Global distribution of tetrafluoromethane (CF4) and hexafluoroethane (C2F6) emissions determined by inverse modeling

Püschel, Benjamin; Vojta, Martin; Kandler, Luise; Crotwell, Molly; Engel, Andreas; Krummel, Paul B.; Lunder, Chris R.; Mühle, Jens; O'Doherty, Simon; Prinn, Ronald G.; Stanley, Kieran M.; Vimont, Isaac; Vollmer, Martin K.; Wagenhäuser, Thomas; Weiss, Ray F.; Young, Dickon; Stohl, Andreas

doi:10.5194/egusphere-2025-5656

Preprints

https://doi.org/10.5194/egusphere-2025-5656

Preprints

18 Dec 2025

| 18 Dec 2025

Global distribution of tetrafluoromethane (CF₄) and hexafluoroethane (C₂F₆) emissions determined by inverse modeling

Benjamin Püschel, Martin Vojta, Luise Kandler, Molly Crotwell, Andreas Engel, Paul B. Krummel, Chris R. Lunder, Jens Mühle, Simon O'Doherty, Ronald G. Prinn, Kieran M. Stanley, Isaac Vimont, Martin K. Vollmer, Thomas Wagenhäuser, Ray F. Weiss, Dickon Young, and Andreas Stohl

Abstract. The perfluorocarbons CF₄ and C₂F₆ are among the most potent greenhouse gases with lifetimes of fifty and ten thousand years, respectively. They are both primarily emitted during aluminum smelting and electronics manufacturing. We perform the first regionally resolved global inversion of CF₄ and C₂F₆, providing atmospheric measurement-based top-down emission estimates for 2006–2023 using the FLEXPART transport model and the FLEXINVERT+ framework.

Introducing a global-total constraint to align the inversion results with the relatively accurate global total emissions from the AGAGE 12-box model stabilizes the emissions in poorly monitored regions. Compared to the global bottom-up inventory EDGAR, the inversion increases global CF₄ and C₂F₆ emissions by factors of 2.6 ± 0.3 and 3.1 ± 0.7, respectively, for 2018–2023. China dominates global emissions, contributing 56 % (CF₄) and 58 % (C₂F₆) in 2018–2023. The contribution of South and Southeast Asia to global emissions rose from about 6 % and 10 % in 2006–2011 to 22 % and 18 % by 2018–2023, respectively, though large uncertainties remain due to a lack of measurements in the region itself. European emissions declined until 2010, then stabilized and contribute 2 %–3 % to global emissions by 2018–2023. U.S. CF₄ emissions remain constant and C₂F₆ emissions decreased steadily (reaching 3 %–4 % by 2018–2023), with a temporary drop in 2009 likely linked to the financial crisis. On the global scale, our results suggest a contribution of 81 % by the aluminum industry to total CF₄ emissions and 48 % by the electronics industry to global C₂F₆ emissions.

Received: 16 Nov 2025 – Discussion started: 18 Dec 2025

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Preprint (PDF, 8352 KB)

Supplement (2368 KB)

Download & links

Status: closed (peer review stopped)

RC1:
'Comment on egusphere-2025-5656', Anonymous Referee #1, 28 Mar 2026
Review of “Global regional inversion of tetrafluoromethane (CF4) and hexafluoroethane (C2F6) emissions determined by inverse modeling”, by Benjamin Püschel et al.
General comments:
The authors use the FLEXPART transport model and the FLEXINVERT⁺ inversion framework, and introduce a post-processing constraint that aligns the sum of regional emissions with global total estimates from the AGAGE 12-box model. While this approach is potentially useful, I have several major concerns regarding the methodology and the associated conclusions. In particular, it is unclear whether the introduction of a global-total constraint can effectively stabilize emissions in poorly monitored regions. Such a constraint may help to constrain the aggregate emissions of these regions, but it does not necessarily improve the ability to distinguish emissions among them, especially under limited observational coverage. The authors are encouraged to provide additional justification and supporting analyses (e.g., posterior error covariance, sensitivity tests, or resolution diagnostics) to demonstrate the effectiveness of this approach. Furthermore, the limitations of the inversion framework in resolving regional emissions should be more clearly acknowledged and discussed. Overall, substantial revisions are required before the manuscript can be considered for publication.
Specific Comments
Abstract: I am not convinced that introducing a global-total constraint can effectively stabilize emissions in poorly monitored regions.While such a constraint, combined with well-constrained regions (e.g., Europe, the USA, and East Asia), may help constrain the aggregate emissions of the remaining poorly monitored regions, it does not necessarily improve the ability to distinguish emissions among these regions (e.g., Africa, South America, South Asia, and Southeast Asia). In other words, the global-total constraint effectively reduces only one degree of freedom (the global sum), but the partitioning of emissions among poorly observed regions likely remains weakly constrained. Have the authors examined the posterior error covariance (or correlation) among these regions? It is likely that substantial negative correlations persist, indicating that the inversion cannot robustly separate their emissions despite the global constraint. Therefore, statements suggesting that the global-total constraint “stabilizes emissions in poorly monitored regions” may be overstated and should be revised or more carefully qualified.

Abstract: It is unclear whether the inversion framework can robustly distinguish emissions from South and Southeast Asia, given the limited observational constraints in these regions. The authors are encouraged to provide supporting evidence (e.g., posterior error correlations, resolution metrics, or sensitivity analyses) to substantiate this claim. In addition, uncertainties in the fractional contributions to the global total should be quantified and reported, given their potentially large magnitude.

Introduction: This statement may give the impression that the present study can close the global gap by robustly constraining regional emissions in South and Southeast Asia, South America, and Africa. However, this may be overstated, as posterior emissions in these regions are likely still strongly influenced by prior assumptions due to limited observational constraints. While the study may provide useful insights, it may not be sufficient to fully resolve emissions in these poorly monitored regions. The authors are encouraged to moderate this statement or clarify its scope.

2.2: The use of “daily intervals” for emission sensitivities is unclear. Earlier, the observations were averaged to 3-hour intervals, could the authors clarify why a different temporal resolution is used here? Is there any inconsistency between the temporal resolution of the observations and that of the sensitivities, and if so, how is this handled in the inversion?

Equation 4: While the formulation shows how the posterior emissions are adjusted after imposing the global-total constraint, it is unclear how the posterior uncertainties for individual grid cells are affected.How does this constraint propagate into the posterior error covariance? In particular, does it reduce uncertainties uniformly or introduce additional correlations among grid cells? The authors are encouraged to provide the corresponding formulation or a more detailed discussion.

Equation 4:The constraint is applied as a post‑processing step using Eq. (4). How sensitive are the final regional emissions to the prescribed uncertainty matrix C? A sensitivity test with, e.g., ±50% changes in C would help readers gauge the influence.

The prior emissions are described as “adjusted BU inventory” (EDGAR with modifications for China, India, South Korea). This adjustment brings the prior closer to the global total but also introduces external BU information.I recommend presenting both the original EDGAR and the adjusted prior separately in the main text (e.g., in Fig. 4 and in the regional time series) to allow readers to distinguish the effect of the adjustment from the pure inversion. How does the use of this adjusted prior affect the attribution of the TD BU discrepancy in Section 3.3? The discrepancy is defined against EDGAR, but the inversion prior is already partly corrected. This should be explicitly noted.

The paper notes a pronounced drop in Chinese CF4emissions (approximately 20%) in 2019, identifying this as a potential inversion artifact. While the authors link it to the US–China trade war’s impact on aluminum production (around 4%), the 20% reduction far exceeds what can be explained by output fluctuations alone. The authors are advised to further investigate whether this anomaly stems from missing observations at specific sites, anomalies in the meteorological fields, or shifts in the allocation weights of the global constraint equation in that year.

The study uses sectoral emission fractions (aluminum industry: 0.10; electronics industry: 0.40) from Kim et al. (2014) to attribute sources. However, given the evolution of aluminum smelting technologies (e.g., anode effect control) and electronics industry abatement technologies (e.g., end-of-pipe treatment) over the past 15 years, retaining these fixed ratios may introduce biases in sectoral attribution. The authors themselves acknowledge unrealistic allocation results for South Korea and Taiwan. It is recommended that the discussion include an analysis of temporal changes in sectoral emission fractions.

The study reports that the share of emissions from South and Southeast Asia (SSEA), particularly India, has surged from 6–10% in the early period to 18–22% in recent years. Given the complete absence of in-situ observational stations in this region, to what extent are these “increases” driven by forced allocation to unconstrained regions via the global constraint equation? The authors are encouraged to explicitly assess the robustness of the growth trendsfor these regions.

Technical corrections:
Section 2.2 mentions the use of a 100-day backward tracking window; whilst this helps to improve far-field constraints, it may also introduce coupling between earlier emission signals and prior emissions. It is recommended that the implications of this choice for interannual variations in emissions and regional allocation be further explained.

2.2: Can the authors further clarify why longer backward simulationperiods (e.g., 50 or 100 days) are expected to improve the robustness of the inversion? It is unclear how emissions from such long time scales contribute additional information to constrain regional emissions. Air masses older than ~30 days are likely well mixed and may primarily influence the background (baseline) rather than providing sensitivity to regional sources.

The observation uncertainty in R is defined with thresholds (2.0 ppt for CF4, 0.4 ppt for C2F6) and a linear increase above those thresholds. Provide a brief justification (e.g., based on model‑observation residuals from a test run) or cite a previous study where this approach was validated.

The paper provides data citations and some DOIs, but does not explicitly state whether the core inversion code, post-processing scripts and structural descriptions of the gridded emission output files are available. Given EGUsphere’s requirements regarding reproducibility, it is recommended that the authors provide this information in the ‘Data availability’ section.

South and Southeast Asia (Fig. 9) show a rapid emission increase, but the uncertainties remain large and the error reduction is low. The conclusion that this region contributes 22% (CF4) and 18% (C2F6) to global emissions in 2018–2023 is important, yet the text should more explicitly caution that these numbers are heavily influenced by the global constraint rather than by local measurements.

The fractional contributions to the TD‑BU discrepancy are shown for EDGAR. Given that the prior already incorporated adjustments for some countries, the discrepancy might be partially reduced by construction. I suggest adding a version of Fig. 10 using the original EDGAR inventory as the BU reference, to show the full extent of underestimation.

Taiwan is a province of China. Revise Taiwan to Taiwan Province of China. Revise China to China mainland.
Citation: https://doi.org/10.5194/egusphere-2025-5656-RC1
RC2:
'Comment on egusphere-2025-5656', Anonymous Referee #2, 02 Apr 2026
This study presents global and regional/national emissions of two perfluorocarbons using measurements from AGAGE, NOAA and the Chinese flask network (previously published data), through 2023.
The manuscript is well-written (I came across almost no typos), which made it very easy to read. The figures and text are polished with appropriate uncertainties, etc.
However, I have two major reservations, which in my opinion, requires the results to be presented differently in order to be considered for publication.
The first is in the level of disaggregation the authors are presenting for the regions with no measurements. The authors present this work in several places as “the first regionally-resolved inversion results”; however, given that there has been no change in the measurements used (over previous studies), the results need to be presented differently. I agree that it may be possible to use a global approach to resolve emissions at higher resolution than the box model resolution of Western et al., (semi-hemispheric) but I am not convinced that it is possible to resolve individual countries (e.g., India, Canada, Russia, Singapore/Malaysia) with those measurements with a global total constraint alone. Much more evidence and testing would be needed. Let alone resolving, as shown in Fig S1, at an even higher spatial resolution within these countries (even with a 100km correlation scale). I’d suggest aggregating the emissions over much larger continental scales. Specific comments on this point:
Even in the "well-constrained" Europe, the measurements do not constrain the whole EU27 but parts of north-western Europe (see for example Say et al., 2021 only presented Northwest Europe). so care should be taken not to over-extrapolate.

In the USA, the flask network sensitivity is relatively low compared to the in situ stations, and more information is needed to demonstrate that the whole country can be resolved (looking at the SI Table on measurements, the flask measurement frequency ranges between 2d and 2m, so still quite sparse).

The second comment is about novelty. As much of this work is already published elsewhere (with the exception of the USA results) I would have expected this manuscript to extensively compare results with those studies. The main difference, for the regions that are well-constrained, is in using a different transport model (which is important). Yet there is little comparison with those studies. I’d suggest focusing this manuscript on the comparisons with other studies for the well-constrained regions (i.e. what is the impact of transport model choice) and doing more on the USA results, since to my knowledge, this is the first study to look at this region.
I provide some specific comments as well, but as there are major issues to do with the presentation of results, I refrain from presenting detailed specific comments on those aspects at this stage.
Line 104- why are the instrumental uncertainties not used? They have important time-varying patterns that the model uncertainties sometimes do not capture.

Line 150 – why is extrapolating C2F6 based on aluminium production only valid when there is a large semiconductor source?

Line 167 – is a spatial correlation scale of 100km appropriate for point sources? Why would one expect point source emissions uncertainties to correlate as the emissions are episodic.

Line 175 – what is the slope of the linear increase in model error?

Section 3.1 – the discussions about global emissions (e.g., financial crisis) are already presented in other papers and does not need to be discussed again.

Fig 4 – it does not make sense to present global UNFCCC numbers when the reporting for UNFCCC does not have global coverage

Line 328 – is there evidence to back up a shift within the USA to aluminum over semiconductor manufacturing?
Citation: https://doi.org/10.5194/egusphere-2025-5656-RC2

Status: closed (peer review stopped)

RC1:
'Comment on egusphere-2025-5656', Anonymous Referee #1, 28 Mar 2026
Review of “Global regional inversion of tetrafluoromethane (CF4) and hexafluoroethane (C2F6) emissions determined by inverse modeling”, by Benjamin Püschel et al.
General comments:
The authors use the FLEXPART transport model and the FLEXINVERT⁺ inversion framework, and introduce a post-processing constraint that aligns the sum of regional emissions with global total estimates from the AGAGE 12-box model. While this approach is potentially useful, I have several major concerns regarding the methodology and the associated conclusions. In particular, it is unclear whether the introduction of a global-total constraint can effectively stabilize emissions in poorly monitored regions. Such a constraint may help to constrain the aggregate emissions of these regions, but it does not necessarily improve the ability to distinguish emissions among them, especially under limited observational coverage. The authors are encouraged to provide additional justification and supporting analyses (e.g., posterior error covariance, sensitivity tests, or resolution diagnostics) to demonstrate the effectiveness of this approach. Furthermore, the limitations of the inversion framework in resolving regional emissions should be more clearly acknowledged and discussed. Overall, substantial revisions are required before the manuscript can be considered for publication.
Specific Comments
Abstract: I am not convinced that introducing a global-total constraint can effectively stabilize emissions in poorly monitored regions.While such a constraint, combined with well-constrained regions (e.g., Europe, the USA, and East Asia), may help constrain the aggregate emissions of the remaining poorly monitored regions, it does not necessarily improve the ability to distinguish emissions among these regions (e.g., Africa, South America, South Asia, and Southeast Asia). In other words, the global-total constraint effectively reduces only one degree of freedom (the global sum), but the partitioning of emissions among poorly observed regions likely remains weakly constrained. Have the authors examined the posterior error covariance (or correlation) among these regions? It is likely that substantial negative correlations persist, indicating that the inversion cannot robustly separate their emissions despite the global constraint. Therefore, statements suggesting that the global-total constraint “stabilizes emissions in poorly monitored regions” may be overstated and should be revised or more carefully qualified.

Abstract: It is unclear whether the inversion framework can robustly distinguish emissions from South and Southeast Asia, given the limited observational constraints in these regions. The authors are encouraged to provide supporting evidence (e.g., posterior error correlations, resolution metrics, or sensitivity analyses) to substantiate this claim. In addition, uncertainties in the fractional contributions to the global total should be quantified and reported, given their potentially large magnitude.

Introduction: This statement may give the impression that the present study can close the global gap by robustly constraining regional emissions in South and Southeast Asia, South America, and Africa. However, this may be overstated, as posterior emissions in these regions are likely still strongly influenced by prior assumptions due to limited observational constraints. While the study may provide useful insights, it may not be sufficient to fully resolve emissions in these poorly monitored regions. The authors are encouraged to moderate this statement or clarify its scope.

2.2: The use of “daily intervals” for emission sensitivities is unclear. Earlier, the observations were averaged to 3-hour intervals, could the authors clarify why a different temporal resolution is used here? Is there any inconsistency between the temporal resolution of the observations and that of the sensitivities, and if so, how is this handled in the inversion?

Equation 4: While the formulation shows how the posterior emissions are adjusted after imposing the global-total constraint, it is unclear how the posterior uncertainties for individual grid cells are affected.How does this constraint propagate into the posterior error covariance? In particular, does it reduce uncertainties uniformly or introduce additional correlations among grid cells? The authors are encouraged to provide the corresponding formulation or a more detailed discussion.

Equation 4:The constraint is applied as a post‑processing step using Eq. (4). How sensitive are the final regional emissions to the prescribed uncertainty matrix C? A sensitivity test with, e.g., ±50% changes in C would help readers gauge the influence.

The prior emissions are described as “adjusted BU inventory” (EDGAR with modifications for China, India, South Korea). This adjustment brings the prior closer to the global total but also introduces external BU information.I recommend presenting both the original EDGAR and the adjusted prior separately in the main text (e.g., in Fig. 4 and in the regional time series) to allow readers to distinguish the effect of the adjustment from the pure inversion. How does the use of this adjusted prior affect the attribution of the TD BU discrepancy in Section 3.3? The discrepancy is defined against EDGAR, but the inversion prior is already partly corrected. This should be explicitly noted.

The paper notes a pronounced drop in Chinese CF4emissions (approximately 20%) in 2019, identifying this as a potential inversion artifact. While the authors link it to the US–China trade war’s impact on aluminum production (around 4%), the 20% reduction far exceeds what can be explained by output fluctuations alone. The authors are advised to further investigate whether this anomaly stems from missing observations at specific sites, anomalies in the meteorological fields, or shifts in the allocation weights of the global constraint equation in that year.

The study uses sectoral emission fractions (aluminum industry: 0.10; electronics industry: 0.40) from Kim et al. (2014) to attribute sources. However, given the evolution of aluminum smelting technologies (e.g., anode effect control) and electronics industry abatement technologies (e.g., end-of-pipe treatment) over the past 15 years, retaining these fixed ratios may introduce biases in sectoral attribution. The authors themselves acknowledge unrealistic allocation results for South Korea and Taiwan. It is recommended that the discussion include an analysis of temporal changes in sectoral emission fractions.

The study reports that the share of emissions from South and Southeast Asia (SSEA), particularly India, has surged from 6–10% in the early period to 18–22% in recent years. Given the complete absence of in-situ observational stations in this region, to what extent are these “increases” driven by forced allocation to unconstrained regions via the global constraint equation? The authors are encouraged to explicitly assess the robustness of the growth trendsfor these regions.

Technical corrections:
Section 2.2 mentions the use of a 100-day backward tracking window; whilst this helps to improve far-field constraints, it may also introduce coupling between earlier emission signals and prior emissions. It is recommended that the implications of this choice for interannual variations in emissions and regional allocation be further explained.

2.2: Can the authors further clarify why longer backward simulationperiods (e.g., 50 or 100 days) are expected to improve the robustness of the inversion? It is unclear how emissions from such long time scales contribute additional information to constrain regional emissions. Air masses older than ~30 days are likely well mixed and may primarily influence the background (baseline) rather than providing sensitivity to regional sources.

The observation uncertainty in R is defined with thresholds (2.0 ppt for CF4, 0.4 ppt for C2F6) and a linear increase above those thresholds. Provide a brief justification (e.g., based on model‑observation residuals from a test run) or cite a previous study where this approach was validated.

The paper provides data citations and some DOIs, but does not explicitly state whether the core inversion code, post-processing scripts and structural descriptions of the gridded emission output files are available. Given EGUsphere’s requirements regarding reproducibility, it is recommended that the authors provide this information in the ‘Data availability’ section.

South and Southeast Asia (Fig. 9) show a rapid emission increase, but the uncertainties remain large and the error reduction is low. The conclusion that this region contributes 22% (CF4) and 18% (C2F6) to global emissions in 2018–2023 is important, yet the text should more explicitly caution that these numbers are heavily influenced by the global constraint rather than by local measurements.

The fractional contributions to the TD‑BU discrepancy are shown for EDGAR. Given that the prior already incorporated adjustments for some countries, the discrepancy might be partially reduced by construction. I suggest adding a version of Fig. 10 using the original EDGAR inventory as the BU reference, to show the full extent of underestimation.

Taiwan is a province of China. Revise Taiwan to Taiwan Province of China. Revise China to China mainland.
Citation: https://doi.org/10.5194/egusphere-2025-5656-RC1
RC2:
'Comment on egusphere-2025-5656', Anonymous Referee #2, 02 Apr 2026
This study presents global and regional/national emissions of two perfluorocarbons using measurements from AGAGE, NOAA and the Chinese flask network (previously published data), through 2023.
The manuscript is well-written (I came across almost no typos), which made it very easy to read. The figures and text are polished with appropriate uncertainties, etc.
However, I have two major reservations, which in my opinion, requires the results to be presented differently in order to be considered for publication.
The first is in the level of disaggregation the authors are presenting for the regions with no measurements. The authors present this work in several places as “the first regionally-resolved inversion results”; however, given that there has been no change in the measurements used (over previous studies), the results need to be presented differently. I agree that it may be possible to use a global approach to resolve emissions at higher resolution than the box model resolution of Western et al., (semi-hemispheric) but I am not convinced that it is possible to resolve individual countries (e.g., India, Canada, Russia, Singapore/Malaysia) with those measurements with a global total constraint alone. Much more evidence and testing would be needed. Let alone resolving, as shown in Fig S1, at an even higher spatial resolution within these countries (even with a 100km correlation scale). I’d suggest aggregating the emissions over much larger continental scales. Specific comments on this point:
Even in the "well-constrained" Europe, the measurements do not constrain the whole EU27 but parts of north-western Europe (see for example Say et al., 2021 only presented Northwest Europe). so care should be taken not to over-extrapolate.

In the USA, the flask network sensitivity is relatively low compared to the in situ stations, and more information is needed to demonstrate that the whole country can be resolved (looking at the SI Table on measurements, the flask measurement frequency ranges between 2d and 2m, so still quite sparse).

The second comment is about novelty. As much of this work is already published elsewhere (with the exception of the USA results) I would have expected this manuscript to extensively compare results with those studies. The main difference, for the regions that are well-constrained, is in using a different transport model (which is important). Yet there is little comparison with those studies. I’d suggest focusing this manuscript on the comparisons with other studies for the well-constrained regions (i.e. what is the impact of transport model choice) and doing more on the USA results, since to my knowledge, this is the first study to look at this region.
I provide some specific comments as well, but as there are major issues to do with the presentation of results, I refrain from presenting detailed specific comments on those aspects at this stage.
Line 104- why are the instrumental uncertainties not used? They have important time-varying patterns that the model uncertainties sometimes do not capture.

Line 150 – why is extrapolating C2F6 based on aluminium production only valid when there is a large semiconductor source?

Line 167 – is a spatial correlation scale of 100km appropriate for point sources? Why would one expect point source emissions uncertainties to correlate as the emissions are episodic.

Line 175 – what is the slope of the linear increase in model error?

Section 3.1 – the discussions about global emissions (e.g., financial crisis) are already presented in other papers and does not need to be discussed again.

Fig 4 – it does not make sense to present global UNFCCC numbers when the reporting for UNFCCC does not have global coverage

Line 328 – is there evidence to back up a shift within the USA to aluminum over semiconductor manufacturing?
Citation: https://doi.org/10.5194/egusphere-2025-5656-RC2

Supplement

https://doi.org/10.5194/egusphere-2025-5656-supplement

Data sets

CF4 & C2F6 global annual emission fields between 2006 and 2023 B. Püschel and L. Kandler https://doi.org/10.25365/phaidra.751

CF4 & C2F6 global daily mole fraction fields between 2006 and 2023 B. Püschel and L. Kandler https://doi.org/10.25365/phaidra.752

The dataset of in-situ measurements of chemically and radiatively important atmospheric gases from the Advanced Global Atmospheric Gas Experiment (AGAGE) and affiliated stations (Version 20250123) R. Prinn et al. https://doi.org/10.60718/0FXA-QF43

Atmospheric Dry Air Mole Fractions of CF4 from the NOAA GML Surface and Aircraft Vertical Profile Network. I. Vimont et al. https://doi.org/10.15138/1sds-1672

Available Data for "Inverse Modeling of High Global Warming Potential Perfluorinated Greenhouse Gases in Southeastern China" Yuyang Chen https://doi.org/10.7910/DVN/BFINPM

Measurement of Halocarbons and greenhouse gases at Trollhaugen C. Lunder et al. https://doi.org/10.48597/P8CA-KWZG

Model code and software

FLEXPART-v11 L. Bakels et al. https://doi.org/10.5281/zenodo.13748655

FLEXINVERT+: Code M. Vojta et al. https://doi.org/10.25365/phaidra.488

Viewed

Total article views: 3,363 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
1,627	1,549	187	3,363	316	236	204

HTML: 1,627
PDF: 1,549
XML: 187
Total: 3,363
Supplement: 316
BibTeX: 236
EndNote: 204

Views and downloads (calculated since 18 Dec 2025)

Month	HTML	PDF	XML	Total
Dec 2025	515	370	75	960
Jan 2026	372	384	25	781
Feb 2026	194	221	40	455
Mar 2026	440	449	44	933
Apr 2026	93	98	3	194
May 2026	13	27	0	40

Cumulative views and downloads (calculated since 18 Dec 2025)

Month	HTML	PDF	XML	Total
Dec 2025	515	370	75	960
Jan 2026	372	384	25	781
Feb 2026	194	221	40	455
Mar 2026	440	449	44	933
Apr 2026	93	98	3	194
May 2026	13	27	0	40

Viewed (geographical distribution)

Total article views: 3,353 (including HTML, PDF, and XML) Thereof 3,353 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 19 May 2026

Short summary

We present global and regionally resolved emissions of the two highly potent greenhouse gases CF₄ and C₂F₆ from 2006 to 2023, based on atmospheric measurements and transport modeling. In nearly all regions studied, our results show that national reports and industry-based data underestimate emissions. While emissions in Europe and the U.S. declined, we find rising emissions in South, Southeast, and East Asia. China and India are the major contributors to global emissions in recent years.


Total:	0
HTML:	0
PDF:	0
XML:	0