Multi-Index spectral analysis of urban green space categories in Bengaluru using Sentinel-2 Data

The rapid densification of metropolitan areas in the Global South has precipitated a critical decline in urban green spaces (UGS), necessitating advanced monitoring frameworks that transcend simple binary vegetation mapping. This study presents a typological classification and multi-index spectral characterization of UGS in Bengaluru, India, utilizing high-resolution Sentinel-2 multispectral imagery. A comprehensive seven-class typology was developed to distinguish between distinct ecological and governance categories, including Urban Forests, Managed Public Parks, and Institutional Green Complexes. Spectral health was assessed using a suite of indices- Normalised Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Soil-Adjusted Vegetation Index (SAVI), and Normalized Difference Built-up Index (NDBI) to capture the structural heterogeneity of these spaces. Results indicate a stark spatial imbalance in the city's green cover. The landscape is co dominated by Grassland/Scrubland Systems (30.77%) and Neighborhood/Community Parks (30.39%), reflecting a fragmented network of transitional and residential open spaces rather than contiguous ecosystems. In contrast, high-ecological-value Urban Forests remain critically fragmented, occupying only 2.20% of the classified area. Spectral analysis reveals that while Institutional Green Complexes (10.96%) function as vital ecological refugia with moderate-to-high vegetation density (μ_NDVI=0.56), Avenue Plantations exhibit significant urban stress (μ_NDVI=0.26) and high interaction with built-up surfaces. These findings demonstrate that a multi-index approach provides a more nuanced assessment of urban ecological quality than single-index methods, highlighting the urgent need to protect institutional lands and restore fragmented forest patches for sustainable urban planning.