Characterizing the provision and inequality of primary school greenspaces in China’s major cities based on multi-sensor remote sensing

Environmental and green justice problems occur globally, especially in cities with unequal access to urban greenspaces. Recently, inequality in school greenspaces has drawn growing attention, given the importance of campus green environments in young students’ health and academic performance. However, the commonly used Normalized Differences Vegetation Index (NDVI) method for measuring greenspace from satellite imagery is hindered by the saturation issue and tend to underestimate greenspace at high vegetation cover areas, causing large uncertainties in greenspace inequality studies at a national scale. Besides, despite the progress on the inequality of public greenspace exposure, our understandings of primary school greenspace provision and inequality, as well as the driving factors, for young students in a developing world (e.g., China) is still limited. To address these issues, we first adapted a spectral unmixing technique based on multi-sensor remote sensing for more accurate measurements of greenspace provision. Then, we evaluated the provision and inequality of greenspace for 19,681 primary schools in China’s 31 major cities and examined the driving factors using an integrated path analysis. Our findings revealed that: (1) Our proposed multi-sensor remote sensing-based method for greenspace measurement is reliable across our study area with a R2 of 0.81 and RMSE of 0.14; in contrast, the traditional NDVI-based greenspace measurement saturated at the range of 0.7–1.0, leading to much lower accuracy (a R2 of 0.72 and RMSE of 0.24). (2) Most of the cities under study had low to moderate levels of inequality in primary school greenspace (Gini index < 0.5), but the overall greenspace provision was relatively low; Five cities under study facing high inequality in greenspace exposure (Gini index ≥ 0.5) as well as low greenspace provision (mean fraction cover < 0.25). (3) The monthly maximum temperature and the mean cover of greenspace in primary schools were identified as variables directly affecting the inequality in primary school greenspace (R2 = 0.76, p-value < 0.05), whereas the city-level government revenue manifests its effects through the mean cover of greenspace in primary schools and city-level mean greenspace cover. By developing a novel framework for examining the provision and inequality of greenspace in all primary schools in China’s major cities, our study provides valuable insights for designing and evaluating school greening programs in support of healthier learning environment development for next generations.