the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 26 Feb 2025
Seasonal isoprene emission estimates over tropical South America inferred from satellite observations of isoprene
Abstract. Isoprene, a volatile organic compound (VOC) emitted by plants, plays a significant role in atmospheric chemistry and climate. The Amazon rainforest is a globally-relevant source of atmospheric isoprene, influencing regional and global atmospheric composition. We report isoprene emissions for 2019 inferred from the Cross-track Infrared Sounder (CrIS) and the local sensitivities between isoprene emissions and isoprene columns determined by the GEOS-Chem chemical transport model. Compared with the MEGAN bottom-up inventory of isoprene emissions, the isoprene emission estimates inferred from CrIS have different spatial and seasonal distributions with generally lower emission rates but with higher emission rates over the north of Amazon basin and southeast of Brazil. Isoprene mole fraction data collected at the Amazon Tall Tower Observatory (ATTO) are invaluable for evaluating isoprene emission estimates. The observed mean isoprene concentration at ATTO, March—December 2019, is 3.0 ± 2.2 ppbv, which is reproduced better by the GEOS-Chem model driven by isoprene emissions inferred from CrIS (2.8 ± 1.4 ppbv) than by the MEGAN inventory (4.1±1.3 ppbv). GEOS-Chem model formaldehyde (HCHO) columns, corresponding to isoprene emissions inferred from CrIS, are generally more consistent with TROPOMI data (normalized mean error, NME = 43 %) than the HCHO columns corresponding to MEGAN isoprene emissions (NME = 50 %), as expected. They also improve the model agreement with regional TROPOMI HCHO:NO2 column ratios that are indicative of changes in photochemical regime. Our results provide confidence that we can use CrIS data to examine future impacts of anthropogenic activities on isoprene emissions from the Amazon.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Atmospheric Chemistry and Physics. The authors have no other competing interests to declare.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.- Preprint
(3578 KB) - Metadata XML
-
Supplement
(1328 KB) - BibTeX
- EndNote
Status: open (until 09 Apr 2025)
-
RC1: 'Comment on egusphere-2025-778', Anonymous Referee #1, 25 Mar 2025
reply
This study employs modeled isoprene column:emission rations from GEOS-Chem to derive monthly isoprene emissions for 2019 over tropical South America based on the RAL IMS CrIS isoprene column retrieval. The authors evaluate their results using in situ isoprene concentrations at the ATTO site in Brazil as well as TROPOMI HCHO columns. The CrIS-based isoprene emissions resulted in modest improvements in model HCHO and isoprene with respect to MEGAN, with a small impact on model ozone over the region.
This study presents a novel use of a new satellite product and is well within the scope of ACP. However, I think the manuscript needs significant work before it can be considered for publication. My main concern is that it fails to consider the potential impact of documented low model NOx biases in the region, and how the resulting high isoprene lifetimes will bias the emissions derived from CrIS based on the model isoprene column:emission slopes. The authors are strongly encouraged to first consider adjusting the model NOx emissions based on a comparison to TROPOMI NO2, or at least perform a sensitivity study in which NOx emissions are adjusted to quantify how much it impacts their findings. I also think that this study would be improved by first including an evaluation of the RAL IMS isoprene product, as it is a new retrieval for which no validation has yet been published. The text could also use some clarification throughout; see below for specific recommendations.
Specific comments
Line 54-55: What is meant by “chemical networks”? Do the authors mean “chemical mechanisms” here?
Line 65: Since there are a few different CrIS isoprene retrievals available, it would be good to mention in the introduction which is used in this work.
Line 105: What soil NOx emission scheme was used in the simulation? This is important to know given the sensitivity of the column:emission relationship to isoprene lifetime.
Line 107-122: Some of this information would be better suited to the introduction rather than contained in the methods, since it reflects general uncertainties in our understanding of isoprene emission processes that are not unique or specific to the MEGAN model being introduced here.
Line 128: What do the authors mean by “replace any a priori information assumed by the retrieval” here? Don’t the authors apply both the averaging kernel and a priori profile from the RAL IMS retrieval to the GEOS-Chem profile in their comparisons?
Line 133: There is also a CrIS instrument onboard the NOAA-21 satellite, launched in 2022.
Line 142-143: “IMS column averages tend to be lower than those derived from surface-based observations when surface level concentrations are high…” I think the authors need to do more to put this statement into context, as it does not necessarily represent a problem and is in fact expected behavior for IR satellite retrievals that are more sensitive to the mid-troposphere. A reader might read this and incorrectly assume the retrieval has a low bias, when there has actually been no published validation (that I know of) for the RAL IMS isoprene retrieval. The latter fact also bears mentioning here.
Line 164: In addition to the bias correction to OMI and TROPOMI, does the HCHO comparison also include an application of observation operators to the GEOS-Chem profiles?
Line 182: I assume that B and S refer to the intercept and slope of the linear regression model here, but the authors should still define these variables.
Line 183-186: As the authors note, the slope in this isoprene column:isoprene emission comparison is mainly determined by the isoprene lifetime, and the slope in the HCHO column:isoprene emission comparison is determined both by the HCHO yield and lifetime. How do the authors account for the fact that these may be biased in the model, due to biases in NOx for example? These would potentially lead to large differences in the slope that is being used to derive emissions from the satellite observations.
Figure 1: Since the GEOS-Chem output has had scene dependent averaging kernels applied, I assume the monthly mean only reflects times when the CrIS data were also available (i.e. model fields were screened when CrIS data were discarded due to cloud or other quality concerns). Is that correct?
Line 270-274: These sentences are unclear. Are the authors saying that they compare the ATTO measurements to an average of the GEOS-Chem predictions for the grid box containing the site, and all adjacent grid boxes?
Lines 309-323: I found this discussion to be too general to be very useful to the reader. Also, the parameters in Figs 3c-f will better correlate with isoprene emissions than concentrations, so if the authors want to include them it would make more sense to do so in Figure 2. Most importantly, however, this section is missing a discussion of model NOx biases (and, thus, biases in isoprene lifetime) as a possible source of discrepancies between the observed and simulation concentrations. Assuming that the Hudman et al. (2012) soil NOx emission scheme was used in this work, there has been at least one study that found it significantly underestimates NOx (by a factor of 30) in the region (Liu et al., 2016). I strongly encourage the authors to include a sensitivity study adjusting the NOx in their simulation to see how it impacts their results.
Lines 335-344: While the CrIS-derived emissions yield modest improvements in the mean bias and error of GEOS-Chem HCHO with respect to TROPOMI, the spatial correlation is often unchanged or even degraded. Do the authors have some ideas as to why this is?
Lines 354-386: This section also needs to consider the potential impact of model NOx biases on the FNR and the resulting model ozone. Could the authors evaluate the model NO2 based on the TROPOMI NO2 and adjust NOx emission accordingly to see how it improves the model ozone with and without the CrIS-based isoprene emissions?
Technical comments
Line 18: Suggest changing “north of Amazon” to “northern Amazon” and changing “southeast of Brazil” to “southeast Brazil”
Line 31: “Influences” should be “influence”
Line 64: Insert “the” before “GEOS-Chem”
Line 132: Insert the word “launched” before “onboard”
Line 147: “am equatorial” should be “an equatorial”
Line 148: “collected” should be “collects”
Line 152: “We refer to the reader to a dedicated reported” should be edited to “We refer the reader to a dedicated report”
Line 175: Add parentheses around the year for the citation here.
Line 178: “relationships these” should be edited to “relationship of these”
Line 222: I think “NMB > 100%” should be “NMB > 0%” here? The NMB values reported earlier in the paragraph are both positive (> 0) but less than 100%.
Line 254: Insert “are” between “hotspots” and “collocated”
Line 271: Insert “the” before “ATTO”.
Line 290: Insert “of” between “because” and “low”
Line 300-302: This sentence is a fragment as is. Consider revising.
Line 338: Consider changing the word “at” to “in” or “of”
Line 345-346: This sentence is a fragment, consider revising.
Line 351: Insert “the” before “boundary layer”
Line 378: Suggest changing “to the central and southeast of Brazil” to “in central and southeast Brazil”
References
Hudman et al. (2012). Steps towards a mechanistic model of global nitric oxide emissions: implementation and space-based constraints. Atmos. Chem. Phys. 12, 7779-7795.
Liu et al. (2016). Isoprene photochemistry over the Amazon rainforest. Proc. Natl Acad. Sci. USA 113, 6125-6130.
Citation: https://doi.org/10.5194/egusphere-2025-778-RC1
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 92 | 20 | 7 | 119 | 11 | 5 | 5 |
- HTML: 92
- PDF: 20
- XML: 7
- Total: 119
- Supplement: 11
- BibTeX: 5
- EndNote: 5
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 62 | 38 |
| United Kingdom | 2 | 18 | 11 |
| Germany | 3 | 16 | 9 |
| China | 4 | 10 | 6 |
| France | 5 | 10 | 6 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 62