Estimation of Net Ecosystem Productivity on the Tibetan Plateau Grassland from 1982 to 2018 Based on Random Forest Model

Round 1

Reviewer 1 Report

Comments to the author

Tibetan Plateau, known as the “Third Pole” of the Earth, is very sensitive to global changing, and plays important roles in the global carbon cycle. The author developed the random forest model to estimate spatial and temporal patterns of NEP on the Qinghai-Tibet Plateau nearly 40 years. They found that the grasslands were weaken carbon sink on the Qinghai-Tibet Plateau and the changes in NEP were mainly driven by different climatic factors in the southeastern and northwestern region. This work offers the supporting evidence to the debate on carbon sink on the Tibet Plateau and can provide an alternative method to estimate the C sink of alpine ecosystems on the Qinghai-Tibet Plateau. I advised a minor revision about some writing details before acceptance for publication, and the following comments should be taken into consideration by the authors.

Main comments

1. In the abstract section, Tibet Plateau is an important study region, why not obviously highlight in the first sentence. The sentence describing the result is the past tense. Please correct it.

2. I had captured more critical information in the introduction section. I think that the order of these information should be adjusted, and in turn is background, knowledge gap, and importance of region.

3. The authors did a lot of literature reading and summary work. I think that the author focused on discussing the main or highlight results and compensated more information. For example, the authors did a good work to compare the difference in C sink between this work and previous studies, but this content was not completely and clearly summary and did not share more information to readers.  

Other comments and suggestions are specific to the following parts of the manuscript:

Line15: “uncertainties” changed as “uncertainty”

Line17-Line19: “Based on the observations from 18 eddy covariance flux sites distributed on the TP grassland, we developed a random forest (RF) model to analyze the spatial-temporal patterns of grassland NEP and its driving factors from 1982 to 2018” Changed as “ Using CO2 eddy covariance flux data from 1982 to 2018 in 18 sites distributed on the TP grassland, we developed……”

Line19: “revealed” changed as “found”

Line20: “is” changed as “was”

Line21: “During the study periods” changed as “during the observing periods”

Line22: “showed significant increasing trend of 0.4 g C m-2 yr-1” changed as “was significantly increased by 0.4 g C m-2 yr-1 ”

Line22: “At regional scale” changed as “At regional scales”

Line23: Can you explain “similar pattern was also found for long-term trends”

Line26: “largely” changed as “dominantly”

Line27: conducted the present tense

Line28: ordering the word based on the letter

Line32-33: changed as “ as critical component………………”

Line37: changed as “experience”

Line42-48: I think this paragraph did not appear here, and put them in the end.

Line42: DELETE “on the TP”

Line45-48: “Therefore, accurate monitor and estimation of carbon budget in grassland ecosystems is a key step to predict future climate scenarios and promote understanding of the feedback between the terrestrial biosphere and atmosphere under climate change [11, 12].” Changed as “Therefore, it is key to accurately monitor and estimate carbon budget in grassland ecosystems under future climate scenarios and promote understanding of the feedback between the terrestrial biosphere and atmosphere under climate change [11, 12].”

Line61-63: Please reconstruct this sentence

Line71-73: delete this sentence

Line76-78: The goals of this study are two or three.

Line120: “44 site year datasets” changed as “44- year data at each site”

Line161-163: “44 s Figure 3a illustrated that our model can accurately reflect the observed data, with R2 of 0.65 and MSE of 3.56 g C m2 8d-1 between the observed and simulated values for validation samples” changed as “We conducted model accurately reflecting the observed data, with R2 of 0.65 and MSE of 3.56 g C m2 8d-1 between the observed and simulated values for validation samples”

Line175: change as exhibited

Line177-178: "occurs” change as "appeared”

Line181: " during 1982-2018” change as "during the period from 1982 to 2018”

Line184: "occurs” change as "arisen”

Line185-186: Please reconstruct this sentence.

Line195: changed as existed

Line197,199: changed as was

Line213-217: Please reconstruct these sentences.

Line240-241: " Our analysis revealed an opposite pattern of NEP controlled by moisture or temperature.” change as " We found an opposite pattern of NEP controlled by moisture or temperature.”

Line246: Please reconstruct these sentences.

Line247: " increase” change as " increases”

Line250-251: " The rising temperature can alleviate the temperature limitations on alpine vegetation growth temperature in high elevation regions” change as " The rising temperature can stimulate plant growth due to alleviate temperature limitations in high elevation regions”

Line253-254: " This result in weak response of NEP to temperature at the regional scale” change as " This result reflects the weak response of NEP to temperature at the regional scale”

Line258: " the time period” change as " the time scale”

Line260-262: “a result of the combined response of photosynthesis and respiration to environmental conditions” is vague.

Line263-264: " the conclusions of the environmental influences on TP grassland based on site-scale studies might be inadequate” change as " the environmental factors impacts on Tibetan Plateau grassland might be inadequate based on site-scale studies”

Line275-276: Please reconstruct these sentences.

Line281: “good” changed as “appropriate”.

Line285: what is “show a clear advantage”. Can you explain that?

Line295: “were non-negligible” changed as “ were not negligible”

Line295-296: “the flux towers can be easily affected by meteorological conditions and exist systematic errors, so interpolation of missing data may lead to the uncertainties” changed as “ CO2 eddy covariance flux observation can be easily affected by meteorological conditions that lead to systematic errors”

Line318: which model? Adding some references

Line344-359:”TEMP ,SHUM” changed as complete spelling

Line345-346 delete “The model was………………….”

Line350-353 changed as “44.7% of Tibetan Plateaus grassland was controlled by SHUM and 28.2% of that was more susceptible to TEMP.”

Line355-359 the sentence is too long. Please combines and simplifies this content.

Reviewer 2 Report

This manuscript is not ready for review due to deficiencies in English composition.  Use of "exacting" instead of "extracting" is just one of many grammatical issues.  The authors should work with a capable editor of English on presentation before I am willing to consider content in detail.

Reviewer 3 Report

The paper takes the grasslands of Qinghai-Tibet Plateau as the research object, and constructs a net primary productivity inversion model for grasslands based on meteorological data, remote sensing data, and carbon flux data, where the research results have some scientific value for accurate estimation of the net primary productivity of grasslands. However, some problems still exist as follows.

1. As the variable for model construction, carbon flux data must be representative spatially and in terms of grassland type; however, the location schematic shows that there are fewer carbon flux points (18 stations) in the northwestern part of the Tibetan Plateau, and mainly concentrated in the southeastern part of the Tibetan Plateau. The representativeness and accuracy of constructing the net primary productivity of grasslands across the Tibetan Plateau with these point data is debatable, while the proposal of adding grassland types to the location schematic;

2. A description of the method or means by which the flux data are converted from daily to 8 days should be added;

3. A need on how much data is used for modeling and how much data is used for validation is required on;

4. In the context of the paper, the indicators used to construct the net primary productivity model for grasslands are mainly meteorological data and remote sensing data; however, the reasons why the paper selects the seven indicators for meteorological and remote sensing need to be described additionally in the paper;

5. How the spatial matching of supplementary meteorological and remote sensing data is addressed is required in the paper;

6. Sources of carbon flux data acquisition are required;

7. What are the causes of negative carbon fluxes;

8. The types and areas of grasslands on the Tibetan Plateau are required;

9. Multiple “section 3.1” appear in the paper;

10. Pay attention to the numbers in Figure 7.