Assessment of Urban and Peri-Urban Green Infrastructure Patterns Using Morphological Spatial Pattern Analysis and Satellite Imagery: Case Studies of Braşov and Oradea, Romania

Sustainable urban development requires a detailed understanding of green infrastructure (GI) and its spatial patterns. Urban growth can occur within existing urban areas or through expansion into peri-urban zones. Maintaining a high percentage of GI contributes to well-being by regulating microclimatic parameters, reducing air pollution, decreasing urban noise, and supporting public health by providing access to functional green spaces. This study aimed to assess the GI patterns in the urban and peri-urban areas of two Romanian cities, Brașov and Oradea, using Morphological Spatial Pattern Analysis (MSPA). Data for evaluating GI patterns were derived from multi-temporal Sentinel-1 (S-1) and Sentinel-2 (S-2) satellite imagery, alongside vegetation indices (NDVI, EVI, NDBI, SAVI, and NDWI). Classification was performed into seven land use/land cover (LULC) classes using the Gradient Tree Boosting (GTB) machine learning algorithm, achieving overall accuracies of 96.02% for Brașov and 95.67% for Oradea. The LULC categories were reclassified into foreground (forest, grassland, cropland, water) and background (built-up, uncultivated, bare land). Seven MSPA classes (core, edge, bridge, branch, islet, perforation, and loop) were evaluated to measure GI's morphological patterns, and Gini coefficients were calculated to assess GI equity. Results showed that Brașov had a moderately equitable GI distribution (0.325) in urban and peri-urban areas, while Oradea displayed a more unequal distribution (0.519). Core areas represented the largest spatial extent in both cities, with Brașov covering 73.1% (urban) and 84.1% (urban and peri-urban), while Oradea covered 59.0% (urban) and 66.5% (urban and peri-urban). The GI edge pattern in Oradea was more complex, indicating higher fragmentation.