General comments:

The study by Ren et al. investigates bacterial diversity and community composition as well as potential deterministic and stochastic processes that shape bacterial communities in three types of thermokarst habitats across the Qinghai-Tibet Plateau. The manuscript is clear and concise, the figures are informative and relevant, and overall the study is an important contribution to understanding bacterial community composition and assembly processes in thermokarst landscapes.

However, the manuscript would benefit from a more detailed explanation of certain analyses in the method section, given that some relevant information is lacking. Moreover, the discussion could be expanded to give a more thorough interpretation of the results. Currenlty, the discussion is often very general and does not address some of the interesting findings of this study concretely. Below are more specific comments and examples.

Specific comments:

L139: The paragraph about PCR and sequencing is missing some important detail that would enable the reader to understand what exactly was done and why certain choices were made. First, even though  PCR preparation, PCR conditions were described in a previous study, it would be good to describe them briefly in this paper. Moreover, could the authors describe other steps of library preparation, how were the PCR products cleaned and quantified? Were paired-end reads sequenced, and how many base pairs? The sequences were trimmed at the end; what was the length of the sequences, and what was the trim length? How were the low-quality sequences detected, and what was the threshold used? Which version of the Silva database was used (release date)? How were the sequences normalized, and why was this particular threshold used (24.251)?

L155: Analyses: In some cases, multiple tests were performed (e.g. correlation tests). Did the authors apply any correction of P values for multiple testing?

L176: Could the authors provide more details about the construction of SEM? How were the paths constructed, and why? The path construction should have some theoretical rationale. Did the authors assess the fit of SEM, and which parameters were used for this? Which package was used to construct SEM?

L348-351: Could the authors explain in more detail how this analysis was performed? E.g. how were homogeneous and heterogeneous selection determined?

L433-435: Could the authors provide some possible explanations for this?

L441-447: Instead of repeating the results in detail, it would perhaps be more useful to focus on discussing these results and interpreting what they mean for each habitat. For instance, the general discussion about dispersal limitation was interesting, but more specifically for the study, it would be interesting to discuss, e.g. why heterogeneous selection influenced PBCs much more strongly than WBCs and why dispersal limitation was the most important for sediments.

L473: Could this be discussed further?

L494-504: The authors write about the general importance of different environmental factors for bacterial communities but do not go into why certain of these factors were important in this study. Could the authors discuss more specifically why particular environmental factors are important in these different habitats and, even more interestingly, why the environment seems to be more important in PBCs than the other habitats? For instance, in this respect, it could be interesting to discuss the results of SEM.

In the study by Ren et al, the authors investigate microbial community assemblages in different degraded environments, degraded permafrost soils, thermokarst lake sediments and lake water, with the aim to identify dispersion and assembly  processes. Although the  communities differed among the environments they nevertheless shared 41% of OTUs which suggests that and taxa disperse among the systems.

The manuscript is very clearly structured and well written, and the introduction provides a good overview of the topic. Moreover, due to the increasing possibility of enhanced thermokarst lake formation the authors elucidate and highlight microbial colonization pathways of those newly formed ecosystems.

As pointed out by reviewer 1, more information on DNA extraction and sequencing should be included, as well as on the assumptions made in the base-SEM. Moreover, the rationale for assembly processes could be explained a more detailed, especially the definitions of homogenous and heterogenous selection and how their contribution, and dispersal limitation were estimated, as well as how deterministic and stochastic processes (and what deterministic processes would that be) were defined? Could this maybe elaborated more in the introduction already

The discussion still has large stretches with results being repeated, which could be streamlined and put into a larger context, by relating to other findings.

More specific comments:

Across the manuscript: the abbreviations for the three studied ecosystems are not very intuitive (Permafrost soil and lake sediments are PCBs and SCBs), maybe they could be simplified?

Fig. 1 the letters in particular in Fig. 1a and b are very small and hard to read, please increase size.

Fig. 6: How exactly was habitat niche breadth determined (based on OTU distribution?) and I am wondering if maybe Fig. 5 and 6 be merged into one, as they seem a bit redundant (as also the nmds is displaying the Bray Curtis distances, if I understood correctly.

Line 501 and following lines: this is a very general statement, are there any studies that could here focus more on bacterial differences in thermokarst lakes, or at least in cold/permafrost ecosystems?