Tibetan Plateau grasslands might increase sequestration of microbial necromass carbon under future warming

Microbial necromass carbon (MNC) can reflect soil carbon (C) sequestration capacity. However, changes in the reserves of MNC in response to warming in alpine grasslands across the Tibetan Plateau are currently unclear. Based on large-scale sampling and published observations, we divided eco-clusters based on dominant phylotypes, calculated their relative abundance, and found that their averaged importance to MNC was higher than most other environmental variables. With a deep learning model based on stacked autoencoder, we proved that using eco-cluster relative abundance as the input variable of the model can accurately predict the overall distribution of MNC under current and warming conditions. It implied that warming could lead to an overall increase in the MNC in grassland topsoil across the Tibetan Plateau, with an average increase of 7.49 mg/g, a 68.3% increase. Collectively, this study concludes that alpine grassland has the tendency to increase soil C sequestration capacity on the Tibetan Plateau under future warming.

The soil carbon sequestration capacity of the Tibetan Plateau grasslands was investigated and it showed an increasing trend under climate warming.By combining measurements of a large number of amino sugars with data on soil microbial communities, carbon-metabolizing function genes, climate parameters and soil properties, the current stock of microbial necromass carbon (MNC) was assessed and the contribution of eco-clusters to MNC was determined.Subsequently, the machine learning model was used to predict the change of MNC storage with climate change.Specifically, climate warming will lead to an average increase of 7.49 mg/g of soil MNC of the Tibetan Plateau grassland.These results are useful for understanding of the overall MNC distribution on the Tibetan Plateau under warming.The results are new and interesting.The manuscript was well prepared.Some expressions in the manuscript need to be modified listed in the following: Line 24: we divided eco-clusters, calculated their Line 28: under current and warming conditions Line 34-37: Two sentences can be merged into one sentence.Line 65-65: Need rephrase this sentence, based on the above statement.Line 81: The correlation analysis of beta diversity of dominant and rare microbial communities and microbial necromass carbon is suggested to further illustrate the possible importance of non-dominant microbial community to MNC.Line 86-87: Abundance of 16S gene copies is not a robust metric for total microbial biomass C. Line 103-104: Please define "core microbiomes" in methods.Explain in detail the relationship between "core microbiomes" and ecological clusters.Line 136: Why does the CE model include the first 19 variables?Please explain.Line 141: K-fold Line 144-148: I don't consider the improvement between the CE model and the WE model to be vital.It did improve, but the WE model is not far behind.Line 159-156: I do not understand the connection you're making between microbial consumption of plant C and the consumption of MNC.Line 173: Why was RCP8.5 chosen?Is the eco-cluster also a key indicator in other RCP scenarios?Line 235-237: "microbial community changes".Please give an example or explanation of which microorganisms.Line 252-256: What microorganisms make up the core microbiome and is it similar to the description in line 213-214?Please complete the relevant information.Line 285-295: Whether this paragraph can be merged with other parts.Line 376: NEO, https://neo.sci.gsfc.nasa.gov/And some questions: 1. Fig1: The altitude symbology is mostly lost behind the steppe vs. meadow symbology.Altitude is often a poor proxy for climate, so perhaps it's better to just drop it form the figure.2. What measures were taken to ensure that the previously published data aligned appropriately with the new field data? 3. It can be compared with the driving factors of microbial necromass carbon change in other ecological environments and the climate simulation results, highlighting the important role of the Tibetan Plateau in carbon sequestration under climate warming.
Reviewer #2 (Remarks to the Author): Zhang et al. have investigated how future climate warming affects microbial necromass carbon sequestration through deep learning model.Their investigation based on large scale sampling and published data.They conducted soil sampling and estimated the MNC by measuring amino sugars, and then they ascertained the soil's capacity for carbon transformation by quantifying the expression of genes related to carbon degradation and fixation using qPCR.Generally, the study is interesting and the manuscript is well written.I recommend publication of the manuscript after a minor revision to include: 1-The term "eco clusters" needs for further explanation in the abstract and introduction sections 2-It would be interesting to show data (in the main text) regarding the community composition of NDVI which highly contribute the MNC Reviewer #3 (Remarks to the Author): Reviewer #1: The carbon sequestration capacity of the Tibetan Plateau under future warming has been a hot topic of concern in recent years.This article uses modeling to give quantitative conclusions, which is very important.However, I still have some concerns and questions about some of the details of the article L2: The climate here may be redundant L28-30: The results and discussion in the article do not quantify the average impact of warming on MNC, how did you arrive at this specific value?L32: Under climate change conditions is a broad concept and this article should focus on future warming.L46: Please abbreviate carbon to C, as you have already expressed the abbreviation in the front of the article (L39).All other abbreviated expressions in the article need to be rechecked.L52: Please abbreviate microbial necromass carbon as MNC, as you have already expressed the abbreviation in the front of the article (L39).L56-58: Such an assertion seems too absolute, and there have been many large-scale studies in the past.L199-207: Suggest inserting references here that support the argument.L237-239: The experimental results do confirm the conclusions of the model fit, but consider whether comparisons with other model simulation results are needed to highlight the sophistication of the model in this article.Whether the sentences L271-272 and L222-223 are contradictory in the discussion.L320-326: There seems to be a contradiction here in that the number of sample replicates you collected (n = 6) does not match the number of sample replicates in the article you referenced (n = 7).(Ding et al) L326-329: Are the 71 samples here at MNC based on the same experimental methods of detection?L369-373: There is a large time gap between when data on climate indicators and soil properties are obtained, and I am concerned that using past aridity index to explain current MNC is biased.This may also contribute to the low effect of AI on MNC (Fig. 3a).
Fig. 1 The samples of microbial necromass carbon (MNC) and microbial related genes don't seem to be perfectly aligned, and I'm confused as to how your linear fit here was achieved?Fig. 3a Here it is proposed to determine the significance of environmental factors on MNC.

Dear Editor and Reviewers:
We are grateful for your insightful and constructive comments on our manuscript entitled "Tibetan Plateau grasslands will increase sequestration of microbial necromass carbon under future warming climate" (COMMSBIO-23-4750-T).We have carefully considered and responded to all reviewers' comments, which help us to improve the manuscript substantially.We have revised the content of the manuscript according to the valuable suggestions from reviewers.All changes in the revised version are highlighted in yellow.The line numbers in our response refer to the revised version of the manuscript.The following are the responses and revisions we have made in response to the reviewers' suggestions on an item-by-item basis.Thanks again to the hard work of the editor and reviewers.

Response to the comments of Reviewer #1:
Comment: The soil carbon sequestration capacity of the Tibetan Plateau grasslands was investigated and it showed an increasing trend under climate warming.
By combining measurements of a large number of amino sugars with data on soil microbial communities, carbon-metabolizing function genes, climate parameters and soil properties, the current stock of microbial necromass carbon (MNC) was assessed and the contribution of eco-clusters to MNC was determined.Subsequently, the machine learning model was used to predict the change of MNC storage with climate change.Specifically, climate warming will lead to an average increase of 7.49 mg/g of soil MNC of the Tibetan Plateau grassland.These results are useful for understanding of the overall MNC distribution on the Tibetan Plateau under warming.The results are new and interesting.The manuscript was well prepared.Some expressions in the manuscript need to be modified listed in the following: Response: We thank the reviewer for the comment and suggestion.We have carefully considered all the comments and responded one by one.As follows: Comment: 1. Line 24: we divided eco-clusters, calculated their Response: Done accordingly on lines 24-25.Response: Thanks for the comment.Based on the comment of the reviewer, we conducted a correlation analysis between the beta diversity of dominant and rare microbial communities and microbial necromass carbon (MNC) (Fig. 1).The results showed that there was a certain relationship between the community structure and MNC, in which the dominant microbial communities were significantly positively correlated with MNC, while the rare microbial communities were negatively correlated with MNC.
However, the fitting effect between dominant microbial communities and MNC (R 2 = 0.42, p = 0.0004) was significantly better than that between rare microbial communities and MNC (R 2 = 0.17, p = 0.0390), indicating the contribution of dominant microbial communities to MNC and laying a foundation for subsequent research.We have described the relevant results in the new manuscript of lines 96-99, and the pictures (Fig. S3) can be found in the supplementary materials.Comment: 6. Line 86-87: Abundance of 16S gene copies is not a robust metric for total microbial biomass C.

Response:
We thank the reviewer for the comment and suggestion.According to published studies, 16S rRNA gene copies can evaluate microbial abundance, indicating a certain linear relationship between the abundance of 16S gene copies and total microbial biomass C. Therefore, the abundance of 16S gene copies can be used as a characterization indicator of total microbial biomass C because it is present in all prokaryotic bacteria.Some published papers have explored and supported the relationship between 16S rRNA gene copies/DNA concentration and total microbial biomass to some extent, and these findings jointly strengthen the application of this indicator in related studies.In sum, abundance of 16S gene copies can reflect or characterize the total microbial biomass C. Reference： 1. Fierer, N., et al. (2007).Toward an ecological classification of soil bacteria.Ecology, 88(6), 1354-1364. 2. Kembel, S.W., et al. (2012).Incorporating 16S gene copy number information improves estimates of microbial diversity and abundance.PLOS Computational Biology, 8(10), e1002743. 3. Liang, C., et al. (2017).The importance of anabolism in microbial control over soil carbon storage.Nature Microbiology, 2, 17105. 4. Chen, Y.L., et al. (2020).Large-scale evidence for microbial response and associated carbon release after permafrost thaw.Global Change Biology, 27, 3218-3229. 5. Gong, H.Y., et al. (2021).Soil microbial DNA concentration is a powerful indicator for estimating soil microbial biomass C and N across arid and semi-arid regions in northern China.Applied Soil Ecology, 160, 103869.
Explain in detail the relationship between "core microbiomes" and ecological clusters.Response: Thanks for the comment.In this study, a total of 28 variables were evaluated using random forest (RF), and the first 19 variables accounted for 70% of the total variables, which have been selected and explained in the manuscript of lines 142-146.Among them, 70% is not a definite or constant parameter, but the first 19 variables are comprehensively selected according to experience and reference to the impact of environmental factor variables on MNC in other studies.For example, NPP and TN, which rank relatively low among variables, have been reported in other studies to have an important impact on MNC.Therefore, in this study, we also included them as input variables for the model.The principle of variable screening is to retain as many variables as possible that have a significant impact on the MNC, so as to ensure that some key influencing factors are not missed.Don't worry too much about redundancy in variable selection, because the deep learning model itself can learn to optimize the input variables to ensure the accuracy of the output results.Reference Liang, C., et al. (2011).Microbial production of recalcitrant organic matter in global soils: implications for productivity and climate policy.Nature Reviews Microbiology, 9(1), 75-79. Kallenbach, C.M., et al. (2016).Direct evidence for microbial-derived soil organic matter formation and its ecophysiological controls.Nature Communications, 7(1), 1-10.Cotrufo, M.F., et al. (2015).Formation of soil organic matter via biochemical and physical pathways of litter mass loss.Nature Geoscience, 8(10), 776-779.
Comment: 12. Line 173: Why was RCP8.5 chosen?Is the eco-cluster also a key indicator in other RCP scenarios?
Response: Thanks for the comment.RCP8.5 is the most severe scenario for climate warming.It is chosen as the background for warming because it can maximize the impact of climate warming on MNC and may make the changes to MNC more obvious.In addition, we also predicted the MNC of alpine meadow and alpine steppe under the RCP4.5 scenario.The results showed that the average predicted MNC under the RCP4.5 scenario was 16.92 mg/g, while the average predicted MNC under the RCP8.5 scenario was 18.45 mg/g.Moreover, the predicted MNC values of alpine meadows and alpine steppe under the RCP8.5 scenario were higher than those under the RCP4.5 scenario (Fig. 3).Based on the research objectives of this article, we have chosen RCP8.5 as a more appropriate option to highlight the importance of warming on MNC and the research topic.The indicator effect of the relative abundance of eco-cluster on MNC is based on the investigation results of existing data and is consistent with the objective facts of MNC distribution in alpine grassland and meadow at present.Therefore, the eco-cluster, as a predictor, is the premise of the model prediction hypothesis, and of course, it can also be used to simulate the distribution changes of MNC under other RCP scenario models.
Response: Thank you for the insightful comment.Corresponding to the previous sentence in the manuscript, Table S3 shows the changes of eco-cluster, which corresponds to the changes in the overall distribution pattern of the microbial community, not that one microbial community changes.According to the reviewer's suggestion, we made a modification in the new manuscript of line 243.
Comment: 14. Line 252-256: What microorganisms make up the core microbiome and is it similar to the description in line 213-214?Please complete the relevant information.

Comment: 2 .
Line 28: under current and warming conditions Response: Done accordingly on line 29.Comment: 3. Line 34-37: Two sentences can be merged into one sentence.Response: Thanks for the suggestion.This sentence was revised in the new manuscript of line 35-38.Comment: 4. Line 65-65: Need rephrase this sentence, based on the above statement.Response:Thanks for the suggestion.Based on the suggestion, this sentence was revised in the new manuscript of line 66-69.Comment: 5. Line 81: The correlation analysis of beta diversity of dominant and rare microbial communities and microbial necromass carbon is suggested to further illustrate the possible importance of non-dominant microbial community to MNC.

Fig. 1
Fig. 1 The correlation analysis of PC1 of dominant (A) and rare (B) microbial communities and microbial necromass carbon (MNC).

Response:
Thanks for this helpful comment.In order to facilitate reader understanding and make the article clearer, we have replaced "core microbiomes" with "dominant phylotypes" in the new manuscript of line25, line 108, line 260, line 261, line 264, line 267, line 313, and line 421.The definition of dominant phylotypes has been explained and defined in the method of line 418-421.In addition, ecological clusters are composed of dominant phylotypes in the new manuscript of line 415-430.Comment: 8. Line 136: Why does the CE model include the first 19 variables?Please explain.

Comment: 9 .
Fig. 2 below (in the new manuscript of Fig. 3).The result shows that the hold-out validation of the CE model is better than that of the WE model in the new manuscript of lines 151-152.

Fig. 2
Fig. 2 RF analysis results and model validation.(a) Importance ranking of RF analysis between MNC and environmental variables.(b) CE model K-fold cross validation result.(c) WE model K-fold cross validation result.(d) CE model hold-out validation result.(e) WE model hold-out validation result.