In this work, the authors revised the state of the art on the impact of climate change on flood risk, posing major attention to Australia.

The review is very well written and in line with the scope of the Journal, and there is no need for significant changes before publication.

In my opinion, the manuscript can benefit from a few minor changes:

- in light of having a coherent structure, maybe it is worth adding the “Systematic review” subsection to sections 4.3.1 and 4.3.6

- please add a reference to the sentence at lines 557-559, to justify the “…it is well known…” part

- please consider adding a (graphical) summary of the review in terms of total number of articles per year, per topic, etc. I know that those numbers are provided within the manuscript, but maybe having a figure summarizing them can help readers

- I would like to see some more comments on the reproducibility/transferability of this study. As the results are very much connected with the authors' expertise, what are the key problems in reproducing (or eventually updating) this review? Are the used methods transferable to other countries? Do you think that the results might be influenced by the experts involved?

- In section 5.2 you briefly addressed the biases involved in the study, and in lines 960-961 you said “The outcomes of the per-researcher analyses were consistently similar”. Is it possible to have a more quantitative picture of such a similarity, as well as some more details on the sensitivity testing (line 959)?

I am confident that addressing the above points could help in further improving an already very good manuscript.

Review of Wasko et al, “A systematic review of climate change science relevant to Australian design flood estimation”

Overall assessment

This is an exceptionally thorough and well-written review, which I think will be of interest to a wide readership. The authors methodically review the different approaches for design flood estimation, with a specific focus on the Australian context. They apply objective methods to assess the level of consensus of different sources of evidence. The findings are mostly uncontroversial, and some are quite important, such as the fact that no defensible method currently exists for factoring climate change into flood frequency estimates. The summary of changes in rainfall with warming is particularly interesting, as is the table summarising all the findings (Table 4).

Main comments

L.81. The authors state that there has been “little research undertaken on the non-stationarity of other factors affecting the design flood estimate”. There has been considerable research on a range of nonstationary factors affecting large floods, so this statement seems a little surprising. Does it matter that the research has focussed on nonstationary factors affecting general floods rather than design floods? Presumably the challenge is more about understanding how to incorporate the nonstationarity in the design estimates, rather than a lack of research on nonstationary factors.

L.97-145. Section 2 is a very helpful overview of the main methods, but it provides almost no citations. I think it would be helpful to include at least the key references for each of the methods described. It might also be helpful for the reader to provide a summary table of the strengths/limitations of each method (I leave this decision with the authors).

The paper reviews different methods for design flood estimation based on streamflow and based on rainfall. At various points in the review, the authors seem to suggest that the streamflow-based flood projections are either too complex (L.1040) or too uncertain (L79) for practitioners to apply. I wonder if it might be worth reflecting a little more on the adequacy of different approaches, and the limitations of the simpler precipitation-based approaches (which make up a very large part of the review). They may work well for certain types of flooding (e.g. pluvial floods in an urban context), but is it fair to say they may not work so well for fluvial flooding, where other factors such as changes in groundwater storage need to be taken into account?

Pg.17-18. It is clear that a lot of effort went into producing the ‘best estimates of central scaling rates’ in which the authors independently assigned weights to different studies to arrive at a median estimate and 66% range. The Supplementary table is an impressive exercise in collating and trying to synthesize information from 179 different sources. However, it is a little difficult to see how the summary values in Table 1 were arrived at, since the weights are not provided and the thresholds shown in Table 1 (<1hr or >24hr) don’t match those shown in Figure 2 (<6hr, >12hr) (perhaps I missed why Table 1 and Figure 2 differ). It would be helpful to provide more data on how the values were obtained, so it doesn’t seem too subjective.

Minor comments

L.99 “the primary differences between methods relating to where in the causal chain of flooding the data are obtained, and where the probability model is fitted” could be rephrased for greater clarity, perhaps being more specific.

L.139. The term “efficacy” in the caption of Figure 1 is a little ambiguous and could be clarified.

L.172/183 “average effect size” – please specify the effect size of what.

L.177 “with variation to storm duration .. and location preserved” could be rephrased for clarity.

L.186 “was weighted by”?

L.206-211. Recent work has shown that groundwater is a more important driver of flooding than either antecedent soil moisture or antecedent extreme rainfall (see work by W.Berghuijs); this may be worth mentioning.

L.237-240 “even the use of physically-based covariates is problematic as the covariates should capture the differing processes”: please clarify/elaborate. Also, which of the “statistical associations may not remain constant with climate change”?- it is worth being more explicit.

L.432 “their application to the future period remains untested”: the phrasing is a little odd; application to the future is always untested. Do you mean their predictive ability (to predict out of sample) is untested?

L.957-961. I am not sure how helpful it is to tell the reader that the papers were assessed independently and through weighting of evidence, if the outcomes of those analyses are not presented. It’s a bit like asking the reader to simply trust that the analysis is objective.

L.1068 “the impact on rare floods diminished” could be rephrased for clarity.

Figures

Figure 1. Nice summary figure. The labels of the x-axis could explained in the caption. The “S” shape of the curve could also be explained.

Figure 2. Again, very interesting (and novel) summary figure. I don’t think 2xCC and 3xCC are defined.