Abstract

GIS provides the opportunity to visualize and analyse the actual geographic distribution and properties of spatial data. Cellular automata provide the opportunity to examine the general behaviour and emergent properties of dynamic spatial processes. Combining the two allows us to compare actual geography with abstract theory. If we can transform between observed spatial distributions and cellular automata, the simplicity and flexibility of the cellular automata allow exploration of alternative scenarios that are almost impossible to conceive based directly on observed spatial data. ‘Cartogram’ or ‘density-equalizing’ transformations are well established methods that provide the necessary bridge between actual space and cellular automata space. We applied the script CartogramCreator in ArcGISTM to map the actual distribution of sheep farms in Great Britain to a cellular automata. Observations on the birth to death movements of sheep that died in the field were used to construct a simple model of the movement of scrapie infected sheep between farms. This formed the basis of a scrapie outbreak model, which generated 1000 simulations of the spatial distribution of scrapie cases. The simulated spatial distributions were compared with the distribution of scrapie cases reported in 2000, which was the last year before the Foot and Mouth outbreak and recent control measures against scrapie affected the distribution and reporting of scrapie. Comparison by spatial analysis suggested that the reported distribution of scrapie could be consistent with a long-lived outbreak, but might also be consistent with an endemic disease. If it is an outbreak, elimination of scrapie in Great Britain is more likely.