Application of Remote Sensing and Machine Learning Technologies to Assess the Impact of Urban Expansion on Forest Degradation and PM2.5 and PM10 Emissions in Hanoi for Urban Health Protection

Urban expansion is an unavoidable phenomenon in the realm of economic and social progress; nonetheless, it brings forth significant environmental repercussions, notably the diminishment of green spaces and the escalation of fine particulate matter concentrations in the atmosphere. The findings of this research indicate that the proportion of developed land has surged from 13.92% in 1995 to 32.27% in 2025, while forest coverage has markedly declined, particularly between 2015 and 2025, witnessing a loss of 6,297.55 ha. This shift not only constricts green areas but also exacerbates the levels of fine particulate matter, especially in regions characterized by high urban density and substantial industrial operations. The Ant Colony Optimization-Generative Adversarial Network (ACO-GAN) model exhibited exceptional efficacy, achieving RMSE values of 7.38 and 9.76 for PM2.5 and PM10, respectively, alongside an R² coefficient surpassing 0.82. Variables such as construction density (CONST DEN), land surface temperature (LST), Normalized Difference Vegetation Index (NDVI), and Soil-Adjusted Vegetation Index (SAVI) have been recognized as the most pivotal determinants affecting fine particulate concentrations. Additionally, the findings reveal that areas with elevated urban density, particularly industrial locales like Noi Bai and Dong Anh, demonstrate alarmingly high levels of fine particulate matter, whereas regions abundant in vegetation, such as Ba Vi Mountain, exhibit considerably lower concentrations. This highlights the essential function of green spaces in alleviating air pollution and safeguarding the living environment. These insights not only elucidate the adverse effects of urbanization on air quality but also establish a scientific foundation for advocating sustainable urban management and planning strategies.