the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Rapid assimilations of O3 observations – Part 1: methodology and tropospheric O3 changes in China in 2015–2020
Rui Zhu
Zhaojun Tang
Xiaokang Chen
Xiong Liu
Abstract. The high computational cost of chemical transport models (CTMs) is a potential bottleneck for the rapid assimilation of ozone (O3) observations. Here we developed a single tracer tagged-O3 mode to build the capability of the GEOS-Chem model for rapid simulation of tropospheric O3. The tagged-O3 mode demonstrates high consistency with GEOS-Chem full-chemistry simulation and dramatic reductions in computational costs by approximately 91–94 %. The tagged-O3 simulation was combined with China Ministry of Ecology and Environment (MEE) and Ozone Monitoring Instrument (OMI) O3 observations to investigate the changes in tropospheric O3 over E. Asia in 2015–2020. The assimilated O3 concentrations demonstrate good agreement with O3 observations: surface O3 concentrations are 42.9, 41.8 and 42.1 ppb; and tropospheric O3 columns are 37.1, 37.9 and 38.0 DU in the simulations, assimilations and observations, respectively. The assimilations indicate rapid increases in surface O3 by 1.60 (spring), 1.16 (summer), 1.47 (autumn) and 0.80 (winter) ppb yr-1 over E. China in 2015–2020, and the increasing trends are underestimated by the a priori simulations. More attention is thus suggested to the rapid increases in O3 pollution in spring and autumn. Furthermore, we find stronger increases in tropospheric O3 columns over highly polluted areas, which may reflect the larger contributions of local emissions. The large discrepancy in the trends in tropospheric O3 columns by assimilating surface and satellite observations further indicates the possible uncertainties in the derived free tropospheric O3 changes. The rapid O3 assimilation capability is a useful tool for the extension and interpretation of atmospheric O3 observations.
- Preprint
(4683 KB) -
Supplement
(832 KB) - BibTeX
- EndNote
Rui Zhu et al.
Status: final response (author comments only)
-
RC1: 'Comment on gmd-2023-35', Anonymous Referee #1, 08 May 2023
This paper describes the rapid assimilation of surface and column ozone into the GEOS-Chem model and uses this approach to quantify changes in ozone over China between 2015 and 2020. It provides useful confirmation of the ozone changes over this important and rapidly-evolving region. However, the paper has a number of major deficiencies that make it unsuitable for publication in GMD in its present form. Specifically, the purpose and rationale of the study are not stated; the motivation for a single-tracer assimilation is not described; the benefits of "rapid" assimilation are not explained; and the results for ozone trends are not compared with those from any previous studies to provide context for the reader. These are major deficiencies that need to be addressed fully before the paper can be considered for publication.
I have looked at Part 2 of this study under consideration for ACP (also a weak paper) and would suggest that the two papers are combined into a single paper. It makes little scientific sense to split the study into two given that the tools, approach and aims are the same, and only the regions of interest differ. This would address a number of obvious problems, and the combined paper would be more complete, making a more useful addition to the published literature.
General Comments
The term "tagged-O3 mode" is not explained here, and is confusing as no tagging is used anywhere in this study. Please explain what this term means in the present context. "Single tracer mode" would be a much clearer and more appropriate description given that the model is run with a single ozone tracer.
What is the purpose of the ozone assimilation? The apparent goal of the study (to identify ozone changes in China) can be performed from the observations alone, so what additional information does assilimation provide? This needs to be explained clearly. The introduction does not provide a rationale for the approach or explain why it is needed.
What are the benefits of "rapid" assimilation? The analysis described in the paper could be performed equally well with assimilation using the full model, so what value does the speed up provide? I can see a benefit if the assimilation is to be repeated a very large number of times, but this does not appear to be the case here.
How do the ozone trends over China compare with previous observational or model-based studies? Many recent studies have quantified these, so it is essential to provide a comparison as context for the reader and to demonstrate the value of the approach adopted here.
There is no consideration of uncertainty in the results, or any attempt to explore the significance of the derived trends. While the a posteriori results are compared with the a priori model results, no attempt is made to investigate or explain why the a priori results might be wrong, and this is a missed opportunity.
There is an over-reliance on referencing recent papers for general material such as that presented at the start of the introduction. This suggests that the authors are not familiar with the wider literature. Where references are needed, please cite original primary references and not recent derivative studies.
Specific Comments:
Line 44: the emphasis of this sentence reveals some weaknesses in perspective. Surface stations provide the most direct information on air quality, while satellite observations are much less important for this.
Line 46: "assimilation of surface observations can effectively improve the predicted surface O3 concentrations". This statement appears obvious; please explain why it is useful.
Line 52: better understanding requires photochemistry, not assimilation; this sentence is incorrect and needs rephrasing.
Lines 96-100: The AQS and AirBase data are not used in this paper, so this information should be removed unless the two parts of this study are combined. It would be more useful to comment on the locations of the MEE sites, in particular on how many are urban and suburban, and if any rural sites are available. What issues are associated with using a measurement network that is predominantly urban?
Line 121-124: How does this convolution approach differ from previous studies? If this is standard, please cite the original literature.
Line 138 states that MEIC emisisons are used, while Line 143 indicates that emissions are scaled corresponding to MEIC. Which of these approaches is used?
Line 148: information on US and Europe is provided here, but are not needed in this paper unless the papers are combined.
Line 168: "The model errors are assumed to be 50%." Why? Some explanation is needed here.
Line 174: If the errors are calculated on a station basis, how is the grid-based superobservation applied? More information is required here.
Line 201: Why is it necesary to "design and perform different assimilation experiments"? Is this just to improve the method? This is an important point, as the rationale for development of a rapid assimilation method depends upon it.
Line 211: The difference between tagged-O3 and tagged-Ox needs to be explained clearly to the reader before this (no one outside the GEOS-Chem community is likely to understand the distinction).
Line 221-224: This point is poorly explained. The issue is associated with the timescales for turblent transport in the PBL vs. chemical timescales, and this justifies the need to adjust ozone throughout the PBL. Please explain how the factor of 0.8 was chosen.
Para 332: This explanation is unconvincing, and clearer justification is required. The differing seasonality of the column over the NCP is interesting, but no evidence of transport differences is provided. It would be possible to diagnose this properly using the tagged-O3 approach for the purpose it was designed for.
Typos and minor issues
Line 91: "have the ability to report" Please rephrase this.
Line 129: "does not cancel out" - this point is poorly described, please rephrase.
Line 226: The "l" and "1" are too similar, please change the notation here (use "n" for the layer?)
Fig S1 is very unclear. What does this framework show, and what information is passed following the arrows? Which aspects are full chemistry and which are single tracer? This diagram needs to be reconsidered and redrawn.
A large number of multi-panel figures are included in the paper, but many of the results are only very briefly mentioned in the text. The paper would be sharper and clearer if the authors were more selective and moved some of these to the supplement.
Citation: https://doi.org/10.5194/gmd-2023-35-RC1 -
RC2: 'Comment on gmd-2023-35', Anonymous Referee #2, 26 May 2023
In this study, the authors developed single tracer tagged-O3 mode of the GEOS-Chem model to investigate the tropospheric and surface NO2 and O3 changes in China. Further data assimilations were performed with both surface and satellite observation. The authors also pointed out a companion paper on ACPD which applied the method developed here to study the ozone changes in US and Europe. Unfortunately, these two papers are not in companion order, either in ACP or GMD. However, reading from this manuscript, it was still drafted to be the case. I suggest the authors spend some time to reorganize the manuscript, so it will be independent of the other one. For example, in line 96-100: I suggest the authors leave these few sentences to the companion paper.
Line 21: give full name of E. Asia, and also E. China in line 25 since they appear for the first time. Pay attention to other abbreviations in the main context.
Citation: https://doi.org/10.5194/gmd-2023-35-RC2
Rui Zhu et al.
Rui Zhu et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 237 | 54 | 9 | 300 | 22 | 2 | 2 |
- HTML: 237
- PDF: 54
- XML: 9
- Total: 300
- Supplement: 22
- BibTeX: 2
- EndNote: 2
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1