the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 24 Mar 2021
Physical and chemical constraints on transformation and mass-increase of fine aerosols in northeast Asia
Abstract. Over the past few decades, northeast Asia has suffered from the extreme levels of PM2.5 (particulate matter with an aerodynamic diameter smaller than 2.5 μm). Despite extensive efforts and the scientific advances in understanding PM2.5 pollution, the fundamental mechanisms responsible for the occurrence of high PM2.5 concentrations have not been comprehensively understood. In this study, we investigated the physical and chemical drivers for the formation and transformation of atmospheric particles using a four-year dataset of nanoparticle number size distributions, PM2.5 chemical composition, gaseous precursors, and meteorological variables in northeast Asia outflows. The empirical orthogonal function (EOF) analyses of size-separated particle numbers extracted two modes representing a burst of nanoparticles (EOF1) and an increase in PM2.5 mass (EOF2) associated with persistent anticyclone and synoptic-scale stagnation, respectively. The vertical structure of the particles demonstrated that the synoptic conditions also affected the daily evolution of boundary layer, promoting either the formation of nanoparticles through deep mixing or conversion into accumulation-mode particles in shallow mixed layers. In the haze-development episode equivalent to EOF2 during the KORUS-AQ (KORea-US Air Quality) campaign, the PM2.5 mass reached 63 μg m−3 with the highest contribution from inorganic constituents, which was accompanied by a thick coating of refractory black carbon (rBC) that linearly increased with condensation-mode particles. This observational evidence suggests that the thick coating of rBC resulted from an active conversion of condensable gases into particle-phase on the BC surface, thereby increasing the mass of the accumulation-mode aerosol. Consequently, this result complies with the strategy to reduce black carbon as a way to effectively mitigate haze pollution as well as climate change in northeast Asia.
This preprint has been withdrawn.
-
Withdrawal notice
This preprint has been withdrawn.
-
Preprint
(1809 KB)
-
Supplement
(826 KB)
-
This preprint has been withdrawn.
- Preprint
(1809 KB) - Metadata XML
-
Supplement
(826 KB) - BibTeX
- EndNote
Interactive discussion
Status: closed
-
RC1: 'Review of Lim et al.', Anonymous Referee #2, 29 Apr 2021
<General comments>
This paper discusses the formation mechanism of aerosols in northeast Asia. The authors used in-situ data obtained in Jeju Island from January 2013 to December 2016 and those during KORUS-AQ (May-June 2016) to investigate the variability in aerosol concentrations in the boundary layer. They used the empirical orthogonal function (EOF) analysis to classify the observed features. While the main topic of this paper is suitable for ACP, I do not think that the main conclusions are fully supported by observational evidence. I suggest that the authors largely reorganize the results and discussion and clarify the robust and new findings from this study. I recommend major revisions.<Specific comments>
Introduction
The formation of aerosols associated with meteorological cycles (e.g., anticyclones, cyclones, and air stagnation) has been extensively studied either in continental source areas or downwind regions. Variability in vertical profiles of aerosols associated with the evolution of the boundary layer has also been extensively investigated. Although this study might be the first to present such data in Korea, I think the fundamental mechanisms are common in many cases. The authors should briefly review previous studies in northeast Asia (e.g., TRACE-P, CARE-Beijing) and also in other regions, and discuss the similarity and difference between this study and previous ones. Here are some examples of the previous studies in northeast Asia.
Weber, R. J., et al. (2003): J. Geophys. Res., 108, 8814, doi:10.1029/2002JD003112.
Matsui, H., et al. (2009): J. Geophys. Res., 114, D00G13, doi:10.1029/2008JD010906.
Takegawa, N., et al. (2009): J. Geophys. Res., 114, D00G05, doi:10.1029/2008JD010857.
Haenel, A., et al. (2012): J. Geophys. Res., 117, D13201, doi:10.1029/2012JD017577.
Wang, J., et al., (2019): Atmos. Chem. Phys., 19, 8845-8861, doi:10.5194/acp-19-8845-2019.The following review paper would also be useful for the interpretation of NPF events in relation with meteorological conditions.
Kerminen, V.-M., et al. (2018), Environ. Res. Lett., 13, 103003. doi:10.1088/1748-9326/aadf3c.L144-146: Uncertainty in the coating thickness
The estimation of coating thickness of BC particles from SP2 data, although it has been used by many investigators, may contain significant uncertainties. The authors selected a BC core diameter of 200 +/- 20 nm. Why did the authors select this specific diameter? Is it reasonable to estimate a coating thickness of > 10 nm with the core diameter uncertainty of 20 nm?L249-251: Boundary layer stability
The dominance of a high-pressure system (subsidence) generally leads to the formation of strong inversions and stable boundary layers. The description in this paragraph seems to be opposite.L259-269: Entrainment
The authors conclude that the rapid increase in PM2.5 was due to the entrainment of particles from upstream areas. The authors state that elevation of aerosol concentrations is “believed” to occur by the intrusion of pollutants from the upper atmosphere. What is the basis for this statement? I do not think that the descriptions in this paragraph are supported by observational evidence.L306-318: Gas-to-particle conversion.
The discussion in this paragraph is highly speculative. The partitioning between HNO3 and NH4NO3 should be explicitly investigated to discuss the gas-to-particle conversion for nitrate aerosols. See, for example, Neuman, J. A., et al. (2003): J. Geophys. Res., 108, 4557, doi:10.1029/2003JD003616. Furthermore, the formation of (NH4)2SO4 might be controlled by aqueous-phase reactions in cloud droplets rather than condensation processes. Please reconsider the interpretation.L353-366: Interpretation of the coating thickness
The authors suggest that the coating thickness of rBC is s useful parameter to understand the formation of secondary aerosols, and also suggest that reducing BC emissions is the effective way to reduce PM2.5 in Asia. I think the descriptions in these paragraphs are also very speculative and not supported by observational evidence. Fig. 7 seems to the basis for this hypothesis, but I find many data points at lower aerosol mass loadings with thick coatings. It may be true that the EOF2 case can be characterized by high PM2.5 and thick coatings, but it does not necessarily mean that the coating thickness is the controlling factor. I would guess the correlation between the PM2.5 concentrations and the coating thickness is rather weak. Please show more convincing data to support the hypothesis. Otherwise I recommend that the authors should remove (at least tone down) this conclusion.<Minor comments>
L73-74, L182: SO2, NOx, and VOCs are not "condensable" gases but precursors.L100-121: Please specify the model number of the SMPS, CPC, and OPC. Please also describe how these instruments were evaluated and calibrated. It is not necessary to capitalize the first character of the name of the instruments.
L178-179: I do not think the estimate of GR values has three significant digits.
L331: "peaking below 100 nm" - Please specify number, surface, or mass.
Citation: https://doi.org/10.5194/acp-2020-1247-RC1 -
AC1: 'Comment on acp-2020-1247', Meehye Lee, 15 Sep 2021
The comment was uploaded in the form of a supplement: https://acp.copernicus.org/preprints/acp-2020-1247/acp-2020-1247-AC1-supplement.pdf
-
AC1: 'Comment on acp-2020-1247', Meehye Lee, 15 Sep 2021
-
RC2: 'Comment on acp-2020-1247', Anonymous Referee #1, 10 Jul 2021
General Comments:
This manuscript describes in situ observations of aerosol size distributions and composition at a rural site on the Korean peninsula, with occasional vertical aerosol concentration information provided by nearby instrumented balloon launches. The work highlights two characteristic daily patterns of aerosol size distributions, a new particle formation EOF and a haze or accumulation-mode dominated EOF. A key question raised by the authors is what drives the development of high PM2.5 loading in this region, and the size distributions observations are compared with meteorological conditions as well as physico-chemical observations of black carbon aerosol.
My main comment is that the argument that meteorological differences define the two EOF features is not well-supported by the analysis presented in the paper. In fact, the meteorological description of the periods is not consistent with the description provided on the same measurement period in this work, which is cited once in the present work:
Peterson, DA, et al. 2019. Meteorology influencing springtime
air quality, pollution transport, and visibility in Korea. Elem Sci
Anth, 7: 57. DOI: https://doi.org/10.1525/elementa.395
Peterson et al. characterize the period of May 17-22 as "stagnation under a persistent anticycle" and the period May 25-21 as "dynamic meteorology, low-level transport, and haze development", whereas in the present work (to the best of my understanding) that earlier period is described as "persistent anticyclone" (associated with EOF1) and the 2nd period as "synoptic-scale stagnation" (associated with EOF2). There seems to be a disconnect here. The caption of Figure 6 is consistent with the Peterson paper, but the abstract and perhaps the rest of the present manuscript are not. Furthermore, to my (perhaps untrained) eye, the meteorological patterns plotted in Fig 4a. and 4b. do not appear to be very different. Both appear to be fairly dynamic, quite distinct from the stagnant/blocking pattern shown in Fig. 4c. of Peterson et al.
A related issue is that the terms EOF1 and EOF2 are used fairly loosely in the manuscript. I understand them to be defined by a statistical treatment of the size distribution data and refer to two specific patterns of aerosol size development over a day. But these terms are used to represent actual time periods as well, e.g. in Figure 4 where the geopotential height averaged over EOF1 and EOF2 is given. What time periods are actually represented there? Are they EOF1 and 2 periods over the multi-year data set or during the KORUS-AQ measurement priod? I advise the authors to use different terms to define time periods in the multi-year data set and the KORUS measurement period.
I found the discussion of black carbon coating thickness as a useful diagnostic tool for the prevalent aerosol formation processes to be a very interesting concept and well-supported by the observations presented.
My bottom line for publishing this work in ACP is that the authors need to either do a lot more work showing the relationship between the characteristic aerosol EOF periods and synoptic scale meteorology, or they need to significantly de-emphasize claims of a relationship between them in the paper. In any case the time periods described need to be more clearly defined and not always tagged simply as EOF1 or EOF2.
Specific comments:
Separating the figures from the captions makes the figures difficult to review.
line 72 seems to imply all aerosol particles start from nucleation. Suggest rephrasing.
line 76 suggest change to "the level of pre-exisiting particles". As it stands, "a level" seems to imply that there's a minimum threshold of CS to achieve NPF, and I suspect that's not what the authors mean.
Line 169 and Figure S4. How were EOF1 and EOF2 periods determined? Is it just chance that 143 days each were found, or was that purposeful? Is there some threshold PCA value that causes a given time period to be included in the EOF1 or 2 bin?
Line 177 not sure what exactly is meant here. Are there >10^4/cm3 particles when only considering 20-30 nm particles?
Line 188. Same comment as line 177.
Line 193 "It turned out..." This sentence is very broad and isn't immediately supported by the details of what is meant so it seems out of place.
Line 243. What is meant by a mid-low cloud base height?
Line 251. Can you elaborate on why you consider EOF2 to correspond to "stagnant" conditions? To me this implies that in EOF1 there may be higher windspeeds, but this was not observed according to Table 1. In general, I find I am not convinced about the clear meteorological differences between the two cases. To my eye, the geopotential height and wind vector plots look fairly similar for the two EOF cases. This issue arises in Table 2 as well, where the boundary layer is just described in words without any analysis.
Line 308. "burst of particle(>3.5 nm) above 10^4" needs to be stated more clearly, at least give units for the concentration.
Line 313-314 "number of >3.5 nm particles tended to be backed up"- not sure what backed up means here.
Line 306-318. It may be helpful to define a particle size range of >3.5 nm to 0.3 um. It's a little confusing talking about >3.5 nm particles (which includes the 0.3-0.5 um and 0.5-1.0 um particles) as distinct from these other size ranges. I understand most of the number in the >3.5 nm particles must be below 300 nm, but you could make this paragraph significantly clearer by removing >3.5 nm particles and including >3.5-300 nm paticles as a size class.
Line 356. How would the weather conditions have suppressed condensation of volatiles onto particle surfaces? Please be more specific. The temperature was lower during EOF1, which seems like it would support more rather than less condensation.
Line 364-366. The claim in the second sentence is a big claim and it does not follow from the first sentence in this paragraph. It is an interesting claim, and I would encourage the authors to expand upon it. What number fraction of the particles is made up of BC particles? If they were not present, what would happen to the materials that would otherwise condense on them?
Figure 4. Maybe the continent outlines could be in a thicker pen? It's a little hard to make them out. Please give units for the geopotential height. What timescale do these back trajectories cover? Please state that as well.
Figure 7. It's a little difficult to know how to compare the sizes of the circles and squares (i.e. volume vs width). Maybe alongside the scale for the circle size vs. coating thickness you could do the same for the squares in the EOF1 case.Citation: https://doi.org/10.5194/acp-2020-1247-RC2 -
AC1: 'Comment on acp-2020-1247', Meehye Lee, 15 Sep 2021
The comment was uploaded in the form of a supplement: https://acp.copernicus.org/preprints/acp-2020-1247/acp-2020-1247-AC1-supplement.pdf
-
AC1: 'Comment on acp-2020-1247', Meehye Lee, 15 Sep 2021
-
AC1: 'Comment on acp-2020-1247', Meehye Lee, 15 Sep 2021
The comment was uploaded in the form of a supplement: https://acp.copernicus.org/preprints/acp-2020-1247/acp-2020-1247-AC1-supplement.pdf
Interactive discussion
Status: closed
-
RC1: 'Review of Lim et al.', Anonymous Referee #2, 29 Apr 2021
<General comments>
This paper discusses the formation mechanism of aerosols in northeast Asia. The authors used in-situ data obtained in Jeju Island from January 2013 to December 2016 and those during KORUS-AQ (May-June 2016) to investigate the variability in aerosol concentrations in the boundary layer. They used the empirical orthogonal function (EOF) analysis to classify the observed features. While the main topic of this paper is suitable for ACP, I do not think that the main conclusions are fully supported by observational evidence. I suggest that the authors largely reorganize the results and discussion and clarify the robust and new findings from this study. I recommend major revisions.<Specific comments>
Introduction
The formation of aerosols associated with meteorological cycles (e.g., anticyclones, cyclones, and air stagnation) has been extensively studied either in continental source areas or downwind regions. Variability in vertical profiles of aerosols associated with the evolution of the boundary layer has also been extensively investigated. Although this study might be the first to present such data in Korea, I think the fundamental mechanisms are common in many cases. The authors should briefly review previous studies in northeast Asia (e.g., TRACE-P, CARE-Beijing) and also in other regions, and discuss the similarity and difference between this study and previous ones. Here are some examples of the previous studies in northeast Asia.
Weber, R. J., et al. (2003): J. Geophys. Res., 108, 8814, doi:10.1029/2002JD003112.
Matsui, H., et al. (2009): J. Geophys. Res., 114, D00G13, doi:10.1029/2008JD010906.
Takegawa, N., et al. (2009): J. Geophys. Res., 114, D00G05, doi:10.1029/2008JD010857.
Haenel, A., et al. (2012): J. Geophys. Res., 117, D13201, doi:10.1029/2012JD017577.
Wang, J., et al., (2019): Atmos. Chem. Phys., 19, 8845-8861, doi:10.5194/acp-19-8845-2019.The following review paper would also be useful for the interpretation of NPF events in relation with meteorological conditions.
Kerminen, V.-M., et al. (2018), Environ. Res. Lett., 13, 103003. doi:10.1088/1748-9326/aadf3c.L144-146: Uncertainty in the coating thickness
The estimation of coating thickness of BC particles from SP2 data, although it has been used by many investigators, may contain significant uncertainties. The authors selected a BC core diameter of 200 +/- 20 nm. Why did the authors select this specific diameter? Is it reasonable to estimate a coating thickness of > 10 nm with the core diameter uncertainty of 20 nm?L249-251: Boundary layer stability
The dominance of a high-pressure system (subsidence) generally leads to the formation of strong inversions and stable boundary layers. The description in this paragraph seems to be opposite.L259-269: Entrainment
The authors conclude that the rapid increase in PM2.5 was due to the entrainment of particles from upstream areas. The authors state that elevation of aerosol concentrations is “believed” to occur by the intrusion of pollutants from the upper atmosphere. What is the basis for this statement? I do not think that the descriptions in this paragraph are supported by observational evidence.L306-318: Gas-to-particle conversion.
The discussion in this paragraph is highly speculative. The partitioning between HNO3 and NH4NO3 should be explicitly investigated to discuss the gas-to-particle conversion for nitrate aerosols. See, for example, Neuman, J. A., et al. (2003): J. Geophys. Res., 108, 4557, doi:10.1029/2003JD003616. Furthermore, the formation of (NH4)2SO4 might be controlled by aqueous-phase reactions in cloud droplets rather than condensation processes. Please reconsider the interpretation.L353-366: Interpretation of the coating thickness
The authors suggest that the coating thickness of rBC is s useful parameter to understand the formation of secondary aerosols, and also suggest that reducing BC emissions is the effective way to reduce PM2.5 in Asia. I think the descriptions in these paragraphs are also very speculative and not supported by observational evidence. Fig. 7 seems to the basis for this hypothesis, but I find many data points at lower aerosol mass loadings with thick coatings. It may be true that the EOF2 case can be characterized by high PM2.5 and thick coatings, but it does not necessarily mean that the coating thickness is the controlling factor. I would guess the correlation between the PM2.5 concentrations and the coating thickness is rather weak. Please show more convincing data to support the hypothesis. Otherwise I recommend that the authors should remove (at least tone down) this conclusion.<Minor comments>
L73-74, L182: SO2, NOx, and VOCs are not "condensable" gases but precursors.L100-121: Please specify the model number of the SMPS, CPC, and OPC. Please also describe how these instruments were evaluated and calibrated. It is not necessary to capitalize the first character of the name of the instruments.
L178-179: I do not think the estimate of GR values has three significant digits.
L331: "peaking below 100 nm" - Please specify number, surface, or mass.
Citation: https://doi.org/10.5194/acp-2020-1247-RC1 -
AC1: 'Comment on acp-2020-1247', Meehye Lee, 15 Sep 2021
The comment was uploaded in the form of a supplement: https://acp.copernicus.org/preprints/acp-2020-1247/acp-2020-1247-AC1-supplement.pdf
-
AC1: 'Comment on acp-2020-1247', Meehye Lee, 15 Sep 2021
-
RC2: 'Comment on acp-2020-1247', Anonymous Referee #1, 10 Jul 2021
General Comments:
This manuscript describes in situ observations of aerosol size distributions and composition at a rural site on the Korean peninsula, with occasional vertical aerosol concentration information provided by nearby instrumented balloon launches. The work highlights two characteristic daily patterns of aerosol size distributions, a new particle formation EOF and a haze or accumulation-mode dominated EOF. A key question raised by the authors is what drives the development of high PM2.5 loading in this region, and the size distributions observations are compared with meteorological conditions as well as physico-chemical observations of black carbon aerosol.
My main comment is that the argument that meteorological differences define the two EOF features is not well-supported by the analysis presented in the paper. In fact, the meteorological description of the periods is not consistent with the description provided on the same measurement period in this work, which is cited once in the present work:
Peterson, DA, et al. 2019. Meteorology influencing springtime
air quality, pollution transport, and visibility in Korea. Elem Sci
Anth, 7: 57. DOI: https://doi.org/10.1525/elementa.395
Peterson et al. characterize the period of May 17-22 as "stagnation under a persistent anticycle" and the period May 25-21 as "dynamic meteorology, low-level transport, and haze development", whereas in the present work (to the best of my understanding) that earlier period is described as "persistent anticyclone" (associated with EOF1) and the 2nd period as "synoptic-scale stagnation" (associated with EOF2). There seems to be a disconnect here. The caption of Figure 6 is consistent with the Peterson paper, but the abstract and perhaps the rest of the present manuscript are not. Furthermore, to my (perhaps untrained) eye, the meteorological patterns plotted in Fig 4a. and 4b. do not appear to be very different. Both appear to be fairly dynamic, quite distinct from the stagnant/blocking pattern shown in Fig. 4c. of Peterson et al.
A related issue is that the terms EOF1 and EOF2 are used fairly loosely in the manuscript. I understand them to be defined by a statistical treatment of the size distribution data and refer to two specific patterns of aerosol size development over a day. But these terms are used to represent actual time periods as well, e.g. in Figure 4 where the geopotential height averaged over EOF1 and EOF2 is given. What time periods are actually represented there? Are they EOF1 and 2 periods over the multi-year data set or during the KORUS-AQ measurement priod? I advise the authors to use different terms to define time periods in the multi-year data set and the KORUS measurement period.
I found the discussion of black carbon coating thickness as a useful diagnostic tool for the prevalent aerosol formation processes to be a very interesting concept and well-supported by the observations presented.
My bottom line for publishing this work in ACP is that the authors need to either do a lot more work showing the relationship between the characteristic aerosol EOF periods and synoptic scale meteorology, or they need to significantly de-emphasize claims of a relationship between them in the paper. In any case the time periods described need to be more clearly defined and not always tagged simply as EOF1 or EOF2.
Specific comments:
Separating the figures from the captions makes the figures difficult to review.
line 72 seems to imply all aerosol particles start from nucleation. Suggest rephrasing.
line 76 suggest change to "the level of pre-exisiting particles". As it stands, "a level" seems to imply that there's a minimum threshold of CS to achieve NPF, and I suspect that's not what the authors mean.
Line 169 and Figure S4. How were EOF1 and EOF2 periods determined? Is it just chance that 143 days each were found, or was that purposeful? Is there some threshold PCA value that causes a given time period to be included in the EOF1 or 2 bin?
Line 177 not sure what exactly is meant here. Are there >10^4/cm3 particles when only considering 20-30 nm particles?
Line 188. Same comment as line 177.
Line 193 "It turned out..." This sentence is very broad and isn't immediately supported by the details of what is meant so it seems out of place.
Line 243. What is meant by a mid-low cloud base height?
Line 251. Can you elaborate on why you consider EOF2 to correspond to "stagnant" conditions? To me this implies that in EOF1 there may be higher windspeeds, but this was not observed according to Table 1. In general, I find I am not convinced about the clear meteorological differences between the two cases. To my eye, the geopotential height and wind vector plots look fairly similar for the two EOF cases. This issue arises in Table 2 as well, where the boundary layer is just described in words without any analysis.
Line 308. "burst of particle(>3.5 nm) above 10^4" needs to be stated more clearly, at least give units for the concentration.
Line 313-314 "number of >3.5 nm particles tended to be backed up"- not sure what backed up means here.
Line 306-318. It may be helpful to define a particle size range of >3.5 nm to 0.3 um. It's a little confusing talking about >3.5 nm particles (which includes the 0.3-0.5 um and 0.5-1.0 um particles) as distinct from these other size ranges. I understand most of the number in the >3.5 nm particles must be below 300 nm, but you could make this paragraph significantly clearer by removing >3.5 nm particles and including >3.5-300 nm paticles as a size class.
Line 356. How would the weather conditions have suppressed condensation of volatiles onto particle surfaces? Please be more specific. The temperature was lower during EOF1, which seems like it would support more rather than less condensation.
Line 364-366. The claim in the second sentence is a big claim and it does not follow from the first sentence in this paragraph. It is an interesting claim, and I would encourage the authors to expand upon it. What number fraction of the particles is made up of BC particles? If they were not present, what would happen to the materials that would otherwise condense on them?
Figure 4. Maybe the continent outlines could be in a thicker pen? It's a little hard to make them out. Please give units for the geopotential height. What timescale do these back trajectories cover? Please state that as well.
Figure 7. It's a little difficult to know how to compare the sizes of the circles and squares (i.e. volume vs width). Maybe alongside the scale for the circle size vs. coating thickness you could do the same for the squares in the EOF1 case.Citation: https://doi.org/10.5194/acp-2020-1247-RC2 -
AC1: 'Comment on acp-2020-1247', Meehye Lee, 15 Sep 2021
The comment was uploaded in the form of a supplement: https://acp.copernicus.org/preprints/acp-2020-1247/acp-2020-1247-AC1-supplement.pdf
-
AC1: 'Comment on acp-2020-1247', Meehye Lee, 15 Sep 2021
-
AC1: 'Comment on acp-2020-1247', Meehye Lee, 15 Sep 2021
The comment was uploaded in the form of a supplement: https://acp.copernicus.org/preprints/acp-2020-1247/acp-2020-1247-AC1-supplement.pdf
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 798 | 338 | 57 | 1,193 | 131 | 49 | 51 |
- HTML: 798
- PDF: 338
- XML: 57
- Total: 1,193
- Supplement: 131
- BibTeX: 49
- EndNote: 51
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Cited
1 citations as recorded by crossref.
Saehee Lim
Paolo Laj
Sang-Woo Kim
Kang-Ho Ahn
Junsoo Gil
Xiaona Shang
Marco Zanatta
Kyeong-Sik Kang
This preprint has been withdrawn.
- Preprint
(1809 KB) - Metadata XML
-
Supplement
(826 KB) - BibTeX
- EndNote