the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 31 Aug 2023
Surface snow bromide and nitrate at Eureka, Canada in early spring and implications for polar boundary layer chemistry
Abstract. This study explores the role of snowpack in polar boundary layer chemistry, especially as a direct source of reactive bromine (BrOX=BrO+Br) and nitrogen (NOX=NO+NO2) in the Arctic springtime. Surface snow samples were collected daily from a Canadian high Arctic location at Eureka, Nunavut (80° N, 86° W) from the end of February to the end of March in 2018 and 2019. The snow was sampled at several sites representing distinct environments: sea ice, inland close to sea level, and a hilltop ~600 m above sea level (asl). At the inland sites, surface snow salinity has a double-peak distribution with the first and lowest peak at 0.001–0.002 practical salinity unit (psu), which corresponds to the precipitation effect, and the second peak at 0.01–0.04 psu, which is likely related to the salt accumulation effect (due to loss of water vapour by sublimation). Snow salinity on sea ice has a triple-peak distribution; its first and second peaks overlap with the inland peaks, and the third peak at 0.2–0.4 psu is likely due to the sea water effect (due to upward migration of brine on sea ice). At all sites, snow sodium and chloride concentrations increase by almost 10-fold from the top 0.2 cm to ~1.5 cm in depth. Surface snow bromide at sea level is significantly enriched, indicating a net sink of atmospheric bromine. Moreover, surface snow bromide at sea level has an increasing trend over the measurement time period, with mean slopes of 0.024 in the 0–0.2 cm layer and 0.016 μM d-1 in the 0.2–0.5 cm layer. Surface snow nitrate at sea level also shows a significant increasing trend, with mean slopes of 0.27, 0.20, and 0.07 μM d-1 in the top 0.2 cm, 0.2–0.5 cm, and 0.5–1.5 cm layers, respectively. Using these trends, an integrated net deposition flux of bromide of 1.01×107 molecules cm-2 s-1 and an integrated net deposition flux of nitrate of 2.6×108 molecules cm-2 s-1 were derived. In addition, nitrate and bromide in the morning samples are significantly higher than the afternoon samples, indicating a strong photochemistry effect. However, the mean bromide loss rate (0.027–0.040 μM) is smaller than the nitrate loss rate (0.23–0.362 μM) by an order of magnitude, implying the reactive bromine emission flux from snowpack is significantly smaller than the reactive nitrogen emission flux, which is consistent with the large difference between their derived net deposition fluxes. After considering the photochemical loss effect, the corrected bromide deposition flux at sea level is 2.73×107 molecules cm-2 s-1; for nitrate, the corrected deposition flux is 5.98×108 molecules cm-2 s-1. In addition, the surface snow nitrate and bromide at inland sites were found to be significantly correlated (R=0.48–0.76), and the [NO3-]/[Br-] ratio of 4–7 indicates a possible acceleration effect of reactive bromine in atmospheric NOX-to-nitrate conversion. This is the first time such an effect has been seen in snow chemistry data obtained with a sampling frequency as short as one day.
- Preprint
(11753 KB) - Metadata XML
-
Supplement
(1159 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2023-1446', Anonymous Referee #1, 14 Sep 2023
Yang et al report observations of snow concentrations of ions along with surface atmosphere ozone and BrO concentrations at an Arctic location in the early spring of 2018 and 2019. To my knowledge, these are the first depth-dependent observations of snow bromide, which may be useful for understanding processes determining reactive bromine emissions from snow. They use time and depth-dependent measurements of bromide and nitrate to calculate the net deposition flux during the observational time period. They find that nitrate and bromide in snow are correlated and suggest that they are linked to one another through the formation and hydrolysis of bromine nitrate. They also find that deposition is confined to the surface skin layer.
This paper is very difficult to read, especially the long results section. There are a lot of details and numbers and it is presented in a way that makes it very difficult to discern the big picture. It reads like a first draft. The paragraph starting on line 424 is a particularly good example of this. There are more numbers than words in this paragraph and it is not readable. In general, the paper needs some reorganization and needs to be presented in a more succinct and readable manner. It often reads as a list of disconnected observations.
A large portion of the results section focuses on salinity, but in the end, it is not clear what they learned from it as the results section is difficult to read and there is no follow-up on the salinity observations in the discussion or conclusions section. It is also unclear how an iceberg will impact snow salinity on sea ice and land.
Abstract line 26: missing a unit after 0.024.
Methods: State the eluants used for the IC measurements.
Line 342: I think you mean to say that the “concentrations” are larger, not the “profiles”.
Line 419: What is a near zero increasing trend? Does this mean that the increasing trend is not statistically different from zero?
Section 3.5: I think these calculations represent a net deposition flux (deposition minus emissions) and this should be explicitly stated.
Citation: https://doi.org/10.5194/egusphere-2023-1446-RC1 -
AC1: 'Reply on RC1', Xin Yang, 01 Jan 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1446/egusphere-2023-1446-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Xin Yang, 01 Jan 2024
-
RC2: 'Comment on egusphere-2023-1446', Anonymous Referee #2, 16 Oct 2023
I find this manuscript confusing and the conclusions over-reaching the actual findings. The authors appear to see a net deposition flux of Br- to snowpack in one year (2019) and a much smaller one in the prior year (2018). The trend of net bromide deposition in 2019 appears to be out of error estimates. Beyond these solid findings, I think that the manuscript does not make clear arguments for how the net deposition flux of Br- is relevant to snowpack emission. Multiple other studies have shown that illuminated snowpack emits reactive halogens, and we know that reactive halogen eventually convert back to halides and deposit to snowpack. If these two processes (emission and deposition) are in balance, then the net deposition to snowpack can be near zero. Their finding of only small net deposition is not in conflict with other studies showing snowpack is a source of reactive bromine, and their finding of small net deposition does not preclude snowpack playing a significant role in reactive halogen production.
Error analysis:
Aspects reported in this manuscript seem to be smaller than detection limits. On line 203, it is stated that the limit of detection (LOD) for Br- is 0.2 micromolar. It appears from the slope analysis (main text says it is Table S4, but it appears to be Table S5) that there are enough data points and a long enough time window (about 5-20 March in 2019) to get the slopes to be statistically significant despite this error on Br-.
Mass balance considerations:
Considering the Br- LOD (0.2 micromole / L) in comparison to the atmospheric column of BrO is instructive. Say the top 1.5cm of snow had density 0.3 g cm^-3 (from Figure S2), then the LOD of Br- (0.2 micromole / L) would be equivalent to: 0.2e-6 mole / 1000 cm^-3 water * 0.3 cm^3 water / 1 cm^3 snow * 1.5cm snow * 6.022e23 molecule / mole = 5e13 molecule cm^-2, which is comparable to the larger BrO partial columns. This calculation just points out that errors on the Br- analysis greatly complicate the interpretation of these data with respect to bromine activation.
Tray comparison:
When considering the surface snow, if there was one-way deposition of a species, it would build up in the snowpack and have an increase as is detected for Br- in the top snow, let's say at the 0.024 micromolar / day rate, so 0.024 micromolar is gained every day. It is not clear in the text, but one might assume that the tray samples are swept clean at each (daily) sampling, so there would not be an integration of the Br- over a longer period, but the tray would only have 0.024 micromolar of Br- in it each day. That would seem to say that there should not be a slope of the tray samples, but only a small fixed amount each day.
On the other hand, if there was an increasing flux of atmospheric Br- over this period, which Figure S1 shows to be a period of greatly increasing UV intensity, it might lead to an increasing amount of Br- in the tray deposit samples. One would then expect that the snowpack would be gaining Br- with an accelerating rate (because more Br- is coming down according to the daily tray samples), but that is not observed, possibly because the snowpack is producing reactive halogens that reduce the concentration of Br- in the snow, and we should not consider the deposition to be "one way".
Morning / afternoon differences:
Lines 463-471 are not very clear to me. They say that "signals are not significant across all sampling sites". From the text, it appears that all of the mentioned differences between morning and afternoon are well below the mutual error of the morning and afternoon samples. For example, on line 464, it says that morning is 0.25+/-0.12 micromolar, and afternoon is 0.23+/-0.21 micromolar. From these error bars, I would say that these numbers are the same. If they want to state that they are different, they would need to give numbers of data points and do careful statistics. Similarly, looking at Figure 8, visually examining the points, it seems 3-4 points on this plot, and the error bars seem to overlap a lot, but I cannot tell which error bar goes with which. I think that the morning/afternoon difference needs a clearer plot and an error analysis to be convincing. This section then concludes by reporting: "Based on the above numbers, a mean daytime bromide loss rate of 0.027 micromolar at sea level was obtained." I don't see how they put these conflicting numbers, all seemingly below error bars to finally result in a number that is about a factor of 7 below their Br- detection limit. Overall, this photochemical difference would need a better explanation to be believable.
Tray samples mass balance problems:
On lines 472-479, it is discussed that trays appear to gain Br- and NO3- over the day. If you wanted to try to compare the trays to snowpack, one would need to consider the amount of snow in each reservoir to calculate the mass of Br- in the tray and then compare to the mass of Br- in say the top of the snow pack. If there is not a lot of water mass in the tray, then a small addition of Br- (mass) could increase its concentration much more than it would affect the larger reservoir of snow pack. If they want to try to make a mass balance consideration of snowpack bromide emission and uptake of atmospheric particles, they need to consider the sizes of the reservoirs.
Deposition fluxes:
Section 3.5 attempts to calculate the net deposition flux of bromide from the increase in bromide in the top layers of snow. They get a deposition flux at sea level, but don't have an error bar on this number. From the standard deviation of the slopes given in Table S5, this should be possible to be calculated. The slope errors appear to be on the order of 0.009 micromolar Br- / day. If I compare that to the surface snow slope of about 0.02 micromolar Br- / day, that would be a fractional error of 0.009 / 0.02 = 0.45 or 45% relative error. Therefore, my ball park calculation would indicate that the deposition flux is (1.01 +/- 0.45) x 10^7 molecule cm^-2 s^-1. It appears unlikely that PEARL's error would be very different, so I'm not at all convinced that the quoted "At PEARL, the integrated flux is 7.9 x 10^-6 molecule cm^-2 s^-1, which is ~20% lower than at sea level." is actually true outside of mutual errors. They then go on to say that this proves that snowpack at sea level is not a large source of reactive bromine.
In Section 3.5, discussing (net) deposition flux of Br-, they make the statement "Therefore, if local snowpack on sea ice in the fiord is a large source of reactive bromine, an enhanced deposition flux at sea level should be detected." Other studies have shown that snowpack produces reactive bromine, which of course depletes snowpack of Br-. Therefore, the snowpack at sea level would be expected to be losing Br- by snowpack photochemistry, which they even claim to observe. Some or all of this later re-deposits, and if the net cycle of snowpack production followed by deposition are in balance, then the net trend of Br- in the snowpack would be very small. They show a very small net deposition flux of Br- in the snowpack, which can be perfectly consistent with snowpack production of reactive Br that then does atmospheric chemistry and eventually is converted back to Br- and deposits back to the snowpack. If snowpack 50km offshore produced reactive bromine through snowpack photochemistry, some of that could transport to their study region in a few hours (at 5m/s wind, 50km is traversed in under 3 hours), and then deposit explaining their small net deposition flux.
Nitrate -- bromide relationship:
I don't understand what the sentence from line 521-523 means, and they say that the data is not shown. If they want to make some claim, they should show data for it. Similarly, the discussion made later in this section states "the ratio of [NO3-]/[Br-] ranges form 3.5-6.8, indicating that one molecule of bromide deposited to the surface is likely accompanied by 4-7 nitrate molecules." They don't take into account that only the net deposition is being measured in their studies. Given that they don't get at underlying emission and deposition, I don't understand how to make sense of this ratio in terms of gas-phase chemistry (R1 and R2).
Two-way fluxes:
Literature has long supported a snowpack source of NOx from nitrate photochemistry. This will cause a flux out of the snowpack. NOx can also convert back to nitrate, which has a fast deposition velocity and will deposit back to snowpack. They don't measure the flux of nitrate being lost from the snowpack photochemically, but only the "net" flux that is the deposition minus the loss. Similarly, for Br-, they only measure the net deposition flux, not either production or loss individually. It is not at all clear that this work has truly quantified the daytime loss of Br- from snowpack, and even if they did measure the net loss during daytime, there could still be faster emission plus some deposition during the day that could make the snowpack production rate faster than their daytime snowpack Br- loss. I think that the discussion in lines 550-578 may be trying to do a calculation to split their net deposition into component true emission and true deposition fluxes, but I cannot follow what they are saying here. In addition to not being able to follow it, the whole discussion seems to be built upon the "daytime loss" of 0.027 micromolar, which had no error analysis and doesn't appear significant from Figure 8. Overall, I think that the discussion in this section is not clear enough that I can even diagnose if their reactive bromine emission flux is realistic or if the range listed is based upon realistic error estimates.
Overall:
I think that this manuscript would need major revisions with improved error analysis and clearer discussion of how the observed net deposition flux is split into emission and deposition fluxes to be acceptable.
Citation: https://doi.org/10.5194/egusphere-2023-1446-RC2 -
AC2: 'Reply on RC2', Xin Yang, 01 Jan 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1446/egusphere-2023-1446-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Xin Yang, 01 Jan 2024
Status: closed
-
RC1: 'Comment on egusphere-2023-1446', Anonymous Referee #1, 14 Sep 2023
Yang et al report observations of snow concentrations of ions along with surface atmosphere ozone and BrO concentrations at an Arctic location in the early spring of 2018 and 2019. To my knowledge, these are the first depth-dependent observations of snow bromide, which may be useful for understanding processes determining reactive bromine emissions from snow. They use time and depth-dependent measurements of bromide and nitrate to calculate the net deposition flux during the observational time period. They find that nitrate and bromide in snow are correlated and suggest that they are linked to one another through the formation and hydrolysis of bromine nitrate. They also find that deposition is confined to the surface skin layer.
This paper is very difficult to read, especially the long results section. There are a lot of details and numbers and it is presented in a way that makes it very difficult to discern the big picture. It reads like a first draft. The paragraph starting on line 424 is a particularly good example of this. There are more numbers than words in this paragraph and it is not readable. In general, the paper needs some reorganization and needs to be presented in a more succinct and readable manner. It often reads as a list of disconnected observations.
A large portion of the results section focuses on salinity, but in the end, it is not clear what they learned from it as the results section is difficult to read and there is no follow-up on the salinity observations in the discussion or conclusions section. It is also unclear how an iceberg will impact snow salinity on sea ice and land.
Abstract line 26: missing a unit after 0.024.
Methods: State the eluants used for the IC measurements.
Line 342: I think you mean to say that the “concentrations” are larger, not the “profiles”.
Line 419: What is a near zero increasing trend? Does this mean that the increasing trend is not statistically different from zero?
Section 3.5: I think these calculations represent a net deposition flux (deposition minus emissions) and this should be explicitly stated.
Citation: https://doi.org/10.5194/egusphere-2023-1446-RC1 -
AC1: 'Reply on RC1', Xin Yang, 01 Jan 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1446/egusphere-2023-1446-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Xin Yang, 01 Jan 2024
-
RC2: 'Comment on egusphere-2023-1446', Anonymous Referee #2, 16 Oct 2023
I find this manuscript confusing and the conclusions over-reaching the actual findings. The authors appear to see a net deposition flux of Br- to snowpack in one year (2019) and a much smaller one in the prior year (2018). The trend of net bromide deposition in 2019 appears to be out of error estimates. Beyond these solid findings, I think that the manuscript does not make clear arguments for how the net deposition flux of Br- is relevant to snowpack emission. Multiple other studies have shown that illuminated snowpack emits reactive halogens, and we know that reactive halogen eventually convert back to halides and deposit to snowpack. If these two processes (emission and deposition) are in balance, then the net deposition to snowpack can be near zero. Their finding of only small net deposition is not in conflict with other studies showing snowpack is a source of reactive bromine, and their finding of small net deposition does not preclude snowpack playing a significant role in reactive halogen production.
Error analysis:
Aspects reported in this manuscript seem to be smaller than detection limits. On line 203, it is stated that the limit of detection (LOD) for Br- is 0.2 micromolar. It appears from the slope analysis (main text says it is Table S4, but it appears to be Table S5) that there are enough data points and a long enough time window (about 5-20 March in 2019) to get the slopes to be statistically significant despite this error on Br-.
Mass balance considerations:
Considering the Br- LOD (0.2 micromole / L) in comparison to the atmospheric column of BrO is instructive. Say the top 1.5cm of snow had density 0.3 g cm^-3 (from Figure S2), then the LOD of Br- (0.2 micromole / L) would be equivalent to: 0.2e-6 mole / 1000 cm^-3 water * 0.3 cm^3 water / 1 cm^3 snow * 1.5cm snow * 6.022e23 molecule / mole = 5e13 molecule cm^-2, which is comparable to the larger BrO partial columns. This calculation just points out that errors on the Br- analysis greatly complicate the interpretation of these data with respect to bromine activation.
Tray comparison:
When considering the surface snow, if there was one-way deposition of a species, it would build up in the snowpack and have an increase as is detected for Br- in the top snow, let's say at the 0.024 micromolar / day rate, so 0.024 micromolar is gained every day. It is not clear in the text, but one might assume that the tray samples are swept clean at each (daily) sampling, so there would not be an integration of the Br- over a longer period, but the tray would only have 0.024 micromolar of Br- in it each day. That would seem to say that there should not be a slope of the tray samples, but only a small fixed amount each day.
On the other hand, if there was an increasing flux of atmospheric Br- over this period, which Figure S1 shows to be a period of greatly increasing UV intensity, it might lead to an increasing amount of Br- in the tray deposit samples. One would then expect that the snowpack would be gaining Br- with an accelerating rate (because more Br- is coming down according to the daily tray samples), but that is not observed, possibly because the snowpack is producing reactive halogens that reduce the concentration of Br- in the snow, and we should not consider the deposition to be "one way".
Morning / afternoon differences:
Lines 463-471 are not very clear to me. They say that "signals are not significant across all sampling sites". From the text, it appears that all of the mentioned differences between morning and afternoon are well below the mutual error of the morning and afternoon samples. For example, on line 464, it says that morning is 0.25+/-0.12 micromolar, and afternoon is 0.23+/-0.21 micromolar. From these error bars, I would say that these numbers are the same. If they want to state that they are different, they would need to give numbers of data points and do careful statistics. Similarly, looking at Figure 8, visually examining the points, it seems 3-4 points on this plot, and the error bars seem to overlap a lot, but I cannot tell which error bar goes with which. I think that the morning/afternoon difference needs a clearer plot and an error analysis to be convincing. This section then concludes by reporting: "Based on the above numbers, a mean daytime bromide loss rate of 0.027 micromolar at sea level was obtained." I don't see how they put these conflicting numbers, all seemingly below error bars to finally result in a number that is about a factor of 7 below their Br- detection limit. Overall, this photochemical difference would need a better explanation to be believable.
Tray samples mass balance problems:
On lines 472-479, it is discussed that trays appear to gain Br- and NO3- over the day. If you wanted to try to compare the trays to snowpack, one would need to consider the amount of snow in each reservoir to calculate the mass of Br- in the tray and then compare to the mass of Br- in say the top of the snow pack. If there is not a lot of water mass in the tray, then a small addition of Br- (mass) could increase its concentration much more than it would affect the larger reservoir of snow pack. If they want to try to make a mass balance consideration of snowpack bromide emission and uptake of atmospheric particles, they need to consider the sizes of the reservoirs.
Deposition fluxes:
Section 3.5 attempts to calculate the net deposition flux of bromide from the increase in bromide in the top layers of snow. They get a deposition flux at sea level, but don't have an error bar on this number. From the standard deviation of the slopes given in Table S5, this should be possible to be calculated. The slope errors appear to be on the order of 0.009 micromolar Br- / day. If I compare that to the surface snow slope of about 0.02 micromolar Br- / day, that would be a fractional error of 0.009 / 0.02 = 0.45 or 45% relative error. Therefore, my ball park calculation would indicate that the deposition flux is (1.01 +/- 0.45) x 10^7 molecule cm^-2 s^-1. It appears unlikely that PEARL's error would be very different, so I'm not at all convinced that the quoted "At PEARL, the integrated flux is 7.9 x 10^-6 molecule cm^-2 s^-1, which is ~20% lower than at sea level." is actually true outside of mutual errors. They then go on to say that this proves that snowpack at sea level is not a large source of reactive bromine.
In Section 3.5, discussing (net) deposition flux of Br-, they make the statement "Therefore, if local snowpack on sea ice in the fiord is a large source of reactive bromine, an enhanced deposition flux at sea level should be detected." Other studies have shown that snowpack produces reactive bromine, which of course depletes snowpack of Br-. Therefore, the snowpack at sea level would be expected to be losing Br- by snowpack photochemistry, which they even claim to observe. Some or all of this later re-deposits, and if the net cycle of snowpack production followed by deposition are in balance, then the net trend of Br- in the snowpack would be very small. They show a very small net deposition flux of Br- in the snowpack, which can be perfectly consistent with snowpack production of reactive Br that then does atmospheric chemistry and eventually is converted back to Br- and deposits back to the snowpack. If snowpack 50km offshore produced reactive bromine through snowpack photochemistry, some of that could transport to their study region in a few hours (at 5m/s wind, 50km is traversed in under 3 hours), and then deposit explaining their small net deposition flux.
Nitrate -- bromide relationship:
I don't understand what the sentence from line 521-523 means, and they say that the data is not shown. If they want to make some claim, they should show data for it. Similarly, the discussion made later in this section states "the ratio of [NO3-]/[Br-] ranges form 3.5-6.8, indicating that one molecule of bromide deposited to the surface is likely accompanied by 4-7 nitrate molecules." They don't take into account that only the net deposition is being measured in their studies. Given that they don't get at underlying emission and deposition, I don't understand how to make sense of this ratio in terms of gas-phase chemistry (R1 and R2).
Two-way fluxes:
Literature has long supported a snowpack source of NOx from nitrate photochemistry. This will cause a flux out of the snowpack. NOx can also convert back to nitrate, which has a fast deposition velocity and will deposit back to snowpack. They don't measure the flux of nitrate being lost from the snowpack photochemically, but only the "net" flux that is the deposition minus the loss. Similarly, for Br-, they only measure the net deposition flux, not either production or loss individually. It is not at all clear that this work has truly quantified the daytime loss of Br- from snowpack, and even if they did measure the net loss during daytime, there could still be faster emission plus some deposition during the day that could make the snowpack production rate faster than their daytime snowpack Br- loss. I think that the discussion in lines 550-578 may be trying to do a calculation to split their net deposition into component true emission and true deposition fluxes, but I cannot follow what they are saying here. In addition to not being able to follow it, the whole discussion seems to be built upon the "daytime loss" of 0.027 micromolar, which had no error analysis and doesn't appear significant from Figure 8. Overall, I think that the discussion in this section is not clear enough that I can even diagnose if their reactive bromine emission flux is realistic or if the range listed is based upon realistic error estimates.
Overall:
I think that this manuscript would need major revisions with improved error analysis and clearer discussion of how the observed net deposition flux is split into emission and deposition fluxes to be acceptable.
Citation: https://doi.org/10.5194/egusphere-2023-1446-RC2 -
AC2: 'Reply on RC2', Xin Yang, 01 Jan 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1446/egusphere-2023-1446-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Xin Yang, 01 Jan 2024
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 339 | 125 | 24 | 488 | 45 | 16 | 17 |
- HTML: 339
- PDF: 125
- XML: 24
- Total: 488
- Supplement: 45
- BibTeX: 16
- EndNote: 17
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1