Aerosols on the Tropical Island of La Réunion (21°S, 55°E): Assessment of Climatology, Origin of Variability and Trend

Round 1

Reviewer 1 Report

Comments are attached in *.doc file.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

In the current manuscript, the authors have presented results on the climatological AOTs and the other retrieved optical properties based on long term AERONET measurements. These results have a very high significance given that such long-term ground based data are either very rare or non existent from remote oceanic locations (SWIO).

I have few suggestions:

1. Since the study is based on long term data, the authors should discuss about the instrument (Cimel Sunphotometer) calibrations (frequency of the calibrations) and if there are any seasonal or annual changes/drifts in the calibration errors.

2. The authors can compare their results with similar AERONET data from other remote locations such as Mouna Loa, Ascension Island, etc.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

The manuscript presents the AERONET measurements (2007-2019) carried out at the Observatory of Atmospheric Physics of La Réunion (OPAR), a remote site in the South-Western Indian Ocean, with an analysis of the aerosol optical thickness at 440 nm (AOT₄₄₀), the Ångtröm exponent between 500 and 870 nm, a, and the aerosol volume size distribution. The seasonal behaviour of the AOT and a is investigated in order to highlight the role of biomass burning aerosol transported from the surrounding continents and to separate them from the background conditions dominated by marine aerosols. Results show that, while background conditions are characterized by low (high) values of AOT₄₄₀ (a) from August to November, contribution of biomass burning determines larger (lower) AOT₄₄₀ (a) during the wildfire season in the Austral summer. The role of fine and coarse particles is also found in the volume size distribution.

An additional analysis employs the liner-regression fitting model TrendRUN  with the double aim of  reproducing the observed AOT monthly time series and of evaluating the contribution to the AOD evolution of different components linked to aerosol properties (biomass burning particle emissions and sea salt AOT) and to large-scale atmospheric phenomena (Quasi-Biennal Oscillation, El Niño Southern Oscillation, Indian Ocean Dipole, Madden-Julian Oscillation). This analysis uses different databases, i.e. emission from the Global Fire Emissions Database, sea salt AOT from CAMS reanalysis, different indices. Here the results give a good performance of the model in reproducing the monthly averaged measured AOT₄₄₀, with a coefficient of determination R² of 0.892, although particular large AOT₄₄₀ peaks can not be modelled. Biomass burning emission provides the largest contribution, with a predominant role of the Southern Hemisphere Africa and the Southern Hemisphere South America wildfires.

The manuscript is valuable in presenting the aerosol measurements in a remote area of the Southern Hemisphere, where ground-based observations are of great importance for the validation of satellite and model/reanalysis products. The description of the data and methods is sufficiently clear. I suggest to anticipate in the abstract the period that has been considered in the study (May 2007-December 2019), which appears only in Section 2 Materials and Methods. In addition, I think that it would be useful to have some details about geography and environment of the site, for example if the Observatory is close to a city or to the sea, in order to have an idea of possible local effects.

The part dedicated to the analysis of the components affecting the AOT evolution and trend is, in my opinion, interesting although my question is why the chosen forcings should be more representative than others. I refer in particular to the regional biomass burning emission database, which provides the emitted total particulate matter (ETPM), but can not take into account any transport mechanism determining an impact on the column AOT on the site. For example, the contribution of the marine aerosol is accounted by the CAMS sea salt AOT averaged on a 2°x2° box centered on the site: did the authors examined the possibility to use the CAMS black carbon AOT instead of the  ETPM? Moreover, it would be interesting to see if the black carbon AOT around La Réunion may present a trend, so to eventually explain the measured AOT trend. Another question is what the results may be using MERRA-2 instead of CAMS aerosols?

Also, since the IOD MJO contribution is very low, may those indices be excluded in the analysis?

Finally, I would like to ask if large volcanic eruption may have affected AOT in the period under investigation and if the authors evaluated the possibility/feasibility to include the contribution of volcanic aerosols.

Specific comments are presented in the following.

Lines 45-55. This sentence mostly refers to studies based on field experiments, while many global and hemispheric studies based on satellite observations (e.g. MODIS) and/or model data exist.

Line 62. Since volcanic aerosol is mentioned later, hints on the transport of this type of aerosol in the area can be given here.

Line 76. Can the authors provide some example of participation to global network or research infrastructure or satellite validation programme or a web link where to find these information?

Line 80-83. It would be useful to the reader to have some information about local aerosol sources and their possible role in the measured AOT variability, since in the manuscript only long-range transport is considered. This makes me believe that, for example, local wildfires are negligible.

Lines 119-120. I guess that instantaneous and not averaged data are used. Am I right?

Line 176. Can the authors specify what DMI is?

Line 214. Negative a values probably result from very low AOT measurements and are thus associated by large uncertainty. Authors should account this when comment the results.

Line 219. I would not use the term “trend” in this case, but “behaviour”.

Line 266-267. Mean effective radius.

Line 284. Change “gives” with “give”.

Lines 369-382. Are all the presented trend statistically significant? The non-statistically significant trend should be specified.

Line 374. Please check the -0.1T trend.

Lines 402-403. The https://www.pa.po.dlr.de/CCMVal/Forcings/qbo_data_ccmval link is not accessible.

Line 405. The www.jamstec.ggo.jp/frsgc/research/d1/iod link is not accessible.

So my advice is to further discuss the choice of the components that should contribute to the observed AOT evolution and variability, in order to make the analysis more robust.

My conclusion is the manuscript in this form has to undergo major revisions.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report

This paper analyzes the aerosol measurements from 2007 to 2019 performed by a Cimel Sun-photometer (part of AERONET network) placed in the Le Reunion island, in the South-Western Indian Ocean. In particular AOT (at 440 nm) and Angstrom parameter were considered, plus a figure with volume size distribution too. Using a linear regression fitting model (TrendRUN), AOT trend was correlated with biomass burning emissions from different relevant areas, marine aerosols and some large-scale atmospheric structures. Through this model the authors were able to show the contributions of each considered phenomenon to AOT variability, highlighting how the greatest one comes from biomass burning in particular from Southern Hemisphere South Africa and from Southern Hemisphere South America.

General comments:

In my opinion the paper is interesting and well written, so I think that it could be accepted for publication after some minor revisions.

Personally the part that I found most interesting and surprising is the one concerning the TrendRUN model and its ability to very accurately simulate the AOT variability from 2007 to 2019 (section 3.2, fig. 8 and table 2). And this is also the part that in my opinion needs some clarifications, especially regarding how the model works and the way to consider the BB contributions. Which BB parameter was used in the model? Was Emitted Total Particulate Matter (ETPM) shown in fig. 1 used? How has it been taken into account that aerosols emitted in various parts of the world reach the island of Le Reunion or not? Was a specific area above the island considered as in the case of sea salt aerosols (2 deg box)?

Specific comments:

L19 and everywhere in the whole paper: please consider to replace AOT with AOD. More and more often in recent years the term "aerosol optical depth" is preferably used over the term "aerosol optical thickness". In Aeronet website AOD is used. Also CAMS EAC4 uses AOD.

L20: define "AOT440", that is Aerosol Optical Thickness at 400 nm

L30: put "440" in the subscript of AOT

L53: define "BB" as biomass burning here and not in line 61 

L54-55: missing a piece in this sentence?

L76-77: "OPAR is the only ground-based observation point in the Indian Ocean": sure of this sentence? Others AERONET sites are present in Indian Ocean Area I think ...

L92: why you cited "volcanic emissions" here? In the paper volcanic emissions were not considered. Please argue more about this sentence or erase it.

L120: put a blank between "870" and "nm"

L127: define "UTLS"

L142: please clarify how do you compute ETPM. I've downloaded the GFED4s readme file and it seems to me that this parameter is not present directly in the database. Did you obtain it from fire carbon and dry matter emissions using emissions factors? So ETPM has a spatial resolution of 0.25 degrees like GFED global files? The original data are daily means or monthly means? The units are per m2, so have you multiply with the area to obtain the total emissions in g? In fig. 1 you show the ETPM that arrives above Le Reunion island only or is it the total emitted from the areas considered? 

L154: put a blank between "550" and "nm". About SSAOT550, please give some more information about its original spatial and temporal resolution.

L167: the text "3.2)." is between the figure and the caption.

L171: as for biomass burning, please give some more explanations on how you managed to consider the contributions of the various large-scale atmospheric structures specifically on the island area.

L174: the QBO link doesn't work

L176: define "DMI"

L176: the IOD link doesn't work

L176 and L178: in others parts of the paper the model is named "TrendRUN" while here is "Trend-Run", please correct

L185 e L189: I think it's better "89150" (without blank)

L198: table 1: the maximum values of AOT for MAM and JJA are less than mean values (that is impossible). Please correct

L203: please give an explanation as to why you have chosen AOD_440 and Angstrom_500-870 and not for examples others wavelengths

L219: I think it's quite difficult to see a "clear trend" in fig. 4 (upper right). I would rather say that for high AOT there are small particles only (high alfa), while for low AOT there are both small and big particles (both high and low alfa).

L255-258: I think that at this point of the paper this sentence (the fact that seasonal variation of AOT and alfa at La Reunion depend mainly by sea salt and BB aerosols) is not sufficiently justified from scientific point of view. AOT and alfa alone do not permits to be sure about the origins of aerosols. This thesis is certainly reasonable but in this case it will be confirmed only after applying the model TrendRUN with the various contributions. I suggest to shift L255-258 considerations at the end of section 3.1 (after fig. 6) and to rephrase the sentence in a hypothetical way and not as a fact that can be inferred with certainty from the data. 

L260: fig.5: x-axis label is missing

L273: fig.6: personally I'd prefer replacing the x values labels (10^-1, 10^0, 10^1) with (0.1, 1, 10) as in AERONET website. I think it would be better and faster understandable.

L275: section 3.2: as said in the general comments, this part is very interesting but needs some more explanations about the characteristics of the parameters put inside the TrendRUN model and how the model works. In particular for biomass burning aerosols. 

L291: in table 2, since there is enough space, I suggest making acronyms explicit to facilitate the understanding of readers, that is writing: Quasi-Biennal Oscillation (QBO), El Niño Southern Oscillation (ENSO), ..., Southern Hemisphere South America (SHSA), and so on ...

L308: after the consideration about the "episodic volcanic aerosol loading, which are not taken into account in this study" I suggest to insert a few words to say that on the island there is an active volcano (Piton de la Fournaise) because a generic reader does not necessarily knows.

L331: put a blank between "800" and "km"

L333: put "550" in the subscript of SSAOT

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 5 Report

Review of "Aerosols on the tropical island of La Réunion (55°E, 21°S): assessment of climatology, origin of variability and trend" by Duflot et al., 2022. The main goal of the submitted paper is to assess from a climatological point of view the characteristics of both local and advected aerosols over Reunion Island using Aeronet sunphotometer observation. The paper makes sense, especially in this remote part of the world. The main criticism is related to the fact that the study is lacking a little bit of originality. Without upsetting the whole study, I would add how the different types of aerosols related to the particular period of the year affect the Earth-atmosphere budget. It is possible to have an estimate also without knowing the vertical profile of the aerosols. Some ideas can be found in Lolli, S., Khor, W. Y., Matjafri, M. Z., & Lim, H. S. (2019). Monsoon season quantitative assessment of biomass burning clear-sky aerosol radiative effect at surface by ground-based lidar observations in Pulau Pinang, Malaysia in 2014. Remote Sensing, 11(22), 2660. and in Gu, Y., Liou, K. N., Jiang, J. H., Su, H., & Liu, X. (2012). Dust aerosol impact on North Africa climate: a GCM investigation of aerosol-cloud-radiation interactions using A-Train satellite data. Atmospheric Chemistry and Physics, 12(4), 1667-1679.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

In revised version of the manuscript, authors incorporated all the comments raised by reviewer and can be accepted in its present form.

Author Response

Thank you.

Reviewer 3 Report

I thank the authors for their reply and for improving the manuscript with the additional descriptions and discussion.

I have some minor points to be addressed.

Introduction: the authors do not mention desert dust as possible aerosol type contributing to AOD. Do dust particles have a completely negligible role in AOD at La Réunion?

Lines 45-46: What do the authors mean with the sentence "where changes in the aerosol concentration can give rise to unexpected results"?

Line 269: change “show” with “shows”;

Section 4; I would suggest to add a comment on why most of the highest peaks in AOD440 are not reproduced by the Trend-Run model.

After these minor revisions the manuscript can be published.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 5 Report

I am not happy with the authors' answers. A few lines just saying it is out of their paper's scope are insufficient. I suggested a couple of references, I would have expected in their new draft a discussion on what I meant in my first review, putting their research in a wide context. 

Author Response

Please see the attachment.

Author Response File: Author Response.docx