the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 25 Jun 2024
Unheralded contributions of biogenic volatile organic compounds from urban greening to ozone pollution: a high-resolution modeling study
Abstract. Urban Green Spaces (UGS) are widely advocated for mitigating urban atmospheric environment. However, this study reveals that it can exacerbate urban ozone (O3) levels under certain conditions, as demonstrated by a September 2017 study in Guangzhou, China. Utilizing the Weather Research and Forecasting Model with the Model of Emissions of Gases and Aerosols from Nature (WRF-MEGAN) and the Community Multiscale Air Quality (CMAQ) model with a high horizontal resolution (1 km), we assessed the impact of UGS-related biogenic volatile organic compound (BVOC) emissions on urban O3. Our findings indicate that UGS-BVOC emissions in Guangzhou amounted to 666.49 Gg, primarily from isoprene (ISOP) and terpenes (TERP). These emissions contribute ~30 % of urban ISOP concentrations and their incorporations to the model significantly reduce the underestimation against observations. The study shows improvements in simulation biases for NO2, from 7.01 µg/m3 to 6.03 µg/m3, and for O3, from 7.77 µg/m3 to -1.60 µg/m3. UGS-BVOC and UGS-LUCC (land use cover changes) integration in air quality models notably enhances surface monthly mean O3 predictions by 3.6–8.0 µg/m3 (+3.8–8.5 %) and contributes up to 18.7 µg/m3 (+10.0 %) to MDA8 O3 during O3 pollution episodes. Additionally, UGS-BVOC emissions alone increase the monthly mean O3 levels by 2.2–3.0 µg/m3 (+2.3–3.2 %) in urban areas and contribute up to 6.3 µg/m3 (+3.3 %) to MDA8 O3 levels during O3 pollution episodes. These impacts can extend to surrounding suburban and rural areas through regional transport, highlighting the need for selecting low-emission vegetation and refining vegetation classification in urban planning.
- Preprint
(21726 KB) - Metadata XML
-
Supplement
(2433 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-1163', Anonymous Referee #3, 31 Oct 2024
General Comments
This paper usefully demonstrates the impacts of updating biogenic VOCs in urban environments (called here UGS-BVOC) as well as updating land use cover data to high resolution maps (called here UGS-LUCC). The impact of each of these changes separately and together on ozone in Guangzhou, China is thoroughly explored. These emissions and model improvements are clearly important for accurately simulating ozone in this region. I recommend publication after major revisions specified below.
Major comment 1: While VOCs are important for ozone production, so are nitrogen oxides. This paper could be mis-leading to policy makers if NOx is not also mentioned as an ozone precursor more clearly in the text.
Major comment 2: Why does the abstract and throughout the text use ug/m3 as the units for gases in the atmosphere. The convention is typically ppb in atmospheric chemistry. Please provide a valid reason or use the conventional unit. Additionally, throughout the paper the authors switch from ug/m3 and ppb throughout the text and are not consistent with one unit. This is confusing for the reader. I would highly recommend switching all units to ppb for all gases.
Major comment 3: The use of urban as both an “urban region” in Figure 1 and “urban landcover type” to mean two different things makes the paper hard to understand at first. This paper selects locations classified with the urban landcover type as urban and this is called UGS (urban green space), but then also classifies by region into 3 types called urban, suburban, and rural. Providing more detail on how the regions are determined in Figure 1 and lines 202 – 206 would be useful. And please consider calling this something other than “urban”, so that readers can more easily differentiate when the authors are referring to urban as a region and or as a landcover type. Perhaps, terminology like “city center”, “suburban”, and “rural” could work.
Major comment 4: There is very little comparison to isoprene observations even though this is the major development in this paper (Table 2 and line 235). It is an excellent opportunity that the model can be compared against these two monitoring sites for isoprene. For reader, clarity, can you please add these two monitoring sites to Figure 1? Looking at the average isoprene concentrations over the entire campaign (Sept – Nov) is not sufficient proof that the isoprene emissions have improved. Also I am confused on the time periods. The observations appear to be for Sept 20 – Nov 20, but in the methods section you state the model is only run for September 1 – Sept 30 (ignoring spin-up). Are you comparing the observations and model during the same time period? It is very important to compare the model and observations over the exact same time period since isoprene emissions are very seasonally dependent. Can you provide additional statistics like you do for the meteorological variables in Table S2 and ozone and NO2 in Table 3? Further evaluation of the diurnal cycle in the model compared to the observations and statistics for the different months would also be very useful in understanding whether the isoprene emissions have improved in the model. Since your main updates are isoprene emissions it is very important that this evaluation is clear on what has been done and that this is a good evaluation to ensure that the updates the author has made are robust.
Can you also provide more information on how the observations were collected at these monitoring sites? What instrument technique? Were any interferences considered if the instrument technique was a PTR (Coggon, et al., 2024 - https://doi.org/10.5194/amt-17-801-2024)? Was the diurnal cycle of the observations consistent with known chemistry of isoprene where isoprene concentrations rise during the day and fall rapidly at night in high NOx urban locations with significant amounts of NO3 radical? This can help you verify that any interferences for isoprene in your instrument technique are appropriately accounted for.
Specific Comments
Title: “Unheralded” here in the title seems to suggest that no study before has attributed biogenic VOCs from urban greening to ozone production before, but there are many past studies that are cited in the introduction that have also concluded this. Could the authors choose a different word that better reflects the advancements and scope of their work?
Line 21 – Can you expand more on what you mean by “advocated for mitigating urban atmospheric environment”? I’m not sure what this means.
Table 3 – I believe that this is an hourly comparison between O3 and NO2. If so, can you add hourly to the title or table description for clarity to the reader.
Line 272 and Table 4 – You mention MDA8 O3 here in the text? Is Table 4 MDA8 O3 or hourly O3? If possible, calculating these statistics on hourly O3 and MDA8 O3 is the most useful for both Table 3 and 4. Hourly O3 especially R helps understand if you have represented the diurnal cycle well. And MDA8 ozone is useful to just investigate daytime ozone when O3 is highest for regulatory applications. I would recommend calculating both MDA8 O3 and hourly O3 for both tables here.
Line 288 – By “monoethylene” here, do you mean “monoterpene”. Is this a typo?
Line 297 – See major comment. I would suggest renaming urban in figure 1 and throughout the text including in this paragraph to something else so that readers can clearly differentiate between urban landcover type and what you are classifying as an urban region.
Line 381 – Can you verify this statement “high-resolution land use cover data increase the estimation of the UGS-BVOC emissions in the urban and suburban regions.” This is different from the text above and Figure 2A.
Throughout 3.2, and all figures and tables therein, can you be clearer in the table and figure descriptions and throughout the text that these are all evaluated over September. I assume this is the case, but it would make it easier for the reader to state this clearly in the table and figure headings and descriptions.
Lines 397 – 400 – These are very strong statements, but it’s hard to confirm that these statements are really accurate without also plotting the absolute concentrations of MDA8 ozone and how much the UGS-BVOC and UGS-LUCC contribute to the total. Your conclusions here would be more impactful, if you plotted the total MDA8 ozone and then the contribution from UGS-BVOC and UGS-LUCC and where the regulatory metric is for MDA8 ozone as well. I would suggest showing a plot of your total MDA8 as well.
Figure 6 – If this is an average over September of the differences? Can you add this to the Figure description for clarity to the reader.
Figure 7A – Why plot your base case scenario (Gdef_N) here for MDA8 ozone. Why not your improved scenario (Ghr_Y) or at least both?
Line 459 – What do you mean by “cumulative effects” here?
Line 519 – I’m not sure these are what everyone would consider significant changes in ozone. Please add the percent changes in ozone in parenthesis here, so that readers can confirm your definition of significant is similar to their definition.
Data availability – ACP journals require at least model code be made public (https://www.atmospheric-chemistry-and-physics.net/policies/data_policy.html). Please upload your WRF-Chem model code to be available online.
Technical Corrections
Line 210 need a space between Table 1 and were
Line 507 – not a complete sentence
Line 520 – typo with a comma instead of a period.
Citation: https://doi.org/10.5194/egusphere-2024-1163-RC1 -
RC2: 'Comment on egusphere-2024-1163', Anonymous Referee #4, 01 Nov 2024
This manuscript describes a numerical modeling study (WRF-CMAQ-MEGAN) to investigate the impact of urban greening on air pollution. This is an important topic that has been the subject of quite a few papers recently but there is certainly more that needs to be done to adequately address this topic. The authors use a relatively high-resolution (1 km) modeling system which is appropriate for advancing our understanding on urban BVOC impacts. They compare their results with observations which is helpful for seeing if there are improvements. They then present three scenarios (landcover data resolution, BVOC emission estimates, urban green space amount) to a base case to show the impacts of each one. This is a useful study but I have some concerns about the manuscript that should be addressed before accepting this paper for publication in ACP.
General comments:
The introduction (Section 1) has statements that do not seem to be supported by the references given and these references do not appear to be very relevant for urban green spaces anyway. Examples include lines 63-64, 67-71, 71-73, 80-81, etc.
The comparison between the “default” low resolution run and the higher resolution data there are other differences in these runs and they do not really show what is the impact of 1) the resolution of the data used to derive landcover and 2) the resolution of the model simulations. It would be more useful to show the individual impacts of these two differences. Also, while it is reasonable to assume that 10-m landcover data is more appropriate than 1-km landcover data, there are still uncertainties associated with the 10-m data. There should be a detailed discussion of the uncertainties for each landcover input (LAI, growth form fractions, ecotypes and ecotype-specific emission factors) and how this compares with 1-km data.
The comparison with ambient isoprene observations is a valuable addition to this study but it should be extended by discussing the issues with this limited comparison including how these concentrations are influenced by model emissions, dilution, chemical losses, etc which have major uncertainties. How do uncertainties compare with the differences in the results for the different scenarios? Also, include a description of the BVOC data used to evaluate the model and an assessment of the quality of the data.
A major missing component is a thorough discussion on what these findings can tell us about urban BVOC and air pollution- more than ozone will increase if there is sufficient NOx which is already well known. This needs to include a discussion of NOx as well as other complexities. Make it clear whether these findings are specific for Guangzhou or are also relevant for other cities. Discuss limitations such as not having representative BVOC emission factor data for these landscapes (that requires tree species composition data and accurate tree species-specific emission factors). Also, what are the uncertainties associated with growth form estimates. If the UGS is mostly grass then that will have a much lower emission than trees.
The authors discuss the impacts of LUCC on temperature and solar radiation- the temperature “heat island” is well known but the impact on solar radiation is less clear. What are the processes in the model and are they realistic? By “urban region receives less solar radiation than other regions likely due to the shading effect of urban canopies” do you mean that you are using the ground surface temperature and light to drive BVOC emissions? The ground surface values are not the light and temperature that you should be using to drive the BVOC emissions. It should be the canopy light and temperature. Reassess whether the impacts you are seeing are influenced by the model is using the wrong light and temperature.
Ozone responds to temperature for multiple reasons. Quantify the impact of BVOC emission response to temperature relative to these other reasons.
Summarize the data shown in section 3.3 in a table that provides an overview of these results
The manuscript would benefit from editing for English language usage.
Specific comments:
Line 1: “Unheralded” is not justified for the title and should be deleted.
27: In addition to the 666 Gg BVOC value for Guangzhou, include the emission per area and also how BVOC compare to AVOC.
28: What does it mean 30% of urban ISOP? Is the other 70% anthropogenic isoprene or is it other biogenic isoprene. Along these lines, is UGS BVOC all BVOC? Or are there BVOC from other vegetation such as street trees which are generally not thought of as green space vegetation.
36: how does this study highlight the need for selecting low-emission vegetation and refining vegetation classification? I see little here that provides specific guidance to air quality managers.
63-64: Clarify the point being made regarding the importance of dispersion over deposition. What do the references say that actually support this?
Line 126 to 129: This is not the default LAI for MEGAN3.1. The LAI data referred to here is probably the MEGAN2.1 LAI data. It can be used with MEGAN3.1 but there is no default MEGAN3.1 LAI data. It is expected that users generate their own LAI data for any MEGAN 3.1 simulation.
Line 155-164: These equations and text describe MEGAN2.1, not MEGAN3.1.
Line 310-311: How do you define “natural area” vs “UGS”.
Line 398- this is not clear. Please reword this sentence
Line 399-400: If there is a “revelation” here then you should make it clear how managers can use this and what they should be doing
486: replace “was” with “is” unless you are referring to a specific episode (in which case make that clear)
488-490: This meaning of this sentence is not clear.
27 and 497: should note that isoprene and terpenes are only half of the total. “primary” could lead readers to think they are almost all of it.
498: are there UGS outside of urban areas? I would expect that they should all be urban.
505-506: Are you saying that managers should not plant EBTs? But some EBTs have low emissions- lower than other vegetation. They should not be all grouped together.
507-508- should also point out that there are other possibilities for ozone underestimation.
Citation: https://doi.org/10.5194/egusphere-2024-1163-RC2 -
RC3: 'Comment on egusphere-2024-1163', Anonymous Referee #5, 06 Nov 2024
This paper uses high-resolution WRF-CMAQ simulations to investigate the impacts of urban greening in Guangzhou, China through the use of different resolution input datasets on land cover and leaf area indices. The impact of these different model configurations is usefully explored through comparisons with observations. Representing urban green spaces in models is important for understanding ozone atmospheric chemistry in urban settings. This study is relevant to ACP. I recommend publication of this manuscript after the concerns specified below have been addressed.
General Comments
The manuscript discusses the use of different resolution input datasets and uses a high resolution, nested model. It would be useful if the authors could include a section on the benefits and limitations of using high resolution. For example, you have 1 km versus 10 m land cover data, but when running the model, you have all of this information in a 1 km grid box. How do these high-resolution datasets impact the calculations of BVOCs in the model? What are the uncertainties here?
Although the authors compare the simulations to observations of meteorological parameters, O3, NO2, and isoprene, there is very little discussion on how these observations were measured and the time resolution. It would be helpful to readers to elaborate more on this.
In this manuscript, the authors are addressing the sensitivity of O3 to changes in BVOCs, but there is little discussion on whether O3 is NOx-limited or VOC-limited in Guangzhou. There is also no discussion on other potential sources affecting the area. The authors mention that there is rapid urbanization happening in Guangzhou. Does this mean that there is more industry, more vehicles, etc. that could be impacting O3 production and loss processes? O3 is nonlinear, and depending on the regime, and other factors like emissions, the composition and ratios of VOCs and NOx in the area will affect O3 differently.
There is an inconsistency with the units, which can be confusing to readers. At times, the authors use “ppb” and at other times they use “µg/m3” for gases. Please use a consistent unit throughout the paper.
For the tables and figures in the paper, please specify the timeframes that the concentrations are averaged over.
Specific Comments
Figure 1: This is a paper based on Guangzhou. For readers who might not be familiar with the area, could the authors provide more description on the urban, suburban, and rural regions? Is there much vegetation in the urban centers?
Section 2.2: Did the authors use online or offline MEGAN? It seems to me that you are referring to MEGANv2.1, not v3.1. Please clarify, as the resolution of the model can have impacts on online emission calculations.
Section 2.3 and onwards: I would recommend that the authors refrain from using the term “scenario” to describe the different simulations. These are model configurations, whereas the term “scenario” tends to highlight a potential future or possible outcomes.
Section 2.4: What kind of instrumentation was used to measure O3?
Table 2: Are these monthly concentrations? Please clarify in the text or in the caption.
Line 260: By “specified requirement”, do the authors mean the observed values? If so, I would refer to them as that because the authors are not referring to a standard of sorts.
Line 265-267: The authors mention that integrating UGS-BVOC can improve the accuracy of NO2 predictions. Can the authors elaborate on why this improvement in simulated NO2 happens?
Table 3: Again, what time frame are these concentrations averaged over? Monthly?
Table 4: Same comment as tables 2 and 3.
Line 288: Do you mean “monoterpene” instead of “monoethylene”?
Line 473-475: This sentence is rather confusing. Can you clarify how an increase and simultaneous decrease leads to an overall increase in wind speed?
Technical Corrections
Technical Correction 1: At the beginning of the paper, the authors establish an abbreviation convention, such as TERP for monoterpenes, ISOP for isoprene, LUCC for land use cover change, etc. Throughout the paper, I found that the abbreviations were being described again on several occasions (i.e., line 288, line 277-278). Please make this consistent throughout the paper.
Technical Correction 2: When naming figures throughout the paper, the authors should use the naming convention of “Figure 1A and 1B” rather than “Figure 1 (A) and (B)”.
Line 74: “the complexity of these interactions, and they demonstrated that vegetation could exert nonlinear effects on” – Remove “they” from sentence
Line 97: Should be “investigated” instead of “investigating”
Line 210: Space between “Table 1” and “were”
Citation: https://doi.org/10.5194/egusphere-2024-1163-RC3
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 450 | 117 | 16 | 583 | 50 | 13 | 11 |
- HTML: 450
- PDF: 117
- XML: 16
- Total: 583
- Supplement: 50
- BibTeX: 13
- EndNote: 11
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1