The approach used here consists of coupling a Lanscape Evolution Model (LEM) developed by the 1^st author with well-known cosmogenic nuclide (CN) production and decay laws, in order to track individual particles (grains) journey from their source to their sink (here, when they leave the model grid). This allows the authors to evaluate statistically how the 10Be signal carried by a population of grains in riverine sands is representative of the average denudation rate of the upstream catchment. For now, I guess the aim of the authors is to demonstrate the model’s ability to achieve the desired goal (i.e., to retrieve denudation rate variations with a limited number of grains and a reasonable computational time), not to address very specific questions on real or synthetic cases. To this respect, this paper is extremely interesting and the Lagrangian approach used here allows to account for the large variability in individual grains histories. The effect of grid size, resolution, time step, and number of grains are tested and the results seem extremely robust. It may however reach its limits for a larger landscape (here the grid is only 10km x 10km, except for fig 7 where it is 20km x 20km), and/or with low denudation rates, for which a large number of grains would be needed.

I only have some minor to moderate comments. There are some points that are unclear to me concerning the LEM itself, independently of the CN part. Some sentences are not very clear and may need to be rephrased. Finally, I think that some outcomes deserve a deeper discussion: the influence of the grains’ origin (Quartz source located close or far from the outlet) and the strong amplitude reduction of the 10Be-derived erosion rate for short period precipitation oscillations.

See detailed comments in the pdf file.

Carretier et al. present a timely development of their Cidre model that supports the rapidly developing field of new lab techniques for sampling cosmogenic concentrations in individual sediment grains. The model allows the numerical exploration of the landscape processes that influence the residence time of sediment grains in mountain catchments and the impact these processes have on the population statistics of cosmogenic concentrations in exported sediment. This is very useful for generating hypotheses that are becoming increasingly testable with field data.

One uncertainty that I have relates to the relationship between relief and CN production in the model and how erosion rate is defined relative to the topographic surface. It would be great if there was a figure clarifying the coordinate system for the attenuation path of cosmic rays emulated together with the surface lowering/mass removal processes. Are the rays attenuated vertical or perpendicular to the topographic surface in the model? Is this important if the model is to be applicable to topographies steeper than those modelled in the paper (>30^O)? I am thinking that steeper slopes do not lower their surface uniformly, so if cosmic rays are attenuated vertically, how might the production rates vary depending on the hillslope model used? E.g. non-linear hillslope diffusion model vs Cidre’s model where detachment rates are proportional to slope and mass removal is modelled with non-local effects above a threshold. Some clarification would be great for us visual learners.

Other minor comments are included in the pdf attached.