This work presents a nice summary of carbonate weathering and denudation using a compilation of 36Cl rates from around the world. Overall, the methods are sound and the results and discussion are interesting.

One particularly interesting result is that bedrock denudation and weathering rates inferred from rivers roughly agree, even though the timescale integrated by these measurements is quite different. The authors argue that this suggests that weathering dominates denudation of exposed bedrock (rightfully, I think). Does this also mean that weathering rates (at least on average) have remained relatively steady over the past several thousand years?

Long-term denudation rates from catchments are faster, and the authors present a straightforward analysis that constrains physical erosion rates from these catchments. Weathering still makes up a substantial portion of the mass flux (much more than in silicate-dominated landscapes), and the authors conclude that carbonate landscapes should generally denude more evenly than silicic ones. 

The weaknesses here lie mostly in some overlooked literature that should be included, and in some cases discussed. I’d also like to see a bit more information on the field sites, particularly the ones that make up the bulk of the analysis.

The literature that first partitioned denudation into weathering and erosion in silicic landscapes should be cited here. The approach presented here builds on that body of literature by applying these ideas to carbonates, but the approach itself isn’t novel. This is also true of some of the weathering bias work (e.g. Riebe and Granger 2012).

More information on field sites would be really useful, especially the Mediterranean ones that form the basis of much of the analysis. Is the bedrock ALL carbonate, or a mix of sedimentary lithologies? How big are the catchments?

I find the argument that “most weathering happens near the surface” in a highly-reactive lithology, even in the presence of caves, to be unconvincing. Does the correction method of Ott et al 2022 include mass loss by weathering at depths > the attenuation lengthscale? Literature on the silicic rock community has identified deep weathering (which isn’t “seen” by cosmogenic nuclides) as potential complication (e.g., Dixon et al., 2009; Campbell et al., 2022). If the correction accounts for deep weathering, it should at least be described in the supplemental material. I actually think it should be addressed briefly in the main text, but I understand that the authors may choose not to devote much space to it if it’s covered in the previous publication. The summary in the supplemental material mentions changes in residence time due to differential mineral weathering (e.g. where quartz and carbonates are present together), but doesn’t address weathering at depth. 

Recent work from Erlanger et al (2021) should absolutely be cited here, and their findings should be considered in the discussion. They found a large fraction of dissolved load was actually re-precipitating as carbonate sand. If this were also happening in the catchments studied here, might the measured dissolved loads actually be a minimum estimate of weathering rates from rivers?

It seems odd to add an example from outside the study (Ireland) in the final figure. I understand that the authors are trying to provide a low-relief end-member, and I recognize that there simply aren’t that many places where 36Cl has been measured in catchments to compare to. Still, I suggest sticking to data reported/analyzed elsewhere in the paper to avoid confusion, rather than bringing in a new setting at the end.

Small tweaks for clarity:

Fig. 1: It looks like the sites from Ott et al 2019 are just catchments in the main figure, but like they’re both catchment and bedrock in the inset.  

I’d love to see a map with catchments in the supplemental – it’s difficult to assess the size, gradient, topography, etc. when they’re only reported as points at the sample location. At the very least, add catchment areas to the table.

Section 4.4: It’s hard to assess this info on these 3 sites without the context of relief or slope. It would be easy to add this info here, so readers don’t have to go to the supplemental table to find it.

Fig. 4: Color-code marker for sites (warm-to-cool?) in order of relief (or average slope). The red and orange dots are a bit close together in color tone, which makes them harder to distinguish on the figure. You might use a simpler color bar for the erosion rate gradient, perhaps? What’s the dashed line?  

Recent references mentioned above:

Campbell et al. (2022). Cosmogenic nuclide and solute flux data from central Cuban rivers emphasize the importance of both physical and chemical mass loss from tropical landscapes. Geochronology, 4, 435–453. https://doi.org/10.5194/gchron-4-435-2022

Erlanger, E. D., Rugenstein, J. K. C., Bufe, A., Picotti, V., & Willett, S. D. (2021). Controls on physical and chemical denudation in a mixed carbonate-siliciclastic orogen. Journal of Geophysical Research: Earth Surface, 126, e2021JF006064. https://doi. org/10.1029/2021JF006064

please see pdf