the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 12 Feb 2024
Changes in South American Surface Ozone Trends: Exploring the Influences of Precursors and Extreme Events
Abstract. In this study, 21st-century ground-level ozone trends and its precursors in South America were examined, which is an understudied region where trend estimates have rarely been comprehensively addressed. Therefore, we provided an updated regional analysis based on validated surface observations. We tested the hypothesis that the recent increasing ozone trends, mostly in urban environments, resulted from intense wildfires driven by extreme meteorological events impacting cities where preexisting volatile organic compound (VOC)-limited regimes dominate. We applied the quantile regression method to estimate trends, quantify their uncertainties, and detect trend change points. Additionally, the maximum daily 8-hour average (MDA8) and peak-season metrics were used to assess present-day short- and long-term exposure levels (2017–2021). Our results showed lower levels in tropical cities (Bogotá and Quito), varying between 39 and 43 ppbv for short-term exposure and between 26 and 27 ppbv for long-term exposure. In contrast, ozone mixing ratios were higher in extratropical cities (Santiago and São Paulo), with a short-term exposure level of 61 ppbv and long-term exposure levels varying between 40 and 41 ppbv. Santiago (since 2017) and São Paulo (since 2008) exhibited positive trends of 0.6 and 0.3 ppbv yr-1, respectively, with very high certainty. We attributed these upward trends, or no evidence of variation, such as in Bogotá and Quito, to a well-established VOC-limited regime. However, we attributed the greater increase in the extreme percentile trends (≥ 90th) to heat waves and, in the case of southwestern South America, to wildfires associated with extreme meteorological events.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(12836 KB)
-
Supplement
(8833 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(12836 KB) - Metadata XML
-
Supplement
(8833 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-328', Anonymous Referee #1, 15 Mar 2024
review of manuscript egusphere-2024-328
Title: Changes in South American Surface Ozone Trends: Exploring the Influences of Precursors and Extreme Events
Author(s): Rodrigo J. Seguel et al.
General comments:
The manuscript presents a comprehensive analysis of the distribution and trends in long-term ozone and ozone precursor observations in cities and background locations in South America. While mainly European and North American ozone records are extensively studied and its interpretation can be found in the peer-reviewed literature, such studies are rather limited for South America. Therefore, the present manuscript provides a valuable contribution to the understanding of ozone trends in this less studied region.
The determination of the trends and the change points of the trends is sound. I would have just liked to learn more about the underlying data (analytical methods, quality control, screening, …) since the quality of the data is a crucial requirement for the analysis.
The paper will fit well into the TOAR-II Community Special Issue. See below a few specific comments that should be addressed prior to publication.
Specific comments:
Line 22-24: not clear which metric the numbers are referring to.
Line 25: reader does not know yet how short-term and log-term exposure levels are defined. Add some information from lines 101 ff.
Line 26: trends refer to which metric?
Line 54: replace "a chemical regime […] has been established …" by "a chemical regime […] has been found …"
Lines 76-79: this sentence reads like a part of the conclusions. Move it below?
Lines 83 ff.: add some brief description of the measurement techniques. All UV absorption for O3? chemiluminescence for NO and NO2? If measurements are done by regulatory networks, I assume that NO2 was converted to NO prior to detection with heated surface (molybdenum) converters. It is known that these converters overestimate the NO2 mole fractions, especially in rural areas. This contribution may change over time when the amount of oxidized nitrogen species decreases. CO measurements with NDIR?
Lines 86-87: please elaborate on the data screening performed by the authors. How was drift (trends in the instruments' response, I suppose) and representativeness assessed? A 75% data coverage criterion is mentioned below. This should be added here. Did you also exclude other data such as outliers, periods with very little or very large noise, … or was the quality of the received data just good. Please add how many data/datasets were rejected prior to your analysis. Lines 189-191 provide some of this information. Still the "quality control test established in the methodology" remains unclear.
Figure 2, caption: "The black dots denote the monitoring stations that do not meet the data quality criteria." I do not see any back dots.
Lines 163 ff.: did I get it right? You attribute the lower O3 levels in Quito to intense vertical mixing that mixes (less O3-rich) air from the free troposphere to the site. Is there no signature from stratospheric intrusions seen at this elevation?
Figure S2: caption reads trend in ppb/yr while ppm/yr is shown in the figure.
Chapter 3.3: for the interpretation of the O3 change points along with the trends of the precursors. I wonder if you looked into the hourly data and the trends of the different percentiles there. At many location worldwide, it is often seen that the lowest values (percentiles) do show a positive trends (due to the reduction in NO and less O3 titration) while the highest values show negative trends.
Lines 299-300: "… These measurements have been accompanied by an increase in ozone since 2008." This refers to my comment just made above. You could doublecheck if you see that in all of your data, too. "… increase in ozone since 2018 …" Which metric are you referring to?
Lines 350-351. "We attributed these observed ozone trends to […] the establishment of volatile organic compound-limited regimes.". This looks like a firm statement that might require some more (model) analysis.
Lines 363-364: "… the lack of quality control, which prevents the inclusion of additional existing measurements." Do you refer to the quality of the measurements here? As a group of South American scientists /experts in high-quality observations, do you have any suggestion to improve the situation? Training, workshops, development of common standards (if not available), development of common tools for quality control, …
Citation: https://doi.org/10.5194/egusphere-2024-328-RC1 - CC1: 'Comment on egusphere-2024-328', Owen Cooper, 18 Mar 2024
-
RC2: 'Comment on egusphere-2024-328', Anonymous Referee #2, 18 Mar 2024
General Comments
This paper reports on surface ozone trends in South America. Data from different locations are shown together with ozone precursor data, trends are reported and reasons for the trends are discussed with the help of the precursor data. Since publications of long time series of station data in South America are rare, the manuscript should be published after these questions have been answered:
In line 152 the authors write:
In these cities, a significant fraction of ozone precursors is emitted by vehicular fleets and has decreased according to air quality control measures such as the introduction of better fuel quality, sulfur content reduction, enforcement of threeway catalytic converters, stricter emission standards for new fleet vehicles and mandatory periodic technical inspection for inuse vehicles.
What is the impact of sulfur content reduction in fuel on ozone trends ?
In Figure 3, small boxes in the left panel could indicate where the zoomed regions in the right panel are.
In Table 2, the authors divide the MDA8 and peak season data into data for 2012-2016 and data for 2017-2021. The reason for this is not clear to me. Would it be possible to treat the MDA8 and peak season data like the datasets in Table 3 and calculate turning points, p and SNR values?
In line 210 the authors write: Regardless of the latitude of each large city analyzed, each urban agglomeration contains subdivisions with high-certainty positive ozone trends.
This is difficult to see from the data in Table 3: in the Bogotá region, the trend is either positive or negative, depending on whether you look at the 5th, 50th or 95th percentile. Also, this is in contradiction to what the authors write in line 243: The ozone mixing ratios in Bogotá showed no evidence of reduction or increase during the last decade despite efforts to reduce primary pollutant emissions, as shown in Figure 4a.
I would suggest that the author rephrase the sentence or indicate which data set the authors are referring to.
Table 3 is central to the manuscript. For better readability, I would suggest including a column where it is easy to see whether a trend is certain according to the criteria in Table 1, e.g. very high certainty, high certainty…
In line 224 the authors write: the trend observed after 2014 was likely impacted by the COVID-19 pandemic in 2020 and possibly in 2021 (Putero et al., 2023).
This would suggest a change point in 2019/2020.
Can this be seen in the data?
In line 239 the authors write: In general terms, we note that these ozone precursor abatement measures have been implemented, ignoring the VOC-to- NOx ratio, suggesting that ozone increases once the VOC-limited regime is reached. The latter, together with the extensive wildfires around the cities studied, could explain the occurrence of trend change points at some sites. This should be discussed more in detail.
The authors write that ozone trend is determined by ozone production. What is the role of titration effects, e.g. an increase in ozone concentration due to a decrease in NO mixing ratio?
The CO trend reflects the trend of VOCs originating from combustion. What is the role of biogenic VOCs in ozone formation here?
A change in NOx or VOC concentration would have an opposite effect on ozone production, depending on whether the chemical regime is VOC-limited or NOx -limited. Is there any indication of whether ozone production is VOC-limited or NOx -limited?
If wildfires play a role. I would expect different trends in different seasons? Has this been studied?
In line 244 the authors write: However, in the northern area of the city, which is impacted by ozone formation in higher proportions, the median ozone trend decreased at a rate of -1.01ppb yr-1 (high certainty) between 2008 and 2013.
Can it be shown that the northern area is more affected by ozone precursors? I would expect the ozone trend of the whole region to be shown and compared with the ozone trend of the northern region. The same comparison should be made for precursors.
In line 264 the authors write: Overall, in the Quito NOx-saturated environment, decreases in NOx precursors were anticorrelated with increases in ozone.
This sentence is unclear to me as it is not clear what NOx saturation means. Is ozone production limited by VOC? Does it mean that O3 is removed by titration with NO? Also, the term NOx precursors is not clear. Are the authors referring to the sources of NOx, e.g. that NOx emissions have decreased?
In line 264 the authors write: decreases in NOx precursors were anticorrelated with increases in ozone.
Shouldn´t it be: NOx mixing ratio was anticorrelated to the ozone mixing ratio.
In line 267 the authors write: This change generally coincided with the time series period when the ozone trend stopped decreasing, leading to a change point.
However, the change point was for ozone was in 2011, whereas the change points for NOx and CO were in 2013 and 2014. Is this within the uncertainty of the change point determination?
In line 267 the authors write: The implementation of this policy probably shifted the composition and proportion of precursors, especially during the morning.
Can this be shown? Here I would expect that the trend should be more significant when looking only at the values during the morning hours at polluted sites.
In line 275 the authors write: Ozone in Santiago decreased for nearly two decades due to public policies focusing mainly on curbing particulate matter.
This implies that PM and O3 have the same sources. Can this be shown? Do PM and O3 have the same trend?
In line 282 the authors write: In other words, until 2017, the policies effectively lowered the highest ozone percentiles.
From the plots it looks that ozone was low because NOx was high and NO reacted with ozone to convert ozone into NO2. The question is what happened in 2017? Did the policy change or was NOx so low that less ozone was removed by titration?
In line 303 the authors write: Notably, many higher anomalies occurred in the warmer months (Jan-Feb) and were more frequent after the ozone change point in 2008 (Figure 7a).
This is not readily apparent from the figure. The orange dots indicating the beginning of the year are on either side of the trend line. Perhaps it is possible to specify where the trend manifests itself.
In line 304 the authors write: As a result, the ozone trends at the 90th and 95th percentiles increased
I would suggest writing: accordingly or correspondingly (as this is not a result, but a different way of presenting the results)
In line 326 the authors write: In this regard, extreme positive ozone anomalies were observed in January 2017 (7.6 ppb) and February 2023 (8.6 ppb), caused by ozone and precursors transported from areas affected by intense wildfires (Fig 6a).
Can these datapoints be shown in the figure?
In line 342 the authors write: Short-term (MDA8) and long-term (peak-season) exposure metrics calculated for the present day (2017-2021) revealed latitudinal differences in South America.
However, in lines 343 and following, the authors argue that there are several factors, and that latitude plays only a minor role. I would suggest omitting the word "latitudinal" here.
In line 350 the authors write: We attributed these observed ozone trends to a greater decrease in nitrogen oxides than in carbon monoxide, which resulted in the establishment of volatile organic compound-limited regimes.
The arguments in favour of this statement, as set out in line 239, are in my opinion too weak, see comments on line 239.
In line 366 the authors write: Finally, our results revealed signs of a climate penalty for ozone in South America
How would this climate penalty be reflected in the data? Is this related to the more frequent forest fires? This could then be mentioned again here.
In line 366 the authors write: and identified extratropical zones as those where the increase in ozone poses the highest risk.
As the authors also argue, latitude is only one of several factors that determine the ozone trend, in addition to local measurements as in Quito, Santiago, and São Paulo or the local wind system as in Bogota, see also line 150 ff. A dependence of the ozone trend on the latitude would have to be shown specifically on the existing data.
Citation: https://doi.org/10.5194/egusphere-2024-328-RC2 - AC1: 'Comment on egusphere-2024-328', Rodrigo Seguel, 01 May 2024
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-328', Anonymous Referee #1, 15 Mar 2024
review of manuscript egusphere-2024-328
Title: Changes in South American Surface Ozone Trends: Exploring the Influences of Precursors and Extreme Events
Author(s): Rodrigo J. Seguel et al.
General comments:
The manuscript presents a comprehensive analysis of the distribution and trends in long-term ozone and ozone precursor observations in cities and background locations in South America. While mainly European and North American ozone records are extensively studied and its interpretation can be found in the peer-reviewed literature, such studies are rather limited for South America. Therefore, the present manuscript provides a valuable contribution to the understanding of ozone trends in this less studied region.
The determination of the trends and the change points of the trends is sound. I would have just liked to learn more about the underlying data (analytical methods, quality control, screening, …) since the quality of the data is a crucial requirement for the analysis.
The paper will fit well into the TOAR-II Community Special Issue. See below a few specific comments that should be addressed prior to publication.
Specific comments:
Line 22-24: not clear which metric the numbers are referring to.
Line 25: reader does not know yet how short-term and log-term exposure levels are defined. Add some information from lines 101 ff.
Line 26: trends refer to which metric?
Line 54: replace "a chemical regime […] has been established …" by "a chemical regime […] has been found …"
Lines 76-79: this sentence reads like a part of the conclusions. Move it below?
Lines 83 ff.: add some brief description of the measurement techniques. All UV absorption for O3? chemiluminescence for NO and NO2? If measurements are done by regulatory networks, I assume that NO2 was converted to NO prior to detection with heated surface (molybdenum) converters. It is known that these converters overestimate the NO2 mole fractions, especially in rural areas. This contribution may change over time when the amount of oxidized nitrogen species decreases. CO measurements with NDIR?
Lines 86-87: please elaborate on the data screening performed by the authors. How was drift (trends in the instruments' response, I suppose) and representativeness assessed? A 75% data coverage criterion is mentioned below. This should be added here. Did you also exclude other data such as outliers, periods with very little or very large noise, … or was the quality of the received data just good. Please add how many data/datasets were rejected prior to your analysis. Lines 189-191 provide some of this information. Still the "quality control test established in the methodology" remains unclear.
Figure 2, caption: "The black dots denote the monitoring stations that do not meet the data quality criteria." I do not see any back dots.
Lines 163 ff.: did I get it right? You attribute the lower O3 levels in Quito to intense vertical mixing that mixes (less O3-rich) air from the free troposphere to the site. Is there no signature from stratospheric intrusions seen at this elevation?
Figure S2: caption reads trend in ppb/yr while ppm/yr is shown in the figure.
Chapter 3.3: for the interpretation of the O3 change points along with the trends of the precursors. I wonder if you looked into the hourly data and the trends of the different percentiles there. At many location worldwide, it is often seen that the lowest values (percentiles) do show a positive trends (due to the reduction in NO and less O3 titration) while the highest values show negative trends.
Lines 299-300: "… These measurements have been accompanied by an increase in ozone since 2008." This refers to my comment just made above. You could doublecheck if you see that in all of your data, too. "… increase in ozone since 2018 …" Which metric are you referring to?
Lines 350-351. "We attributed these observed ozone trends to […] the establishment of volatile organic compound-limited regimes.". This looks like a firm statement that might require some more (model) analysis.
Lines 363-364: "… the lack of quality control, which prevents the inclusion of additional existing measurements." Do you refer to the quality of the measurements here? As a group of South American scientists /experts in high-quality observations, do you have any suggestion to improve the situation? Training, workshops, development of common standards (if not available), development of common tools for quality control, …
Citation: https://doi.org/10.5194/egusphere-2024-328-RC1 - CC1: 'Comment on egusphere-2024-328', Owen Cooper, 18 Mar 2024
-
RC2: 'Comment on egusphere-2024-328', Anonymous Referee #2, 18 Mar 2024
General Comments
This paper reports on surface ozone trends in South America. Data from different locations are shown together with ozone precursor data, trends are reported and reasons for the trends are discussed with the help of the precursor data. Since publications of long time series of station data in South America are rare, the manuscript should be published after these questions have been answered:
In line 152 the authors write:
In these cities, a significant fraction of ozone precursors is emitted by vehicular fleets and has decreased according to air quality control measures such as the introduction of better fuel quality, sulfur content reduction, enforcement of threeway catalytic converters, stricter emission standards for new fleet vehicles and mandatory periodic technical inspection for inuse vehicles.
What is the impact of sulfur content reduction in fuel on ozone trends ?
In Figure 3, small boxes in the left panel could indicate where the zoomed regions in the right panel are.
In Table 2, the authors divide the MDA8 and peak season data into data for 2012-2016 and data for 2017-2021. The reason for this is not clear to me. Would it be possible to treat the MDA8 and peak season data like the datasets in Table 3 and calculate turning points, p and SNR values?
In line 210 the authors write: Regardless of the latitude of each large city analyzed, each urban agglomeration contains subdivisions with high-certainty positive ozone trends.
This is difficult to see from the data in Table 3: in the Bogotá region, the trend is either positive or negative, depending on whether you look at the 5th, 50th or 95th percentile. Also, this is in contradiction to what the authors write in line 243: The ozone mixing ratios in Bogotá showed no evidence of reduction or increase during the last decade despite efforts to reduce primary pollutant emissions, as shown in Figure 4a.
I would suggest that the author rephrase the sentence or indicate which data set the authors are referring to.
Table 3 is central to the manuscript. For better readability, I would suggest including a column where it is easy to see whether a trend is certain according to the criteria in Table 1, e.g. very high certainty, high certainty…
In line 224 the authors write: the trend observed after 2014 was likely impacted by the COVID-19 pandemic in 2020 and possibly in 2021 (Putero et al., 2023).
This would suggest a change point in 2019/2020.
Can this be seen in the data?
In line 239 the authors write: In general terms, we note that these ozone precursor abatement measures have been implemented, ignoring the VOC-to- NOx ratio, suggesting that ozone increases once the VOC-limited regime is reached. The latter, together with the extensive wildfires around the cities studied, could explain the occurrence of trend change points at some sites. This should be discussed more in detail.
The authors write that ozone trend is determined by ozone production. What is the role of titration effects, e.g. an increase in ozone concentration due to a decrease in NO mixing ratio?
The CO trend reflects the trend of VOCs originating from combustion. What is the role of biogenic VOCs in ozone formation here?
A change in NOx or VOC concentration would have an opposite effect on ozone production, depending on whether the chemical regime is VOC-limited or NOx -limited. Is there any indication of whether ozone production is VOC-limited or NOx -limited?
If wildfires play a role. I would expect different trends in different seasons? Has this been studied?
In line 244 the authors write: However, in the northern area of the city, which is impacted by ozone formation in higher proportions, the median ozone trend decreased at a rate of -1.01ppb yr-1 (high certainty) between 2008 and 2013.
Can it be shown that the northern area is more affected by ozone precursors? I would expect the ozone trend of the whole region to be shown and compared with the ozone trend of the northern region. The same comparison should be made for precursors.
In line 264 the authors write: Overall, in the Quito NOx-saturated environment, decreases in NOx precursors were anticorrelated with increases in ozone.
This sentence is unclear to me as it is not clear what NOx saturation means. Is ozone production limited by VOC? Does it mean that O3 is removed by titration with NO? Also, the term NOx precursors is not clear. Are the authors referring to the sources of NOx, e.g. that NOx emissions have decreased?
In line 264 the authors write: decreases in NOx precursors were anticorrelated with increases in ozone.
Shouldn´t it be: NOx mixing ratio was anticorrelated to the ozone mixing ratio.
In line 267 the authors write: This change generally coincided with the time series period when the ozone trend stopped decreasing, leading to a change point.
However, the change point was for ozone was in 2011, whereas the change points for NOx and CO were in 2013 and 2014. Is this within the uncertainty of the change point determination?
In line 267 the authors write: The implementation of this policy probably shifted the composition and proportion of precursors, especially during the morning.
Can this be shown? Here I would expect that the trend should be more significant when looking only at the values during the morning hours at polluted sites.
In line 275 the authors write: Ozone in Santiago decreased for nearly two decades due to public policies focusing mainly on curbing particulate matter.
This implies that PM and O3 have the same sources. Can this be shown? Do PM and O3 have the same trend?
In line 282 the authors write: In other words, until 2017, the policies effectively lowered the highest ozone percentiles.
From the plots it looks that ozone was low because NOx was high and NO reacted with ozone to convert ozone into NO2. The question is what happened in 2017? Did the policy change or was NOx so low that less ozone was removed by titration?
In line 303 the authors write: Notably, many higher anomalies occurred in the warmer months (Jan-Feb) and were more frequent after the ozone change point in 2008 (Figure 7a).
This is not readily apparent from the figure. The orange dots indicating the beginning of the year are on either side of the trend line. Perhaps it is possible to specify where the trend manifests itself.
In line 304 the authors write: As a result, the ozone trends at the 90th and 95th percentiles increased
I would suggest writing: accordingly or correspondingly (as this is not a result, but a different way of presenting the results)
In line 326 the authors write: In this regard, extreme positive ozone anomalies were observed in January 2017 (7.6 ppb) and February 2023 (8.6 ppb), caused by ozone and precursors transported from areas affected by intense wildfires (Fig 6a).
Can these datapoints be shown in the figure?
In line 342 the authors write: Short-term (MDA8) and long-term (peak-season) exposure metrics calculated for the present day (2017-2021) revealed latitudinal differences in South America.
However, in lines 343 and following, the authors argue that there are several factors, and that latitude plays only a minor role. I would suggest omitting the word "latitudinal" here.
In line 350 the authors write: We attributed these observed ozone trends to a greater decrease in nitrogen oxides than in carbon monoxide, which resulted in the establishment of volatile organic compound-limited regimes.
The arguments in favour of this statement, as set out in line 239, are in my opinion too weak, see comments on line 239.
In line 366 the authors write: Finally, our results revealed signs of a climate penalty for ozone in South America
How would this climate penalty be reflected in the data? Is this related to the more frequent forest fires? This could then be mentioned again here.
In line 366 the authors write: and identified extratropical zones as those where the increase in ozone poses the highest risk.
As the authors also argue, latitude is only one of several factors that determine the ozone trend, in addition to local measurements as in Quito, Santiago, and São Paulo or the local wind system as in Bogota, see also line 150 ff. A dependence of the ozone trend on the latitude would have to be shown specifically on the existing data.
Citation: https://doi.org/10.5194/egusphere-2024-328-RC2 - AC1: 'Comment on egusphere-2024-328', Rodrigo Seguel, 01 May 2024
Peer review completion
Post-review adjustments
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 587 | 158 | 36 | 781 | 50 | 20 | 20 |
- HTML: 587
- PDF: 158
- XML: 36
- Total: 781
- Supplement: 50
- BibTeX: 20
- EndNote: 20
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 261 | 31 |
| Chile | 2 | 137 | 16 |
| China | 3 | 105 | 12 |
| India | 4 | 38 | 4 |
| Germany | 5 | 29 | 3 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 261
Rodrigo J. Seguel
Lucas Castillo
Charlie Opazo
Néstor Y. Rojas
Thiago Nogueira
María Cazorla
Mario Gavidia-Calderón
Laura Gallardo
René Garreaud
Tomás Carrasco-Escaff
Yasin Elshorbany
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(12836 KB) - Metadata XML
-
Supplement
(8833 KB) - BibTeX
- EndNote
- Final revised paper