Estimation of the PM2.5 Pollution Levels in Beijing Based on Nighttime Light Data from the Defense Meteorological Satellite Program-Operational Linescan System

• Main Authors: Runya Li, Xiangnan Liu, Xuqing Li
• Format: Article
• Language: English
• Published: MDPI AG 2015-05-01
• Series: Atmosphere
• Subjects: PM2.5; DMSP-OLS; nighttime light data; BP neural-network; Beijing
• Online Access: http://www.mdpi.com/2073-4433/6/5/607

Nighttime light data record the artificial light on the Earth’s surface and can be used to estimate the degree of pollution associated with particulate matter with an aerodynamic diameter of less than 2.5 μm (PM2.5) in the ground-level atmosphere. This study proposes a simple method for monitoring PM2.5 concentrations at night by using nighttime light imagery from the Defense Meteorological Satellite Program-Operational Linescan System (DMSP-OLS). This research synthesizes remote sensing and geographic information system techniques and establishes a back propagation neural-network (BP network) model. The BP network model for nighttime light data performed well in estimating the PM2.5 pollution in Beijing. The correlation coefficient between the BP network model predictions and the corrected PM2.5 concentration was 0.975; the root mean square error was 26.26 μg/m3, with a corresponding average PM2.5 concentration of 155.07 μg/m3; and the average accuracy was 0.796. The accuracy of the results primarily depended on the method of selecting regions in the DMSP nighttime light data. This study provides an opportunity to measure the nighttime environment. Furthermore, these results can assist government agencies in determining particulate matter pollution control areas and developing and implementing environmental conservation planning.