An ecological approach to land-cover mapping is often absent in studies into the area-level relationship between the natural environment and human well-being. Research within public health often bundles the natural environment into simplistic metrics such as proportion cover by green space, without consideration of wider ecological parameters such as quality, structure, type or succession. Although measures related to population well-being, socio-economic status and cultural background are increasingly sophisticated in related studies, such an in-depth approach is still rarely applied to characterisation of the physical landscape. This paucity of detail in characterising green space in public health research persists despite the availability of remote sensing and GIS techniques which permit the relatively rapid creation of data towards describing landscape characteristics at a range of scales. Applying the use of relevant landscape metrics at a scale congruent with data available on population health remains a challenge in research into the natural environment and human well-being.
Using high resolution (10m) remotely sensed data (Sentinel 2A; 2016), a supervised classification method was used to characterise the landscape of the conurbation of Greater Manchester stratified according to recognized woodland habitat successional layers. Additional data on canopy cover (Red Rose Forest, 2011) and water bodies and courses (OS VectorMap, 2016) were incorporated and, after final manual editing, the resulting dataset was cross-validated using Edina Digimap aerial photography (2017).
Subsequent measures of land-cover were used in the calculation of a range of landscape indices using the QGIS landscape ecology plug-in (LecoS 2.0.7). These were then reduced, through factor analysis, to reveal thematic properties such as cohesion/fragmentation, diversity, evenness, complexity and succession throughout the landscape. This allowed a typology of landscape character to be generated using K-means clustering. The resulting typology presents a practical resource for use, alone or in combination with other socio-demographic typologies, in a range of potential social-ecological analyses.
