the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Impacts of land cover changes on biogenic emission and its contribution to ozone and secondary organic aerosol in China
Jinlong Ma
Shengqiang Zhu
Siyu Wang
Jianmin Chen
Abstract. The greening impacts in China led to an increase in natural emissions, which may further aggravate the local burden of secondary organic aerosol and ozone. Thus, a robust natural emissions inventory is crucial for assessing the contribution of biogenic volatile organic compounds (BVOCs) to air quality. However, using different satellites will introduce uncertainties in BVOC estimations. Quantitative analysis is needed for evaluating the accuracy of BVOC estimations and their impacts on air quality. In this study, Model of Emissions of Gases and Aerosols from Nature (MEGAN) v2.1 was used to investigate the impact of different LAI and LC datasets on BVOC emissions in China in 2016 and impacts on O3 and SOA were evaluated based on the Community Multiscale Air Quality Modelling System (CMAQ). Three LAI satellite datasets of the Global LAnd Surface Satellite (GLASS), the Moderate Resolution Imaging Spectroradiometer (MODIS) MOD15A2H version 6 (MOD15), and the Copernicus Global Land Service (CGLS), as well as three LC satellite datasets of the MODIS MCD12Q1 LC products, the Copernicus Climate Change Service (C3S) LC products, and the CGLS LC products were used in five parallel experiments (cases: C1-C5). Results show that BVOC emissions in China range from 25.42 to 37.39 Tg in 2016 and are mainly concentrated in central and south-eastern China, with the highest in C5 (using GLASS and CGLS LC) and the least in C4 (using GLASS and C3S LC). According to model validations, C4 is the better choice for estimating BVOC emissions in China compared to other scenarios. In addition, the use of different LAI and LC inputs has a further impact on the concentrations of O3 and SOA, especially in central and eastern China, where the differences are up to 4.8–6.9 ppb and 5.3–8.4 μg m-3 in O3 and biogenic SOA (BSOA), respectively. Due to the summer monsoon, a large amount of O3 is also transported from southern China to the Yangtze River Delta (YRD) and North China Plain (NCP), resulting in a large difference in O3 concentrations between different cases. For example, the O3 in C1 (using GLASS and MCD12Q1) is 3.6 ppb higher than in C4 in the YRD. Furthermore, isoprene contributes most to BSOA concentrations in China, which is 1-fold higher than monoterpenes and 1.5-fold higher than sesquiterpenes. The BSOA formed by isoprene is the main reason resulting in the difference between C1 and C5.
- Preprint
(2104 KB) -
Supplement
(1675 KB) - BibTeX
- EndNote
Jinlong Ma et al.
Status: closed
-
RC1: 'Comment on acp-2022-739', Anonymous Referee #1, 05 Dec 2022
General comments:
This paper explores the impact of different land cover data sets on biogenic emissions and its contribution to ozone and secondary organic aerosol in China with the Model of Emissions of Gases and Aerosols from Nature (MEGAN) v2.1 employing a further developed version of the Community Multiscale Air Quality Modelling System (CMAQ). For that, a set of experiments with three different leaf area index (LAI) and plant functional type (PFT) input were conducted to show the impact of the different input to BVOC emissions, extreme ozone and SOA. The BVOC emissions estimated from all simulations are within the reported range of literature values whereas, however, only the changes of LC data lead to a significant impact on BVOC emissions and related air quality.
This investigation will be valuable, also on a global scale to assess further the uncertainty of the input data to BVOC emissions and their products and finally improve the model representation of air quality. Besides the following suggestions, I advise the authors to improve the manuscript regarding the language and the structure, adapt the visualization more to the analysis purpose and strengthen the conclusion.
Specific comments:
Abstract: In particular, the beginning is a bit unstructured. Some sentences don’t have enough content to understand the link between different aspects see the following points. If you structure it more following a red line it may also lead to more comprehension of the text.
- Time period of ‘greening impacts in China’? Which role does this play for BVOC emissions? How does BVOCs act for O3 and SOA? Why are these important species to study? – air quality/health impact. How is the BVOC estimation based on inventories, and how does this link to satellites? If you want to mention the uncertainty of satellites include a relative estimate.
- A more general formulation of the results would make the abstract more appealing. Also, including relative changes helps the reader to judge the impact order of your changes.
Introduction: It’s confusing that some sentences are lacking scientific background and with some you go in very detail as can be seen in the following points. In particular the last paragraph could be more conclusive by linking the mentioned studies more together:
- ‘wide range of warming and cooling climate pollutants’ (l.39) is to unspecific. The different radiative forcing/role in climate between aerosols and greenhouse gases like O3 should be mentioned
- (l.51) The role of emission and activity factors for BVOC emissions need to be explained before. In general, I suggest to describe in short how BVOC emissions are calculated by the model
- (l. 69) which CTM was used in the cited study?
- - (l. 71 f.) Why is the particular SOA concentration in Sichuan mentioned here? Is this a high number?
Section 2: I suggest to re-structure this section (Model description: CMAQ, MEGAN, WARF; Data description, experiments) to improve readability. Along this, I have some question:
- (Line 100/101) According to which maps are the PFT classified?
- (l.112) To which LAI maps does GLASS show better consistency?
- 140-153 belongs to ‘Model description’
- 150 f. Difficult sentence structure. Is MEIC based on EDGAR?
- 147: The chemistry does not impact the meteorology, right?
Section 3: The statistical analysis is useful, but in the later analysis I often miss the conclusion/interpretation of the simulation results instead the authors often only describe the figures. What can we learn from the data/results?
- 158: criteria -> uncertainty? How is the uncertainty calculated?
- 170/171: The day-to-day variability in Fig S2 are hard to see. Think about another visualization when you want to show this.
- 172 f.: The MB and he GE of the T2 exceeds the 'benchmark', you have to explain this, and why the model performance is still acceptable. Temperature is a crucial driver of BVOC emissions.
- 179: What is your motivation to investigate both, MDA1O3 AND MDA8O3?
- 181: What are cut-off concentrations?
- 199: I wouldn’t say that air quality is not correlated to LAI input only because of this result. Not only BVOC emissions rely on LAI input (also e.g. dry deposition)
- 214: definition or reference for LAIv=2
- 215: In which figure can this be seen?
- 220 and before: Are these findings in agreement with other studies?
- 221: seasonal variations are not easy to see in Fig. 3. I suggest to show/add the annual cycle
- 224-226: the two sentences are contradicting.
- 225 ‘seasonal change trend’. What do you mean? You do not consider multiple years
- 228: Is this in agreement with other findings?
- 230/31: Are the different species emitted by different PFTs? (Could also be a reason for the difference). What is the temperature coefficient?
- 233/234: Why have the monoterpenes only little sensitivity to LAI sensitivity
- 240 and before: I suggest to show the emission change per each PFT which would make it easier to interpret and draw general conclusions
- 267: For which year did the studies, which are referred here, estimate the emissions?
- Table 4: It would make more sense to compare the emission estimates from approximately the same time period. Also, adding one more column in the table with the average LAI value would improve the clarity or alternatively list the BVOC emission estimates normalized by the LAI. Why is the estimate for 2018, which is quite close to your study period, almost double?
- Line 286: How do you estimate the BVOC emissions only formed from O3? (Information could be added to the Methods section
- 6: The seasonal variation is hardly visible and might be better displayed by seasonal cycles (with uncertainty bars)
- 327: Why don't you use the same unit as in the Fig. 7?
- 352/53: Do you want to say that changing the MEGAN inputs has a large impact on isoprene emissions and since isoprene is the main contributor to BVOCs this also impact BVOC emissions?
- 364: Why is summer BSOA in C1 2.5 times higher than in C4?
- From Fig. 4 you found that BSOA changed for different LAI as well as for different PFT inputs although the BVOC emissions (analyzed before) don’t change much with different LAI. Why is that?
Conclusion: l.370: How do you judge that C4 was better for BVOC emissions? You haven't done an evaluation of the BVOC emissions, or?
- 372: is this consistent with other studies?
- 380: name the concrete impact (e.g. relative changes)
Technical corrections:
- Abstract: introduce abbreviations LAL and LC
- 49: as in PRD …
- Grammar and tense issues: l.52/53: ‘quantifi(ed)/determine(d)’, l. 55 ‘affect(ing)’, l.99&l.111 &l.145&l.122 ‘were’ -> are
- 104 ‘major(ity)’
- Sentence in line 62 f. (‘Wang et al. (2018a) misses one verb
- Sentence in l. 71 has a complicated sentence structure, integrate the last part in the main sentence
- -(l.100) ‘of each LC type […] and [this]’
- 158 ‘based on five different cases’ -> with 5 different sets of LC input data for MEGAN
- Footnote in Table S3 not found, in caption: ‘ The values are bolded without meeting the benchmarks’ -> ‘The values without meeting the benchmarks are bolded’
- S3: scale could be improved
- Table S5: units are not given in the caption, information on benchmarks in the table, distinction between MDO3 and MDA8 could be clearer
- 252: are occurred -> occur
- 260: shouldn’t be the sentence refer to Fig. S4 ?
- 344/45: the plots from the three different species are hard to compare since they have different scales
- 347: Reformulate
- 350: attribute -> be attributed
- 355: delete ‘amount of’
- 359: ‘temporal trend’ -> seasonal cycle ?
- 369: sometimes you mean both LAI and PFT when you write LC datasets
- 374-76: More comprehensive formulation needed for the conclusion
Citation: https://doi.org/10.5194/acp-2022-739-RC1 -
AC1: 'Reply on RC1', Hongliang Zhang, 21 Feb 2023
Dear Referee #1,
We appreciate your comments to help improve the manuscript. We tried our best to address your comments and detailed responses and related changes are shown below and in the attached PDF file.
Comments: This paper explores the impact of different land cover data sets on biogenic emissions and its contribution to ozone and secondary organic aerosol in China with the Model of Emissions of Gases and Aerosols from Nature (MEGAN) v2.1 employing a further developed version of the Community Multiscale Air Quality Modelling System (CMAQ). For that, a set of experiments with three different leaf area index (LAI) and plant functional type (PFT) input were conducted to show the impact of the different input to BVOC emissions, extreme ozone and SOA. The BVOC emissions estimated from all simulations are within the reported range of literature values whereas, however, only the changes of LC data lead to a significant impact on BVOC emissions and related air quality.
This investigation will be valuable, also on a global scale to assess further the uncertainty of the input data to BVOC emissions and their products and finally improve the model representation of air quality. Besides the following suggestions, I advise the authors to improve the manuscript regarding the language and the structure, adapt the visualization more to the analysis purpose and strengthen the conclusion.
Response: Thanks for the recognition of our study. Below is the response to each specific comment.
Comments:
- Abstract: In particular, the beginning is a bit unstructured. Some sentences don’t have enough content to understand the link between different aspects see the following points. If you structure it more following a red line it may also lead to more comprehension of the text.
1.1 Time period of ‘greening impacts in China’? Which role does this play for BVOC emissions? How does BVOCs act for O3 and SOA? Why are these important species to study? – air quality/health impact. How is the BVOC estimation based on inventories, and how does this link to satellites? If you want to mention the uncertainty of satellites include a relative estimate.
Response: Thanks for pointing out these important issues. The period of ‘greening impacts in China’ is from 2000 to 2017 (Chen et al., 2019). The greening impacts increase the vegetated area, further increasing the biogenic volatile organic compounds (BVOC) emissions. Since BVOCs are the important precursor for ozone (O3) and secondary organic aerosol (SOA), the increase in BVOC emissions further changes concentrations of O3 and SOA. In this study, the BVOC emissions are estimated by the Model of Emissions of Gases and Aerosols from Nature (MEGAN) and the satellite datasets are necessary inputs for the MEGAN model. The sentences in the abstract were modified and shown below.
Changes in manuscript:
Abstract (Lines 11-14 in the revision): The greening impacts in China from 2000 to 2017 led to an increase in vegetated areas and thus enhanced biogenic volatile organic compounds (BVOC) emissions. BVOCs are regarded as important precursors for ozone (O3) and secondary organic aerosol (SOA). As a result, accurate estimation of BVOC emissions is critical to understanding their impacts on air quality.
1.2 A more general formulation of the results would make the abstract more appealing. Also, including relative changes helps the reader to judge the impact order of your changes.
Response: Thanks for your suggestions. The results were modified to be more general and the differences between cases were presented with the relative difference in abstract.
Changes in manuscript:
Abstract (Lines 23-31 in the revision): Changing the LC inputs for the MEGAN model has a more significant difference in BVOC estimates than using different LAI datasets. The C4 case has better model performance, indicating that it is the better choice for BVOC estimations in China. Changing the MEGAN inputs further impacts the concentrations of O3 and SOA. The highest O3 and biogenic SOA (BSOA) concentrations appear in the C1 (using GLASS and MCD12Q1 LC) simulation, which can reach 12 ppb and 9.8 μg m-3, respectively. Due to the combined effect of local BVOC emissions and the summer monsoon, the relative difference between C1 and C4 is over 52% and 140% in O3 and BSOA in central and eastern China. The BSOA difference between C1 and C4 is mainly attributed to the isoprene SOA (ISOA), which is a major contributor to BSOA. Particularly, the relative difference in ISOA between these two cases is up to 160% in eastern China. Therefore, our results suggest that the uncertainties in MEGAN inputs should be fully considered in future O3 and SOA simulations.
- Introduction: It’s confusing that some sentences are lacking scientific background and with some you go in very detail as can be seen in the following points. In particular the last paragraph could be more conclusive by linking the mentioned studies more together:
2.1 ‘wide range of warming and cooling climate pollutants’ (l.39) is to unspecific. The different radiative forcing/role in climate between aerosols and greenhouse gases like O3 should be mentioned.
Response: The pollutants and their effects on climate were determined in the manuscript and listed below.
Changes in manuscript:
Introduction (Lines 38-41 in the revision): In addition, changes in emissions of BVOCs will alter the capacity of a wide range of warming and cooling climate pollutants, such as O3, methane (CH4) and aerosols. O3 and CH4 can warm the climate, while the aerosols have a cooling effect by scattering solar radiation (Unger, 2014b, a).
2.2 (l.51) The role of emission and activity factors for BVOC emissions need to be explained before. In general, I suggest to describe in short how BVOC emissions are calculated by the model.
Response: Thanks for your suggestion. The calculation of BVOC emission was added in the manuscript and explained below.
Changes in manuscript:
Introduction (Lines 49-51 in the revision): The model determined the vegetation types according to model inputs and then use the activity factor multiplied with the emission factor to calculate emissions for each vegetation type (Guenther et al., 2012).
2.3 (l. 69) which CTM was used in the cited study?
Response: The specific name for the chemical transport model (CTM) of the cited study was written in the manuscript and shown below.
Changes in manuscript:
Introduction (Lines 69-72 in the revision): Fu and Liao (2012) used the Goddard Earth Observing System chemical transport model (GEOS-Chem) to quantitate the impact of biogenic emissions on O3 in China over the year 2001-2006 and found that the difference in O3 concentrations induced by interannual variability of BVOCs could be 2-5%.
2.4 (l. 71 f.) Why is the particular SOA concentration in Sichuan mentioned here? Is this a high number?
Response: Yes, it is. The sentence was modified to clear its point in the manuscript and shown below.
Changes in manuscript:
Introduction (Lines 76-77 in the revision): Qin et al. (2018) investigated the biogenic SOA (BSOA) during summertime in 2012 and found that a high level of BSOA concentration appeared in Sichuan Basin.
- Section 2: I suggest to re-structure this section (Model description: CMAQ, MEGAN, WARF; Data description, experiments) to improve readability. Along this, I have some question:
Response: Thanks for your suggestions. The structure of the section ‘Methodology’ was changed from Data description, Model description, and Model application to Model setup, and Data description.
3.1 (Line 100/101) According to which maps are the PFT classified?
Response:
The plant function type (PFT) scheme used in the MEGAN is classified based on Community Land Model v4.0 (CLM4) (Guenther et al., 2012). The sentence was modified and Figure R1 (named as Figure S2 in the revision) was added to the manuscript to make it clear. Detail see attachment.
Changes in manuscript:
Methodology (Lines 142-145 in the revision): PFTs used in the MEGAN model adopt the scheme used for Community Land Model v4.0 (CLM4) (Guenther et al., 2012). Three LC maps are first re-gridded to the CMAQ domain (Fig. S2). Secondly, LC types are categorized into eight vegetation types according to legend descriptions of LC maps. Lastly, eight vegetation types are further reclassified into CLM-15 PFTs based on the climate rules described in Bonan et al. (2002).
Changes in supplementary material:
Figure R1 see attachment.
3.2 (l.112) To which LAI maps does GLASS show better consistency?
Response: In leaf area index (LAI) satellite products, the Global LAnd Surface Satellite (GLASS) products show better consistency than the MODIS MOD15A2H version 6 (MOD15) products according to the study in Xiao et al. (2016), while the accuracy of the Copernicus Global Land Service (CGLS) products are slightly less than MOD15 products (Fuster et al., 2020). Therefore, GLASS products have a better consistency than MOD15 products and CGLS products.
Changes in manuscript: No changes were made for this comment.
3.3 140-153 belongs to ‘Model description’
Response: Thanks for your comments. The subheading ‘Model description’ was changed to ‘Model setup’ and sentences in line140-153 were moved to ‘Model setup’.
3.4 150 f. Difficult sentence structure. Is MEIC based on EDGAR?
Response: No, it is not. The modelling domain in CMAQ was 36 km × 36 km in horizontal spatial resolution, which covers China and its surrounding countries in East Asia (Fig. S1). Since Multiresolution Emission Inventory for China (MEIC) only provided anthropogenic emissions for China, while anthropogenic emissions for regions excluding China were provided by the Emissions Database for Global Atmospheric Research (EDGAR) v4.3. The sentence was modified to make it clear and shown below.
Changes in manuscript:
Methodology (Lines 127-134 in the revision): The anthropogenic emissions of China used the datasets from Multiresolution Emission Inventory for China (MEIC; available at http://www. meicmodel.org, last access:3 May 2022). Since the MEIC only provides anthropogenic emissions for China, anthropogenic emissions from foreign countries were provided by the Emissions Database for Global Atmospheric Research (EDGAR) v4.3 (available at http://edgar.jrc.ec.europa.eu/overview.php? v=_431, last access: 10 May 2022). The MEIC inventory is widely used in air quality studies in China (Li et al., 2017;Hu et al., 2016;Wu et al., 2020). It had an improvement in a vehicle emission inventory with high resolution (Zheng et al., 2014), and a non-methane VOC mapping approach for different chemical mechanisms (Li et al., 2014). The EDGAR is a grided emissions inventory with a high horizontal resolution of 0.1°×0.1° (Saikawa et al., 2017).
3.5 147: The chemistry does not impact the meteorology, right?
Response: The chemistry does not impact the meteorology. We made it clear in the manuscript that chemistry has no effect on meteorology.
Changes in manuscript:
Methodology (Lines 171 in the revision): The model chemistry has no effect on meteorological conditions when simulating.
- 4. Section 3: The statistical analysis is useful, but in the later analysis I often miss the conclusion/interpretation of the simulation results instead the authors often only describe the figures. What can we learn from the data/results?
Response: In this study, we concluded that changing the MEGAN inputs has an impact on BVOC estimates, and this further influences the formation of O3 and secondary organic aerosol (SOA). Besides, changing land cover datasets for the model shows more conspicuous differences in BVOC emissions than using different LAI datasets.
4.1 158: criteria -> uncertainty? How is the uncertainty calculated?
Response: The criterion is a way of evaluating the model performance and is obtained based on the observations. Simulations with different parameter schemes will generate uncertainties that can lead to bias between predictions and observations, hence, the need to use the criterion to measure model performance. The criteria of 2 is a benchmark value for wind speed. The criteria for other meteorological conditions and calculation methods are described by Emery et al. (2001).
Changes in manuscript: No changes were made for this comment.
4.2 170/171: The day-to-day variability in Fig S2 are hard to see. Think about another visualization when you want to show this.
Response: Thanks for your advice. We changed the line chart of day-to-day variability for 2 m temperature into the scatter chart, which is Figure R2 (named as Figure S3 in the revision). Detail see attachment.
Changes in supplementary material:
Figure R2 see attachment.
4.3 172 f.: The MB and he GE of the T2 exceeds the 'benchmark', you have to explain this, and why the model performance is still acceptable. Temperature is a crucial driver of BVOC emissions.
Response: The result of the mean bias and the gross error of temperature at 2 m (T2) exceeding the benchmark is possible due to the overestimation of cloud coverage in the WRF model leading to an underestimated T2 (Wu et al., 2020). These biases of T2 are relatively small compared with previous studies with a yearly long WRF simulation in China (Wu et al., 2020;Wang et al., 2018a). The reasons were explained in the manuscript and shown below.
Changes in manuscript:
Results and discussion (Lines 178-181 in the revision): It is possible due to the overestimation of cloud coverage in the WRF model leading to an underestimated T2 (Wu et al., 2020). These biases are relatively small compared with previous studies with a yearly long WRF simulation in China (Wu et al., 2020;Wang et al., 2018a).
4.4 179: What is your motivation to investigate both, MDA1O3 AND MDA8O3?
Response: The maximum daily averaged 1h (MDA1) O3 and maximum daily averaged 8h (MDA8) O3 are different indicators for O3 in the National Ambient Air Quality Standard (NAAQS) proposed by EPA (1997, 2015). They are used to represent concentrations for the short-term (1 to 3 hours) and the prolonged (6 to 8 hours) exposures to O3 according to the NAAQS. They differ spatially and temporally due to differences in calculation methods (Bell and Ellis, 2003). China, the study area in this study, also uses them to evaluate the O3 pollution (MEP, 2012). Therefore, they are both important and need to be investigated.
Changes in manuscript: No changes were made for this comment.
4.5 181: What are cut-off concentrations?
Response: The cut-off concentration is a minimum threshold value used to restrict observations to mitigate the bias of the model performance analysis (Emery et al., 2017). The 60 ppb as the threshold for O3 is suggested by EPA (2007) to focus bias and error statistics on times of higher O3.
Changes in manuscript: No changes were made for this comment.
4.6 199: I wouldn’t say that air quality is not correlated to LAI input only because of this result. Not only BVOC emissions rely on LAI input (also e.g. dry deposition)
Response: The dry deposition is a process modelled by the Community Multiscale Air Quality Modelling System (CMAQ). Although this process will be influenced by the satellite-derived land use and vegetation products (Pleim and Ran, 2011), these parameters provided by Weather Research and Forecasting model (WRF) remain unchanged in this study. According to the scope of this paper, only inputs for the Model of Emissions of Gases and Aerosols from Nature (MEGAN) were changed, which did not have impacts on the WRF model, and thus the dry deposition would not be influenced. We modified the sentence to make it appropriate.
Changes in manuscript:
Results and discussion (Lines 211-213 in the revision): Although C1, C2, and C3 adopt LAI satellite products with different accuracies, the accuracies of these products have no significant impact on the model performance due to similar statistics values.
4.7 214: definition or reference for LAIv=2
Response: LAIv means the average LAI for vegetated areas, which is estimated by dividing the grid average LAI by the fraction of the grid that is covered by vegetation (Guenther et al., 2006). The related sentence was modified.
Changes in manuscript:
Results and discussion (Lines 224-225 in the revision): The BVOC emissions in C1 are about 5 Tg higher than C2.
4.8 215: In which figure can this be seen?
Response: The remarkable difference in isoprene between C1 and C2 can be seen in Table 2. The related sentences were modified.
Changes in manuscript:
Results and discussion (Lines 224-225 in the revision): The BVOC emissions in C1 are about 5 Tg higher than C2.
4.9 220 and before: Are these findings in agreement with other studies?
Response: Yes, these findings agreed with the study of Wang et al. (2018). Wang et al. (2018) used three different land cover datasets to run the MEGAN model, which were the Finer Resolution Observation and Monitoring of Global Land Cover (E1), Moderate-Resolution Imaging Spectroradiometer (MODIS)
MCD12Q1 PFT products (E4), and the Climate Change Initiative Land Cover (CCI LC) products (E5) in his study. The higher fractions of broadleaf tree in E4 resulted in higher isoprene emissions than E1 and E5, but the lower fractions of needleleaf tree and shrub in E4 resulted in lower monoterpene and sesquiterpene emissions than E1. We added the reference to the sentence and showed below. Figure R3 (named as Figure S3 in the revision) was added to the supplementary material. Detail see attachment.
Changes in manuscript:
Results and discussion (Lines 227-230 in the revision): Although the total BVOC emissions in C5 are 1.29 Tg higher than those in C1, the emissions of monoterpenes, sesquiterpenes, and other VOCs are lower than in C1. This is induced by the discrepancy in the distribution of needleleaf tree and shrubs between C1 and C5, which is in agreement with the result in Wang et al. (2018) (Fig. S3 and Fig. S5).
Changes in supplementary material:
Figure R3 see attachment.
4.10 221: seasonal variations are not easy to see in Fig. 3. I suggest to show/add the annual cycle
Response: Thanks for your suggestion. We changed the Figure R4 (named as Figure 3 in the revision) from the histogram to the stacked column chart to represent the annual emissions of BVOC as shown below. As shown in the Figure R4, the BVOC emissions are mainly concentrated in summer, accounting for 60.9%~63.8% of total emissions compared to 2.9%~3.4% in winter. Detail see attachment.
Changes in manuscript:
Figure R4 see attachment.
4.11 224-226: the two sentences are contradicting.
Response: These two sentences were corrected in manuscript and shown below.
Changes in manuscript:
Results and discussion (Lines 233-235 in the revision): In general, using different LAI and LC products does have an impact on the temporal variability in the BVOC emissions. The BVOC emissions show similar seasonal variations in all cases, which are mainly concentrated in summer, accounting for 60.9%-63.8% of total BVOC emissions compared to 2.9%~3.4% in winter
4.12 225 ‘seasonal change trend’. What do you mean? You do not consider multiple years
Response: We wanted to express that the different cases showed similar seasonal change characteristics. The expression was corrected and shown below.
Changes in manuscript:
Results and discussion (Lines 233-235 in the revision): In general, using different LAI and LC products does have an impact on the temporal variability in the BVOC emissions. The BVOC emissions show similar seasonal variations in all cases, which are mainly concentrated in summer, accounting for 60.9%-63.8% of total BVOC emissions compared to 2.9%~3.4% in winter
4.13 228: Is this in agreement with other findings?
Response: Yes, it is. In the study of Ibrahim et al. (2010), the sesquiterpene emitted from birch and aspen increased faster than monoterpene with temperature rising. And Bai et al. (2015) concluded that the increasing rates of isoprene with temperature were much higher than for monoterpenes. These references were added to the sentence.
Changes in manuscript:
Results and discussion (Lines 238-241 in the revision): The percentage of winter monoterpenes in the total monoterpenes is higher than that of isoprene and sesquiterpenes, probably because isoprene and sesquiterpenes are more sensitive to temperature changes than monoterpenes (Ibrahim et al., 2010;Bai et al., 2015).
4.14 230/31: Are the different species emitted by different PFTs? (Could also be a reason for the difference). What is the temperature coefficient?
Response: No, they aren’t. In the MEGAN model, different PFTs can emit all BVOC species with different emission factors. The emission ( ) of chemical species i from vegetation type j according to
Where is the emission factor at standard conditions for different vegetation type j with fractional grid box areal coverage .The emission activity factor ( ) accounts for the processes controlling emission responses to environmental and phenological conditions (Guenther et al., 2012). The temperature coefficient is an empirical coefficient used to determine the light-independent fraction, and the light-independent is a part of the temperature activity factor in the MEGAN model (Guenther et al., 1993;Guenther et al., 2012;Helmig et al., 2006).
Changes in manuscript: No changes were made for this comment.
4.15 233/234: Why have the monoterpenes only little sensitivity to LAI sensitivity
Response: We made a mistake here and the sentence was deleted.
Changes in manuscript:
Results and discussion (Lines 241 in the revision): C4 used the C3S LC and GLASS shows the lowest emissions in total BVOCs and its main species among each season.
4.16 240 and before: I suggest to show the emission change per each PFT which would make it easier to interpret and draw general conclusions
Response: Thanks for your suggestion. The figure about the BVOC emissions of main PFT was added in the supplementary material (Figure R3, named as Figure S3 in the revision). In Figure R3, the highest fraction of broadleaf trees in C5 contributing large BVOC emissions. Although grass areas are higher than broadleaf tree areas, it does not contribute significantly to the BVOC emissions due to its lower the emission factor. Detail see attachment.
Changes in supplementary material:
Figure R3 see attachment.
4.17 Table 4: It would make more sense to compare the emission estimates from approximately the same time period. Also, adding one more column in the table with the average LAI value would improve the clarity or alternatively list the BVOC emission estimates normalized by the LAI. Why is the estimate for 2018, which is quite close to your study period, almost double?
Response: Many thanks for your comments. We changed the study comparison with Li et al. (2013) to a comparison with Li et al. (2020). This column about the LAI could not be added because the LAI data from previous studies were not available. The reason for the last question is that Li et al. (2020) produced the basal emission rates for 192 plant species and categorized them into 82 PFTs for China resulting in more BVOC estimates.
Changes in manuscript:
Results and discussion (Lines 280-283 in the revision): There is a considerable difference in BVOC emissions between this study and those of Li et al. (2020). The difference is mainly due to the combined effect of emission rate and PFTs. Liu et al. (2020) produced the basal emission rates for 192 plant species and categorized them into 82 PFTs for China resulting in more BVOC estimates.
4.18 Line 286: How do you estimate the BVOC emissions only formed from O3? (Information could be added to the Methods section
Response: The concentrations of O3 from different VOC sources (henceforth ) were determined by the source-oriented method (Ying and Krishnan, 2010). Based on the method, the non-reactive O3 tracer is used to track O3 attributed to BVOCs, which is tagged as and directly predicted. The descriptions of O3 source apportionment see detailed in Wang et al. (2019).
Changes in manuscript:
Methodology (Lines 105-108 in the revision): The concentrations of O3 from different VOC sources (henceforth ) were determined by the source-oriented method (Ying and Krishnan, 2010). Based on the method, the non-reactive O3 tracer is used to track O3 attributed to BVOCs, which is tagged as and directly predicted. The descriptions of O3 source apportionment see detailed in Wang et al. (2019).
4.19 6: The seasonal variation is hardly visible and might be better displayed by seasonal cycles (with uncertainty bars)
Response: Thanks for your suggestion. The seasonal histogram charts contain information not only on seasonal changes but seasonal differences between cases. Therefore, using the seasonal cycles will not get any better.
Changes in manuscript: No changes were made for this comment.
4.20 327: Why don't you use the same unit as in the Fig. 7?
Response: Ppb and μg m-3 are both the commonly used units (Wang et al., 2021;Zhang et al., 2021;Hu et al., 2016).
Changes in manuscript: No changes were made for this comment.
4.21 352/53: Do you want to say that changing the MEGAN inputs has a large impact on isoprene emissions and since isoprene is the main contributor to BVOCs this also impact BVOC emissions?
Response: We want to express that changing the MEGAN inputs has a large impact on isoprene emissions, which are the main contributor to BVOCs. And this impact further changes the concentration of secondary organic aerosol (SOA). The sentence was modified to make it clear and shown below.
Changes in manuscript:
Results and discussion (Lines 348-349 in the revision): Changing the MEGAN inputs has a large impact on isoprene emissions, which are the main contributor to BVOC emissions. This further impact the formation of SOA.
4.22 364: Why is summer BSOA in C1 2.5 times higher than in C4?
Response: This is due to the higher BVOC emissions in the C1. The reason was added to the manuscript. Figure R5 (named as Figure S8 in the revision) was added to the supplementary material. Detail see attachment.
Changes in manuscript:
Results and discussion (Lines 357-358 in the revision): This is because that the summer BVOC emissions in C1 are higher than those in C4 in the YRD (Fig. S8) and thus formed more BSOA.
Changes in supplementary material:
Figure R5 see attachment.
4.23 From Fig. 4 you found that BSOA changed for different LAI as well as for different PFT inputs although the BVOC emissions (analyzed before) don’t change much with different LAI. Why is that?
Response: BVOC is an important contributor to SOA, so changing BVOC emissions will have an impact on the formation of SOA, and the magnitude of the impact depends on the amount of change in BVOC emissions. In this study, although changing the LAI for the MEGAN model has little effect on BVOC estimates, it could be amplified in BSOA through complex chemical process involving gas-particle reactions (Mahilang et al., 2021).
Changes in manuscript: No changes were made for this comment.
4.24 267: For which year did the studies, which are referred here, estimate the emissions?
Response: The study estimated the BVOC emissions in 2018. The study year for the reference was added to the sentence.
Changes in manuscript:
Results and discussion (Lines 270-271 in the revision): However, results in this study are lower than 58.9 Tg for 2018 estimated by Li et al. (2020).
Technical corrections:
- Abstract: introduce abbreviations LAL and LC.
Response: Revised accordingly.
- 49: as in PRD …
Response: Revised accordingly.
- Grammar and tense issues: l.52/53: ‘quantifi(ed)/determine(d)’, l. 55 ‘affect(ing)’, l.99&l.111 &l.145&l.122 ‘were’ -> are
Response: Revised accordingly.
- 104 ‘major(ity)’
Response: Revised accordingly.
- Sentence in line 62 f. (‘Wang et al. (2018a) misses one verb
Response: Revised accordingly.
- Sentence in l. 71 has a complicated sentence structure, integrate the last part in the main sentence
Response: Revised accordingly.
- -(l.100) ‘of each LC type […] and [this]’
Response: Revised accordingly.
- 158 ‘based on five different cases’ -> with 5 different sets of LC input data for MEGAN
Response: Revised accordingly.
- Footnote in Table S3 not found, in caption: ‘The values are bolded without meeting the benchmarks’ -> ‘The values without meeting the benchmarks are bolded’
Response: Revised accordingly.
- S3: scale could be improved
Response: Revised accordingly.
- Table S5: units are not given in the caption, information on benchmarks in the table, distinction between MDO3 and MDA8 could be clearer
Response: Revised accordingly.
- 252: are occurred -> occur
Response: Revised accordingly.
- 260: shouldn’t be the sentence refer to Fig. S4 ?
Response: Revised accordingly.
- 344/45: the plots from the three different species are hard to compare since they have different scales
Response: The sentence was deleted.
- 347: Reformulate
Response: The sentence was deleted.
- 350: attribute -> be attributed
Response: Revised accordingly.
- 355: delete ‘amount of’
Response: Revised accordingly.
- 359: ‘temporal trend’ -> seasonal cycle ?
Response: Revised accordingly.
- 369: sometimes you mean both LAI and PFT when you write LC datasets
Response: We corrected the mistakes.
- 374-76: More comprehensive formulation needed for the conclusion
Response: Revised accordingly.
-
RC2: 'Comment on acp-2022-739', Anonymous Referee #2, 11 Jan 2023
Manuscript Review
In the manuscript titled “Impacts of land cover changes on biogenic emission and its contribution to ozone and secondary organic aerosol in China” the authors have analysed the impact of LAI and LC datasets on biogenic VOCs emission over China. The study has further attempted to understand its contribution to O3 and secondary organic aerosols by performing five different set of simulations using MEGAN and WRF-CMAQ models.
This study provides an interesting insight into the uncertainties in regard to the biogenic VOC emission input data and the impact it can have on air quality. However, there are a few minor comments which the authors can address in the revised manuscript.
Minor Comments:
- Page 1 Abstract Section: The results from the study should be stated for each experimental design before drawing out the inference as to which input dataset is best suited for the study region along with a sentence on the significance/impact of this study in the concluding statement of the abstract.
- Page 1 Line 17: Define LAI, LC, SOA and BSOA abbreviation in the Abstract.
- Page 1, Line 44 and 45: Change “isoprene emission ranged” and “monoterpene emissions ranged” to “isoprene emissions are ranged” and “monoterpene emissions are ranged”
- Page 1 Line 52 and 53: Either “quantified” to “quantify” and “determined” to “determine” or rephrase the sentence accordingly to fit the tense.
- Page 1 Line 55: Remove ‘that’
- Page 3 Line 93: Remove “land cover” from “land cover (LC)”. Abbreviation only needs to be defined at its first instance in the manuscript, authors need not define it again for every section.
- Page 4 Section 2.1 Data description: Mention of supplementary material Table S2 and Fig S2 comes before the Table or Figure S1. To ensure a chronical order to the flow of figures and tables in text and supplementary, authors should renumber the supplementary figures and tables as it appears in the manuscript.
- Page 4 Line 100-101: The text mentions the 16 PFTs classifications whereas, Figure 1 only shows 15 classifications. Correct accordingly.
- Supplementary Table S2 lists the sources of the LC and LAI datasets. The table should include more information to better describe all the datasets used in the study such as temporal resolution spatial resolution and years for which data is available etc.
- Page 4 Line 106: The description of Figure S2 in text seems to be Figure S4 in supplementary material. Check and correct.
- Page 5 Line 145: States the physical schemes adopted in WRF model. Authors are suggested to support the choice of these WRF physical parameterizations through literature review for the same study region.
- Page 5 Line 148: Number of grid cells of CMAQ model are given. Is the CMAQ model running at the same resolution as the WRF model? Provide horizontal spatial resolution for CMAQ model too.
- Page 5 Lines 150-153: The sources of anthropogenic emissions used for China and other countries is provided. Few sentences on MEIC and EDGAR emission inventories should be added to understand the reason behind selection of these inventories.
- Table S4 gives the benchmarks for some statistical indices for meteorological parameters. It is suggested to provide the reference of these benchmarks. Similarly, provide the reference for O3 benchmark as well.
- Page 7 Line 212: Define the term ‘LAIv’.
- Page 9 Line 260: “Fig S3” should be “S4”.
- Page 11 Lines 323-325: The results analyzed in Section 3.4.1, have a special mention to the BSOA concentrations noted over Sichuan basin. As it is also stated in the Conclusion Section on Page 12, it would be nice to mark the location of Sichuan basin on the spatial figures of both main manuscript and supplementary material.
Citation: https://doi.org/10.5194/acp-2022-739-RC2 -
AC2: 'Reply on RC2', Hongliang Zhang, 21 Feb 2023
Dear Referee #2,
We appreciate your comments to help improve the manuscript. We tried our best to address your comments and detailed responses and related changes are shown below and in the attached PDF file.
Comments: In the manuscript titled “Impacts of land cover changes on biogenic emission and its contribution to ozone and secondary organic aerosol in China” the authors have analysed the impact of LAI and LC datasets on biogenic VOCs emission over China. The study has further attempted to understand its contribution to O3 and secondary organic aerosols by performing five different set of simulations using MEGAN and WRF-CMAQ models.
This study provides an interesting insight into the uncertainties in regard to the biogenic VOC emission input data and the impact it can have on air quality. However, there are a few minor comments which the authors can address in the revised manuscript.
Response: Thanks for the recognition of our study. Below is the response to each specific comment.
Minor Comments:
- Page 1 Abstract Section: The results from the study should be stated for each experimental design before drawing out the inference as to which input dataset is best suited for the study region along with a sentence on the significance/impact of this study in the concluding statement of the abstract.
Response: Thanks for your comments. A brief summary of model validation and the sentence about the significant of this study were added to the Abstract.
Changes in manuscript:
Abstract (Lines 23-31 in the revision): Changing the LC inputs for the MEGAN model has a more significant difference in BVOC estimates than using different LAI datasets. The C4 case has better model performance, indicating that it is the better choice for BVOC estimations in China. Changing the MEGAN inputs further impacts the concentrations of O3 and SOA. The highest O3 and biogenic SOA (BSOA) concentrations appear in the C1 (using GLASS and MCD12Q1 LC) simulation, which can reach 12 ppb and 9.8 μg m-3, respectively. Due to the combined effect of local BVOC emissions and the summer monsoon, the relative difference between C1 and C4 is over 52% and 140% in O3 and BSOA in central and eastern China. The BSOA difference between C1 and C4 is mainly attributed to the isoprene SOA (ISOA), which is a major contributor to BSOA. Particularly, the relative difference in ISOA between these two cases is up to 160% in eastern China. Therefore, our results suggest that the uncertainties in MEGAN inputs should be fully considered in future O3 and SOA simulations.
- Page 1 Line 17: Define LAI, LC, SOA and BSOA abbreviation in the Abstract.
Response: The abbreviations of these terms were added to the Abstract.
Changes in manuscript:
Abstract (Lines 14-15 in the revision): “In this study, Model of Emissions of Gases and Aerosols from Nature (MEGAN) v2.1 was used to investigate the impact of different leaf area index (LAI) and land cover (LC) …”
- Page 1, Line 44 and 45: Change “isoprene emission ranged” and “monoterpene emissions ranged” to “isoprene emissions are ranged” and “monoterpene emissions are ranged”:
Response: Revised accordingly.
Changes in manuscript:
Abstract (Lines 45-47 in the revision): Global annual inventories of the isoprene emission are ranged from 500 to 750 Tg yr-1 (Guenther et al., 2006) and those of monoterpene emissions are ranged from 74.4-157 Tg yr-1 (Guenther et al., 2012;Messina et al., 2016).
- Page 1 Line 52 and 53: Either “quantified” to “quantify” and “determined” to “determine” or rephrase the sentence accordingly to fit the tense.
Response: Revised accordingly.
Changes in manuscript:
Introduction (Lines 53-54 in the revision): Therefore, it is necessary to quantify the influence of those factors and determine the bias in BVOC emissions.
- Page 1 Line 55: Remove ‘that’.
Response: Revised accordingly.
Changes in manuscript:
Introduction (Lines 56-57 in the revision): Land cover (LC), including leaf area index (LAI) and plant function types (PFTs) fractions, is a major factor affecting the BVOC emissions in the MEGAN model.
- Page 3 Line 93: Remove “land cover” from “land cover (LC)”. Abbreviation only needs to be defined at its first instance in the manuscript, authors need not define it again for every section.
Response: Revised accordingly.
Changes in manuscript:
Methodology (Lines 136 in the revision): “Three LC datasets were applied as PFTs inputs…”.
- Page 4 Section 2.1 Data description: Mention of supplementary material Table S2 and Fig S2 comes before the Table or Figure S1. To ensure a chronical order to the flow of figures and tables in text and supplementary, authors should renumber the supplementary figures and tables as it appears in the manuscript.
Response: Revised accordingly.
Changes in manuscript:
Methodology (Lines 118, 122, 142, and 143 in the revision):
“…China and its surrounding countries in East Asia (Fig. S1) …”
“… Table S1 briefly lists the physical options used for the WRF model.”
“Sources of these products were listed in Table S2.”
“Three LC maps are first re-gridded to the CMAQ domain (Fig. S2).”
- Page 4 Line 100-101: The text mentions the 16 PFTs classifications whereas, Figure 1 only shows 15 classifications. Correct accordingly.
Response: Revised accordingly.
Changes in manuscript:
Methodology (Lines 144-145 in the revision): Lastly, eight vegetation types are further reclassified into CLM-15 PFTs based on the climate rules described in Bonan et al. (2002).
- Supplementary Table S2 lists the sources of the LC and LAI datasets. The table should include more information to better describe all the datasets used in the study such as temporal resolution spatial resolution and years for which data is available etc.
Response: Thanks for your suggestions. The temporal resolution, spatial resolution and available years for satellite datasets used in this study were added to the Table R1 (named as Table S2 in the revision). Detail see attachment.
Changes in supplementary material:
Table R1 see attachment.
- Page 4 Line 106: The description of Figure S2 in text seems to be Figure S4 in supplementary material. Check and correct.
Response: Revised accordingly.
Changes in manuscript:
Methodology (Lines 148-150 in the revision): Although MCD12Q1 and CGLS LC both show a large area of broadleaf tree in central and southern China, the area fraction of broadleaf tree in CGLS LC is higher than that in MCD12Q1 (Fig 1 and Fig. S3).
- Page 5 Line 145: States the physical schemes adopted in WRF model. Authors are suggested to support the choice of these WRF physical parameterizations through literature review for the same study region.
Response: Thanks for your suggestions. The references were added to the manuscript.
Changes in manuscript:
Methodology (Lines 121-122 in the revision): The model configurations are similar to the previous studies (Wang et al., 2018;Wang et al., 2020;Zhu et al., 2021) and Table S1 briefly lists the physical options used for the WRF model.
- Page 5 Line 148: Number of grid cells of CMAQ model are given. Is the CMAQ model running at the same resolution as the WRF model? Provide horizontal spatial resolution for CMAQ model too.
Response: Yes, it is. The horizontal spatial resolution for the Community Multiscale Air Quality Modelling System (CMAQ) model was added to the sentence and shown below.
Changes in manuscript:
Methodology (Lines 125-126 in the revision): The CMAQ model used the same horizontal resolution as WRF with a horizontal domain of 197 × 127 grid cells. This domain covers China and its surrounding areas (Fig. S1).
- Page 5 Lines 150-153: The sources of anthropogenic emissions used for China and other countries is provided. Few sentences on MEIC and EDGAR emission inventories should be added to understand the reason behind selection of these inventories.
Response: Thanks for your comments. The reasons for choosing MEIC and EDGAR were added to the manuscript.
Changes in manuscript:
Methodology (Lines 127-134 in the revision): The anthropogenic emissions of China used the datasets from Multiresolution Emission Inventory for China (MEIC; available at http://www. meicmodel.org, last access:3 May 2022). Since the MEIC only provides anthropogenic emissions for China, anthropogenic emissions from foreign countries were provided by the Emissions Database for Global Atmospheric Research (EDGAR) v4.3 (available at http://edgar.jrc.ec.europa.eu/overview.php? v=_431, last access: 10 May 2022). The MEIC inventory is widely used in air quality studies in China (Li et al., 2017;Hu et al., 2016;Wu et al., 2020). It had an improvement in a vehicle emission inventory with high resolution (Zheng et al., 2014), and a non-methane VOC mapping approach for different chemical mechanisms (Li et al., 2014). The EDGAR is a grided emissions inventory with a high horizontal resolution of 0.1°×0.1° (Saikawa et al., 2017).
- 14. Table S4 gives the benchmarks for some statistical indices for meteorological parameters. It is suggested to provide the reference of these benchmarks. Similarly, provide the reference for O3 benchmark as well.
Response: The references for the benchmarks were added to the Table S4 and Table S5.
Changes in supplementary material:
“Note: * are benchmarks limits suggested by (Emery et al., 2001).”
“Note: * are criteria suggested by EPA (2007).”
- Page 7 Line 212: Define the term ‘LAIv’.
Response: LAIv means the LAI of vegetation covered surface, which is calculated by dividing the grid average LAI with the fraction of grid that is covered by vegetation (Guenther et al., 2006). This term has been explained in the manuscript in Line 161.
Changes in manuscript: No changes were made for this comment.
- Page 9 Line 260: “Fig S3” should be “S4”.
Response: Revised accordingly.
Changes in manuscript:
Results and discussion (Lines 261-263 in the revision): The spatial distribution of isoprene emission in C5 is conspicuously different than in C1, which is consistent with a difference in the broadleaf tree distribution in the inputs (Fig. S5).
- Page 11 Lines 323-325: The results analyzed in Section 3.4.1, have a special mention to the BSOA concentrations noted over Sichuan basin. As it is also stated in the Conclusion Section on Page 12, it would be nice to mark the location of Sichuan basin on the spatial figures of both main manuscript and supplementary material.
Response: Thanks for your comments. We marked the location of the Sichuan basin in the Figure R1 (named as Figure S1 in the revision). Detail see attachment.
Changes in supplementary material:
Figure R1 see attachment.
Status: closed
-
RC1: 'Comment on acp-2022-739', Anonymous Referee #1, 05 Dec 2022
General comments:
This paper explores the impact of different land cover data sets on biogenic emissions and its contribution to ozone and secondary organic aerosol in China with the Model of Emissions of Gases and Aerosols from Nature (MEGAN) v2.1 employing a further developed version of the Community Multiscale Air Quality Modelling System (CMAQ). For that, a set of experiments with three different leaf area index (LAI) and plant functional type (PFT) input were conducted to show the impact of the different input to BVOC emissions, extreme ozone and SOA. The BVOC emissions estimated from all simulations are within the reported range of literature values whereas, however, only the changes of LC data lead to a significant impact on BVOC emissions and related air quality.
This investigation will be valuable, also on a global scale to assess further the uncertainty of the input data to BVOC emissions and their products and finally improve the model representation of air quality. Besides the following suggestions, I advise the authors to improve the manuscript regarding the language and the structure, adapt the visualization more to the analysis purpose and strengthen the conclusion.
Specific comments:
Abstract: In particular, the beginning is a bit unstructured. Some sentences don’t have enough content to understand the link between different aspects see the following points. If you structure it more following a red line it may also lead to more comprehension of the text.
- Time period of ‘greening impacts in China’? Which role does this play for BVOC emissions? How does BVOCs act for O3 and SOA? Why are these important species to study? – air quality/health impact. How is the BVOC estimation based on inventories, and how does this link to satellites? If you want to mention the uncertainty of satellites include a relative estimate.
- A more general formulation of the results would make the abstract more appealing. Also, including relative changes helps the reader to judge the impact order of your changes.
Introduction: It’s confusing that some sentences are lacking scientific background and with some you go in very detail as can be seen in the following points. In particular the last paragraph could be more conclusive by linking the mentioned studies more together:
- ‘wide range of warming and cooling climate pollutants’ (l.39) is to unspecific. The different radiative forcing/role in climate between aerosols and greenhouse gases like O3 should be mentioned
- (l.51) The role of emission and activity factors for BVOC emissions need to be explained before. In general, I suggest to describe in short how BVOC emissions are calculated by the model
- (l. 69) which CTM was used in the cited study?
- - (l. 71 f.) Why is the particular SOA concentration in Sichuan mentioned here? Is this a high number?
Section 2: I suggest to re-structure this section (Model description: CMAQ, MEGAN, WARF; Data description, experiments) to improve readability. Along this, I have some question:
- (Line 100/101) According to which maps are the PFT classified?
- (l.112) To which LAI maps does GLASS show better consistency?
- 140-153 belongs to ‘Model description’
- 150 f. Difficult sentence structure. Is MEIC based on EDGAR?
- 147: The chemistry does not impact the meteorology, right?
Section 3: The statistical analysis is useful, but in the later analysis I often miss the conclusion/interpretation of the simulation results instead the authors often only describe the figures. What can we learn from the data/results?
- 158: criteria -> uncertainty? How is the uncertainty calculated?
- 170/171: The day-to-day variability in Fig S2 are hard to see. Think about another visualization when you want to show this.
- 172 f.: The MB and he GE of the T2 exceeds the 'benchmark', you have to explain this, and why the model performance is still acceptable. Temperature is a crucial driver of BVOC emissions.
- 179: What is your motivation to investigate both, MDA1O3 AND MDA8O3?
- 181: What are cut-off concentrations?
- 199: I wouldn’t say that air quality is not correlated to LAI input only because of this result. Not only BVOC emissions rely on LAI input (also e.g. dry deposition)
- 214: definition or reference for LAIv=2
- 215: In which figure can this be seen?
- 220 and before: Are these findings in agreement with other studies?
- 221: seasonal variations are not easy to see in Fig. 3. I suggest to show/add the annual cycle
- 224-226: the two sentences are contradicting.
- 225 ‘seasonal change trend’. What do you mean? You do not consider multiple years
- 228: Is this in agreement with other findings?
- 230/31: Are the different species emitted by different PFTs? (Could also be a reason for the difference). What is the temperature coefficient?
- 233/234: Why have the monoterpenes only little sensitivity to LAI sensitivity
- 240 and before: I suggest to show the emission change per each PFT which would make it easier to interpret and draw general conclusions
- 267: For which year did the studies, which are referred here, estimate the emissions?
- Table 4: It would make more sense to compare the emission estimates from approximately the same time period. Also, adding one more column in the table with the average LAI value would improve the clarity or alternatively list the BVOC emission estimates normalized by the LAI. Why is the estimate for 2018, which is quite close to your study period, almost double?
- Line 286: How do you estimate the BVOC emissions only formed from O3? (Information could be added to the Methods section
- 6: The seasonal variation is hardly visible and might be better displayed by seasonal cycles (with uncertainty bars)
- 327: Why don't you use the same unit as in the Fig. 7?
- 352/53: Do you want to say that changing the MEGAN inputs has a large impact on isoprene emissions and since isoprene is the main contributor to BVOCs this also impact BVOC emissions?
- 364: Why is summer BSOA in C1 2.5 times higher than in C4?
- From Fig. 4 you found that BSOA changed for different LAI as well as for different PFT inputs although the BVOC emissions (analyzed before) don’t change much with different LAI. Why is that?
Conclusion: l.370: How do you judge that C4 was better for BVOC emissions? You haven't done an evaluation of the BVOC emissions, or?
- 372: is this consistent with other studies?
- 380: name the concrete impact (e.g. relative changes)
Technical corrections:
- Abstract: introduce abbreviations LAL and LC
- 49: as in PRD …
- Grammar and tense issues: l.52/53: ‘quantifi(ed)/determine(d)’, l. 55 ‘affect(ing)’, l.99&l.111 &l.145&l.122 ‘were’ -> are
- 104 ‘major(ity)’
- Sentence in line 62 f. (‘Wang et al. (2018a) misses one verb
- Sentence in l. 71 has a complicated sentence structure, integrate the last part in the main sentence
- -(l.100) ‘of each LC type […] and [this]’
- 158 ‘based on five different cases’ -> with 5 different sets of LC input data for MEGAN
- Footnote in Table S3 not found, in caption: ‘ The values are bolded without meeting the benchmarks’ -> ‘The values without meeting the benchmarks are bolded’
- S3: scale could be improved
- Table S5: units are not given in the caption, information on benchmarks in the table, distinction between MDO3 and MDA8 could be clearer
- 252: are occurred -> occur
- 260: shouldn’t be the sentence refer to Fig. S4 ?
- 344/45: the plots from the three different species are hard to compare since they have different scales
- 347: Reformulate
- 350: attribute -> be attributed
- 355: delete ‘amount of’
- 359: ‘temporal trend’ -> seasonal cycle ?
- 369: sometimes you mean both LAI and PFT when you write LC datasets
- 374-76: More comprehensive formulation needed for the conclusion
Citation: https://doi.org/10.5194/acp-2022-739-RC1 -
AC1: 'Reply on RC1', Hongliang Zhang, 21 Feb 2023
Dear Referee #1,
We appreciate your comments to help improve the manuscript. We tried our best to address your comments and detailed responses and related changes are shown below and in the attached PDF file.
Comments: This paper explores the impact of different land cover data sets on biogenic emissions and its contribution to ozone and secondary organic aerosol in China with the Model of Emissions of Gases and Aerosols from Nature (MEGAN) v2.1 employing a further developed version of the Community Multiscale Air Quality Modelling System (CMAQ). For that, a set of experiments with three different leaf area index (LAI) and plant functional type (PFT) input were conducted to show the impact of the different input to BVOC emissions, extreme ozone and SOA. The BVOC emissions estimated from all simulations are within the reported range of literature values whereas, however, only the changes of LC data lead to a significant impact on BVOC emissions and related air quality.
This investigation will be valuable, also on a global scale to assess further the uncertainty of the input data to BVOC emissions and their products and finally improve the model representation of air quality. Besides the following suggestions, I advise the authors to improve the manuscript regarding the language and the structure, adapt the visualization more to the analysis purpose and strengthen the conclusion.
Response: Thanks for the recognition of our study. Below is the response to each specific comment.
Comments:
- Abstract: In particular, the beginning is a bit unstructured. Some sentences don’t have enough content to understand the link between different aspects see the following points. If you structure it more following a red line it may also lead to more comprehension of the text.
1.1 Time period of ‘greening impacts in China’? Which role does this play for BVOC emissions? How does BVOCs act for O3 and SOA? Why are these important species to study? – air quality/health impact. How is the BVOC estimation based on inventories, and how does this link to satellites? If you want to mention the uncertainty of satellites include a relative estimate.
Response: Thanks for pointing out these important issues. The period of ‘greening impacts in China’ is from 2000 to 2017 (Chen et al., 2019). The greening impacts increase the vegetated area, further increasing the biogenic volatile organic compounds (BVOC) emissions. Since BVOCs are the important precursor for ozone (O3) and secondary organic aerosol (SOA), the increase in BVOC emissions further changes concentrations of O3 and SOA. In this study, the BVOC emissions are estimated by the Model of Emissions of Gases and Aerosols from Nature (MEGAN) and the satellite datasets are necessary inputs for the MEGAN model. The sentences in the abstract were modified and shown below.
Changes in manuscript:
Abstract (Lines 11-14 in the revision): The greening impacts in China from 2000 to 2017 led to an increase in vegetated areas and thus enhanced biogenic volatile organic compounds (BVOC) emissions. BVOCs are regarded as important precursors for ozone (O3) and secondary organic aerosol (SOA). As a result, accurate estimation of BVOC emissions is critical to understanding their impacts on air quality.
1.2 A more general formulation of the results would make the abstract more appealing. Also, including relative changes helps the reader to judge the impact order of your changes.
Response: Thanks for your suggestions. The results were modified to be more general and the differences between cases were presented with the relative difference in abstract.
Changes in manuscript:
Abstract (Lines 23-31 in the revision): Changing the LC inputs for the MEGAN model has a more significant difference in BVOC estimates than using different LAI datasets. The C4 case has better model performance, indicating that it is the better choice for BVOC estimations in China. Changing the MEGAN inputs further impacts the concentrations of O3 and SOA. The highest O3 and biogenic SOA (BSOA) concentrations appear in the C1 (using GLASS and MCD12Q1 LC) simulation, which can reach 12 ppb and 9.8 μg m-3, respectively. Due to the combined effect of local BVOC emissions and the summer monsoon, the relative difference between C1 and C4 is over 52% and 140% in O3 and BSOA in central and eastern China. The BSOA difference between C1 and C4 is mainly attributed to the isoprene SOA (ISOA), which is a major contributor to BSOA. Particularly, the relative difference in ISOA between these two cases is up to 160% in eastern China. Therefore, our results suggest that the uncertainties in MEGAN inputs should be fully considered in future O3 and SOA simulations.
- Introduction: It’s confusing that some sentences are lacking scientific background and with some you go in very detail as can be seen in the following points. In particular the last paragraph could be more conclusive by linking the mentioned studies more together:
2.1 ‘wide range of warming and cooling climate pollutants’ (l.39) is to unspecific. The different radiative forcing/role in climate between aerosols and greenhouse gases like O3 should be mentioned.
Response: The pollutants and their effects on climate were determined in the manuscript and listed below.
Changes in manuscript:
Introduction (Lines 38-41 in the revision): In addition, changes in emissions of BVOCs will alter the capacity of a wide range of warming and cooling climate pollutants, such as O3, methane (CH4) and aerosols. O3 and CH4 can warm the climate, while the aerosols have a cooling effect by scattering solar radiation (Unger, 2014b, a).
2.2 (l.51) The role of emission and activity factors for BVOC emissions need to be explained before. In general, I suggest to describe in short how BVOC emissions are calculated by the model.
Response: Thanks for your suggestion. The calculation of BVOC emission was added in the manuscript and explained below.
Changes in manuscript:
Introduction (Lines 49-51 in the revision): The model determined the vegetation types according to model inputs and then use the activity factor multiplied with the emission factor to calculate emissions for each vegetation type (Guenther et al., 2012).
2.3 (l. 69) which CTM was used in the cited study?
Response: The specific name for the chemical transport model (CTM) of the cited study was written in the manuscript and shown below.
Changes in manuscript:
Introduction (Lines 69-72 in the revision): Fu and Liao (2012) used the Goddard Earth Observing System chemical transport model (GEOS-Chem) to quantitate the impact of biogenic emissions on O3 in China over the year 2001-2006 and found that the difference in O3 concentrations induced by interannual variability of BVOCs could be 2-5%.
2.4 (l. 71 f.) Why is the particular SOA concentration in Sichuan mentioned here? Is this a high number?
Response: Yes, it is. The sentence was modified to clear its point in the manuscript and shown below.
Changes in manuscript:
Introduction (Lines 76-77 in the revision): Qin et al. (2018) investigated the biogenic SOA (BSOA) during summertime in 2012 and found that a high level of BSOA concentration appeared in Sichuan Basin.
- Section 2: I suggest to re-structure this section (Model description: CMAQ, MEGAN, WARF; Data description, experiments) to improve readability. Along this, I have some question:
Response: Thanks for your suggestions. The structure of the section ‘Methodology’ was changed from Data description, Model description, and Model application to Model setup, and Data description.
3.1 (Line 100/101) According to which maps are the PFT classified?
Response:
The plant function type (PFT) scheme used in the MEGAN is classified based on Community Land Model v4.0 (CLM4) (Guenther et al., 2012). The sentence was modified and Figure R1 (named as Figure S2 in the revision) was added to the manuscript to make it clear. Detail see attachment.
Changes in manuscript:
Methodology (Lines 142-145 in the revision): PFTs used in the MEGAN model adopt the scheme used for Community Land Model v4.0 (CLM4) (Guenther et al., 2012). Three LC maps are first re-gridded to the CMAQ domain (Fig. S2). Secondly, LC types are categorized into eight vegetation types according to legend descriptions of LC maps. Lastly, eight vegetation types are further reclassified into CLM-15 PFTs based on the climate rules described in Bonan et al. (2002).
Changes in supplementary material:
Figure R1 see attachment.
3.2 (l.112) To which LAI maps does GLASS show better consistency?
Response: In leaf area index (LAI) satellite products, the Global LAnd Surface Satellite (GLASS) products show better consistency than the MODIS MOD15A2H version 6 (MOD15) products according to the study in Xiao et al. (2016), while the accuracy of the Copernicus Global Land Service (CGLS) products are slightly less than MOD15 products (Fuster et al., 2020). Therefore, GLASS products have a better consistency than MOD15 products and CGLS products.
Changes in manuscript: No changes were made for this comment.
3.3 140-153 belongs to ‘Model description’
Response: Thanks for your comments. The subheading ‘Model description’ was changed to ‘Model setup’ and sentences in line140-153 were moved to ‘Model setup’.
3.4 150 f. Difficult sentence structure. Is MEIC based on EDGAR?
Response: No, it is not. The modelling domain in CMAQ was 36 km × 36 km in horizontal spatial resolution, which covers China and its surrounding countries in East Asia (Fig. S1). Since Multiresolution Emission Inventory for China (MEIC) only provided anthropogenic emissions for China, while anthropogenic emissions for regions excluding China were provided by the Emissions Database for Global Atmospheric Research (EDGAR) v4.3. The sentence was modified to make it clear and shown below.
Changes in manuscript:
Methodology (Lines 127-134 in the revision): The anthropogenic emissions of China used the datasets from Multiresolution Emission Inventory for China (MEIC; available at http://www. meicmodel.org, last access:3 May 2022). Since the MEIC only provides anthropogenic emissions for China, anthropogenic emissions from foreign countries were provided by the Emissions Database for Global Atmospheric Research (EDGAR) v4.3 (available at http://edgar.jrc.ec.europa.eu/overview.php? v=_431, last access: 10 May 2022). The MEIC inventory is widely used in air quality studies in China (Li et al., 2017;Hu et al., 2016;Wu et al., 2020). It had an improvement in a vehicle emission inventory with high resolution (Zheng et al., 2014), and a non-methane VOC mapping approach for different chemical mechanisms (Li et al., 2014). The EDGAR is a grided emissions inventory with a high horizontal resolution of 0.1°×0.1° (Saikawa et al., 2017).
3.5 147: The chemistry does not impact the meteorology, right?
Response: The chemistry does not impact the meteorology. We made it clear in the manuscript that chemistry has no effect on meteorology.
Changes in manuscript:
Methodology (Lines 171 in the revision): The model chemistry has no effect on meteorological conditions when simulating.
- 4. Section 3: The statistical analysis is useful, but in the later analysis I often miss the conclusion/interpretation of the simulation results instead the authors often only describe the figures. What can we learn from the data/results?
Response: In this study, we concluded that changing the MEGAN inputs has an impact on BVOC estimates, and this further influences the formation of O3 and secondary organic aerosol (SOA). Besides, changing land cover datasets for the model shows more conspicuous differences in BVOC emissions than using different LAI datasets.
4.1 158: criteria -> uncertainty? How is the uncertainty calculated?
Response: The criterion is a way of evaluating the model performance and is obtained based on the observations. Simulations with different parameter schemes will generate uncertainties that can lead to bias between predictions and observations, hence, the need to use the criterion to measure model performance. The criteria of 2 is a benchmark value for wind speed. The criteria for other meteorological conditions and calculation methods are described by Emery et al. (2001).
Changes in manuscript: No changes were made for this comment.
4.2 170/171: The day-to-day variability in Fig S2 are hard to see. Think about another visualization when you want to show this.
Response: Thanks for your advice. We changed the line chart of day-to-day variability for 2 m temperature into the scatter chart, which is Figure R2 (named as Figure S3 in the revision). Detail see attachment.
Changes in supplementary material:
Figure R2 see attachment.
4.3 172 f.: The MB and he GE of the T2 exceeds the 'benchmark', you have to explain this, and why the model performance is still acceptable. Temperature is a crucial driver of BVOC emissions.
Response: The result of the mean bias and the gross error of temperature at 2 m (T2) exceeding the benchmark is possible due to the overestimation of cloud coverage in the WRF model leading to an underestimated T2 (Wu et al., 2020). These biases of T2 are relatively small compared with previous studies with a yearly long WRF simulation in China (Wu et al., 2020;Wang et al., 2018a). The reasons were explained in the manuscript and shown below.
Changes in manuscript:
Results and discussion (Lines 178-181 in the revision): It is possible due to the overestimation of cloud coverage in the WRF model leading to an underestimated T2 (Wu et al., 2020). These biases are relatively small compared with previous studies with a yearly long WRF simulation in China (Wu et al., 2020;Wang et al., 2018a).
4.4 179: What is your motivation to investigate both, MDA1O3 AND MDA8O3?
Response: The maximum daily averaged 1h (MDA1) O3 and maximum daily averaged 8h (MDA8) O3 are different indicators for O3 in the National Ambient Air Quality Standard (NAAQS) proposed by EPA (1997, 2015). They are used to represent concentrations for the short-term (1 to 3 hours) and the prolonged (6 to 8 hours) exposures to O3 according to the NAAQS. They differ spatially and temporally due to differences in calculation methods (Bell and Ellis, 2003). China, the study area in this study, also uses them to evaluate the O3 pollution (MEP, 2012). Therefore, they are both important and need to be investigated.
Changes in manuscript: No changes were made for this comment.
4.5 181: What are cut-off concentrations?
Response: The cut-off concentration is a minimum threshold value used to restrict observations to mitigate the bias of the model performance analysis (Emery et al., 2017). The 60 ppb as the threshold for O3 is suggested by EPA (2007) to focus bias and error statistics on times of higher O3.
Changes in manuscript: No changes were made for this comment.
4.6 199: I wouldn’t say that air quality is not correlated to LAI input only because of this result. Not only BVOC emissions rely on LAI input (also e.g. dry deposition)
Response: The dry deposition is a process modelled by the Community Multiscale Air Quality Modelling System (CMAQ). Although this process will be influenced by the satellite-derived land use and vegetation products (Pleim and Ran, 2011), these parameters provided by Weather Research and Forecasting model (WRF) remain unchanged in this study. According to the scope of this paper, only inputs for the Model of Emissions of Gases and Aerosols from Nature (MEGAN) were changed, which did not have impacts on the WRF model, and thus the dry deposition would not be influenced. We modified the sentence to make it appropriate.
Changes in manuscript:
Results and discussion (Lines 211-213 in the revision): Although C1, C2, and C3 adopt LAI satellite products with different accuracies, the accuracies of these products have no significant impact on the model performance due to similar statistics values.
4.7 214: definition or reference for LAIv=2
Response: LAIv means the average LAI for vegetated areas, which is estimated by dividing the grid average LAI by the fraction of the grid that is covered by vegetation (Guenther et al., 2006). The related sentence was modified.
Changes in manuscript:
Results and discussion (Lines 224-225 in the revision): The BVOC emissions in C1 are about 5 Tg higher than C2.
4.8 215: In which figure can this be seen?
Response: The remarkable difference in isoprene between C1 and C2 can be seen in Table 2. The related sentences were modified.
Changes in manuscript:
Results and discussion (Lines 224-225 in the revision): The BVOC emissions in C1 are about 5 Tg higher than C2.
4.9 220 and before: Are these findings in agreement with other studies?
Response: Yes, these findings agreed with the study of Wang et al. (2018). Wang et al. (2018) used three different land cover datasets to run the MEGAN model, which were the Finer Resolution Observation and Monitoring of Global Land Cover (E1), Moderate-Resolution Imaging Spectroradiometer (MODIS)
MCD12Q1 PFT products (E4), and the Climate Change Initiative Land Cover (CCI LC) products (E5) in his study. The higher fractions of broadleaf tree in E4 resulted in higher isoprene emissions than E1 and E5, but the lower fractions of needleleaf tree and shrub in E4 resulted in lower monoterpene and sesquiterpene emissions than E1. We added the reference to the sentence and showed below. Figure R3 (named as Figure S3 in the revision) was added to the supplementary material. Detail see attachment.
Changes in manuscript:
Results and discussion (Lines 227-230 in the revision): Although the total BVOC emissions in C5 are 1.29 Tg higher than those in C1, the emissions of monoterpenes, sesquiterpenes, and other VOCs are lower than in C1. This is induced by the discrepancy in the distribution of needleleaf tree and shrubs between C1 and C5, which is in agreement with the result in Wang et al. (2018) (Fig. S3 and Fig. S5).
Changes in supplementary material:
Figure R3 see attachment.
4.10 221: seasonal variations are not easy to see in Fig. 3. I suggest to show/add the annual cycle
Response: Thanks for your suggestion. We changed the Figure R4 (named as Figure 3 in the revision) from the histogram to the stacked column chart to represent the annual emissions of BVOC as shown below. As shown in the Figure R4, the BVOC emissions are mainly concentrated in summer, accounting for 60.9%~63.8% of total emissions compared to 2.9%~3.4% in winter. Detail see attachment.
Changes in manuscript:
Figure R4 see attachment.
4.11 224-226: the two sentences are contradicting.
Response: These two sentences were corrected in manuscript and shown below.
Changes in manuscript:
Results and discussion (Lines 233-235 in the revision): In general, using different LAI and LC products does have an impact on the temporal variability in the BVOC emissions. The BVOC emissions show similar seasonal variations in all cases, which are mainly concentrated in summer, accounting for 60.9%-63.8% of total BVOC emissions compared to 2.9%~3.4% in winter
4.12 225 ‘seasonal change trend’. What do you mean? You do not consider multiple years
Response: We wanted to express that the different cases showed similar seasonal change characteristics. The expression was corrected and shown below.
Changes in manuscript:
Results and discussion (Lines 233-235 in the revision): In general, using different LAI and LC products does have an impact on the temporal variability in the BVOC emissions. The BVOC emissions show similar seasonal variations in all cases, which are mainly concentrated in summer, accounting for 60.9%-63.8% of total BVOC emissions compared to 2.9%~3.4% in winter
4.13 228: Is this in agreement with other findings?
Response: Yes, it is. In the study of Ibrahim et al. (2010), the sesquiterpene emitted from birch and aspen increased faster than monoterpene with temperature rising. And Bai et al. (2015) concluded that the increasing rates of isoprene with temperature were much higher than for monoterpenes. These references were added to the sentence.
Changes in manuscript:
Results and discussion (Lines 238-241 in the revision): The percentage of winter monoterpenes in the total monoterpenes is higher than that of isoprene and sesquiterpenes, probably because isoprene and sesquiterpenes are more sensitive to temperature changes than monoterpenes (Ibrahim et al., 2010;Bai et al., 2015).
4.14 230/31: Are the different species emitted by different PFTs? (Could also be a reason for the difference). What is the temperature coefficient?
Response: No, they aren’t. In the MEGAN model, different PFTs can emit all BVOC species with different emission factors. The emission ( ) of chemical species i from vegetation type j according to
Where is the emission factor at standard conditions for different vegetation type j with fractional grid box areal coverage .The emission activity factor ( ) accounts for the processes controlling emission responses to environmental and phenological conditions (Guenther et al., 2012). The temperature coefficient is an empirical coefficient used to determine the light-independent fraction, and the light-independent is a part of the temperature activity factor in the MEGAN model (Guenther et al., 1993;Guenther et al., 2012;Helmig et al., 2006).
Changes in manuscript: No changes were made for this comment.
4.15 233/234: Why have the monoterpenes only little sensitivity to LAI sensitivity
Response: We made a mistake here and the sentence was deleted.
Changes in manuscript:
Results and discussion (Lines 241 in the revision): C4 used the C3S LC and GLASS shows the lowest emissions in total BVOCs and its main species among each season.
4.16 240 and before: I suggest to show the emission change per each PFT which would make it easier to interpret and draw general conclusions
Response: Thanks for your suggestion. The figure about the BVOC emissions of main PFT was added in the supplementary material (Figure R3, named as Figure S3 in the revision). In Figure R3, the highest fraction of broadleaf trees in C5 contributing large BVOC emissions. Although grass areas are higher than broadleaf tree areas, it does not contribute significantly to the BVOC emissions due to its lower the emission factor. Detail see attachment.
Changes in supplementary material:
Figure R3 see attachment.
4.17 Table 4: It would make more sense to compare the emission estimates from approximately the same time period. Also, adding one more column in the table with the average LAI value would improve the clarity or alternatively list the BVOC emission estimates normalized by the LAI. Why is the estimate for 2018, which is quite close to your study period, almost double?
Response: Many thanks for your comments. We changed the study comparison with Li et al. (2013) to a comparison with Li et al. (2020). This column about the LAI could not be added because the LAI data from previous studies were not available. The reason for the last question is that Li et al. (2020) produced the basal emission rates for 192 plant species and categorized them into 82 PFTs for China resulting in more BVOC estimates.
Changes in manuscript:
Results and discussion (Lines 280-283 in the revision): There is a considerable difference in BVOC emissions between this study and those of Li et al. (2020). The difference is mainly due to the combined effect of emission rate and PFTs. Liu et al. (2020) produced the basal emission rates for 192 plant species and categorized them into 82 PFTs for China resulting in more BVOC estimates.
4.18 Line 286: How do you estimate the BVOC emissions only formed from O3? (Information could be added to the Methods section
Response: The concentrations of O3 from different VOC sources (henceforth ) were determined by the source-oriented method (Ying and Krishnan, 2010). Based on the method, the non-reactive O3 tracer is used to track O3 attributed to BVOCs, which is tagged as and directly predicted. The descriptions of O3 source apportionment see detailed in Wang et al. (2019).
Changes in manuscript:
Methodology (Lines 105-108 in the revision): The concentrations of O3 from different VOC sources (henceforth ) were determined by the source-oriented method (Ying and Krishnan, 2010). Based on the method, the non-reactive O3 tracer is used to track O3 attributed to BVOCs, which is tagged as and directly predicted. The descriptions of O3 source apportionment see detailed in Wang et al. (2019).
4.19 6: The seasonal variation is hardly visible and might be better displayed by seasonal cycles (with uncertainty bars)
Response: Thanks for your suggestion. The seasonal histogram charts contain information not only on seasonal changes but seasonal differences between cases. Therefore, using the seasonal cycles will not get any better.
Changes in manuscript: No changes were made for this comment.
4.20 327: Why don't you use the same unit as in the Fig. 7?
Response: Ppb and μg m-3 are both the commonly used units (Wang et al., 2021;Zhang et al., 2021;Hu et al., 2016).
Changes in manuscript: No changes were made for this comment.
4.21 352/53: Do you want to say that changing the MEGAN inputs has a large impact on isoprene emissions and since isoprene is the main contributor to BVOCs this also impact BVOC emissions?
Response: We want to express that changing the MEGAN inputs has a large impact on isoprene emissions, which are the main contributor to BVOCs. And this impact further changes the concentration of secondary organic aerosol (SOA). The sentence was modified to make it clear and shown below.
Changes in manuscript:
Results and discussion (Lines 348-349 in the revision): Changing the MEGAN inputs has a large impact on isoprene emissions, which are the main contributor to BVOC emissions. This further impact the formation of SOA.
4.22 364: Why is summer BSOA in C1 2.5 times higher than in C4?
Response: This is due to the higher BVOC emissions in the C1. The reason was added to the manuscript. Figure R5 (named as Figure S8 in the revision) was added to the supplementary material. Detail see attachment.
Changes in manuscript:
Results and discussion (Lines 357-358 in the revision): This is because that the summer BVOC emissions in C1 are higher than those in C4 in the YRD (Fig. S8) and thus formed more BSOA.
Changes in supplementary material:
Figure R5 see attachment.
4.23 From Fig. 4 you found that BSOA changed for different LAI as well as for different PFT inputs although the BVOC emissions (analyzed before) don’t change much with different LAI. Why is that?
Response: BVOC is an important contributor to SOA, so changing BVOC emissions will have an impact on the formation of SOA, and the magnitude of the impact depends on the amount of change in BVOC emissions. In this study, although changing the LAI for the MEGAN model has little effect on BVOC estimates, it could be amplified in BSOA through complex chemical process involving gas-particle reactions (Mahilang et al., 2021).
Changes in manuscript: No changes were made for this comment.
4.24 267: For which year did the studies, which are referred here, estimate the emissions?
Response: The study estimated the BVOC emissions in 2018. The study year for the reference was added to the sentence.
Changes in manuscript:
Results and discussion (Lines 270-271 in the revision): However, results in this study are lower than 58.9 Tg for 2018 estimated by Li et al. (2020).
Technical corrections:
- Abstract: introduce abbreviations LAL and LC.
Response: Revised accordingly.
- 49: as in PRD …
Response: Revised accordingly.
- Grammar and tense issues: l.52/53: ‘quantifi(ed)/determine(d)’, l. 55 ‘affect(ing)’, l.99&l.111 &l.145&l.122 ‘were’ -> are
Response: Revised accordingly.
- 104 ‘major(ity)’
Response: Revised accordingly.
- Sentence in line 62 f. (‘Wang et al. (2018a) misses one verb
Response: Revised accordingly.
- Sentence in l. 71 has a complicated sentence structure, integrate the last part in the main sentence
Response: Revised accordingly.
- -(l.100) ‘of each LC type […] and [this]’
Response: Revised accordingly.
- 158 ‘based on five different cases’ -> with 5 different sets of LC input data for MEGAN
Response: Revised accordingly.
- Footnote in Table S3 not found, in caption: ‘The values are bolded without meeting the benchmarks’ -> ‘The values without meeting the benchmarks are bolded’
Response: Revised accordingly.
- S3: scale could be improved
Response: Revised accordingly.
- Table S5: units are not given in the caption, information on benchmarks in the table, distinction between MDO3 and MDA8 could be clearer
Response: Revised accordingly.
- 252: are occurred -> occur
Response: Revised accordingly.
- 260: shouldn’t be the sentence refer to Fig. S4 ?
Response: Revised accordingly.
- 344/45: the plots from the three different species are hard to compare since they have different scales
Response: The sentence was deleted.
- 347: Reformulate
Response: The sentence was deleted.
- 350: attribute -> be attributed
Response: Revised accordingly.
- 355: delete ‘amount of’
Response: Revised accordingly.
- 359: ‘temporal trend’ -> seasonal cycle ?
Response: Revised accordingly.
- 369: sometimes you mean both LAI and PFT when you write LC datasets
Response: We corrected the mistakes.
- 374-76: More comprehensive formulation needed for the conclusion
Response: Revised accordingly.
-
RC2: 'Comment on acp-2022-739', Anonymous Referee #2, 11 Jan 2023
Manuscript Review
In the manuscript titled “Impacts of land cover changes on biogenic emission and its contribution to ozone and secondary organic aerosol in China” the authors have analysed the impact of LAI and LC datasets on biogenic VOCs emission over China. The study has further attempted to understand its contribution to O3 and secondary organic aerosols by performing five different set of simulations using MEGAN and WRF-CMAQ models.
This study provides an interesting insight into the uncertainties in regard to the biogenic VOC emission input data and the impact it can have on air quality. However, there are a few minor comments which the authors can address in the revised manuscript.
Minor Comments:
- Page 1 Abstract Section: The results from the study should be stated for each experimental design before drawing out the inference as to which input dataset is best suited for the study region along with a sentence on the significance/impact of this study in the concluding statement of the abstract.
- Page 1 Line 17: Define LAI, LC, SOA and BSOA abbreviation in the Abstract.
- Page 1, Line 44 and 45: Change “isoprene emission ranged” and “monoterpene emissions ranged” to “isoprene emissions are ranged” and “monoterpene emissions are ranged”
- Page 1 Line 52 and 53: Either “quantified” to “quantify” and “determined” to “determine” or rephrase the sentence accordingly to fit the tense.
- Page 1 Line 55: Remove ‘that’
- Page 3 Line 93: Remove “land cover” from “land cover (LC)”. Abbreviation only needs to be defined at its first instance in the manuscript, authors need not define it again for every section.
- Page 4 Section 2.1 Data description: Mention of supplementary material Table S2 and Fig S2 comes before the Table or Figure S1. To ensure a chronical order to the flow of figures and tables in text and supplementary, authors should renumber the supplementary figures and tables as it appears in the manuscript.
- Page 4 Line 100-101: The text mentions the 16 PFTs classifications whereas, Figure 1 only shows 15 classifications. Correct accordingly.
- Supplementary Table S2 lists the sources of the LC and LAI datasets. The table should include more information to better describe all the datasets used in the study such as temporal resolution spatial resolution and years for which data is available etc.
- Page 4 Line 106: The description of Figure S2 in text seems to be Figure S4 in supplementary material. Check and correct.
- Page 5 Line 145: States the physical schemes adopted in WRF model. Authors are suggested to support the choice of these WRF physical parameterizations through literature review for the same study region.
- Page 5 Line 148: Number of grid cells of CMAQ model are given. Is the CMAQ model running at the same resolution as the WRF model? Provide horizontal spatial resolution for CMAQ model too.
- Page 5 Lines 150-153: The sources of anthropogenic emissions used for China and other countries is provided. Few sentences on MEIC and EDGAR emission inventories should be added to understand the reason behind selection of these inventories.
- Table S4 gives the benchmarks for some statistical indices for meteorological parameters. It is suggested to provide the reference of these benchmarks. Similarly, provide the reference for O3 benchmark as well.
- Page 7 Line 212: Define the term ‘LAIv’.
- Page 9 Line 260: “Fig S3” should be “S4”.
- Page 11 Lines 323-325: The results analyzed in Section 3.4.1, have a special mention to the BSOA concentrations noted over Sichuan basin. As it is also stated in the Conclusion Section on Page 12, it would be nice to mark the location of Sichuan basin on the spatial figures of both main manuscript and supplementary material.
Citation: https://doi.org/10.5194/acp-2022-739-RC2 -
AC2: 'Reply on RC2', Hongliang Zhang, 21 Feb 2023
Dear Referee #2,
We appreciate your comments to help improve the manuscript. We tried our best to address your comments and detailed responses and related changes are shown below and in the attached PDF file.
Comments: In the manuscript titled “Impacts of land cover changes on biogenic emission and its contribution to ozone and secondary organic aerosol in China” the authors have analysed the impact of LAI and LC datasets on biogenic VOCs emission over China. The study has further attempted to understand its contribution to O3 and secondary organic aerosols by performing five different set of simulations using MEGAN and WRF-CMAQ models.
This study provides an interesting insight into the uncertainties in regard to the biogenic VOC emission input data and the impact it can have on air quality. However, there are a few minor comments which the authors can address in the revised manuscript.
Response: Thanks for the recognition of our study. Below is the response to each specific comment.
Minor Comments:
- Page 1 Abstract Section: The results from the study should be stated for each experimental design before drawing out the inference as to which input dataset is best suited for the study region along with a sentence on the significance/impact of this study in the concluding statement of the abstract.
Response: Thanks for your comments. A brief summary of model validation and the sentence about the significant of this study were added to the Abstract.
Changes in manuscript:
Abstract (Lines 23-31 in the revision): Changing the LC inputs for the MEGAN model has a more significant difference in BVOC estimates than using different LAI datasets. The C4 case has better model performance, indicating that it is the better choice for BVOC estimations in China. Changing the MEGAN inputs further impacts the concentrations of O3 and SOA. The highest O3 and biogenic SOA (BSOA) concentrations appear in the C1 (using GLASS and MCD12Q1 LC) simulation, which can reach 12 ppb and 9.8 μg m-3, respectively. Due to the combined effect of local BVOC emissions and the summer monsoon, the relative difference between C1 and C4 is over 52% and 140% in O3 and BSOA in central and eastern China. The BSOA difference between C1 and C4 is mainly attributed to the isoprene SOA (ISOA), which is a major contributor to BSOA. Particularly, the relative difference in ISOA between these two cases is up to 160% in eastern China. Therefore, our results suggest that the uncertainties in MEGAN inputs should be fully considered in future O3 and SOA simulations.
- Page 1 Line 17: Define LAI, LC, SOA and BSOA abbreviation in the Abstract.
Response: The abbreviations of these terms were added to the Abstract.
Changes in manuscript:
Abstract (Lines 14-15 in the revision): “In this study, Model of Emissions of Gases and Aerosols from Nature (MEGAN) v2.1 was used to investigate the impact of different leaf area index (LAI) and land cover (LC) …”
- Page 1, Line 44 and 45: Change “isoprene emission ranged” and “monoterpene emissions ranged” to “isoprene emissions are ranged” and “monoterpene emissions are ranged”:
Response: Revised accordingly.
Changes in manuscript:
Abstract (Lines 45-47 in the revision): Global annual inventories of the isoprene emission are ranged from 500 to 750 Tg yr-1 (Guenther et al., 2006) and those of monoterpene emissions are ranged from 74.4-157 Tg yr-1 (Guenther et al., 2012;Messina et al., 2016).
- Page 1 Line 52 and 53: Either “quantified” to “quantify” and “determined” to “determine” or rephrase the sentence accordingly to fit the tense.
Response: Revised accordingly.
Changes in manuscript:
Introduction (Lines 53-54 in the revision): Therefore, it is necessary to quantify the influence of those factors and determine the bias in BVOC emissions.
- Page 1 Line 55: Remove ‘that’.
Response: Revised accordingly.
Changes in manuscript:
Introduction (Lines 56-57 in the revision): Land cover (LC), including leaf area index (LAI) and plant function types (PFTs) fractions, is a major factor affecting the BVOC emissions in the MEGAN model.
- Page 3 Line 93: Remove “land cover” from “land cover (LC)”. Abbreviation only needs to be defined at its first instance in the manuscript, authors need not define it again for every section.
Response: Revised accordingly.
Changes in manuscript:
Methodology (Lines 136 in the revision): “Three LC datasets were applied as PFTs inputs…”.
- Page 4 Section 2.1 Data description: Mention of supplementary material Table S2 and Fig S2 comes before the Table or Figure S1. To ensure a chronical order to the flow of figures and tables in text and supplementary, authors should renumber the supplementary figures and tables as it appears in the manuscript.
Response: Revised accordingly.
Changes in manuscript:
Methodology (Lines 118, 122, 142, and 143 in the revision):
“…China and its surrounding countries in East Asia (Fig. S1) …”
“… Table S1 briefly lists the physical options used for the WRF model.”
“Sources of these products were listed in Table S2.”
“Three LC maps are first re-gridded to the CMAQ domain (Fig. S2).”
- Page 4 Line 100-101: The text mentions the 16 PFTs classifications whereas, Figure 1 only shows 15 classifications. Correct accordingly.
Response: Revised accordingly.
Changes in manuscript:
Methodology (Lines 144-145 in the revision): Lastly, eight vegetation types are further reclassified into CLM-15 PFTs based on the climate rules described in Bonan et al. (2002).
- Supplementary Table S2 lists the sources of the LC and LAI datasets. The table should include more information to better describe all the datasets used in the study such as temporal resolution spatial resolution and years for which data is available etc.
Response: Thanks for your suggestions. The temporal resolution, spatial resolution and available years for satellite datasets used in this study were added to the Table R1 (named as Table S2 in the revision). Detail see attachment.
Changes in supplementary material:
Table R1 see attachment.
- Page 4 Line 106: The description of Figure S2 in text seems to be Figure S4 in supplementary material. Check and correct.
Response: Revised accordingly.
Changes in manuscript:
Methodology (Lines 148-150 in the revision): Although MCD12Q1 and CGLS LC both show a large area of broadleaf tree in central and southern China, the area fraction of broadleaf tree in CGLS LC is higher than that in MCD12Q1 (Fig 1 and Fig. S3).
- Page 5 Line 145: States the physical schemes adopted in WRF model. Authors are suggested to support the choice of these WRF physical parameterizations through literature review for the same study region.
Response: Thanks for your suggestions. The references were added to the manuscript.
Changes in manuscript:
Methodology (Lines 121-122 in the revision): The model configurations are similar to the previous studies (Wang et al., 2018;Wang et al., 2020;Zhu et al., 2021) and Table S1 briefly lists the physical options used for the WRF model.
- Page 5 Line 148: Number of grid cells of CMAQ model are given. Is the CMAQ model running at the same resolution as the WRF model? Provide horizontal spatial resolution for CMAQ model too.
Response: Yes, it is. The horizontal spatial resolution for the Community Multiscale Air Quality Modelling System (CMAQ) model was added to the sentence and shown below.
Changes in manuscript:
Methodology (Lines 125-126 in the revision): The CMAQ model used the same horizontal resolution as WRF with a horizontal domain of 197 × 127 grid cells. This domain covers China and its surrounding areas (Fig. S1).
- Page 5 Lines 150-153: The sources of anthropogenic emissions used for China and other countries is provided. Few sentences on MEIC and EDGAR emission inventories should be added to understand the reason behind selection of these inventories.
Response: Thanks for your comments. The reasons for choosing MEIC and EDGAR were added to the manuscript.
Changes in manuscript:
Methodology (Lines 127-134 in the revision): The anthropogenic emissions of China used the datasets from Multiresolution Emission Inventory for China (MEIC; available at http://www. meicmodel.org, last access:3 May 2022). Since the MEIC only provides anthropogenic emissions for China, anthropogenic emissions from foreign countries were provided by the Emissions Database for Global Atmospheric Research (EDGAR) v4.3 (available at http://edgar.jrc.ec.europa.eu/overview.php? v=_431, last access: 10 May 2022). The MEIC inventory is widely used in air quality studies in China (Li et al., 2017;Hu et al., 2016;Wu et al., 2020). It had an improvement in a vehicle emission inventory with high resolution (Zheng et al., 2014), and a non-methane VOC mapping approach for different chemical mechanisms (Li et al., 2014). The EDGAR is a grided emissions inventory with a high horizontal resolution of 0.1°×0.1° (Saikawa et al., 2017).
- 14. Table S4 gives the benchmarks for some statistical indices for meteorological parameters. It is suggested to provide the reference of these benchmarks. Similarly, provide the reference for O3 benchmark as well.
Response: The references for the benchmarks were added to the Table S4 and Table S5.
Changes in supplementary material:
“Note: * are benchmarks limits suggested by (Emery et al., 2001).”
“Note: * are criteria suggested by EPA (2007).”
- Page 7 Line 212: Define the term ‘LAIv’.
Response: LAIv means the LAI of vegetation covered surface, which is calculated by dividing the grid average LAI with the fraction of grid that is covered by vegetation (Guenther et al., 2006). This term has been explained in the manuscript in Line 161.
Changes in manuscript: No changes were made for this comment.
- Page 9 Line 260: “Fig S3” should be “S4”.
Response: Revised accordingly.
Changes in manuscript:
Results and discussion (Lines 261-263 in the revision): The spatial distribution of isoprene emission in C5 is conspicuously different than in C1, which is consistent with a difference in the broadleaf tree distribution in the inputs (Fig. S5).
- Page 11 Lines 323-325: The results analyzed in Section 3.4.1, have a special mention to the BSOA concentrations noted over Sichuan basin. As it is also stated in the Conclusion Section on Page 12, it would be nice to mark the location of Sichuan basin on the spatial figures of both main manuscript and supplementary material.
Response: Thanks for your comments. We marked the location of the Sichuan basin in the Figure R1 (named as Figure S1 in the revision). Detail see attachment.
Changes in supplementary material:
Figure R1 see attachment.
Jinlong Ma et al.
Jinlong Ma et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 369 | 126 | 14 | 509 | 35 | 3 | 10 |
- HTML: 369
- PDF: 126
- XML: 14
- Total: 509
- Supplement: 35
- BibTeX: 3
- EndNote: 10
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1