Comments on: Evidence of a recent decline in UK emissions of HFCs determined by the InTEM inverse model and atmospheric measurements, Manning et al., 2021, https://doi.org/10.5194/acp-2021-261

This paper describes results from an inverse model study of HFC emissions from the UK using the Numerical Atmospheric dispersion Modelling Environment (NAME) model incorporated into the Inversion Technique for Emissions Modelling (InTEM) model system. This study uses data from long-term monitoring sites at Mace Head, Ireland; Jungfraujoch, Switzerland; and Monte Cimone, Italy.  In addition, the study incorporates recent measurements from Tacolneston (UK), Carnsore Point (Ireland), and Taunus (Germany) to better constrain UK emissions and provide higher spatial resolution. The authors found that total UK emissions of the HFCs studied were lower in 2019-2020 (expressed in CO2-equivalent units) than in the previous decade.  They also found that UK emissions based on top-down methods are about 30-50% lower than bottom-up, inventory-based emissions for several HFCs.

This work and the underlying methods provide important constraints and feedback on policy decisions related to reducing the consumption of HFCs in order to limit the climate impacts of these high-global-warming potential gases. This work should also help inform and possibly improve bottom-up inventory estimates of these gases.

General Comments:

The paper is well-written and the methods are well-established. I have only a few comments related to the work and presentation.

1. On page 4, line 94, you say that you are applying an updated version of the InTEM model framework. Since models are updated occasionally, it would help the reader to know if the version of InTEM used in this study is the same as that described in Arnold et al (2018). If updates have been made since Arnold et al (2018), perhaps these could be summarized somewhere.
2. The method of running the inversion in 1-yr and 2-yr blocks could use a bit more explanation. Why is this done?  Is this new or typical?  
3. On page 7, line 159, you discuss the use of a uniform prior emission field for HFC-23, and then later explain that you use 100% uncertainty on UK emissions in the prior for other gases. I would assume that a uniform (flat) prior would also lead to similar results for other gases (little dependence on the prior). Perhaps you can comment if that is indeed the case.
4. In section 2.3 you refer to both “background” and “baseline” mole fractions. Are these the same thing?  Or is “background” a model term and “baseline” a measurement term?  It is a minor point, but these terms seem to be used interchangeably.  And you also refer to background mole fractions as “the prior” (pg. 6, line 145).  Not to be confused with the emissions “prior”.  Is there a better way to distinguish between these different things?

Specific Comments:

P 8, L194:  “Time-varying background levels of mole fractions are required as prior information for InTEM for three stations: MHD,  JFJ and CMN.”

This implies that background mole fractions are not needed for the other stations (TAC, CSP, TOB). Is that the message you intended? 

P 17, L280:  “ …however the growth in the Northern Hemisphere appears to have peaked in 2017 …”

Given that the background mole fraction HFC-134a shows interannual variability, one could have also said the same thing in ~2006 and ~2013.  So perhaps it is too soon to say?  Still, I like figure 4 because it provides global context.

Sec. 3.2, HFC-134a:  Perhaps a comment on COVID-19 is also applicable here.  Is there any relation between HFC-134a emissions and vehicle miles driven?  Could the lock-downs have played a role in 2020?  Perhaps not, since 2019 emissions were also lower.  It might also be relevant to mention when the EU began the transition to HFO alternatives to HFC-134a in mobile air conditioning.

Figure 6:  The emission source maps for most HFCs shown seem closely related to population density, except for HFC-227ea, which shows a region of emissions in NW England that does not correspond to a large population center.  And the correlation coefficients in Table S2 are smaller for some observing sites.  Is this worth a comment, especially with respect to results from different years?  Is that region of emissions in NW England an anomaly in 2019-2020?

Technical Corrections:

P, L15:  remove “in the” after “decline in UK emissions”.     “… decline in UK emissions in the since 2018”.

Comments on: Evidence of a recent decline in UK emissions of HFCs determined by the InTEM inverse model and atmospheric measurements, Manning et al., 2021, https://doi.org/10.5194/acp-2021-261

The manuscript assesses the emissions rate of HFCs from the UK through a top down approach. The inversion method is based on the Numerical Atmospheric dispersion Modelling Environment (NAME) model in conjunction with measurement observations  and the Inversion Technique for Emissions Modelling (InTEM) model system. The inverse modelling results point out a decreasing trend of the aggregated HFCs emissions since 2018, in response to the UK implementation of European Union regulation of those gasses emissions. Moreover, the aggregated HFCs emissions calculated are significantly lower (73%) than those reported from the UK to National Greenhouse Gas Inventories (GHGI) and submitted annually to the United Nations Framework Convention on Climate Change (UNFCCC).

General comments

This work is well written, and it is very much appreciated.

Few general comments  :

1. On page 4 line 98, Does the Carnsore Point, Taunus and Tacolneston measurement data have different calibration scales compared to Mace Head and the two mountain stations ? If the authors provided the intercalibration between those scales it could be interesting to show it on Table 2.  
2. It could be useful add more information about the Carnsore Point, Taunus and Tacolneston measurement stations (e.g. sub urban, rural or remote).
3. On page 4 line 98, “which greatly enhances the ability…” As you do not show any test that proves it, I would suggest rephrasing it or report some specific test (on Supplementary Material S.M.) that proves how those stations improved the inversion performance.
4. As the Taunus station is located in the center of Germany, and it has only 1 weekly measurement, does it really greatly improve the ability of the inversion system for the UK?
5. Are the two mountain stations affected by HFCs UK sources ? If not, why did you use them on your inversion system ?
6. Did you run the inversions for the central Europe (Fig 1) or only for UK and IE (fig 6)  ?  I think showing (in fig 1) the European domain as the inversion model domain, but then reporting emission values only for the UK and Ireland (Fig 6) could generate confusion. Unless, you better explain this approach
7. On page 5 line 135 , If I well understand the system, using the passive tracer over 30 days of the simulations, it should imply an underprediction of ~5% of the HFC-152a estimates. If so, you shuld mention it.
8. On pag 6 line 140, I agree with referee 1, could you please describe the updates on InTem system?
9. On page 9 line 204. “when the population under the surface footprint is small”, Did you use the same threshold value used for MHD ?

Specific comments:

1. On page 4 line 103, I think you could introduce the Site Acronym, and then use them from this point on.
2. On pag 12 paragraph 3.1 I think in this paragraph you should report more quantitative analysis of the main important trends described, instead of qualitative indication .
3.    On page 12 paragraph 3.1 I think it could be useful to show (on S.M.) the plot of the three baseline data and also the MHD baseline data overlapping the observed time-series of TAC, CSP and TOB.
4. On page 18 315 “increase rapidly” please report the value of it.
5. On page 18 334 “accelerating rate” please report the value of acceleration or trend.
6. On page 19 3.6 HFC-227ea “It is also likely to be less tied to population distribution because of its specialist uses” In this case why did you not use a uniform land-based prior as for HFC-23?
7. On page 19 348 “HFC-227ea decline markedly from 2018 to 2020” please report the value to indicate it
8. I agree with referee 1, a few times the authors use “background” but it seems they rather refer to “baseline”. See below the difference and the definition of both words: “Baseline concentrations refer to observations made at a site when it is not influenced by recent, locally emitted or produced anthropogenic pollution. The term global or hemispheric background concentration is a model construct that estimates the atmospheric concentration of a pollutant due to natural sources only.” References: HTAP, T., 2010. Hemispheric Transport of Air Pollution 2010 Part A: Ozone And Particulate Matter, Air Pollution Studies No. 17. Cooper, O. R., Parrish, D. D., Ziemke, J., Cupeiro, M., Galbally, I. E., Gilge, S., ... & Oltmans, S. J. (2014).
9. On page 2 Table 2 of Supplementary Material. Do the 1 yr inversion correlation values are similar to those ? For clarity, if so, just mention it on the paper or report the Table of 1 yr inversion correlation values.
10. On page 6 line 19 of S. M. . “ a significant decline “ please report a value of                                  significant decline
11. On page 6 line 33 of S.M. “ sharp decline from 2017 to 2020” I would suggest to report the value of the sharp decline
12. On page 7 of S.M. “small rise”.. “dropping sharply” I suggest corroborating these descriptions with quantitative values.