the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 20 Jul 2023
Tropical tropospheric aerosol sources and chemical composition observed at high-altitude in the Bolivian Andes
C. Isabel Moreno
Radovan Krejci
Jean-Luc Jaffrezo
Gaëlle Uzu
Andrés Alastuey
Marcos F. Andrade
Valeria Mardóñez
Alkuin Maximilian Koenig
Diego Aliaga
Claudia Mohr
Laura Ticona
Fernando Velarde
Luis Blacutt
Ricardo Forno
David N. Whiteman
Alfred Wiedensohler
Patrick Ginot
Paolo Laj
Abstract. The chemical composition of PM10 and PM2.5 was studied at the summit of Mt. Chacaltaya (5380 masl, lat.-16.346950º, lon. -68.128250º) providing a unique long-term record spanning from December 2011 to March 2020. The chemical composition of aerosol at the Chacaltaya GAW site is representative of the regional background, seasonally affected by biomass burning practices and by nearby anthropogenic emissions from the metropolitan area of La Paz – El Alto. Concentration levels are clearly influenced by seasons with minimum occurring during the wet season (December to March) and maxima occurring during the dry and transition seasons (April to November). Ions, total carbon (EC+OC) and saccharide concentrations range between 558–1785, 384–1120 and 4.3–25.5 ng m-3 for bulk PM10 and 917–2308, 519–1175 and 3.9–24.1 ng m-3 for PM2.5, respectively. Such concentrations are overall lower compared to other high-altitude stations around the globe, but higher than Amazonian remote sites (except for OC). For PM10, there is dominance of insoluble mineral matter (33–56 % of the mass), organic matter (7–34 %) and secondary inorganic aerosol (15–26 %). Chemical composition profiles were identified for different origins: EC, NO3-, NH4+, glucose, C2O4-2 for the nearby urban and rural areas; OC, EC, NO3-, K+, acetate, formiate, levoglucosan, some F- and Br- for biomass burning; MeSO3-, Na+, Mg2+, Br- for aged marine emissions from the Pacific Ocean; arabitol, mannitol, K+ for biogenic emissions; Na+, Ca2+, Mg2+ for soil dust, and SO42-, F-, and some Cl- for volcanism. Regional biomass-burning practices influence the soluble fraction of the aerosol particularly between July and September. The organic fraction is present all year round and has both anthropogenic (biomass burning and other combustion sources) and natural (primary and secondary biogenic emissions) origins, with the OC/EC mass ratio being practically constant all year round (10.5±38.9). Peruvian volcanism dominates the SO42- concentration since 2014, though it presents a strong temporal variability due to the intermittence of the sources and seasonal changes on the transport patterns. These measurements represent some of the first long-term observations of aerosol chemical composition at a continental high-altitude site in the tropical Southern hemisphere.
- Preprint
(1469 KB) - Metadata XML
-
Supplement
(3261 KB) - BibTeX
- EndNote
C. Isabel Moreno et al.
Status: final response (author comments only)
-
CC1: 'Comment on egusphere-2023-1298', Hector Jorquera, 12 Oct 2023
Publisher’s note: the content of this comment was removed on 16 October 2023 since the comment was posted by mistake.
Citation: https://doi.org/10.5194/egusphere-2023-1298-CC1 -
RC1: 'Comment on egusphere-2023-1298', Héctor Jorquera, 13 Oct 2023
General comment
This manuscript reports the chemical speciation of a long-term set of ambient PM10 and PM2.5 samples taken at the Chacaltaya observatory, Bolivia (5240ma.s.l.), the highest GAW station.
The manuscript is clear in describing the sampling protocols and associated speciation analyses, along with meteorological analyses regarding backward wind trajectory analyses and a methodology to assess influence of convective PBL air masses arriving at the station, as compared with free troposphere, longer-ranged trajectories. The amount of new information presented is substantial, and it does not overlap with previous reports of measurements at the same site. Hence, I find this manuscript suitable for ACP’s scope.
Nonetheless, the manuscript needs a revision to be considered acceptable for publication in ACP.
Specific comments:
- Several statements regarding potential sources of PM10 and PM5 are qualitative (Abstract, lines 34-40). A source apportionment (SA) should have been carried out to achieve quantitative conclusions about major sources impacting the monitoring site. Several manuscript authors have already done so for the closest urban area of La Paz – El Alto (Mardoñez et al, Source apportionment study on particulate air pollution in two high-altitude Bolivian cities: La Paz and El Alto, Atmospheric Chemistry and Physics Discussions, 1–41, https://doi.org/10.5194/acp-2022-780, 2022). A comparison of Chacaltaya SA results with those already published for the neighbor urban area would provide an in-depth quantitative analysis and would enhance the manuscript’s scientific value. Most published SA studies present a chemical speciation campaign followed by application of a receptor model.
- In close connection with the above comment, the discussion in section 3.3 would benefit of presenting SA results beforehand, so seasonality would be discussed in terms of sources rather than by species (that may come from several sources).
- In Section 3.1, two estimates of OM/OC ratio were used, because of seasonality. Is it possible to estimate that ratio by linear regression of (PM-inorganic mass) against OC? This could be carried out by season to account for such variability. In this way, the uncertainty in OM would be reduced.
- In section 3.2.2 (lines 363-376) it is discussed that OC/EC is ≈ 10 with little seasonality, and this is ascribed to long-range, aged aerosol dominates with a high SOA contribution to OC. I do not understand the hypothesis stated in lines 369-370: why is this hypothesis needed to explain these OC/EC ~ constant results?
- Section 3.5: the discussion that ends with Table 6 would have improved with a SA result for Chacaltaya beforehand.
- Conclusion section: I think there are contradictory statements here. First, on lines 630-631, it is mentioned that “La Paz and El Alto … activities… affect the aerosol chemical composition (at Chacaltaya) with EC, NO3 … as traffic indicators… ”. Then, in lines 636-637 it is stated that “OC/EC ratio … does not have a marked seasonality … likely due the permanent influence of long-range transport”. However, OC is also emitted by traffic, and it is mentioned that OC/EC ratios for La Paz – El Alto range between 2 – 3.5 (approx.). Then, I do not understand why EC from La Paz -El Alto would impact Chacaltaya but not OC emitted from the very same area — given that in lines 369-379 the authors hypothesized that “… the high UV of the tropical atmosphere over the Altiplano could play a role in the impressively fast aging of the organic matter at this site when transported from the nearby urban area.” This issue needs to be clarified.
Technical corrections
- I think figure S12 should be referred to instead of S10 (line 122).
- In Section 3.3, Figure 5 is hard to visualize. I would recommend splitting it in several graphs, perhaps moving some to supplementary information.
- Since this is not the first report about Chacaltaya measurements, sections 2.1 and 2.2 could be shortened by moving some paragraphs to Supplementary Information.
Citation: https://doi.org/10.5194/egusphere-2023-1298-RC1 -
RC2: 'Comment on egusphere-2023-1298', Anonymous Referee #2, 18 Oct 2023
This manuscript presents the composition of aerosol collected over nearly 9 years at the summit of Mt.Chacaltaya (elevation 5380 m) in the Bolivian Andes. The sampling program was quite demanding, making the data set valuable at least in part due to uniqueness. Data interpretation focuses largely on establishing seasonal variations and assessing contributions from different sources, as well as comparing concentrations to those observed at other remote stations around the world.
It is not entirely clear how many new insights have resulted from considering the full 2011-2020 record now available, compared to findings already published from shorter campaigns at the same site. This issue should be addressed in a revised and partly reorganized version, that also seeks to clarify some confusion in sections of the current draft that will be described below.
To me, the most significant example of this problem is section 3.5, which seeks to ascribe small (largely not significant) differences between the concentrations observed during the daytime versus nighttime to boundary layer dynamics. Issues that strike me are that 1) all earlier subsections of Results and Discussion combine daytime, nighttime, and 24-hour filter samples precisely because any diurnal effects are so small, 2) details of the sampling protocol in this study seem to indicate that there were large temporal offsets between any given pair of day/night samples that will complicate comparisons and then averaging 6-8 daytime and 7-12 nighttime samples before comparing would seem to combine multiple possible factors controlling concentrations and likely obscure any impact dominated by boundary layer variations, (more on this later), and 3) several other short term studies that made faster measurements are cited that provide stronger evidence of vertical mixing bringing local pollution to the site for relatively short episodes. I recommend that Section 3.5 be removed from the revised version.
Following are a list of less substantial suggested changes that should improve clarity of the presentation, referenced to line numbers in the pdf.
30-32 I found it disconcerting to read in the abstract that concentrations in PM2.5 tended to be higher than in PM10. This should not be possible for samples collected simultaneously. Much later (line 161 and in Table 1) it becomes clear that sampling was either behind PM 10 or PM 2.5 impactors but not both at same time. The inference that most of the measured compounds were dominantly on smaller particles, based on the similar concentrations measured in PM 2.5 and PM 10 samples, seems well founded. However, the abstract needs to mention that the PM 2.5 and PM 10 sampling occurred during non-overlapping periods, and might point out that it seems most aerosol mass is found on the PM 2.5 fraction.
40-41 A mean (or median) of 10.5 with standard deviation of 38.9 does not seem consistent with the statement that the EC/OC ratio was "practically constant" year-round.
95-97 Are La Paz and El Alto 2 different cities that are close to each other, or is one specifically a city and the other the surrounding metropolitan area that includes the city (like Los Angeles and the Los Angeles basin)? At the start of this sentence I thought El Alto was a city in the La Paz metro area because the population in 1976 was much larger in La Paz. But then it says that El Alto population in 2012 was much larger than the La Paz population (not possible if El Alto is part of La Paz).
105 Since you note that much of the precipitation is solid you should clearly state that 865 mm annual average is water equivalent depth, or separately report the depth of rain and snow if necessary.
108-109 "Wet-to- transition" should be "Dry-to-wet transition"
110-111 It does not seem to make sense for winds coming to the station from the SE and E to be channeled through valleys N of the station. Could work for NE winds.
117-125 This paragraph is part of the reason I suggest deleting section 3.5. Findings based on the long data set in that section are kind of weak, and largely already established
145-155 In combination with Table 1, this paragraph claims to provide details of the sampling schedule. However, I feel some important details are missing. In particular, it is not clear how you alternated between day, night, and 24-hour samples. If, for example, you collected 8 day, then 7 night, then 2 24-hour samples during the wet season in PM10-A, the mid points of the first day and last night samples would have been separated by ~42 – 98 days. This would make it hard to ascribe any differences in concentrations solely to boundary layer dynamics. Even if you alternated between day and night samples the temporal offset poses a challenge in terms of attributing differences to specific process(es).
Not sure whether it would be easier to describe the actual sampling schedule in the text or in Table 1.
198 "elementary" should be "elemental"
218-219 Why was preservation of fluoride, chloride, and nitrate not a problem in periods A and B? Related question, how confident should the reader be in the fluoride, chloride, and nitrate concentrations in the earlier periods, including for PM 2.5?
237 HYSPLIT is Hybrid Single-Particle Integrated Trajectory (add underlined parts)
262-263 Consider providing more details on the comparison between trajectories driven by WRF versus ERA-5. The short statement here suggests that the ones with ERA are suspect, yet they are used for most analyses.
265-273 This paragraph is jumpy, making it hard to know what you are trying to emphasize. I also note that it is not customary to refer to Table 4 prior to mentioning Tables 2 and 3, then Table 5 also before Table 3. In addition, it is not clear what hypothesis is tested with the Mann-Kendall test (Table 5). Here it is suggested the test was whether concentrations day/night were different, Section 3.5 suggests you were comparing across the 4 sampling periods, and the words in the table suggest somehow this test allowed source attribution.
271 Is it surprising or notable that the very low concentrations during wet season were the most statistically similar?
278-296 I do not understand the decision to begin Results and Discussion with Section 3.1 which focus on a very small subset of data. Would seem better to start with 3.2 and 3.3 and come back to this as another "special case" before or after Section 3.4.
Table 3. Suggest reporting TE in micrograms/m^3 like everything else.
304-306 Would be helpful to remind reader here that PM 10 and PM 2.5 samples were collected at different times.
315 seems either there are words missing after "and" or that "and" should be deleted in "nitrate and stands"
337-339 This summary of prior work on sulfate sources at Chacaltaya is part of the reason I said it is not so clear what the new long data set brings to the story.
363-365 Two questions here. What about April, it is notable by not being included in any of the seasons? And how can the standard deviation be 38.9 over the entire study (line 40-41) but never higher than 7 in any season.
369-370 It does not seem that the OC/EC ratio provides strong evidence for "impressively fast aging" so not sure why this sentence is inserted in this paragraph.
378-385 This is nice background information, but what is interesting or notable about the Chacaltaya results in Fig. 3?
Figure 3 might work better if the scaling on Y axis was same in both panels.
403 "does not to have" should be "does not" or "does not seem to"
414-415 Does it make sense to aggregate all 4 sampling periods to assess seasonality? In particular I wonder about sulfate, which you later show has a step change due to volcanic emissions that may obscure seasonality.
It might make sense to check seasonality in each period separately. Might not want/need to show these results, but you could note whether the seasonality is persistent across the study (and perhaps focus on species for which it is not to see if there is information there).
424 In section 2.1 you seemed to imply that westerly winds were quite rare, so a little surprising to hear there is a season with significant westerly flow.
435-440 Interesting that you find significant marine-sourced MeSO3^-, but suggest that nearly all sodium and magnesium are crustal. One might expect some sea-salt with the MeSO3^-. Might be worth looking at case studies rather than the monthly averages.
445-447 This sentence almost contradicts the one I pointed to immediately above.
Section 3.3.2 It is quite surprising to me, and may be to others, that you identify a biomass burning (BB)cluster that does not include ammonium. You may want to confront this in this section, rather than just noting that ammonium peaks in the dry season, often correlated with sulfate, and coming back to it in section 3.5 which I suggest be deleted.
460 Not sure "notorious" is the correct word here.
485-510 Text here seems to muddle your story. I grant that most of the things measured have more than a single source, but this section is supposed to be focused on JAS when smoke seems a significant if not dominant source. My point is why would possible marine, urban, volcanic sources contribute to peaks in selected compounds in late summer, but other compounds that also come from some of these sources do not show significant enhancements.
515-516 If lithium is often near detection limits, why focus on it? And why suggest it may come from BB in SON when previous section points to JAS peak in BB influence?
520-523 Speculation about glucose, mannitol, and ararbitol seems weak. Why would high variability indicate continuous influence from the Amazon. The March peaks are not striking in Fig 5, in fact all seem enhanced in Aug-Nov nearly as much as in March.
Section 3.4 How does the proposed increase in volcanic emissions after period A fit with the earlier finding that sulfate peaks in dry season (section 3.3.1)? Seems unlikely that the volcanoes track seasons. Would the seasonal variation of sulfate change if you removed samples with W or NW trajectories before calculating monthly averages? Main point is that different sections of this manuscript need to be somehow connected.
Fig 6. Why show the volcanic emissions from 2005 through 2011 (before you have aerosol measurements) in these plots?
Section 3.5. No detailed comments given recommendation that entire section should be deleted.
636-638 Confusing to claim important year-round influence of long-range transport immediately after emphasizing local sources.
Citation: https://doi.org/10.5194/egusphere-2023-1298-RC2
C. Isabel Moreno et al.
C. Isabel Moreno et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 384 | 170 | 22 | 576 | 38 | 12 | 15 |
- HTML: 384
- PDF: 170
- XML: 22
- Total: 576
- Supplement: 38
- BibTeX: 12
- EndNote: 15
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1