How suitable are discrete choice experiments based on landscape indicators for estimating landscape preferences?: Fid-Bau Portal

How suitable are discrete choice experiments based on landscape indicators for estimating landscape preferences?

Schirpke, Uta / Mölk, Felix / Feilhauer, Elisabeth / Tappeiner, Ulrike / Tappeiner, Gottfried

Highlights • We assess the stability of landscape preferences over time and across picture sets. • Based on 11 landscape indicators, we derive importances and part-worth utilities. • Landscape preferences remain sufficiently stable over time. • Landscape preferences are less stable across different picture sets. • Our approach has potential to map aesthetic values at the landscape scale.

Abstract Appealing landscapes contribute to human well-being, but landscape changes require the attention of management and planning. Although aesthetic landscape values are increasingly addressed in the context of cultural ecosystem services, their assessment is particularly challenging due to the subjectivity of landscape preferences. To map aesthetic landscape values across landscapes, discrete choice experiments in combination with landscape indicators may be a powerful tool. However, it is still not sufficiently understood (1) whether there are time effects on landscape preferences and (2) how much the stated landscape preferences depend on the selection of landscape pictures that are used as stimuli. To address these open questions, we used an experimental approach involving 252 respondents. Based on photo-based questionnaires with 98 landscapes photographs from the Central Alps, we conducted two survey rounds with one month difference. The visual properties of the photographs were quantified through 11 landscape indicators describing spatial patterns and features. We applied conjoint analysis to derive importances of the attributs and part-worth utilities of the corresponding levels for evaluating the stability of landscape preferences. Our results indicate that landscape preferences remain sufficiently stable over a short period (correlation of 0.890 for mean importances), while the results are more sensitive to the selection of the pictures. This suggests that the methodological approach can provide meaningful insights into landscape preferences if pictures are used that are representative of a specific landscape. Our findings contribute to identifying limitations and opportunities for assessing and mapping aesthetic landscape values to support management and decision-making.