Blue-green infrastructure assessment for sustainable urban open space planning

Abstract

Background: Rapid urban expansion exerts significant environmental pressure that degrades the quality and distribution of urban green open spaces (GOS). Urbanisation displaces vegetation cover, fragments ecological connectivity, and exacerbates the urban heat island effect, making comprehensive GOS quality assessment a critical planning priority. This study evaluates the quality of GOS distribution at the sub-district level in Semarang City, Indonesia, using the building blue-green index (BBGI). Methods: The BBGI framework integrates five parameters—vegetation index (VI), green index (GI), blue index (BI), impervious surface index (ISI), and high-rise building index (HBI)—weighted through analytical hierarchy process (AHP). Landsat satellite imagery (2010, 2020, 2024) was analysed using normalized difference vegetation index (NDVI)-based land-cover classification in ArcGIS 10.8. Findings: Nearest neighbor analysis (NNA) was applied to characterise the spatial distribution pattern of GOS, and spatial multi-criteria analysis (SMCA) was used to compute composite BBGI scores for all 16 sub-districts. GOS quality in Semarang City is spatially uneven, following a concentric pattern. Only three sub-districts achieved the high-quality category: Gunungpati (BBGI = 5.214), Mijen (4.540), and Tugu (4.522). Four sub-districts fell in the moderate category, and nine sub-districts (56.25%) were classified as low quality, with BBGI scores ranging from 1.479 to 2.608, concentrated in the urban core. The NNA z-score of −44.82 confirms a statistically significant clustered distribution of GOS, indicating systematic spatial inequality. Despite Semarang's relatively high aggregate GOS proportion (44.12%), qualitative distribution remains highly inequitable, constituting a form of environmental injustice for densely populated central districts. A zone management strategy—encompassing conservation, buffer, and intensification zones—is recommended to optimise GOS in accordance with the specific characteristics of each sub-district. Conclusion: This study represents the first application of the BBGI framework at sub-district scale in Semarang City, integrating multitemporal Landsat data, NNA spatial analysis, and AHP-weighted multicriteria evaluation. Novelty/Originality of this article: The findings provide a spatially explicit, evidence-based foundation for differentiated GOS optimisation strategies in rapidly urbanising Indonesian cities.