Xiaohu XIONG^1,2,,,
Weijian ZHOU^1,2,3,4,5,
Shugang WU^1,2,
Peng CHENG^1,2,
Hua DU^1,2,
Yaoyao HOU^1,2,
Zhenchuan NIU^1,2,4,
Peng WANG^1,2,
Xuefeng LU^1,2,
Yunchong FU^1,2

Corresponding author: Xiaohu XIONG,xiongxh@ieecas.cn;

1.State Key Laboratory of Loess and Quaternary Geology, Institute of Earth Environment, Chinese Academy of Sciences, Xi'an 710061, China
2.Shaanxi Key Laboratory of AMS Technology and Application, Xi'an AMS Center, Xi'an 710061, China
3.Xi’an Jiaotong University, Xi’an 710049, China
4.Chinese Academy of Sciences Center for Excellence in Quaternary Science and Global Change, Xi’an 710061, China.
5.Open Studio for Oceanic-Continental Climate and Environment Changes, Pilot National Laboratory for Marine Science and Technology (Qingdao), Qingdao 266237, China
Manuscript received: 2019-11-01
Manuscript revised: 2020-01-06
Manuscript accepted: 2020-01-09
Abstract:The need for atmospheric carbon dioxide (CO₂) reduction in the context of global warming is widely acknowledged by the global scientific community. Fossil fuel CO₂ (CO_2ff) emissions occur mainly in cities, and can be monitored directly with radiocarbon (¹⁴C). In this research, annual plants [Setaria viridis (L.) Beauv.] were collected from 26 sites in 2013 and 2014 in the central urban district of Xi'an City. The Δ¹⁴C content of the samples were analyzed using a 3 MV Accelerator Mass Spectrometer, and CO_2ff concentrations were calculated based on mass balance equations. The results showed that the CO_2ff mixing ratio ranged from 15.9 to 25.0 ppm (part per million, equivalent to μmol mol^?1), with an average of 20.5 ppm in 2013. The range of measured values became larger in 2014, from 13.9 ppm to 33.1 ppm, with an average of 23.5 ppm. The differences among the average CO_2ff concentrations between the central area and outer urban areas were not statistically significant. Although the year-to-year variation of the CO_2ff concentration was significant (P < 0.01), there was a distinctly low CO_2ff value observed in the northeast corner of the city. CO_2ff emissions from vehicle exhaust and residential sources appeared to be more significant than two thermal power plants, according to our observed CO_2ff spatial distribution. The variation of pollution source transport recorded in our observations was likely controlled by southwesterly winds. These results could assist in the optimal placement of regional CO₂ monitoring stations, and benefit the local government in the implementation of efficient carbon emission reduction measures.
Keywords: fossil fuel CO₂,
¹⁴C,
Xi’an, carbon emissions reduction,
Setaria viridis

HTML

--> --> -->

Atmospheric carbon dioxide (CO₂) reached a new level in 2017 (405.5 ppm), equal to 146% of the pre-industrial level, and is mainly attributed to anthropogenic fossil fuel emissions (WMO, 2018). Global net anthropogenic CO₂ emissions should decline by about 45% from 2010 levels by 2030 to limit global warming to below 1.5°C (IPCC, 2018). Therefore, reliable methods to quantify anthropogenic CO₂ emissions are essential. Two methods have been commonly applied: “bottom-up” and “top-down.” The former is based on activity data and emission factors (IPCC, 2006). The reliability of this method not only depends on the quality of statistical activity data, but also on the accuracy of emission factors (Liu et al., 2015). The estimated uncertainty of the “bottom-up” method has been estimated to range from a few percent to about 50% on an individual national scale (Andres et al., 2012); and may be as high as 100% on an urban scale (Gurney et al., 2009; Peylin et al., 2011; Turnbull et al., 2011). The “top-down” method usually measures the atmospheric CO₂ mixing ratio from ground stations (Levin et al., 2003), aircraft (Turnbull et al., 2016) or satellites (Buchwitz et al., 2017). This approach can produce high temporal and spatial resolutions using CO₂ data. However, observed CO₂ fluctuations do not directly measure fossil fuel CO₂ (CO_2ff) emissions, as both CO_2ff and natural CO₂ (e.g., from heterotrophic respiration) can contribute to atmospheric CO₂ variations.
Radiocarbon (¹⁴C) is a unique and powerful tracer for CO_2ff and a useful tool to verify “bottom-up” inventory estimates (Levin et al., 2003; Nisbet and Weiss, 2010; Turnbull et al., 2016). Δ¹⁴C in plants has been widely used to represent mean daytime atmospheric Δ¹⁴C during their growth period for quantification of atmospheric CO_2ff (e.g. Lichtfouse et al., 2005; Bozhinova et al., 2013; Turnbull et al., 2014). Plant samples are easier to collect in a large area at low cost, as compared with air samples (Turnbull et al., 2014), and are especially advantageous when focusing on the spatial distribution of CO_2ff (Hsueh et al., 2007; Riley et al., 2008). Cities, with concentrated carbon emissions and frequently affected by serious air pollution meanwhile, have been recognized as playing a key role in the action of climate change mitigation (Rosenzweig et al., 2010; Hoornweg et al., 2011). The spatial distribution of CO_2ff indicated by plants has been investigated in a few cities (Xi et al., 2011, 2013; Djuricin et al., 2012; Zhou et al., 2014; Beramendi-Orosco et al., 2015; Niu et al., 2016); however, very few city centers have been targeted, because of the lack of particular species (e.g., corn leaves were chosen in many studies) in city centers. Interannual CO_2ff variations and their spatial distribution based on plants is also scarce (Bozhinova et al., 2016). Green foxtail samples have been successfully used to reflect CO_2ff differences in some areas in Xi’an City in 2010 (Zhou et al., 2014). In this study, green foxtail was sampled across a wide distribution of sites in the central urban area of Xi’an in 2013 and 2014 to investigate the atmospheric CO_2ff spatial distribution and its year-to-year variation. These data were used to quantify the spatial distribution and to estimate the influencing factors. The main goal was to provide quantitative information that can be used to improve local carbon emission reduction measures.

2

-->

2.1. Study sites and materials

Xi’an is a famous tourist city and a significant node city of “the Belt and Road” initiative in the west of China. Its resident population reached 10 million in 2018. This city experiences a semiarid continental monsoon climate, with a mean annual temperature of 13.0°C–13.7°C. The mean annual rainfall is 522.4–719.5 mm, with over 80% of the precipitation falling in May to October. The winds are characteristically mild in Xi’an, with a prevailing northeasterly wind direction.
Annual plant samples [green foxtail, Setaria viridis (L.) Beauv.] from 26 sites inside the Xi’an Bypass Highway were collected in parks or scenic spots in October of 2013 and 2014. The distribution of these sites is shown in Fig. 1. Green foxtail is a C4 plant species, widely distributed in the city green belt as weeds. Its growing period is April to September in Xi’an City; thus, it reflects average atmospheric CO_2ff in the late spring–summer–early autumn in general. Every location sample was mixed by two or three individual plants with seed maturation, and all of them grew in open areas with good ventilation, as far as possible from urban trunk roads. Plant materials were air-dried before pretreatment.
Figure1. The distribution of green foxtail sampling sites in the Xi’an urban area. Twenty-six sites (solid circles) were in the parks or in the scenic spots in the central urban district of Xi’an City. Two thermal power plants (solid triangles) were in the east and west area outside of the second Ring Road. The region inside the Bypass Highway was divided into three parts by the Ring Road and the second Ring Road, namely, Area I, II and III, respectively



2 -->

2.2. Experiment

The grass leaves were soaked with 1-mol L^?1 hydrogen chloride at 70°C for 2 h to remove any adhering carbonate, then rinsed with deionized water until neutral, and dried at 60°C. The dry plant materials (4–5 mg) were mixed with copper oxide (~200 mg) in a quartz tube and combusted at 850°C for 2 h under vacuum (~100 Pa). Finally, the CO₂ was extracted using liquid nitrogen as a cryogen, and further purified with liquid nitrogen/alcohol (?75°C) to remove water. The purified CO₂ was reduced to graphite carbon using zinc as a reductant over an iron catalyst. The produced graphite was pressed into an aluminum holder and measured in a 3-MV Accelerator Mass Spectrometer (High Voltage Engineering Europa B.V.) at the Xi’an AMS Center. Oxalic acid-II (NIST SRM 4990c) and coal were used as reference and blank, respectively. The measurement precision of a single sample is about 3‰ (1σ).
The ¹⁴C results were calculated as Δ¹⁴C values, according to the equation (Stuiver and Polach, 1977)
where (¹⁴C/¹²C)_samp is the ¹⁴C/¹²C ratio of samples corrected for radioactive decay and normalized to a δ¹³C of ?25‰, and (¹⁴C/¹²C)_ref is the ¹⁴C/¹²C ratio of the reference standard.

2 -->

2.3. Calculation of CO_2ff

Atmospheric CO₂ (C_obs) can be considered to consist of three sources: (1) background CO₂ (C_bg), (2) fossil fuel-derived CO₂ (C_ff), and (3) other CO₂ sources (C_other). According to mass balance:
For time-integration, as in the case of seasonal plants, C_obs cannot be determined. Thus, Eq. (2) is substituted into Eq. (3), resulting in
The second term in this expression is usually omitted, because the error introduced is small (Turnbull et al., 2015). Equation (4) can then be simplified, and C_ff can be calculated by the following widely used formula:
In this equation, C_ff and C_bg are CO₂ derived from fossil fuel and CO₂ represented by background measurements, respectively; and Δ¹⁴C_obs is measured in the plant samples. Because there is negligible ¹⁴C in fossil fuel, the end-member value of Δ¹⁴C_ff = ?1000‰ is assumed according to the definition of Δ¹⁴C [Eq. (1)]. C_bg records were obtained from measurements made in the Mauna Loa Atmospheric Baseline Observatory in Hawaii, the United States (Tans and Keeling, 2019), and background Δ¹⁴C data were from Jungfraujoch Atmospheric Baseline Observatory in Switzerland (Hammer and Levin, 2017).

2

-->

3.1. Spatial variation of CO_2ff in 2013

CO_2ff results from all sites in 2013 are shown in Fig. 2. The range of CO_2ff was 15.6–24.6 ppm, with an average of 20.1 ± 2.4 ppm. The average inside the Ring Road (5 sites, Area I), the Ring Road to the second Ring Road (6 sites, Area II) and the second Ring Road to the Bypass Highway segment (15 sites, Area III) were 19.8 ± 1.9 ppm, 19.6 ± 1.9 ppm and 20.4 ± 2.6 ppm, respectively. There were some sites with high CO_2ff values, such as the Southwest Xi’an City Wall (SWXW), Changle Park (CLP), Yongyang Park (YYP) and Yanming Park (YMP). These sites were scattered, but mainly located in the south of the Xi’an urban area. In contrast, sites with relatively low CO_2ff values were in the north, near the Bypass Highway. For example, the City Sports Park (CSP), Yuanjiabu Village (YJB), Xi’an World Horticulture Exposition Park (XEP), and the Epang Palace Relic Park (EPRP).
Figure2. CO_2ff mixing ratio spatial distribution in 2013. The value of CO_2ff is proportional to the height of the light blue column



2 -->

3.2. Spatial variation of CO_2ff in 2014

The spatial variations of CO_2ff in 2014 are shown in Fig. 3. In 2014, the average CO_2ff of all sites was 22.1 ± 5.4 ppm. The highest CO_2ff value was 31.7 ± 1.5 ppm, and the lowest was 12.5 ± 1.5 ppm. The average CO_2ff values of the above-mentioned three regions divided by the city’s ring roads were 22.9 ± 5.5 ppm, 20.1 ± 2.5 ppm and 22.7 ± 6.1 ppm, respectively, without statistically significant differences. As shown in Fig. 3, both of the higher, and lower, CO_2ff values existed in the region inside the Ring Road, with a similar pattern within the second Ring Road to Bypass Highway area. It is interesting to note that the sites with higher CO_2ff values were widely distributed in the area adjacent the Bypass Highway, except for the northeast corner, with the two lowest values (City Sports Park and Xi’an World Horticulture Exposition Park).
Figure3. CO_2ff mixing ratio spatial distribution in 2014. The value of CO_2ff is proportional to the height of the pink column



2 -->

3.3. CO_2ff variation from 2013 to 2014

A comparison of both years showed that CO_2ff variation in 2014 was larger than in 2013; the variation coefficients were 24% and 12%, respectively. Moreover, the differences between 2013 and 2014 were statistically significant (P < 0.01, paired t-test) (Fig. 4). In detail, CO_2ff values at eleven sites increased from 2013 to 2014, and six decreased. Only nine sites showed minor changes. CO_2ff variations in Area II sites were smaller than observed for the other two areas. It is worth noting that most sites close to the Bypass Highway showed obvious increases, with two exceptional cases in the northeast: the City Sports Park and Xi’an World Horticulture Exposition Park. Further, CO_2ff from these two sites had the lowest values observed in both 2013 and 2014.
Figure4. CO_2ff variation from 2013 to 2014. The sites with a CO_2ff decrease and increase from 2013 to 2014 are shown in blue and red, respectively



2 -->

3.4. CO_2ff differences between Chinese cities

The CO_2ff mixing ratio was about 10–30 ppm in the Xi’an central urban district in 2013 and 2014, similar to that reported for 2010 (Zhou et al., 2014), and very close to downtown Beijing in 2009 (Niu et al., 2016). However, the Xi'an value was higher than that observed in some small and medium-sized cities (e.g., Lhasa, Jiuquan, Erdos, Luochuan, Linfen and Yantai); CO_2ff in these cities were generally lower than 10 ppm in 2010 (Xi et al., 2013). The likely implication is that cities with large populations have higher CO₂ emissions.

2 -->

3.5. The effect of emission sources

The dominant CO_2ff sources in the Xi’an urban district include thermal power plants, vehicle exhaust, and residential emissions. There are two thermal power plants in the investigated area, as shown in Fig. 1: Baqiao Thermal Power Plant (BTPP) in the east, and Western Thermal Power Plant (WTPP) in the west. We found that the CO_2ff mixing ratio of the sites in the downwind area (southwest) near BTPP and WTPP, such as Changle Park (CLP), Xingqing Palace Park (XQP) and Epang Palace Relic Park (EPRP), was not especially elevated, when compared to results from the other regions. This might be attributable to enhanced diffusion of exhaust gas during transport from the very high power plant vents to the ground (Zhang, 2013). Our results suggest, therefore, that these two power plants are not primary contributors to CO_2ff spatial variation.
Vehicle exhaust and residential emissions are non-point sources of CO_2ff emissions. In general, they are widely distributed within the urban area. In addition, vehicle exhaust is a mobile emission source, which is distinct from residential emissions. Nevertheless, both are closely associated with population density. The two sites (CSP and XEP) with the lowest measured CO_2ff values were located in an area in the northeast with low population density. In contrast, relatively high CO_2ff values were frequently observed in the south of Xi’an, which includes two of six population centers in this district (Mi et al., 2014). The rapid increase of motor vehicles during 2013–14 (from 1.6349 million to 1.9260 million) but overall slow residential population growth (from 5.8060 million to 5.8716 million) suggested that CO_2ff increases observed for many sites were likely related to rapidly growing regional traffic conditions. For example, the Ring Road was under construction, with the addition of three tunnels, to improve the traffic in 2013. This project was completed in April 2014, at which point traffic flow increased rapidly. As a result, at sites SWXW and NWXW, adjacent to the Ring Road, the CO_2ff mixing ratio rose from ~20 ppm to ~30 ppm.

2 -->

3.6. The effect of weather conditions

Wind speed and wind direction have a strong influence on CO₂ transport and diffusion, which may be reflected in spatial CO_2ff variations. The average wind speed during April to September in Xi’an City (Xi’an Jinghe National Meteorological Station, No. 57131) in 2013 and 2014 was 2.66 m s^?1 and 2.54 m s^?1, respectively, with typically low winds of about the same magnitude. Figure 5 shows wind rose diagrams for April to September in 2013 and 2014. The wind differences show an increase in southwesterly winds, and a decrease in north-northeasterly winds in 2014, as compared to 2013. There are some industrial parks in the southwest Xi’an urban area, including the large Xi’an Hi-Tech Industries Development Zone. Emissions from this zone could be carried to urban areas by southwesterly winds (Guo et al., 2014). CO_2ff values of most sites near the west and south Bypass Highway did increase from 2013 to 2014, possibly due to more frequent southwesterly winds.
Figure5. Wind rose diagrams during April to September in 2013 and 2014. Wind speeds are shown in different colors

The spatial distribution of CO_2ff indicated by plant Δ¹⁴C measurements in the central Xi'an urban district in the April to September of 2013 and 2014 is presented in this paper. Three characteristics in our observed period based on the above results and discussion are as follows. First, the year-to-year variation of CO_2ff from different sites was significant. Second, there was no obvious difference in average CO_2ff between the three districts divided by the Ring Road, 2nd Ring Road and the Bypass Highway. Third, the northeast corner of the study area featured distinctly lower CO_2ff values than the rest of the sites studied. These spatial and temporal variations were influenced by emission sources and weather conditions. Vehicle exhaust emissions and residential emissions appeared to control the spatial differences. Weather conditions (wind direction in particular) also influenced regional temporal CO_2ff variations.
Our results show high spatial and temporal CO_2ff variations in Xi’an. Implicit in these results is the need for high spatial resolution, and long-term CO_2ff monitoring. To characterize CO_2ff spatial distributions in urban areas, plant samples are economical and convenient. To obtain reliable results, there are some recommendations when adopting this method: (a) sampling sites should be chosen carefully, and over a large area, with sites distant from CO₂ emission sources being suitable; (b) the plants should grow in an open environment; and (c) widespread plants with a known growth period are preferable. Such a program would support timely carbon emission reduction measures and efficient policies to cope with future change.
Acknowledgments. The authors would like to thank the anonymous reviewers and Dr. George S. BURR for their helpful comments. This work was jointly supported by the National Natural Science Foundation of China (Grant No. NSFC41730108, 41773141, 41573136, and 41991250), National Research Program for Key Issues in Air Pollution Control (Grant No. DQGG0105-02), the Strategic Priority Research Program of the Chinese Academy of Sciences (Grant No. XDA23010302), the Youth Innovation Promotion Association CAS (Grant No.2016360) and the Natural Science Basic Research Program of Shaanxi (Program No. 2019JCW-20).