the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Nitrate chemistry in the northeast US part I: nitrogen isotope seasonality tracks nitrate formation chemistry
Claire Bekker
Wendell W. Walters
Lee T. Murray
Meredith G. Hastings
Abstract. Despite significant precursor emission reductions in the US over recent decades, atmospheric nitrate deposition remains an important terrestrial stressor. Here we utilized statistical air mass back trajectory analysis and nitrogen stable isotope deltas (δ(15N)) to investigate atmospheric nitrate spatiotemporal trends in the northeastern US from samples collected at three US EPA Clean Air Status and Trends Network (CASTNET) sites from December 2016–2018. For the considered sites, similar seasonal patterns in nitric acid (HNO3) and particulate nitrate (pNO3) concentrations were observed with spatial differences attributed to nitrogen oxide (NOx) emission densities in source contributing regions that were typically ≤1000 km. Significant spatiotemporal δ(15N) variabilities in HNO3 and pNO3 were observed with higher values during winter relative to summer, like previous reports from CASTNET samples collected in the early 2000s for our study region. In the early 2000s, δ(15N) of atmospheric nitrate in the Northeast US had been suggested to be driven by NOx emissions; however, we did not find significant spatiotemporal changes in the modeled NOx emissions by sector and fuel type or δ(15N, NOx) for the source regions of the CASTNET sites. Instead, the spatiotemporal trends were driven by δ(15N) fractionation associated with nitrate formation. Under the field conditions of low NOx relative to O3 concentrations and when δ(15N, NOx) emission sources do not have significant variability, we demonstrate that δ(15N) of atmospheric nitrate can be a robust tracer for diagnosing nitrate formation.
Claire Bekker et al.
Status: closed
-
RC1: 'Comment on acp-2022-621', Eva Stueeken, 15 Nov 2022
The authors present new isotopic data and model calculations for atmospheric nitrate phases in the northeastern USA. The results reveal seasonal isotopic patterns in the measured data, which cannot be explained by source variability alone. Instead, the authors propose that the isotopic composition of atmospheric nitrate is significantly impacted by secondary reactions within the atmosphere and that the temporal variability can be explained by changes in the formation pathway.
The manuscript is well written and underpinned by a strong dataset. I’m not an atmospheric chemist and cannot evaluate the model calculations, but as an isotope geochemist I found the paper interesting to read and overall compelling.
My main comment is to better discuss the importance of varying nitrate formation pathways. What causes them to change seasonally, and what can that tell us about the broader environment (either anthropogenic or natural processes)?
Apart from that, I only have a few minor points for clarification (to the non-specialist):
l. 60-61: This sentence needs to be simplified. Maybe write “Accounting for these isotope effects is important when using for δ(15N) as quantitative tracker…”
l. 85-86: This sampling description is lacking lots of details. Which containers were used? How were the filters applied? How much sample was collected? Please expand this paragraph and provide appropriate references.
l. 93: Perhaps also state that the data agree to within XX%. A 1:1 relationship could also arise if there were a significant but consistent offset between the datasets.
ll. 94-95: Does this mean that samples from four weeks were mixed together into one container? Please state this more clearly.
ll. 104-105: Name the model of the mass spectrometer. Was the same instrument used to measure D17O?
l. 185: Provide some quantitative comparison here for how high NOx levels used to be in the past. Otherwise, the reader is left wondering.
l. 228: Better rephrase to “This increase is likely due to a significant heating demand during this period”. Please also explain why other reasons can be ruled out, leaving heating (i.e., coal combustion?) as the most likely explanation. NOTE: Later in the discussion it is argued that the observed seasonal variability is caused by secondary fractionation effects rather than source variability. Hence this sentence here should be rewritten. Otherwise, it is very confusing.
l. 229: As above, change to “… possibly due to increased emissions related to electricity generation for cooling”. Explain why other reasons can be ruled out. NOTE (same as above): Later in the discussion it is argued that the observed seasonal variability is caused by secondary fractionation effects rather than source variability. Hence this sentence here should be rewritten. Otherwise, it is very confusing.
ll. 231-238: How could all these endmembers be calculated so accurately? The introduction mentions isotope data for only soils, liquid fuel combustion, vehiclces and coal combustion. How were all these other sources listed here isolated from mixed isotopic signals? Please explain how this was done.
l. 276: parenthesis missing in the equation
l. 285: change fractionation to reaction
l. 289: How low is low? Provide a quantitative threshold, so that one can compare this to the data.
ll. 331-332: How are the pathways changing? Please expand briefly.
l. 362: Again, remind the reader here how large this reduction was, in percent.
ll. 372-374: What is the importance of knowing the nitrate formation pathway? Can knowledge over the pathway help understand air quality or other parameters? Please expand.
Figure 3: Is there any significance to the nitrate speciation between HNO3 and particulate? Does that relate back to the nitrate formation pathway as well? This would be helpful to discuss in the manuscript.
Figure 4: Make the y-axes shorter in all plots, so that the data are more spread out.
Eva Stüeken
Citation: https://doi.org/10.5194/acp-2022-621-RC1 -
RC2: 'Comment on acp-2022-621', Pete D. Akers, 05 Jan 2023
General comments
In this manuscript, the authors examine atmospheric nitrate concentrations and d15N values at three sites in the northeastern United States over two years. They find clear seasonal cycles in total NO3− concentration and in d15N values. Modelling suggests that these cycles cannot be explained by seasonally changing NOx sources or by isotopic fractionation during NOx cycling. Rather, isotopic fractionation during the formation of nitrate provides the best match to observed d15N values and thus seasonally changing nitrogen formation pathways appear to drive d15N variability over a year. This is in contrast to previous studies which more prominently focused on NOx emissions as the key d15N driver.
Overall, I found the manuscript to be well-written and generally clear. The narrative is straightforward and generally well-balanced between providing enough information to follow the methodology and authors’ thoughts while not getting too bogged down in technical and modelling details. The work appears to have been performed well with a comprehensive set of analyses and models to investigate this field data. The figures are generally good, although Figure 1 needs substantial changes to meet the quality shown throughout the rest of the manuscript.
My comments are generally minor. My largest concern is that while the paper states that it is examining spatiotemporal variability, there is actually very little discussion of the spatial variability between sites. The WST site, in particular, has much lower NO3− concentrations than the other two sites, and smaller but consistent differences in d15N values are also mentioned across the three sites. Are there geographical or environmental differences between these sites that could explain these observations? Can these differences help you understand or interpret the model predictions better? Also, be careful describing things as spatiotemporal variability if you actually just discussing a broad temporal variability observed in a similar manner across the three sites.
Additionally, while the authors identify that the NO3 formation pathways are the most likely drivers of the temporal variability, the final discussion could use a bit more elaboration (for those with less expertise in atmospheric chemistry and modelling) on why these formation pathways differ seasonally. Also, it could be useful if the authors spoke to whether these seasonal pathways are expected to have been affected by air pollution and other atmospheric chemistry changes over the past several decades (i.e., would you expect these to have largely had the same effect on NO3− isotopes in the pre-industrial or during the peak of the NOx pollution as now?). This need not be an intense discussion, but rather adding some context for readers to better understand how applicable your findings could be and get a better idea of the importance.
Specific comments
Notation for d15N is non-standard and should be simply d15N. For specific chemical species, notation could either be d15N(NOx) or d15NNOx.
21: I know this is just the abstract, but “Instead, the spatiotemporal trends were driven by δ(15N) fractionation associated with formation” is pretty vague for being one of the primary findings of your paper. Could you give a little more specific information on what part of the formation process or reaction is driving this fractionation?
80: What are the elevations for these three sites? I see that they are located away from cities and point source pollution, but are they in primarily agricultural or natural settings? Do the three sites differ by surrounding land use or other environmental factors other than geographic coordinate?
107: This is a fairly small range of d15N values for your standards, and many of your samples have d15N values outside this calibration standard range. Can you speak to the quality of the corrections outside this range? Is there solid reason or evidence to assume that an extrapolated correction is still accurate and precise?
175: There’s no discussion of why there is such a dramatic difference in concentrations between WST and the other two sites. This would seemingly indicate that either the NO3 supplies and/or flux are very different for this site. Would this affect your ability to make region-wide conclusions? To that end, how much spatial variance in NO3 concentration and d15N are you expecting to see across the region? Should they all be fairly similar, or might you expect substantial local variations?
200: Again, there’s no discussion of what might be causing these spatial differences in the d15N values between sites.
205: This close match to theoretical values is nice to see in the field data!
255: Positively or negatively correlated?
Figure 1. This map needs substantial changes or wholesale remaking, as it is not up to the quality of the rest of the manuscript. There are no indications for the symbology and coloring of the map. There is no scale for the map, nor any indication of the particular projection of the map (it is probably Web Mercator based on Google Maps, but this is generally a poor projection choice for scientific use). Likewise, there are no geographic coordinates nor an inset map indicated to give geographical context to the selected area of the map. These introduction map figures give good opportunity to provide supplemental environmental and geographic information about the region, such as elevation or land cover, and as such you should consider having this figure give the reader more context and information than simply the three locations on a transportation map.
Figure 5: The maps in C, F, and I should also have Canada and Mexico present, particularly since the source regions extend into Canada.
Technical corrections
This may be caught in later proofing, but I believe the editorial style for Copernicus is one space after periods, and here throughout it appears to be 2 spaces (and in some cases, maybe 3 spaces).
12: Comma after Here
30: Consider writing this as “total atmospheric nitrate” so that it is more clear why the abbreviation has a “t” in it.
103: P. aureofaciens should be italicized. And the genus should probably be fully spelled out since it is not spelled out elsewhere. It may also be worth noting that this is a specific modified strain of the bacteria to lack N2O reductase, and not simply the wild-type.
156: I think Copernicus publications usually have these model Zenodo DOI citations included in the literature citations.
204: Parenthesis aren’t needed when simply stating a mean value difference.
Table 2: Standard deviations are more easily and typically shown as ± after the mean (i.e., Mean±SD, 0.526±0.200)
Figure 4: Is there a true need to distinguish between p <0.001, p < 0.0001, and p < 0.00001? Are these scientifically significant differences?
Figure 8. The color schemes should ideally be different between A and B, because they are not showing the same data groupings (e.g., the teal color in A doesn’t represent the same data as the teal in B).
For Figures 3+4 and 6+7, the same data is shown but in different forms. However, the grouping is different between the paired plots, as Fig 3 and 6 have columns by site and Fig 4 and 7 have columns by NO3 type. If it does not substantially affect the visual point you are trying to make with the plots, it would make it easier for the reader if the layout was the same for both paired plots.
Citation: https://doi.org/10.5194/acp-2022-621-RC2 - AC1: 'Comment on acp-2022-621', Wendell Walters, 16 Feb 2023
Status: closed
-
RC1: 'Comment on acp-2022-621', Eva Stueeken, 15 Nov 2022
The authors present new isotopic data and model calculations for atmospheric nitrate phases in the northeastern USA. The results reveal seasonal isotopic patterns in the measured data, which cannot be explained by source variability alone. Instead, the authors propose that the isotopic composition of atmospheric nitrate is significantly impacted by secondary reactions within the atmosphere and that the temporal variability can be explained by changes in the formation pathway.
The manuscript is well written and underpinned by a strong dataset. I’m not an atmospheric chemist and cannot evaluate the model calculations, but as an isotope geochemist I found the paper interesting to read and overall compelling.
My main comment is to better discuss the importance of varying nitrate formation pathways. What causes them to change seasonally, and what can that tell us about the broader environment (either anthropogenic or natural processes)?
Apart from that, I only have a few minor points for clarification (to the non-specialist):
l. 60-61: This sentence needs to be simplified. Maybe write “Accounting for these isotope effects is important when using for δ(15N) as quantitative tracker…”
l. 85-86: This sampling description is lacking lots of details. Which containers were used? How were the filters applied? How much sample was collected? Please expand this paragraph and provide appropriate references.
l. 93: Perhaps also state that the data agree to within XX%. A 1:1 relationship could also arise if there were a significant but consistent offset between the datasets.
ll. 94-95: Does this mean that samples from four weeks were mixed together into one container? Please state this more clearly.
ll. 104-105: Name the model of the mass spectrometer. Was the same instrument used to measure D17O?
l. 185: Provide some quantitative comparison here for how high NOx levels used to be in the past. Otherwise, the reader is left wondering.
l. 228: Better rephrase to “This increase is likely due to a significant heating demand during this period”. Please also explain why other reasons can be ruled out, leaving heating (i.e., coal combustion?) as the most likely explanation. NOTE: Later in the discussion it is argued that the observed seasonal variability is caused by secondary fractionation effects rather than source variability. Hence this sentence here should be rewritten. Otherwise, it is very confusing.
l. 229: As above, change to “… possibly due to increased emissions related to electricity generation for cooling”. Explain why other reasons can be ruled out. NOTE (same as above): Later in the discussion it is argued that the observed seasonal variability is caused by secondary fractionation effects rather than source variability. Hence this sentence here should be rewritten. Otherwise, it is very confusing.
ll. 231-238: How could all these endmembers be calculated so accurately? The introduction mentions isotope data for only soils, liquid fuel combustion, vehiclces and coal combustion. How were all these other sources listed here isolated from mixed isotopic signals? Please explain how this was done.
l. 276: parenthesis missing in the equation
l. 285: change fractionation to reaction
l. 289: How low is low? Provide a quantitative threshold, so that one can compare this to the data.
ll. 331-332: How are the pathways changing? Please expand briefly.
l. 362: Again, remind the reader here how large this reduction was, in percent.
ll. 372-374: What is the importance of knowing the nitrate formation pathway? Can knowledge over the pathway help understand air quality or other parameters? Please expand.
Figure 3: Is there any significance to the nitrate speciation between HNO3 and particulate? Does that relate back to the nitrate formation pathway as well? This would be helpful to discuss in the manuscript.
Figure 4: Make the y-axes shorter in all plots, so that the data are more spread out.
Eva Stüeken
Citation: https://doi.org/10.5194/acp-2022-621-RC1 -
RC2: 'Comment on acp-2022-621', Pete D. Akers, 05 Jan 2023
General comments
In this manuscript, the authors examine atmospheric nitrate concentrations and d15N values at three sites in the northeastern United States over two years. They find clear seasonal cycles in total NO3− concentration and in d15N values. Modelling suggests that these cycles cannot be explained by seasonally changing NOx sources or by isotopic fractionation during NOx cycling. Rather, isotopic fractionation during the formation of nitrate provides the best match to observed d15N values and thus seasonally changing nitrogen formation pathways appear to drive d15N variability over a year. This is in contrast to previous studies which more prominently focused on NOx emissions as the key d15N driver.
Overall, I found the manuscript to be well-written and generally clear. The narrative is straightforward and generally well-balanced between providing enough information to follow the methodology and authors’ thoughts while not getting too bogged down in technical and modelling details. The work appears to have been performed well with a comprehensive set of analyses and models to investigate this field data. The figures are generally good, although Figure 1 needs substantial changes to meet the quality shown throughout the rest of the manuscript.
My comments are generally minor. My largest concern is that while the paper states that it is examining spatiotemporal variability, there is actually very little discussion of the spatial variability between sites. The WST site, in particular, has much lower NO3− concentrations than the other two sites, and smaller but consistent differences in d15N values are also mentioned across the three sites. Are there geographical or environmental differences between these sites that could explain these observations? Can these differences help you understand or interpret the model predictions better? Also, be careful describing things as spatiotemporal variability if you actually just discussing a broad temporal variability observed in a similar manner across the three sites.
Additionally, while the authors identify that the NO3 formation pathways are the most likely drivers of the temporal variability, the final discussion could use a bit more elaboration (for those with less expertise in atmospheric chemistry and modelling) on why these formation pathways differ seasonally. Also, it could be useful if the authors spoke to whether these seasonal pathways are expected to have been affected by air pollution and other atmospheric chemistry changes over the past several decades (i.e., would you expect these to have largely had the same effect on NO3− isotopes in the pre-industrial or during the peak of the NOx pollution as now?). This need not be an intense discussion, but rather adding some context for readers to better understand how applicable your findings could be and get a better idea of the importance.
Specific comments
Notation for d15N is non-standard and should be simply d15N. For specific chemical species, notation could either be d15N(NOx) or d15NNOx.
21: I know this is just the abstract, but “Instead, the spatiotemporal trends were driven by δ(15N) fractionation associated with formation” is pretty vague for being one of the primary findings of your paper. Could you give a little more specific information on what part of the formation process or reaction is driving this fractionation?
80: What are the elevations for these three sites? I see that they are located away from cities and point source pollution, but are they in primarily agricultural or natural settings? Do the three sites differ by surrounding land use or other environmental factors other than geographic coordinate?
107: This is a fairly small range of d15N values for your standards, and many of your samples have d15N values outside this calibration standard range. Can you speak to the quality of the corrections outside this range? Is there solid reason or evidence to assume that an extrapolated correction is still accurate and precise?
175: There’s no discussion of why there is such a dramatic difference in concentrations between WST and the other two sites. This would seemingly indicate that either the NO3 supplies and/or flux are very different for this site. Would this affect your ability to make region-wide conclusions? To that end, how much spatial variance in NO3 concentration and d15N are you expecting to see across the region? Should they all be fairly similar, or might you expect substantial local variations?
200: Again, there’s no discussion of what might be causing these spatial differences in the d15N values between sites.
205: This close match to theoretical values is nice to see in the field data!
255: Positively or negatively correlated?
Figure 1. This map needs substantial changes or wholesale remaking, as it is not up to the quality of the rest of the manuscript. There are no indications for the symbology and coloring of the map. There is no scale for the map, nor any indication of the particular projection of the map (it is probably Web Mercator based on Google Maps, but this is generally a poor projection choice for scientific use). Likewise, there are no geographic coordinates nor an inset map indicated to give geographical context to the selected area of the map. These introduction map figures give good opportunity to provide supplemental environmental and geographic information about the region, such as elevation or land cover, and as such you should consider having this figure give the reader more context and information than simply the three locations on a transportation map.
Figure 5: The maps in C, F, and I should also have Canada and Mexico present, particularly since the source regions extend into Canada.
Technical corrections
This may be caught in later proofing, but I believe the editorial style for Copernicus is one space after periods, and here throughout it appears to be 2 spaces (and in some cases, maybe 3 spaces).
12: Comma after Here
30: Consider writing this as “total atmospheric nitrate” so that it is more clear why the abbreviation has a “t” in it.
103: P. aureofaciens should be italicized. And the genus should probably be fully spelled out since it is not spelled out elsewhere. It may also be worth noting that this is a specific modified strain of the bacteria to lack N2O reductase, and not simply the wild-type.
156: I think Copernicus publications usually have these model Zenodo DOI citations included in the literature citations.
204: Parenthesis aren’t needed when simply stating a mean value difference.
Table 2: Standard deviations are more easily and typically shown as ± after the mean (i.e., Mean±SD, 0.526±0.200)
Figure 4: Is there a true need to distinguish between p <0.001, p < 0.0001, and p < 0.00001? Are these scientifically significant differences?
Figure 8. The color schemes should ideally be different between A and B, because they are not showing the same data groupings (e.g., the teal color in A doesn’t represent the same data as the teal in B).
For Figures 3+4 and 6+7, the same data is shown but in different forms. However, the grouping is different between the paired plots, as Fig 3 and 6 have columns by site and Fig 4 and 7 have columns by NO3 type. If it does not substantially affect the visual point you are trying to make with the plots, it would make it easier for the reader if the layout was the same for both paired plots.
Citation: https://doi.org/10.5194/acp-2022-621-RC2 - AC1: 'Comment on acp-2022-621', Wendell Walters, 16 Feb 2023
Claire Bekker et al.
Claire Bekker et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 366 | 127 | 13 | 506 | 8 | 11 |
- HTML: 366
- PDF: 127
- XML: 13
- Total: 506
- BibTeX: 8
- EndNote: 11
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1