[A-AS09] Atmospheric Chemistry

[AAS09-17] Enhancing Global Air Quality Monitoring through Remote Sensing and Machine Learning

*TANIA SEPTI ANGGRAINI^1,2, HITOSHI IRIE¹, ANJAR DIMARA SAKTI^3,4, KETUT WIKANTIKA^3,4 (1.Center for Environmental Remote Sensing, Chiba University, 1-33 Yayoicho, Inage-Ku, Chiba 263-8522, Japan, 2.Doctoral Program in Geodesy and Geomatics Engineering, Faculty of Earth Science and Technology, Institut Teknologi Bandung, Bandung, 40132, Indonesia , 3.Center for Remote Sensing, Institut Teknologi Bandung, Bandung, 40132, Indonesia , 4.Remote Sensing and Geographic Information Sciences Research Group, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, 40132, Indonesia)

Keywords:Air Quality Index, Remote Sensing, Machine Learning Linear Regression, Monitoring Stations

The global distribution of air pollution monitoring stations is notably sparse, primarily due to the high costs associated with establishing and maintaining ground-based stations. This limitation has led to the concentration of these stations in urban areas, often identified as air pollution hotspots, leaving vast regions without adequate monitoring. This study proposes the utilization of Remote Sensing observations as a Low-Cost Monitoring and spatially comprehensive solution to this challenge. The research aims to predict the Air Quality Index (AQI) by integrating ground station measurements with Remote Sensing data through the application of Machine Learning Linear Regression (MLLR). MLLR is chosen for its efficiency in data processing and effective learning capabilities, will incorporate satellite-derived air pollution metrics, socio-economic factors, and environmental variables. The methodology is segmented into four scenarios: Single Linear Regression (SLR), Multiple Linear Regression (MLR), Single Random Forest Regression (SRFR), and Multiple Random Forest Regression (MRFR). Preliminary results indicate that the Multiple methods, particularly MLR and MRFR, show promising accuracy and reliability in predicting AQI. The implications of this study are significant, particularly for regions lacking ground-based air pollution monitoring stations. By leveraging Remote Sensing and advanced machine learning techniques, areas previously unmonitored can gain insights into their air quality status, facilitating informed decision-making and effective air pollution mitigation planning.