I found this manuscript to be very well written and easy to follow. The experiments were carefully designed and the analysis was thorough and insightful. I have a few substantive comments on the analysis and discussion and several recommendations for clarification, all of which should be fairly straightforward to address. Therefore, my recommendation is for the authors to make minor revisions prior to publication.  Please see the attached pdf supplement for my complete review.

Review of The upstream-downstream mechanism of North Atlantic and Mediterranean cyclones in semi-idealized simulations by Alexander Scherrmann, Heini Wernli, and Emmanouil Flaounas

The paper investigates the dynamical linkage between North Atlantic and Mediterranean cyclones, which is known to involve (upstream) warm conveyor belts over the Atlantic that amplify (downstream) Rossby wave breaking into potential vorticity streamers over the Mediterranean. The results are based on semi-idealized simulations for the Euro-Atlantic region with the WRF model. Increasing potential vorticity along the upper-level waveguide is shown to enhance Atlantic cyclogenesis, amplify Rossby wave breaking and trigger Mediterranean cyclogenesis as expected. Sensitivity experiments show that a stronger increase in potential vorticity induces stronger Atlantic cyclogenesis but the impact on the Mediterranean is more complex, while the response also depends on the season and is weak in summer.

The numerical framework based on semi-idealized simulations is very interesting and provides clear results that help understanding dynamical processes, which are often obscured by the complexity of the large-scale flow in real case studies. The paper is well structured and written and the results are generally clear and well supported by the numerical experiments. However, they would deserve more discussion about the role of large-scale Rossby wave dynamics and small-scale cloud microphysics in the upstream-downstream mechanism, as well as the relevance of sensitivity experiments for the different seasons.

I recommend the paper for publication after consideration of the general comments below. Specific comments are also provided to help improve the paper.

General comments

1. The paper seems to claim that Atlantic cyclones are the cause of Mediterranean cyclones. However, in the used numerical framework the initial perturbation is applied to upper-levels dynamics, which contributes to Atlantic cyclogenesis but also triggers Rossby waves that propagate further downstream, thus the link between Atlantic and Mediterranean cyclones may not be causal. Hovmöller plots of flow anomalies compared to the background climatological state may help investigating the respective contributions of Atlantic cyclogenesis and Rossby wave propagation.
2. The abstract and introduction emphasize the role of warm conveyor belts in the chain of events that leads to Mediterranean cyclogenesis but these are not discussed further in the paper (although they are implicitly accounted for via the model microphysics). Quantifying their contribution would likely need new simulations with the microphysics scheme modified or switched off, which may be too demanding, but some estimate or investigation would be appreciated.
3. The results do not really reflect the actual seasonal and geographical distribution of Mediterranean cyclones, especially the hotspot over the Gulf of Genoa. This questions how the location and intensity of PV/jet perturbations in the semi-idealized framework compare with the observed variability around the climatological composites. Again, re-running the simulations would likely be too demanding but the perturbations should at least be compared to the actual variability for each season.
4. The added value of running perturbed experiments with the SPPT scheme is unclear, because the results are not discussed much. Perhaps this would contribute answering comment 2 on the role of warm conveyor belts, as suggested in Section 2.2.
5. I recommend writing spring, summer, autumn and winter in the text instead of MAM, JJA, SON and DJF but I leave it to the authors’ preference.

Specific comments

Title: “upstream-downstream mechanism of cyclones” sounds awkward; for instance, l. 15–16 rather define the “upstream-downstream mechanism of cyclogenesis” and l. 69–70 refer to an “upstream-downstream connection of cyclones”

The first third of the abstract solely describes results from previous studies and with specific information, which appears unnecessary, while more details are expected on the actual results of the paper described in the last third of the abstract. For instance, what is the sensitivity of the Mediterranean cyclone characteristics to the dynamical structure and intensity of the intruding PV streamer? How does the seasonal cycle of Mediterranean cyclogenesis depend on the large-scale atmospheric circulation?

l. 29–30 Which are “The spatially distinct regions of cyclogenesis in the Mediterranean”?

l. 35–36 Not sure what is meant by “south of the Alps” for Argence et al. (2008): in their abstract, “This study explores the predictability of a heavy rainfall event that struck North Africa on 9 and 10 November 2001”

l. 51 A tropical cyclone is involved in this specific case

l. 112 Repetition of l. 102

l. 125 Repetition of l. 115

l. 131 Remove “both” here

l. 135 northeastward?

l. 137 upper troposphere

l. 146 x, y and z − zh should be |x-xh|, |y-yh| and |z − zh| in Equation 1

Table 1 The added value of indicating “simulations that are not explicitly discussed as they provided no additional insight” appears limited

l. 171–175 A brief description is expected here for the method of Wernli and Schwierz (2006), and for the modifications of Sprenger et al. (2017) if relevant

l. 178–194 Identifying the relevant features on Figure 3 is not obvious; labeling them would help

l. 236 The terminology should be introduced in Section 2 and included in Table 1

l. 260–261 This contradicts the above statement that “similar results are found for the strong and weak initial perturbation”

l. 273–274 This is already described in Section 2.2 (but with 10 ensemble simulations)

l. 276 How large is the spread between simulations?

l. 282 Where?

l. 305 Is sigma the correlation coefficient?

l. 306–307 Please explicit the dynamical link between jet streak intensity and large-scale forcing for upward motion that amplifies the development of the cyclone

l. 322-323 Is the “heavy influence of local factors like sea surface temperature and orography in the Mediterranean” shown somewhere, or is it an assumption based on previous studies?

l. 387 Refer to Fig. 11f

l. 390 Repetition of l. 321–322

l. 392 What are the cyclogenesis hotspots found by Bartholy et al. (2009)?

l. 401 This contradicts the discussion in l. 318–317

l. 404 higher

l. 411-412 Any reference for that? I would expect the strongest cyclones to develop in winter

l. 412–414 SSTs are lowest in spring

l. 418–419 This appears to contradict l. 412–414

l. 433 “the the”

l. 475 Why is the static stability reduced in MAM? SSTs are lowest in spring (see l. 412–414)

l. 481 Not exactly: the Atlantic cyclogenesis is triggered by the upper-level PV anomaly, which also impacts Rossby wave propagation, thus there may be no strictly causal link between the Atlantic and Mediterranean cyclones