Window-Based Filtering Aerosol Retrieval Algorithm of Fine-Scale Remote Sensing Images: A Case Using Sentinel-2 Data in Beijing Region
Round 1
Reviewer 1 Report (New Reviewer)
In this study, Authors proposed an aerosol retrieval algorithm using two-day images including a reference day and retrieval day instead of constructing LSR database. Considering subtle LSR changes and the BRDF effects, they design a filtering process in this algorithm to remove unreliable pixels in a retrieval window. Benefiting from the filtering process and high-resolution images, the retrieval AODs show very continuous and detail spatial distributions relative to MAIAC AODs during low AOD and high AOD days.
They conclude that this algorithm is designed for retrieving AODs using a reference image, although, it could also be used in algorithms based on LSR database for the purpose of obtaining greater precision.
The revised version is acceptable for publication due to sounded topic and good presentation with new ideas ..
Author Response
We sincerely appreciate your valuable comments and positive feedback on our work throughout the review process. Your evaluation has been immensely helpful for us and has boosted our confidence in our research.
We are delighted that you find our proposed aerosol retrieval algorithm innovative and that you recognized its performance qualities. We will continue to strive to improve our algorithm in order to obtain even more accurate results.
Thank you once again for your thoughtful review and support. We hope our research can bring more value to the related fields.
Reviewer 2 Report (New Reviewer)
Dear Authors,
congratulations on your work. I saw that the version of the paper that was sent to me already had numerous corrections, so I recommend only minor revisions, especially in the format of the Figures. I also have some questions about the data being used and suggest adding a table or two.
In particular
Figure 1. Enlarge the coordinates, scale and name of the analyzed sites.
Figure 3. Enlarge the font of the Cartesian axes and legends.
Furthermore, with reference to this and some subsequent Figures, specify which AERONET Data Level was used. In addition, the AOD is defined between 0<AOD<1. How do you explain values >1?
Figures 4 and 5. Enlarge the coordinates on the axes, specify in the text if the MODIS data are from the TERRA or AQUA satellite. Additionally, specify whether local or universal time is in the text and caption.
Figure 6. The Figure is very interesting, I recommend enlarging it considerably to make it more readable (Especially to better define the points in relation to the scale colors in the graph). Also specify in the caption the definition of region 1 and 2 (vegetable and urban). Explain the values of AOD>1.
Better explain in the text the difference between interpolated and non-interpolated image and highlight the differences. At the moment, they are really imperceptible.
Figure 7. Enlarge axis coordinates. Reiterate the reason for the cropped images of the 3, 13 and 66 days in the text.
Figure 8. The Figure is very interesting, I recommend enlarging it considerably to make it more readable (Especially to better define the points in relation to the scale colors in the graph).
Finally, I recommend adding a table with the coordinates of the analyzed sites. Instead, I believe a complete summary table is necessary with all the ground and satellite data used in the paper. You should report in the table the time horizon of the analyzed series, the spatial resolution, the observation bands, the number of data, the type of data (Level), the characteristics of the satellite, etc.
I reserve the right to read the conclusions once these minor revisions have been made and my few doubts have been clarified. Congratulations again for the work.
Finally, I enclose a list of important references for the study of aerosols also in relation to black carbon aerosols, which I recommend quoting in the text.
Aerosol
• Atmospheric Environment
Volume 287, 15 October 2022, 119288
Secondary inorganic aerosol dominated the light absorption enhancement of black carbon aerosol in Wuhan, Central China
Zheng, Huang ; Kong, Shaofei ; Chen, Nan ; Wu, Cheng
• Atmospheric Environment
Volume 282, 1 August 2022, 119174
Aerosols optical and radiative properties in Indonesia based on AERONET version 3
Kusumaningtyas, Sheila Dewi Ayu ; Tonokura, Kenichi ; Aldrian, Edvin ; Giles, David M. ; Holben, Brent N. ; Gunawan, Dodo ; Lestari, Puji ; Iriana, Windy
• Carbon and Trace Element Compositions of Total Suspended Particles (TSP) and Nanoparticles (PM0.1) in Ambient Air of Southern Thailand and Characterization of Their Sources
Atmosphere, vol. 13, issue 4, p. 626
Pub Date: April 2022
Inerb, Muanfun ; Phairuang, Worradorn search by orcid ; Paluang, Phakphum ; Hata, Mitsuhiko ; Furuuchi, Masami ; Wangpakapattanawong, Prasit
• Background Optical Depth Correction over Urban Areas to Improve Land Aerosol Retrieval from Himawari-8
IOP Conference Series: Earth and Environmental Science, Volume 893, Issue 1, id.012060, 8 pp.
• Monthly Notices of the Royal Astronomical Society, Volume 499, Issue 4, December 2020, Pages 5075–5089, https://doi.org/10.1093/mnras/staa3157
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.
Round 1
Reviewer 1 Report
Comments for author File: 
Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 2 Report
Dear authors,
This reviewer acknowledges the importance of the topics about the AOD retrieval from images at medium spatial resolution like Sentinel 2 as well as the efforts put by the authors on this work.
However, a complete and in-depth revision of the manuscript is required.
A) The bibliography is inaccurate, the authors should add reference addressed to the state-of-the-art about remote sensing data (satellite and ground-based) for AOD retrieval specifically for medium spatial resolution images.
Please, include a complete and coherent reference list.
The authors have to complete and correct the references in accordance with the aim of the manuscript.
The bibliography can be completed adding also the articles listed below:
"About aerosol description and their radiative effects" (3 articles):
1) The Intergovernmental Panel on Climate Change (IPCC). The Physical Science Basis. IPCC Fourth Assessment Report: Climate Change 2007 (AR4); IPCC: Geneva, Switzerland, 2007.
Haywood, J.; Boucher, O. Estimates of the direct and indirect radiative forcing due to tropospheric aerosols: A review. Rev. Geophys. 2000, 38, 513–543.
2) D’Almeida, G.A.; Koepke, P.; Shettle, E.P. Atmospheric Aerosols: Global Climatology and Radiative Characteristics; A. DEEPAK Publishing: Hampton, VA, USA, 1991.
3) The Intergovernmental Panel on Climate Change (IPCC). The Physical Science Basis. IPCC Fifth Assessment Report: Climate Change 2013 (AR5); IPCC: Geneva, Switzerland, 2013.
"About aerosol from space" (1 article):
1) King, M., Kaufman, Y., Tanre, D., and Nakajima, T.: Remote sensing of tropospheric aerosols from space: Past, present, and future, B. Am. Meteorol. Soc., 88, 2229–2259, doi:10.1175/1520-0477(1999)080<2229:RSOTAF>2.0.CO;2, 1999.
"About aerosol on atmospheric correction processing and sensitivity of sensor signal to aerosol" (1 article):
1) Bassani, C., Cavalli, R. M., and Antonelli, P.: Influence of aerosol and surface reflectance variability on hyperspectral observed radiance, Atmos. Meas. Tech., 5, 1193–1203, https://doi.org/10.5194/amt-5-1193-2012, 2012.
"About theoretical basis" (2 articles):
1) Vermote, Eric F., et al. "Second simulation of the satellite signal in the solar spectrum, 6S: An overview." IEEE transactions on geoscience and remote sensing 35.3 (1997): 675-686.
2) D. Tanre, J. l. Deuze, M. Herman, R. Santer, and E. Vermote, “Second simulation of the satellite signal in the solar spectrum—6S code,” in Proc. 10th Annu. Int. Symp. Geosci. Remote Sens., May 1990, p. 187.
B) The authors have to revised the paper to provide more clarity on the method and the results which also make easier for readers to understand the aims of this work. Especially the Section 3.2.1 "Filtering" and 3.2.2 "Retrieval".
Furthermore, the reviewer suggests to justify each assumption introduced in the paper. For example, the method is based on negligible (small) multiple ground-atmosphere interaction and constant surface reflectance during the period between t1 and t2. Please, deepen these aspects into the manuscript.
C) Please, answer to these questions:
What about the variation of surface reflectance for meteorological and environmental conditions?
Are you sure that 30 days can be used to consider your window as invariant target?
Is the spherical albedo of the atmosphere considered zero also for hazy days?
How do you recognized a priori the sunny and hazy days?
What about the adjacency effect?
Please explain why the AOD depends on a single aerosol model adding also consistent bibliography.
Are you sure that a single aerosol is enough to describe all data acquired in an urban site? How many data have you used? How long is the period with available data?
The microphysical properties of the aerosol depend on the AOD. Please explain this assumption and adding also robust reference.
Why the microphysical properties of Table 1 are not compared with AERONET inverse products?
Why are not reported the AOD values of hazy days in function of the corresponding sunny days? Please, add to the manuscript plots with AOD values for hazy and corresponding sunny days.
The Discussion Section seems to present an additional application of the algorithm without mentioning other works in this research field. Please, rewrite the Section highlights these aspects.
The monitoring of air pollution sources needs often to work with values higher than ones used in this work. Why do you address your algorithm to the atmospheric pollution monitoring?
The air quality is based on PM data. How do you correlate the retrieved AOD with PM? PM2.5? PM10?
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Specific comments:
p. 2 l. 61: Please, correct R2=0.932 with 0.93
p. 2 l. 66: Please, correct R2=0.920 with 0.92
p. 3 l. 90-106: Please, add robust bibliography to Sentinel 2 mission
p. 3 l. 115: Please, correct "not at 550nm but at 550nm, 500nm and 675nm"
p. 3 l. 132: Please, correct the bibliography of the equation [Vermote, Eric F., et al. "Second simulation of the satellite signal in the solar spectrum, 6S: An overview." IEEE transactions on geoscience and remote sensing 35.3 (1997): 675-686.]
p. 4 l. 165: Please, rewrite "Instead of constructing Land Surface Reflectance (LSR) datasets, the LSR is obtained
directly from an image of sunny day with as close as possible time"
p. 5 l. 177-219: Please, rewrite the Section "Filtering" to make easier to understand your method
p. 6 l. 221-225: Please, rewrite the Section "Retrieval" to make easier to understand your method
p. 7 l. 262: Explain o rewrite "These illustrates both the stability of the algorithm and the rationality of the aerosol model."
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
Review of “Window-based Filtering Aerosol Retrieval Algorithm of Fine-scale Remote Sensing images: a case using Sentinel-2 data in Beijing region” by Jian Zhou et al. For publication in MDPI Remote Sensing
The paper describes an algorithm designed to retrieve the aerosol optical depth (AOD) from Sentinael-2 MSI images over land surfaces. The algorithm is applied to hazy images acquired over Beijing region (PRC) between 2017 and 2021. The algorithm uses a reference image for which the AOD is known and then derives the AOD for other hazy pictures over the same area. The retrieved AODs are then compared to AERONET AOD measurements and MODIS MAIAC algorithm retrievals.
Major remarks :
- It is unclear for me where you get the AOD for the so-called sunny pictures. I suggest that you put better explain the choice of your reference image and called “reference image” throughout all the paper.
- As mentioned in the paper, a 30-day period is quite long and the surface reflectance can change rapidly due to vegetation cycle. It would be interesting for the reader to have more information on a possible reduction of the time period and how it impacts the results.
- the comparison of your retrievals to MAIAC is rather difficult to evaluate (high AODs in saturated red color). I suggest to provide a scatter plot of MAIAC resultst and your retrieval for selected regions. You can also extract values below selected transects that goes from low to high AODs on both images.
Minor remarks :
L28 : use units
L33 : both natural and anthropogenic aerosols have short atmospheric lifetime
L55 : please revise how you present the 2 methods. I guess (1) refers to one of the method used to estimate LSR ?
L62. What is the meaning of (2) and (3) here ?
L155-L157. Error on subscribes. Please revise the sentence.
L157. 30 days is not a small time interval.
L233. Interpolate window ?
L281. Revise sentence.
L282. “Reasonable” doesn’t sound like a scientific expression. Please explain.
L346. The impact of a change in the spatial resolution of the AODS on air quality monitoring over urban areas is not demonstrated in your paper. So it’s not a conclusion of your paper.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
