Exploring individual preferences for periurban landscapes

International audienceLandscape aesthetic quality has recently become an important dimension of land use policies since it is likely to play a role in human well-being. Several scientific studies have shown that landscape is a significant component of residential satisfaction of inhabitants in urban and suburban areas (Hur & Nasar, 2010; Kweon et al., 2010). Although neglected by land-use planning for decades (Gallent et al., 2004), urban fringes are particularly interesting for analyzing landscape changes. These spaces where town meets countryside (Scott et al., 2013) are indeed particularly concerned with urban sprawl that occurs in all European cities from 1960s (EEA, 2006). Such changes contribute to impact landscapes, and thus affect the living environment of inhabitants.Two main quantitative methods allow to assess landscape aesthetic quality, based (i) on monetary values in econometric evaluations (Cavailhès et al., 2009), or (ii) onn individual cognitive processes in environmental psychology (Kaplan & Kaplan, 1989). In each case, the construction of landscape indicators is an essential step. They make it possible to establish an objective basis for identifying landscape character by dividing all of our visual perception of the physical landscape into quantifiable characteristics (Ode et al., 2008). In this framework, visibility analysis based on GIS has emerged in 1990s (Fisher, 1992; Franklin & Ray, 1994) and has been progressively used to model landscape preferences and landscape satisfaction (De la Fuente et al., 2006 ; Youssoufi & Foltête, 2013). Landscape visibility computations can be split into two main methods. Metrics can be computed by (i) the sight-line method (Fisher, 1996; Joly et al., 2009), consisting in characterizing visible landscape according to a count of pixels seen from a virtual viewpoint, or (ii) the solid angles method (Chamberlain & Meitner, 2013; Domingo-Santos et al., 2011), consisting to quantify the surfaces occupied by landscape features on the observer’s retina. However, such analyses have a well-recognized limitation: usual landscape metrics in any methods are almost exclusively based on landscape composition, and does not allow to model landscape structure.This contribution aims to propose new visibility metrics computed by solid angles method to model landscape visibility structure. The capacity of these metrics to explain landscape preferences is evaluated by means of a perception survey filled by a panel of individuals (n = 1420), and based on a corpus of 30 photographs showing diversity of urban fringe’s landscapes. Results from linear regression showed that the set of selected visibility metrics is interesting to create robust models of landscape aesthetic potential. The main perspective of this study would be its use for decision and planning support systems, for implications in landscape management and urban planning with the objective of sustainable development of urban fringes.Study areas are two urban fringes, near the medium sized city of Besançon (eastern France) and in the eastern part of the Parisian metropolitan region, where urban development has increased since the 1960s, but according differentiated dynamics