The Peter Burrough medal for senior scientists
Peter was among the first scholars to draw attention to the problem of error and uncertainty in spatial modelling with geographic information systems. He dedicated a full chapter to the topic in Principles of Geographical Information Systems for Land Resources Assessment (1986), and made major contributions to the topic over the following decades. His pioneering work defined the subject, and he was able to enthuse students and fellow scientists and let the subject develop into a major field of research within geo-information science. In doing so, Peter developed computer-based methods for landscape classification and display that led to numerous publications about, among other topics, error propagation in spatial modelling. He contributed a presentation titled “Identification of an error model for quantitative spatial attributes under different models of spatial variation” to the first “Spatial Accuracy” symposium, and was co-chair of the local organizing committee for the 4th symposium.
Peter’s scientific training sensitized him to the implications of the high levels of error and uncertainty present in much geospatial data. His early work about digitizing and scanning errors discussed in the first edition of his textbook merely scratches the surface of the problem, being concerned with the additionalerrors introduced when maps are converted to digital form. At the end of Chapter 1 in the first edition of his book, in a section about future directions and trends in GIS, Peter writes (p. 11): “the problems of errors and data description [that] arise largely as a consequence of the imprecision of the world in general, coupled with the requirement of current conceptual models to be able to represent natural variation in terms of compartmentalized models and watertight logic . . . we should be looking for better ways in which to describe the vagaries of the world, and new methods for dealing with the imprecision of qualitative judgements that are an integral part of human thought processes.” In June 1994, Peter co-organized a workshop that ultimately produced Geographic Objects with Indeterminate Boundaries (Burrough and Frank 1996), an edited collection of papers about the problem of uncertainty and its manifestation in vague boundaries. The issue had long festered in cartography, but as long as boundaries had to be drawn on paper using a pen, no alternative existed to portraying them as precisely known, despite the obvious falsity of this assumption in such fields as soil or ecological habitat mapping. Concepts of fuzzy and rough sets, the representation of transition zones, and the “egg-yolk” model of a certain centre and uncertain periphery entered GIScience at this time, largely stimulated by this meeting.
Many other significant contributions followed. Peter worked on the problem of error propagation, resulting in a series of techniques, both analytic and computational, for examining how uncertainty in geographic data produces uncertainty in the outputs of a GIS.
Peter passed away in 2009. For his outstanding contributions to the academic interests of the International Symposia on Spatial Accuracy Assessment in Natural Resources and Environmental Sciences, ISARA established a medal in honour of Peter Burrough.