Assessing landscape composition using visual assessment: accuracy of rapid description compared to digital mapping

We compared the accuracy of visual assessment of landscape composition to that of Geographic Information Systems (GIS) mapping in (agro)ecology studies that focus on processes in which landscape factors are assumed to be insignificant and are often ignored. Seven land-use types were visually estimated from aerial photographs of contrasting rural landscapes in France. Visual assessment showed good agreement with results of GIS mapping for all land-use types, regardless of the observer. Estimates were accurate for forests and water areas, which are more visible than other land-use types. In contrast, observers often failed to distinguish crops from grassland, which decreased agreement for these land-use types; however, the estimate became fairly accurate when these land-use types were combined. Overall, this study confirms the relevance of using visual assessment to obtain simple landscape indices to improve studies that might otherwise ignore the landscape, provided that the loss of precision is compatible with the purpose of the study.