Urban-rural gradient detection using multivariate spatial analysis and landscape metrics

The gradient approach allows for an innovative representation of landscape composition and configuration not presupposing spatial discontinuities typical of the conventional methods of analysis. Also the urban-rural dichotomy can be better understood through a continuous landscape gradient whose characterization changes accordingly to natural and anthropic variables taken into account and to the spatio-temporal scale adopted for the study. The research was aimed at the analysis of an urban-rural gradient within a study area located in central Italy, using spatial indicators associated with urbanization, agriculture and natural elements. A multivariate spatial analysis (MSA) of such indicators enabled the identification of urban, agricultural and natural dominated areas, as well as specific landscape transitions where the most relevant relationships between agriculture and other landscape components were detected. Landscapes derived from MSA were studied by a set of key landscape pattern metrics within a framework oriented to the structural characterization of the whole urban-rural gradient. The results showed two distinct sub-gradients: one urban-agricultural and one agricultural-natural, both characterized by different fringe areas. This application highlighted how the proposed methodology can represent a reliable approach supporting modern landscape planning and management.