Thanks for submitting the revised version of your paper. The new version has been reviewed by one of the original reviewers, and I have also had a close look at it. The paper does show substantial improvement compared to the previous version, but I am in agreement with the reviewer that further revisions are needed before this paper is suitable for publication.

The main point that you need to address is why you find the relationship between D95/D50 and the probability of transport occurrence. You need to convince a reader that this is a real relationship, rather than just an artefact of the analysis, and I don’t think that the paper in its current form does this. The data that you present in Table 1 is very useful, and there is a missed opportunity to use these data to explain your findings. Both the reviewer and I have been plotting the data, and identified possible relationships that you should consider when revising the paper.

The probability of transport in each river depends on the difference between the applied shear stresses (which are a function of discharge, width and slope) and the critical shear stresses (which are a function of D84 and the Shield’s criteria). You should be able to use the data in Table 1 to explore which of these parameters are changing most with D95/D50, and therefore whether variations in shear stress or critical shear stress between the rivers are controlling the relationship that you identified. Exploring this would help with explaining the underlying processes.

There seem to be possible relationships between both shear stress and critical shear stress and D95/D50. For the Swiss rivers, there is a positive relationship between D95/D50 and slope. For both sets of rivers there is a negative relationship between D84 and D95/D50, producing a negative relationship between D95/D50 and critical shear stress. These two relationships would both act to produce your identified pattern of mobility. It would be helpful to identify which of these relationships (or any others that you can identify) contributes most to the observed pattern of sediment mobility. This is also necessary to address the reviewers question about the extent to which the observed sediment mobility is controlled by slope, which has already been observed in previous work.

It would also be helpful to think more about why the observed relationships might occur. It’s not entirely clear to me why there should be a negative relationship between D95/D50 and D84; has this been reported elsewhere in the literature? Otherwise, could it be a measurement artefact, with the coarsest grains being under-represented by the sampling in the sites with largest grain sizes?

Finally, you need to address the reviewer’s comments about better comparing your analysis to previous work that has been done, and making sure that that work is correctly represented.

Thanks for your thorough set of revisions. I am happy to recommend that this is published subject to some minor revisions that I have identified out in the attached document. The majority of these are points of clarification. One point to consider throughout is the use of the phrase ‘grain size specific threshold’, when it would be more precise to use the phrase ‘grain size specific critical shear stress’.

The only more substantial point is a query about the extent to which you need to consider the relationship between D84 and D95/D50. It seems to me that there is a negative relationship between the two. (I have to admit that I’ve not checked it statistically though). I think that this relationship occurs because between the finest and coarsest samples D50 increases by 8 times (0.009 to 0.07 m) whereas D95 increases by 4 times (0.029 to 0.128 m), so as the beds get coarser, D50 increases at a faster rate than D95, and the ratio between the two gets smaller.

At the moment, the paper implies that sorting is a completely separate new variable that needs to be considered; but I would argue that, because of this relationship, when you are referring to D95/D50, you are actually also referring to absolute grain size. If you recast the findings in this light (high probability of transport when D95/D50 is high because D84 is small), they become less remarkable. Is there any way in which you could separate out the relative impacts of sorting and D84? At the least, I think that you do need to acknowledge that there is a relationship between the two. Or if I am incorrect and they are unrelated it would be useful to acknowledge that too.

Thanks for your engagement with the review process, and for your thorough revisions in response to each set of reviews. I'm happy that the paper now provides an complete representation of your data, and recommend that it is published. I hope that it will be of interest to a broad range of readers.

Thanks for engaging thoughtfully with the review process.