The reliability of 'Phase 1' habitat mapping in the UK: the extent and types of observer bias

The recording of broad vegetation types in habitat or land cover maps is often the starting point of ecological investigations of sites and landscapes. Habitat maps inform wildlife evaluations and management decisions, and can aid in the identification of areas requiring further investigation. Little data are currently available on the reliability of maps recorded by field surveyors, or the types of error that they contain. In this study, maps of the same upland area in northern England produced by six experienced ecologists are compared. The surveyors used a standard method applied widely within the United Kingdom, but agreement between pairs of maps averaged only 25.6% of the study site's area. The majority of differences were due to differing interpretations of the types of vegetation present, rather than spatial errors - although the latter were also detected. The range of vegetation types which were confused with each other was great, but ecologically related vegetation types were most often confused. The causes and consequences of differences between maps are discussed, and quality control and quality assurance are identified as important, but neglected issues in ecological survey. (C) 1999 Elsevier Science B.V. All rights reserved.