Dear Authors,

I was invited to review the Manuscript Number: egusphere-2024-1088, titled "Influence of building collapse on pluvial and fluvial flood inundation of metro stations in central Shanghai."

The manuscript investigates the local-scale effect of building collapse on the propagation of both pluvial and fluvial flooding. The study focuses on the Huangpu district in Shanghai, analyzing 17 points representing metro stations. For hydraulic modeling, the authors employed SERGHEI-SWE, an open-source high-performance hydrodynamic model that solves the two-dimensional shallow water equations on Cartesian grids using Godunov-type finite-volume methods. For modeling building collapse, the authors adopted an approach similar to Takabatake et al. (2022), wherein 50% of the original building volume, estimated from the building footprint area and height data, is distributed to the surroundings.

The manuscript is well-written, easy to read, and the overall approach is clear and well-represented. The main methodological steps and the results section are presented clearly. Assumptions are stated and discussed in the discussion section. However, the discussion should further clarify the novelty of this work, particularly in relation to the specific objectives defined in the introduction. I suggest revising the initial section to define the aims and objectives more explicitly and structuring the discussion to reflect these objectives (authors state "We aim at understanding if, how and why earthquake-induced building collapse affect the inundation of the metro stations. The findings will serve as the cornerstone for multi-hazard risk assessment of urban metro systems")

It is important to specify that the analysis focuses on hazard rather than risk, as no vulnerability assessments of the exposed assets were considered. The manuscript correctly discusses multi-hazard rather than multi-risk; however, a more detailed description emphasizing this distinction is recommended. Additionally, it is crucial to note in the discussion that the call for more local-scale analysis, highlighted in the article, must be accompanied by a detailed analysis of both exposure and vulnerability. Otherwise, the effort in hazard assessment may be rendered ineffective. Moreover, greater attention should be given to the term multi-hazard: the work is preparatory for multi-hazard assessment, as the earthquake is not studied directly in this work but rather random collapses are considered.

Finally, I have a few minor observations:

• Figure 1 appears incomplete and does not describe the entire complexity. If it is self-produced, the various choices need to be explained and justified; otherwise, the source should be cited (e.g., for river the driver is only SLR, bearing body are only those?). I do not believe it helps clarify the manuscript's work; I suggest rethinking it to focus more on the conducted work rather than a general framework.
• Add a discussion on the assumption of constant rainfall over the entire area of study: while it is a cautious approach, is it reasonable? This should be discussed in the manuscript.
• Line 236: "Pluvial floods originate from heavy rainfall, which is spatially distributed" -> This is due to the assumption of constant rainfall, but in reality, rainfall varies from area to area and maybe the difference from the fluvial flood will become less relevant.
• Line 127: What magnitude scale was used to indicate the earthquake?
• Line 140: Provide more information on the collapse mechanism. Although the model is cited, it would be helpful to provide a few more details on the adopted collapse mechanism. For instance, how is the presence of nearby buildings considered? How far do the debris reach?
• Line 170: The collapse mechanism becomes crucial for considering the impact on a station located within a collapsing building.
• Line 213: The written content does not correspond to what is reported in Table 2; please verify and correct.

I believe that these suggestions require a major revision of the paper, and if you are willing to consider these revisions, I will be very happy to revise the new submission. 

The authors present an interesting article where they investigate the impact of building collapse on flooding of metro stations. While the topic is quite interesting, several things only became clear to me in a late stage of reading the paper. More explanation of the simplifications and assumptions is required.

• When I finished to read the introduction (chap. 1) it did not become clear how the building collapse is taken into account; this should at least be addressed briefly here and then refer to chap 2.3 for more details; add reference(s) to the literature
• Chap 2.2: how are the metro stations taken into account, as holes in the model, i.e. inner open boundaries, where water can flow out of the domain? If I see correctly, this is not the case and you just observe the water level in the cells where metro stations are; this simplification should be discussed and explained, possibly you already comment on this in chap. 1
• 141: The sentencse does not become clear, more explanation is required.
• chap. 2.3: what happens during collapse is not fully clear; better description or a figure where 1 building is shown before and after collapse, give length, width, height, volume etc
• You should more clearly state that your approach is a very first step to investigate impacts of building collapse on flooding at metro stations; what is the value of your results so far; you arbitrarily selected n=100 simulations with different number of collapses, what happens, if the number n is higher; you arbitrarily selected 3 break points for the fluvial flooding, what happens if this number differs; there is no need to make further simulations, but you should discuss this (more), include that in chap. 4, overall more justification

Minor:

• some more information on OpenMP, Cuda, HIP, MPI; not everyone knows this
• in Tab 1, 2: give [%] here, then no need to repeat in each row

Further minor comments are in the attached pdf.