned with the modelling and analysis of surface data relevant in a GIScience context. Although the included articles encompass a variety of both methods used and application contexts, they have been divided into two groups: those concerned with socio-economic surfaces, and those concerned with topographic surfaces. Such a grouping reflects the two main foci of interest at this time for surface modelling in GIScience. The advantages of surface representations of socio-economic variables such as population are generally well-known in that they can provide a more realistic continuous model of settlement pattern, reflecting changes in density and occupation which are often difficult to represent using zone-based models. A zone-based model that employs discrete and often arbitrary boundaries forms the basis of many of the census of population products in both the UK and the US. Such zone-based models are also prone to the modifiable areal unit problem that is a major constraint to the tasks of rescaling and pattern identification often undertaken with such data. Interest in surface representations of variables such as population, often created using some form of kernel density estimation (KDE), is clearly on the increase as evidenced by the recent workshop on gridding population data held on 2±3 May at Columbia University. The utility of surfaces in a topographic context is perhaps more intuitive in that elevation is more commonly represented as a variable which varies continuously across space in the form of a Digital Terrain Model (DTM). Research issues here are primarily concerned with the methods of generation, manipulation, interpretation, visualisation and application of DTMs (Weibel 1997). Similar to the socio-economic context, scale issues are also important. These can be the identification of scaledependent and scale-independent elements of real topographical surfaces as well as the Transactions in GIS, 2000, 4(4): 301±303