Next Article in Journal
Machine-Learning Classification of SAR Remotely-Sensed Sea-Surface Petroleum Signatures—Part 1: Training and Testing Cross Validation
Previous Article in Journal
Cold Atom Interferometry for Enhancing the Radio Science Gravity Experiment: A Phobos Case Study
Previous Article in Special Issue
Land Use Quantile Regression Modeling of Fine Particulate Matter in Australia
 
 
Article
Peer-Review Record

Atmospheric Formaldehyde Monitored by TROPOMI Satellite Instrument throughout 2020 over São Paulo State, Brazil

Remote Sens. 2022, 14(13), 3032; https://doi.org/10.3390/rs14133032
by Arthur Dias Freitas * and Adalgiza Fornaro
Reviewer 1:
Reviewer 2:
Reviewer 3: Anonymous
Remote Sens. 2022, 14(13), 3032; https://doi.org/10.3390/rs14133032
Submission received: 29 May 2022 / Revised: 17 June 2022 / Accepted: 19 June 2022 / Published: 24 June 2022

Round 1

Reviewer 1 Report

Summary: Authors present satellite-derived HCHO column-integrated concentrations at several areas with a variety of land surface types in a southern part of Brazil. About 200 days worth of column measurements are averaged for a year (which year?). Satellite observations are compared with approximately comparable surface-based concentration measurements, showing rough consistency. Concentrations in the 1-10 ppb range are reported,  along with seasonal and spatial patterns. The measurements are new and interesting, and the authors have done an excellent analysis and presentation. The results should be published with some revisions and clarifications noted here.

1) I'm confused about the timing of when measurements are reported. It appears that there are only 1 to 3 column measurement made per day during the time noon-4PM local time in the domain the authors are studying, but this is reported as a "daily" value. Clearly one measurement made at 2PM is not a daily average. My understanding is the satellite has approximately a 2-hr orbit time, and a ~2400 km measurement swath, so theoretically there might be 3 observations (e. g. noon, 2PM, 4PM) averaged to create each 6 km x 4 km "pixel" on the maps shown here? If there are 2-3 "observations" per day at individual points, this finer time resolution might be interesting to analyze. Please describe the raw measurements & how the maps showing HCHO columns in this paper are created

2) Furthermore these "daily observations" of ~6x4 km^2 area happen only under clear conditions? Which looks like it might be only half the time? Obviously HCHO concentrations probably exhibit significant diurnal variations, and also HCHO abundance probably depends on whether it is cloudy or clear. So really the authors are only presenting 1-3 mid-afternoon concentrations on clear days, which is OK, but must be clearly described. Later, the authors compare surface measurements that probably are for true 24-hour averages. This discrepancy in timing should be noted.

3) Table 1 must clearly state the time period (e. g. Sep 2019-Dec 2020) over which measurements are presented. Do the 180 days of measurements at Zona Leste encompass the same days as the 140 reported PETAR1? Are all these days in 2020? Do the authors know the number of actual satellite overpass measurements included, or just the days where at least a single valid overpass measurement is available?

4) Table 2 and Table 1 imply these are truly annual or seasonal averages, when the concentrations are really for just one-three "observations" per day measured during the satellite overfly times (12-4PM local). I'm sure there are diurnal fluctuations of the column totals shown, and the limited observation times must be explicitly stated. A more correct description would be these are "Annual-averaged column concentrations of mid-late afternoon HCHO (12-4 PM local time)"

5) Table 2 should show in parentheses the NUMBER of observations (number of overpasses is better than number of days with a valid overpass) involved in each seasonal average presented like: "1.8 (89)"

6) Figure 4 shows column totals on two days, Again the time of these observations (noon-4PM must be clearly stated in the figure caption. If in fact the maps shown are averages of 2-3 overfly times, then it might be interesting to compare maps during the individual satellite passage times, so readers could note correlations and/or changes from each sequential orbital overpass. However, maybe the measurements from individual satellite passages is not available.

7) Please show the larger boxed "analysis areas" (rectangles) shown in Fig 3 on the much larger domain shown in Fig 4, and identify those focussed areas in the caption of Fig. 4.

8) At line 268 the authors compare surface measurements with satellite column measurements, and they simply convert both the mean and standard deviation to concentration units by assuming a fixed PBL depth. However, there is considerably uncertainty in the PBL depth used to convert column density to concentration.  The standard deviation or uncertainty should INCREASE substantially (by 50% or more) due to uncertainties in the assumed PBL depth. I understand the standard deviation of the observations and the uncertainty of the PBL depth are two different sources of any "+" or "-" ranges, but even here, the authors should use a reasonable low and high limit (say 700 m or 1500 m), then show maybe two concentrations derived from the satellite columns using the two PBL depths.

9) Fig 5 shows daily HCHO columns. As noted before, the caption should read: "Column mean averaged during noon-4PM each cloud-free day during 2020". It would be interesting to show individual column measurements if there were more than one per day at these sights. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

GENERAL COMMENT

The work presented in this paper is related to Copernicus TROPOMI / Sentinel 5P and fully falls within the scope of the Remote Sensing Journal. TROPOMI data were intensively used to study the COVID on air quality improvement due to the lockdown and were focused on NO2. Here the authors are used TROPOMI data to study HCHO in Brasilian context. That is particularly interesting because Brazil brings together several characteristic sources of formaldehyde emitters: forest (several kinds), biomass fire (forest and sugarcane crops), intensive used of biofuels

The paper is correctly built is exhaustive references. I find the introduction particularly clear and very well suited to a journal that is not necessarily specialized in atmospheric chemistry

SPECIFIC COMMENTS

1.     Introduction : well done !

2.       Data and Methods

2.1.  Formaldehyde columns :

2.2.  Study area and analysis windows

·       The different areas of interest are discerning to cover the main HCHO sources  and choices are well explained

·       Scale is missing on figure 2

2.3.   Meteorological parameters and auxiliary data

2.4.  Estimate of HCHO surface concentration

·       At this satellite scale, considering the concentration homogeneous in the PBL is acceptable. Our understanding is that the authors had data of two radiosoundings per day, one day and one night, as well as the observations of HCHO columns. Can the authors specify if they have done the average separating day and night data?

3.     Results and Discussion

3.1.  Annual and seasonal statistics

·       Please precise how the confidence interval is computed ?

·       Could you make the use of Turkey Test most explicit. The use of this test for multiple comparison may be a problem (See : Gurvich, V.; Naumova, M. Logical Contradictions in the One-Way ANOVA and Tukey–Kramer Multiple Comparisons Tests with More Than Two Groups of Observations. Symmetry 2021, 13, 1387. https://doi.org/10.3390/sym13081387 )

·       The averaged maps for each season are missing. I strongly encourage authors to produce them. They would allow to visualize both the  seasonal variations and the impact of the dense urban area (city of Sao Paulo)

·       Why no test to see if there is a significant difference between seasonal mean is done ?

3.2.  Case studies

·       Good initiative to study the highest value. As there are only few measurements per day, episode of several days might be more robust ?

·       The backward trajectories are interesting, but the NOAA Hysplit is not adequate (typical lifetime of HCHO is a few hours)

·       Figure 6  a) It should be interesting not to limited the fires inside MASP ?

·       Figure 7a) We do not recognize the MASP borders?

4.     Conclusion

·       An interesting prospect would be to follow the evolution over several years. an annual update will make it possible to follow the conjunction of several phenomena: forest fire, evolution of anthropogenic pollution, climate change

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

This is a well written paper about saellite based formaldehyde burdens in the Sao Paulo Brazil region.  The paper contains no major issues and is essentially ready for publication now.  

Thesre are one or two minor comments that the authors might consider:

1.  Under Fu\ig. 2 Para 2--PM10 is not coarse PM but measured as coarse plus fine PM.

2. Clapeyron equation--might use ideal gas law --readers would better understand this name.

3.  Paras 3 below Table 1--might  include HCHO concentration for China; next para also for SE US and Amazon for comparison.

4. Para below Fig. 7 might add examples of high O3 concentration examples compare with wet season.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Back to TopTop