the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 26 Apr 2022
Diagnosing the stratospheric proportion in tropospheric ozone using triple oxygen isotopes as tracers
Abstract. Using a multistep nitrite-coated filter-pack system for sampling, we determined the seasonal variations in the triple oxygen isotopic composition (Δ17O) of tropospheric ozone (O3) in the terminal positions (Δ17Oterm(O3)) in the cities Nagoya and Niigata (Japan) in the eastern Asia region to quantify the mixing ratio of stratospheric O3 within the total tropospheric O3 supplied by stratosphere–troposphere transport (STT). In Nagoya, diurnal variations have also been studied. Both the average Δ17Oterm(O3) and their 1σ variation ranges agreed well with previous studies, (37.5 ± 1.4) ‰ in Nagoya and (37.0 ± 1.7) ‰ in Niigata. The average difference in Δ17Oterm(O3) between daytime (higher) and nighttime (lower) was (1.4 ± 0.7) ‰ (1σ) in Nagoya, which was responsible for the formation of a stable boundary layer at night, reducing mixing with high Δ17Oterm(O3) from the free troposphere. We also found a significant correlation between 7Be activity concentrations and the Δ17Oterm(O3), implying that STT was responsible for the elevated Δ17Oterm of O3 in the troposphere. By using the relationship between the reciprocal of concentrations and Δ17Oterm of tropospheric O3, we estimated the Δ17O of stratospheric O3 supplied through the STT (Δ17OSTT), together with that produced through photochemical reactions at surface altitude (Δ17Osur). Moreover, using Δ17OSTT and Δ17Osur, we estimated the mixing ratios of stratospheric O3 (i.e., O3 produced in the stratosphere and supplied to the troposphere through STT) in each tropospheric O3 (fSTT), as well as the absolute concentrations of stratospheric O3 supplied through STT in the troposphere (CSTT(O3)). The CSTT(O3) exhibited minimum values in summer ((5.3 ± 1.0) ppb) and maximum values in late winter to spring ((15.9 ± 2.1) ppb). Although the fSTT values were higher than those estimated using the chemistry climate models from past studies, the trends of the seasonal variations were consistent with them. We concluded that Δ17O successfully provided observational constraints on the STT of O3.
- Preprint
(915 KB) - Metadata XML
-
Supplement
(355 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on acp-2021-1099', Joel Savarino, 10 Jun 2022
This paper follows a first publication in RCMS in 2021 that described a new methodology for ozone isotope analysis. In this new manuscript, the authors implement this technique and document the isotopic composition of ozone at two urban sites in Japan. Documenting the isotopic composition of ozone in time and space is fundamental because it is the source of the propagation during oxidation reactions of 17O-excess in the atmosphere. Constraining the isotopic composition of ozone allows to establish the oxidation mechanisms and to model them.
There is no doubt that the new data are of very high quality and deserve to be published. However, I have some fundamental criticisms on the interpretation of the data which unfortunately prevents the publication of the article as it is.
At no point in the article is the contradiction brought about by an interpretation involving ozone mixing from two regions of the atmosphere with radically different P and T conditions discussed. Indeed, it is well established that ozone formation and destruction are in dynamic equilibrium and that its isotopic composition is controlled by P and T conditions. Although the lifetime of ozone is a few days to a few weeks, at the molecular scale, ozone is destroyed and then reformed at the half-hour scale (i.e. 1/J(O1D)). Moreover, the isotopic exchanges resulting from the equilibrium O + O2 <--> O3* are themselves 50 times faster than the stabilization of O3* into O3. Therefore, ozone is considered to be permanently in isotopic equilibrium with the environment in which it is found (Vicars 2014, Yeung 2012). As a result, stratospheric ozone entering the troposphere almost instantaneously acquires an isotopic composition corresponding to the P and T conditions of the troposphere. The interpretation of the Δ17O peak observed in spring is therefore in contradiction with what we know about ozone formation and it is never discussed.
On the other hand, since the sampling is not continuous but at the frequency of about one weekly collection per month, it is difficult to measure the natural variability of the isotopic composition of ozone, especially in spring when the stratospheric contributions are considered as maximum, and thus to see the dynamics of this spring peak and its statistical significance between years.
The interpretation of the diurnal variability based again on air-mass mixing could have been easily tested. I have no doubt that there is a possibility in Japan to install their very simple collection device at an altitude above the boundary layer and to test the hypothesis that the free troposphere has a higher Δ17O. This hypothesis is currently based on modeling work (Lyons, 2001) that has otherwise never been confronted with observation. Here the authors have missed an occasion to improve our understanding of the ozone isotopic composition and check the hypothesis they are using. Could for instance the difference between day and night been the result of the temperature difference alone? All chemical sinks of O3 is currently considered as mass-dependent but is it true? For instance we know that the CO+OH sink is not.
There may be some processes other than P & T that influence the ozone Δ 17O. There is now a body of evidence (e.g. diurnal variation, seasonal maximum, hemispheric difference) that ozone isotopic composition varies well beyond what PBL P and T range allows but the isotopic equilibrium during ozone formation cannot be ignore in the discussion.
In Vicars 2014, nitrogen isotopes of the nitrate produced by ozone is used as a quality check, and to correct ozone Δ 17O variability, why in Xu et al 2021 and in this paper, nitrogen isotopes are not reported? Checking Δ 17O versus 15N may reveals some artefacts.
A revised version where ozone formation is discussed in light of its isotopic equilibrium is definitely required before the paper can be published. Without it, it gives the false impression that interpretating Δ 17O of ozone is just a question of air-mass mixing.
Joel Savarino
Citation: https://doi.org/10.5194/acp-2021-1099-RC1 -
AC1: 'Reply on RC1', Hao Xu, 29 Jul 2022
Dear Savarino,
Thank you very much for your valuable comments on our manuscript. We have responded to each of your comments and questions.
1) At no point in the article is the contradiction brought about by an interpretation involving ozone mixing from two regions of the atmosphere with radically different P and T conditions discussed. Indeed, it is well established that ozone formation and destruction are in dynamic equilibrium and that its isotopic composition is controlled by P and T conditions. Although the lifetime of ozone is a few days to a few weeks, at the molecular scale, ozone is destroyed and then reformed at the half-hour scale (i.e. 1/J(O1D)). Moreover, the isotopic exchanges resulting from the equilibrium O + O2 <--> O3* are themselves 50 times faster than the stabilization of O3* into O3. Therefore, ozone is considered to be permanently in isotopic equilibrium with the environment in which it is found (Vicars 2014, Yeung 2012). As a result, stratospheric ozone entering the troposphere almost instantaneously acquires an isotopic composition corresponding to the P and T conditions of the troposphere. The interpretation of the Δ17O peak observed in spring is therefore in contradiction with what we know about ozone formation and it is never discussed.
We will make the suggested revision and discuss it in Section 4.2. Because the Δ17Oterm(O3) values in April were 2.2–4.6 ‰ greater than the annual mean, a significant increase when considering the < ±1 ‰ uncertainty associated with the analytical technique. Moreover, the Δ17O of O3 increased at high altitudes (Lyons, 2001). Furthermore, the significant correlation between 7Be data and Δ17Oterm(O3) implies that the STT is highly responsible for the elevated Δ17Oterm(O3) in the troposphere. However, as you point out, the Δ17Oterm(O3) value of stratospheric O3 changes rapidly when it enters the troposphere. Thus, to clarify the reasons for this contradiction, we must collect samples on the different altitudes of tropospheric O3 in future studies (e.g., planetary boundary layer, free troposphere, upper troposphere).
2) On the other hand, since the sampling is not continuous but at the frequency of about one weekly collection per month, it is difficult to measure the natural variability of the isotopic composition of ozone, especially in spring when the stratospheric contributions are considered as maximum, and thus to see the dynamics of this spring peak and its statistical significance between years.
Because the archived samples were precious and the measurements of the Δ17O values of O3 were costly and time-consuming, the number of samples for stable isotopes was limited to one weekly collection per month. Despite this, 210 Δ17O values (all nitrite-coated filters) are reported in this manuscript (supplement). We repeated the analysis for each sample at least three times to attain high precision (see section 2.3). We hope our results, including these many data values, are worthy of publication. We will collect samples with higher frequency (in spring) and high-altitude resolution in future studies.
3) The interpretation of the diurnal variability based again on air-mass mixing could have been easily tested. I have no doubt that there is a possibility in Japan to install their very simple collection device at an altitude above the boundary layer and to test the hypothesis that the free troposphere has a higher Δ17O. This hypothesis is currently based on modeling work (Lyons, 2001) that has otherwise never been confronted with observation. Here the authors have missed an occasion to improve our understanding of the ozone isotopic composition and check the hypothesis they are using. Could for instance the difference between day and night been the result of the temperature difference alone? All chemical sinks of O3 is currently considered as mass-dependent but is it true? For instance we know that the CO+OH sink is not.
We appreciate the referee’s comments. We will collect O3 samples at different altitudes in future studies. Because the diurnal temperature variability is generally less than 10°C in Nagoya city. Moreover, isotopic enrichment (Δ17O) increases very slowly with increasing temperature (Krankowsky et al., 2007). Thus, we do not think the temperature is the main reason for the diurnal variation in Δ17Oterm(O3). Because oxygen isotopic fractionations associated with the most important chemical reaction processes are mass-dependent, such as the reactions of NO + O3 and NO2 + O3 (Berhanu et al., 2012; Chakraborty and Chakraborty, 2003; Savarino et al., 2008), we assumed that Δ17O was almost stable during the sink of O3.
4) There may be some processes other than P & T that influence the ozone Δ17O. There is now a body of evidence (e.g. diurnal variation, seasonal maximum, hemispheric difference) that ozone isotopic composition varies well beyond what PBL P and T range allows but the isotopic equilibrium during ozone formation cannot be ignore in the discussion.
We will make the suggested revision and discuss it in Section 4.2.
5) In Vicars 2014, nitrogen isotopes of the nitrate produced by ozone is used as a quality check, and to correct ozone Δ17O variability, why in Xu et al 2021 and in this paper, nitrogen isotopes are not reported? Checking Δ17O versus 15N may reveals some artefacts.
In the study of Vicars and Savarino (2014), the single nitrite-coated filter method was corrected using nitrogen isotopes, because of nitrate blank produced (NaNO2 reagent and 2NO2– + O2 → 2NO3– reaction) and the kinetic isotope fractionation (for δ18O and δ15N) during the collections. In this study, we used a multistep nitrite-coated filter-pack method developed by Xu et al. (2021) for the correction of nitrate blank produced (2NO2– + O2 → 2NO3– reaction) and the kinetic isotope fractionation (for δ18O). Moreover, we found that the NO3– on the filters of the control group (only NaNO2 reagent) was negligible at less than 0.08 μmol, which was less than 1% on average of that produced by the reaction of NO2– and O3. Furthermore, Xu et al. (2021) also verified the accuracy of this method through the measurement of artificial O3 with known Δ17Oterm(O3) that had been determined from the changes in Δ17O of O2. Thus, we did not measure the nitrogen isotopes in this study.
6) A revised version where ozone formation is discussed in light of its isotopic equilibrium is definitely required before the paper can be published. Without it, it gives the false impression that interpretation Δ17O of ozone is just a question of air-mass mixing.
We will make the suggested revision; the following sentences are added to the manuscript (Section 4.2).
Considering that the lifetime of O3 is important for controlling the levels of tropospheric O3, variations in the isotopic lifetimes of the O3 molecules are the potential factor to impact Δ17O values of O3. Although the chemical lifetime of the O3 molecule can be of the order of months, the isotopic lifetime will generally be much shorter (Johnston and Thiemens, 1997; Vicars and Savarino, 2014; Yeung et al., 2012). Previous studies have suggested that the isotopic lifetime of the O3 molecule is approximately 30 min in the daytime by measuring the absolute O3 photolysis frequency, j(O3P); however, in the absence of photolytic cycling, the isotopic lifetime of O3 is extended at night (Vicars and Savarino, 2014). The new O3 molecules formed through the photolytic cycling depend only on local temperature and pressure, which determine the isotope enrichments. As a result, stratospheric O3 entering the troposphere may lose the isotopic compositions of the original O3. However, the Δ17Oterm(O3) values in April were 2.2–4.6 ‰ greater than the annual mean, a significant increase when considering the < ±1 ‰ uncertainty associated with the analytical technique. Moreover, the significant correlation between 7Be data and Δ17Oterm(O3) implies that the STT is highly responsible for the elevated Δ17Oterm(O3) in the troposphere (see Section 4.3). Thus, to clarify the reasons for this contradiction, we must collect samples on the different altitudes of tropospheric O3 in future studies (e.g., planetary boundary layer, free troposphere, upper troposphere).
Reference
Berhanu, T. A., Savarino, J., Bhattacharya, S. K. and Vicars, W. C.: 17O excess transfer during the NO2 + O3 → NO3 + O2 reaction, J. Chem. Phys., 136(4), doi:10.1063/1.3666852, 2012.
Chakraborty, S. and Chakraborty, S.: Isotopic fractionation of the O3-nitric oxide reaction, Curr. Sci., 85(8), 1210–1212, 2003.
Johnston, J. C. and Thiemens, M. H.: The isotopic composition of tropospheric ozone in three environments, J. Geophys. Res. Atmos., 102(21), 25395–25404, doi:10.1029/97jd02075, 1997.
Krankowsky, D., Lämmerzahl, P., Mauersberger, K., Janssen, C., Tuzson, B. and Röckmann, T.: Stratospheric ozone isotope fractionations derived from collected samples, J. Geophys. Res. Atmos., 112(8), 1–7, doi:10.1029/2006JD007855, 2007.
Lyons, J. R.: Transfer of mass-independent fractionation in ozone to other oxygen-containing radicals in the atmosphere, Geophys. Res. Lett., 28(17), 3231–3234, doi:10.1029/2000GL012791, 2001.
Savarino, J., Bhattacharya, S. K., Morin, S., Baroni, M. and Doussin, J. F.: The NO+ O3 reaction: A triple oxygen isotope perspective on the reaction dynamics and atmospheric implications for the transfer of the ozone isotope anomaly, J. Chem. Phys., 128(19), doi:10.1063/1.2917581, 2008.
Vicars, W. C. and Savarino, J.: Quantitative constraints on the 17O-excess (Δ17O) signature of surface ozone: Ambient measurements from 50°N to 50°S using the nitrite-coated filter technique, Geochim. Cosmochim. Acta, 135, 270–287, doi:10.1016/j.gca.2014.03.023, 2014.
Xu, H., Tsunogai, U., Nakagawa, F., Li, Y., Ito, M., Sato, K. and Tanimoto, H.: Determination of the triple oxygen isotopic composition of tropospheric ozone in terminal positions using a multistep nitrite‐coated filter‐pack system, Rapid Commun. Mass Spectrom., 35(15), 1–15, doi:10.1002/rcm.9124, 2021.
Yeung, L. Y., Young, E. D. and Schauble, E. A.: Measurements of 18O18O and 17O18O in the atmosphere and the role of isotope-exchange reactions, J. Geophys. Res. Atmos., 117(17), doi:10.1029/2012JD017992, 2012.
Citation: https://doi.org/10.5194/acp-2021-1099-AC1
-
AC1: 'Reply on RC1', Hao Xu, 29 Jul 2022
-
RC2: 'Comment on acp-2021-1099', Yunting Fang, 12 Jun 2022
The manuscript submitted by Hao Xu et al. reports seasonal variations in triple oxygen isotopes of tropospheric ozone for two sites in Japan, and quantified the proportion of stratosphere-troposphere transport. The authors used a multistep nitrite-coated filter-pack system, which was newly-developed, to collect air samples, for the two selected sites in Japan. They found that the two sites had a similar seasonal pattern in triple oxygen isotopic composition of tropospheric ozone in the terminal position, with averages of 37 permil in at both sites and the highest values in April. The day and night difference had also been examined for one of the two sites, and a difference of 1.4 permil was observed. Using the relationship between concentration and the triple oxygen isotopic composition for collected samples over study period, the authors identified the triple oxygen isotopic composition was 44.3 permil for stratospheric ozone and 34.8 permil for the ozone produced in troposphere, respectively. With these values, the proportion of stratosphere-troposphere transport was quantified with the simple mixing model to be 23% to 36%. The stratosphere-troposphere transport was further evidenced by a positive relationship between triple oxygen isotopic composition of tropospheric ozone and the radionuclide Beryllium-7 activities over the two study sites. The study was well designed and the manuscript was clearly written.
I have one major concern about the partition equations (equations 7 to 9) for tropospheric ozone with oxygen isotopic composition. As stated in the section 4.4, tropospheric ozone can be considered to include three components, i.e., background ozone in the troposphere, stratospheric ozone supplied to the troposphere via stratosphere-troposphere transport (STT) and the ozone produced in situ in the troposphere through photochemical reaction. However, in the equations 7 to 9, the tropospheric ozone was considered to only include the later two components. I suggest the authors to modify these equations, by first subtracting the background ozone from the troposphere ozone with results of concentration and oxygen isotopic composition in figure 4 (orange circles), then partitioning them into components with equations 7 to 9. After that, combine the results and the proportion of background ozone which is STT orignal. By so doing, it may reduce the proportion of STT and solve the discrepancy between the observation from the present study and the CHASER model.
Specific comments:
1, To check if the difference between day and night time or between two sites is statistically significant, the authors could use repeated measurement ANOVA.
2, For lines 215-218, I suggest move them into the method sections or discussion section, because they did not state any results on Beryllium-7 activities.
Citation: https://doi.org/10.5194/acp-2021-1099-RC2 -
AC2: 'Reply on RC2', Hao Xu, 29 Jul 2022
Dear Fang,
Thank you very much for your valuable comments on our manuscript. We have responded to each of your comments and questions.
1) I have one major concern about the partition equations (equations 7 to 9) for tropospheric ozone with oxygen isotopic composition. As stated in the section 4.4, tropospheric ozone can be considered to include three components, i.e., background ozone in the troposphere, stratospheric ozone supplied to the troposphere via stratosphere-troposphere transport (STT) and the ozone produced in situ in the troposphere through photochemical reaction. However, in the equations 7 to 9, the tropospheric ozone was considered to only include the later two components. I suggest the authors to modify these equations, by first subtracting the background ozone from the troposphere ozone with results of concentration and oxygen isotopic composition in figure 4 (orange circles), then partitioning them into components with equations 7 to 9. After that, combine the results and the proportion of background ozone which is STT orignal. By so doing, it may reduce the proportion of STT and solve the discrepancy between the observation from the present study and the CHASER model.
Thank you for the comments on the background O3 in the troposphere. In this study, the background O3 in the troposphere means the tropospheric O3 with background isotope properties (not only concentrations but also isotopes). It can be explained by the concentration of O3 from the stratosphere (STT) and the troposphere (produced through photochemical reactions) becoming very low over time. Thus, we did not subtract the background ozone from the troposphere ozone with the results of concentration and oxygen isotopic composition in Figure 4 (orange circles). However, it should be noted that the conclusions presented are preliminary because of the missing contributions of the free troposphere and upper troposphere. We will collect samples on the different altitudes of tropospheric O3 in future studies.
2) To check if the difference between day and night time or between two sites is statistically significant, the authors could use repeated measurement ANOVA.
According to the referee’s suggestion, the following sentences are modified as follows (Section 4.1)
The variations in the Δ17Oterm(O3) between the daytime and nighttime during 2019 and 2020 in Nagoya (1.4 ‰ on average) exceeded the uncertainty of the Δ17Oterm(O3) measurements (±0.8 ‰ on average), implying that the diurnal variations were significant (p < 0.01, ANOVA).
3) For lines 215-218, I suggest move them into the method sections or discussion section, because they did not state any results on Beryllium-7 activities.
We moved them into Section 4.3 of the manuscript.
Citation: https://doi.org/10.5194/acp-2021-1099-AC2
-
AC2: 'Reply on RC2', Hao Xu, 29 Jul 2022
Status: closed
-
RC1: 'Comment on acp-2021-1099', Joel Savarino, 10 Jun 2022
This paper follows a first publication in RCMS in 2021 that described a new methodology for ozone isotope analysis. In this new manuscript, the authors implement this technique and document the isotopic composition of ozone at two urban sites in Japan. Documenting the isotopic composition of ozone in time and space is fundamental because it is the source of the propagation during oxidation reactions of 17O-excess in the atmosphere. Constraining the isotopic composition of ozone allows to establish the oxidation mechanisms and to model them.
There is no doubt that the new data are of very high quality and deserve to be published. However, I have some fundamental criticisms on the interpretation of the data which unfortunately prevents the publication of the article as it is.
At no point in the article is the contradiction brought about by an interpretation involving ozone mixing from two regions of the atmosphere with radically different P and T conditions discussed. Indeed, it is well established that ozone formation and destruction are in dynamic equilibrium and that its isotopic composition is controlled by P and T conditions. Although the lifetime of ozone is a few days to a few weeks, at the molecular scale, ozone is destroyed and then reformed at the half-hour scale (i.e. 1/J(O1D)). Moreover, the isotopic exchanges resulting from the equilibrium O + O2 <--> O3* are themselves 50 times faster than the stabilization of O3* into O3. Therefore, ozone is considered to be permanently in isotopic equilibrium with the environment in which it is found (Vicars 2014, Yeung 2012). As a result, stratospheric ozone entering the troposphere almost instantaneously acquires an isotopic composition corresponding to the P and T conditions of the troposphere. The interpretation of the Δ17O peak observed in spring is therefore in contradiction with what we know about ozone formation and it is never discussed.
On the other hand, since the sampling is not continuous but at the frequency of about one weekly collection per month, it is difficult to measure the natural variability of the isotopic composition of ozone, especially in spring when the stratospheric contributions are considered as maximum, and thus to see the dynamics of this spring peak and its statistical significance between years.
The interpretation of the diurnal variability based again on air-mass mixing could have been easily tested. I have no doubt that there is a possibility in Japan to install their very simple collection device at an altitude above the boundary layer and to test the hypothesis that the free troposphere has a higher Δ17O. This hypothesis is currently based on modeling work (Lyons, 2001) that has otherwise never been confronted with observation. Here the authors have missed an occasion to improve our understanding of the ozone isotopic composition and check the hypothesis they are using. Could for instance the difference between day and night been the result of the temperature difference alone? All chemical sinks of O3 is currently considered as mass-dependent but is it true? For instance we know that the CO+OH sink is not.
There may be some processes other than P & T that influence the ozone Δ 17O. There is now a body of evidence (e.g. diurnal variation, seasonal maximum, hemispheric difference) that ozone isotopic composition varies well beyond what PBL P and T range allows but the isotopic equilibrium during ozone formation cannot be ignore in the discussion.
In Vicars 2014, nitrogen isotopes of the nitrate produced by ozone is used as a quality check, and to correct ozone Δ 17O variability, why in Xu et al 2021 and in this paper, nitrogen isotopes are not reported? Checking Δ 17O versus 15N may reveals some artefacts.
A revised version where ozone formation is discussed in light of its isotopic equilibrium is definitely required before the paper can be published. Without it, it gives the false impression that interpretating Δ 17O of ozone is just a question of air-mass mixing.
Joel Savarino
Citation: https://doi.org/10.5194/acp-2021-1099-RC1 -
AC1: 'Reply on RC1', Hao Xu, 29 Jul 2022
Dear Savarino,
Thank you very much for your valuable comments on our manuscript. We have responded to each of your comments and questions.
1) At no point in the article is the contradiction brought about by an interpretation involving ozone mixing from two regions of the atmosphere with radically different P and T conditions discussed. Indeed, it is well established that ozone formation and destruction are in dynamic equilibrium and that its isotopic composition is controlled by P and T conditions. Although the lifetime of ozone is a few days to a few weeks, at the molecular scale, ozone is destroyed and then reformed at the half-hour scale (i.e. 1/J(O1D)). Moreover, the isotopic exchanges resulting from the equilibrium O + O2 <--> O3* are themselves 50 times faster than the stabilization of O3* into O3. Therefore, ozone is considered to be permanently in isotopic equilibrium with the environment in which it is found (Vicars 2014, Yeung 2012). As a result, stratospheric ozone entering the troposphere almost instantaneously acquires an isotopic composition corresponding to the P and T conditions of the troposphere. The interpretation of the Δ17O peak observed in spring is therefore in contradiction with what we know about ozone formation and it is never discussed.
We will make the suggested revision and discuss it in Section 4.2. Because the Δ17Oterm(O3) values in April were 2.2–4.6 ‰ greater than the annual mean, a significant increase when considering the < ±1 ‰ uncertainty associated with the analytical technique. Moreover, the Δ17O of O3 increased at high altitudes (Lyons, 2001). Furthermore, the significant correlation between 7Be data and Δ17Oterm(O3) implies that the STT is highly responsible for the elevated Δ17Oterm(O3) in the troposphere. However, as you point out, the Δ17Oterm(O3) value of stratospheric O3 changes rapidly when it enters the troposphere. Thus, to clarify the reasons for this contradiction, we must collect samples on the different altitudes of tropospheric O3 in future studies (e.g., planetary boundary layer, free troposphere, upper troposphere).
2) On the other hand, since the sampling is not continuous but at the frequency of about one weekly collection per month, it is difficult to measure the natural variability of the isotopic composition of ozone, especially in spring when the stratospheric contributions are considered as maximum, and thus to see the dynamics of this spring peak and its statistical significance between years.
Because the archived samples were precious and the measurements of the Δ17O values of O3 were costly and time-consuming, the number of samples for stable isotopes was limited to one weekly collection per month. Despite this, 210 Δ17O values (all nitrite-coated filters) are reported in this manuscript (supplement). We repeated the analysis for each sample at least three times to attain high precision (see section 2.3). We hope our results, including these many data values, are worthy of publication. We will collect samples with higher frequency (in spring) and high-altitude resolution in future studies.
3) The interpretation of the diurnal variability based again on air-mass mixing could have been easily tested. I have no doubt that there is a possibility in Japan to install their very simple collection device at an altitude above the boundary layer and to test the hypothesis that the free troposphere has a higher Δ17O. This hypothesis is currently based on modeling work (Lyons, 2001) that has otherwise never been confronted with observation. Here the authors have missed an occasion to improve our understanding of the ozone isotopic composition and check the hypothesis they are using. Could for instance the difference between day and night been the result of the temperature difference alone? All chemical sinks of O3 is currently considered as mass-dependent but is it true? For instance we know that the CO+OH sink is not.
We appreciate the referee’s comments. We will collect O3 samples at different altitudes in future studies. Because the diurnal temperature variability is generally less than 10°C in Nagoya city. Moreover, isotopic enrichment (Δ17O) increases very slowly with increasing temperature (Krankowsky et al., 2007). Thus, we do not think the temperature is the main reason for the diurnal variation in Δ17Oterm(O3). Because oxygen isotopic fractionations associated with the most important chemical reaction processes are mass-dependent, such as the reactions of NO + O3 and NO2 + O3 (Berhanu et al., 2012; Chakraborty and Chakraborty, 2003; Savarino et al., 2008), we assumed that Δ17O was almost stable during the sink of O3.
4) There may be some processes other than P & T that influence the ozone Δ17O. There is now a body of evidence (e.g. diurnal variation, seasonal maximum, hemispheric difference) that ozone isotopic composition varies well beyond what PBL P and T range allows but the isotopic equilibrium during ozone formation cannot be ignore in the discussion.
We will make the suggested revision and discuss it in Section 4.2.
5) In Vicars 2014, nitrogen isotopes of the nitrate produced by ozone is used as a quality check, and to correct ozone Δ17O variability, why in Xu et al 2021 and in this paper, nitrogen isotopes are not reported? Checking Δ17O versus 15N may reveals some artefacts.
In the study of Vicars and Savarino (2014), the single nitrite-coated filter method was corrected using nitrogen isotopes, because of nitrate blank produced (NaNO2 reagent and 2NO2– + O2 → 2NO3– reaction) and the kinetic isotope fractionation (for δ18O and δ15N) during the collections. In this study, we used a multistep nitrite-coated filter-pack method developed by Xu et al. (2021) for the correction of nitrate blank produced (2NO2– + O2 → 2NO3– reaction) and the kinetic isotope fractionation (for δ18O). Moreover, we found that the NO3– on the filters of the control group (only NaNO2 reagent) was negligible at less than 0.08 μmol, which was less than 1% on average of that produced by the reaction of NO2– and O3. Furthermore, Xu et al. (2021) also verified the accuracy of this method through the measurement of artificial O3 with known Δ17Oterm(O3) that had been determined from the changes in Δ17O of O2. Thus, we did not measure the nitrogen isotopes in this study.
6) A revised version where ozone formation is discussed in light of its isotopic equilibrium is definitely required before the paper can be published. Without it, it gives the false impression that interpretation Δ17O of ozone is just a question of air-mass mixing.
We will make the suggested revision; the following sentences are added to the manuscript (Section 4.2).
Considering that the lifetime of O3 is important for controlling the levels of tropospheric O3, variations in the isotopic lifetimes of the O3 molecules are the potential factor to impact Δ17O values of O3. Although the chemical lifetime of the O3 molecule can be of the order of months, the isotopic lifetime will generally be much shorter (Johnston and Thiemens, 1997; Vicars and Savarino, 2014; Yeung et al., 2012). Previous studies have suggested that the isotopic lifetime of the O3 molecule is approximately 30 min in the daytime by measuring the absolute O3 photolysis frequency, j(O3P); however, in the absence of photolytic cycling, the isotopic lifetime of O3 is extended at night (Vicars and Savarino, 2014). The new O3 molecules formed through the photolytic cycling depend only on local temperature and pressure, which determine the isotope enrichments. As a result, stratospheric O3 entering the troposphere may lose the isotopic compositions of the original O3. However, the Δ17Oterm(O3) values in April were 2.2–4.6 ‰ greater than the annual mean, a significant increase when considering the < ±1 ‰ uncertainty associated with the analytical technique. Moreover, the significant correlation between 7Be data and Δ17Oterm(O3) implies that the STT is highly responsible for the elevated Δ17Oterm(O3) in the troposphere (see Section 4.3). Thus, to clarify the reasons for this contradiction, we must collect samples on the different altitudes of tropospheric O3 in future studies (e.g., planetary boundary layer, free troposphere, upper troposphere).
Reference
Berhanu, T. A., Savarino, J., Bhattacharya, S. K. and Vicars, W. C.: 17O excess transfer during the NO2 + O3 → NO3 + O2 reaction, J. Chem. Phys., 136(4), doi:10.1063/1.3666852, 2012.
Chakraborty, S. and Chakraborty, S.: Isotopic fractionation of the O3-nitric oxide reaction, Curr. Sci., 85(8), 1210–1212, 2003.
Johnston, J. C. and Thiemens, M. H.: The isotopic composition of tropospheric ozone in three environments, J. Geophys. Res. Atmos., 102(21), 25395–25404, doi:10.1029/97jd02075, 1997.
Krankowsky, D., Lämmerzahl, P., Mauersberger, K., Janssen, C., Tuzson, B. and Röckmann, T.: Stratospheric ozone isotope fractionations derived from collected samples, J. Geophys. Res. Atmos., 112(8), 1–7, doi:10.1029/2006JD007855, 2007.
Lyons, J. R.: Transfer of mass-independent fractionation in ozone to other oxygen-containing radicals in the atmosphere, Geophys. Res. Lett., 28(17), 3231–3234, doi:10.1029/2000GL012791, 2001.
Savarino, J., Bhattacharya, S. K., Morin, S., Baroni, M. and Doussin, J. F.: The NO+ O3 reaction: A triple oxygen isotope perspective on the reaction dynamics and atmospheric implications for the transfer of the ozone isotope anomaly, J. Chem. Phys., 128(19), doi:10.1063/1.2917581, 2008.
Vicars, W. C. and Savarino, J.: Quantitative constraints on the 17O-excess (Δ17O) signature of surface ozone: Ambient measurements from 50°N to 50°S using the nitrite-coated filter technique, Geochim. Cosmochim. Acta, 135, 270–287, doi:10.1016/j.gca.2014.03.023, 2014.
Xu, H., Tsunogai, U., Nakagawa, F., Li, Y., Ito, M., Sato, K. and Tanimoto, H.: Determination of the triple oxygen isotopic composition of tropospheric ozone in terminal positions using a multistep nitrite‐coated filter‐pack system, Rapid Commun. Mass Spectrom., 35(15), 1–15, doi:10.1002/rcm.9124, 2021.
Yeung, L. Y., Young, E. D. and Schauble, E. A.: Measurements of 18O18O and 17O18O in the atmosphere and the role of isotope-exchange reactions, J. Geophys. Res. Atmos., 117(17), doi:10.1029/2012JD017992, 2012.
Citation: https://doi.org/10.5194/acp-2021-1099-AC1
-
AC1: 'Reply on RC1', Hao Xu, 29 Jul 2022
-
RC2: 'Comment on acp-2021-1099', Yunting Fang, 12 Jun 2022
The manuscript submitted by Hao Xu et al. reports seasonal variations in triple oxygen isotopes of tropospheric ozone for two sites in Japan, and quantified the proportion of stratosphere-troposphere transport. The authors used a multistep nitrite-coated filter-pack system, which was newly-developed, to collect air samples, for the two selected sites in Japan. They found that the two sites had a similar seasonal pattern in triple oxygen isotopic composition of tropospheric ozone in the terminal position, with averages of 37 permil in at both sites and the highest values in April. The day and night difference had also been examined for one of the two sites, and a difference of 1.4 permil was observed. Using the relationship between concentration and the triple oxygen isotopic composition for collected samples over study period, the authors identified the triple oxygen isotopic composition was 44.3 permil for stratospheric ozone and 34.8 permil for the ozone produced in troposphere, respectively. With these values, the proportion of stratosphere-troposphere transport was quantified with the simple mixing model to be 23% to 36%. The stratosphere-troposphere transport was further evidenced by a positive relationship between triple oxygen isotopic composition of tropospheric ozone and the radionuclide Beryllium-7 activities over the two study sites. The study was well designed and the manuscript was clearly written.
I have one major concern about the partition equations (equations 7 to 9) for tropospheric ozone with oxygen isotopic composition. As stated in the section 4.4, tropospheric ozone can be considered to include three components, i.e., background ozone in the troposphere, stratospheric ozone supplied to the troposphere via stratosphere-troposphere transport (STT) and the ozone produced in situ in the troposphere through photochemical reaction. However, in the equations 7 to 9, the tropospheric ozone was considered to only include the later two components. I suggest the authors to modify these equations, by first subtracting the background ozone from the troposphere ozone with results of concentration and oxygen isotopic composition in figure 4 (orange circles), then partitioning them into components with equations 7 to 9. After that, combine the results and the proportion of background ozone which is STT orignal. By so doing, it may reduce the proportion of STT and solve the discrepancy between the observation from the present study and the CHASER model.
Specific comments:
1, To check if the difference between day and night time or between two sites is statistically significant, the authors could use repeated measurement ANOVA.
2, For lines 215-218, I suggest move them into the method sections or discussion section, because they did not state any results on Beryllium-7 activities.
Citation: https://doi.org/10.5194/acp-2021-1099-RC2 -
AC2: 'Reply on RC2', Hao Xu, 29 Jul 2022
Dear Fang,
Thank you very much for your valuable comments on our manuscript. We have responded to each of your comments and questions.
1) I have one major concern about the partition equations (equations 7 to 9) for tropospheric ozone with oxygen isotopic composition. As stated in the section 4.4, tropospheric ozone can be considered to include three components, i.e., background ozone in the troposphere, stratospheric ozone supplied to the troposphere via stratosphere-troposphere transport (STT) and the ozone produced in situ in the troposphere through photochemical reaction. However, in the equations 7 to 9, the tropospheric ozone was considered to only include the later two components. I suggest the authors to modify these equations, by first subtracting the background ozone from the troposphere ozone with results of concentration and oxygen isotopic composition in figure 4 (orange circles), then partitioning them into components with equations 7 to 9. After that, combine the results and the proportion of background ozone which is STT orignal. By so doing, it may reduce the proportion of STT and solve the discrepancy between the observation from the present study and the CHASER model.
Thank you for the comments on the background O3 in the troposphere. In this study, the background O3 in the troposphere means the tropospheric O3 with background isotope properties (not only concentrations but also isotopes). It can be explained by the concentration of O3 from the stratosphere (STT) and the troposphere (produced through photochemical reactions) becoming very low over time. Thus, we did not subtract the background ozone from the troposphere ozone with the results of concentration and oxygen isotopic composition in Figure 4 (orange circles). However, it should be noted that the conclusions presented are preliminary because of the missing contributions of the free troposphere and upper troposphere. We will collect samples on the different altitudes of tropospheric O3 in future studies.
2) To check if the difference between day and night time or between two sites is statistically significant, the authors could use repeated measurement ANOVA.
According to the referee’s suggestion, the following sentences are modified as follows (Section 4.1)
The variations in the Δ17Oterm(O3) between the daytime and nighttime during 2019 and 2020 in Nagoya (1.4 ‰ on average) exceeded the uncertainty of the Δ17Oterm(O3) measurements (±0.8 ‰ on average), implying that the diurnal variations were significant (p < 0.01, ANOVA).
3) For lines 215-218, I suggest move them into the method sections or discussion section, because they did not state any results on Beryllium-7 activities.
We moved them into Section 4.3 of the manuscript.
Citation: https://doi.org/10.5194/acp-2021-1099-AC2
-
AC2: 'Reply on RC2', Hao Xu, 29 Jul 2022
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 1,172 | 239 | 32 | 1,443 | 50 | 20 | 26 |
- HTML: 1,172
- PDF: 239
- XML: 32
- Total: 1,443
- Supplement: 50
- BibTeX: 20
- EndNote: 26
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1