Creating Landscape Preference Models Using Internet Survey Techniques

Abstract

The use of the Internet to collect data for predictive landscape preference models is described and tested, including the issues involved in using such a method for data collection and the functionality of the questionnaire. The creation of a psychophysical landscape preference model for rural landscapes in Scotland is described. Ninety landscape images were analysed by digitizing each image into seven landform types, based on elevation and land cover. The colours present in each image were also recorded and variables representing complexity and coherence were measured. The most important predictors of visual landscape preference were found to be complexity, coherence and the presence or absence of water and mountainous landscape. These were among the variables used in two preference models, which explained over 65% of the variation in landscape preference and which produced correlations of over 0.6 with a second data set of preference scores.