Ecological Indicators (Oct 2023)
Description of the multi-dimensional environment at the territorial scale: A holistic framework using cluster analysis and open data in France
Abstract
The impact of the environment on health is usually studied in a segmented manner, with a focus on a single source, pollutant, or exposure medium. To better understand spatial health inequalities, it is necessary to adopt multi-dimensional approaches to comprehensively describe the environment, especially at the territorial level. Clustering methods, which allow for the development of territorial typologies, are particularly interesting for this purpose. By simplifying complex datasets, these methods may reveal spatial patterns and geographical phenomena that would otherwise be difficult to observe. Based on the existing literature, there is a clear need for large-scale territorial typologies that comprehensively address the physical and outdoor environment.A robust and transposable framework was developed and applied to 3,041 municipalities in Northern France using open environmental data. It consists of five main steps: data collection, data selection, data preparation, cluster analysis, and cluster interpretation. This methodology allows for the development of an environmental classification of municipalities by identifying the primary environmental profiles represented in the study area. Cluster detection was performed based on 39 spatialized indicators that describe the level of environmental contamination (air, water, soil), the level of pollutant emissions, the proximity to emission sources, the land use, the agricultural practices, and the degree of naturalness in every municipality. As a result, municipalities were allocated into one of the seven following environmental profiles: (i) Dense urban centers; (ii) Peripheral urban municipalities; (iii) Intensive agricultural municipalities under urban influence; (iv) Intensive agricultural municipalities beyond urban influence; (v) More extensive and diversified agricultural municipalities; (vi) Municipalities with predominant livestock activities and significant natural areas; (vii) Municipalities with predominant natural areas: forests, wetlands, and water surfaces. The resulting typology goes far beyond a simple description of the urban–rural continuum. Five profiles of rural municipalities were identified, primarily distinguished by agricultural practices, degree of naturalness, and intensity of urban pressure.This approach enables researchers to identify the combination of environmental factors that shape a territory. It provides a more comprehensive and nuanced understanding of how environmental pressures and amenities are distributed in space and overlap with each other. By linking these typologies with health data, it could provide new insights into the etiology of complex diseases with unidentified environmental risk factors. Relying on open data, this framework is a valuable tool to assess etiological hypotheses at the territorial level.
