the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
A quantitative decoupling analysis (QDA v1.0) method for the assessment of meteorological, emission and chemical contributions to fine particulate pollution
Abstract. A comprehensive understanding of the effects of meteorology, emission and chemistry on severe haze is critical in the mitigation of air pollution. However, such understanding is largely hindered by the nonlinearity of atmospheric chemistry systems. Here, we developed a novel quantitative decoupling analysis (QDA) method to quantify the effects of emission, meteorology, chemical reaction, and their nonlinear interactions on the fine particulate matter (PM2.5) pollution based on the accompanying simulations for different atmospheric processes. Via embedding the QDA method into the Weather Research and Forecasting-Nested Air Quality Prediction Modeling System (WRF-NAQPMS) model, we first employed this method into a typical heavy haze episode in Beijing. Different from the previously sensitive simulation method, which usually linked to a certain period, the QDA achieves the fully decomposing analysis of PM2.5 concentration during any pollution event into seven different parts, including meteorological contribution (M), emission contribution (E), chemical contribution (C), and interactions among these drivers (i.e., ME, MC, EC and MCE). The results show that the meteorology contribution varied significantly at different stages of episode, from 0.21 µg·m−3·h−1 during accumulation period to −11.82 µg·m−3·h−1 during the removal period, dominating the hourly changes of PM2.5 concentrations. The chemical contributions were shown to increase with the level of haze, which become largest (0.37 µg·m−3·h−1) at the maintenance period, 25 % higher than that during the clean period. The contribution of primary emission is relatively stable in all stages due to the use of fixed emission during the simulation. Besides, the QDA method highlights that there exist nonnegligible coupling effects of meteorology, emission and chemistry on PM2.5 concentrations (−1.83 to 2.44 µg·m−3·h−1), which were commonly ignored in previous studies and the development of heavy-pollution control strategies. These results indicate that the QDA method can not only provide researchers and policy makers with valuable information for understanding of key factors to heavy pollution, but also help the modelers to find out the sources of uncertainties among numerical models.
- Preprint
(907 KB) - Metadata XML
-
Supplement
(1191 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on gmd-2021-259', Anonymous Referee #1, 22 Sep 2021
General comments
The authors describe a method for determining the contributions of different processes to PM2.5 formation and the couplings between the processes. The authors call this method Quantitative Decoupling Analysis (QDA) and apply the method to a haze episode in the Beijing-Tianjin-Hebei region from 17-28 February 2014. The manuscript is generally well-written.
However, there are three significant problems with the work:
* The QDA method is not new. This is the Factor Separation method introduced in 1993 and applied in later work. See references below.
* The emissions are constant throughout the simulation period, which is not realistic. As a consequence, the contribution of emissions to the PM2.5 concentration change is constant throughout the episode, and all the time-variation in the factors and couplings is driven by the meteorology.
* The authors consider the influence of three factors on PM2.5, total emissions, chemistry, and meteorology, and indicate that their work provides valuable information to decision makers (lines 64-74). PM2.5 pollution episodes are driven by anthropogenic emissions and meteorology. Chemistry is a secondary factor that responds to emissions and meteorology but can be controlled by decision makers only by regulating the anthropogenic emissions. It would be much more relevant to decision makers if the authors had chosen biogenic emissions, anthropogenic emissions, and meteorology as the three factors. Then the full effect of anthropogenic emissions on PM2.5 during the episode would be apparent, rather than burying some of the effect in the chemistry factor. Also, the results would help determine if emergency anthropogenic emission controls during episodes would reduce PM2.5, which is a goal of the authors’ work (lines 71-72) but not a result of their work.
Specific comments
p. 3, lines 64-72. Decision makers can control anthropogenic emissions and possibly have a minor impact on some biogenic emissions (e.g., types of trees planted in urban areas). Understanding the impact of meteorology on atmospheric concentrations is also important and useful to decision makers. But separating out the impact of chemical reactions does not help regulators reduce atmospheric concentrations. The chemistry factor is controlled by emissions and meteorology, so some (not all) of the chemistry factor represents the impact of emissions. The decision makers need to understand the full impact of the emissions, but that cannot be obtained from the factors that the authors chose.
p. 4, lines 89-91. Eq. 2 is incorrect. There should be a factor of 2 in front of the cross terms Δx1Δx2, Δx2Δx3, and Δx1Δx3 in the group of second-order terms and other non-unity factors derived from the binomial coefficients in front of the cross terms in the group of third-order terms. The authors may not have used Eq. 2 and used only Eq. 3. However, if the authors actually used Eq. 2 in their calculations and analyses, they should verify that they used the correct equation, and, if not, the calculations and analyses must be re-done. In any case, Eq. 2 should be corrected.
p. 4, line 104. The interaction between emissions and meteorology is bi-directional. Higher temperatures increase evaporative emissions from gasoline vehicles, higher temperatures and greater sunlight increase isoprene emissions from plants, etc.
p.4, lines 106-112 and Table 1. There should be a more complete description of the simulations and what is different between simulation M1 and the other simulations. In particular for simulation M4, are the meteorological processes and emissions absent for the entire simulation? If so, what PM2.5 could there be in a grid cell other than the initial PM2.5 concentration, which is stationary in space because no meteorological processes are included? Is the PM2.5 concentration at the start of a specific time step taken from simulation M1, simulation M4 is run over that time step without including meteorology processes or emissions anywhere in the modeling domain, then a new PM2.5 concentration is obtained from simulation M1 for the next time step? What the authors did is very unclear.
p. 4, lines 111-114. Did the authors run the base simulation 6 times, each time with one of the “accompanying” simulations? That would effectively be 12 simulations. A simpler approach would seem to be running the base simulation once, recording the timesteps used, and then running each of the “accompanying” simulations once with the same timesteps used for the base case. That would reduce the number of simulations needed to 7.
p. 4, lines 119-120. There should be a detailed explanation of how IPR is applied to the results for each factor. It is unclear how this was done. Simulation M4 (C factor) does not contain emissions, so the chemistry will be different from that when emissions are present. Are the IPR results then meaningful for C?
pp. 3-5. The QDA method is not new. This is the Factor Separation method introduced by U. Stein and P. Alpert, Factor separation in numerical simulations, J. Atmos. Sci. 50, 2107-2115 (1993). Subsequently, Tao et al. applied the method to separate the contributions of area, mobile, and point source emissions to ozone and their interactions(Tao et al., Area, mobile, and point source contributions to ground level ozone: a summer simulation across the continental USA, Atmos. Environ. 39, 1869-1877 (2005)). The authors should not refer to QDA as a new method and should credit Stein and Alpert and Tao et al. by including their references in the manuscript.
p. 7, line 187. What is MBE? This is not defined in Table S1. MB is -13.7 μg/m3 and ME is 42.1 μg/m3 (Table S2) so it cannot be either of those two statistics.
p. 7, line 191. Again, what is MBE and where are these values (7.1 and 5.3 μg/m3 ) in Table S2? If the values are discussed in the manuscript, they should be in Table S2.
p. 7, line 194. There is a more recent paper (L. Huang et al., Atmos. Chem. Phys. 21, 2725-2743 (2021)) that gives goals and criteria specifically for PM 2.5 simulations in China.
p. 7, lines 210-212. For their analyses, the authors fixed the emissions to be constant in time. It is unclear why this is necessary for the method, and it is a serious limitation of their work. Neither the anthropogenic nor the biogenic emissions are constant in time; there are large variations over the diurnal cycle. As a consequence of the authors’ assumption of constant emissions, their calculated emission contribution is constant over all 12 days of the episode (Figures 7 and 8, Tables 3 and 4). This is not an interesting or very valuable result, especially for the decision makers/regulators. We cannot control the meteorology, only the anthropogenic emissions, so the important question is to what extent instituting greater emission controls during stagnation events will improve air quality. The authors’ results do not provide any insight on that question. Further, the assumption of constant emissions also influences the chemical contribution because time-varying emissions would very likely give much greater variation in the chemistry contribution.
p. 8, lines 234-239. These conclusions are well-known from many previous studies.
p. 9, line 276. It is unclear what the range of -0.86 to 1.86 represents. It is much wider than what the results in Table 3 suggest.
pp. 10-11, Section 3.4. Many of the conclusions here are well-known from prior work, and the Factor Separation method (QDA) adds little new information to the prior work. At most, this section shows consistency between the Factor Separation method and the results of previous studies, but there is no detailed evaluation of the Factor Separation method.
p. 10, lines 281-282. Yes, the results in the paper do not give much information about the importance of emissions and therefore are not of much use to decision makers.
pp. 11-12. Again, QDA is not a new method and most of the conclusions here are not new.
Technical corrections
p. 2, lines 54-55. “However, due to the nonrepeatability of individual pollution cases, … .” Not clear what is meant here. If one has an estimate of the meteorological fields from a weather model and an estimate of the emissions, the air quality model can estimate the atmospheric concentration of PM for days in different years (“individual cases”). Not clear why sensitivity experiments are necessary to “fully reproduce the individual cases.”
p. 2, line 57. Define the PLMA acronym.
p. 5, line 133. Should be “nitrate” not “nitrite”?
p. 5, line 149. Was it MOZART v 2.4 or v 2.5?
p. 6, line 159. “The” should be “the”
Figure 2 (b). The legend should be larger.
p. 6, lines 179-180. It would be clearer to use the same nomenclature for these statistics as in Table S1. Table S1 has the conventional names.
Figure 6 caption. There is no solid line in the figure, only 3 dashed lines. Do the points represent 24 hour averages? For which days?
Tables 3 and 4 should be in the Supporting Information because they repeat the information in Figure 8.
p. 9, line 259. Not clear why one limit is -3%. This seems to be a comparison of the magnitudes of the two quantities, in which case the limit would be +3%.
p. 10, line 293. Should it be “from stages 1 to 2” instead of :from stages 2 to 1”?
p. 11, line 311. Define the acronym PLAM.
p. 14, lines 436-437. The title of the paper should not be in all capitals.
Citation: https://doi.org/10.5194/gmd-2021-259-RC1 -
AC1: 'Reply on RC1', Baozhu Ge, 05 Dec 2021
The comment was uploaded in the form of a supplement: https://gmd.copernicus.org/preprints/gmd-2021-259/gmd-2021-259-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Baozhu Ge, 05 Dec 2021
-
RC2: 'Comment on gmd-2021-259', Anonymous Referee #2, 28 Sep 2021
Review comment on gmd-2021-259
General Comments:
The manuscript ‘A quantitative decoupling analysis (QDA v1.0) method for the assessment of meteorological, emission and chemical contributions to fine particulate pollution’ written by Junhua Wang presented the QDA method as novel way to evaluate meteorology, emission, and chemistry processes involved for the aerosol formation. Although the concept of this method is interesting, I cannot fully understand the description of method itself and therefore go through to result and discussion section well. At the current presentation quality, this manuscript cannot be considered for publication. At this round, I would like to reject this manuscript. Before considering the possible publication, I sincerely request the fundamental amendments. I wish the following major and minor comments will help to revise this manuscript.
Major Comments:
1. The description of QDA and its relation to IPR
The newly developed QDA method is just the using of six accompanying simulations to calculate M, E, and C terms. In this sense, for example, to drive M term, this method seems to be identical to the SAA as described in the introduction. Actually, how to conduct six accompanying simulations is unclear. Under each time-step simulation, how can do the base-model derive each process? The detailed description of M2-M7 is required to understand the QDA method. In addition, without E term, C term cannot be driven due to the absence of precursors. Therefore I guess that EC term inherently connected and it could be hard to be divided. Moreover, on P4, L118, it was stated that “The above QDA method can also be combine with the IPR method to resolve more detailed information…”. This statement is confusing to me because this impressed that QDA is just the using of IPR. Under the current presentation quality, it is difficult to understand QDA method and I cannot recognize this method as novel way in modeling analysis.
2. Results and discussion of QDA.
Because the description of QDA is insufficient, I also cannot follow the result and discussion section. Why E term showed same values through analyzed stages? Is this because emissions did not consider temporal variation through analyzed episode? The meteorological field are shown in Fig. 4, but how about the precipitation? Because the term of “wetdep” was 0.00 through stages, I felt that there was no rain. Although this was the severe haze event, without the wet deposition analysis, this episode seems to be not interesting as test case to show the QDA result. As evaluated using NOR and SOR, I like the idea to consider the formation process from the viewpoint of each specie. The result of QDA is now discussed for PM2.5; however, each specie have been evolved as different E and C terms. I would like to strongly recommend to show the same kind of analysis of Figs. 7-9 for each specie. This analysis will offer the insight into C roles on chemical formation during haze episode.
3. The application of QDA.
As found in the abstract, this QDA method could help modelers to understand the each process and find these uncertainties. I have briefly checked the source code of QDA distributed in ZENODO, but I felt that the fortran90 codes seems to be incorporated into the NAQPMS source codes. How can we apply this source code into other models? If the authors claimed that “QDA is a universal tool”, the explanation for how to use this QDA method in other models codes should be kindly introduced within this distribution.
Minor Comments:
P2, L42: CMAQ have to be introduced after the definition of CTM (P2, L56). The organization of introduction for second and third paragraphs should be reconsidered.
P2, L41-46: Under this context, IRR should be also carefully introduced. The IRR can be used to define the role of reaction rate, and this will relate C term in this study.
P4, L92-L99 and Eq. 2: How can we treat the second- and third-order partial differential of x1. x2, and x3? Does this represent the nonlinear term of M, E, and C? What stands for them?
P4, L104: For example, higher temperature will relate activated plant, and change the biogenic emissions intensity. Why E to M is unidirectional?
P4, L112-113: As commented in major point 1, how did conduct accompanying simulations at each time step? The detailed description of each scenario should be explained.
P4, L113-115: However, even though each accompanying simulation conducted at each time step, the result is merely derived from the difference (subtraction) from baseline simulation. What was the advantages to embed these accompanying simulations? How about the computational burden? It was not clearly stated here. Therefore, I cannot follow the importance of QDA method as novel way.
P4, L119-120: In case of sulfate, this will be also produced in aqueous-phase oxidation pathways. How this process was incorporated?
P5, Section 2.2: The core mechanisms configured NAQPMS seems to be outdated over 20 years as stated in this section. Despite the recent progress of modeling components, I cannot follow “… has been widely used in scientific research and air quality prediction practice (Wang et al., 2014) due to its good performance in simulating the emission, meteorological and chemical processes in the atmosphere.”. Detailed introductions of research examples are required, because the modeling performance itself will be important to discuss this manuscript.
P5, L132: Typo in “ISORRPIA”.
P5, L142: What was this year? It should be defined first here.
P5, L143-144: What was the height of lowermost layer? It should be explicitly stated to consider the modeling performance at surface level.
P5, L145: Was the MEIC also targeted to the analyzed year?
P5, L147: What kind of biomass burning emissions was used? If not used, why?
P5, L149: Confirm the version of MOZART 2.4 or 2.5?
P5, L151: Again, WRF version 3.7 seems to be also outdated. What is the exact reason to use this version to generate meteorological field despite the authors’ claim of the importance of meteorology.
P5, L150-152: Does NAQPMS model online-coupled to WRF meteorological field? It was not clarified here.
P6, L171: Need the definition of LST. What is the difference from GMT?
P7. L193-194: Over China, recommendations of modeling standards have been updated (https://acp.copernicus.org/articles/21/2725/2021/), and it is better to use this criteria because this study targeted BTH region.
Figure 5 and 9: Without the explicit information of vertical layer height, this presentation is weak. Please clarify these information on Section 2.2 or 2.3.
Figure 8: The contribution of M and E terms are larger compared to other terms. I would like to recommend to use different scale for them, especially for (e)-(h). Again as I have commented as major comments of 1 and 2, this result impressed me that QDA was just the usage of IPR method. Please clarify this point in introduction and methodology.
Figure 10: Was the vertical axis used log-scale? It seems to be used unusually scaled axis.
Citation: https://doi.org/10.5194/gmd-2021-259-RC2 -
AC2: 'Reply on RC2', Baozhu Ge, 05 Dec 2021
The comment was uploaded in the form of a supplement: https://gmd.copernicus.org/preprints/gmd-2021-259/gmd-2021-259-AC2-supplement.pdf
-
AC3: 'Reply on RC2', Baozhu Ge, 05 Dec 2021
The comment was uploaded in the form of a supplement: https://gmd.copernicus.org/preprints/gmd-2021-259/gmd-2021-259-AC3-supplement.pdf
-
AC2: 'Reply on RC2', Baozhu Ge, 05 Dec 2021
Status: closed
-
RC1: 'Comment on gmd-2021-259', Anonymous Referee #1, 22 Sep 2021
General comments
The authors describe a method for determining the contributions of different processes to PM2.5 formation and the couplings between the processes. The authors call this method Quantitative Decoupling Analysis (QDA) and apply the method to a haze episode in the Beijing-Tianjin-Hebei region from 17-28 February 2014. The manuscript is generally well-written.
However, there are three significant problems with the work:
* The QDA method is not new. This is the Factor Separation method introduced in 1993 and applied in later work. See references below.
* The emissions are constant throughout the simulation period, which is not realistic. As a consequence, the contribution of emissions to the PM2.5 concentration change is constant throughout the episode, and all the time-variation in the factors and couplings is driven by the meteorology.
* The authors consider the influence of three factors on PM2.5, total emissions, chemistry, and meteorology, and indicate that their work provides valuable information to decision makers (lines 64-74). PM2.5 pollution episodes are driven by anthropogenic emissions and meteorology. Chemistry is a secondary factor that responds to emissions and meteorology but can be controlled by decision makers only by regulating the anthropogenic emissions. It would be much more relevant to decision makers if the authors had chosen biogenic emissions, anthropogenic emissions, and meteorology as the three factors. Then the full effect of anthropogenic emissions on PM2.5 during the episode would be apparent, rather than burying some of the effect in the chemistry factor. Also, the results would help determine if emergency anthropogenic emission controls during episodes would reduce PM2.5, which is a goal of the authors’ work (lines 71-72) but not a result of their work.
Specific comments
p. 3, lines 64-72. Decision makers can control anthropogenic emissions and possibly have a minor impact on some biogenic emissions (e.g., types of trees planted in urban areas). Understanding the impact of meteorology on atmospheric concentrations is also important and useful to decision makers. But separating out the impact of chemical reactions does not help regulators reduce atmospheric concentrations. The chemistry factor is controlled by emissions and meteorology, so some (not all) of the chemistry factor represents the impact of emissions. The decision makers need to understand the full impact of the emissions, but that cannot be obtained from the factors that the authors chose.
p. 4, lines 89-91. Eq. 2 is incorrect. There should be a factor of 2 in front of the cross terms Δx1Δx2, Δx2Δx3, and Δx1Δx3 in the group of second-order terms and other non-unity factors derived from the binomial coefficients in front of the cross terms in the group of third-order terms. The authors may not have used Eq. 2 and used only Eq. 3. However, if the authors actually used Eq. 2 in their calculations and analyses, they should verify that they used the correct equation, and, if not, the calculations and analyses must be re-done. In any case, Eq. 2 should be corrected.
p. 4, line 104. The interaction between emissions and meteorology is bi-directional. Higher temperatures increase evaporative emissions from gasoline vehicles, higher temperatures and greater sunlight increase isoprene emissions from plants, etc.
p.4, lines 106-112 and Table 1. There should be a more complete description of the simulations and what is different between simulation M1 and the other simulations. In particular for simulation M4, are the meteorological processes and emissions absent for the entire simulation? If so, what PM2.5 could there be in a grid cell other than the initial PM2.5 concentration, which is stationary in space because no meteorological processes are included? Is the PM2.5 concentration at the start of a specific time step taken from simulation M1, simulation M4 is run over that time step without including meteorology processes or emissions anywhere in the modeling domain, then a new PM2.5 concentration is obtained from simulation M1 for the next time step? What the authors did is very unclear.
p. 4, lines 111-114. Did the authors run the base simulation 6 times, each time with one of the “accompanying” simulations? That would effectively be 12 simulations. A simpler approach would seem to be running the base simulation once, recording the timesteps used, and then running each of the “accompanying” simulations once with the same timesteps used for the base case. That would reduce the number of simulations needed to 7.
p. 4, lines 119-120. There should be a detailed explanation of how IPR is applied to the results for each factor. It is unclear how this was done. Simulation M4 (C factor) does not contain emissions, so the chemistry will be different from that when emissions are present. Are the IPR results then meaningful for C?
pp. 3-5. The QDA method is not new. This is the Factor Separation method introduced by U. Stein and P. Alpert, Factor separation in numerical simulations, J. Atmos. Sci. 50, 2107-2115 (1993). Subsequently, Tao et al. applied the method to separate the contributions of area, mobile, and point source emissions to ozone and their interactions(Tao et al., Area, mobile, and point source contributions to ground level ozone: a summer simulation across the continental USA, Atmos. Environ. 39, 1869-1877 (2005)). The authors should not refer to QDA as a new method and should credit Stein and Alpert and Tao et al. by including their references in the manuscript.
p. 7, line 187. What is MBE? This is not defined in Table S1. MB is -13.7 μg/m3 and ME is 42.1 μg/m3 (Table S2) so it cannot be either of those two statistics.
p. 7, line 191. Again, what is MBE and where are these values (7.1 and 5.3 μg/m3 ) in Table S2? If the values are discussed in the manuscript, they should be in Table S2.
p. 7, line 194. There is a more recent paper (L. Huang et al., Atmos. Chem. Phys. 21, 2725-2743 (2021)) that gives goals and criteria specifically for PM 2.5 simulations in China.
p. 7, lines 210-212. For their analyses, the authors fixed the emissions to be constant in time. It is unclear why this is necessary for the method, and it is a serious limitation of their work. Neither the anthropogenic nor the biogenic emissions are constant in time; there are large variations over the diurnal cycle. As a consequence of the authors’ assumption of constant emissions, their calculated emission contribution is constant over all 12 days of the episode (Figures 7 and 8, Tables 3 and 4). This is not an interesting or very valuable result, especially for the decision makers/regulators. We cannot control the meteorology, only the anthropogenic emissions, so the important question is to what extent instituting greater emission controls during stagnation events will improve air quality. The authors’ results do not provide any insight on that question. Further, the assumption of constant emissions also influences the chemical contribution because time-varying emissions would very likely give much greater variation in the chemistry contribution.
p. 8, lines 234-239. These conclusions are well-known from many previous studies.
p. 9, line 276. It is unclear what the range of -0.86 to 1.86 represents. It is much wider than what the results in Table 3 suggest.
pp. 10-11, Section 3.4. Many of the conclusions here are well-known from prior work, and the Factor Separation method (QDA) adds little new information to the prior work. At most, this section shows consistency between the Factor Separation method and the results of previous studies, but there is no detailed evaluation of the Factor Separation method.
p. 10, lines 281-282. Yes, the results in the paper do not give much information about the importance of emissions and therefore are not of much use to decision makers.
pp. 11-12. Again, QDA is not a new method and most of the conclusions here are not new.
Technical corrections
p. 2, lines 54-55. “However, due to the nonrepeatability of individual pollution cases, … .” Not clear what is meant here. If one has an estimate of the meteorological fields from a weather model and an estimate of the emissions, the air quality model can estimate the atmospheric concentration of PM for days in different years (“individual cases”). Not clear why sensitivity experiments are necessary to “fully reproduce the individual cases.”
p. 2, line 57. Define the PLMA acronym.
p. 5, line 133. Should be “nitrate” not “nitrite”?
p. 5, line 149. Was it MOZART v 2.4 or v 2.5?
p. 6, line 159. “The” should be “the”
Figure 2 (b). The legend should be larger.
p. 6, lines 179-180. It would be clearer to use the same nomenclature for these statistics as in Table S1. Table S1 has the conventional names.
Figure 6 caption. There is no solid line in the figure, only 3 dashed lines. Do the points represent 24 hour averages? For which days?
Tables 3 and 4 should be in the Supporting Information because they repeat the information in Figure 8.
p. 9, line 259. Not clear why one limit is -3%. This seems to be a comparison of the magnitudes of the two quantities, in which case the limit would be +3%.
p. 10, line 293. Should it be “from stages 1 to 2” instead of :from stages 2 to 1”?
p. 11, line 311. Define the acronym PLAM.
p. 14, lines 436-437. The title of the paper should not be in all capitals.
Citation: https://doi.org/10.5194/gmd-2021-259-RC1 -
AC1: 'Reply on RC1', Baozhu Ge, 05 Dec 2021
The comment was uploaded in the form of a supplement: https://gmd.copernicus.org/preprints/gmd-2021-259/gmd-2021-259-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Baozhu Ge, 05 Dec 2021
-
RC2: 'Comment on gmd-2021-259', Anonymous Referee #2, 28 Sep 2021
Review comment on gmd-2021-259
General Comments:
The manuscript ‘A quantitative decoupling analysis (QDA v1.0) method for the assessment of meteorological, emission and chemical contributions to fine particulate pollution’ written by Junhua Wang presented the QDA method as novel way to evaluate meteorology, emission, and chemistry processes involved for the aerosol formation. Although the concept of this method is interesting, I cannot fully understand the description of method itself and therefore go through to result and discussion section well. At the current presentation quality, this manuscript cannot be considered for publication. At this round, I would like to reject this manuscript. Before considering the possible publication, I sincerely request the fundamental amendments. I wish the following major and minor comments will help to revise this manuscript.
Major Comments:
1. The description of QDA and its relation to IPR
The newly developed QDA method is just the using of six accompanying simulations to calculate M, E, and C terms. In this sense, for example, to drive M term, this method seems to be identical to the SAA as described in the introduction. Actually, how to conduct six accompanying simulations is unclear. Under each time-step simulation, how can do the base-model derive each process? The detailed description of M2-M7 is required to understand the QDA method. In addition, without E term, C term cannot be driven due to the absence of precursors. Therefore I guess that EC term inherently connected and it could be hard to be divided. Moreover, on P4, L118, it was stated that “The above QDA method can also be combine with the IPR method to resolve more detailed information…”. This statement is confusing to me because this impressed that QDA is just the using of IPR. Under the current presentation quality, it is difficult to understand QDA method and I cannot recognize this method as novel way in modeling analysis.
2. Results and discussion of QDA.
Because the description of QDA is insufficient, I also cannot follow the result and discussion section. Why E term showed same values through analyzed stages? Is this because emissions did not consider temporal variation through analyzed episode? The meteorological field are shown in Fig. 4, but how about the precipitation? Because the term of “wetdep” was 0.00 through stages, I felt that there was no rain. Although this was the severe haze event, without the wet deposition analysis, this episode seems to be not interesting as test case to show the QDA result. As evaluated using NOR and SOR, I like the idea to consider the formation process from the viewpoint of each specie. The result of QDA is now discussed for PM2.5; however, each specie have been evolved as different E and C terms. I would like to strongly recommend to show the same kind of analysis of Figs. 7-9 for each specie. This analysis will offer the insight into C roles on chemical formation during haze episode.
3. The application of QDA.
As found in the abstract, this QDA method could help modelers to understand the each process and find these uncertainties. I have briefly checked the source code of QDA distributed in ZENODO, but I felt that the fortran90 codes seems to be incorporated into the NAQPMS source codes. How can we apply this source code into other models? If the authors claimed that “QDA is a universal tool”, the explanation for how to use this QDA method in other models codes should be kindly introduced within this distribution.
Minor Comments:
P2, L42: CMAQ have to be introduced after the definition of CTM (P2, L56). The organization of introduction for second and third paragraphs should be reconsidered.
P2, L41-46: Under this context, IRR should be also carefully introduced. The IRR can be used to define the role of reaction rate, and this will relate C term in this study.
P4, L92-L99 and Eq. 2: How can we treat the second- and third-order partial differential of x1. x2, and x3? Does this represent the nonlinear term of M, E, and C? What stands for them?
P4, L104: For example, higher temperature will relate activated plant, and change the biogenic emissions intensity. Why E to M is unidirectional?
P4, L112-113: As commented in major point 1, how did conduct accompanying simulations at each time step? The detailed description of each scenario should be explained.
P4, L113-115: However, even though each accompanying simulation conducted at each time step, the result is merely derived from the difference (subtraction) from baseline simulation. What was the advantages to embed these accompanying simulations? How about the computational burden? It was not clearly stated here. Therefore, I cannot follow the importance of QDA method as novel way.
P4, L119-120: In case of sulfate, this will be also produced in aqueous-phase oxidation pathways. How this process was incorporated?
P5, Section 2.2: The core mechanisms configured NAQPMS seems to be outdated over 20 years as stated in this section. Despite the recent progress of modeling components, I cannot follow “… has been widely used in scientific research and air quality prediction practice (Wang et al., 2014) due to its good performance in simulating the emission, meteorological and chemical processes in the atmosphere.”. Detailed introductions of research examples are required, because the modeling performance itself will be important to discuss this manuscript.
P5, L132: Typo in “ISORRPIA”.
P5, L142: What was this year? It should be defined first here.
P5, L143-144: What was the height of lowermost layer? It should be explicitly stated to consider the modeling performance at surface level.
P5, L145: Was the MEIC also targeted to the analyzed year?
P5, L147: What kind of biomass burning emissions was used? If not used, why?
P5, L149: Confirm the version of MOZART 2.4 or 2.5?
P5, L151: Again, WRF version 3.7 seems to be also outdated. What is the exact reason to use this version to generate meteorological field despite the authors’ claim of the importance of meteorology.
P5, L150-152: Does NAQPMS model online-coupled to WRF meteorological field? It was not clarified here.
P6, L171: Need the definition of LST. What is the difference from GMT?
P7. L193-194: Over China, recommendations of modeling standards have been updated (https://acp.copernicus.org/articles/21/2725/2021/), and it is better to use this criteria because this study targeted BTH region.
Figure 5 and 9: Without the explicit information of vertical layer height, this presentation is weak. Please clarify these information on Section 2.2 or 2.3.
Figure 8: The contribution of M and E terms are larger compared to other terms. I would like to recommend to use different scale for them, especially for (e)-(h). Again as I have commented as major comments of 1 and 2, this result impressed me that QDA was just the usage of IPR method. Please clarify this point in introduction and methodology.
Figure 10: Was the vertical axis used log-scale? It seems to be used unusually scaled axis.
Citation: https://doi.org/10.5194/gmd-2021-259-RC2 -
AC2: 'Reply on RC2', Baozhu Ge, 05 Dec 2021
The comment was uploaded in the form of a supplement: https://gmd.copernicus.org/preprints/gmd-2021-259/gmd-2021-259-AC2-supplement.pdf
-
AC3: 'Reply on RC2', Baozhu Ge, 05 Dec 2021
The comment was uploaded in the form of a supplement: https://gmd.copernicus.org/preprints/gmd-2021-259/gmd-2021-259-AC3-supplement.pdf
-
AC2: 'Reply on RC2', Baozhu Ge, 05 Dec 2021
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
1,526 | 274 | 64 | 1,864 | 111 | 39 | 41 |
- HTML: 1,526
- PDF: 274
- XML: 64
- Total: 1,864
- Supplement: 111
- BibTeX: 39
- EndNote: 41
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1