Error in GIS is a ‘fact of life’ but not an intractable problem. Adopting good practice in terms of data capture and analysis should, in most instances, be sufficient to ensure errors are kept to a...

Error in GIS is a ‘fact of life’ but not an intractable problem. Adopting good practice in terms of data capture and analysis should, in most instances, be sufficient to ensure errors are kept to a minimum. Conscientious documentation and the incorporation of lineage information and quality statements on GIS output will help to ensure that end users are aware of the limitations of GIS products. Box 10.9 outlines some of the known errors in the nuclear waste case study by way of an example of the types of errors that can be documented. Despite the existence and documentation of errors such as those presented in Box 10.9, some GIS users, like car drivers, press on with their work despite the GIS equivalent of the flashing of the oil warning light, the clouds of steam coming from the radiator and the clanking noise from the rear axle. Like (petrol engine) motor cars, GIS require the correct data (petrol not diesel), data which are fit for the purpose and system (unleaded or leaded petrol), appropriate choice of analysis method (careful driving) and maintenance of correct lineage (good service history). As GIS become widely used in everyday decision making, the issue of legal liability raises its head. If a GIS is used to support a decision, which subsequently results in loss or injury where the GIS analysis has been proven to be at fault, who or what is to blame? Is it the decision maker who decided to use GIS? Is it the GIS analyst who directed the GIS work? Is it the GIS technician who pushed the keys? Is it the GIS or the vendor who supplied the software? Is it the data or the data supplier? Chances are that the ‘buck’ stops with the GIS analyst, for it is they who will have made the crucial decisions regarding what data and methods of analysis to use. Precedents exist in the USA, where people have been successfully sued over faulty GIS analyses – think back to the example of the Happy Valley GIS team and the risk zones produced using their avalanche model. Openshaw (1993) gives an interesting account of these and other issues in his editorial ‘GIS crime and criminality’. With expertise in GIS comes responsibility. Responsibility to ensure that the data and the chosen method of analysis are appropriate to the problem in hand. Responsibility to ensure that possible alternative solutions have been considered and are presented to the decision maker. Responsibility to ensure that the best available data are used and any potential errors are accounted for and communicated clearly to the end user. As suggested in Chapter 8, GIS output never fails to impress management and decision makers; it is colourful, it has been produced using a computer, it is often complex. Is it correct? Any doubt in the mind of the GIS analyst, whether due to potential data error or to uncertainty about the analysis, should be clearly stated in the output. Be careful in your choice of data. Be careful in your chosen analysis procedure. Communicate any doubts clearly in the results and output.