A composite index for integrated assessment of urban green space exposure

Urban green areas contribute to healthier cities by improving air quality, promoting physical activity and social cohesion, and mitigating the urban heat island effect. Despite this, exposure to green areas is often estimated using metrics that focus on different dimensions of greenery, leading to heterogeneous exposure estimates. In this study, we compared traditional green space indices and developed a composite Green Exposure Index (GEI) that integrates vegetation cover, density, and accessibility within a single quantitative framework to improve exposure assessment. We applied these indices to a population-based amyotrophic lateral sclerosis (ALS) case-control dataset from a Northern Italy community. We computed the index values for all residential locations across an 8400 km² urban-peri-urban domain from 1985 to 2020, using high-resolution remote sensing and land cover data. Comparisons between traditional indices showed high agreement between NDVI and Tasseled Cap Greenness (r ≥ 0.94), and exposure estimates derived from 100 m and 200 m buffers also remained strongly correlated (r = 0.94 - 0.96). Seasonal NDVI better captured vegetation patterns than annual values (r = 0.77 - 0.99), and spatial aggregation restricted to vegetated areas reduced the overestimation observed with circular buffers, improving classification accuracy while maintaining strong correlations (r > 0.80). The GEI consists of three components: seasonal NDVI, the Green Coverage Ratio (GCR), and an accessibility index defined for this application. Accessibility was calculated by assigning a value to each green area based on its type, with values decreasing with a logarithmic function as distance from the green area increased, reaching zero for distances beyond 1200 m. This threshold corresponds to the average distance traveled within a 15-minute walk, in line with the 15-minute city planning approach. The GEI was evaluated under three weighting scenarios, which produced substantial differences in exposure classification and confirmed that metric choice strongly influences results. The GCR alone classified 61.7% of the population as Not Exposed, whereas accessibility alone classified 86.1% as Exposed or Highly Exposed. The equally weighted GEI3 placed 79.7% of the population in the intermediate Mildly Exposed and Exposed categories, resulting in a balanced distribution. Analysis of the GEI time series revealed green space changes over the 36-year study period, reliably identifying areas affected by urbanization or green redevelopment. Findings from this case study demonstrate the added value of composite indices such as the GEI for characterizing green space exposure, enabling more comprehensive and robust assessments of the benefits and effects of green infrastructure, with applications in public health policy and urban planning.