Öttl and colleagues present a very comprehensive modelling study on the effects of soil redistribution on soil organic carbon stocks over millennial timescales. The manuscript provides detailed information on changes in agricultural practices, erosion intensities and the effect on SOC fluxes over this period.

The authors simulate a large variety of modelling scenarios, with varying erosion intensities and SOC dynamics. I think this is one of the first times that such a sensitivity analysis is performed in soil-landscape evolution studies. I’m impressed with the amount and quality of the work from the authors, but I have two main concerns which should be addressed before the paper can be published. These relate to the structure of the manuscript and missing information on certain data and definitions.

I detailed these concerns below and added more explanation and additional smaller comments in the attached PDF.

With best regards,

Marijn van der Meij

Structure of the manuscript

I feel that a clear structure is missing in parts of the manuscript, which makes it difficult to understand what is presented. With a revision of the structure, this should be easy to improve. Examples are

• the Methods section, where inputs, parameters, calibrations and justifications are presented in a mixed manner;
• The Results section, where a logical order of the results is missing;
• A missing overview of all the different scenarios that have been simulated. For example, only in the Results and Discussion it becomes clear that there are also simulations with and without deep C burial. A table with all scenarios would be helpful here.

Missing information

When reading the manuscript, I found several terms, data sources and modelling settings that were not introduced in the text. Examples are:

• The definition of a heavily eroded soil, both for the classification from remote sensing and in the model results (line 285, Section 3.1);
• The SOC-depth profile from forest soils that were used as initial condition (Fig. 3). This data is not introduced in the text;
• The absence of updating the topography after soil redistribution. This is mentioned for the first time in Section 4.4.1. I was surprised this was not included, as it has a big impact on erosion and SOC dynamics over 1000 years and should be relatively easy to include;
• A Discussion on how the size of the simulated landscape affects the model results. In other words, what are the benefits of simulating the entire Quillow catchment instead of the test sites?

Öttl et al. explored the long-term impact of water and tillage erosion on landscape-scale carbon dynamics based on a modified version of the spatially explicit soil redistribution and carbon (C) turnover model SPEROS-C, moreover, the model parameterisation uncertainty was estimated. The results indicate that in young moraine areas, SOC patterns and dynamics are substantially affected by tillage-induced soil redistribution processes, and it was estimated that after 1000 years of arable land use, SOC redistribution by tillage and water erosion results in a landscape-scale C sink of up to 0.66‰ per year. The MS is well-written and should be of interest to readers of SOIL. I have two suggestions that may improve the manuscript.

First, more details should be provided in Methods, so the readers are easily able to understand the paper, e.g., how the data was obtained and the models constructed.

Second, in Figure 6, while a significant correlation was observed, however, there are still spaces to improve prediction accuracy (R²). Moreover, if the data is suitable for regression analysis? e.g., if the residuals of the regression line are normally distributed?