10Be/7Be implies the contribution of stratosphere-troposphere transport to the winter-spring surface O3 variation observed on the Tibetan Plateau

Abstract10Be/7Be is a stratospheric sensitive tracer. In this paper, measurements of 10Be/7Be and surface O3 from October 2005 to May 2006 at Mt. Waliguan (hereafter WLG, 100.898°E, 39.287°N, 3810 m, a.s.l.), China global atmospheric watch (GAW) observatory, are introduced and used to investigate the stratosphere-troposphere transport (STT) and its impact on surface O3 on the Tibetan Plateau. The results show that the magnitude of STT is weak in winter, followed by strengthening from the end of winter to the middle of spring (from mid February to mid April) with large increases in 10Be,7Be,10Be/7Be and surface O3. At the end of spring (from the end of April to mid May in this paper), the STT weakened, and the continuous increase of surface O3 at WLG is produced by tropospheric photochemistry reactions.

Be and 10 Be are produced mainly from collisions of cosmic rays with oxygen and nitrogen atoms in the upper atmosphere, including the upper troposphere and stratosphere. Both 7 Be and 10 Be are absorbed easily by sub-micrometersize aerosol particles once produced. 7 Be as a natural tracer of downward transport from the upper atmosphere had been gradually established since the 1960s, and the first application of 10 Be/ 7 Be as a stratospheric tracer was addressed by Rasebeck et al. [1] in 1981. The half -lives of 7 Be and 10 Be are 53.3 d and 1.5×10 6 a, respectively, while the average age of the stratospheric air mass is ~1.5 a. In the stratosphere, although the initial production rate of 10 Be is lower than that of 7 Be, 10 Be/ 7 Be increases as the 7 Be isotope naturally decays *Corresponding author (email: zhengxd@cams.cma.gov.cn) to a lower concentration until reaching its equilibrium concentration, resulting in higher 10 Be/ 7 Be. In the troposphere, the ratio of 10 Be/ 7 Be is little influenced by wet deposition scavenging of aerosol particles because they are from the same source, making the ratio a more sensitive tracer of stratosphere than using 7 Be or 10 Be alone. Research on 10 Be measurements is mainly conducted in the field of paleoclimatology. However, as a natural stratospheric tracer, 10 Be/ 7 Be effectively indicates the contribution of Stratosphere-Troposphere Transport (STT) to surface-level O 3 variations as observed in the European Alps and Arctic regions [2,3]. Jordan et al. [4] reported latitudinal distribution of 10 Be/ 7 Be values in the upper troposphere and lower stratosphere over several regions through analysis of extensive aircraft aerosol samples. However, publications of at-mospheric 10 Be measurements are still limited because of the complexities in 10 Be target preparation and 10 Be detection by Accelerator Mass Spectrometry (AMS). Shen et al. [5] reported short-term 10 Be concentrations in dust deposition sampled in Beijing and Ningbo. Measurements of 10 Be/ 7 Be in atmospheric aerosol samples have not yet been reported in China.
Tropospheric O 3 is a significant trace gas due its oxidation and greenhouse characteristics. Tropospheric photochemical production and the direct or indirect STT are the two primary sources of tropospheric O 3 . The debate on which source is dominant is ongoing, especially on the extensive spring peaks in surface O 3 in the northern hemisphere [6]. With respect to atmospheric circulation, the seasonally intensified downward transport branch (presented as STT) of the stratospheric Brewer-Dobson circulation in the northern hemisphere increases the fraction of stratosphere O 3 in troposphere, contributing to the rise of surface O 3 [7]. However, atmospheric chemistry suggests that as the solar radiation seasonally increases in spring, tropospheric O 3 may be largely produced in the short-term through photochemical reactions with precursors, such as NO x , which are emitted in significant amounts in winter [8].
Mt. Waliguan (hereafter WLG) is located on the northeast margin of the Tibetan Plateau. It was selected as a site for China's Global Atmospheric Watch (GAW) observatory. Long-term observations reveal that surface O 3 at WLG increases from winter to spring, reaching its maximum in June. There are several publications presenting possible mechanisms of the summer O 3 maximum at WLG [9][10][11][12][13]. However, the mechanisms for winter and spring surface variations at WLG require elucidation. In this paper, weekly concentrations of 7 Be and 10 Be measured on aerosol samples obtained from October 2005 to May 2006 at WLG are presented, and 10 Be/ 7 Be is used as a sensitive stratospheric tracer to investigate the STT and its contribution to winter-spring variation in surface O 3 .

Measurements and data set
Measurements of 7 Be and 210 Pb analysis were described in an earlier study [14]. Every aerosol particle sample was continuously collected over one week by a pump with flow rate of ~1.3 m 3 min -1 . The collector is a three-layered polypropylene filter. The sampled filters were then sent to the State Key Laboratory of Earth Geochemistry (SKLEG), Chinese Academy of Sciences, to measure 7 Be and 210 Pb. The total relative measurement errors are from ±1.4% to ±6.4% [14]. Complying with the standard operation procedures issued by the World Meteorological Organization (WMO), surface O 3 at WLG is routinely measured by UV absorption analyzers (Thermal Environment, Model 49C) with a precision of ±2×10 −9 volume mixing ratio [15]. The long-term surface O 3 data is also periodically audited by the assigned WMO laboratory with stringent inter-comparison measurements [16].
After 7 Be and 210 Pb were analyzed, the aerosol samples were sent to the Key Laboratory of Isotope Geochronology and Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, for target preparation for AMS 10 Be analysis. The detailed procedures for making 10 Be targets have been described in other references [5,17]. Briefly, the filters were alternately pulverized and leached to extract final 10 Be residues in BeO. The 10 Be AMS target was made by mixing Nb powder with the BeO. These targets were sent to the 6 MeV EN tandem AMS system, National Key Laboratory of Nuclear Physics and Technology, Peking University for 10 Be measurements. The detection limit of 10 Be in the AMS system is 6×10 -15 . The standard, 10 Be/ 9 Be, is 2.67×10 -11 with a precision higher than 5‰, and was imported from the US National Institute of Standards and Technology [18]. The total relative error for the 10 Be determination is estimated to within ±5%, including errors caused by the efficiency of leaching filters and 10 Be measurements [17]. Because precipitation is the most significant sink for the Be and Pb isotopes, weekly precipitation and occurrence fog are also shown in Figure 1. The variations in 7 Be and 10 Be concentrations are generally consistent. 10 Be concentration was lower than that of 7 Be before February 2006, and then rose, exceeding 7 Be concentration. After mid April, both 10 Be and 7 Be concentrations began to decrease. The weekly variations in Be isotope concentration are somewhat associated with synoptic processes indicated by precipitation and fog records. Be isotope concentration increased almost continuously during two periods with little precipitation, one was from circa December 2005 to February 2006, another was from circa mid of March to mid of April in 2006. However, in the second dry period, Be concentrations increased as a result of STT ( Figure 2). Be isotope concentrations began to reduce after mid April as precipitation increased.

Results and analysis
The main source of 210 Pb is from land surface emission. Figure 1 shows that 210 Pb concentration in the spring did not increase as did Be isotopes. When maximum 210 Pb concentrations occurred in winter, concentrations of Be isotope did not rise correspondingly. These comparisons suggest that dust or soil emissions have little impact on variation in Be isotope concentration at WLG, and the cause of spring 10 Be and 7 Be enhancement is only the intensified STT. The temporal increases of 210 Pb, 10 Be and 7 Be concentration during the two non-precipitation periods are due to weak wet scavenging. To some extent, the general decreasing trend of 210 Pb concentration from winter to spring is associated with  seasonal increase in precipitation. Figure 2 presents weekly variations of surface O 3 concentration, 10 Be/ 7 Be, and 7 Be/ 210 Pb. 7 Be/ 210 Pb is considered to be a tracer unaffected by wet scavenging processes. The STT during the period from 15 February to the end of April is evidenced by the simultaneous elevations of surface O 3 and 10 Be/ 7 Be. However, for the last five samples starting from the end of April, the clear reductions of 10 Be/ 7 Be reveal a weakening of STT. At the same time, the continuous increase of surface O 3 at WLG supports dominant tropospheric photochemical O 3 production. This transfer is consistent with estimations of monthly average surface O 3 budget using the 7 Be observation alone. In that study, the contribution from STT to the surface O 3 budget in May 2003 was concluded to be less than that in April and June, and the input from tropospheric chemical production was dominant for the surface O 3 budget [19].
Using high values of 7 Be/ 210 Pb, studies have plausibly interpreted that summer surface O 3 at WLG is influenced by downward transport from the upper atmosphere [12,13]. However, the last four samples in Figure 2 demonstrate the different physical processes implied by 7 Be/ 210 Pb, and 10 Be/ 7 Be. Obviously, the implication of strengthening downward transport from the stratosphere is not reasonable based on the observed increase of 7 Be/ 210 Pb alone. Due to differences in the source distribution, 210 Pb is more sensitive to wet scavenging than Be isotopes [20], and this caused a higher 7 Be/ 210 Pb ratio at the end of spring and even throughout the summer at WLG, but a corresponding strengthening STT did not occur.
The inter-comparisons of 10 Be, and 10 Be/ 7 Be ratio between WLG and the other sites are given in Table 1. It has been noted that 7 Be concentrations tend to have the highest values in comparisons with data obtained at the sites with elevations or latitudes similar to WLG [23]. In Table 1, 10 Be also tends to have the highest concentration at WLG, but the case for 10 Be/ 7 Be is much different given the fact that WLG is located at lower latitude. Although being at high elevation,

Discussion
Some transport processes at time scales of less than one week would be smoothed in the data because of the fixed weekly sampling for 10 Be/ 7 Be determination, especially in the case of deep STT that generally persists for just 1-2 d or less. For an individual sampling week with a lasting precipitation event, the correlations between surface O 3 and 10 Be/ 7 Be before, during or after the precipitation would be also filtered out by the averaging. Therefore, in situ aerosol sampling schedule based on synoptic changes is recommended. In addition, 10 Be/ 7 Be is considered to be little influenced by wet scavenging. However, because Be isotopes reside in aerosol particles, and the fact that behavior of aerosol particle differs from atmospheric trace gases such as O 3 may also influence the interpretation of 10 Be/ 7 Be. To utilize a tracer residing in the gaseous phase should better describe the contribution of STT to surface O 3 , for example, 18 O in CO 2 [11], while 17 O in CO 2 which is entirely produced in the stratosphere deserves to be more considered [25,26].

Conclusion
Winter and spring concentrations of 10 Be and 7 Be at WLG are generally higher than those measured at other sites. However, in winter the relatively lower 10 Be/ 7 Be and surface O 3 concentration demonstrates the weakness of STT and its impact on surface O 3 . From the end of winter to the middle of spring (from mid February to mid April), continuous elevations of 10 Be, 7 Be, and surface O 3 concentration and 10 Be/ 7 Be support the seasonal intensified STT and its positive contribution to the surface O 3 budget. At the end of spring (from the end of April to mid of May in this study), the reductions of 10 Be and 7 Be concentration and 10 Be/ 7 Be suggest weakening of the STT, while the continuous increase of surface O 3 at WLG is from the positive input of tropospheric photochemical reactions.