the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 03 Jul 2024
Disentangling the chemistry and transport impacts of the Quasi-Biennial Oscillation on stratospheric ozone
Abstract. The quasi-biennial oscillation (QBO) in tropical winds perturbs stratospheric ozone throughout much of the atmosphere via changes in transport of ozone and other trace gases and via temperature changes that alter chemical processes. Here we separate the temperature-driven changes using the Department of Energy’s Energy Exascale Earth System Model version 2 (E3SMv2) with linearized stratospheric ozone chemistry. E3SM produces a natural QBO cycle in winds, temperature, and ozone. Our analysis defines climatological QBO patterns of ozone for the period 1979–2020 using both nonlinear principal component analysis and monthly composites centered on QBO phase shift. As a climate model, E3SM cannot predict the timing of the phase shift, but it does match these climatological patterns. We develop an offline version of our stratospheric chemistry module to calculate the steady-state response of ozone to temperature and overhead ozone perturbations, assuming that other chemical families involved in ozone chemistry remain fixed. We find a clear demarcation: ozone perturbations in the upper stratosphere (above 20-hPa) are predicted by the steady-state response of the ozone column to the temperature changes; while those in the lower stratosphere show no temperature response and are presumably driven by circulation changes. These results are important for diagnosing model-model differences in the QBO-ozone responses for climate projections.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(17949 KB)
-
Supplement
(8244 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(17949 KB) - Metadata XML
-
Supplement
(8244 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-1927', Anonymous Referee #1, 29 Jul 2024
Review of “Disentangling the chemistry and transport impacts of the Quasi-Biennial Oscillation on stratospheric ozone” by Xie et al.
The authors derive a new QBO index. They then use observations and tailored versions of two global models to investigate the relative impacts of photochemistry and transport on the response of stratospheric ozone to the QBO.
Major comments:
- While it is certainly important to understand the processes that contribute to the response of ozone to the QBO, the question addressed in this paper is quite limited and the answer has been known for decades. The authors’ goal is to determine the relative roles of chemistry and transport. However, they do not address what chemistry (i.e., which reactions; what leads them to change with the QBO) and what transport (meridional or vertical advection, diffusion due to wave breaking) are responsible. Instead, their assumptions seem to be that all photochemical effects are caused by temperature-dependent rates of ozone loss and all transport is due to the vertical component of the transformed Eulerian mean circulation. Beyond this, the important cross-over possibilities (i.e., transport of one reactant affects the photochemistry of another that is differently affected by transport) seem not to have been considered.
- Using simplified, constrained, or nudged models can often be a valuable tool for probing causative mechanisms. However, this paper does not present a clear motivation for why these model constraints were used in the present case. Overall, the paper is lacking in motivation for the specific modeling experiments that were set up. What can we learn from these simulations that we would not learn from either a middle atmosphere chemical transport model or a GCM with interactive chemistry? Moreover, additional evidence is needed to show that processes omitted due to their assumptions, such as neglect of full Ox chemistry, are not important.
- As a related comment, the authors do not take advantage of the simplified and controlled nature of their simulations to perform additional testing of mechanisms. For example, the last paragraph of Section 5.3 speculates about which differences in the setup of the different models might account for differences in the results. Why not actually test these?
- The derivation of a new QBO index is interesting. I especially appreciate Figure 1a, which displays the temporal development nicely. I would have been interested to see how well this index is able to capture the QBO interruptions of 2016 and 2020. My concern about this index is that its description needs work. It appears in the figure captions as “0 is centered on the month when QBO transits from QBOe to QBOw (determined by when current QBO index<0 and next QBO index>0). The QBO phase is determined by 5S-5N average of the zonal wind.” It's confusing because of the mention of the month when the QBO transits as if there was a single transit month rather than a different month at each pressure. Please take care to use language that distinguishes between the QBO winds and the QBO index. Perhaps something like this would be clearer: “The x-axis is the QBO index, defined by Eq. (1).” Also, please label the axis, e.g. “QBO index”.
Additional comments
- Supplemental figures 2 and 4 are not mentioned in the text. Are they necessary or helpful for understanding the results?
- (l. 18-19) It is not clear what “phase shift” refers to nor why E3SM cannot predict it.
- (l. 33-34) “assigning the pattern of ozone perturbations over the QBO cycle to specific processes is not easy”: This gets at the gist of the issue raised by my first major concern. One reason that it is not easy is that no simple either/or assignment can capture the complexity.
- (l. 125-126) “The parameters for the convective GW momentum transport were tuned especially for this version”: Can you provide more details, such as a link where the tuning choices and parameters can be found? Some other users of this model might be interested.
- (l. 407-408) “this demonstrates the overall effectiveness of the nudging strategy” I would say “overall effectiveness” is too strong since the extratropical response is not well simulated.
- (l. 419-420) “The results shown here are important for diagnosing model-model and model-observation differences” Can you provide an example?
- (l. 447-448) “using interactive ozone or not in the simulation does not significantly alter the results for QBO simulations” This has not been shown. Also, I suggest you take care with the term “significantly”, which is best reserved for quantitative measures of statistical significance.
Citation: https://doi.org/10.5194/egusphere-2024-1927-RC1 -
AC1: 'Reply on RC1', Jinbo Xie, 29 Jan 2025
Item-by-item response to all review comments
NOTE: To facilitate the evaluation of our responses, original review comments are listed first in their originals, followed by our itemized response.
We thank the reviewers’ comments, which are in text below. Our response is followed.
------------------------------------------------------------------------------
We thank the reviewers’ comments on helping to improve this manuscript. We have made substantial modifications to address the reviewer’s concerns as follows:
- The motivation of this paper is to demonstrate the novel use of the steady-state ozone metric on nudged climate model simulations to separate the chemical and transport impact of QBO on ozone. This is applicable to other models and would help to further diagnose model-to-model difference in future projection of QBO-ozone response, thereby contributing to improved ozone and climate projections. To make this clear, we made substantial modification to the introduction, and added additional sub-sections in section 6 and in discussions. Please refer to section 5.4 and section 6.1.
- To address the reviewers’ concerns on potential cross-over chemistry and uncertainty in the climate models, we’ve made two additional simulations, one using nudged E3SMv2 with Linoz-v3 (adding NOy-N2O-CH4-H2O chemistry), another using nudged E3SMv2 with fixed-ozone. It is shown that the NOy contributes to QBO-ozone between 6-hPa to 20-hPa. Please refer to section 5.4.
- Another new understanding from the comparison of the interactive/fixed ozone nudged E3SMv2 simulations show that the interactive ozone feedback damps the QBO-temperature, this is detailed in the discussion section. Please refer to section 6.1.
------------------------------------------------------------------------------
Reviewer 1
The authors derive a new QBO index. They then use observations and tailored versions of two global models to investigate the relative impacts of photochemistry and transport on the response of stratospheric ozone to the QBO.
Major comments:
1. While it is certainly important to understand the processes that contribute to the response of ozone to the QBO, the question addressed in this paper is quite limited and the answer has been known for decades. The authors’ goal is to determine the relative roles of chemistry and transport. However, they do not address what chemistry (i.e., which reactions; what leads them to change with the QBO) and what transport (meridional or vertical advection, diffusion due to wave breaking) are responsible. Instead, their assumptions seem to be that all photochemical effects are caused by temperature-dependent rates of ozone loss and all transport is due to the vertical component of the transformed Eulerian mean circulation. Beyond this, the important cross-over possibilities (i.e., transport of one reactant affects the photochemistry of another that is differently affected by transport) seem not to have been considered.
Response: Thank you for the comments. To further separate the role of chemistry/transport and examine impact of potential cross-over chemistry species, we performed additional simulations and analysis.
a. Transport: The Brewer-Dobson circulation (BDC) characterized by the residual mean circulation ( , ) is the major source of transport for ozone from the tropics to high-latitudes. In this study, we mainly focused on the anomalous as it has good correspondence with the ozone change below 20-hPa. Additional analysis on the anomalous (Figure S6) shows that the meridional transport of BDC is stronger above 20-hPa compared to below 20-hPa, which may have minor contribution to the ozone transport below 20-hPa. Please refer to line 503-510, Figures 4d-f, and Figure S6 for detail.
b. Chemistry species: The additional simulation is produced for 1979-2020 using the nudged E3SMv2 with the “Linoz-v3 configuration” (Hsu and Prather, 2010). This “Linoz-v3 configuration” further adds stratospheric N2O, NOy, H2O, and CH4 chemistry to the current Linoz-v2 setup in E3SMv2. Our simulations showed that E3SMv2 with “Linoz-v3 configuration” better captures the QBO-ozone change between 6-hPa to 20-hPa than those with Linoz-v2. This indicates the impact of added chemistry including NOy in this pressure range. This is added in the text and figure, please refer to section 5.3 and Figures 8 and 9 for detail.
Hsu, J., and M. J. Prather (2010), Global long-lived chemical modes excited in a 3-D chemistry transport model: Stratospheric N2O, NOy, O3 and CH4 chemistry, Geophys. Res. Lett., 37, L07805, doi:10.1029/2009GL042243.
2. Using simplified, constrained, or nudged models can often be a valuable tool for probing causative mechanisms. However, this paper does not present a clear motivation for why these model constraints were used in the present case. Overall, the paper is lacking in motivation for the specific modeling experiments that were set up. What can we learn from these simulations that we would not learn from either a middle atmosphere chemical transport model or a GCM with interactive chemistry? Moreover, additional evidence is needed to show that processes omitted due to their assumptions, such as neglect of full Ox chemistry, are not important.
Response: We thank the reviewer’s comment. Despite large studies of QBO-ozone, the current challenges remain in terms of the model’s ability to simulate QBO variability (Richter et al., 2020) and its phase asymmetry (Scaife et al., 2014). To address this issue, we produce a set of nudged simulations following the protocol of QBOi. Unlike other nudging experiments, the current setup is intended to nudge only the tropical QBO dynamic variability while allowing other variabilities (e.g. semi-annual oscillation) outside of the tropical lower-stratosphere and other variables (e.g. temperature) freely evolve, thereby to ensure better assessments of its impact on global climate under a more realistic QBO without too much constraint on other processes.
To further examine the potential impact of other species on QBO-ozone, we conducted an additional nudged simulation of E3SMv2 with “Linoz-v3 configuration” (adding the N2O-NOy-H2O-CH4 chemistry). It is shown that NOy chemistry does contribute to a non-negligible impact the QBO-ozone relationship at between 6-hPa to 20-hPa, we thus refined our conclusions on NOy, we thus refined our conclusions based on the new results. Please refer to section 5.1 and Figures 8 and 9 for further detail.
3. As a related comment, the authors do not take advantage of the simplified and controlled nature of their simulations to perform additional testing of mechanisms. For example, the last paragraph of Section 5.3 speculates about which differences in the setup of the different models might account for differences in the results. Why not actually test these?
Response: To address the mentioned issue, an additional E3SMv2 nudged experiment with fixed ozone (E3SMv2 nudged fixed-ozone, using ozone input from CESM2) is added to further separate the impact of different setups. It is shown that the results still exhibit QBO in-phase changes, indicating a stronger constraint of nudging. The temperature phase change pattern in E3SMv2 nudged fixed-ozone (Figure S7) shows larger amplitude than that of the E3SMv2 nudged (Figures 4 and 5), indicating the damping effect of interactive ozone on temperature. Please refer to discussion section 6.1, and Figures 4, 5, and S7 for further detail.
4. The derivation of a new QBO index is interesting. I especially appreciate Figure 1a, which displays the temporal development nicely. I would have been interested to see how well this index is able to capture the QBO interruptions of 2016 and 2020. My concern about this index is that its description needs work. It appears in the figure captions as “0 is centered on the month when QBO transits from QBOe to QBOw (determined by when current QBO index<0 and next QBO index>0). The QBO phase is determined by 5S-5N average of the zonal wind.” It's confusing because of the mention of the month when the QBO transits as if there was a single transit month rather than a different month at each pressure. Please take care to use language that distinguishes between the QBO winds and the QBO index. Perhaps something like this would be clearer: “The x-axis is the QBO index, defined by Eq. (1).” Also, please label the axis, e.g. “QBO index”.
Response: We thank the reviewer for the comment. To avoid the confusion in description, we changed the figure caption to “0 is centered on the month when QBO index shifts from QBOe to QBOw (determined by when current QBO index<0 and next QBO index>0).” Thereby to ensure a single month centered on in the figure. Since the month -14 to 14 are offsets of month from the center 0 month, we’d like to request to change the axis label to “offset (month)” as is done now to make this description clearer. Please refer to the Figures 3 to 12 for detail.
Additional comments
1. Supplemental figures 2 and 4 are not mentioned in the text. Are they necessary or helpful for understanding the results?
Response: We thank the reviewer for the comment. Supplementary Figure 2 is used to depict the difference between NLPCA and PCA QBO index. The description is added in the text, please refer to line 308-311. Supplementary Figure 4 is removed. The supplementary Figures are also tagged Figure S… for more concise labelling.
2. (l. 18-19) It is not clear what “phase shift” refers to nor why E3SM cannot predict it.
Response: We have changed to “timing of the phase transitions” since it refers to the a transition point in time. E3SM has a shorter period and does not capture the right timing for these QBO phase transitions. Please refer to line 18-19.
3. (l. 33-34) “assigning the pattern of ozone perturbations over the QBO cycle to specific processes is not easy”: This gets at the gist of the issue raised by my first major concern. One reason that it is not easy is that no simple either/or assignment can capture the complexity.
Response: Agreed. We added further Linoz-v3 tests so as to further analyze the impact of N2O-NOy-CH4-H2O chemistry on ozone simulation. It is shown that NOy plays a large role around 10 to 20-hPa as the reviewers have mentioned.
4. (l. 125-126) “The parameters for the convective GW momentum transport were tuned especially for this version”: Can you provide more details, such as a link where the tuning choices and parameters can be found? Some other users of this model might be interested.
Response: The reference for the tuning of GW momentum transport were added to the text. Please refer to lines 125-126.
5. (l. 407-408) “this demonstrates the overall effectiveness of the nudging strategy” I would say “overall effectiveness” is too strong since the extratropical response is not well simulated.
Response: We thank the reviewer’s comment. This statement is removed from the text.
6. (l. 419-420) “The results shown here are important for diagnosing model-model and model-observation differences” Can you provide an example?
Response: Thank you for the comment. The main goal of the current study is to disentangle the chemical and transport impact of QBO on ozone using a novel steady-state ozone metric. Section 5.4 documents the application of this new metric on the nudged E3SMv2 simulations, where we were able to separate the impact of QBO on ozone from temperature, NOy variation, and circulation change. This may also be applied to other models to trace their uncertainty from these impacts, thereby to diagnose their difference among different models and observations.
7. (l. 447-448) “using interactive ozone or not in the simulation does not significantly alter the results for QBO simulations” This has not been shown. Also, I suggest you take care with the term “significantly”, which is best reserved for quantitative measures of statistical significance.
Response: Thank you for pointing this out. We have removed the statement.
Citation: https://doi.org/10.5194/egusphere-2024-1927-AC1
-
RC2: 'Comment on egusphere-2024-1927', Anonymous Referee #2, 07 Aug 2024
General comments
The paper explores the chemistry and transport impact of the QBO on stratospheric ozone. They make use of climate model simulations as well as an offline ozone calculation to investigate the response of ozone to temperature and overhead ozone changes. They claim that ozone perturbations above 20hPa are predicted by the ozone column response to temperature whilst those below show no temperature response and are presumed to be driven by the circulation changes.
1) The motivation for the choice of methodology is not made clear in the paper. It is not apparent what is being gained by it. Far too many other uncertainties are being introduced by the methodology that lead to questions about the robustness of the results. The work makes use of two different climate models, EAMv2 and CESM2. These two models do not simply differ in one having a QBO, they will have different ozone climatologies. This makes it difficult to compare the experiment as interactive and non-interactive runs as the paper is trying to do. Furthermore, there are inconsistencies as to how ensembles are being used (different numbers for historical and future).
2) The new QBO index is being calculated using a complicated method. It would be good to see more validation of this being performed. How related is this new index to the underlying dynamical variability? On line 173, “asymmetric” presumably refers to two phases rather than in latitude.
3) The Linoz calculation is problematic. The major assumption that NOy is omitted from the chemical families whilst it has been shown by Chipperfield et al. (1994) and subsequent work such as by Tian et al. (2006) that NOy variations are the primary driver of ozone QBO changes above 20hPa. This major assumption will affect the conclusion of ozone behaviour above 20 hPa.
4) The authors speculate that the changes below 20 hPa are driven by dynamics. Using their tools, it would be possible to quantify this.
Specific comments
1) Line 271 and Figure 3a. I am having difficulty seeing a monopole to tripole pattern. It looks like a tripole in each phase?
2) Much more care needs to be taken with the figures. The colorbars especially do not have sensible divisions. Some figures are too small to be easily readable.
References:
Chipperfield, M. P., L. J. Gray, J. S. Kinnersley, and J. Zawodny (1994), A two-dimensional model study of QBO signal in SAGE II NO2 and O3, Geophys. Res. Lett., 21, 589 – 592.
Tian, W., M. P. Chipperfield, L. J. Gray, and J. M. Zawodny (2006), Quasi-biennial oscillation and tracer distributions in a coupled chemistry-climate model, J. Geophys. Res., 111, D20301, doi:10.1029/2005JD006871.
Citation: https://doi.org/10.5194/egusphere-2024-1927-RC2 -
AC2: 'Reply on RC2', Jinbo Xie, 29 Jan 2025
Item-by-item response to all review comments
NOTE: To facilitate the evaluation of our responses, original review comments are listed first in their originals, followed by our itemized response.
We thank the reviewers’ comments, which are in text below. Our response is followed.
------------------------------------------------------------------------------
We thank the reviewers’ comments on helping to improve this manuscript. We have made substantial modifications to address the reviewer’s concerns as follows:
- The motivation of this paper is to demonstrate the novel use of the steady-state ozone metric on nudged climate model simulations to separate the chemical and transport impact of QBO on ozone. This is applicable to other models and would help to further diagnose model-to-model difference in future projection of QBO-ozone response, thereby contributing to improved ozone and climate projections. To make this clear, we made substantial modification to the introduction, and added additional sub-sections in section 6 and in discussions. Please refer to section 5.4 and section 6.1.
- To address the reviewers’ concerns on potential cross-over chemistry and uncertainty in the climate models, we’ve made two additional simulations, one using nudged E3SMv2 with Linoz-v3 (adding NOy-N2O-CH4-H2O chemistry), another using nudged E3SMv2 with fixed-ozone. It is shown that the NOy contributes to QBO-ozone between 6-hPa to 20-hPa. Please refer to section 5.4.
- Another new understanding from the comparison of the interactive/fixed ozone nudged E3SMv2 simulations show that the interactive ozone feedback damps the QBO-temperature, this is detailed in the discussion section. Please refer to section 6.1.
------------------------------------------------------------------------------
Reviewer 2
General comments
The paper explores the chemistry and transport impact of the QBO on stratospheric ozone. They make use of climate model simulations as well as an offline ozone calculation to investigate the response of ozone to temperature and overhead ozone changes. They claim that ozone perturbations above 20hPa are predicted by the ozone column response to temperature whilst those below show no temperature response and are presumed to be driven by the circulation changes.
- The motivation for the choice of methodology is not made clear in the paper. It is not apparent what is being gained by it. Far too many other uncertainties are being introduced by the methodology that lead to questions about the robustness of the results. The work makes use of two different climate models, EAMv2 and CESM2. These two models do not simply differ in one having a QBO, they will have different ozone climatologies. This makes it difficult to compare the experiment as interactive and non-interactive runs as the paper is trying to do. Furthermore, there are inconsistencies as to how ensembles are being used (different numbers for historical and future).
Response: We thank the reviewer’s comment. The main goal for the current paper is to disentangle the chemical and transport impact of QBO on ozone using a novel steady-state ozone metric. The nudging experiments are produced following the QBOi protocol intended to nudge only the tropical QBO dynamic variability while allowing other variabilities and variables freely evolve, thereby to ensure better assessments of its impact on global climate under a more realistic QBO without too much constraint on other processes. Despite large difference in model structure and ozone climatologies, both the E3SMv2 and CESM2 models show reasonable QBO and related impact (Figures 3 and 4), indicating the strong constraint of nudging despite of the ozone impact. Please refer to section 5.1 and Figures 3 and 4.
To further isolate the impact on interactive ozone, we conducted a separate E3SMv2 simulation of fixed-ozone (ozone input taken from CESM2). It is shown that we’d still get a reasonable QBO but with a stronger amplitude in temperature phase change (Figure S7). This indicates the stronger impact of nudging and the damping effect of interactive ozone on temperature. Please refer to Figure S7 and section 6.1 for further detail.
For the ensembles used in the current study, we’d like to clarify that each of the E3SMv2 and CESM2 nudged simulations used, we have been using ensemble mean of the 3 members (both historical of 1979-2014 and future of 2015-2020). It is only ozone input for CESM2 that had been put together using three members of WACCM for historical period (1979-2014) and one member of SSP370 (2014-2020). We have modified the text to make this clear. Please refer to line 142-148 and line 191-192 for further detail.
- The new QBO index is being calculated using a complicated method. It would be good to see more validation of this being performed. How related is this new index to the underlying dynamical variability? On line 173, “asymmetric” presumably refers to two phases rather than in latitude.
Response: The new index is relevant to the underlying dynamical variability in that it reasonably captures the asymmetric phase response of the climate to QBO (Figures 1 and S3). Therefore, we used this index to identify the QBO phase shift and form the composite analysis.
- The Linoz calculation is problematic. The major assumption that NOy is omitted from the chemical families whilst it has been shown by Chipperfield et al. (1994) and subsequent work such as by Tian et al. (2006) that NOy variations are the primary driver of ozone QBO changes above 20hPa. This major assumption will affect the conclusion of ozone behaviour above 20 hPa.
Response: To address the issue of chemistry species, we produced additional simulation for 1979-2020 using the nudged E3SMv2 with the “Linoz-v3 configuration” (Hsu and Prather, 2010). This “Linoz-v3 configuration” further adds stratospheric N2O, NOy, H2O, and CH4 chemistry to the current Linoz-v2 setup in E3SMv2. Adding the new chemistry better captures the ozone change in the positive QBO phase between 6-hPa to 20-hPa than those with Linoz-v2, and including especially the NOy variations alone can reproduce large parts of observed ozone change between 6-hPa to 20-hPa. These all indicate the important of NOy this region, and we should include NOy in steady-state ozone calculation where available. We have added this in the text (including the above citations in section 5.4) and figure and revised our conclusion, please refer to section 5.4 and 6, and Figures 10-12 for further detail.
- The authors speculate that the changes below 20 hPa are driven by dynamics. Using their tools, it would be possible to quantify this.
Response: Thank you for the comment. The main goal of the current study is to disentangle the chemical and transport impact of QBO on ozone using this novel steady-state ozone metric. We added a section 5.4 to specifically document how this metric is applied to the nudged E3SMv2 simulations. And we were able to separate the impact of QBO on ozone from temperature, NOy variation, and circulation change. The tool here can also be applied to other models to diagnose their difference among different models and observations. Please refer to section 5.4 for further detail.
Specific comments
- Line 271 and Figure 3a. I am having difficulty seeing a monopole to tripole pattern. It looks like a tripole in each phase?
Response: Thank you for pointing this out. We had changed to “tripole pattern”. Please refer to line 341, 342, and 346.
- Much more care needs to be taken with the figures. The colorbars especially do not have sensible divisions. Some figures are too small to be easily readable.
Response: We thank the reviewer for the comments. The figures are re-arranged and redrawn to make figures clearer. The colorbars are also revised as suggested.
References:
Chipperfield, M. P., L. J. Gray, J. S. Kinnersley, and J. Zawodny (1994), A two-dimensional model study of QBO signal in SAGE II NO2 and O3, Geophys. Res. Lett., 21, 589 – 592.
Tian, W., M. P. Chipperfield, L. J. Gray, and J. M. Zawodny (2006), Quasi-biennial oscillation and tracer distributions in a coupled chemistry-climate model, J. Geophys. Res., 111, D20301, doi:10.1029/2005JD006871.
Citation: https://doi.org/10.5194/egusphere-2024-1927-AC2
-
AC2: 'Reply on RC2', Jinbo Xie, 29 Jan 2025
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-1927', Anonymous Referee #1, 29 Jul 2024
Review of “Disentangling the chemistry and transport impacts of the Quasi-Biennial Oscillation on stratospheric ozone” by Xie et al.
The authors derive a new QBO index. They then use observations and tailored versions of two global models to investigate the relative impacts of photochemistry and transport on the response of stratospheric ozone to the QBO.
Major comments:
- While it is certainly important to understand the processes that contribute to the response of ozone to the QBO, the question addressed in this paper is quite limited and the answer has been known for decades. The authors’ goal is to determine the relative roles of chemistry and transport. However, they do not address what chemistry (i.e., which reactions; what leads them to change with the QBO) and what transport (meridional or vertical advection, diffusion due to wave breaking) are responsible. Instead, their assumptions seem to be that all photochemical effects are caused by temperature-dependent rates of ozone loss and all transport is due to the vertical component of the transformed Eulerian mean circulation. Beyond this, the important cross-over possibilities (i.e., transport of one reactant affects the photochemistry of another that is differently affected by transport) seem not to have been considered.
- Using simplified, constrained, or nudged models can often be a valuable tool for probing causative mechanisms. However, this paper does not present a clear motivation for why these model constraints were used in the present case. Overall, the paper is lacking in motivation for the specific modeling experiments that were set up. What can we learn from these simulations that we would not learn from either a middle atmosphere chemical transport model or a GCM with interactive chemistry? Moreover, additional evidence is needed to show that processes omitted due to their assumptions, such as neglect of full Ox chemistry, are not important.
- As a related comment, the authors do not take advantage of the simplified and controlled nature of their simulations to perform additional testing of mechanisms. For example, the last paragraph of Section 5.3 speculates about which differences in the setup of the different models might account for differences in the results. Why not actually test these?
- The derivation of a new QBO index is interesting. I especially appreciate Figure 1a, which displays the temporal development nicely. I would have been interested to see how well this index is able to capture the QBO interruptions of 2016 and 2020. My concern about this index is that its description needs work. It appears in the figure captions as “0 is centered on the month when QBO transits from QBOe to QBOw (determined by when current QBO index<0 and next QBO index>0). The QBO phase is determined by 5S-5N average of the zonal wind.” It's confusing because of the mention of the month when the QBO transits as if there was a single transit month rather than a different month at each pressure. Please take care to use language that distinguishes between the QBO winds and the QBO index. Perhaps something like this would be clearer: “The x-axis is the QBO index, defined by Eq. (1).” Also, please label the axis, e.g. “QBO index”.
Additional comments
- Supplemental figures 2 and 4 are not mentioned in the text. Are they necessary or helpful for understanding the results?
- (l. 18-19) It is not clear what “phase shift” refers to nor why E3SM cannot predict it.
- (l. 33-34) “assigning the pattern of ozone perturbations over the QBO cycle to specific processes is not easy”: This gets at the gist of the issue raised by my first major concern. One reason that it is not easy is that no simple either/or assignment can capture the complexity.
- (l. 125-126) “The parameters for the convective GW momentum transport were tuned especially for this version”: Can you provide more details, such as a link where the tuning choices and parameters can be found? Some other users of this model might be interested.
- (l. 407-408) “this demonstrates the overall effectiveness of the nudging strategy” I would say “overall effectiveness” is too strong since the extratropical response is not well simulated.
- (l. 419-420) “The results shown here are important for diagnosing model-model and model-observation differences” Can you provide an example?
- (l. 447-448) “using interactive ozone or not in the simulation does not significantly alter the results for QBO simulations” This has not been shown. Also, I suggest you take care with the term “significantly”, which is best reserved for quantitative measures of statistical significance.
Citation: https://doi.org/10.5194/egusphere-2024-1927-RC1 -
AC1: 'Reply on RC1', Jinbo Xie, 29 Jan 2025
Item-by-item response to all review comments
NOTE: To facilitate the evaluation of our responses, original review comments are listed first in their originals, followed by our itemized response.
We thank the reviewers’ comments, which are in text below. Our response is followed.
------------------------------------------------------------------------------
We thank the reviewers’ comments on helping to improve this manuscript. We have made substantial modifications to address the reviewer’s concerns as follows:
- The motivation of this paper is to demonstrate the novel use of the steady-state ozone metric on nudged climate model simulations to separate the chemical and transport impact of QBO on ozone. This is applicable to other models and would help to further diagnose model-to-model difference in future projection of QBO-ozone response, thereby contributing to improved ozone and climate projections. To make this clear, we made substantial modification to the introduction, and added additional sub-sections in section 6 and in discussions. Please refer to section 5.4 and section 6.1.
- To address the reviewers’ concerns on potential cross-over chemistry and uncertainty in the climate models, we’ve made two additional simulations, one using nudged E3SMv2 with Linoz-v3 (adding NOy-N2O-CH4-H2O chemistry), another using nudged E3SMv2 with fixed-ozone. It is shown that the NOy contributes to QBO-ozone between 6-hPa to 20-hPa. Please refer to section 5.4.
- Another new understanding from the comparison of the interactive/fixed ozone nudged E3SMv2 simulations show that the interactive ozone feedback damps the QBO-temperature, this is detailed in the discussion section. Please refer to section 6.1.
------------------------------------------------------------------------------
Reviewer 1
The authors derive a new QBO index. They then use observations and tailored versions of two global models to investigate the relative impacts of photochemistry and transport on the response of stratospheric ozone to the QBO.
Major comments:
1. While it is certainly important to understand the processes that contribute to the response of ozone to the QBO, the question addressed in this paper is quite limited and the answer has been known for decades. The authors’ goal is to determine the relative roles of chemistry and transport. However, they do not address what chemistry (i.e., which reactions; what leads them to change with the QBO) and what transport (meridional or vertical advection, diffusion due to wave breaking) are responsible. Instead, their assumptions seem to be that all photochemical effects are caused by temperature-dependent rates of ozone loss and all transport is due to the vertical component of the transformed Eulerian mean circulation. Beyond this, the important cross-over possibilities (i.e., transport of one reactant affects the photochemistry of another that is differently affected by transport) seem not to have been considered.
Response: Thank you for the comments. To further separate the role of chemistry/transport and examine impact of potential cross-over chemistry species, we performed additional simulations and analysis.
a. Transport: The Brewer-Dobson circulation (BDC) characterized by the residual mean circulation ( , ) is the major source of transport for ozone from the tropics to high-latitudes. In this study, we mainly focused on the anomalous as it has good correspondence with the ozone change below 20-hPa. Additional analysis on the anomalous (Figure S6) shows that the meridional transport of BDC is stronger above 20-hPa compared to below 20-hPa, which may have minor contribution to the ozone transport below 20-hPa. Please refer to line 503-510, Figures 4d-f, and Figure S6 for detail.
b. Chemistry species: The additional simulation is produced for 1979-2020 using the nudged E3SMv2 with the “Linoz-v3 configuration” (Hsu and Prather, 2010). This “Linoz-v3 configuration” further adds stratospheric N2O, NOy, H2O, and CH4 chemistry to the current Linoz-v2 setup in E3SMv2. Our simulations showed that E3SMv2 with “Linoz-v3 configuration” better captures the QBO-ozone change between 6-hPa to 20-hPa than those with Linoz-v2. This indicates the impact of added chemistry including NOy in this pressure range. This is added in the text and figure, please refer to section 5.3 and Figures 8 and 9 for detail.
Hsu, J., and M. J. Prather (2010), Global long-lived chemical modes excited in a 3-D chemistry transport model: Stratospheric N2O, NOy, O3 and CH4 chemistry, Geophys. Res. Lett., 37, L07805, doi:10.1029/2009GL042243.
2. Using simplified, constrained, or nudged models can often be a valuable tool for probing causative mechanisms. However, this paper does not present a clear motivation for why these model constraints were used in the present case. Overall, the paper is lacking in motivation for the specific modeling experiments that were set up. What can we learn from these simulations that we would not learn from either a middle atmosphere chemical transport model or a GCM with interactive chemistry? Moreover, additional evidence is needed to show that processes omitted due to their assumptions, such as neglect of full Ox chemistry, are not important.
Response: We thank the reviewer’s comment. Despite large studies of QBO-ozone, the current challenges remain in terms of the model’s ability to simulate QBO variability (Richter et al., 2020) and its phase asymmetry (Scaife et al., 2014). To address this issue, we produce a set of nudged simulations following the protocol of QBOi. Unlike other nudging experiments, the current setup is intended to nudge only the tropical QBO dynamic variability while allowing other variabilities (e.g. semi-annual oscillation) outside of the tropical lower-stratosphere and other variables (e.g. temperature) freely evolve, thereby to ensure better assessments of its impact on global climate under a more realistic QBO without too much constraint on other processes.
To further examine the potential impact of other species on QBO-ozone, we conducted an additional nudged simulation of E3SMv2 with “Linoz-v3 configuration” (adding the N2O-NOy-H2O-CH4 chemistry). It is shown that NOy chemistry does contribute to a non-negligible impact the QBO-ozone relationship at between 6-hPa to 20-hPa, we thus refined our conclusions on NOy, we thus refined our conclusions based on the new results. Please refer to section 5.1 and Figures 8 and 9 for further detail.
3. As a related comment, the authors do not take advantage of the simplified and controlled nature of their simulations to perform additional testing of mechanisms. For example, the last paragraph of Section 5.3 speculates about which differences in the setup of the different models might account for differences in the results. Why not actually test these?
Response: To address the mentioned issue, an additional E3SMv2 nudged experiment with fixed ozone (E3SMv2 nudged fixed-ozone, using ozone input from CESM2) is added to further separate the impact of different setups. It is shown that the results still exhibit QBO in-phase changes, indicating a stronger constraint of nudging. The temperature phase change pattern in E3SMv2 nudged fixed-ozone (Figure S7) shows larger amplitude than that of the E3SMv2 nudged (Figures 4 and 5), indicating the damping effect of interactive ozone on temperature. Please refer to discussion section 6.1, and Figures 4, 5, and S7 for further detail.
4. The derivation of a new QBO index is interesting. I especially appreciate Figure 1a, which displays the temporal development nicely. I would have been interested to see how well this index is able to capture the QBO interruptions of 2016 and 2020. My concern about this index is that its description needs work. It appears in the figure captions as “0 is centered on the month when QBO transits from QBOe to QBOw (determined by when current QBO index<0 and next QBO index>0). The QBO phase is determined by 5S-5N average of the zonal wind.” It's confusing because of the mention of the month when the QBO transits as if there was a single transit month rather than a different month at each pressure. Please take care to use language that distinguishes between the QBO winds and the QBO index. Perhaps something like this would be clearer: “The x-axis is the QBO index, defined by Eq. (1).” Also, please label the axis, e.g. “QBO index”.
Response: We thank the reviewer for the comment. To avoid the confusion in description, we changed the figure caption to “0 is centered on the month when QBO index shifts from QBOe to QBOw (determined by when current QBO index<0 and next QBO index>0).” Thereby to ensure a single month centered on in the figure. Since the month -14 to 14 are offsets of month from the center 0 month, we’d like to request to change the axis label to “offset (month)” as is done now to make this description clearer. Please refer to the Figures 3 to 12 for detail.
Additional comments
1. Supplemental figures 2 and 4 are not mentioned in the text. Are they necessary or helpful for understanding the results?
Response: We thank the reviewer for the comment. Supplementary Figure 2 is used to depict the difference between NLPCA and PCA QBO index. The description is added in the text, please refer to line 308-311. Supplementary Figure 4 is removed. The supplementary Figures are also tagged Figure S… for more concise labelling.
2. (l. 18-19) It is not clear what “phase shift” refers to nor why E3SM cannot predict it.
Response: We have changed to “timing of the phase transitions” since it refers to the a transition point in time. E3SM has a shorter period and does not capture the right timing for these QBO phase transitions. Please refer to line 18-19.
3. (l. 33-34) “assigning the pattern of ozone perturbations over the QBO cycle to specific processes is not easy”: This gets at the gist of the issue raised by my first major concern. One reason that it is not easy is that no simple either/or assignment can capture the complexity.
Response: Agreed. We added further Linoz-v3 tests so as to further analyze the impact of N2O-NOy-CH4-H2O chemistry on ozone simulation. It is shown that NOy plays a large role around 10 to 20-hPa as the reviewers have mentioned.
4. (l. 125-126) “The parameters for the convective GW momentum transport were tuned especially for this version”: Can you provide more details, such as a link where the tuning choices and parameters can be found? Some other users of this model might be interested.
Response: The reference for the tuning of GW momentum transport were added to the text. Please refer to lines 125-126.
5. (l. 407-408) “this demonstrates the overall effectiveness of the nudging strategy” I would say “overall effectiveness” is too strong since the extratropical response is not well simulated.
Response: We thank the reviewer’s comment. This statement is removed from the text.
6. (l. 419-420) “The results shown here are important for diagnosing model-model and model-observation differences” Can you provide an example?
Response: Thank you for the comment. The main goal of the current study is to disentangle the chemical and transport impact of QBO on ozone using a novel steady-state ozone metric. Section 5.4 documents the application of this new metric on the nudged E3SMv2 simulations, where we were able to separate the impact of QBO on ozone from temperature, NOy variation, and circulation change. This may also be applied to other models to trace their uncertainty from these impacts, thereby to diagnose their difference among different models and observations.
7. (l. 447-448) “using interactive ozone or not in the simulation does not significantly alter the results for QBO simulations” This has not been shown. Also, I suggest you take care with the term “significantly”, which is best reserved for quantitative measures of statistical significance.
Response: Thank you for pointing this out. We have removed the statement.
Citation: https://doi.org/10.5194/egusphere-2024-1927-AC1
-
RC2: 'Comment on egusphere-2024-1927', Anonymous Referee #2, 07 Aug 2024
General comments
The paper explores the chemistry and transport impact of the QBO on stratospheric ozone. They make use of climate model simulations as well as an offline ozone calculation to investigate the response of ozone to temperature and overhead ozone changes. They claim that ozone perturbations above 20hPa are predicted by the ozone column response to temperature whilst those below show no temperature response and are presumed to be driven by the circulation changes.
1) The motivation for the choice of methodology is not made clear in the paper. It is not apparent what is being gained by it. Far too many other uncertainties are being introduced by the methodology that lead to questions about the robustness of the results. The work makes use of two different climate models, EAMv2 and CESM2. These two models do not simply differ in one having a QBO, they will have different ozone climatologies. This makes it difficult to compare the experiment as interactive and non-interactive runs as the paper is trying to do. Furthermore, there are inconsistencies as to how ensembles are being used (different numbers for historical and future).
2) The new QBO index is being calculated using a complicated method. It would be good to see more validation of this being performed. How related is this new index to the underlying dynamical variability? On line 173, “asymmetric” presumably refers to two phases rather than in latitude.
3) The Linoz calculation is problematic. The major assumption that NOy is omitted from the chemical families whilst it has been shown by Chipperfield et al. (1994) and subsequent work such as by Tian et al. (2006) that NOy variations are the primary driver of ozone QBO changes above 20hPa. This major assumption will affect the conclusion of ozone behaviour above 20 hPa.
4) The authors speculate that the changes below 20 hPa are driven by dynamics. Using their tools, it would be possible to quantify this.
Specific comments
1) Line 271 and Figure 3a. I am having difficulty seeing a monopole to tripole pattern. It looks like a tripole in each phase?
2) Much more care needs to be taken with the figures. The colorbars especially do not have sensible divisions. Some figures are too small to be easily readable.
References:
Chipperfield, M. P., L. J. Gray, J. S. Kinnersley, and J. Zawodny (1994), A two-dimensional model study of QBO signal in SAGE II NO2 and O3, Geophys. Res. Lett., 21, 589 – 592.
Tian, W., M. P. Chipperfield, L. J. Gray, and J. M. Zawodny (2006), Quasi-biennial oscillation and tracer distributions in a coupled chemistry-climate model, J. Geophys. Res., 111, D20301, doi:10.1029/2005JD006871.
Citation: https://doi.org/10.5194/egusphere-2024-1927-RC2 -
AC2: 'Reply on RC2', Jinbo Xie, 29 Jan 2025
Item-by-item response to all review comments
NOTE: To facilitate the evaluation of our responses, original review comments are listed first in their originals, followed by our itemized response.
We thank the reviewers’ comments, which are in text below. Our response is followed.
------------------------------------------------------------------------------
We thank the reviewers’ comments on helping to improve this manuscript. We have made substantial modifications to address the reviewer’s concerns as follows:
- The motivation of this paper is to demonstrate the novel use of the steady-state ozone metric on nudged climate model simulations to separate the chemical and transport impact of QBO on ozone. This is applicable to other models and would help to further diagnose model-to-model difference in future projection of QBO-ozone response, thereby contributing to improved ozone and climate projections. To make this clear, we made substantial modification to the introduction, and added additional sub-sections in section 6 and in discussions. Please refer to section 5.4 and section 6.1.
- To address the reviewers’ concerns on potential cross-over chemistry and uncertainty in the climate models, we’ve made two additional simulations, one using nudged E3SMv2 with Linoz-v3 (adding NOy-N2O-CH4-H2O chemistry), another using nudged E3SMv2 with fixed-ozone. It is shown that the NOy contributes to QBO-ozone between 6-hPa to 20-hPa. Please refer to section 5.4.
- Another new understanding from the comparison of the interactive/fixed ozone nudged E3SMv2 simulations show that the interactive ozone feedback damps the QBO-temperature, this is detailed in the discussion section. Please refer to section 6.1.
------------------------------------------------------------------------------
Reviewer 2
General comments
The paper explores the chemistry and transport impact of the QBO on stratospheric ozone. They make use of climate model simulations as well as an offline ozone calculation to investigate the response of ozone to temperature and overhead ozone changes. They claim that ozone perturbations above 20hPa are predicted by the ozone column response to temperature whilst those below show no temperature response and are presumed to be driven by the circulation changes.
- The motivation for the choice of methodology is not made clear in the paper. It is not apparent what is being gained by it. Far too many other uncertainties are being introduced by the methodology that lead to questions about the robustness of the results. The work makes use of two different climate models, EAMv2 and CESM2. These two models do not simply differ in one having a QBO, they will have different ozone climatologies. This makes it difficult to compare the experiment as interactive and non-interactive runs as the paper is trying to do. Furthermore, there are inconsistencies as to how ensembles are being used (different numbers for historical and future).
Response: We thank the reviewer’s comment. The main goal for the current paper is to disentangle the chemical and transport impact of QBO on ozone using a novel steady-state ozone metric. The nudging experiments are produced following the QBOi protocol intended to nudge only the tropical QBO dynamic variability while allowing other variabilities and variables freely evolve, thereby to ensure better assessments of its impact on global climate under a more realistic QBO without too much constraint on other processes. Despite large difference in model structure and ozone climatologies, both the E3SMv2 and CESM2 models show reasonable QBO and related impact (Figures 3 and 4), indicating the strong constraint of nudging despite of the ozone impact. Please refer to section 5.1 and Figures 3 and 4.
To further isolate the impact on interactive ozone, we conducted a separate E3SMv2 simulation of fixed-ozone (ozone input taken from CESM2). It is shown that we’d still get a reasonable QBO but with a stronger amplitude in temperature phase change (Figure S7). This indicates the stronger impact of nudging and the damping effect of interactive ozone on temperature. Please refer to Figure S7 and section 6.1 for further detail.
For the ensembles used in the current study, we’d like to clarify that each of the E3SMv2 and CESM2 nudged simulations used, we have been using ensemble mean of the 3 members (both historical of 1979-2014 and future of 2015-2020). It is only ozone input for CESM2 that had been put together using three members of WACCM for historical period (1979-2014) and one member of SSP370 (2014-2020). We have modified the text to make this clear. Please refer to line 142-148 and line 191-192 for further detail.
- The new QBO index is being calculated using a complicated method. It would be good to see more validation of this being performed. How related is this new index to the underlying dynamical variability? On line 173, “asymmetric” presumably refers to two phases rather than in latitude.
Response: The new index is relevant to the underlying dynamical variability in that it reasonably captures the asymmetric phase response of the climate to QBO (Figures 1 and S3). Therefore, we used this index to identify the QBO phase shift and form the composite analysis.
- The Linoz calculation is problematic. The major assumption that NOy is omitted from the chemical families whilst it has been shown by Chipperfield et al. (1994) and subsequent work such as by Tian et al. (2006) that NOy variations are the primary driver of ozone QBO changes above 20hPa. This major assumption will affect the conclusion of ozone behaviour above 20 hPa.
Response: To address the issue of chemistry species, we produced additional simulation for 1979-2020 using the nudged E3SMv2 with the “Linoz-v3 configuration” (Hsu and Prather, 2010). This “Linoz-v3 configuration” further adds stratospheric N2O, NOy, H2O, and CH4 chemistry to the current Linoz-v2 setup in E3SMv2. Adding the new chemistry better captures the ozone change in the positive QBO phase between 6-hPa to 20-hPa than those with Linoz-v2, and including especially the NOy variations alone can reproduce large parts of observed ozone change between 6-hPa to 20-hPa. These all indicate the important of NOy this region, and we should include NOy in steady-state ozone calculation where available. We have added this in the text (including the above citations in section 5.4) and figure and revised our conclusion, please refer to section 5.4 and 6, and Figures 10-12 for further detail.
- The authors speculate that the changes below 20 hPa are driven by dynamics. Using their tools, it would be possible to quantify this.
Response: Thank you for the comment. The main goal of the current study is to disentangle the chemical and transport impact of QBO on ozone using this novel steady-state ozone metric. We added a section 5.4 to specifically document how this metric is applied to the nudged E3SMv2 simulations. And we were able to separate the impact of QBO on ozone from temperature, NOy variation, and circulation change. The tool here can also be applied to other models to diagnose their difference among different models and observations. Please refer to section 5.4 for further detail.
Specific comments
- Line 271 and Figure 3a. I am having difficulty seeing a monopole to tripole pattern. It looks like a tripole in each phase?
Response: Thank you for pointing this out. We had changed to “tripole pattern”. Please refer to line 341, 342, and 346.
- Much more care needs to be taken with the figures. The colorbars especially do not have sensible divisions. Some figures are too small to be easily readable.
Response: We thank the reviewer for the comments. The figures are re-arranged and redrawn to make figures clearer. The colorbars are also revised as suggested.
References:
Chipperfield, M. P., L. J. Gray, J. S. Kinnersley, and J. Zawodny (1994), A two-dimensional model study of QBO signal in SAGE II NO2 and O3, Geophys. Res. Lett., 21, 589 – 592.
Tian, W., M. P. Chipperfield, L. J. Gray, and J. M. Zawodny (2006), Quasi-biennial oscillation and tracer distributions in a coupled chemistry-climate model, J. Geophys. Res., 111, D20301, doi:10.1029/2005JD006871.
Citation: https://doi.org/10.5194/egusphere-2024-1927-AC2
-
AC2: 'Reply on RC2', Jinbo Xie, 29 Jan 2025
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 529 | 166 | 35 | 730 | 54 | 29 | 47 |
- HTML: 529
- PDF: 166
- XML: 35
- Total: 730
- Supplement: 54
- BibTeX: 29
- EndNote: 47
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Michael Prather
Jadwiga Richter
Shixuan Zhang
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(17949 KB) - Metadata XML
-
Supplement
(8244 KB) - BibTeX
- EndNote
- Final revised paper