Determining appropriate neighborhood shapes and sizes for modeling landscape satisfaction

Landscape is a key factor in neighborhood satisfaction. Our aim here is to derive indicators of landscape satisfaction from land cover data; consequently the definition of the neighborhood around residential locations is of prime importance. Three shapes of neighborhood are compared in the paper, depending on individuals’ relationships with their day-to-day environment: (1) a circular neighborhood representing an isotropic space; (2) a network neighborhood taking into account the potential spatial practices around the residential location; and (3) a visible neighborhood based on whether or not landscape features are in view. These neighborhoods are determined in a suburban area located near Besançon (France), where an indicator of landscape satisfaction was generated from a survey. Each neighborhood shape was used to define a series of landscape variables from a land cover map, by varying the neighborhood size to identify the most relevant spatial scales. Statistical models using correlation and regression were run to evaluate the connections between satisfaction and landscape variables. Comparisons between the models built with each neighborhood shape shows that the circular neighborhood is globally the most efficient. The results also emphasize the combination of several scales in the perception of residential environment.

► Exploration of several neighborhood sizes reveals that residential satisfaction with landscape is scale-dependent.
► Circular neighborhoods provide effective indicators for representing residential satisfaction with landscape.
► Network and visible neighborhoods do not improve the modeling of residential satisfaction using land cover data.
► Very local scales of neighborhood are heavily dependent on the spatial resolution of data.