THE APPLICATION OF GIS AND REMOTE SENSING BASED MODELLING TECHNIQUES, FOR USE IN THE ECONOMIC AND EPIDEMIOLOGICAL ASSESSMENT OF DISEASE CONTROL INTERVENTIONS, AT A REGIONAL OR NATIONAL LEVEL PATERSON, A. D. 1 , OTTE, M. J. 2 , SLINGENBERGH, J. 2 , WINT, W. 3 , ROGERS, D. 3 The quantitative economic assessment of interventions in livestock production systems requires the analysis of the outputs and resource requirements of all the production systems within the target area. Such analyses are not possible without accurate data describing both livestock and human populations, and the production systems within which the livestock are managed. When, as is often the case, such information is unobtainable, this missing data poses a significant constraint to planned livestock development. The approach described, demonstrates a cost-effective solution for augmenting existing, conventionally gathered field data, using techniques appropriate to the needs and resources of developing countries. Novel techniques based on geographical information system and remote sensing technology are used to classify livestock farming systems, agro-ecological zones, and predict animal populations at a high level of resolution, using a combination of highly objective, remotely sensed satellite and eco-climatic data, combined with existing animal population information. This results in the significant ability to predict the distribution of livestock, and relate them to the production parameters of the farming systems within which they are managed, over a very large geographical area. A major assumption of this work is that all the animals of one species that are predicted to lie within each of the 0.05 o pixels (6 km x 6 km) are managed in the same way e.g. as small- scale, intensive, East African, coastal, dairy cattle, and thus have a recognisable set of production parameters that can be attributed to them. In the context of developing countries, where intensive, housed farming systems are comparatively rare, this is not an unreasonable assumption The techniques described have the potential for application at two levels: • At the national and sub-national level – in this case the main advantage gained is due to the increased resolution (i.e. small unit area) of the predicted data, and the freedom from analytical constraints imposed on data gathered using administrative boundaries (political and administrative boundaries rarely have any relationship to the eco-climatic determinants of human cattle cropping and farming practices within them). This freedom permits novel analyses on the basis of time, space, ecological, physical and other features. This flexibility in analysis allows more accurate and meaningful assessment of disease control strategies, and other types of interventions in production systems, such as: improvements in infrastructure, provision of marketing resources, changes in restocking policies and breed improvement. 1 The Veterinary Epidemiology and Economics Research Unit (VEERU), University of Reading, UK 2 The Animal Production and Health Division (AGAH), The Food and Agriculture Organisation, Rome, Italy 3 The Environmental Research Group Oxford (ERGO), University of Oxford, UK