the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 21 Jul 2022
Nitrogen oxides in the free troposphere: Implications for tropospheric oxidants and the interpretation of satellite NO2 measurements
Abstract. Satellite-based retrievals of tropospheric NO2 columns are used to infer NOx (NO+NO2) emissions at the surface. These retrievals rely on model information for the vertical distribution of NO2. The free tropospheric background above 2 km is particularly important because the sensitivity of the retrievals increases with altitude. Free tropospheric NOx also has a strong effect on tropospheric OH and ozone concentrations. Here we use observations from three aircraft campaigns (SEAC4RS, DC3, and ATom) and four atmospheric chemistry models (GEOS-Chem, GMI, TM5, and CAMS) to evaluate the model capabilities for simulating background NOx and attribute this background to sources. NO2 measurements over the southeast US during SEAC4RS and DC3 show increasing concentrations in the upper troposphere above 10 km, which is not replicated by GEOS-Chem although the model is consistent with the NO measurements. Using concurrent NO, NO2 and ozone observations from a DC3 flight in a thunderstorm outflow, we show that NO2 measurements in the upper troposphere are biased high, plausibly due to interference from thermally labile NO2 reservoirs, such as peroxynitric acid (HNO4) and methyl peroxy nitrate (MPN). We find that NO2 concentrations calculated from the NO measurements and NO-NO2 photochemical steady state (PSS) are more reliable to evaluate the vertical profiles of NO2 in models. GEOS-Chem reproduces the shape of the PSS-inferred NO2 profiles throughout the troposphere for SEAC4RS and DC3 but overestimates NO2 concentrations by about a factor of 2. The model underestimates MPN and alkyl nitrate concentrations, suggesting missing organic NOx chemistry. On the other hand, the standard GEOS-Chem model underestimates NO observations from the ATom campaigns over the Pacific and Atlantic Oceans, indicating a missing NOx source over the oceans. We find that we can account for this missing source by including in the model the photolysis of particulate nitrate on sea salt aerosols at rates inferred from laboratory studies and field observations of nitrous acid (HONO) over the Atlantic. The average NO2 column density for the ATom campaign in the GEOS-Chem simulation is 2.4×1014 molec cm-2 with particulate nitrate photolysis and 1.5×1014 molec cm-2 without, compared to 1.9×1014 molec cm-2 in the observations (using PSS NO2) and 1.4–2.4×1014 molec cm-2 in the GMI, TM5 and CAMS models. We find from GEOS-Chem that lightning is the main primary NOx source in the free troposphere over the tropics and southern midlatitudes, but aircraft emissions dominate at northern midlatitudes in winter and in summer over the oceans. Particulate nitrate photolysis increases ozone concentrations by up to 5 ppbv in the free troposphere in the northern extratropics in the model, which would largely correct the low model bias relative to ozonesonde observations. Global tropospheric OH concentrations increase by 19 %. The contribution of the free tropospheric background to the tropospheric NO2 columns observed by satellites over the contiguous US increases from 25 % in winter to 65 % in summer according to the GEOS-Chem vertical profiles. This needs to be accounted for when deriving NOx emissions from satellite NO2 column measurements.
-
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
(3099 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(3099 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-656', Anonymous Referee #1, 16 Aug 2022
This manuscript contains important analyses with regard to observations and model calculations of NOx in the free troposphere. The paper convincingly proves that NO2 from both laser-induced fluoresence and pholoysis/chemiluminescence instruments have significant high biaes due to interferences from other NOy species. NO2 calculated from photostationary state (PSS) assumptions likely yield a better estimate. However, the GEOS-Chem model overestimates the PSS-NO2 in the free troposphere in the southeast US during the SEAC4RS and DC3 experiments. In the remote free troposphere the model underestimates NO during ATom, but inclusion of of photolysis of particulate nitrate greatly improves the simulations. The implications of these findings for NO2 satellite retrievals are discussed. As found in previous studies, lightning is noted as the primary NOx source to the free troposphere over the tropics and southern midlatitudes in all seasons and over the US in summer. The free tropospheric component of the NO2 column over the US in summer (65%) is sufficiently large to make surface emissions estimates in this season difficult. This is an important conclusion of the manuscript. The paper is very well written and should be published with just a few minor revisions as noted below:
line 131: ....column retrievals, if the airborne measurements are assumed to be correct.
line 167: NOy/NO > 3 seems like this would be aged emssions, not fresh. Maybe this should be < 3 ?
line 563: "....errors in modeled tropospheric NO2 columns over clean areas in relatively small." This doesn't seem correct based on the model results shown in Figure 6. The difference between models is ~1 x 10^14 and the PSS-based NO2 column is ~1.9 x 10^14. Wouldn't this imply an uncertainty greater than 50%?
Citation: https://doi.org/10.5194/egusphere-2022-656-RC1 -
RC2: 'Comment on egusphere-2022-656', Anonymous Referee #2, 31 Aug 2022
This manuscript analyzes NOx concentrations and the vertical distribution in the troposphere based on three aircraft campaigns - SEAC4RS, DC3 & ATom – and four atmospheric chemistry models – GEOS-Chem, GMI, TM5 & CAMS. The authors present that measurements via LIF and P-CL overestimate NO2 concentrations in the upper troposphere due to thermal interferences and can be better represented by PSS calculations from measured NO. pNO3- photolysis as a missing NOx source in models is evaluated and is found to have a significant contribution, particularly over the oceans which improves the model performance for the ATom mission. Lightning and aircraft emissions are identified as main sources of NOx in the free troposphere, with individual contributions varying by latitude and season.
This manuscript is well written and presents an interesting and comprehensive analysis of free tropospheric NOx from measurements and models.
My main concerns are that the P-CL NO2 measurements were not corrected for thermal interferences from e.g. MPN and HNO4. This should be done and compared to the LIF measurements and the model results. I suggest some changes to the Figures, i.a. adding letters (a), (b), … to the subfigures for easier distinction, changing some of the axis scales and adding labels to all axis, but most importantly including error bars. Finally, I recommend the authors to elaborate in more detail how the NO2 columns were calculated from the measurements and what the uncertainties are. You will find my detailed comments and questions below. Once these points are addressed properly, this paper would be a valuable contribution to the current literature.
Major comments:
Lines 156 ff.: I assume this interference results from thermal decomposition of MPN. Could you go a bit more into detail? How was the correction determined? What’s the temperature in the instrument and what is the residence time of the sample gas?
Lines 162 ff.: What’s the residence time and the temperature in the photolysis cell (I assume the operation of the LEDs inevitably increases the temperature in the cell)? And how large is the resulting interference? Does thermal decay of HNO4 play a role? From Bourgeois et al. (2022), I understand that the wavelength of the LEDs in the photolysis cell is 385nm – do you experience and correct for photolytic interference from HONO?
Table 1: It would be helpful to add the uncertainties of each measurement in the table (e.g. as a fifth column).
Line 196: Do you perform a separate model run for the data along the flight track?
Line 252: How is the factor of 2.39 determined?
Line 278: It looks like the DC3 NO observations show a minimum for the lowest altitude bin. Is this real and if yes, why? Or do you have a small number of observations at this altitude?
Line 284: Do I understand correctly that the P-CL NO2 measurements were not corrected for the MPN and HNO4 interferences?
Figure 1: I recommend adding letters (a)-(f) to the subfigures for easier distinction. Could you lower the upper x-limit for NO to e.g. 200 or 250 pptv – so the profile is better to see. I find it a bit confusing that black was chosen as the observation in the NO vertical profiles and for the PSS calculation in the NO2 vertical profile. The LIF and P-CL observations are also hard to distinguish. I suggest using different colors and potentially decreasing the line width to prevent overlapping. Please add error bars to the modeled data, too.
Lines 307 ff.: So, the PSS is not entirely calculated from observations? - Maybe clarify in the legend of Figure 1.
Lines 312: It might be more accurate to use k(NO+CH3O2) as surrogate for k3=k(NO+RO2).
Line 320: Does jNO2 increase with altitude and plays a role in this, too?
Lines 326 ff.: You can estimate the interference from HNO4 and MPN with the temperature and the residence time in the instrument assuming first order decay (presented by Reed et al. (2016) and Nussbaumer et al. (2021)). If I see correctly you have measurements of both species available for SEAC4RS & DC3 and can calculate the resulting artifact and correct the measurements. For ATom (if needed), you could estimate HNO4 and MPN via PSS calculations. It could be helpful to compare your measurements with the PSS calculations for SEAC4RS & DC3 to see whether using the PSS to calculate HNO4 and MPN is a valid assumption. You can find the rate constants for the decay here: https://iupac-aeris.ipsl.fr/#.
Line 329: I don’t agree with the generalized assumption that HNO4 and MPN dissociate with the same efficiency as NO2. This largely depends on the temperature and the residence time. NO2 is subject to photolytic dissociation by the LEDs in the converter (which is mostly temperature independent), while HNO4 and MPN produce artifacts through thermal dissociation which is strongly temperature-dependent. MPN dissociates at lower temperatures compared to HNO4, and therefore produces the NO2 artifact at a higher efficiency than HNO4.
Line 332: How does it change the agreement between PSS NO/NO2 and P-CL NO/NO2 when accounting for the thermal interferences?
Line 341: Do you observe fresh lightning peaks in your NO signal which support this statement?
Line 343: It would be helpful to also show the trace gas concentrations as a function of time of the measurement period. Does NO continuously decrease over the period? How long is the measurement period?
Line 345: This likely looks different when correcting the P-CL NO2 data as described above. Are the changes described for the NO2 observations with increasing NOy/NO molar ratio significant? It looks like the scatter of the 1-minute observations is approximately the same as the increase/decrease of the NO2 lines.
Line 348: Is [HO2]=[RO2] a valid assumption? What’s the error arising from this assumption?
Equation (4): How did you calculate the ozone loss term via photolysis?
Figure 2: Again, letters (a)-(c) for the subfigures would be helpful to follow in the text. I recommend using the same colors as in Figure 1 for the according measurements / calculations. Again, LIF NO2 and P-CL NO2 are hard to distinguish and it is impossible to see the difference between the 1-minute observations. It could be helpful to use error bars instead. Please add error bars or the 1-minute data points for all species.
Lines 394 ff.: This seems a little bit like a circular argument to me. Wasn’t the PPS NO2 applied for the observations because the model identified a bias in the P-CL & LIF NO2 measurements (Lines 281 ff.)? And now the model performance is evaluated in comparison to the PSS NO2?
Line 403: How do these values compare to PSS calculations of MPN (production via CH3O2+NO2 and loss via decay and photolysis)?
Lines 411 ff.: This whole paragraph describes a method which is used in the subsequent section. I recommend shifting this paragraph to section 3.2.
Figure 3 and 4: Consider using letters for the subfigures. Why was the log scale chosen for the x-axis? I would find it more straight forward with a linear scale and it would also be easier to compare to Figure 1. Please add horizontal error bars for all traces.
Lines 447 ff.: How does this compare to the satellite measurements?
Line 471: Could you show a vertical profile of pNO3- somewhere - either in the manuscript or the supplement?
Figure 5: Please add labels to the axis and the color bar.
Line 533: I cannot follow this argument. It looks like measured pNO3- is much larger than the modeled values. Please clarify.
Lines 552 ff.: How are the PSS NO2 column density and the corresponding AMF determined? Could you provide a more detailed procedure? From Eq. (5) it looks like the viewing geometry of the satellite is required to determine the AMF. How does this apply to PSS or (in situ) measured values?
Lines 562 ff.: Considering a column density of ~1.3x1014 molec/cm² for GMI, isn’t a difference of ~1x1014 molec/cm² quite large (~80%)?
Figure 6: What are the errors on these values?
Line 582: Do these values refer to the annual mean over all layers? Looking at the NOx changes in Figure 7 (up to 400%), the 9% value probably has a large uncertainty. It could be helpful to add the uncertainty e.g. as the 1σ standard deviation.
Figure 7: Why do you show two plots with % changes and one with ppb changes?
Lines 615 ff.: Please state the 1σ values (or something comparable) for all averages.
Line 672: It looks like for the 30-60°N oceans and the U.S., absolute aircraft emissions are slightly larger in February compared to August, e.g. at 12km altitude for the U.S. ~70pptv NOx in February and ~30-40pptv NOx in August. Is this significant? Do you have an explanation for that?
Figure 9: Please add error bars, e.g. 1σ from averaging at each altitude bin. Please consider using letters (a) and (b) instead of referring to the right and the left panel.
Lines 739 ff.: I am not sure I can completely follow this argument. So, the right panel from Figure 9 show the summed contribution at each altitude plus everything that’s above up to 12km? Why do the winter profiles not show 100% at ground level?
Minor comments:
Line 346: shows
Line 613: consider replacing ‘worsening’ with ‘reducing’
Citation: https://doi.org/10.5194/egusphere-2022-656-RC2 - RC3: 'Comment on egusphere-2022-656', Anonymous Referee #3, 11 Sep 2022
- AC1: 'Author response to reviewer comments', Viral Shah, 11 Oct 2022
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-656', Anonymous Referee #1, 16 Aug 2022
This manuscript contains important analyses with regard to observations and model calculations of NOx in the free troposphere. The paper convincingly proves that NO2 from both laser-induced fluoresence and pholoysis/chemiluminescence instruments have significant high biaes due to interferences from other NOy species. NO2 calculated from photostationary state (PSS) assumptions likely yield a better estimate. However, the GEOS-Chem model overestimates the PSS-NO2 in the free troposphere in the southeast US during the SEAC4RS and DC3 experiments. In the remote free troposphere the model underestimates NO during ATom, but inclusion of of photolysis of particulate nitrate greatly improves the simulations. The implications of these findings for NO2 satellite retrievals are discussed. As found in previous studies, lightning is noted as the primary NOx source to the free troposphere over the tropics and southern midlatitudes in all seasons and over the US in summer. The free tropospheric component of the NO2 column over the US in summer (65%) is sufficiently large to make surface emissions estimates in this season difficult. This is an important conclusion of the manuscript. The paper is very well written and should be published with just a few minor revisions as noted below:
line 131: ....column retrievals, if the airborne measurements are assumed to be correct.
line 167: NOy/NO > 3 seems like this would be aged emssions, not fresh. Maybe this should be < 3 ?
line 563: "....errors in modeled tropospheric NO2 columns over clean areas in relatively small." This doesn't seem correct based on the model results shown in Figure 6. The difference between models is ~1 x 10^14 and the PSS-based NO2 column is ~1.9 x 10^14. Wouldn't this imply an uncertainty greater than 50%?
Citation: https://doi.org/10.5194/egusphere-2022-656-RC1 -
RC2: 'Comment on egusphere-2022-656', Anonymous Referee #2, 31 Aug 2022
This manuscript analyzes NOx concentrations and the vertical distribution in the troposphere based on three aircraft campaigns - SEAC4RS, DC3 & ATom – and four atmospheric chemistry models – GEOS-Chem, GMI, TM5 & CAMS. The authors present that measurements via LIF and P-CL overestimate NO2 concentrations in the upper troposphere due to thermal interferences and can be better represented by PSS calculations from measured NO. pNO3- photolysis as a missing NOx source in models is evaluated and is found to have a significant contribution, particularly over the oceans which improves the model performance for the ATom mission. Lightning and aircraft emissions are identified as main sources of NOx in the free troposphere, with individual contributions varying by latitude and season.
This manuscript is well written and presents an interesting and comprehensive analysis of free tropospheric NOx from measurements and models.
My main concerns are that the P-CL NO2 measurements were not corrected for thermal interferences from e.g. MPN and HNO4. This should be done and compared to the LIF measurements and the model results. I suggest some changes to the Figures, i.a. adding letters (a), (b), … to the subfigures for easier distinction, changing some of the axis scales and adding labels to all axis, but most importantly including error bars. Finally, I recommend the authors to elaborate in more detail how the NO2 columns were calculated from the measurements and what the uncertainties are. You will find my detailed comments and questions below. Once these points are addressed properly, this paper would be a valuable contribution to the current literature.
Major comments:
Lines 156 ff.: I assume this interference results from thermal decomposition of MPN. Could you go a bit more into detail? How was the correction determined? What’s the temperature in the instrument and what is the residence time of the sample gas?
Lines 162 ff.: What’s the residence time and the temperature in the photolysis cell (I assume the operation of the LEDs inevitably increases the temperature in the cell)? And how large is the resulting interference? Does thermal decay of HNO4 play a role? From Bourgeois et al. (2022), I understand that the wavelength of the LEDs in the photolysis cell is 385nm – do you experience and correct for photolytic interference from HONO?
Table 1: It would be helpful to add the uncertainties of each measurement in the table (e.g. as a fifth column).
Line 196: Do you perform a separate model run for the data along the flight track?
Line 252: How is the factor of 2.39 determined?
Line 278: It looks like the DC3 NO observations show a minimum for the lowest altitude bin. Is this real and if yes, why? Or do you have a small number of observations at this altitude?
Line 284: Do I understand correctly that the P-CL NO2 measurements were not corrected for the MPN and HNO4 interferences?
Figure 1: I recommend adding letters (a)-(f) to the subfigures for easier distinction. Could you lower the upper x-limit for NO to e.g. 200 or 250 pptv – so the profile is better to see. I find it a bit confusing that black was chosen as the observation in the NO vertical profiles and for the PSS calculation in the NO2 vertical profile. The LIF and P-CL observations are also hard to distinguish. I suggest using different colors and potentially decreasing the line width to prevent overlapping. Please add error bars to the modeled data, too.
Lines 307 ff.: So, the PSS is not entirely calculated from observations? - Maybe clarify in the legend of Figure 1.
Lines 312: It might be more accurate to use k(NO+CH3O2) as surrogate for k3=k(NO+RO2).
Line 320: Does jNO2 increase with altitude and plays a role in this, too?
Lines 326 ff.: You can estimate the interference from HNO4 and MPN with the temperature and the residence time in the instrument assuming first order decay (presented by Reed et al. (2016) and Nussbaumer et al. (2021)). If I see correctly you have measurements of both species available for SEAC4RS & DC3 and can calculate the resulting artifact and correct the measurements. For ATom (if needed), you could estimate HNO4 and MPN via PSS calculations. It could be helpful to compare your measurements with the PSS calculations for SEAC4RS & DC3 to see whether using the PSS to calculate HNO4 and MPN is a valid assumption. You can find the rate constants for the decay here: https://iupac-aeris.ipsl.fr/#.
Line 329: I don’t agree with the generalized assumption that HNO4 and MPN dissociate with the same efficiency as NO2. This largely depends on the temperature and the residence time. NO2 is subject to photolytic dissociation by the LEDs in the converter (which is mostly temperature independent), while HNO4 and MPN produce artifacts through thermal dissociation which is strongly temperature-dependent. MPN dissociates at lower temperatures compared to HNO4, and therefore produces the NO2 artifact at a higher efficiency than HNO4.
Line 332: How does it change the agreement between PSS NO/NO2 and P-CL NO/NO2 when accounting for the thermal interferences?
Line 341: Do you observe fresh lightning peaks in your NO signal which support this statement?
Line 343: It would be helpful to also show the trace gas concentrations as a function of time of the measurement period. Does NO continuously decrease over the period? How long is the measurement period?
Line 345: This likely looks different when correcting the P-CL NO2 data as described above. Are the changes described for the NO2 observations with increasing NOy/NO molar ratio significant? It looks like the scatter of the 1-minute observations is approximately the same as the increase/decrease of the NO2 lines.
Line 348: Is [HO2]=[RO2] a valid assumption? What’s the error arising from this assumption?
Equation (4): How did you calculate the ozone loss term via photolysis?
Figure 2: Again, letters (a)-(c) for the subfigures would be helpful to follow in the text. I recommend using the same colors as in Figure 1 for the according measurements / calculations. Again, LIF NO2 and P-CL NO2 are hard to distinguish and it is impossible to see the difference between the 1-minute observations. It could be helpful to use error bars instead. Please add error bars or the 1-minute data points for all species.
Lines 394 ff.: This seems a little bit like a circular argument to me. Wasn’t the PPS NO2 applied for the observations because the model identified a bias in the P-CL & LIF NO2 measurements (Lines 281 ff.)? And now the model performance is evaluated in comparison to the PSS NO2?
Line 403: How do these values compare to PSS calculations of MPN (production via CH3O2+NO2 and loss via decay and photolysis)?
Lines 411 ff.: This whole paragraph describes a method which is used in the subsequent section. I recommend shifting this paragraph to section 3.2.
Figure 3 and 4: Consider using letters for the subfigures. Why was the log scale chosen for the x-axis? I would find it more straight forward with a linear scale and it would also be easier to compare to Figure 1. Please add horizontal error bars for all traces.
Lines 447 ff.: How does this compare to the satellite measurements?
Line 471: Could you show a vertical profile of pNO3- somewhere - either in the manuscript or the supplement?
Figure 5: Please add labels to the axis and the color bar.
Line 533: I cannot follow this argument. It looks like measured pNO3- is much larger than the modeled values. Please clarify.
Lines 552 ff.: How are the PSS NO2 column density and the corresponding AMF determined? Could you provide a more detailed procedure? From Eq. (5) it looks like the viewing geometry of the satellite is required to determine the AMF. How does this apply to PSS or (in situ) measured values?
Lines 562 ff.: Considering a column density of ~1.3x1014 molec/cm² for GMI, isn’t a difference of ~1x1014 molec/cm² quite large (~80%)?
Figure 6: What are the errors on these values?
Line 582: Do these values refer to the annual mean over all layers? Looking at the NOx changes in Figure 7 (up to 400%), the 9% value probably has a large uncertainty. It could be helpful to add the uncertainty e.g. as the 1σ standard deviation.
Figure 7: Why do you show two plots with % changes and one with ppb changes?
Lines 615 ff.: Please state the 1σ values (or something comparable) for all averages.
Line 672: It looks like for the 30-60°N oceans and the U.S., absolute aircraft emissions are slightly larger in February compared to August, e.g. at 12km altitude for the U.S. ~70pptv NOx in February and ~30-40pptv NOx in August. Is this significant? Do you have an explanation for that?
Figure 9: Please add error bars, e.g. 1σ from averaging at each altitude bin. Please consider using letters (a) and (b) instead of referring to the right and the left panel.
Lines 739 ff.: I am not sure I can completely follow this argument. So, the right panel from Figure 9 show the summed contribution at each altitude plus everything that’s above up to 12km? Why do the winter profiles not show 100% at ground level?
Minor comments:
Line 346: shows
Line 613: consider replacing ‘worsening’ with ‘reducing’
Citation: https://doi.org/10.5194/egusphere-2022-656-RC2 - RC3: 'Comment on egusphere-2022-656', Anonymous Referee #3, 11 Sep 2022
- AC1: 'Author response to reviewer comments', Viral Shah, 11 Oct 2022
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 781 | 389 | 14 | 1,184 | 7 | 16 |
- HTML: 781
- PDF: 389
- XML: 14
- Total: 1,184
- BibTeX: 7
- EndNote: 16
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Cited
3 citations as recorded by crossref.
- Multidecadal increases in global tropospheric ozone derived from ozonesonde and surface site observations: can models reproduce ozone trends? A. Christiansen et al. 10.5194/acp-22-14751-2022
- Change in Tropospheric Ozone in the Recent Decades and Its Contribution to Global Total Ozone J. Liu et al. 10.1029/2022JD037170
- Evaluating NOx emissions and their effect on O3 production in Texas using TROPOMI NO2 and HCHO D. Goldberg et al. 10.5194/acp-22-10875-2022
Daniel J. Jacob
Ruijun Dang
Lok N. Lamsal
Sarah A. Strode
Stephen D. Steenrod
K. Folkert Boersma
Sebastian D. Eastham
Thibaud M. Fritz
Chelsea Thompson
Jeff Peischl
Ilann Bourgeois
Ilana B. Pollack
Benjamin A. Nault
Ronald C. Cohen
Pedro Campuzano-Jost
Jose L. Jimenez
Simone T. Andersen
Lucy J. Carpenter
Tomás Sherwen
Mat J. Evans
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(3099 KB) - Metadata XML