Characteristics and source analysis of greenhouse gas concentration changes at Akedala Station in Central Asia

Mole fractions of atmospheric carbon dioxide (CO2), methane (CH4), nitrous oxide (N2O), and sulfur hexafluoride (SF6) have been continuously measured since September 2009 at the Akedala Station (47°06′N, 87°58′E, 563.3 masl) in China. The station is located in the Central Asia and northwest of China, and it is the only station in that region with background conditions for long-term greenhouse gas observations. Characteristics of the mole fractions, growth rates, and influence of long-distance transport were studied considering data from September 2009 to December 2019. The greenhouse gases concentrations at Akedala Station show a trend of year-on-year growth, with CO2 concentrations ranging from 389.80 × 10−6 to 408.79 × 10−6 (molar fraction of substances, same below), CH4 concentrations ranging from 1890.07 × 10−9 to 1976.32 × 10−9, N2O concentrations ranging from 321.26 × 10−9 to 332.03 × 10−9, and SF6 concentrations ranging from 7.04 × 10−12 to 10.07 × 10−12, the growth rate of which is similar to the decadal average growth rate in the Northern Hemisphere. There exist obvious seasonal variations, with CO2 concentrations showing high in winter and low in summer and CH4 showing a distinct “W”-shaped trend while N2O and SF6 showing little difference between the four seasons. A relatively strong correlation and homology exist among the four greenhouse gases except in summer, and the analysis based on backward trajectories model shows that the Akedala Station is influenced by the airflow from northwest or southwest throughout the year. The Akedala Station is an important atmospheric background station in Central Asia, and its greenhouse gas concentration levels and variation characteristics are significantly different from those of the background stations in the monsoon region. Its degree changes are closely related to local source emissions, monsoon transport, and atmospheric photochemical processes.


Introduction
In recent times, the technology of human society has developed very rapidly. With the accelerating global industrialization following the industrial revolution, the concentrations of greenhouse gases in the atmosphere have shown a significant trend of increase (Hu and Sun 2021;WMO 2020). Carbon dioxide (CO 2 ), methane (CH 4 ), nitrous oxide (N 2 O), and sulfur hexafluoride (SF 6 ) are the gases in the atmosphere with long life and potential to heat up (WMO 2020). Greenhouse gas emissions from human production activities have become an important driver of climate change. The increase in global temperature is caused by the increase in greenhouse gas concentration, which brings about a series of extreme weather events such as droughts, floods, cold waves, and sea level rise, leading to a negative impact on people's production and life and bringing great damage to the natural environment Ou-Yang et al. 2014;He et al. 2020). To achieve the 2 °C warming target of the Paris Agreement, the study of the characteristics and sources of long-term changes of greenhouse gases in typical regions is essential for formulating energy conservation and emission reduction policies and actively responding to climate change such as global warming.
At present, domestic and foreign atmospheric background stations have carried out continuous observations of greenhouse gas concentrations for many years (Lioubimtseva and Henebry 2009), and long-term atmospheric composition observations are necessary to assess the impact of human activities on climate and the environment, and to provide basic data for regional climate change research work. The study of CO 2 and CH 4 concentration levels and variation characteristics from 2009 to 2013 at Shangdianzi atmospheric background station in North China background region and Longfeng Mountain atmospheric background station in Northeast China background region shows that there are obvious differences in CO 2 and CH 4 variations between North China and Northeast China (Dong et al. 2016;Fang et al. 2017). Kim et al. (2020) analyzed the characteristics and sources of CO 2 and CH 4 concentrations in the southern highlands of Korea, and found that CH 4 concentrations showed a significant upward trend and seasonal variation, and contributed to the development of environmental protection measures by using a backward trajectory model to identify its winter sources as Mongolia, northeastern China, and industrial cities in northwestern Korea. The Valiguan Global Atmospheric Background Station represents the atmospheric background characteristics of the Asian and European continents, and the changes in greenhouse gas concentrations are an important data support for global climate change (Xu et al. 2018). While the atmospheric background stations in Asia and Europe show an uneven distribution, the number of atmospheric background stations in Europe and the eastern monsoon region of Asia is high and concentrated in the research work (Cheng et al. 2015). Studies on the concentration characteristics and sources of greenhouse gases in Central Asia and the arid zone of Northwest China are limited, and the time scales of the studies are relatively short, and there are insufficient studies on the correlation of atmospheric CO 2 , CH 4 , N 2 O, and SF 6 concentration changes and their sources.
Central Asia is one of the most important arid regions in the world, and its climate characteristics are significantly different from those of East Asia and Europe, with interspersed mountain basins, deserts, and oases Li et al. 2017). At the same time, the Central Asian region is the upper reaches of China's weather and the core area of the "Belt and Road" construction. Although important progress has been made in climate and environmental research on precipitation and temperature in the Central Asian region (Sorg et al. 2012), there is insufficient research on the characteristics of long series changes of greenhouse gases such as CO 2 . The Akedala National Atmospheric Background Station is one of the six regional atmospheric background stations established by the China Meteorological Administration, located in the Gobi of the Altai region of Xinjiang, China, and the background parameters have been scientifically demonstrated to be representative of the background conditions in Central Asia and the northern border of Xinjiang (Xu and Tang 2004). This paper presents the long-term trends and seasonal variations in Akedala atmospheric CO 2 , CH 4 , N 2 O, and SF 6 concentrations and discusses the possible influencing mechanisms during 2009 to 2019. The linear trend analysis method, the contrastive and statistical analysis method, and Mann-Kendall method are applied to explore characteristics of the time series and seasonal trends of greenhouse gases at this station and tested whether abrupt change exists. In addition, Pearson correlation coefficient is used to determine the correlation and homology among the four greenhouse gases and backward trajectories model is applied to explore the source of greenhouse gases at Akedala Station in different seasons. Our study is expected to provide valuable information for the study of the longterm atmospheric composition evolution and the impact of atmospheric composition on the climate, ecology, and environment of Central Asia and Northwest China.

Overview of the station and data sources
Central Asia is a globally important arid zone, which has formed a unique climate system due to its unique topography, substratum type, and water cycle variability characteristics. During the global warming process, Central Asia has shown a jump in temperature increase in the last 20 years, which has attracted the attention of many scholars (Chen et al. 2016;Hu et al. 2014). Akedala Background Atmosphere Station (AKDL, 47°06′N, 87°58′E, 563.3 m above sea level) is located in the Gobi wetland in the eastern part of Fuhai County, Altay Prefecture, Xinjiang, also the northern margin of Junggar Basin. Within 50 km around the station are alluvial flat land and Gobi. The nearest village is 5 km away from the site, with few people around and no trails of complex human activities (Feng et al. 2021). Akedala Station is located in an area of temperate continental arid climate, which is a typical arid and semi-arid region. The climate in this region is characterized by hot summer with little rain, large temperature difference between day and night, relatively cold winter, short spring and autumn, and frequent windy days. The vegetation type of its underlying surface is desert steppe, which is representative in North Xinjiang and in the hinterland of Asia. Westerly and northwesterly winds prevail throughout the year, with more easterly winds in winter. Akedala Station is located in the hinterland of Asia and Europe, far from the ocean, and belongs to the westerly wind belt control area, which is also the upstream area of Chinese weather (Wang et al. 2012). The geographical location of Akedala Station is shown in Fig. 1.

Data sources and quality control
The sampling point is located at Akedala Regional Background Atmosphere Station, which is one of the field test bases of the China Meteorological Administration. The gas sampling equipment consists of portable pumping installations and glass flasks, and the air inlet of Akedala Station is located on a steel tower at a height of 50 m without obstructions. According to the working standards of the sampling (Meteorology Industry Standard of the People's Republic of China QX/T 164-2012), the sampling was carried out at 14:00 every Tuesday (Beijing Time), and the wind speed of the sampling room was ensured to be higher than 2 m·s −1 to avoid the influence of anomalous weather phenomena such as rain, snow, and dust. During the sampling process, the flasks were firstly connected in series, inflated with local air for more than 10 min, and then pressurized to 1.2-1.5 atm for sampling. Gas samples using flask sampling were mailed to the Atmospheric Constituent Laboratory of the Chinese Academy of Meteorological Sciences for analysis according to the working standards, whose technical parameters meet the GWO/GAW quality goals as well as the requirements of background atmospheric constituent analysis. The samples were sent to the greenhouse gas laboratory of the China Meteorological Administration for the analysis of trace components of greenhouse gases. The concentrations of CO 2 , CH 4 , N 2 O, and SF 6 were analyzed by utilizing Picarro G2401 based on Optical Cavity Decay Spectroscopy (CRDS) and Gas Chromatograph (GC) Agilent 7890, and the measurements were in line with the data quality goals of World Meteorological Organization. The data used in this study were obtained from the China Meteorological Administration after quality control and the study period was from September 2009 to December 2019.
The data for the same period of the Waliguan Global Background Atmosphere Station, Hohenpeissenberg Global Background Atmosphere Station in Germany, and Tae-ahn Peninsula Regional Background Atmosphere Station in Korea were obtained from the WMO World Data Centre for Greenhouse Gases (https:// gaw. kishou. go. jp/) for the comparative analysis of time series changes. The data sources and location of each station are shown in Table 1.

Research methods
The trend analysis method, contrastive, and statistical analysis method were used for analysis of time series characteristics of greenhouse gases, and Mann-Kendall method was used to study abrupt change (Mann 1945;Kendall 1975;Wei 2007;Xing et al. 2015;Yang et al. 2018). Pearson correlation coefficient was used to research the correlation of between greenhouse gases, and HYSPLIT backward trajectory analysis was adopted to study of potential source regions.  Figure 3 shows that the concentrations of CO 2 , CH 4 , N 2 O, and SF 6  showed a clear increasing trend at the four stations, and there were differences in their average annual concentration levels. The CO 2 concentration ranged from 387.38 × 10 −6 to 411.06 × 10 −6 at Waliguan Station, from 389.80 × 10 −6 to 408.79 × 10 −6 at Akedala Station, from 387.88 × 10 −6 to 413.21 × 10 −6 at Hohenpeissenberg Station, and from 390.24 × 10 −6 to 418.67 × 10 −6 at Tae-ahn Peninsula Station (Fig. 3). The CO 2 concentration at TAP station was significantly higher than the other three stations, with the concentration in these stations ranked from high to low as TAP > HPB > AKDL > WLG. And the growth rates were 2.37 × 10 −6 year −1 at WLG, 1.90 × 10 −6 year −1 at AKDL, 2.53 × 10 −6 year −1 at HPB, and 2.85 × 10 −6 year −1 at TAP (Fig. 4a). The growth rate of CO 2 in the Northern Hemisphere was 2.27 × 10 −6 year −1 in the last decade with the growth rate of Akedala Station slightly lower than the average level, which may be related to the location of the background station and the source and sink. The WLG Station is located on the Qinghai-Tibet Plateau, which is sparsely populated and is less affected by the source region of emission (Liu et al. 2014). The HPB Station is located in the mountainous region of southern Germany, and it is an important crop production area. And the Altay Prefecture of Xinjiang, where the AKDL Station is located, has a long heating period, which may be a factor influencing the concentration of CO 2 , while the TAP Station is located in the coastal area of South Korea, with a relatively developed economy and more human activities (Chung and Tans 2000;Lee et al. 1993).
The CH 4 concentration variation range was from 1899.92 × 10 −9 to 1991.08 × 10 −9 at WLG, from 1890.07 × 10 −9 to 1976.32 × 10 −9 at AKDL, from 1912.42 × 10 −9 to 1984.67 × 10 −9 at HPB, and from 1912.35 × 10 −9 to 1991.07 × 10 −9 at TAP (Fig. 3). Overall, the CH 4 concentrations at HPB Station and TAP Station were slightly higher than those at WLG Station and AKDL Station. The growth rates were 9.12 × 10 −9 year −1 at WLG, 8.62 × 10 −9 year −1 at AKDL, 7.22 × 10 −9 year −1 at HPB, and 7.87 × 10 −9 year −1 at TAP (Fig. 4b) with a decadal average growth rate of 7.3 × 10 −9 year −1 for the Northern Hemisphere (WMO 2020), and the growth rates of WLG Station and AKDL Station were significantly higher than those of HPB Station and TAP Station. The variation and growth of CH 4 concentration were affected by source region, biological factors, and photochemical factors (Zhang 2011). HPB Station and TAP Station are located in wet areas, which are major crop production areas and urban zones, while WLG Station and AKDL Station are relatively less influenced by source regions, so differences in concentrations arose. The location of HPB Station and TAP Station might have strong photochemical action, resulting in relatively low growth rates of CH 4 concentration, and the relatively high growth rates of CO 2 were due to the formation of CO 2 from CH 4 under photochemical action (Zhang et al. 2021). The variation range of N 2 O concentration was from 322.88 × 10 −9 to 332.43 × 10 −9 at WLG, from 321.26 × 10 −9 to 332.03 × 10 −9 at AKDL, from 323.46 × 10 −9 to 333.16 × 10 −9 at HPB, and from 323.35 × 10 −9 to 333.42 × 10 −9 at TAP (Fig. 3), with the growth rate of 0.95 × 10 −9 year −1 at WLG, 1.08 × 10 −9 year −1 at AKDL, 0.96 × 10 −9 year −1 at HPB, and 1.01 × 10 −9 year −1 at TAP (Fig. 4c), not significantly different from the decadal average growth rate of 0.96 × 10 −9 year −1 in the Northern Hemisphere. The SF 6 concentration ranged from 6.97 × 10 −12 to 10.19 × 10 −12 at WLG, from 7.04 × 10 −12 to 10.07 × 10 −12 at AKDL, from 7.01 × 10 −12 to 10.21 × 10 −12 at HPB, and from 7.14 × 10 −12 to 10.36 × 10 −12 at TAP (Fig. 4d). The growth rate of SF 6 at Akedala Station is much higher than those of the other three gases. SF 6 is a synthetic gas widely used in the electric power industry, whose chemical property is stable. The trend of rapid increasing concentration of SF 6 at Akedala Station is the same as those of the other three atmospheric background stations, with an increase of 42.90% during 11a. SF 6 comes from power equipment such as high-voltage switches and circuit breakers. With the economic development, the demand for power equipment grows, leading to the increasing SF 6 emissions. Moreover, SF 6 has a long duration and strong stability, which also causes its accumulation in the atmosphere. Although the contents of N 2 O and SF 6 are relatively low in the atmosphere, it is quite evident that they had a clear rising trend in the last decade, and the results from four stations had a strong consistency on this point. More attention is needed for the variation of N 2 O and SF 6 concentrations in the atmosphere.

Seasonal variation of greenhouse gases concentrations
Affected by biological and non-biological sources, the emissions of greenhouse gases showed a pronounced seasonal variation. Figure 5 represents the variation of the average values in every month of greenhouse gases concentrations at Akedala Station over the years. At this station, CO 2 and CH 4 concentrations showed more obvious seasonal variation while N 2 O and SF 6 concentrations showed less. The CO 2 concentration rose to the highest level in March and April, with a highest value of 406.0 × 10 −6 in April, and then fell to the lowest level in August, with the lowest value of 389.0 × 10 −6 . The change of CH 4 concentration presented an approximate "W"-shaped trend, with a downward trend from January to June, and a rising to a peak in July and August, followed by a gradual decrease till to the bottom in September, and an increase afterwards. The "W"-shaped trend was roughly similar to the trend of CH 4 concentrations at stations like Shangdianzi Station and Longfenshan Station (Mueller et al. 2018;Dai and Dai 2018). The CH 4 concentration showed two peaks in July and January of the following year (Zhang 2011). The CH 4 concentration at Akedala Station reached 1924.6 × 10 −9 in August and 1975.6 × 10 −9 in January. The monthly average concentration of N 2 O and SF 6 In China, winter is from December to February, spring is from March to May, summer is from June to August, and winter is from September to November. Based on the average values in every month of greenhouse gases concentrations at Akedala Station, the analysis of the data of different seasons was carried out. Figure 6 shows the seasonal variation of greenhouse gases at Akedala Station. It can be seen that there were obvious seasonal differences in CO 2 and CH 4 concentrations, while N 2 O and SF 6 concentrations did not vary significantly from season to season. The CO 2 concentration at Akedala Station in the four seasons was in the order of winter, spring, autumn, and summer from high to low, which were roughly consistent with the variation of CO 2 concentration at Shangdianzi Station ). Similar to Shangdianzi Station, Akedala Station is located in the northern part of Xinjiang where winter is cold and residents need to heat their homes by means like burning coal. The burning of fossil fuels increased the CO 2 content in the atmosphere in winter, while in summer, ground living plants consumed CO 2 in the atmosphere through comparatively strong photosynthesis, so CO 2 concentration showed obvious seasonal variation (Ou- Yang et al. 2014). CH 4 concentration showed an overall trend of high in winter and autumn and low in spring and summer, and this trend was opposite to the seasonal variation at Waliguan Station and Shangdianzi Station (Zhang et al. 2013;Fang et al. 2017). The air masses in winter and autumn were mostly influenced by the northwest airflow, and in winter, the airflow from the northern economic zone increased significantly. In summer, both rainfall and UV light were more intense, contributing to the photochemical action of CH 4 (Xu et al. 2011).
N 2 O concentrations in summer were slightly lower than those in the other three seasons, while SF 6 concentrations remained essentially unchanged in all seasons, which were mainly influenced by various factors such as photochemical action and long-range air mass movement.
The seasonal variation of greenhouse gases at Akedala Station reflected the influence and effect of the periodic variations in the terrestrial biosphere in typical regions of Northwest China and Central Asia on the atmosphere.

Mann-Kendall test on greenhouse gases concentrations
Mann-Kendall method is frequently used to analyze the internal characteristics of hydrological time series. The results of M-K test for greenhouse gases concentrations at Akedala Station are shown in Fig. 7. The values of U F and U B of CO 2 concentrations at Akedala Station were both higher than 0 from 2009 to 2017, while the difference between U F and U B appeared from 2017 to 2019, and the value of U B started to be lower than 0 with a clear rising trend of the annual average concentration over the 11 years, reaching a maximum in 2018 and then a decrease in 2019. The change trends of CH 4 , N 2 O, and SF 6 concentrations were roughly the same, and U F and U B were higher than 0 during the observation period, indicating that their concentrations showed a rising trend. The results of the It is worth noting that CO 2 concentration might have shown a decreasing trend since 2019, which might be related to the strong control policies in the source regions. Since 2018, the Altay Prefecture has realized electric heating in the whole region for environmental protection and ecological construction. In addition, some progress has been made in mine management, and work related to the disposal of abandoned mines has been started in order to realize the goal of building a beautiful China.
To further validate the relationship between GDP and greenhouse gases growth, the correlation between Xinjiang's quarterly GDP and the quarterly average concentrations of CO 2 , CH 4 , N 2 O, and SF 6 has been analyzed (Fig. 8). The first quarter is from January to March, the second quarter April to June, the third quarter July to September, and the fourth quarter October to December. In addition, the correlation results show that the regional quarterly GDP has a good correlation with the quarterly average concentrations of CO 2 , CH 4 , N 2 O, and SF 6 . Moreover, the correlation between concentrations of CH 4 , N 2 O, and SF 6 and regional quarterly GDP is better than that of CO 2 , indicating that the increasing greenhouse gases in the atmosphere are closely related to human activities. The correlation between CO 2 and regional quarterly GDP is relatively weak, which may be related to the complex source-sink mechanism of CO 2 itself in different seasons. And the regional quarterly GDP only represents some sources of CO 2 emissions, which does not take into account the sinks of CO 2 in different seasons.

Correlation analysis of greenhouse gases concentrations
To better explore the potential source regions of greenhouse gases at Akedala Station, it is necessary to make further investigation of the correlation between various greenhouse gases concentrations. Correlation analysis was performed on the monthly averages of greenhouse gases concentrations at the Akedala Station from 2009 to 2019, and the numbers of data involved in the analysis totaled 32 of spring, 34 of summer, 34 of autumn, and 32 of winter. Pearson correlation Fig. 7 The results of M-K test for greenhouse gases concentrations at Akedala Station (a CO 2 ; b CH 4 ; c N 2 O; d SF 6 ) coefficients were calculated between the greenhouse gases concentrations throughout the year and in each season, and the results are presented in Table 2. Pearson correlation between greenhouse gases was more significant in spring, autumn, and winter than in summer at the Akedala Station, which was particularly evident between CO 2 and CH 4 . Pearson correlation coefficients between CO 2 and CH 4 were 0.746 (p < 0.01), 0.880 (p < 0.01), 0.328 (p < 0.05), 0.870 (p < 0.01), and 0.860 (p < 0.01) for a year-round, spring, summer, autumn, and winter, respectively. From the data, it is obvious that CO 2 and CH 4 were significantly correlated except in summer. The reason for the exception in summer was mainly because the CO 2 concentration in the atmosphere in summer was more influenced by source-sink effect of terrestrial ecosystems, while the photochemical action of CH 4 was stronger in summer, and the OH free radical concentrations were higher in this season, leading to the conversion of part of CH 4 into CO 2 through photochemical action. Pearson correlations between CO 2 and N 2 O, CO 2 and SF 6 , CH 4 and N 2 O, and CH 4 and SF 6 were significant throughout the year as well as in all seasons, respectively. In summary, the Akedala background station is located at the margin of Junggar Basin, the south of which is the core economic region of northern Xinjiang. Greenhouse gases concentrations were affected by anthropogenic greenhouse gas emissions and terrestrial ecosystem. The homology of greenhouse gases was more obvious in spring, autumn, and winter, while the influence of terrestrial ecosystems and photochemical action in summer made the homology of greenhouse gases less obvious in this season. Therefore, the subsequent analysis of air mass pathways and potential source regions needs to be researched according to different seasons.

Analysis of potential source regions in different reasons
The variation of greenhouse gases concentrations at background atmosphere station is also closely related to the external air masses. In order to study the effects of longrange transport and atmospheric boundary layer conditions on the variation of greenhouse gases concentrations at Akedala Station in different seasons, a backward trajectory model was used to trace the trajectories of air masses in different seasons. Based on the result of the backward trajectory, six main trajectories were identified for each season (Fig. 9), and their pressure distribution is shown in Fig. 9. From Fig. 8, differences in the trajectories of air masses in four seasons at Akedala Station with significant seasonal transport characteristics. The Akedala Station was mainly influenced by the air mass from the northwest, accounting for 46.22%, 61.90%, 52.90%, and 72.82% of all trajectories in winter, spring, summer, and autumn, respectively (Table 3), and air pressures of most of air masses were above 800 hPa (Fig. 10). Therefore, the northwest path originating from Central Asia and parallel to the valley of the Irtysh River is the main channel for the movement of air masses. In winter, air masses 1, 2, and 3 played a dominant role, accounting for 75.80% of all trajectories (Table 3). Air mass 1 came from the northern Xinjiang, arriving at Akedala Station from the southeast, which was in line with the direction Fig. 8 The correlation between Xinjiang's quarterly GDP and the quarterly average concentrations of greenhouse gas  (Li et al. 2020). The air mass 1 might have high greenhouse gases concentrations due to the need to burn biomass fuel for heating in northern Xinjiang. Air mass 2 arrived here after a long distance. It came from the industrial zone of eastern Kazakhstan, where heavy industries such as mining and non-ferrous metal smelting were concentrated. Under the combined effect of westerly winds, it was transported eastward to arrive the Akedala Station, potentially bringing more greenhouse gases. Air mass 3 came from Mongolia. Air masses 4 and 6 came from the northern Xinjiang, passing through cities with developed industries such as Karamay, and air mass 5 came from the southern part of Russia after a long-distance transport. Except for air mass 5, the air pressures of the rest of the air masses were all above 750 hPa (Fig. 10).
In the spring, air masses 4, 5, and 6 played a dominant role, and they accounted for 72.07% of all trajectories (Table 3). Air masses 4 and 5 came from the eastern part of Kazakhstan, while air mass 6 came from the northern part of Xinjiang, located to the southeast of Akedala Station. Therefore, the greenhouse gas concentrations in the spring were mainly affected by industrial emissions transported from eastern Kazakhstan for a long distance and from northern Xinjiang for a short distance, which might carry some of the greenhouse gases to Akedala Station (Zhao et al. 2021). Air mass 1 crossed the industrially developed cities of Karamay, while air masses 2 and 3 came from the south of Russia, passed the border of Mongolia, and reached Akedala Station Fig. 9 Cluster-mean backtrajectories by using the hybrid single-particle Lagrangian integrated trajectory (HYSPLIT) model in Akedala Station from January 2010 to December 2019 from the northeast. The air pressures of these two air masses were below 750 hPa, which were relatively low, and the air pressures of other air masses were above 750 hPa (Fig. 10). In summer, air masses 1, 2, and 4 became the dominant, accounting for 82.37% of all trajectories. Air masses 1 and 2 both came from the eastern part of Kazakhstan, accounting for about 67.23% of all trajectories (Table 3), and air mass 4 came from southern Russia and reached here after a long distance. Because the Akedala Station is located in the arid zone of Central Asia, far from the ocean, it is significantly affected by the westerly wind belt in summer, and almost all air masses come from Central Asia or further away from Russia. There existed noticeable difference in terms of pressure distribution. Air masses 2 and 4 were dominant and air pressures of them were mainly above 860 hPa, while those of others were basically below 860 hPa (Fig. 10).
In autumn, the predominant air masses were numbers 2, 4, and 6, accounting for about 68.47% of all trajectories (Table 3), mainly from the eastern part of Kazakhstan. Air masses 2 and 4 were formed when airflow from distant Kazakhstan reached the Akedala Station under the influence of the westerly wind, so their air pressures were relatively close to each other. In addition, air masses 1 and 3, which came from the northern Xinjiang, accounted for 27.19% of the total trajectory (Table 3). Air mass 5 reached the Akedala Station over the long distance of Altai Mountains. Since southern Russia was sparsely populated, the amount of greenhouse gases carried in the airflow might be relatively low, and the air mass pressure was significantly lower (Fig. 10).

Conclusion
CO 2 , CH 4 , N 2 O, and SF 6 , as common greenhouse gases, are among the key contributors to global climate change and extreme weather events. In this study, greenhouse gases concentrations at Akedala Station from 2009 to 2019 were analyzed, and conclusions were drawn as follows.
(2) Changes in greenhouse gas concentrations were closely related to factors including biological and non-biolog- ical sources, atmospheric boundary layer conditions, and source regions. The results of the M-K test for greenhouse gases concentrations at Akedala Station showed a significant rising trend during the observation period. The results indicated that CO 2 concentration showed a distinct increase since 2014, while CH 4 , N 2 O, and SF 6 showed a clear increase since 2012. The greenhouse gases concentrations at Akedala Station showed obvious seasonal variations, with CO 2 concentration being high in winter and low in summer, CH 4 concentration showing a clear "W"-shaped trend, and N 2 O and SF 6 concentrations showing little difference between the four seasons. A strong correlation and homology were observed between the different greenhouse gases concentrations at the Akedala Station in spring, autumn, and winter, while a poor homology was observed in summer. The Akedala Station is critically located in the core area of Central Asia and can reflect the trend of background atmospheric background concentration in Central Asia, which is of great significance for climate change research work in Central Asia. (3) The results of the backward trajectory analysis showed that the Akedala Station was influenced by strong northwesterly airflow in all seasons, with eastern Kazakhstan being one of the main sources of the airflow. And the station was also affected by airflow from the northern Xinjiang. The airflow from the south of Russia was noticeably stronger in summer, which might have an impact on the greenhouse gases concentrations at Akedala Station.