Applying Trait-Based Modeling to Achieve Functional Targets during the Ecological Restoration of an Arid Mine Area
Round 1
Reviewer 1 Report
General Comments
This paper applies a functional ecology lens to minesite restoration in China. The language of the paper suggests that this study informs restoration in order to achieve certain functional targets. The study itself, however, is an observational study with limited application. Though it’s framework is interesting, and the trait work seems well done, I feel like it both needs further clarification on the purpose, and more detail on some of the methods.
Introduction
Oh goodness. There are some problematic claims in this intro that may highlight differences in restoration work in China from many other parts of the world:
Lines 30-31: There are many examples of theory informing restoration, both through the restoration activities of researchers (check many papers in Restoration Ecology if examples are needed), and in the practitioner realm (for example, succession is often cited in forest restoration activities).
Lines 32-33: The use of the word “engineer” here made my eyebrows raise. The vast majority of restoration practitioners globally are not engineers. That doesn’t mean there isn’t a science-practice gap, but it exists between many practitioner sectors and research scientists/outputs.
Lines 47-48: By self-restoration, do you mean natural (unassisted) recovery? The word “artificial” also is jarring and I’m not sure what it means. All management actions taken in an ecosystem are anthropogenic by their very nature, so is there a distinct meaning to the term “artificial” here?
Lines 78-79: The authors are really underselling trait-based research. There are many fabulous examples of applied trait-based research (for examples, check out research by Funk, Laughlin, or Brudvig).
Lines 80-81: Is the CATS model an existing method, or one developed specifically for this piece?
Lines 90-91: I think this highlights my general confusion about the purpose of this study. This sentence is so general, it says almost nothing. There is no application of practice in this study – it is observational, and mathematical. It also does not test any theories. It is still a useful study, but should be described accurately.
Methods
There needs to be a better description of the mine, if it’s going to be included as a study area. What were the disturbances on site? What were the restoration measures?
I also need clarification on the methods:
Why were 14 reference sites selected, and only two sites on the mine?
How were roots collected? Was the whole plant harvested? How was this done?
Please provide more detail on the glms that were used to analyze CWM (which I assume is a single value per plot) against soil values. What was the distribution chosen? Did you validate the model? Were all three variables used? Did you check for collinearity between those variables?
How did the vegetation results from the mine area get incorporated into the modeling? If you are trying to understand the plant communities in reference systems, I think it is problematic to include them in the models. They underwent a dramatically different disturbance/restoration history than the reference sites, and thus are likely to skew the model results by being unique from the reference community composition (that is the argument in this paper, correct? That current restoration targets do not align with things like functional-based recommendations?).
Potentially, you could model just the reference system, get the optimal solution for the mine plots, and compare that solution with the actual vegetation. This would be imperfect, however, because it is complicated by what and how much was seeded/planted in the mine plots, and the age of those plots is significantly younger than the age of the reference ecosystem.
Given that dust capture is an effect trait, how was it considered in the modeling?
How was Tk calculated in the CATS protocol?
Results
The results imply that you ran one model per functional trait, not one model for the CWM of each plot. That is in contrast to how the methods are written.
Discussion
I would like to see some consideration for how different the CATS recommendations are from the alternative restoration methods, which usually include trying to match the composition of nearby reference sites. It seems like these CATS outputs are not much different. That does not negate the usefulness of the method, but it needs to be addressed.
I also would like to see, throughout the paper, a clearer development of the response-effect framing. The effect trait (dust capture) seems to just get thrown into the modeling and CATS with all of the other response traits. Thus, it isn’t actually part of the optimization process, and there’s no clear assessment of how much better one solution might be at dust capture compared to a different solution.
Author Response
General Comments
This paper applies a functional ecology lens to minesite restoration in China. The language of the paper suggests that this study informs restoration in order to achieve certain functional targets. The study itself, however, is an observational study with limited application. Though it’s framework is interesting, and the trait work seems well done, I feel like it both needs further clarification on the purpose, and more detail on some of the methods.
Response 1: Thank you for the constructive suggestions and efforts in reviewing this paper; we have modified the related parts of paper accordingly.
Introduction
Oh goodness. There are some problematic claims in this intro that may highlight differences in restoration work in China from many other parts of the world:
Lines 30-31: There are many examples of theory informing restoration, both through the restoration activities of researchers (check many papers in Restoration Ecology if examples are needed), and in the practitioner realm (for example, succession is often cited in forest restoration activities).
Response 2: We agree with your opinion and have modified this sentence.
Lines 32-33: The use of the word “engineer” here made my eyebrows raise. The vast majority of restoration practitioners globally are not engineers. That doesn’t mean there isn’t a science-practice gap, but it exists between many practitioner sectors and research scientists/outputs.
Response 3: Perhaps the meaning of the word "engineer" in China is different from that in many other parts of the world. To make a clear expression, this word has been replace by “practitioner” in the related place. Thank you for pointing that out.
Lines 47-48: By self-restoration, do you mean natural (unassisted) recovery? The word “artificial” also is jarring and I’m not sure what it means. All management actions taken in an ecosystem are anthropogenic by their very nature, so is there a distinct meaning to the term “artificial” here?
Response 4: To avoid confusion, we have modified this sentence. The word “self-restoration” has been replaced by “natural recovery”, and the word “artificial” has been deleted.
Lines 78-79: The authors are really underselling trait-based research. There are many fabulous examples of applied trait-based research (for examples, check out research by Funk, Laughlin, or Brudvig).
Response 5: We agree with your opinion. This inappropriate expression has been deleted.
Lines 80-81: Is the CATS model an existing method, or one developed specifically for this piece?
Response 6: CATS is an existing method, and has been used in the related research carried out by Laughlin et al. (2013). There is a mistake in reference format. We have modified it.
Lines 90-91: I think this highlights my general confusion about the purpose of this study. This sentence is so general, it says almost nothing. There is no application of practice in this study – it is observational, and mathematical. It also does not test any theories. It is still a useful study, but should be described accurately.
Response 7: Followed your good suggestion, this part has been changed.
Methods
There needs to be a better description of the mine, if it’s going to be included as a study area. What were the disturbances on site? What were the restoration measures?
Response 8: We followed this good suggestion, and a paragraph has been added in the 2.1 section to descript the abandoned mine.
I also need clarification on the methods:
Why were 14 reference sites selected, and only two sites on the mine?
Response 9: Followed the requirements of CATS, on one hand, 14 sites, outside the abandoned mine areas, were selected to build the linear models between soil factors and CWMs. On the other hand, a shady slope sample plot and a sunny slope sample plot in the mine area were selected as soil survey plots in restoration area. Then, we can calculate the targets values of CWMs under the soil conditions of abandoned mine areas in the shady or sunny slope. The related information can be found in section 2.2 and 2.6.4.
How were roots collected? Was the whole plant harvested? How was this done?
Response 10: The whole plant was harvested by a shovel, and the soil was backfill after that. This information has been added in section 2.3.
Please provide more detail on the glms that were used to analyze CWM (which I assume is a single value per plot) against soil values. What was the distribution chosen? Did you validate the model? Were all three variables used? Did you check for collinearity between those variables?
Response 11: We first calculated the value of CWMs and soil factors in each of the 14 herb plots outside of the abandoned mine. Then, generalized linear model was used to analyze the relationship between each soil factor and CWM from the 14 herb plots. Because there are 4 soil factors and 7 CWMs in each plot, 28 correlation curves were fitted by generalized linear model. After comparing different types of correlation curves, 6 correlation curves with the highest value of R2 were selected to build CATS model. Since we did not build a multivariate linear model, collinearity is not involved in the study. So, we did not check for collinearity between different variables. Thank you for the suggestion. We have provided these details on the section 2.6.3 and 3.5.
How did the vegetation results from the mine area get incorporated into the modeling? If you are trying to understand the plant communities in reference systems, I think it is problematic to include them in the models. They underwent a dramatically different disturbance/restoration history than the reference sites, and thus are likely to skew the model results by being unique from the reference community composition (that is the argument in this paper, correct? That current restoration targets do not align with things like functional-based recommendations?).
Response 12: Due to intense human disturbance, there is nearly no plant in the abandoned mining area (Fig. C). So, the value of vegetation coverage and species richness in mine area is default (Table 1), and there is no vegetation results from mine area. The related information can be found in section 2.6.4. Maybe the paper is not clear enough. So, we added a paragraph to descript mine area in section 2.1.
Potentially, you could model just the reference system, get the optimal solution for the mine plots, and compare that solution with the actual vegetation. This would be imperfect, however, because it is complicated by what and how much was seeded/planted in the mine plots, and the age of those plots is significantly younger than the age of the reference ecosystem.
Response 13: Yes, in this paper, we have modeled the reference system, and got the optimal solution for the mine area. But we didn’t compare the solution with the actual vegetation in mine plots, because the mine area is nearly bare land. The purpose of the model simulation is to obtain a plant community suitable for growing in the bare land of the mining area.
Given that dust capture is an effect trait, how was it considered in the modeling?
Response 14: Refer to Response 15.
How was Tk calculated in the CATS protocol?
Response 15: We want to get an appropriative community that can not only adapt to the environmental conditions of mine area, but also play a large enough dust retention function. So, on one hand, based on the linear relationship between CWMs and soil factors, and the value of soil factors in the mine area, the response trait target values for the mine area were estimated. On the other hand, based on the distribution of dust retention amount per unit leaf area of different plants, the maximum value was set as the effect trait target for the mine area. The trait target values (Tk) can be got from above two groups.
Results
The results imply that you ran one model per functional trait, not one model for the CWM of each plot. That is in contrast to how the methods are written.
Response 16: Actually, we have ran generalized linear model for soil factors and CWMs. It can also be said that we have ran generalized linear model for soil factors and plant functional trait in plot scale. Due to CWM is the average value of a trait within a plot, we suggest CWM is an expression type of plant functional trait. However, the inconsistent expression in different parts of paper will affect reading. So the related parts in section 3.5 has been change to make a consistent expression. Thank you for pointing that out.
Discussion
I would like to see some consideration for how different the CATS recommendations are from the alternative restoration methods, which usually include trying to match the composition of nearby reference sites. It seems like these CATS outputs are not much different. That does not negate the usefulness of the method, but it needs to be addressed.
Response 17: We agree with you, and the related information has been added in the second paragraph section 4.3.
I also would like to see, throughout the paper, a clearer development of the response-effect framing. The effect trait (dust capture) seems to just get thrown into the modeling and CATS with all of the other response traits. Thus, it isn’t actually part of the optimization process, and there’s no clear assessment of how much better one solution might be at dust capture compared to a different solution.
Response 18: Thank you for this good suggestion. We have added section 4.3 in Discussion.
Reviewer 2 Report
All suggestions are in the file of the paper.
Comments for author File: 
Comments.pdf
Author Response
Line 94: Fig. 1. (Fig. 1 a, b)
Response 1: Done.
Line 99-100: Whose results are these? Yours? If these are yours then they are part of the Results section, not the Material and methods section.
Response 2: This sentence has been moved to the Results section, followed your suggestion.
Line 107: The research was conducted only for one year (2 months)? I think it is a very short time to make some concrete conclusions about such an important topic.
Response 3: We agree with your opinion. The experiment has been started since 2018, but the main field survey data comes from 2020. The year “2020” has been replaced by “2018-2020” to avoid misunderstand.
Line 109: (Fig. 1c)
Response 4: Done.
Line 119: (Fig. 1 e, d); Plant determination,...;
Response 5: Done.
Line 119: What method was used? Please, explain. Is the word "community" suitable? Perhaps the word "population" is more appropriate (in whole the text of the paper).
Response 6: Community richness means how many species within a community. It can be recorded by the number of species.
Line 120: What method was used? Please, explain.
Response 7: All the vegetation coverage in quadrats were estimated by a same person. The method has been added in section 2.3.
Line 136: Did you take only one sample from each quadrate? It is necessary to take at least 3 soil samples from different places in one quadrate.
Response 8: We take 3 soil sample from different places in one plot, but not in one quadrate.
Line 162: How this is calculated? Which formula was used? Please, explain.
Response 9: Done.
Line 187-188: Species coverage is not seen from Table 1.
Response 10: Yes, but the coverage of plant community or plot can be found in Table 1. For example, plant community coverage/ vegetation coverage in plot 7 is 76.8%. We didn’t calculated species coverage.
Line 318: italic
Response 11: Done.
Line 333, 336 338: I think this is unnecessary.
Response 12: These sentences have been removed, followed your suggestion.
Reviewer 3 Report
Dear authors,
The manuscript is written very well. I did not find any serious errors. Some minor remarks:
Line 318 – „K. tragus” should be in italics
Is this possible to explain a reason for using „dust retention per unit leaf area” as one of plant functional traits? I can imagine that this is important in arid regions, but I think it will be clearer if you explain that shortly.
Author Response
Dear authors,
The manuscript is written very well. I did not find any serious errors. Some minor remarks:
Line 318 – „K. tragus” should be in italics
Response 1: Done.
Is this possible to explain a reason for using „dust retention per unit leaf area” as one of plant functional traits? I can imagine that this is important in arid regions, but I think it will be clearer if you explain that shortly
Response 2: Thank you for the good opinion. We have added the related information in the second paragraph of section 4.3.
Reviewer 4 Report
Comment on 'Applying trait-based modeling to achieve functional targets during the ecological restoration of an arid mine area'
This is a good study and interesting ecological study.
I have a few concerns:
1) You only focus on a) Soil total nitrogen, b) Soil total phosphorus, c) Soil water content and TSP, and TSP Comment: Why is your title not related to these parameters in your study?
2) What is the relationships of these parameters to the abandoned mining sitres? Are these levels positively correlated to the mining sites? Why not?
3) You said 'By combining ecological theory with practice, this study has provided a useful example that advances the study and practical application of ecological theory.' [Comment: How could your findings change and make a novel ecological theory? Please discuss !]
4) can your study relate to conservation effort in the abandoned mining area? Why?
5) Can you discuss your CATS model output is good enough? what things are need to improve it? How?
Thank you
Author Response
Comment on 'Applying trait-based modeling to achieve functional targets during the ecological restoration of an arid mine area'
This is a good study and interesting ecological study.
I have a few concerns:
1) You only focus on a) Soil total nitrogen, b) Soil total phosphorus, c) Soil water content and TSP, and TSP Comment: Why is your title not related to these parameters in your study?
Response 1: Because these factors refer to specific factors of the environment, the title cannot include these specific details.
2) What is the relationships of these parameters to the abandoned mining sitres? Are these levels positively correlated to the mining sites? Why not?
Response 2: “parameters” means soil factor? On one hand, 14 sites, outside the abandoned mine areas, were selected to build the linear model between soil factors and CWMs. On the other hand, a shady slope sample plot and a sunny slope sample plot in the mine area were selected as soil survey plots in restoration area. Then, we can calculate the targets values of CWMs under the soil conditions of abandoned mine areas in the shady or sunny slope.
3) You said 'By combining ecological theory with practice, this study has provided a useful example that advances the study and practical application of ecological theory.' [Comment: How could your findings change and make a novel ecological theory? Please discuss !]
Response 3: This study is just an example of the practice of existing ecological theories, and cannot provide a novel ecological theory, but emphasizes the important role of response-effect trait framework in ecosystem stability and social development. The related discussion and figure have been added in first paragraph of section 4.5.
4) can your study relate to conservation effort in the abandoned mining area? Why?
Response 4: Yes, the ecological restoration methods involved in this study can determine the species composition of the restoration community by setting specific ecosystem functions. The setting specific ecosystem functions can be related to the abandoned mining area.
5) Can you discuss your CATS model output is good enough? what things are need to improve it? How?
Response 5: Thank you for this good suggestion. The related information has been added in the second paragraph of section 4.5.
Thank you
