This paper is well written with thorough citations, documenting all the before-and-after details of the model development, explaining all the methods used for data analysis especially trying to have apple-to-apple comparison with satellite data.

However, the results part still need some attention. The current SAD analysis is not convincing; there’s missing necessary plots and analysis for Arctic and Water vapor. Once these two issures are address, more discussion need to be added in the summary session. I provide several major concerns here for your consideration.

1. Line 317-318: This statement is not correct. NATmix is always the dominant composition category in the CALIPSO Antarctic data (Pitts 2008). From Figure 3, it's hard to know if your new simulation of PSC categories are consistent with what CALIPSO shows for STS and NAT. CALIPSO shows NATmix usually dominant the mid-winter, while the STS dominant the early winter. you should show three panels of absolute values of STS, NAT and ice from RUN_HH. You ice might be ok, maybe a little bit too long into September. Your RUN_HH is not showing early winter SAD (later may to June), which should be dominant by STS. Also why does you simulated PSCs so low in altitude in RUN_HH? The CONTROL has better altitude representation (match the temperature threshold better and more consistent with CALIPSO cloud area). Is it because your background HNO3 is too low?  What about your background water vapor? From your SAD plot, I am not convinced your new PSC scheme is better than before.

2. Section 3 misses dehydration comparison with MLS. Also, what about Arctic denitrification and dehydration?

3. Figure 1 and Figure 2: The color bars between the model and CALIPSO are completely different. It's hard to compare. You should not use log scale color bar. If your concern is your total counts are different, at least, you should interpolate your color bar proportional to CALISPO color bar. For example, the cyan color is 700 for CALIPSO, you don't have a cyan color but maybe it is like 20 in your color bar. That's a huge difference.

Other smaller points:

Line 26: There's no such a thing called Tsts. STS grows as the temperature decrease, not suddenly at a certain temperature. You could say 191 K is a typical temperature that HNO3 condenses rapidly in a typcial stratospheric chemical conditions (i.e. HNO3 ~ 10 ppb, H2O~ 5ppm and 50 hpa), but HNO3 still condenses slowly on the particle at higher temperatures based on vapor pressure.

Line 33: Are you talking about H2SO4-H2O sulfate binary? or also other binary?

Line 76: atmospheric variables -> atmospheric gas variables

Figure 1: Zhu et al. 2017a talked about how to adding noise to the model run to have a closer comparison with the CALIPSO data. Adding noise will spread the model results to more categories. You may try that.

The paper titled “Assessment of an updated polar stratospheric cloud parameterisation for the UK Earth System Model (UKESM1.1) within the UK Met Office Unified Model (v13.9) using CALIOP and MLS observations” by Isabelle Sangha and co-authors presents a comprehensive and well-executed update of the polar stratospheric cloud (PSC) parameterisation in the UK Earth System Model (UKESM1.1). By incorporating a more physically based representation of PSC formation processes, the authors enhance the model’s ability to simulate PSCs. Overall, the manuscript is very well written, logically structured, and carefully researched, with a clear and coherent presentation of methods and results. However, I see some weaknesses in the presentation of the results. Here are my comments:

Line 65 ff: PSCs can also be observed with ballon-borne measurements.

Line 130 ff: Peter et al. (1991) is a pretty old publication. Since 1991, much has happened in PSC research and Thomas Peter would not agree anymore to be cited with the statement of “homogeneous NAT nucleation”. According to Lowe and MacKenzie (2008), “it appears that, from current knowledge, homogeneous nucleation of NAT/NAD from STS is not significant in the stratosphere, although there is not universal consensus on this subject.” Already Koop et al. (1995) showed for bulk solutions that liquid stratospheric aerosol droplets do not freeze under polar winter conditions for temperatures higher than the water ice saturation temperature. Instead, it was discussed in recent years that NAT particles nucleate heterogeneously on foreign nuclei (e.g. Hoyle et al., 2013). For your model simulations, it is not important if NAT nucleates homogeneously or heterogeneously. However, you should discuss this in slightly more detail and cite up-to-date literature here.

Figure 1 and 2: It is hard to compare the model results to the observations because of the different color bars.

Figure 3: Why are you not comparing your results to CALIOP surface area densities? This is a product from Pitts et al.

Figure 4 and 6: The different colors are difficult to see. They lie on top of each other. I have difficulty distinguishing the shades of blue and green. Could you use dashed and dotted lines, for example? 

Generally, I have the impression that the comparison with CALIOP could be much more detailed.

Line 329 ff: The model's underestimation of HNO3 is extreme. Should this not be solved first? It could bias the entire analysis.

What does H2O look like? There are no plots referring to H2O!

Line 406: “high-number-ddensity”

In summary, I have the impression that the evaluation against CALIOP satellite observations and Aura MLS measurements is not yet fully convincing. As it stands, the plots tend to give the impression that the agreement between model and observations is weaker than intended. The plots should be improved to strengthen the overall conclusions.

This paper compares new polar stratospheric cloud (psc) models to the standard model (control) used in the UK Earth system model. Assessments of improvements to the representation of pscs and impacts on nitric acid and ozone are made through comparisons with CALIOP and MLS observations. The paper is mostly ready for publication, but there are a few serious issues that need to be addressed before that.

Here by line, table, figure number are comments the authors should consider as they prepare a revision.

30-35 To what extent is the difference between NAT and STS/sulfate aerosol activation chemistry due to the chemistry, or the surface area available? Generally STS/sulfate aerosol are significantly smaller and more numerous than NAT thereby usually providing significantly more surface area.

110 The sedimentation velocity is fixed for all pressure levels? This is clearly wrong but that’s what the text indicates. Some explanation should be added.

155 Isn’t this the effective radius?

Table 2 is confusing. A is set to orders of magnitude. But then the maximum number densities are given to three significant figures: once in the table description, and then a different set of three significant figure number densities, times some value of A, in the table. In the table the n_NAT,lim,b exceed the total number density for bins 1-6. The point of this table is not at all clear, and is certainly not going to be understood by any reader outside of the authors’ modeling group.

Table 3. Now in the simulations, the number densities are 2 significant figures which still seems a bit exact, but okay, and they all correspond to the numbers in the title of Table 2, rather than the numbers in Table 2.

Figure 1 caption, line 4 should be … indicate the beta_perpendicular and R_532 thresholds.

Figures 1 and 2 have serious problems. For ease of viewing the panel size of the model results, control and all the runs should match the CALIOP observation panels. This can be done by reducing the size of the model figures or moving the color bar explanation to the top or bottom of the CALIOP figures and increase their size to match the model results.

The more serious problem is that the boundaries established by Pitts et al. are not reproduced in the model figures. In the Pitts et al. Antarctic observations, minimums for ice (beta_perp, R_532) are: (2.3e-6, 2.8). In the model panels these minimums are (3e-6, 3.4). Further the NAT mixture boundary from Pitts is beta_perp = ~2.4e-5, while in the model panels this line is at 2e-5. In addition the color scales for the number of counts do not match. In Pitts et al the color scale (red/orange) peaks at 2810. In the model panels the same color scale is at 1000, which for the Pitts et al. panel is in the transition from cyan to green. These differences make it difficult to compare the model panels with the observation panel, yet isn’t that the point of the figure? If it is then a better job needs to be done to reproduce the boundaries and color scales in the model panels so they match, to the extent possible, the observations. In model, and log, space these differences are significant and confounds an easy comparison of the figures. These discrepancies are somewhat lessened in the Arctic cases, but still to the extent possible efforts should be made to make them match.

The discussion of these figures devotes just once sentence to the striking result that none of the model scenarios reproduce the large beta_perpendicular and R_532 space covered by CALIOP in the ice region. In contrast the model results occupy a very small fraction of this space in a very narrow beta_perp and R_532 space. In the Antarctic this space is only penetrated with n_ice high while in the Arctic only if n_nat is high.

308-309 on what basis is the parenthetical clause justified?

307-310, Figure 3. Now we find that in comparison of surface area to the control all the new model results are similar in surface area, independent of n_nat and n_ice? This figure also does not provide any assessment of the improved psc model. It is just a comparison of one model run with a control. Presumably there are no corresponding surface area densities available from CALIOP? If there were that would be a much more useful comparison.

Figure 4 The colors for the 4 runs need to be modified so they can be differentiated. Now they are just some different shade of blue, so the reader can only easily see the MLS observations and the control run. The rest are a guessing game and the legend is no help. Better more differentiated colors are an option, so are different lime styles.

Figure 5 For a real comparison of run HH to control wouldn’t both the solid and liquid pscs have to be added? It would be easy and useful to include such a line.

341 It is a significant stretch to suggest that run HH starts partitioning hno3 to the solid phase prior to the control. Rather the timing of onset of both, and the amounts, are far more comparable than they are different, until the control takes off. To suggest differently is to use information not accessible in Fig. 5.

Figure 6 Same criticism as Figure 5 concerning the choice of colors to represent the model runs. There are a lot of options than just different shades of blue, although in this figure there is almost no discrimination between the two model runs shown, _HH and _LL.

368 Figures 1 and 2 do not show surface area densities.

371 Do the authors mean 70-80 S? In the 60-70 S band, calling the slightly higher hno3 in the model runs, compared to the control, an improvement with MLS hno3 is a stretch.

391-414 The conclusions conveniently skip over the most glaring discrepancy between the model and CALIOP: The region tagged as ice bounded by beta_perp 2.3e-6 to 2e-3 and R_532 from 2.8 – 50. The region only appears in Antarctic runs if n_ice is high, but even then the space occupied is limited to bet_perp 4e-5 to 2e-4 and R_532 3.3 to 5.3, and only appears in the Arctic if n_nat is high. This discrepancy rates just half a sentence in the conclusions. While the rest of that paragraph, and two more (!), are devoted to wave ice, a much smaller region of the beta_perp, R_532 sample space, with only a smattering of CALIOP observations. It is not a surprise that an Earth systems model cannot reproduce wave ice. What is a surprise is that the authors choose to discuss at length such a small region of psc space. The much more important difference is the difficulty in the models producing much regular ice, a much larger region of the sample space, with numerous CALIOP observations, yet very little penetration by the new psc models. The current emphasis on discussing wave ice is entirely out of place.