SPATIO-TEMPORAL VARIATION OF PM2.5 CONCENTRATION IN CHINA FROM 1998 TO 2016 AND ITS RESPONSE TO ECONOMIC DEVELOPMENT

China's economy has experienced rapid development in the past few decades, and economic development has also brought serious pollution problems, which has attracted wide attention of scholars at home and abroad. Based on the data of global PM2.5 remote sensing products and China's economic development from 1998 to 2016, the temporal and spatial variations of PM2.5 concentration in China from 1998 to 2016 were analyzed, and the response of PM2.5 concentration in China to economic development was studied. The results showed that the average annual PM2.5 concentration in 1998-2016 showed the spatial distribution characteristics of high in the East and low in the west; during 1998-2016, PM2.5 increased significantly in most regions, but decreased significantly in Inner Mongolia, Shaanxi, Ningxia and Gansu, while PM2.5 did not change significantly in some parts of the central region; during 1998-2007, PM2.5 concentration in most regions of China experienced rapid economic development. The concentration of PM2.5 in a few areas such as Inner Mongolia decreased significantly, while that in Yunnan, Sichuan and Inner Mongolia did not change significantly. During the 10 years of economic slowdown in China (2008-2016), the downward trend of PM2.5 concentration in China was expanding. The concentration of PM2.5 in the central and southern regions decreased or did not change significantly, except in the northwest and a few northeast regions. The change of PM2.5 concentration responds obviously to economic development, but the response of different regional economic development to the change of PM2.5 concentration is different. * Corresponding author


INTRODUCTION
In recent years, the problem of air pollution caused by China's rapid economic development has attracted extensive attention from all walks of life. In particular, fine particulate matter PM 2.5 has great influence on atmospheric visibility and air quality (Lin et al, 2016), and high concentration PM 2.5 is also the main cause of smog (Liu et al, 2014;Huang et al,2014;Zhao et al,2013). In addition, it poses a great threat to human health (Wang et al,2012;Han et al,2014;Zhang et al,2014). Since the aerodynamic diameter is less than 2.5µm, PM 2.5 can stay in the air for a long time and adhere to harmful substances. Prolonged exposure to PM 2.5 will lead to human health problems and even increase human mortality (Bell et al,2007;Pope et al,2002;Chen et al,2013). In 2013, the International Cancer Research Institute has listed PM 2.5 as a human carcinogen (Yang et al,2015;Chun et al,2010).
Therefore, a better and clearer understanding of the temporal and spatial changes of PM 2.5 is the basis for studying the impact of air pollution and is also helpful to take effective measures against air pollution (Bin et al, 2019).
Currently, there are mainly two ways to obtain PM 2.5 concentration data: ground monitoring and remote sensing inversion. Traditional ground monitoring methods include air sample analysis and automatic monitoring by station instruments. This method can obtain high-precision local PM 2.5 pollution data, but due to the limitation of its coverage, there is no way to reflect the PM 2.5 pollution situation in the whole region (Wang, 2016). However, remote sensing inversion has just made up for this defect. The most widely used method is to use satellite remote sensing atmospheric aerosol optical thickness (AOD) data to retrieve the PM 2.5 concentration on the ground (Zhang et al, 2005). At present, the commonly used inversion methods include: dark pixel method, structural function method and Gao Fancha surface method (Li et al, 2012). Dark pixel method is used to retrieve aerosol optical thickness over dense vegetation. For example, Pennina et al.
used dark pixel method to retrieve aerosol optical thickness The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-3/W10, 2020 International Conference on Geomatics in the Big Data Era (ICGBD), 15-17 November 2019, Guilin, Guangxi, China This contribution has been peer-reviewed. https://doi.org/10.5194/isprs-archives-XLII-3-W10-49-2020 | © Authors 2020. CC BY 4.0 License. (Peng et al, 2008). The structural function method uses dry and clean aerosol information as the background, and obtains the aerosol optical thickness on the pollution day from the change of the projection function. For example, Zhu Lin and the like use the structural function method to retrieve the pixel interval setting in the aerosol optical thickness (Zhu et al, 2016). The Gao Fancha surface method is used to retrieve the aerosol optical thickness of two regions with relatively close spatial positions and stable atmospheric optical characteristics. For example, Sun Lin used Gao Fancha surface method to retrieve the aerosol particle spectrum (Sun, 2006).
Since satellite remote sensing inversion can retrieve the distribution characteristics of PM 2.5 concentration in the whole region, based on the global annual average PM 2.5 remote sensing product data and China's economic development data from 1998 to 2016, this study uses trend analysis and significance test methods to study the temporal and spatial variation characteristics of PM 2.5 concentration in China and preliminarily discusses its response to economic development.

Survey of Research Area
China is located in the east of Asia and the west coast of the Pacific Ocean. The territory is vast and the total area is about 9.6 million square kilometers. China's terrain is high in the West and low in the east, with a stepped distribution.
Mountainous and plateau areas are vast. The distance between the East and the west is about 5000 kilometers, and the coastline of the mainland is more than 18,000 kilometers long.

Research Methods
In this study, the downloaded aerosol optical thickness product data are preprocessed in ARCGIS software. Firstly, the

Spatial Distribution Characteristics of PM 2.5 in China
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-3/W10, 2020 International Conference on Geomatics in the Big Data Era (     have also played a good role in environmental governance.