the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 19 Apr 2024
Assessment of satellite observation-based wildfire emissions inventories using TROPOMI data and IFS-COMPO model simulations
Abstract. Fires are a key component of the global carbon cycle and humans are changing their characteristics. Fire emission monitoring is important to keep track of those changes and TROPOMI satellite observations of tropospheric nitrogen dioxide, carbon monoxide and the absorbing aerosol index can be used to quantify and verify the accuracy and precision of global wildfire emission estimates on a daily basis. Here we use TROPOMI observations to evaluate a new fire emission database based on Global Fire Atlas input for the Sense4Fire project (GFA-S4F) and from the Copernicus Atmosphere Monitoring (CAMS) Global Fire Assimilation System (GFAS) for a number of test regions worldwide representative of the most important wildfire type environments. The main focus is on Amazon and Cerrado biomes (tropical rain forests and deforestation) during August–September 2020, but analyses are also made for a region in sub-Saharan Africa (savannah) as well as two regions in Siberia (steppe and boreal forests/tundra). GFA-S4F and GFAS fire emissions are used as input for global atmospheric composition model simulations based on IFS-COMPO, i.e. an extension of ECMWF’s Integrated Forecasting System (IFS) for simulating atmospheric composition. Comparing the model output with the TROPOMI observations then provides an indirect check on the realism of these emission estimates. Furthermore, for tropospheric nitrogen dioxide the IFS-COMPO model simulations are also used to estimate the model sensitivity of tropospheric nitrogen dioxide columns with respect to fire emission changes. This local relationship is used to optimize the fire NOx emissions directly using the Sentinel-5p nitrogen dioxide observations.
The results reveal that for small fires emission nitrogen dioxide estimates are realistic on average albeit with a large spread, i.e. for individual fires emissions can be significantly under or overestimated regardless of emission database. However, for large fires nitrogen dioxide emissions are systematically and largely overestimated in all four regions. The overestimation can be an order of magnitude or even more. For area total nitrogen dioxide emissions this “large fire bias” is of minor importance, i.e. total nitrogen dioxide emissions are dominated by small fires. The GFA-S4F emission estimates were characterized by a larger positive bias for large fire NO2 emission cases compared to GFAS. The source of this bias is not well understood. With optimized NO2 emissions by direct adjustment of emission using TROPOMI nitrogen dioxide observations the large positive bias can efficiently be resolved. Combined with an update of soil NOx emissions – causing too low background NOx levels – a fairly good agreement between IFS-COMPO and TROPOMI was reached.
Carbon monoxide was generally underestimated using GFAS emission (~50 % on average for the selected regions). Updating carbon monoxide emissions over the Amazon region by incorporating more Sentinel satellite data (GFA-S4F) did reduce this fire CO bias significantly (to ~25 % on average).
Overall, the results show that TROPOMI data allows for systematically identifying uncertainties and errors in satellite-data based fire emissions. The results also suggest that the use of dynamic emission factors may further improve satellite based global emissions inventories. In addition, the results also highlight that the use of TROPOMI data could be much more detailed and refined towards assessing individual fires on a daily basis for better understanding fire dynamics and to improve and diversify fire emission factors.
- Preprint
(4882 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed (peer review stopped)
-
RC1: 'Comment on egusphere-2024-732', Anonymous Referee #1, 25 May 2024
General Comments:
The article, “Assessment of satellite observation-based wildfire emissions inventories using TROPOMI data and IFS-COMPO model simulations” by de Laat, et al., reports an evaluation of NO2 and CO simulations using a chemistry transport model, IFS-COMPO, based on two fire emissions datasets derived from satellite observations (GFAS and GFA-S4F). Using NO2 and CO products retrieved from Sentinel 5p TROPOMI, the authors also applied a so-called b-method that relate changes in emissions to changes in the column amounts of these smoke constituents to scale the GFAS and GFA-S4F emissions datasets. Then, they performed several simulation experiments by varying the model processes/parameters and using the original and scaled emissions datasets in four regions representative of the Amazon/Cerrado, Siberian boreal, Eurasian steppe, and African savannah biomes. Finally, they compared their results with TROPOMI NO2 and CO data and found overall better agreement with the results of simulations based on the TROPOMI-scaled emissions.
The manuscript is poorly written with so many statements that are superfluous, inaccurate, unsubstantiated, or unreferenced. Some examples of the weaknesses include the following:
- The authors state that their study was focused on the Amazon/Cerrado region, although they added three additional regions to achieve some degree of biome representativeness. However, instead of explaining why they chose those specific regions or period of study (August–September, 2020), they went into a long explanation about their original plan to study 5°´5° regions but later decided to make them larger. That long explanation does to offer any value to the reader. It would have been more useful for them to provide logical explanations on why they chose their study regions despite admitting that TROPOMI validation data in those regions were nonexistent for NO2and only minimally available for CO (see lines 133–143 and 172–174). A valid question is why they did not choose regions where validation data are available? If the essence of this study is to evaluate the emissions inventories, how can they be evaluated without basic validation data?
- The authors referred to the application of their b-method as “emission optimization” (e.g., Line 322). This terminology is misleading, as they do not “optimize” the emission or the emission data. Instead, they are simply attributing all errors and uncertainties in the model to the input emission data by scaling them to agree with TROPOMI observations. This seems like a type of “bias adjustment” of the model simulations to match observations. The authors should not refer to their b-method as emission “optimization”, as what they have done does not in any way constitute optimization. Their b-method is correctly described in Lines 704–706 as follows: “A scaling approach was adopted to constrain bottom-up fire NOx emissions with TROPOMI NO2 observations, which relies on the local sensitivity of tropospheric NO2 column changes with respect to NOx emission changes (the β-method). This brought the emission variability much closer in line with those from GFAS, independent of which prior emission estimate was used.” Therefore, it is a “scaling” approach and should be described as such throughout the article, not as an “optimization” approach.
- Under the results section, in describing the satellite images showing fires in the Amazon, the authors write “There is some shallow convection present, but weather conditions are mainly dry.” It is not clear whether this statement is referring to the images or the season. Either way, the statement needs to be referenced or substantiated with facts/data. If this is just referring to Figure 1, the authors should have specifically described how they determined the presence of shallow convection and dry weather. If it is the season, they needed to provide references attesting to the predominance of such ambient conditions in the Amazon in September. On line 363, this same region is described as “moist and sunlit”. How were these attributes determined?
- In their various scatterplots and linear regression fitting that produce fair to poor correlation coefficients (e.g., Figures 4, 6, 7, and 8), the authors have resorted to data binning as a way to improve correlation coefficients without any justification. This is a very unusual, artificial, and confusing way to ‘improve’ results in data intercomparison/validation. It is more useful to simply present the correlation coefficients, even if they are not optimal, and try to explain why they are so and explore ways to improve the results, rather than simply binning and averaging data arbitrarily. Therefore, I recommend that the correlation coefficients of the binned data be excluded from the analysis.
- I do not see any justification for capping emissions, as described in Lines 443–449. The authors make this sound like a regulatory process of capping emissions at source. Truncating emission values in such a trial-and-error manner is not a sound scientific approach to improve correlation. If the authors have reason(s) to believe that high NOx emission values may be erroneous, they can state their justification(s) and filter them out, if this turns out to be the most reasonable and justifiable step.
- There is a pervasive use of the phrase “large tropospheric NO2 column” bias throughout the article. This should be avoided, as it does not appear to be a phenomenon that needs to be highlighted.
Specific Comments:
Line 1: The title should reflect the fact that the authors are evaluating only two of the multiple emissions inventories in existence. See example of the title of a prior ACP publication (Pan et al., 2020), which should also be cited in this article. If possible, it should also reflect the fact that the study is focused on NO2 and CO emissions.
Line 47: Delete ‘significantly’ and continue the sentence without parenthesis to read “… reduce this fire CO bias to ~25% on average”.
Line 51: It is not clear what is meant by “dynamic emission factors”. Such terms that are not common in the literature or obvious to readers should be explained, even in the abstract.
Line 60: “The associated time scales …” is vague. It is not clear what it is in reference to. This needs to be clarified.
Line 66: When citing just a few examples of references among multiple others that could be cited, it is preferable to precede the citations with ‘e.g.,’. In this specific case, there are tens, if not hundreds, of publications that attest to the ability of satellite sensors to provide fire-related observations. This is also applicable to many of the other citations in this manuscript.
Lines 70–72: “Each individual dataset …combining these datasets into one information system”. This statement appears extremely presumptuous, as the authors have not pursued this objective in this article, nor have they discussed ways to do it.
Lines 84–85: “…uncertainties in bottom-up emission estimates due to emission factors …” The authors should have clarified what it is about emission factors that are causing the uncertainties. If you mean the uncertainties in the emission factors, emission process, and tracer lifetime, it should have been clearly stated.
Line 93: “…possibilities for monitoring and studying fires.” TROPOMI does not offer possibilities for studying “fires”. You probably meant “possibilities for observing fire emissions”.
Lines 104–105: “Atmospheric chemical composition modelling is used as an interface between Sentinel 2 and sentinel 3 based emissions vs. TROPOMI observations.” This statement is not helpful. It is not obvious what is meant by “Sentinel 2 and sentinel 3 based emissions”. You need to name the model and the exact emissions data products that are being discussed, as there may be some other emissions products that use Sentinel 2 and Sentinel 3 observations besides the two that are being evaluated in this manuscript.
Line 122: Something is missing between ‘–‘ and ‘of’. Perhaps modify to “– allowing for the retrieval of …”
Line 134: Add “to” between “due” and “lack”
Line 167: Delete the comma after ‘observations’.
Line 257–258: It is not clear what is meant by “… all relevant aspects that are required when matching observation data to model data …”. Please list/describe those relevant aspects, so the reader can understand what is being discussed.
Line 363: There was an earlier statement that “weather conditions are mostly dry”. What do the authors mean by describing this as a “moist” region? If you mean an evergreen forested region, please simply say so and explain its relevance (if any) to the lifetime of tropospheric NO2.
Line 372: “2-D probability distributions”. As far as I can see, these are scatterplots representing the point density in color. I have seen it called “2-D density plot” or “2-D histogram”.
Lines 396 – 397: It is not clear what is meant by “…twenty equally distributed TROPOMI data intervals …” This needs to be explained better. Also, there needs to be a better justification of why the data were averaged and why the averaging window/box size is 20. Why is it not 9 or 11 or 15 or 21 or 25?
Line 424–425: “Strongly enhanced IFS-COMPO tropospheric NO2 column values in this region are predominantly associated with fire emissions rather than emissions from other sources.” This statement needs to be substantiated with evidence or relevant references.
Line 488: “Siberian steppe”: The area shown on Figure A1 as representing this vegetation type is not in Siberia. This vegetation type in the region indicated on the map is normally referred to as “Eurasian steppe”.
Line 520: “which is opposite from the Amazon/Cerrado region.” The first part of this sentence is missing.
Line 523: What do you mean by “dynamical range” in this context?
Lines 528 – 529: “Given that there are fewer fires in Siberia in the particular period studied here, …” Fires are seasonal in different parts of the world and the fire season differs from region to region. What prevented the authors from studying this region during its own fire season or selecting another region whose fire season coincides with the selected study period?
Line 541: This subsection title for Carbon Monoxide is good. Why was there not a similar subsection title for NO2?
Lines 555–558: “This suggests…in the NOx emission factor”. This sentence is long and confusing and needs to be rephrased and made clearer.
Line 566: “amplitude of emissions” is not the appropriate phrase to describe the emission pattern, as it is not a sinusoidal wave, at least not within the time range covered in the study.
Line 600: The use of the term “state-of-the-art” here is misleading and would convey the impression that the community regards GFAS as “state-of-the-art”. This type of potential confusion should be avoided.
Line 682: It is not clear what the authors mean by “a first analysis”. Many recent studies, including several of those cited in this article, have analyzed “TROPOMI satellite observations of fire plumes affecting atmospheric composition…” Are the authors saying that this is the first time anyone has done this?
Line 696: “…as it reflects a lack of understanding.” It is not clear who lacks the understanding of what.
Lines 702 – 703: “…the emission factors that define the ratio between CO and NOx emissions.” This conveys the impression that emission factors define the ratio between these two gases, whereas they each have a different emission factor, which is the ratio between each emission and the corresponding dry biomass burned.
Line 715: Not “perse” but “per se”.
Figures and Captions:
Figure 1: There is no need for both 1a and 1b. Just one of them, perhaps 1b, should be enough. NASA WorldView is made to be user friendly, allowing image clips to be easily saved. I do not see any need to use a separate Python script by Brian Blaylock.
Figures 2, 3, and 4: The bright green circles representing fire detections and FRP are difficult to see. Use a different color scheme or give the green circles black outlines to make them appear more prominent. The figure caption is confusing without labeling the four panels a, b, c, and d. Indicating the FRP as having arbitrary units is a misrepresentation. You may say that the radii (or diameters) of the circles are scaled to be roughly proportionally to the magnitudes of the FRP, without adding “arbitrary unit”.
Figure 5: The caption mentions “(lower plot)”, but the plots are side-by-side and there is no upper or lower plot. This type of confusion is also why it is better to distinguish panels in a figure by labeling them with a, b, c …, so that their description in the caption and text would be easier to follow.
Figure A9: Panels (A, B) and (E, F) are the same. The plots for sub-equatorial African region are missing.
Tables and Captions:
Table 1: This Table caption is too long. A large middle portion of the caption “All simulations used …of the average values.” can be moved to the main text.
References:
Pan X, Ichoku C, Chin M, Bian H, Darmenov A, Colarco P, Ellison L, Kucsera T, da Silva A, Wang J, Oda T., 2020. Six global biomass burning emission datasets: intercomparison and application in one global aerosol model. Atmospheric Chemistry and Physics. 20(2), 969-994.
Citation: https://doi.org/10.5194/egusphere-2024-732-RC1 -
RC2: 'Comment on egusphere-2024-732', Anonymous Referee #2, 11 Jun 2024
The study by de Laat et al. makes use of two satellite-based fire emissions datasets (GFAS and GFA-S4F – the latter based on the Global Fire Atlas) in order to explore the performance of an atmospheric model (IFS-COMPO) when it comes to simulating NO2 and CO, as observed from the TROPOMI instrument on board the Sentinel 5p satellite. The authors performed sensitivity simulations with the atmospheric model, where the different emissions were employed, and also scaled or capped, in order to explore how performance may be improving or degrading following such perturbations. Central to the work presented is also the use of the ‘β-method’ for updating the emissions, based on the agreement of the resulting columns of NO2 and CO with TROPOMI.
The manuscript is on an interesting topic, certainly within the scope of ACP, and will be a useful contribution to progressing the science of fire emissions estimations, making use of cutting-edge Earth observation datasets. It is clearly written in terms of Egnlish and reasonably structured. However, it is also somewhat ‘loose’ when it comes to explaining or justifying certain aspects of the work pursued. Just some examples:
- Why was that period chosen? Was that a ‘typical’ period? Was it anomalous in any way(s)?
- Why those regions? The non-Amazon regions seem a bit like an add-on.
- Why the emissions capping? How can it be justified?
I recommend that the authors read the manuscript carefully and identify such places where statements could be more clear and well-supported. I have also made an effort to highlight some further specific areas for improvement below. Following the amendments suggested (I ticked the ‘minor revisions’ box, but it is somewhere between minor and major, I would say), I believe that the article would be suitable for publication in ACP.
SPECIFIC COMMENTS:
Page 2, Lines 19-22: Here I suggest applying the following rephrasing, so that it becomes clearer to the reader what was done:
“Here we use TROPOMI observations to evaluate a new fire emission database based on Global Fire Atlas input for the Sense4Fire project (GFA-S4F) and from the Copernicus Atmosphere Monitoring (CAMS) Global Fire Assimilation System (GFAS) for a number of test regions worldwide representative of the most important wildfire type environments.”
to
“Here we use TROPOMI observations to evaluate two fire emissions datasets: a new database that was developed for the Sense4Fire project based on Global Fire Atlas input (GFA-S4F), and the Copernicus Atmosphere Monitoring (CAMS) Global Fire Assimilation System (GFAS). We perform the evaluation for a number of test regions worldwide representative of the most important wildfire type environments.”
Page 4, Lines 56-64: Would be good to mention the importance of fire for atmospheric composition and chemistry as well. Both because it is another important angle, and also because the species used here for evaluation/optimization are short-lived pollutants rather than greenhouse gases. So it would be good if the reader reads about how such constituents are affected by fire up front.
Page 5, Line 101: Here I suggest starting the sentence as follows “Here, we show results from the ESA Sense4Fire project (S4F)…”, in order to make it clear that the manuscript presents work from S4F.
Page 5, Line 106: It later becomes clear that aerosols (AAI) is also utilized. Would be good to mention it here.
Page 7, Line 128: What does ‘offline’ mean in this context?
Page 10, Line 205: Suggest changing ‘show’ to ‘feature’.
Pages 10-11, Lines 216-228: There is no mention of the acronym ‘GFA-S4F’ in this section, which is an acronym used in many parts of this manuscript when referring to the Global Fire Atlas-based emissions (including the Abstract, Discussion, Conclusion, main tables etc). Therefore it should be mentioned here, where those emissions are actually described.
Page 11, Line 223: Why that specific period. And was there anything anomalous for the year/those months for the regions explored? Some more discussion and description is needed.
Page 12, Line 262: OK, but what is the time resolution of model output?
Page 12, Line 249: Expand text in parenthesis to be ‘(see Appendix for further description of the model)’.
Page 15, Line 311: ‘based’ -> ‘based on’
Page 15, Line 315: ‘the larger emission’ -> ‘the larger the emissions’
Page 15, Line 320: ‘given the’ can be removed.
Page 18, Line 354: Should ‘cloud pressure’ be ‘cloud base pressure’?
Page 19, Line 363: Wouldn’t say this is entirely the case for aerosols – both in general and in the particular depiction.
Page 18, Figure 2: The circles are not particularly visible. Wouldn’t simply black circles (no fill and thin black line at the periphery) of varying sizes work better?
Page 22, Figure 4: It will not be clear to the reader what CIFS is, as it is not explained earlier in the text. Generally best to use the same naming as in other figures (e.g. Fig. 6). Also, this figure would benefit from panel numbering (letters), so as to refer to each one specifically in the text, and avoid any potential confusion.
Page 23, Lines 401-onwards: It is not clear which part of the figure the reader should start looking at first.
Page 24, Lines 424-425: How is this statement supported? I believe it, but it needs some reference or pointing to another part of the paper.
Page 26, Lines 443-449: Capping emissions this way seems a bit too crude. Is there any justification for excluding emissions above those limits?
Page 26, Line 451: For consistency with GFA.BETA, why is BASE.BETA not called GFAS.BETA?
Page 27, Lines 461-462: Or that the emissions datasets are commonly biased?
Page 27, Line 466: ‘were identified to be comparatively low’ – provide evidence/support statement.
Page 29, Line 495: What does ‘dynamical range’ mean here?
Page 32, Table 2: It is not clear what ‘but for baseline simulations’ means. Please clarify.
Page 35, Line 569: ‘area-and time-averaged’ -> ‘area- and time-averaged’
Page 35, Lines 568-569: ‘and provides results that are very similar compared to the other estimates (GFAS and GFAS β-optimized) in terms of temporal variability’ – is this a good thing?
Page 35, Lines 569-570: ‘while the area-and time-averaged emission totals, quantified in terms of daily mean emissions, are overall reduced’ – again, is this a good thing?
Page 36, Line 572: ‘The β-optimization nevertheless can largely correct for this bias’ – where do we see this?
Page 37, Figure 9: Please also describe the lower panel in the caption. Also, it is again very confusing that GFAS is called BASE in the figure.
Page 38, Line 599: Would be good to outline explicitly the ‘two significant biases’ here in order to aid the reader.
Page 39, Lines 629-631: Can this statement about laboratory measurements be supported by a reference?
Page 40-41, Lines 657-659: Again, reference is needed to support these statements.
Page 43, Lines 714-717: This sentence needs revising to become more clearer. Also ‘perse’ -> ‘per se’.
Citation: https://doi.org/10.5194/egusphere-2024-732-RC2 - RC3: 'Comment on egusphere-2024-732', Anonymous Referee #3, 08 Jul 2024
Status: closed (peer review stopped)
-
RC1: 'Comment on egusphere-2024-732', Anonymous Referee #1, 25 May 2024
General Comments:
The article, “Assessment of satellite observation-based wildfire emissions inventories using TROPOMI data and IFS-COMPO model simulations” by de Laat, et al., reports an evaluation of NO2 and CO simulations using a chemistry transport model, IFS-COMPO, based on two fire emissions datasets derived from satellite observations (GFAS and GFA-S4F). Using NO2 and CO products retrieved from Sentinel 5p TROPOMI, the authors also applied a so-called b-method that relate changes in emissions to changes in the column amounts of these smoke constituents to scale the GFAS and GFA-S4F emissions datasets. Then, they performed several simulation experiments by varying the model processes/parameters and using the original and scaled emissions datasets in four regions representative of the Amazon/Cerrado, Siberian boreal, Eurasian steppe, and African savannah biomes. Finally, they compared their results with TROPOMI NO2 and CO data and found overall better agreement with the results of simulations based on the TROPOMI-scaled emissions.
The manuscript is poorly written with so many statements that are superfluous, inaccurate, unsubstantiated, or unreferenced. Some examples of the weaknesses include the following:
- The authors state that their study was focused on the Amazon/Cerrado region, although they added three additional regions to achieve some degree of biome representativeness. However, instead of explaining why they chose those specific regions or period of study (August–September, 2020), they went into a long explanation about their original plan to study 5°´5° regions but later decided to make them larger. That long explanation does to offer any value to the reader. It would have been more useful for them to provide logical explanations on why they chose their study regions despite admitting that TROPOMI validation data in those regions were nonexistent for NO2and only minimally available for CO (see lines 133–143 and 172–174). A valid question is why they did not choose regions where validation data are available? If the essence of this study is to evaluate the emissions inventories, how can they be evaluated without basic validation data?
- The authors referred to the application of their b-method as “emission optimization” (e.g., Line 322). This terminology is misleading, as they do not “optimize” the emission or the emission data. Instead, they are simply attributing all errors and uncertainties in the model to the input emission data by scaling them to agree with TROPOMI observations. This seems like a type of “bias adjustment” of the model simulations to match observations. The authors should not refer to their b-method as emission “optimization”, as what they have done does not in any way constitute optimization. Their b-method is correctly described in Lines 704–706 as follows: “A scaling approach was adopted to constrain bottom-up fire NOx emissions with TROPOMI NO2 observations, which relies on the local sensitivity of tropospheric NO2 column changes with respect to NOx emission changes (the β-method). This brought the emission variability much closer in line with those from GFAS, independent of which prior emission estimate was used.” Therefore, it is a “scaling” approach and should be described as such throughout the article, not as an “optimization” approach.
- Under the results section, in describing the satellite images showing fires in the Amazon, the authors write “There is some shallow convection present, but weather conditions are mainly dry.” It is not clear whether this statement is referring to the images or the season. Either way, the statement needs to be referenced or substantiated with facts/data. If this is just referring to Figure 1, the authors should have specifically described how they determined the presence of shallow convection and dry weather. If it is the season, they needed to provide references attesting to the predominance of such ambient conditions in the Amazon in September. On line 363, this same region is described as “moist and sunlit”. How were these attributes determined?
- In their various scatterplots and linear regression fitting that produce fair to poor correlation coefficients (e.g., Figures 4, 6, 7, and 8), the authors have resorted to data binning as a way to improve correlation coefficients without any justification. This is a very unusual, artificial, and confusing way to ‘improve’ results in data intercomparison/validation. It is more useful to simply present the correlation coefficients, even if they are not optimal, and try to explain why they are so and explore ways to improve the results, rather than simply binning and averaging data arbitrarily. Therefore, I recommend that the correlation coefficients of the binned data be excluded from the analysis.
- I do not see any justification for capping emissions, as described in Lines 443–449. The authors make this sound like a regulatory process of capping emissions at source. Truncating emission values in such a trial-and-error manner is not a sound scientific approach to improve correlation. If the authors have reason(s) to believe that high NOx emission values may be erroneous, they can state their justification(s) and filter them out, if this turns out to be the most reasonable and justifiable step.
- There is a pervasive use of the phrase “large tropospheric NO2 column” bias throughout the article. This should be avoided, as it does not appear to be a phenomenon that needs to be highlighted.
Specific Comments:
Line 1: The title should reflect the fact that the authors are evaluating only two of the multiple emissions inventories in existence. See example of the title of a prior ACP publication (Pan et al., 2020), which should also be cited in this article. If possible, it should also reflect the fact that the study is focused on NO2 and CO emissions.
Line 47: Delete ‘significantly’ and continue the sentence without parenthesis to read “… reduce this fire CO bias to ~25% on average”.
Line 51: It is not clear what is meant by “dynamic emission factors”. Such terms that are not common in the literature or obvious to readers should be explained, even in the abstract.
Line 60: “The associated time scales …” is vague. It is not clear what it is in reference to. This needs to be clarified.
Line 66: When citing just a few examples of references among multiple others that could be cited, it is preferable to precede the citations with ‘e.g.,’. In this specific case, there are tens, if not hundreds, of publications that attest to the ability of satellite sensors to provide fire-related observations. This is also applicable to many of the other citations in this manuscript.
Lines 70–72: “Each individual dataset …combining these datasets into one information system”. This statement appears extremely presumptuous, as the authors have not pursued this objective in this article, nor have they discussed ways to do it.
Lines 84–85: “…uncertainties in bottom-up emission estimates due to emission factors …” The authors should have clarified what it is about emission factors that are causing the uncertainties. If you mean the uncertainties in the emission factors, emission process, and tracer lifetime, it should have been clearly stated.
Line 93: “…possibilities for monitoring and studying fires.” TROPOMI does not offer possibilities for studying “fires”. You probably meant “possibilities for observing fire emissions”.
Lines 104–105: “Atmospheric chemical composition modelling is used as an interface between Sentinel 2 and sentinel 3 based emissions vs. TROPOMI observations.” This statement is not helpful. It is not obvious what is meant by “Sentinel 2 and sentinel 3 based emissions”. You need to name the model and the exact emissions data products that are being discussed, as there may be some other emissions products that use Sentinel 2 and Sentinel 3 observations besides the two that are being evaluated in this manuscript.
Line 122: Something is missing between ‘–‘ and ‘of’. Perhaps modify to “– allowing for the retrieval of …”
Line 134: Add “to” between “due” and “lack”
Line 167: Delete the comma after ‘observations’.
Line 257–258: It is not clear what is meant by “… all relevant aspects that are required when matching observation data to model data …”. Please list/describe those relevant aspects, so the reader can understand what is being discussed.
Line 363: There was an earlier statement that “weather conditions are mostly dry”. What do the authors mean by describing this as a “moist” region? If you mean an evergreen forested region, please simply say so and explain its relevance (if any) to the lifetime of tropospheric NO2.
Line 372: “2-D probability distributions”. As far as I can see, these are scatterplots representing the point density in color. I have seen it called “2-D density plot” or “2-D histogram”.
Lines 396 – 397: It is not clear what is meant by “…twenty equally distributed TROPOMI data intervals …” This needs to be explained better. Also, there needs to be a better justification of why the data were averaged and why the averaging window/box size is 20. Why is it not 9 or 11 or 15 or 21 or 25?
Line 424–425: “Strongly enhanced IFS-COMPO tropospheric NO2 column values in this region are predominantly associated with fire emissions rather than emissions from other sources.” This statement needs to be substantiated with evidence or relevant references.
Line 488: “Siberian steppe”: The area shown on Figure A1 as representing this vegetation type is not in Siberia. This vegetation type in the region indicated on the map is normally referred to as “Eurasian steppe”.
Line 520: “which is opposite from the Amazon/Cerrado region.” The first part of this sentence is missing.
Line 523: What do you mean by “dynamical range” in this context?
Lines 528 – 529: “Given that there are fewer fires in Siberia in the particular period studied here, …” Fires are seasonal in different parts of the world and the fire season differs from region to region. What prevented the authors from studying this region during its own fire season or selecting another region whose fire season coincides with the selected study period?
Line 541: This subsection title for Carbon Monoxide is good. Why was there not a similar subsection title for NO2?
Lines 555–558: “This suggests…in the NOx emission factor”. This sentence is long and confusing and needs to be rephrased and made clearer.
Line 566: “amplitude of emissions” is not the appropriate phrase to describe the emission pattern, as it is not a sinusoidal wave, at least not within the time range covered in the study.
Line 600: The use of the term “state-of-the-art” here is misleading and would convey the impression that the community regards GFAS as “state-of-the-art”. This type of potential confusion should be avoided.
Line 682: It is not clear what the authors mean by “a first analysis”. Many recent studies, including several of those cited in this article, have analyzed “TROPOMI satellite observations of fire plumes affecting atmospheric composition…” Are the authors saying that this is the first time anyone has done this?
Line 696: “…as it reflects a lack of understanding.” It is not clear who lacks the understanding of what.
Lines 702 – 703: “…the emission factors that define the ratio between CO and NOx emissions.” This conveys the impression that emission factors define the ratio between these two gases, whereas they each have a different emission factor, which is the ratio between each emission and the corresponding dry biomass burned.
Line 715: Not “perse” but “per se”.
Figures and Captions:
Figure 1: There is no need for both 1a and 1b. Just one of them, perhaps 1b, should be enough. NASA WorldView is made to be user friendly, allowing image clips to be easily saved. I do not see any need to use a separate Python script by Brian Blaylock.
Figures 2, 3, and 4: The bright green circles representing fire detections and FRP are difficult to see. Use a different color scheme or give the green circles black outlines to make them appear more prominent. The figure caption is confusing without labeling the four panels a, b, c, and d. Indicating the FRP as having arbitrary units is a misrepresentation. You may say that the radii (or diameters) of the circles are scaled to be roughly proportionally to the magnitudes of the FRP, without adding “arbitrary unit”.
Figure 5: The caption mentions “(lower plot)”, but the plots are side-by-side and there is no upper or lower plot. This type of confusion is also why it is better to distinguish panels in a figure by labeling them with a, b, c …, so that their description in the caption and text would be easier to follow.
Figure A9: Panels (A, B) and (E, F) are the same. The plots for sub-equatorial African region are missing.
Tables and Captions:
Table 1: This Table caption is too long. A large middle portion of the caption “All simulations used …of the average values.” can be moved to the main text.
References:
Pan X, Ichoku C, Chin M, Bian H, Darmenov A, Colarco P, Ellison L, Kucsera T, da Silva A, Wang J, Oda T., 2020. Six global biomass burning emission datasets: intercomparison and application in one global aerosol model. Atmospheric Chemistry and Physics. 20(2), 969-994.
Citation: https://doi.org/10.5194/egusphere-2024-732-RC1 -
RC2: 'Comment on egusphere-2024-732', Anonymous Referee #2, 11 Jun 2024
The study by de Laat et al. makes use of two satellite-based fire emissions datasets (GFAS and GFA-S4F – the latter based on the Global Fire Atlas) in order to explore the performance of an atmospheric model (IFS-COMPO) when it comes to simulating NO2 and CO, as observed from the TROPOMI instrument on board the Sentinel 5p satellite. The authors performed sensitivity simulations with the atmospheric model, where the different emissions were employed, and also scaled or capped, in order to explore how performance may be improving or degrading following such perturbations. Central to the work presented is also the use of the ‘β-method’ for updating the emissions, based on the agreement of the resulting columns of NO2 and CO with TROPOMI.
The manuscript is on an interesting topic, certainly within the scope of ACP, and will be a useful contribution to progressing the science of fire emissions estimations, making use of cutting-edge Earth observation datasets. It is clearly written in terms of Egnlish and reasonably structured. However, it is also somewhat ‘loose’ when it comes to explaining or justifying certain aspects of the work pursued. Just some examples:
- Why was that period chosen? Was that a ‘typical’ period? Was it anomalous in any way(s)?
- Why those regions? The non-Amazon regions seem a bit like an add-on.
- Why the emissions capping? How can it be justified?
I recommend that the authors read the manuscript carefully and identify such places where statements could be more clear and well-supported. I have also made an effort to highlight some further specific areas for improvement below. Following the amendments suggested (I ticked the ‘minor revisions’ box, but it is somewhere between minor and major, I would say), I believe that the article would be suitable for publication in ACP.
SPECIFIC COMMENTS:
Page 2, Lines 19-22: Here I suggest applying the following rephrasing, so that it becomes clearer to the reader what was done:
“Here we use TROPOMI observations to evaluate a new fire emission database based on Global Fire Atlas input for the Sense4Fire project (GFA-S4F) and from the Copernicus Atmosphere Monitoring (CAMS) Global Fire Assimilation System (GFAS) for a number of test regions worldwide representative of the most important wildfire type environments.”
to
“Here we use TROPOMI observations to evaluate two fire emissions datasets: a new database that was developed for the Sense4Fire project based on Global Fire Atlas input (GFA-S4F), and the Copernicus Atmosphere Monitoring (CAMS) Global Fire Assimilation System (GFAS). We perform the evaluation for a number of test regions worldwide representative of the most important wildfire type environments.”
Page 4, Lines 56-64: Would be good to mention the importance of fire for atmospheric composition and chemistry as well. Both because it is another important angle, and also because the species used here for evaluation/optimization are short-lived pollutants rather than greenhouse gases. So it would be good if the reader reads about how such constituents are affected by fire up front.
Page 5, Line 101: Here I suggest starting the sentence as follows “Here, we show results from the ESA Sense4Fire project (S4F)…”, in order to make it clear that the manuscript presents work from S4F.
Page 5, Line 106: It later becomes clear that aerosols (AAI) is also utilized. Would be good to mention it here.
Page 7, Line 128: What does ‘offline’ mean in this context?
Page 10, Line 205: Suggest changing ‘show’ to ‘feature’.
Pages 10-11, Lines 216-228: There is no mention of the acronym ‘GFA-S4F’ in this section, which is an acronym used in many parts of this manuscript when referring to the Global Fire Atlas-based emissions (including the Abstract, Discussion, Conclusion, main tables etc). Therefore it should be mentioned here, where those emissions are actually described.
Page 11, Line 223: Why that specific period. And was there anything anomalous for the year/those months for the regions explored? Some more discussion and description is needed.
Page 12, Line 262: OK, but what is the time resolution of model output?
Page 12, Line 249: Expand text in parenthesis to be ‘(see Appendix for further description of the model)’.
Page 15, Line 311: ‘based’ -> ‘based on’
Page 15, Line 315: ‘the larger emission’ -> ‘the larger the emissions’
Page 15, Line 320: ‘given the’ can be removed.
Page 18, Line 354: Should ‘cloud pressure’ be ‘cloud base pressure’?
Page 19, Line 363: Wouldn’t say this is entirely the case for aerosols – both in general and in the particular depiction.
Page 18, Figure 2: The circles are not particularly visible. Wouldn’t simply black circles (no fill and thin black line at the periphery) of varying sizes work better?
Page 22, Figure 4: It will not be clear to the reader what CIFS is, as it is not explained earlier in the text. Generally best to use the same naming as in other figures (e.g. Fig. 6). Also, this figure would benefit from panel numbering (letters), so as to refer to each one specifically in the text, and avoid any potential confusion.
Page 23, Lines 401-onwards: It is not clear which part of the figure the reader should start looking at first.
Page 24, Lines 424-425: How is this statement supported? I believe it, but it needs some reference or pointing to another part of the paper.
Page 26, Lines 443-449: Capping emissions this way seems a bit too crude. Is there any justification for excluding emissions above those limits?
Page 26, Line 451: For consistency with GFA.BETA, why is BASE.BETA not called GFAS.BETA?
Page 27, Lines 461-462: Or that the emissions datasets are commonly biased?
Page 27, Line 466: ‘were identified to be comparatively low’ – provide evidence/support statement.
Page 29, Line 495: What does ‘dynamical range’ mean here?
Page 32, Table 2: It is not clear what ‘but for baseline simulations’ means. Please clarify.
Page 35, Line 569: ‘area-and time-averaged’ -> ‘area- and time-averaged’
Page 35, Lines 568-569: ‘and provides results that are very similar compared to the other estimates (GFAS and GFAS β-optimized) in terms of temporal variability’ – is this a good thing?
Page 35, Lines 569-570: ‘while the area-and time-averaged emission totals, quantified in terms of daily mean emissions, are overall reduced’ – again, is this a good thing?
Page 36, Line 572: ‘The β-optimization nevertheless can largely correct for this bias’ – where do we see this?
Page 37, Figure 9: Please also describe the lower panel in the caption. Also, it is again very confusing that GFAS is called BASE in the figure.
Page 38, Line 599: Would be good to outline explicitly the ‘two significant biases’ here in order to aid the reader.
Page 39, Lines 629-631: Can this statement about laboratory measurements be supported by a reference?
Page 40-41, Lines 657-659: Again, reference is needed to support these statements.
Page 43, Lines 714-717: This sentence needs revising to become more clearer. Also ‘perse’ -> ‘per se’.
Citation: https://doi.org/10.5194/egusphere-2024-732-RC2 - RC3: 'Comment on egusphere-2024-732', Anonymous Referee #3, 08 Jul 2024
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 410 | 126 | 32 | 568 | 20 | 21 |
- HTML: 410
- PDF: 126
- XML: 32
- Total: 568
- BibTeX: 20
- EndNote: 21
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1