The authors attempt to assess which percentage of flood events can theoretically be observed by the satellites Sentinel-1 and -2. They do so by a rather coarse synthetic study, with several optimistic assumptions, but I very much appreciate the research question and the clarity with which the authors approach the topic. The paper is very well written, key messages are listed in the conclusions, figures are good quality. The authors discuss most of their assumptions, there are just a few points on which I request further clarification:

1. A lot of SAR data is unfortunately commercial, Sentinel-1 being a notable exception
2. It is absolutely fine as a “working framework” to assume that it is always possible to map inundated areas from satellite images, however to state that this assumption is “quite confirmed by several examples in literature” is misleading. There are still big issues with satellite-based flood mapping! A general discussion on the limitations can for example be found in Schumann 2021: https://doi.org/10.1016/B978-0-12-819412-6.00014-6 and a more specific evaluation on a well-documented flood, mapped by the Copernicus EMS and other scientific products with a good ground-truth reference, revealed that rapid mapping products can be of very poor quality (Table 3): https://doi.org/10.3390/rs13112042 This is not only a problem of the classification algorithms, but indeed also of the image quality / observability. In sensor design there is a trade-off between image quality and coverage, so when investigating the potential coverage of specific satellites, we have to deal with the true quality of the sensor. Another key difference between optical and SAR data is the viewing geometry. Operational SAR sensors are side-looking, which causes additional issues in urban areas, radar shadows, layover effects etc. Therefore it is in theory able to detect water below vegetation, however that depends on wavelength and is usually not part of operational flood detection algorithms. Ignoring flooded vegetation can obscure the true land-water boundary, and most impacts occur in urban areas. Products like EMSR detect almost exclusively open water, and even that is not always convincing! There should be a paragraph in the paper on these limitations, to avoid the impression that the 58% potentially observable flood events by Sentinel-1 actually translate to 58% of flood events being mapped in sufficient quality, I assume it is only a small fraction of that.
3. The authors made an assumption on cloud coverage based on a dataset by Wilson and Jetz 2016. They do not take into account that flood events are typically triggered by rain, which requires clouds, and therefore should expect a correlation between the presence of clouds and a flood event. The assumption of the authors is therefore optimistic, which is ok as long as it is clearly stated.  I would find it interesting to actually check how often and how long floods are accompanied by clouds, depending on the geolocation/climate, but I understand that this was not aim of the study to do so. My feeling is that there could be quite significant spatial differences on the percentage of floods that optical sensors may detect (while for SAR it should be the same percentage in all places). Detectability on SAR images probably depends more on topography or built-up density. As the two percentages are your primary results, please briefly discuss this point and whether you think it is useful/possible to put a number on that spatial variability.
4. Another debatable assumption (which the authors do mention) is the definition of a flood event by placing a percentile threshold on a 10-year discharge (!) observation time series. Whether a flood occurs or not is of course dependent on the protective measures, which drastically vary in their design level, up to > 1/10000 years in the Netherlands https://www.deltares.nl/app/uploads/2014/12/kind2014_JFRM1.pdf In rural areas, if there is nothing except agricultural crops to protect, actual flood defense might be much lower, but this will probably not be at the location of the measurement station? There are footprints of real flood events, e.g. from the Dartmouth flood observatory that could potentially be used for such a purpose. See Figure 6 in Lüdtke et al. 2019: https://doi.org/10.1029/2019WR026213 I am not entirely sure why the authors have not used validated flood locations, but I do find the synthetic approach also very interesting. Maybe you can make this more clear?

I am happy to recommend the article for publication, if the abovementioned discussion points are addressed.

Tarpanelli et al., address an important question in the submitted manuscript, namely the suitability of the Sentinel 1/2 Satellites for flood inundation mapping in Europe. Through synthetic assessments based on discharge values to detect flood events, the authors simulate the satellite coverages and calculate the probability of capturing inundation events through the Sentinel-1 Synthetic Aperture Radar sensor and the Sentinel-2 optical sensor. The study also tests the findings from the synthetic study for three real world flood events and find most of their conclusions supported through this analysis. In principle the paper is well written, easy to follow, and of interest to the larger flood mapping community in Europe. I only have some minor comments as summarized below, which I believe will help improve the quality of the manuscript.

1. Referencing: the introduction cites papers from over a decade ago to establish current flood impacts and scientific advances. Given the rapidly evolving body of literature in this topic, I think this should be improved as newer publications are sometimes tackling the newer challenges in the field which arose out of the rise of big data and machine learning. I have provided some reference suggestions, but the authors are welcome to seek out some more. I also found some references cited wrongly – e.g. Clement et al. 2018 is cited for the extraction of inundation water levels which they did not actually do in the paper. The authors should check such oversights in the referencing throughout and correct these before publication.
2. Recent relevant developments: For a study looking to assess the suitability of the Sentinels for flood monitoring in Europe, the paper misses two very important and relevant new developments. These are namely, the launch of the Global Flood Monitoring Service from the Copernicus Emergency Management Services and the failure of Sentinel-1b. While it might not be possible to account for the latter in the analysis at this time without needing to reproduce the figures, I think it is still relevant to acknowledge this issue either in the introduction or in the discussion/conclusions and the potential impact this has on the conclusions of this study.
3. Relationship between discharge and inundation: The authors assume that the discharge peaks represent the inundation peaks as well, this is not true in most cases due to the lag between the channel and floodplain peaks, as well as the highly non-linear relationship between discharge and inundation. Again, I do not think there is any need to alter the analysis, however, it would be nice to have the authors acknowledge this point while stating their assumptions and then assess the potential impact this may have on their conclusions in the discussion section.

Technical corrections are very few and included in the reviewed PDF file. On incorporating these minor comments, I think this article would form a valuable addition to the published literature in this direction. I look forward to seeing the final version online and thank the authors for their time and efforts.