The paper has successfully addressed the issue of estimating the consequences of high-impact-low-frequency events. The research applied a combination of profound models to generate multiple scenarios of global warming, sea level rise, and storm tracks, as well as simulate corresponding flood events. The adoption of multiple stochastic process-based models greatly reduced the inherent uncertainty and arbitrariness of the Storyline Approach. Furthermore, the study managed to include and compare the factors of climate change and internal variability. This further contributes to the understanding of major driving force of high-impact events in the future. With mostly positive feedbacks and pleasant learning experience, there are some minor comments for the current manuscript.

1. Climate scenario constructions. The paper in general well explained how the researchers employed spectral nudging to recreate climate events under various climate conditions. However,

    

   a.What is the necessity of creating a pre-industrial (PI) climate scenario, instead of building a climate change scenario warmer than 2 degree?

   b. More reflection is recommended on the validity of these reconstructed tracks. As the authors pointed out in Figure 2 and line 221, the simulated storm tracks did not well represent the MSLP, highest wind speed, or the variation in flood volumn, though they in fact have rather successfully reproduced the situation when the storm hits NYC. Such discrepancies should be better discussed.

2. Vulnerability of CI to different water levels. It is well understandable that precise estimation of the fully continuous vulnerability curve of various CI is nowhere to find. Therefore it is a common approach to use a discrete and qualitative impact function. However, it remains quite confusing to me how the authors in Section 3.5 managed to give a quantify the change of impacts. I assume the authors actually assigned a percentage of damage to each level of exposure defined in the paragraph between line 172 and 175. It may be more clear to give this numerical relationship.
3. Results presenting. Figure 2 on Page 8 could probably have been polished, such that the simulated results of MSLP, wind speed and precipitation in NYC could be highlighted.

The authors present an approach for understanding the potential impacts of high-consequence tropical cyclones by generating alternate, realistic tracks of a historical event subjected to natural variability and different climate states. These alternate tracks are then simulated using a high-resolution coastal flood model to investigate how differences in the TC tracks/hazards result in differences in flood depth/extent. The framework is applied to Hurricane Sandy, demonstrating that different potential tracks of Sandy could have resulted in widely different flood dynamics. Overall, the paper is well-written and the storyline approach presented here could be very useful in developing coastal hazard scenarios in support of decision-making. I have several questions/comments that should be addressed to improve the interpretability of the manuscript. 

Section 2.2: Please add a table summarizing all the scenarios. At some point I lost track of the number of storylines. 3 climate states x 3 internal variability runs x 2 SLR scenarios x 2 precipitation scenarios...?

Line 75: need a clear one-sentence description of spectral nudging. 

Section 2.2.1 and 2.2.2: Why are the climatology and SLR scenarios not consistent? They use very different time periods (i.e. 2044-2053 for the storm climatology and 2080-2150 for SLR), which does not make sense to me. I understand that both the climatology projections and SLR projections are taken for a 2 degree C (roughly) scenario. But couldn't the authors take the SLR projections from the same GCM (i.e. MPI model) so that the SLR and climatology are consistent with each other. See Lockwood et al. (2022). 

Section 2.3.1: Do you control for the timing between the peak surge and peak astronomical tide? As the landfall timing could be different in each of the climate/internal variability scenarios, how do you account for potential differences in the timing of the tide and surge? If the peak surge occurs at low tide, then overall water levels would be lower (but not due to differences in the climate state, just by chance). Also, as you have precipitation maximization scenarios, what about a surge maximization scenario? 

Line 174: Are these water level thresholds based on any impact literature? For example, 2ft (~61 cm) of water is typically considered the point at which most roads become inaccessible (according to US National Weather Service). If the categories can be linked to any rough impact level that would improve the results. 

Section 3.1: Would be helpful to generate an image showing the potential intensity over the North Atlantic during the storm for each scenario. Are there changes in PI due to global warming that are ultiimately not translated to the track? Or does the similarity of the tracks stem from similar large-scale TC-favorability conditions? 

Lines 183-184: The simulations also underestimate the wind speed at landfall, which would cause underestimation of storm surge. Is the underestimation due to the horizontal resolution of the ECHAM model? 

Line 190: I'm surprised there is no change in rainfall at the study area. Just by the Clausius Clapeyron relation one would expect to see an increase in rainfall associated with a 2C increase in mean global temp. Also, previous work by Liu et al. (2018) projected a large increase in rainfall from extra-tropical transitioning TCs under future warming. I understand that the specific spatio-temporal conditions during the storm may not reflect mean projections, but I think the authors need to add some discussion/comparison with previous studies. Also, as I mentioned earlier, figures and discussion about the differences in the large-scale conditions stemming from each climate scenario are needed. That way we can understand how changes (or lack thereof) in the regional conditions affect the features of Sandy's track

Liu, M., Vecchi, G. A., Smith, J. A., & Murakami, H. (2018). Projection of landfalling-tropical cyclone rainfall in the eastern United States under anthropogenic warming. Journal of Climate, 31(18), 7269–7286. https://doi.org/10.1175/JCLI-D-17-0747.1

Figure 3: add the tracks to b) and c) 

Figure 4: add the track to b)

Line 225: In this case, maximizing the rainfall (which occurred on the left-hand side of the storm) also minimizes the storm surge (as minimal or negative storm surge values are usually observed on the left side of landfalling TCs). This is because winds are pointed away from the coast on the left side of the storm.  

Line 229 "The MP scenario results in the highest number of flooded assets...": But how much of this increase is locations with very low inundation (i.e. on the order of 0.05 m)? I feel that such a low impact threshold of 0.05 m somewhat inflates/exaggerates the impacts of rainfall. Would 0.05 m of water actually cause damage to any building? I assume that most buildings are elevated more than 0.05 m from the bare earth surface