the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 Apr 2024
Identifying airborne snow metamorphism with stable water isotopes
Abstract. Wind-blown snow is a frequent phenomenon in high-elevation and polar regions which impacts the surface energy and mass balance of these areas. Loose surface snow gets eroded and transported by wind which influences the snow particle’s physical properties (size, shape, optical properties) that determine the characteristics of the emerging wind-impacted snowpack layer. During airborne snow transport, the governing processes are happening on the micro-scale, while the particles are transported over long distances. The unfolding processes and the evolution of the particle’s physical properties are thus difficult to observe in-situ. Here we used cold-laboratory ring wind tunnel experiments as an interim solution to study the governing processes during airborne snow transport with stable water isotopes as tracers for these micro-scale processes. Repeated analysis of airborne-sampled snow by micro-computed tomography (μCT) documented a growing and rounding of snow particles with transport time with a concurrent decrease in specific surface area. Stable water isotope analysis of airborne snow and water vapour allowed us to attribute this evolution to the process of airborne snow metamorphism. The changes observed in the snow isotopic composition showed a clear isotopic signature of metamorphic deposition, which requires particle-air temperature gradients. These results question the validity of the thermal equilibrium assumption between particles and air inside the saltation layer of wind-blown snow events where the conditions are similar to the ones found in the wind tunnel. Our results thus refine the understanding of the governing processes in the saltation layer and suggest that the snow’s isotopic composition can inform on local wind-blown snow events as the original snow isotope signal gets overprinted by airborne snow metamorphism. Thus, airborne snow metamorphism has the potential to influence the climate signal stored in snow and ice core stable water isotope records.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(10989 KB)
-
Supplement
(1183 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(10989 KB) - Metadata XML
-
Supplement
(1183 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-745', Anonymous Referee #1, 24 May 2024
This novel study employs ring wind tunnel experiments combined with stable water isotope analysis to investigate airborne snow particle metamorphism. A main finding is that vapour deposition drives snow particle growth and rounding supported by the observed isotopic fractionation and concurrent SSA decrease. It is inferred that particles and air inside the saltation layer are not in thermal equilibrium as is commonly assumed in blowing snow models. Any mechanical particle fragmentation or coalescence likely play a smaller role in the observed particle size changes as they would not induce isotope fractionation. In turn, the water stable isotopic fractionation induced by airborne snow metamorphism needs to be taken into account when extracting climate information from ice cores, especially at dry and windy locations.
General Comments
This is a carefully designed laboratory experiment, with sound methods and data analysis, and with some interesting conclusions, and should published after addressing minor comments listed below.
While this may be common knowledge a very brief description of isothermal versus temperature gradient snow metamorphism as relevant to this study is warranted in the introduction. In particular to clarify the statement that a particle-air temperature gradient must exist to explain depositional particle growth and isotopic fractionation. The alternative would be vapour fluxes (sublimation/deposition) across an individual particle but also between particles driven by the curvature (Kelvin) effect resulting in local water vapour pressure gradients and super(or sub)-saturation.
These fluxes occur at thermal equilibrium and may also induce isotopic fractionation between solid and the remaining vapour phase. I may be convinced that the bulk isotopic composition of snow remains constant but some further discussion is warranted. To do this I'd suggest to better illustrate the temporal co-evolution of the stable H2O isotopes in both snow and also water vapour. E.g. add a similar figure as Fig.3 showing d18O, d2H and d-exx in the vapour phase. Some of the behaviour seen in experiment No.9 (Fig.4) is puzzling, e.g. O18 in vapour and snow shows correlation, whereas 2H shows anti-correlation (significant?). Was this behaviour observed also in other experiments and is this related to the mentioned non-equilibrium conditions?
Detailed Comments
L145 - Mention here what temp was the wind tunnel set to?
Table 1: Clarify in the caption that DELTA T means change in mean wind tunnel T over the duration of the experiment
L182 - cm3
L337 - Be specific: significant enrichment by how many permil?
L395 - In order to illustrate the concurrent vapour isotopic composition change across all experiments I suggest a similar figure as Fig3. (see above)
L431 - Shouldn't mechanic fragmentation lead to a SSA increase if it was the dominating process?
L441 - Please explain "higher SSA decay rates for isothermal snowpack metamorphism", how much higher? Higher than T-gradient metamorphism? reference?
L448 - Are particles in the saltation layer subject to a different metamorphism regime than those in the suspension layer? Please expand & add any relevant reference
L465 - What about the vapour flux between particles, i.e. sublimation of small snow particles, which may eventually disappear, followed by deposition to larger particles. See comment above.
L528 - Except that particle-to-particle vapour flux can occur also at T-gradient = 0 and RH_ice = 1 due to Kelvin effect (equivalent to isothermal metamorphism I think)
Fig.6 - yes, this is related to the curvature (Kelvin) effect
Conclusions - list here and possibly in the abstract the order of magnitude of the observed isotope fractionation attributed to airborne snow metamorphism in permil, a result relevant for the interpretation of field data.Citation: https://doi.org/10.5194/egusphere-2024-745-RC1 -
AC1: 'Reply on RC1', Sonja Wahl, 28 Jun 2024
We thank the reviewer for the detailed feedback and are delighted about the short, yet accurate and precise summary of our manuscript which agrees very well with our own perception of the main take-home messages of this study. We address the reviewer’s comments individually and in detail in the pdf (answes in green). In addition to edits based on the reviewers’ comments, we updated a few inconsistencies in the text and figures, such as the color code in Fig. 3 to be consistent throughout the manuscript. In summary, the major changes made are related to:
1) New Fig. 5 to describe the co-evolution of d18O and dD during and after snow introduction in more detail
2) the statistics of observed isotope changes in vapour and snow. We included a table (Table 2) to group the information and declutter the corresponding Sec 3.2
3) A short paragraph in the introduction to define temperature-gradient and isothermal snow metamorphism
-
AC1: 'Reply on RC1', Sonja Wahl, 28 Jun 2024
-
RC2: 'Comment on egusphere-2024-745', Anonymous Referee #2, 28 May 2024
Dear authors and editors,
thank you for the opportunity to review this interesting study. The work presents a laboratory experiment exploring snow processes during aeolian transport using a ring-shaped wind tunnel. The experiment leverages two technologies, microCT scanning to determine the change in physical properties of the snow grain during transport, and stable water isotope analysis of both the transported snow and the air water vapor to study the physical processes in phase change, i.e. sublimation and deposition during transport. They find that grain snow SSA decreases and grain size increases, and snow particles experience longer time of wind transport. The isotope data suggest both net sublimation in the early stage of the experiment, and vapor deposition on the transported snow particles.
The work is an innovative way to study the changes in snow during wind transport, which is difficult to do in-situ in field conditions. The work used a unique combination of microCT scanning and stable water isotope measurements, allowing novel insights to micro scale processes. I think the experiment is carefully planned and executed, and the findings have implications for both snow physics research, and research using stable water isotopes as tracers of paleoclimate or hydrology in snow-influenced regions. I recommend the work to be published, after addressing my comments below.
L32: Cite some large and small scale wind drift modeling studies.
L55: start new paragraph
L85: also melt and freeze
L89: suggest to cite papers that use water isotopes as tracers in snow studies.
L104: in my understanding the d18O_ice < d18O_vapor in any deposition process, not only supersaturated, fractionation factor in Eq (2) take values >1.
L135: changes in what?
L140: tab -> tap?
L140-141: Don’t get this – how it was not in operation, but was producing new snow minimum four days before the experiment (L144)?
L143: did you analyze the influence of storage time? And how to determine mixing success? If not, suggest to remove the start of this sentence.
L170: out of interest: what was the mass balance of your 600g of added snow? x g sampled, y g deposited, the residual z g sublimated? Maybe not very important for your results and findings, but this would give an idea where did the snow in the air column end up, and allow better imagining the experimental setup?
Table 1: please specify the variables, not only units for the DeltaH2O column. Also, what does DeltaT mean?
section 2.2.1: a carefully though of experimental equipment and setup: trying hard to think of points of criticism but cannot find any😊
L228: what is the difference specifically between drifting and blowing snow in your experiment?
L371-375: cannot locate this data in your plots
L395-406: this difficult to follow. First you stare that most experiments follow example in Fig. 4. On L402 you talk about subsequent evolution, subsequent to snow addition? to which of the categories does the example in Fig.4 belong to, for example? Not sure what is the best way to summarize this data, perhaps in a table, but the current way is difficult for me to digest.
L404: does the reversed evolution pertain only to 18O?
Chapter 3.2: I was expecting also the snow samples and their temporal evolution in Fig.4 to be described in the this paragraph.
L459: start a new paragraph to give rhythm to the section?
L476: What’s your view: would the full sublimation of small particles conceptually lead to isotope fractionation in the suspended snow?
L564: replace “in other words” by this is demonstrated by … or similar. Because you haven’t really show the evidence for the statement.
L582?: can you propose a way to conceptualize the processes you have found into a modeling context, where isotope values in the snowpack are important, such as isotope-enabled climate of hydrological models?
L613: start new paragraph
L659: can you find any field studies that would have observed similar (or any) change in the isotope values of wind-transported snow? Or can you propose an experiment that could study this in the field conditions?
L684: You do not have data for this? the conclusions are indirect from processes.
Citation: https://doi.org/10.5194/egusphere-2024-745-RC2 -
AC2: 'Reply on RC2', Sonja Wahl, 28 Jun 2024
We thank the reviewer for the time that was spent on this review which helped to improve this manuscript. We detailed our changes in reply to the comments in the pdf (answers in green). In addition to edits based on the reviewers’ comments, we updated a few inconsistencies in the text and figures, such as the color code in Fig. 3 to be consistent throughout the manuscript. In summary the major changes made are related to:
1) New Fig. 5 to describe the co-evolution of d18O and dD during and after snow introduction in more detail and an adjustment of the vapour isotope change results section 3.2.2
2) the statistics of observed isotope changes in vapour and snow. We included a table (Table 2) to group the information and declutter section 3.2
3) A short paragraph in the introduction to define temperature-gradient and isothermal snow metamorphism
-
AC2: 'Reply on RC2', Sonja Wahl, 28 Jun 2024
-
RC3: 'Comment on egusphere-2024-745', Anonymous Referee #3, 03 Jun 2024
The review of manuscript “Identifying airborne snow metamorphism with stable water isotopes” by Dr. Sonja Wahl and colleagues.
The manuscript presents and discusses the results of laboratory experiments that simulate blowing snow events. The authors show that the blowing snow particles are modified as a result of “airborne snow metamorphism”. The isotopic composition of the snow particles and of the surrounding water vapor is changing as well (due to sublimation and re-sublimation fluxes), although the sign and value of the isotopic transformations differ from one experiment to another.
This study shows that the snow drift before the newly precipitated snow is finally deposited onto the snow surface, is an important part of “post-depositional” processes that alter the initial isotopic content of the precipitation. Thus, this work is an important step towards a deeper understanding the whole complexity of the post-depositional snow evolution, which is crucial for the interpretation of the deep ice core isotopic signal.
I have only minor correction to the manuscript:
Lines 658-659 (“Thus, it could be possible to use the snow isotopic composition to differentiate between wind-blown snow and precipitated snow”) – firstly, I am not sure why one could need to make such differentiation. Secondly, freshly precipitated snow stays “fresh” not for long time, it is involved to the post-depositional processes immediately after deposition, so its isotopic signature would be modified quickly. Thirdly, in precipitation there is a huge variability of d18O and dxs, as seen from the observation (see data from Concordia station, as an example).
Line 215 – ml min-1 (put a space between ml and min). The same in line 223.
Figure 3 – does the grey background in the upper row have any particular meaning? If not, it’s better to delete it.
Line 481 – do you need the word “explained” here? Suggest to delete it.
Citation: https://doi.org/10.5194/egusphere-2024-745-RC3 -
AC3: 'Reply on RC3', Sonja Wahl, 28 Jun 2024
We thank the reviewer for this very positive review and have noted our replies to the comments in the pdf (answers in green). In addition to edits based on the reviewers’ comments, we updated a few inconsistencies in the text and figures, such as the color code in Fig. 3 to be consistent throughout the manuscript. In summary the changes made are related to:
1) New Fig. 5 to describe the co-evolution of d18O and dD during and after snow introduction in more detail and an adjustment of the vapour isotope change results section 3.2.2
2) the statistics of observed isotope changes in vapour and snow. We included a table (Table 2) to group the information and declutter the corresponding text.
3) A short paragraph in the introduction to define temperature-gradient and isothermal snow metamorphism
-
AC3: 'Reply on RC3', Sonja Wahl, 28 Jun 2024
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-745', Anonymous Referee #1, 24 May 2024
This novel study employs ring wind tunnel experiments combined with stable water isotope analysis to investigate airborne snow particle metamorphism. A main finding is that vapour deposition drives snow particle growth and rounding supported by the observed isotopic fractionation and concurrent SSA decrease. It is inferred that particles and air inside the saltation layer are not in thermal equilibrium as is commonly assumed in blowing snow models. Any mechanical particle fragmentation or coalescence likely play a smaller role in the observed particle size changes as they would not induce isotope fractionation. In turn, the water stable isotopic fractionation induced by airborne snow metamorphism needs to be taken into account when extracting climate information from ice cores, especially at dry and windy locations.
General Comments
This is a carefully designed laboratory experiment, with sound methods and data analysis, and with some interesting conclusions, and should published after addressing minor comments listed below.
While this may be common knowledge a very brief description of isothermal versus temperature gradient snow metamorphism as relevant to this study is warranted in the introduction. In particular to clarify the statement that a particle-air temperature gradient must exist to explain depositional particle growth and isotopic fractionation. The alternative would be vapour fluxes (sublimation/deposition) across an individual particle but also between particles driven by the curvature (Kelvin) effect resulting in local water vapour pressure gradients and super(or sub)-saturation.
These fluxes occur at thermal equilibrium and may also induce isotopic fractionation between solid and the remaining vapour phase. I may be convinced that the bulk isotopic composition of snow remains constant but some further discussion is warranted. To do this I'd suggest to better illustrate the temporal co-evolution of the stable H2O isotopes in both snow and also water vapour. E.g. add a similar figure as Fig.3 showing d18O, d2H and d-exx in the vapour phase. Some of the behaviour seen in experiment No.9 (Fig.4) is puzzling, e.g. O18 in vapour and snow shows correlation, whereas 2H shows anti-correlation (significant?). Was this behaviour observed also in other experiments and is this related to the mentioned non-equilibrium conditions?
Detailed Comments
L145 - Mention here what temp was the wind tunnel set to?
Table 1: Clarify in the caption that DELTA T means change in mean wind tunnel T over the duration of the experiment
L182 - cm3
L337 - Be specific: significant enrichment by how many permil?
L395 - In order to illustrate the concurrent vapour isotopic composition change across all experiments I suggest a similar figure as Fig3. (see above)
L431 - Shouldn't mechanic fragmentation lead to a SSA increase if it was the dominating process?
L441 - Please explain "higher SSA decay rates for isothermal snowpack metamorphism", how much higher? Higher than T-gradient metamorphism? reference?
L448 - Are particles in the saltation layer subject to a different metamorphism regime than those in the suspension layer? Please expand & add any relevant reference
L465 - What about the vapour flux between particles, i.e. sublimation of small snow particles, which may eventually disappear, followed by deposition to larger particles. See comment above.
L528 - Except that particle-to-particle vapour flux can occur also at T-gradient = 0 and RH_ice = 1 due to Kelvin effect (equivalent to isothermal metamorphism I think)
Fig.6 - yes, this is related to the curvature (Kelvin) effect
Conclusions - list here and possibly in the abstract the order of magnitude of the observed isotope fractionation attributed to airborne snow metamorphism in permil, a result relevant for the interpretation of field data.Citation: https://doi.org/10.5194/egusphere-2024-745-RC1 -
AC1: 'Reply on RC1', Sonja Wahl, 28 Jun 2024
We thank the reviewer for the detailed feedback and are delighted about the short, yet accurate and precise summary of our manuscript which agrees very well with our own perception of the main take-home messages of this study. We address the reviewer’s comments individually and in detail in the pdf (answes in green). In addition to edits based on the reviewers’ comments, we updated a few inconsistencies in the text and figures, such as the color code in Fig. 3 to be consistent throughout the manuscript. In summary, the major changes made are related to:
1) New Fig. 5 to describe the co-evolution of d18O and dD during and after snow introduction in more detail
2) the statistics of observed isotope changes in vapour and snow. We included a table (Table 2) to group the information and declutter the corresponding Sec 3.2
3) A short paragraph in the introduction to define temperature-gradient and isothermal snow metamorphism
-
AC1: 'Reply on RC1', Sonja Wahl, 28 Jun 2024
-
RC2: 'Comment on egusphere-2024-745', Anonymous Referee #2, 28 May 2024
Dear authors and editors,
thank you for the opportunity to review this interesting study. The work presents a laboratory experiment exploring snow processes during aeolian transport using a ring-shaped wind tunnel. The experiment leverages two technologies, microCT scanning to determine the change in physical properties of the snow grain during transport, and stable water isotope analysis of both the transported snow and the air water vapor to study the physical processes in phase change, i.e. sublimation and deposition during transport. They find that grain snow SSA decreases and grain size increases, and snow particles experience longer time of wind transport. The isotope data suggest both net sublimation in the early stage of the experiment, and vapor deposition on the transported snow particles.
The work is an innovative way to study the changes in snow during wind transport, which is difficult to do in-situ in field conditions. The work used a unique combination of microCT scanning and stable water isotope measurements, allowing novel insights to micro scale processes. I think the experiment is carefully planned and executed, and the findings have implications for both snow physics research, and research using stable water isotopes as tracers of paleoclimate or hydrology in snow-influenced regions. I recommend the work to be published, after addressing my comments below.
L32: Cite some large and small scale wind drift modeling studies.
L55: start new paragraph
L85: also melt and freeze
L89: suggest to cite papers that use water isotopes as tracers in snow studies.
L104: in my understanding the d18O_ice < d18O_vapor in any deposition process, not only supersaturated, fractionation factor in Eq (2) take values >1.
L135: changes in what?
L140: tab -> tap?
L140-141: Don’t get this – how it was not in operation, but was producing new snow minimum four days before the experiment (L144)?
L143: did you analyze the influence of storage time? And how to determine mixing success? If not, suggest to remove the start of this sentence.
L170: out of interest: what was the mass balance of your 600g of added snow? x g sampled, y g deposited, the residual z g sublimated? Maybe not very important for your results and findings, but this would give an idea where did the snow in the air column end up, and allow better imagining the experimental setup?
Table 1: please specify the variables, not only units for the DeltaH2O column. Also, what does DeltaT mean?
section 2.2.1: a carefully though of experimental equipment and setup: trying hard to think of points of criticism but cannot find any😊
L228: what is the difference specifically between drifting and blowing snow in your experiment?
L371-375: cannot locate this data in your plots
L395-406: this difficult to follow. First you stare that most experiments follow example in Fig. 4. On L402 you talk about subsequent evolution, subsequent to snow addition? to which of the categories does the example in Fig.4 belong to, for example? Not sure what is the best way to summarize this data, perhaps in a table, but the current way is difficult for me to digest.
L404: does the reversed evolution pertain only to 18O?
Chapter 3.2: I was expecting also the snow samples and their temporal evolution in Fig.4 to be described in the this paragraph.
L459: start a new paragraph to give rhythm to the section?
L476: What’s your view: would the full sublimation of small particles conceptually lead to isotope fractionation in the suspended snow?
L564: replace “in other words” by this is demonstrated by … or similar. Because you haven’t really show the evidence for the statement.
L582?: can you propose a way to conceptualize the processes you have found into a modeling context, where isotope values in the snowpack are important, such as isotope-enabled climate of hydrological models?
L613: start new paragraph
L659: can you find any field studies that would have observed similar (or any) change in the isotope values of wind-transported snow? Or can you propose an experiment that could study this in the field conditions?
L684: You do not have data for this? the conclusions are indirect from processes.
Citation: https://doi.org/10.5194/egusphere-2024-745-RC2 -
AC2: 'Reply on RC2', Sonja Wahl, 28 Jun 2024
We thank the reviewer for the time that was spent on this review which helped to improve this manuscript. We detailed our changes in reply to the comments in the pdf (answers in green). In addition to edits based on the reviewers’ comments, we updated a few inconsistencies in the text and figures, such as the color code in Fig. 3 to be consistent throughout the manuscript. In summary the major changes made are related to:
1) New Fig. 5 to describe the co-evolution of d18O and dD during and after snow introduction in more detail and an adjustment of the vapour isotope change results section 3.2.2
2) the statistics of observed isotope changes in vapour and snow. We included a table (Table 2) to group the information and declutter section 3.2
3) A short paragraph in the introduction to define temperature-gradient and isothermal snow metamorphism
-
AC2: 'Reply on RC2', Sonja Wahl, 28 Jun 2024
-
RC3: 'Comment on egusphere-2024-745', Anonymous Referee #3, 03 Jun 2024
The review of manuscript “Identifying airborne snow metamorphism with stable water isotopes” by Dr. Sonja Wahl and colleagues.
The manuscript presents and discusses the results of laboratory experiments that simulate blowing snow events. The authors show that the blowing snow particles are modified as a result of “airborne snow metamorphism”. The isotopic composition of the snow particles and of the surrounding water vapor is changing as well (due to sublimation and re-sublimation fluxes), although the sign and value of the isotopic transformations differ from one experiment to another.
This study shows that the snow drift before the newly precipitated snow is finally deposited onto the snow surface, is an important part of “post-depositional” processes that alter the initial isotopic content of the precipitation. Thus, this work is an important step towards a deeper understanding the whole complexity of the post-depositional snow evolution, which is crucial for the interpretation of the deep ice core isotopic signal.
I have only minor correction to the manuscript:
Lines 658-659 (“Thus, it could be possible to use the snow isotopic composition to differentiate between wind-blown snow and precipitated snow”) – firstly, I am not sure why one could need to make such differentiation. Secondly, freshly precipitated snow stays “fresh” not for long time, it is involved to the post-depositional processes immediately after deposition, so its isotopic signature would be modified quickly. Thirdly, in precipitation there is a huge variability of d18O and dxs, as seen from the observation (see data from Concordia station, as an example).
Line 215 – ml min-1 (put a space between ml and min). The same in line 223.
Figure 3 – does the grey background in the upper row have any particular meaning? If not, it’s better to delete it.
Line 481 – do you need the word “explained” here? Suggest to delete it.
Citation: https://doi.org/10.5194/egusphere-2024-745-RC3 -
AC3: 'Reply on RC3', Sonja Wahl, 28 Jun 2024
We thank the reviewer for this very positive review and have noted our replies to the comments in the pdf (answers in green). In addition to edits based on the reviewers’ comments, we updated a few inconsistencies in the text and figures, such as the color code in Fig. 3 to be consistent throughout the manuscript. In summary the changes made are related to:
1) New Fig. 5 to describe the co-evolution of d18O and dD during and after snow introduction in more detail and an adjustment of the vapour isotope change results section 3.2.2
2) the statistics of observed isotope changes in vapour and snow. We included a table (Table 2) to group the information and declutter the corresponding text.
3) A short paragraph in the introduction to define temperature-gradient and isothermal snow metamorphism
-
AC3: 'Reply on RC3', Sonja Wahl, 28 Jun 2024
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 356 | 117 | 38 | 511 | 51 | 26 | 29 |
- HTML: 356
- PDF: 117
- XML: 38
- Total: 511
- Supplement: 51
- BibTeX: 26
- EndNote: 29
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 145 | 31 |
| Switzerland | 2 | 44 | 9 |
| China | 3 | 44 | 9 |
| France | 4 | 37 | 7 |
| Germany | 5 | 28 | 5 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 145
Cited
1 citations as recorded by crossref.
Benjamin Walter
Franziska Aemisegger
Luca Bianchi
Michael Lehning
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(10989 KB) - Metadata XML
-
Supplement
(1183 KB) - BibTeX
- EndNote
- Final revised paper