the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Oct 2025
Global Observations and European emissions of the halogenated olefins HFO-1234yf, HFO-1234ze(E), and HCFO-1233zd(E) from the AGAGE (Advanced Global Atmospheric Gases Experiment) network
Abstract. Atmospheric observations of the widely used hydrofluoroolefins (HFOs) HFO-1234yf (2,3,3,3-tetrafluoroprop-1-ene), and HFO-1234ze(E) (E-1,3,3,3-tetrafluoroprop-1-ene), and the hydrochlorofluoroolefin (HCFO) HCFO-1233zd(E) (E-1-chloro-3,3,3-trifluoroprop-1-ene) are reported from the Advanced Global Atmospheric Gases Experiment (AGAGE) network. Since 2011, pollution events have grown in magnitude and frequency at sites which are influenced by regional emissions, while remote stations show first appearances of these substances. For HFO-1234yf and HFO-1234ze(E) winter peak mole fractions in background northern hemisphere air have grown from 0.03 ppt (picomol mol-1, parts-per-trillion in dry air) in the mid-2010s to 0.25 ppt in 2024, while the atmospherically more stable HCFO-1233zd(E) showed an increase from 0.06 ppt to 0.45 ppt. This suggests increasing usage of these haloolefins to replace hydrofluorocarbons (HFCs), which are regulated for phase-down over the next decades. Using European observations and the inverse modeling frameworks InTEM, ELRIS, and RHIME we determine emission trends and regional distribution. For Northwest Europe, emissions of HFO-1234yf increased steadily and rapidly from <0.1 Gg yr-1 in 2014 to 1.50 Gg yr-1 by 2023, presumably due to its introduction in the mobile air conditioning and refrigeration sectors. HFO-1234ze(E) emissions were low during 2014–2017, followed by a rapid increase in 2018/2019, potentially due its introduction as aerosol propellant, after which they increased more slowly to 0.96 Gg yr-1 by 2023. HCFO-1233zd(E) emissions are derived from 2017 onwards, showing a steady increase from 0.2 Gg yr-1 to 1.0 Gg yr-1 in 2023.
-
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
(17400 KB)
-
Supplement
(12 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(17400 KB) - Metadata XML
-
Supplement
(12 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-4824', Isaac Vimont, 24 Nov 2025
- AC2: 'Reply on RC1', Martin Vollmer, 06 Mar 2026
-
RC2: 'Comment on egusphere-2025-4824', Anonymous Referee #2, 07 Jan 2026
Review of “Global Observations and European emissions of the halogenated oelfins HFO-1234yf, HFO-1234ze(E), and HCFO-1234zd(E) from the AGAGE network”
Overview: This paper reports on global observations of two HFOs and one HCFO that are being increasingly used as replacements for HFCs. Further, the authors use multiple inverse modeling frameworks to estimate emissions and emissions trends in NW Europe. Authors report increasing prevalence of detectable levels of these gases since 2011, with increasing levels observed in pollution plumes (normalized to a HFC observation). Inverse modeling suggests increasing emissions for these gases in NW Europe, with somewhat rapid increases inferred in the time series. Haloolefins are being increasingly used worldwide, but emissions magnitudes and environmental and health impacts are poorly known. Given the reactive nature of these gases, they are harder to observe and track with sparse global networks. This paper nicely present important measurements of these gases and infers reasonable conclusions. The multi-method inverse assessment of Europe is a strong addition to the paper, and illustrates the need to consider more measurements near emissions regions to truly characterize HFO’s moving forwards. The paper is appropriate for ACP. I have some modest concerns and suggestions for the authors, which should be straightforward to address and would strengthen the manuscript.
General comments:
Add uncertainties: In reporting concentrations, enhancements, and flux numbers, please add uncertainties. For example, in the abstract when stating the increase in ppt’s, or when stating estimated emissions, confidence intervals or uncertainties should be added. This shouldn’t be overly complicated to add, as you have uncertainty assessments on measurements you can apply, and you could create confidence intervals on fluxes in a number of different statistical ways leveraging the multiple inverse model methods.
Consider/address OH and lifetimes in both ratios of pollution magnitudes (e.g. Fig. 3) and slopes of pollution ratios (e.g. Fig B1, B2). As discussed elsewhere in the paper, and included in the inverse modeling, the lifetime of some of these gases for some times of year is sufficiently short to potentially significantly impact observe ratios/slopes, which could strongly impact these values and interpretations. For example, is Gosan lower because it is sampling after greater chemical loss? Please consider doing addition calculations/corrections to account for OH loss and/or consider sensitivity studies.
Regarding the slope analysis, this is not a case where OLS regression is appropriate (Fig. B1,B2). A Type II model regression should be conducted as variance (uncertainty) on both axes are large and comparable (if the y-axes variable variance was 3x> than x, than OLS would be ok, but here that isn’t the case). Further justification would also be needed for forcing the intercept thru 0. I don’t think updating the regression approach will impact the conclusion, but would improve the analysis. Also please include uncertainty/significance/CI for the slopes as well.
For the gridded flux output (such as in Fig. 7), please also provide a spatial figure that shows where the inverse modeling work has sensitivity/constraint. This could be a footprint type figure with contours, or a uncertainty reduction type figure, or Fig. 7 could have regions not constrained left in white or gray. It is important to show where observations are constraining emissions and where not (to prevent false impressions that say the Iberian Peninsula has no emissions when it is simply not observed by this network).
Detailed comments:
Lines 205-208: It is unclear how this release height adjustment was determined to be appropriate – please explain further/justify.
Lines 211-212: Please at some point discuss the possible impact of this simplified chemical loss approach.
Line 340: I’m not sure you can conclude this means delayed replacement. Could you simple be further away in a warmer wetter environment where HFO’s are thus not seen at the site as well? Could you have growth in both HFOs and HFCs here?
Line 346-7: I’m not sure this has been established, particular since chemistry was not accounted for in this analysis.
Fig 3.: Perhaps also do this considering decay of the substances during transport.
Citation: https://doi.org/10.5194/egusphere-2025-4824-RC2 - AC3: 'Reply on RC2', Martin Vollmer, 06 Mar 2026
- AC1: 'Comment on egusphere-2025-4824', Martin Vollmer, 06 Mar 2026
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-4824', Isaac Vimont, 24 Nov 2025
General Comments
“Global Observations and European emissions of the halogenated olefins HFO-1234yf, HFO-1234ze(E), and HCFO-1233zd(E) from the AGAGE (Advanced Global Atmospheric Gases Experiment) network” is an overview of measurements from the AGAGE network of 3 key haloolefins that are increasingly used in foam blowing and refrigeration, heating, and air-conditioning. The authors provide a nice overview of the global measurements, and provide more detailed analysis and emissions estimates for north western Europe. Additionally this paper easily meets every requirement set forth by ACP for publication in their journal.
This paper is excellent, and I have no major comments for the authors to address. I would like to thank the authors for providing such a well-written and thorough manuscript, it was a pleasure to read and review. This is a welcome change from many of the recent papers on halocarbons that I am asked to review. I further appreciate that the authors have taken the time to completely describe their calibration and uncertainty estimates (save for one small detail that I mention below). This is so often overlooked in manuscripts these days, and I found it a welcome change.
I did not find any issue with any of the modeling methods or results, however, the details of the models are outside my area of expertise and therefore I am not properly qualified to provide in-depth review of that portion of the text. It is a shame that these compounds are too short-lived for estimates of global emissions using traditional box modeling methods, as it would be nice to have global estimations of these species going forward as they are used increasingly throughout the world. However, an analysis of global emissions using more complex, full chemistry global models is outside the scope of this paper (indeed, likely a paper on its own).
In my opinion, the paper is nearly ready to publish in ACP. I have included a small number of minor, admittedly nit-picky, comments below, easily addressed, but overall, the paper is excellent and I look forward to its imminent publication.
Specific Comments
Line 35: This part of the sentence (...currently an intense debate…) is ambiguous (is the debate about banning PFA’s or about adding HFO’s to the ban?). To me, it doesn’t add much to the paper nor the point being made in the paragraph. Suggest:
“Haloolefins are also within the scope of the definition of the very stable anthropogenic per- and polyfluoroalkyl substances (PFAS). In January 2023, authorities from Denmark, Germany, the Netherlands, Norway, and Sweden submitted a REACH (Registration, Evaluation, Authorisation and Restriction of Chemicals) dossier for a restriction proposal for PFAS in the EU to the European Chemicals Agency (ECHA, European Chemicals Agency, 2024) which suggests a wide-ranging ban of PFAS from usage in many applications. At the time of publication, this dossier is still being considered by the ECHA.”
Line 51: This paragraph seems out of place. We are 4 paragraphs into the introduction and only now are the three species listed and their uses explained. I suggest moving this up to be the second paragraph (currently the second paragraph starts at line 30). This would help clarify several points being made in the current second and third paragraphs.
Line 56: “...and also difficult to reconstruct the temporal changes in regulations and applications in various parts of the world”. To me, this sentence is unclear. What regulations do you refer to here? Above you have stated that haloolefins are up for consideration for regulation in the EU and that the MP does not restrict their use at this time. Or is this meant to say that it will be difficult in the future?
Section 2.2: Monte Cimone has 5 years of data with a non-Medusa system, yet there is no description of any difference in measurement precisions for the three compounds, or detectability. From what I can see in the data presented here, there seems to be no change in the measurement quality, nor the detection limits, when the Monte Cimone site was switched to the Medusa system, which is excellent.
However, though this system is described in Maione et al., 2013), I suggest the authors provide a little bit more information on the comparability of these two instruments. Given the ~14 months of overlap, this may be easy, and can be as little as to say that the measurement quality is comparable (assuming that it is), or a short mention in Section 2.4 if there is any small difference.
Section 2.4: What are the uncertainties for clean air samples? Unless misread here, I see an excellent detailing of polluted sample uncertainty, but no mention of unpolluted background air uncertainties? I would expect that the uncertainty will increase somewhat as the chromatographic peak approaches the detection limit.
Line 338: “...large magnitudes of pollution…” is awkward. Suggest changing to “...large enhancements relative to background…”
Sections 3.1.1 and 3.1.2: I have some minor skepticism of the conclusions drawn in these two sections relating to the slower adoption of HFO’s in the region of influence for the Gosan station relative to the European stations, particularly with respect to 134a. The Gosan Station has significant influence from both the DPRK, the Republic of Korea, and Eastern China. Both the Korean and Chinese automobile industries have exploded in the last 10 years, with Korea’s beginning even before this. Though my experience with specific vehicles is quite limited, both modern Chinese and Korean vehicles are coming with haloolefin refrigerants, and have been for some time. Despite the DPRK, China, and the Republic of Korea being A5 countries and thus having longer phase down times, both the Korean and Chinese automobile manufacturers are targeting non-A5 countries for sales, including the US, Europe, and Australia. They would be prevented from selling vehicles in these regions with 134a, and from a manufacturing perspective, it is difficult to see how they would tool their manufacturing plants differently depending on whether the vehicle was to be sold domestically or internationally.
If we assume the conclusions drawn here are true, it would be good to discuss why the vehicle fleets in China and Korea would be expected to be largely running 134a despite their new vehicles being produced with HFO-1234yf. Unlike the US, Europe, Australia, etc… where the vehicle fleets might be expected to contain significant numbers of older vehicles (though from this paper it is clear that Europe's fleet is changing over), China’s fleet is likely quite a bit newer, as evidenced by their historically low, but rapidly increasing, vehicle ownership statistics (e.g. https://doi.org/10.1007/s11356-024-34344-0). I think the authors could expand a bit on this, as while their conclusions are certainly plausible, to me they have not addressed this additional factor which could run counter to their hypothesis.
Citation: https://doi.org/10.5194/egusphere-2025-4824-RC1 - AC2: 'Reply on RC1', Martin Vollmer, 06 Mar 2026
-
RC2: 'Comment on egusphere-2025-4824', Anonymous Referee #2, 07 Jan 2026
Review of “Global Observations and European emissions of the halogenated oelfins HFO-1234yf, HFO-1234ze(E), and HCFO-1234zd(E) from the AGAGE network”
Overview: This paper reports on global observations of two HFOs and one HCFO that are being increasingly used as replacements for HFCs. Further, the authors use multiple inverse modeling frameworks to estimate emissions and emissions trends in NW Europe. Authors report increasing prevalence of detectable levels of these gases since 2011, with increasing levels observed in pollution plumes (normalized to a HFC observation). Inverse modeling suggests increasing emissions for these gases in NW Europe, with somewhat rapid increases inferred in the time series. Haloolefins are being increasingly used worldwide, but emissions magnitudes and environmental and health impacts are poorly known. Given the reactive nature of these gases, they are harder to observe and track with sparse global networks. This paper nicely present important measurements of these gases and infers reasonable conclusions. The multi-method inverse assessment of Europe is a strong addition to the paper, and illustrates the need to consider more measurements near emissions regions to truly characterize HFO’s moving forwards. The paper is appropriate for ACP. I have some modest concerns and suggestions for the authors, which should be straightforward to address and would strengthen the manuscript.
General comments:
Add uncertainties: In reporting concentrations, enhancements, and flux numbers, please add uncertainties. For example, in the abstract when stating the increase in ppt’s, or when stating estimated emissions, confidence intervals or uncertainties should be added. This shouldn’t be overly complicated to add, as you have uncertainty assessments on measurements you can apply, and you could create confidence intervals on fluxes in a number of different statistical ways leveraging the multiple inverse model methods.
Consider/address OH and lifetimes in both ratios of pollution magnitudes (e.g. Fig. 3) and slopes of pollution ratios (e.g. Fig B1, B2). As discussed elsewhere in the paper, and included in the inverse modeling, the lifetime of some of these gases for some times of year is sufficiently short to potentially significantly impact observe ratios/slopes, which could strongly impact these values and interpretations. For example, is Gosan lower because it is sampling after greater chemical loss? Please consider doing addition calculations/corrections to account for OH loss and/or consider sensitivity studies.
Regarding the slope analysis, this is not a case where OLS regression is appropriate (Fig. B1,B2). A Type II model regression should be conducted as variance (uncertainty) on both axes are large and comparable (if the y-axes variable variance was 3x> than x, than OLS would be ok, but here that isn’t the case). Further justification would also be needed for forcing the intercept thru 0. I don’t think updating the regression approach will impact the conclusion, but would improve the analysis. Also please include uncertainty/significance/CI for the slopes as well.
For the gridded flux output (such as in Fig. 7), please also provide a spatial figure that shows where the inverse modeling work has sensitivity/constraint. This could be a footprint type figure with contours, or a uncertainty reduction type figure, or Fig. 7 could have regions not constrained left in white or gray. It is important to show where observations are constraining emissions and where not (to prevent false impressions that say the Iberian Peninsula has no emissions when it is simply not observed by this network).
Detailed comments:
Lines 205-208: It is unclear how this release height adjustment was determined to be appropriate – please explain further/justify.
Lines 211-212: Please at some point discuss the possible impact of this simplified chemical loss approach.
Line 340: I’m not sure you can conclude this means delayed replacement. Could you simple be further away in a warmer wetter environment where HFO’s are thus not seen at the site as well? Could you have growth in both HFOs and HFCs here?
Line 346-7: I’m not sure this has been established, particular since chemistry was not accounted for in this analysis.
Fig 3.: Perhaps also do this considering decay of the substances during transport.
Citation: https://doi.org/10.5194/egusphere-2025-4824-RC2 - AC3: 'Reply on RC2', Martin Vollmer, 06 Mar 2026
- AC1: 'Comment on egusphere-2025-4824', Martin Vollmer, 06 Mar 2026
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 3,359 | 2,778 | 265 | 6,402 | 276 | 160 | 177 |
- HTML: 3,359
- PDF: 2,778
- XML: 265
- Total: 6,402
- Supplement: 276
- BibTeX: 160
- EndNote: 177
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Joseph R. Pitt
Dickon Young
Stephan Henne
Blagoj Mitrevski
Jens Mühle
Anita Ganesan
Jgor Arduini
Alistair J. Manning
Thomas Wagenhäuser
Alison L. Redington
Brendan Murphy
Ray Gluckmann
Kieran M. Stanley
Paul B. Krummel
Chris R. Lunder
Jaegeun Yun
Dominique Rust
Angelina Wenger
Myriam Guillevic
Jooil Kim
Ray H. J. Wang
Tae Siek Rhee
Lionel Constantin
Arnoud Frumau
Christina M. Harth
Peter K. Salameh
Ove Hermansen
Andreas Engel
Simon O'Doherty
Sunyoung Park
Michela Maione
Paul J. Fraser
Ronald G. Prinn
Ray F. Weiss
Stefan Reimann
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(17400 KB) - Metadata XML
-
Supplement
(12 KB) - BibTeX
- EndNote
- Final revised paper
General Comments
“Global Observations and European emissions of the halogenated olefins HFO-1234yf, HFO-1234ze(E), and HCFO-1233zd(E) from the AGAGE (Advanced Global Atmospheric Gases Experiment) network” is an overview of measurements from the AGAGE network of 3 key haloolefins that are increasingly used in foam blowing and refrigeration, heating, and air-conditioning. The authors provide a nice overview of the global measurements, and provide more detailed analysis and emissions estimates for north western Europe. Additionally this paper easily meets every requirement set forth by ACP for publication in their journal.
This paper is excellent, and I have no major comments for the authors to address. I would like to thank the authors for providing such a well-written and thorough manuscript, it was a pleasure to read and review. This is a welcome change from many of the recent papers on halocarbons that I am asked to review. I further appreciate that the authors have taken the time to completely describe their calibration and uncertainty estimates (save for one small detail that I mention below). This is so often overlooked in manuscripts these days, and I found it a welcome change.
I did not find any issue with any of the modeling methods or results, however, the details of the models are outside my area of expertise and therefore I am not properly qualified to provide in-depth review of that portion of the text. It is a shame that these compounds are too short-lived for estimates of global emissions using traditional box modeling methods, as it would be nice to have global estimations of these species going forward as they are used increasingly throughout the world. However, an analysis of global emissions using more complex, full chemistry global models is outside the scope of this paper (indeed, likely a paper on its own).
In my opinion, the paper is nearly ready to publish in ACP. I have included a small number of minor, admittedly nit-picky, comments below, easily addressed, but overall, the paper is excellent and I look forward to its imminent publication.
Specific Comments
Line 35: This part of the sentence (...currently an intense debate…) is ambiguous (is the debate about banning PFA’s or about adding HFO’s to the ban?). To me, it doesn’t add much to the paper nor the point being made in the paragraph. Suggest:
“Haloolefins are also within the scope of the definition of the very stable anthropogenic per- and polyfluoroalkyl substances (PFAS). In January 2023, authorities from Denmark, Germany, the Netherlands, Norway, and Sweden submitted a REACH (Registration, Evaluation, Authorisation and Restriction of Chemicals) dossier for a restriction proposal for PFAS in the EU to the European Chemicals Agency (ECHA, European Chemicals Agency, 2024) which suggests a wide-ranging ban of PFAS from usage in many applications. At the time of publication, this dossier is still being considered by the ECHA.”
Line 51: This paragraph seems out of place. We are 4 paragraphs into the introduction and only now are the three species listed and their uses explained. I suggest moving this up to be the second paragraph (currently the second paragraph starts at line 30). This would help clarify several points being made in the current second and third paragraphs.
Line 56: “...and also difficult to reconstruct the temporal changes in regulations and applications in various parts of the world”. To me, this sentence is unclear. What regulations do you refer to here? Above you have stated that haloolefins are up for consideration for regulation in the EU and that the MP does not restrict their use at this time. Or is this meant to say that it will be difficult in the future?
Section 2.2: Monte Cimone has 5 years of data with a non-Medusa system, yet there is no description of any difference in measurement precisions for the three compounds, or detectability. From what I can see in the data presented here, there seems to be no change in the measurement quality, nor the detection limits, when the Monte Cimone site was switched to the Medusa system, which is excellent.
However, though this system is described in Maione et al., 2013), I suggest the authors provide a little bit more information on the comparability of these two instruments. Given the ~14 months of overlap, this may be easy, and can be as little as to say that the measurement quality is comparable (assuming that it is), or a short mention in Section 2.4 if there is any small difference.
Section 2.4: What are the uncertainties for clean air samples? Unless misread here, I see an excellent detailing of polluted sample uncertainty, but no mention of unpolluted background air uncertainties? I would expect that the uncertainty will increase somewhat as the chromatographic peak approaches the detection limit.
Line 338: “...large magnitudes of pollution…” is awkward. Suggest changing to “...large enhancements relative to background…”
Sections 3.1.1 and 3.1.2: I have some minor skepticism of the conclusions drawn in these two sections relating to the slower adoption of HFO’s in the region of influence for the Gosan station relative to the European stations, particularly with respect to 134a. The Gosan Station has significant influence from both the DPRK, the Republic of Korea, and Eastern China. Both the Korean and Chinese automobile industries have exploded in the last 10 years, with Korea’s beginning even before this. Though my experience with specific vehicles is quite limited, both modern Chinese and Korean vehicles are coming with haloolefin refrigerants, and have been for some time. Despite the DPRK, China, and the Republic of Korea being A5 countries and thus having longer phase down times, both the Korean and Chinese automobile manufacturers are targeting non-A5 countries for sales, including the US, Europe, and Australia. They would be prevented from selling vehicles in these regions with 134a, and from a manufacturing perspective, it is difficult to see how they would tool their manufacturing plants differently depending on whether the vehicle was to be sold domestically or internationally.
If we assume the conclusions drawn here are true, it would be good to discuss why the vehicle fleets in China and Korea would be expected to be largely running 134a despite their new vehicles being produced with HFO-1234yf. Unlike the US, Europe, Australia, etc… where the vehicle fleets might be expected to contain significant numbers of older vehicles (though from this paper it is clear that Europe's fleet is changing over), China’s fleet is likely quite a bit newer, as evidenced by their historically low, but rapidly increasing, vehicle ownership statistics (e.g. https://doi.org/10.1007/s11356-024-34344-0). I think the authors could expand a bit on this, as while their conclusions are certainly plausible, to me they have not addressed this additional factor which could run counter to their hypothesis.