Review of "From Sea to Sky: Understanding the sea surface temperature impact on an atmospheric blocking event with the ICON model"

This study uses idealized model sensitivity experiments to investigate the impact of Gulf Stream sea surface temperatures (SSTs) on the development of an intense atmospheric blocking event over Europe that occured during February 2019. The authors combine traditional eulerian diagnostics with analysis of lagrangian trajectories to diagnose the response of the upper-level blocked circulation to rapidly ascending air streams in extratropical cyclones (i.e. Warm Conveyor Belts) and their sensitivity to changing SST boundary conditions. The authors provide a compelling argument that, at least for this case study, Gulf Stream SSTs and associated gradients can substantially modify the development of the upper-atmosphere circulation anomalies through their impact on air-sea fluxes and the temperature and moisture of the WCB inflow regions. The manucript is clearly written and will be of interest many readers of Weather and Climate Dynamics. I have several comments that should be straightforward for the authors to address, but otherwise I think this manuscript is suitable for publication in WCD.

Main comments:

(1) One limitation of this study is the use of single-member deterministic sensitivity experiments for a single start date (00 UTC 18 February 2019) and the lack of uncertainty estimates. Ensemble forecasts (or analysis of multiple case studies/start dates) would provide additional uncertainty information that would allow the authors to better distinguish between systematic effects and chaotic variations that are consequence of the intrinisic predictability limits. However, I appreciate that not all research groups have access to significant HPC resources required to run such experiments. The authors should acknowledge these limitations and also emphasise those aspects of their results that give confidence that the diagnosed impacts are systematic. For example, the consistent differences between experiments at all lead times (e.g. figure 9b and others) and the physically plasuible (thermo)dynamical interpretation give me some confidence that the impacts in this cases study represent signal rather than noise. However, I think the authors could acknowledge these limitations and justify their approach more explicitly in the intro/discussion/conclusions.

(2) Another potential limitation, acknowledged by the authors, is the potential role for ocean coupling. I think it is justifiable to use a prescribed SST boundary condition to explore the sensitivity to different SST anomalies. However, I am not sure why the authors use a constant (i.e. persisted) SST throughout the 10-day forecasts rather than a time-varying boundary condition. I would like the authors to comment on whether they expect either time-varying SST boundary condition or coupled ocean feedbacks might amplify or damp the diagnosed impacts? For instance, how would ocean feedbacks impact the estimated surface turbulent heat fluxes that are important for modifying the properties of the marine atmospheric boundary layer than feed WCBs? Perhaps one possibility is to compare the fluxes from ICON sensitivity experiments with those from a coupled NWP forecast of the same case (e.g. from operational ECMWF forecasts, which have been coupled in ENS/HRES since 2018).

(3) It is not clear to me why "IDEA" is chosen as a reference experiment for comparisons with other sensitivity experiments rather than "CNTRL". I think the comparison of "IDEA" and "CNTRL" is scientifically interesting as it gives some indication of the sensitivity of the atmospheric circulation to the presence of mesoscale ocean eddies. The relative magnitude of the difference between "IDEA" and "CNTRL" is also useful context for the other comparisons, which could be a useful benchmark when considering the significance of results as described in (1).

Minor comments:

Figure 1 - the differet SST gradient contours are difficult to distinguish in panel (a).

Table 1 - Please state the original source of the SST used in the IFS analysis. SST is not part of the IFS 4DVar so it is taken from another high-resolution satellite product (I think OSTIA).

Figure 2 - It is very difficult to distinguish trajectories coloured by pressure and PV contours. Perhap separate into separate panels for PV and trajectories?

Line 272 - "a decrease in surface heat fluxes" -> "reduced magnitude (i.e. less negative) heat fluxes"?

Line 292 - Is the WCB inflow region fixed across all lead times or time varying?

Figure 4 - I suggest removing "of vertical profiles" from the caption to read "Evolution of air temperature differences [...] spatially averaged over the fixed Eulerian WCB inflow region". I initially interpreted this as the averages along the lagrangian trajectories. 

Line 380-382 - The authors describe the difference in WCB trajectory numbers between IDEA and CNTRL as "relatively small" but it is non-zero and the discussion could be more quantitative. From Table 3 it seems that removing the small scale eddies reduces WCBs by 4%, which is not negligible compared to the 9% reduction with WEAK. The authors could comment more about the potential impact of small-scale ocean eddies on the upper-level circulation, which has been suggested in other publications. e.g. "Ocean fronts and eddies force atmospheric rivers and heavy precipitation in western North America" - https://www.nature.com/articles/s41467-021-21504-w.

“From Sea to Sky: Understanding the sea surface temperature impact on an atmospheric blocking event using sensitivity experiments with the ICOsahedral Nonhydrostatic (ICON) model”

By S. Christ et al.

Submitted to EGUsphere

The author presents work on using the ICON model for sensitivity experiments on the impact of SST fronts, namely how the smoothing of these fronts (as well as an increase in SSTs) will impact atmospheric blocking events, extratropical cyclone development (downstream and upstream), as well as general air-sea interactions. The research presented here is extremely well thought out, dense, well executed, and with significant literature for their background, as well as comparing their results to previously published papers. I believe the science is quite sound, the results are fascinating, and it is worthy of acceptance and publication.

My recommendations are mainly technical and with the figures; I believe they will help future readers better understand the work these authors present.

Figure 1: Flip Figures 1d (P1.5K) and 1e (extWEAK). In the rest of the paper, P1.5K is referred to after the WEAK and extWEAK cases. In order to be consistent, I think it is best to list them in the same order in this figure.

Line 172-177 (P1.5K description): Is this case study adding the 1.5K SST increase to the IDEA or CNTRL case? Or a mixture of the two? It seems similar to IDEA, but it wasn’t clear which experiment this temperature increase was being applied to, or if this should be treated completely separately from the rest.

Figure 2:

• Both on screen and printed out, it is difficult to see the trajectories in Figure 2c,f,i as some of the colors for the trajectories and background PV overlap or their hues are close. I do agree it’s important to have these in the same figure, but I would suggest colormaps that do not overlap so details are not missed.
• Add large headers at the top of each column/left of each row (i.e., SST for Column 1, CY1 for Row 1, and so on) and only keep the date/time above each individual figure.
• Double check the dates and times listed in the figure description (a,c,e), as this does not appear to line up with the panels.

Figure 3, 6, 7, 9, 10, 11, 13: The light green line for extWEAK can sometimes be difficult to see (especially in the blue shading to indicate CY1 and CY2). I would suggest a different color or line type (i.e. dashed/dotted) so that it is easier to view.

Figures 4 and 5: While the light blue shading for CY1 and CY2 works in other figures, it clashes with the colormap here. I would suggest a different color or hash marks to indicate when CY1/2 occurs.

Figure 8:

• The letters indicating each figure are hard to see. I would increase the size and reposition them
• The description for Figures 8c and 8d appears to be flipped
• The shadings in Figure 8c are too similar to each other. I would pick a different color for liquid or ice.

Figures 10 and 11: The overlapping shading for the mean standard deviation on every single line is distracting and confusing. Remove the shading and have only lines indicating the ±σ.