A contribution to the assessment of scenic quality of landscapes based on preferences expressed by the public

The purpose of this paper is to validate a model for predicting the public's preference for a landscape using simple statistical techniques. The model assigns numerical values to 42 landscape variables grouped into physical, aesthetic and psychological attributes. The landscape value is obtained using an additive value function. The results of the model on certain landscapes are compared with the preference expressed by the public in a survey of 183 people.The homogeneity of the survey responses was checked in each photograph. It was determined that means and standard deviations of the scores represent the true preference.A strong positive association was observed between preference and certain landscape attributes such as expression, soil use or colour. A marked negative association was detected with respect to landscape alterations.A linear regression was carried out to analyse the predictive capacity of the model. The independent variable was the global score assigned by the model to each photograph and the dependant variable was the mean of the scores assigned by the survey respondents. The high level of correlation obtained indicates that the model is a good predictor of the public's preferences in relation to the set of photographs shown in the survey.To complete the analysis of the model as a preference predictor, a multiple linear regression was carried out between the mean score obtained in the survey and the model elements. Attributes and variables that have the greatest influence on pubic preference were detected.