Integration of Remote Sensing and Social Sensing Data in a Deep Learning Framework for Hourly Urban PM2.5 Mapping.

Shen Shen,Zhou Zhou,Zeng Zeng,Li Li

doi:10.3390/ijerph16214102

Shen Shen, Zhou Zhou + Show 2 more

Open Access

https://doi.org/10.3390/ijerph16214102

Copy DOI

Abstract

Fine spatiotemporal mapping of PM2.5 concentration in urban areas is of great significance in epidemiologic research. However, both the diversity and the complex nonlinear relationships of PM2.5 influencing factors pose challenges for accurate mapping. To address these issues, we innovatively combined social sensing data with remote sensing data and other auxiliary variables, which can bring both natural and social factors into the modeling; meanwhile, we used a deep learning method to learn the nonlinear relationships. The geospatial analysis methods were applied to realize effective feature extraction of the social sensing data and a grid matching process was carried out to integrate the spatiotemporal multi-source heterogeneous data. Based on this research strategy, we finally generated hourly PM2.5 concentration data at a spatial resolution of 0.01°. This method was successfully applied to the central urban area of Wuhan in China, which the optimal result of the 10-fold cross-validation R2 was 0.832. Our work indicated that the real-time check-in and traffic index variables can improve both quantitative and mapping results. The mapping results could be potentially applied for urban environmental monitoring, pollution exposure assessment, and health risk research.

Highlights

Fine particles with an aerodynamic diameter of less than 2.5 micrometers (PM2.5 ), which correspond to the “high-risk respirable convention”, as defined in [1], have aroused worldwide concern [2].The troubling thing is that 92% of the world population are exposed to PM2.5 air pollution concentration that is above the annual mean World Health Organization Air Quality Guidelines (WHO AQG) level of 10 μg/m3 [3]
We find that: (1) PM2.5 is negatively correlated with RH and TEM, which is exactly in line with a previous study in Wuhan [55]; (2) PBLH and normalized difference vegetation index (NDVI) presents the negative relationships with PM2.5, in that a low atmospheric boundary layer height is not conducive to the diffusion and dilution of PM2.5, while vegetation can clean and purify the atmospheric environment; and (3) there are almost no linear correlations between social sensing variables and PM2.5
We mainly considered remote sensing data, social sensing data, meteorological data and the spatiotemporal features of PM2.5 to estimate hourly PM2.5 in the central area of Wuhan

Summary

Introduction

The troubling thing is that 92% of the world population are exposed to PM2.5 air pollution concentration that is above the annual mean World Health Organization Air Quality Guidelines (WHO AQG) level of 10 μg/m3 [3]. In addition to the health effects, PM2.5 has significant impacts on climate change, agricultural production and ecological environment [4]. A number of studies have explored the effects of PM2.5 on health and evaluated population exposure to PM2.5 , based on a continuous distribution of PM2.5 [5,6]. The results showed that the accuracy of the PM2.5 concentration estimation has a great impact on the research conclusions, and the spatiotemporal PM2.5 distribution data are very important basic data. Generating accurate fine spatiotemporal mapping of PM2.5 concentration is important to meet the practical demand

Methods

Results

Conclusion