The authors created high spatial resolution data of organic carbon distribution in the Third pole by compiling all the field data and using machine learning methods. The dataset can be very useful to help the scientific community to understand the carbon cycle. I found the paper is well organized. My major concern is that the authors should clearly explained what are the new findings in comparison with several previous reports on the plateau, e.g., Ding et al. 2016, Wang et al. 2020.

Specific comments:

Figure 1 ,2 and 4 are never mentioned in the manuscript, please add reference mark in right place.

Many SOC data were collected in this study, however, the data of China’s national soil survey were not included in this manuscript, why?

Please add the accuracy assessment in manuscript between your dataset and other global or regional SOC datasets, such as: SoilGrids and HWSD.

L 62-64, Permafrost degradation will not only cause serious geological disasters and affect engineering construction in cold areas…” needs citations here.

L.87-90. “Furthermore, the large-scale maps of vegetation and soil types…”  needs citations

L.33, L.216, and L.367. what is SOCSs?

L.302 & L.302. Change “in the area of” into “in the areas of”

L.223. Change “Fig. A1” to “Fig. S1” Also in L.323.

L.248. “To test the predictive effects of the two machine learning methods…”, two or three?

L.289-291. “The estimated SOC storage at a depth interval of 0–300 cm in forest, shrub, cropland, grassland, and desert areas was 3.30 Pg, 0.85 Pg, 31.67 Pg, 9.77 Pg, and 0.59 Pg, thus accounting for 7.15%, 1.84%, 68.58%, 21.57%, and 1.28% of the total, respectively”. Mismatching in different vegetation types and SOC storage values.

L.296. “lithosols” Inconsistent font.

L.302. Change “QTP” into “Third pole region”

L.313-314. Hence, most terrestrial SOCS studies have focused on the shallow soil layer within 100 cm …especially that of permafrost zones (Ding et al., 2016; Mu et al., 2015; Wang et al., 2020; Zhao et al., 2018). Wrong citations. The soil depth is deeper down to 100cm in those articles

L.475. Change “In” into “in”.

This is a very useful dataset that can be used and cited in the future. But the manuscript is written as a scientific research paper rather than a data description paper. I suggest the authors to restructure the paper to describe what the data actually contain. It is unclear to me whether the dataset contains the climate data, which are collected from some other source. So these things need to be more clearly described. Please also clarify what are modeled and what are in situ data. The title should also reflect this.

For the same reason, the introduction, results, and discussion sections are unnecessarily long and not very useful in actually understanding what the data are.

Minor comments:

- Please describe how and why the sampling area (soil pits) were selected.

- How did the authors deal with gravel and what was the gravel %?