Evaluation and Tradeoff Analysis of Ecosystem Service for Typical Land-Use Patterns in the Karst Region of Southwest China

Round 1

Reviewer 1 Report

I carefully read the submitted manuscript.
This study evaluated tradeoff relationship of ecosystem services among different agroforestry patterns in southwest China. I think manuscript is well constructed,topic is suitable for forests.However, in material and method section, I think some revision should be needed.

line90-92, location of study sites: It would be helpful for reader the map or figure of location of study sites.

Table2: I think the units were lacking in this table.

line134: for ICP system, the information of model, company,country should be added.

line 138: farmers.

That's all.

Author Response

We would like to thank you first for all the positive comments of our manuscript Manuscript ID: forests-738952) entitled “Evaluation and tradeoff analysis of ecosystem service for ten agroforestry patterns in the karst region of southwest China”. We really appreciate your help and patience.

1. line90-92, location of study sites: It would be helpful for reader the map or figure of location of study sites.

A location diagram for each mode has been added to the manuscript as figure 1. It can be seen in line 114.

2. Table2: I think the units were lacking in this table.

The corresponding units have been added in table 2.

3. line134: for ICP system, the information of model, company, country should be added.

We have added specific model and device names to the manuscript. It can be seen in line 138.

4. line 138: farmers.

Yes, we have corrected the problem. It can be seen in line 142.

Reviewer 2 Report

I was excited to review this article because there is little research on evaluating ecosystems services and their potential trade-offs in agroforestry systems.  The overall approach is intriguing and may offer one way to assess multifunctionality and potential compromises.  That said, I have several key concerns that need to be addressed.

One of my first major concerns is focused on terminology. The title and text of the article refers to ten agroforestry patterns; however seven of the patterns are not what most practitioners or scientists would refer to as agroforestry based on standard definitions.

FAO definition: Agroforestry is a collective name for land-use systems and technologies where woody perennials (trees, shrubs, palms, bamboos, etc.) are deliberately used on the same land-management units as agricultural crops and/or animals, in some form of spatial arrangement or temporal sequence. In agroforestry systems there are both ecological and economical interactions between the different components.

So in Table 1 (Ln 113), these seven agricultural patterns are not agroforestry based on accepted definitions.  I have referred to Zou et al. 2019, and I am still not convinced these are agroforestry.  The actual agroforestry patterns evaluated in this study include peach + pumpkin, apple + soybean, and pomegranate + grass.  That said, it does not negate the study or the comparison of the 3 agroforestry patterns in relation to the 7 traditional monoculture systems.  It is still a worthwhile comparison, but the text needs to be revised to make this important clarification. 

In Ln 83-84, it should be 7 traditional monoculture systems and 3 agroforestry or intercropping systems.   

Methods

I realize there is always the challenge to provide enough detail in a manuscript while trying to keep it to a reasonable length. The section describing the evaluation of the ecosystem services (Ln123-143) lacks critical detail to evaluate. This part of the research is foundational for assessing the rest of the results and currently leaves many unanswered questions and concerns.  

From the text, 50m x 50 m quadrats were used for each pattern and 5 plants were “planted” in each quadrat.  After rereading this multiple times, did you mean “sampled” instead of “planted”? What size and how old were the different plants when the data was collected? How was this sample framework applied to the intercropping patterns?  For instance, were 5 plants of pumpkin and 5 peach trees sampled?  How were the ten mature leaves used in the assessment? Foliar assessment? How was this standardized across the different species?

There are also many questions regarding the evaluation methods and formulas.

For provisioning services, what was the spatial unit for yield (e.g., plant, ha, etc)? It seems that Would it be more useful to include production costs to get at a net income per unit of area?  As shown in Table 1, the fruit trees required higher inputs than the corn and marigold which can skew results.  Were the fruit trees at similar production age to allow for appropriate comparison of yields? As it was mentioned in Ln 243, there was little difference in the provisioning services between corn and orange because the orange trees had not reached peak yield of the time.

For reduction of topsoil loss, what is the “area of equivalent reserved land” and how is this formula getting at this particular service? How was the amount of nutrient loss calculated?  Was this based on runoff and/or leaching?  Similarly, how was the retention of NPK fertilizer calculated? How was the amount of oxygen released determined?  

How was carbon fixation calculated? Assume it was based on biomass but how was this done given the many different plant forms? It was mentioned that biomass was calculated based on chest diameter, height, and canopy width.  Generally, volume needs to be calculated and used with previously established allometric equations for individual species to get at biomass and then carbon storage. How was this done? Was the method focused only on aboveground biomass/carbon? How did age of the plants figure into carbon/biomass calculations? How was this standardized for a unit of area (i.e., corn plant vs pomegranate tree)? 

Regarding cultural services and education function, should the replacement cost be considered a negative or positive?  Based on the text, local government was providing training for the various patterns and this cost is borne by society.  Also, how was the training costs calculated for the individual patterns?

Can another table be provided that shows the input data, formulas, and results for each of the seven ecosystem services based on the ten cropping patterns? For the intercropping patterns, how was the data aggregated from the herbaceous crop and woody crop?  If the table cannot fit within the regular manuscript, consider it as an online supplemental table. Please provide references for the economic evaluation methods used (shadow price method, opportunity cost method, etc). 

Results

Without better clarification of the research methods, it is difficult to fully assess the results.

In terms of the intercropping systems, it is difficult to determine the value of the combinations since the crops added with the trees (soybeans, pumpkin, grass) did not have a standalone data. For instance, there was a 63% increase in provisioning services with the pumpkin and peach intercropping. Was this simply due to a yield/price of pumpkin? How would the results change if pumpkin was a standalone crop? I do appreciate Figure 3 which gets at an important comparison although I would have like to have seen soybeans, pumpkin, and grass included. 

Should Table 3 also include relative ecosystem services of three intercropping patterns? It would allow for a better understanding of the primary drivers that result in the TES/EMF in Table 4.

Author Response

We would like to thank you first for all the positive comments of our manuscript Manuscript ID: forests-738952) entitled “Evaluation and tradeoff analysis of ecosystem service for ten agroforestry patterns in the karst region of southwest China”. We really appreciate your help and patience.

1. One of my first major concerns is focused on terminology. The title and text of the article refers to ten agroforestry patterns; however seven of the patterns are not what most practitioners or scientists would refer to as agroforestry based on standard definitions.

FAO definition: Agroforestry is a collective name for land-use systems and technologies where woody perennials (trees, shrubs, palms, bamboos, etc.) are deliberately used on the same land-management units as agricultural crops and/or animals, in some form of spatial arrangement or temporal sequence. In agroforestry systems there are both ecological and economical interactions between the different components.

So in Table 1 (Ln 113), these seven agricultural patterns are not agroforestry based on accepted definitions.  I have referred to Zou et al. 2019, and I am still not convinced these are agroforestry.  The actual agroforestry patterns evaluated in this study include peach + pumpkin, apple + soybean, and pomegranate + grass.  That said, it does not negate the study or the comparison of the 3 agroforestry patterns in relation to the 7 traditional monoculture systems.  It is still a worthwhile comparison, but the text needs to be revised to make this important clarification.

In Ln 83-84, it should be 7 traditional monoculture systems and 3 agroforestry or intercropping systems.

First of all, thank you for your advice on the concept of agroforestry. After discussion, we revised all the relevant descriptions, including the title of the manuscript and the title of each figure and table. We believe this will do enhance the rigor of this manuscript. It can be seen in line 83-84, figure1&3 and table 1&3.

2. From the text, 50m x 50 m quadrats were used for each pattern and 5 plants were “planted” in each quadrat. After rereading this multiple times, did you mean “sampled” instead of “planted”?

Yes, we have revised it. It can be seen in line130.

3. What size and how old were the different plants when the data was collected? How was this sample framework applied to the intercropping patterns? For instance, were 5 plants of pumpkin and 5 peach trees sampled?

On the whole, since various fruit planting patterns have similar life cycles, the samples we want to choose have reached the stable stage, including yield and biomass, etc., so we chose sample plots with size and tree age as the important criteria. The planting scale of 0.66 ha or more ensures the stability of management level such as fertilization and medicine, and the 5 years' tree age can guarantee the stability of biomass and yield of fruit trees. However, in real field surveys, we found citrus orchards in the same area that were only four years old in 2016.

As for sampling, 5 fruit trees were used to measure the biomass of the trees. Intercropping samples were taken using the full excavation method, with five pumpkins or five 1m* 1m plots of grass or soybeans being dug up for weighing.

These details have been added to appendix A.

4. How were the ten mature leaves used in the assessment? Foliar assessment? How was this standardized across the different species?

The mature leaves did not end up being used in the calculation, so we deleted it.

5. For provisioning services, what was the spatial unit for yield (e.g., plant, ha, etc)? It seems that Would it be more useful to include production costs to get at a net income per unit of area? As shown in Table 1, the fruit trees required higher inputs than the corn and marigold which can skew results.

First of all, due to space constraints, we cannot present every calculation or sampling detail used in the text. I do thank you for making us aware that this information is also important for peer review. Therefore, we have added appendix A. In our study, yield, soil loss, biomass, etc., were calculated by converting to per hectare and then standardized. In the process of literature investigation, we found that, due to the subtle differences in various quantitative methods, the standardized values are more likely to show the characteristics of different patterns when comparing them horizontally, especially when analyzing the trade-offs, TES and EMF. Therefore, the results provided in this paper are all standardized. The initial investment of fruit trees is indeed higher than that of corn and marigolds, but this study focuses only on the ecological performance of fruit trees that have reached a stable yield period. At the same time, several fruit tree planting patterns studied in this article are also promoted by the government. The government has given different forms of subsidies such as fertilizers and saplings, and these villages with different subsidies have different quotas, which makes the analysis of the early investment very complicated. It is also difficult to be representative. On the other hand, the planting of fruit trees generally goes through the prenatal, stable, and declining periods. Therefore, a comprehensive comparison of the total cost of each pattern requires at least fifteen years of continuous observation to accurately compare the investment returns of different patterns. It may be the next direction for our team.

6. Were the fruit trees at similar production age to allow for appropriate comparison of yields? As it was mentioned in Ln 243, there was little difference in the provisioning services between corn and orange because the orange trees had not reached peak yield of the time.

Indeed, tree age is a very important influencing factor in the study of fruit trees, which will not only affect fruit yield, but also biomass. Therefore, in the process of selecting typical samples of each pattern, it is an important criterion for the fruit to reach a stable yield. In the end, the age of the plots we chose for each pattern was 5 or 6 years, except for orange. Since there are no suitable scale plots of similar age in this area. In order to avoid the influence of different climate and geological backgrounds on the analysis, we chose a 4-year-old orange orchard. That is to say, just reached the period of stable production. During the final data analysis, we found that the output was lower.

7. For reduction of topsoil loss, what is the “area of equivalent reserved land” and how is this formula getting at this particular service? How was the amount of nutrient loss calculated? Was this based on runoff and/or leaching? Similarly, how was the retention of NPK fertilizer calculated?

The equivalent land area is calculated based on the amount of soil loss, and the amount of soil loss is calculated from the soil erosion modulus. In the process of converting to equivalent land area, the average soil layer thickness of the effective cultivated land is 45 cm. The amount of nutrient loss is equal to the amount of soil loss * soil nutrient content. The fixed soil nutrient is equal to the amount of soil that each model reduces relative to off-farm or wasteland times the respective soil nutrient content.

These details have been added to appendix A.

8 How was the amount of oxygen released determined? How was carbon fixation calculated? Assume it was based on biomass but how was this done given the many different plant forms? It was mentioned that biomass was calculated based on chest diameter, height, and canopy width.  Generally, volume needs to be calculated and used with previously established allometric equations for individual species to get at biomass and then carbon storage. How was this done? Was the method focused only on aboveground biomass/carbon? How did age of the plants figure into carbon/biomass calculations? How was this standardized for a unit of area (i.e., corn plant vs pomegranate tree)?

Oxygen and carbon sequestration are calculated by the carbon tax method and the oxygen cost method. The specific formulas are:

• Carbon fixation:

U_C =1.63*C_C*A*B_y

Notes: 1.63 is the carbon sequestration coefficient; U_C is the annual carbon sequestration value (Yuan / a); C_C is the carbon fixation price (yuan / t); A is the area (ha); B is the annual stand net productivity (t / (ha · a)).

• Oxygen release:

U_O = 1.19*C_O*A*B_y

Notes: 1.19 is the oxygen release coefficient; U_O is the oxygen release value (yuan /a); C_O is the oxygen production cost (yuan /t); A and B are the same as above.

The carbon sequestration price adopts the Swedish carbon tax rate of USD 150 / t (RMB 961.275 / t at the exchange rate of USD 6.4085); The average price of industrial oxygen production was 1000 yuan/t. And the conversion coefficient between biomass and net productivity of economic forest in Yunnan was 2.576.

The biomass is based on the results of our team's previous research on different conversion patterns of farmland to forest and grassland in karst areas. The patterns were divided into trees and shrubs. The growth relation based on chest diameter and tree height was obtained by full excavation. The full excavation means to dig out and collect all the roots, stems, leaves, etc. after selecting standard wood in the plot. The specific formula is:

B_T=0.0699 *（D²_C *H）* 0.8037  r=0.8154^**

B_S=0.0463 *（D²_g* H）* 0.0791  r=0.7860^**

Notes: B_T is the biomass of trees; B_S is the biomass of shrubs; D_c is the chest diameter; D_g is the ground diameter; H is the height.

In this method, all biomass above and below ground is calculated.

The calculation method for intercropping plants is even simpler. For each plot, we selected five 1-square-meter areas and used the full excavation method to obtain their biomass. For pumpkins, five typical plants were selected and their biomass was obtained by the full excavation method. Then use the following formula to get their biomass in the plot.

B_u = B_T * A/N

Notes: B_u is the understory biomass; N is the number of pumpkins; A is the same as above.

These details have been added to appendix A.

The differences in tree age are not taken into account in the calculation. Therefore, in the process of selecting sample plots, achieving stable output is one of the first considerations, that is, about five years.

9. Regarding cultural services and education function, should the replacement cost be considered a negative or positive? Based on the text, local government was providing training for the various patterns and this cost is borne by society. Also, how was the training costs calculated for the individual patterns?

Firstly, it is difficult to directly calculate the value generated by the improvement of knowledge and technology after training, so the replacement cost method is adopted, which is a common method when estimating the value that is difficult to quantify. In this study, according to the number, form and effect of training, the local agricultural technical guidance experts classified each mode to the corresponding education level for calculation. After 15 field training sessions, for example, the average farmer's planting skills for a particular crop are equivalent to a person who has completed six years of primary school education. According to the education statistics yearbook of Yunnan province, the cost of a student's six-year primary education is known. So the value of training is quantified. In addition, this required us to count the number of training sessions for each pattern, which is one reason why we interviewed farmers, companies, cooperatives and relevant government departments.

These details have been added to appendix A.

10. Can another table be provided that shows the input data, formulas, and results for each of the seven ecosystem services based on the ten cropping patterns? For the intercropping patterns, how was the data aggregated from the herbaceous crop and woody crop? If the table cannot fit within the regular manuscript, consider it as an online supplemental table. Please provide references for the economic evaluation methods used (shadow price method, opportunity cost method, etc).

We have added the above methods to appendix A. We also provide three new references in appendix A to provide more details.

11. In terms of the intercropping systems, it is difficult to determine the value of the combinations since the crops added with the trees (soybeans, pumpkin, grass) did not have a standalone data. For instance, there was a 63% increase in provisioning services with the pumpkin and peach intercropping. Was this simply due to a yield/price of pumpkin? How would the results change if pumpkin was a standalone crop? I do appreciate Figure 3 which gets at an important comparison although I would have like to have seen soybeans, pumpkin, and grass included.

This study mainly focuses on the existing monoculture patterns and the representative intercropping patterns based on these patterns. And in the process of questionnaire survey on local farmers and field survey conducted by our research group, it was not found that these intercropping plants were planted separately on similar land conditions in a single large-scale, so such comparison was not included in this study. Of course, this is a very important point of further study, whether the growth characteristics of intercropping plants themselves have changed and to what extent, which may require more rigorous control experiments.

12. Should Table 3 also include relative ecosystem services of three intercropping patterns? It would allow for a better understanding of the primary drivers that result in the TES/EMF in Table 4.

Because the form of table cannot directly show the changes caused by intercropping, and the addition of three intercropping patterns in table 3 will lead to too many columns in this table, which is not easy to read. Therefore, the results of this part are shown in figure 2.

Author Response File: Author Response.docx

Reviewer 3 Report

The manuscript describes novel data about the Evaluation and tradeoff analysis of ecosystem service for ten agroforestry patterns in the karst region of southwest China.

The most important shortcoming is a lack of statistical analysis for all the measured parameters.

Secondly, it is compulsory to ameliorate the quality of the tables.

The are some unclear sentences and lack of references.

Specific improvement are indicate in the following list:

Line 27: “Total”

Line 99: apple or peach?

Line 104-105: scientific names are requested

Line144: “Ecosystem TradeOffs”

Table 3: Insert the standard error of the means

Table 4: as for 3

Figure 2 and 3: coloured could be better to explain differences among patterns

Line 239 - 241: statistical analysis is compulsory to affirm that

Line 246: explain better what means loosely

Line 250 - 252: in the cited references authors indicate perennial crop not PASTURE

Line 250 - 252: could be added a reference about the potential of perennial crops in reducing soil erosion respect to annual crops

Line 271-274: not clear, explain better

Line 301-302: not clear, explain better

Comments for author File: Comments.pdf

Author Response

We would like to thank you first for all the positive comments of our manuscript Manuscript ID: forests-738952) entitled “Evaluation and tradeoff analysis of ecosystem service for ten agroforestry patterns in the karst region of southwest China”. We really appreciate your help and patience.

1. The most important shortcoming is a lack of statistical analysis for all the measured parameters.

For this shortcoming, after our discussion, appendix A was added to describe the calculation method of those parameters.

2. Line 27: “Total”

Yes, we have corrected the problem. It can be seen in line 27.

3. Line 99: apple or peach?

It is “apple” in this sentence, we have corrected the corresponding area in line 100.

4. Line 104-105: scientific names are requested

We have added a Latin name to each plant. It can been seen in 104-106

5. Line144: “Ecosystem TradeOffs”

Yes, we have corrected the problem. It can be seen in line 148.

6&7 Table 3: Insert the standard error of the means; Table 4: as for 3

In the process of literature investigation, we found that, due to the subtle differences in various quantitative methods, the standardized values are more powerful to show the characteristics of different patterns when comparing them horizontally, especially when analyzing the trade-offs, TES and EMF. Therefore, the results provided in this paper are all standardized. And we cannot provide standard errors in the table. However, at your suggestion, we realized that we should provide more details about the calculation progress of these figures, so we purposely added appendix A to show more information about that.

8. Figure 2 and 3: colored could be better to explain differences among patterns

We have redrawn Figure 2 and 3. It can be seen in line 223 and 234.

9. Line 239 - 241: statistical analysis is compulsory to affirm that

In fact, this is a summary of the comparison of multiple indicators in the results section.We only modified this sentence to avoid repetition. It can be seen in line 244.

10. Line 246: explain better what means loosely

We have rewritten the sentence. It can be seen in 250.

11. Line 250 - 252: in the cited references authors indicate perennial crop not PASTURE

After rereading the references, we identified the perennial plants mentioned in this reference, mainly grasslands. To avoid misunderstanding, we modified the PASTURE to perennial grass. It can be seen in line 255-257.

12. Line 250 - 252: could be added a reference about the potential of perennial crops in reducing soil erosion respect to annual crops

According to your suggestion, we have added a reference. It can be seen in line 255-257.

13. Line 271-274: not clear, explain better

We have rewritten the sentence to make it clearer. It can be seen in line 276-278.

14. Line 301-302: not clear, explain better

We have rewritten the sentence to make it clearer. It can be seen in line 306-308.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

I appreciate the effort and thought that went into the revision of the manuscript.  

I am overall satisfied with your efforts to reframe the study as investigation of 7 monoculture patterns and 3 agroforestry patterns.  That said, I would still like to see this explicitly mentioned in the abstract. 

I appreciate the addition of Appendix A which helps address some of my questions and concerns. I still have a few questions and concerns as described below. 

Soil and nutrient loss

I understand that the equivalent land area is based on the amount of soil loss calculated from the “soil erosion modulus”.  However, there is no information on the soil loss modulus? What are the variables and equation for the soil loss modulus? Is it based on the Universal Soil Loss Equation (USLE)?  How did you standardize the soil erosion if the plots do not have underlying similar features (i.e. soil types, slopes)? Consider adding these critical details to 4 in Appendix A.  

Of course, these concerns feed into the nutrient loss which is tied directly to soil loss in your equations.  This can be problematic because one is not taking into account any leaching of nutrients. 

Carbon fixation

How was the annual stand net productivity (B) derived? I see from your equation that it is t (tons) divided by area (ha)? I understand that productivity is tied to the biomass calculation and that the conversion coefficient between biomass and net productivity of 2.576 was used. Given that net productivity changes over time for perennial crops, how was this addressed? I understand that biomass was calculated for a single point in time (5-6 years old trees).

Oxygen release:

As I understand this equation, annual stand net productivity is the surrogate for oxygen production.  So the questions around (B) also pertained to this calculation.

General comments

In the methods section, I think it would help readers if you mention that trees were of 5-6 years of age will the exception of the orange trees.

I definitely understand the challenge of providing a profit- or investment-based assessment due the complexities of subsidies, changes on fruit production over time. etc. I think a simple statement about this challenge would be adequate and followed by your statement that this may be future research direction.

I appreciate the detail provided for the cultural services. 

I still have significant concerns that comparing provisioning services between the intercropping patterns and monoculture patterns may be a false or poor comparison due to different species being used. Not really comparing equal things. I understand that you could not similar plots and this issue/drawback should be mentioned in the discussion.

Author Response

We would like to thank you again for all the comments of our manuscript Manuscript ID: forests-738952) entitled “Evaluation and tradeoff analysis of ecosystem service for ten agroforestry patterns in the karst region of southwest China”. We really appreciate your help and patience.

We have also revised the sentence about agroforestry in the abstract. It can be seen in line20-23.

1. Soil and nutrient loss

I understand that the equivalent land area is based on the amount of soil loss calculated from the “soil erosion modulus”.  However, there is no information on the soil loss modulus? What are the variables and equation for the soil loss modulus? Is it based on the Universal Soil Loss Equation (USLE)?  How did you standardize the soil erosion if the plots do not have underlying similar features (i.e. soil types, slopes)? Consider adding these critical details to 4 in Appendix A.

Of course, these concerns feed into the nutrient loss which is tied directly to soil loss in your equations.  This can be problematic because one is not taking into account any leaching of nutrients.

In fact, all the erosion modules were provided by the national positioning observation and research station of Jianshui desert ecosystem in Yunnan, college of soil and water conservation, Beijing forestry university. Since 2016, they have set up a number of runoff plots in Yunnan to monitor soil erosion, including the research samples we studied in this paper. In other words, these data are from actual measurements in the runoff plot of our study site. We used the average of 2016 and 2017 to calculate.

More importantly, these data have not yet been published, so there are no specific numbers in the previous manuscript. After further communication with the director of the observatory, he agreed that I could provide the specific values in appendix A, but it should be marked as unpublished data.

We've done that already, and it can be seen in appendix A.

As mentioned above, the research content of the biochemistry part of the various patterns will be published by the research team of Beijing forestry university after five years of sample monitoring. Therefore, nutrient leaching was not considered in this paper.

2. Carbon fixation

How was the annual stand net productivity (B) derived? I see from your equation that it is t (tons) divided by area (ha)? I understand that productivity is tied to the biomass calculation and that the conversion coefficient between biomass and net productivity of 2.576 was used. Given that net productivity changes over time for perennial crops, how was this addressed? I understand that biomass was calculated for a single point in time (5-6 years old trees).

2.576 is the conversion coefficient between the biomass and net productivity of economic forest in Yunnan recommended by Fang jingyun’s team after establishing the functional relationships between the biomass and net productivity of each forest type in different regions based on forest inventory data.

The paper also points out that the conversion coefficient of economic forest is basically constant after it matures. This means that we only need to calculate the biomass at a certain year to get the net productivity at the same time.

Therefore, we adopted 2.576 in the calculation. In addition, we have added a reference in appendix A.

3. Oxygen release:

As I understand this equation, annual stand net productivity is the surrogate for oxygen production.  So the questions around (B) also pertained to this calculation.  

In this paper, based on Fang's conclusion, we can get the net productivity in a given year, and therefore the oxygen release in a given year.

4. In the methods section, I think it would help readers if you mention that trees were of 5-6 years of age will the exception of the orange trees.

We have added this information at the end of the methods section. It can be seen in line 143-144.

5.I definitely understand the challenge of providing a profit- or investment-based assessment due the complexities of subsidies, changes on fruit production over time. etc. I think a simple statement about this challenge would be adequate and followed by your statement that this may be future research direction.

We have added this to the discussion section. It can be seen in line 293-299.

6. I appreciate the detail provided for the cultural services.

I still have significant concerns that comparing provisioning services between the intercropping patterns and monoculture patterns may be a false or poor comparison due to different species being used. Not really comparing equal things. I understand that you could not similar plots and this issue/drawback should be mentioned in the discussion. 

We have added this to the discussion section. It can be seen in line 293-299.

Reviewer 3 Report

Dear Authors,
thanks for the efforts but your answer about statistical analysis is not robust.

Your manuscript is based on the methodology proposed by Pan et al., 2013. http://dx.doi.org/10.1016/j.ecoser.2013.06.002.

As you can see in  Pan et al in 2013 manuscript, authors proposed a statistical analysis of trade-off in order to confirm their hypothesis. In your work there is a lack of crucial analysis and it is only a description of the pattern ES.

I suggest to analyse your data with statistical methods and: (i) to explain better how many replicates you have for each pattern; (ii) to explain better as you combined data from year 2016 and 2017 data; and (iii) to improve the discussion in relation to statistical analysis results.

Author Response

We would like to thank you again for all the comments of our manuscript Manuscript ID: forests-738952) entitled “Evaluation and tradeoff analysis of ecosystem service for ten agroforestry patterns in the karst region of southwest China”. We really appreciate your help and patience.

After our detailed discussion, there are two points that need to be mentioned. First, our research focuses on the differences in ecosystem services and trade-offs among various patterns in order to provide the necessary decision-making information for various stakeholders. Therefore, in our results section we mainly compare the various ecosystem services of each pattern. We also focused on the agroforestry systems currently advocated, so we compared the three monocultural systems with agroforestry systems in terms of the performance of ecosystem services. However, pan et al. focused on the spatial differences of ecosystem services, and more importantly, their influencing factors, which is why they did a lot of correlation analysis. They hoped to explore whether environmental factors or land use factors caused the impact ecosystem services through such a process. Such correlation analysis is also very applicable to the Jinghe watershed, which contains 31 counties. For our research, environmental differences are exactly what we try to avoid in the process of selecting sample plots. What we hope to compare is the differences of planting patterns in similar environments.

Second, the reason we refer to this article is the TO and TES indices introduced by pan et al. in this article. During our previous survey, we found that there are many studies on ecosystem services based on sub-categories (such as yield, fertility maintenance, etc.), but such results are too complicated in practical applications, and ordinary stakeholders do not have sufficient knowledge background to understand the findings. After the introduction of the TO index, only one value can show the overall size of the whole system trade-off. Similarly, TES can evaluate the performance of ecosystem services across the system. This makes it possible to judge the pros and cons of each planting pattern simply by comparing these two indicators.

1. to explain better how many replicates you have for each pattern;

In the first paragraph of the method section 2.2 Soil plant sampling and questionnaire survey, we have written that "three 50 m × 50 m quadrats were selected for each pattern", that is, there are three replicates of each pattern, with a total of 30 sample points.

We have modified this sentence to "three 50 m × 50 m quadrats were selected as replicates for each pattern" to make it clearer. It can be seen in line 129.

2. to explain better as you combined data from year 2016 and 2017 data

In the calculation process, in order to avoid the contingency of a single measurement and survey, we used the average value of data in 2016 and 2017 for calculation, mainly for production-related data, because there will be inter-annual fluctuations of pesticides, fertilizers and labor etc.

We have added this information in the method section 2.2. soil plant sampling and questionnaire survey.  It can be seen in line 142.

3. to improve the discussion in relation to statistical analysis results.

Based on the above explanation, we decided not to make extensive changes to the results and discussion sections to maintain the core intent of the article after our discussion. But, we add the information about Appendix B in line 301-305.

At your suggestion, we added appendix B " Correlations between pairs of ecosystem services among different planting patterns.". As expected, only four of the 21 pairs of ecosystem service functions were strongly correlated. This shows that the seven ecosystem services we selected are basically independent and can well represent all the four kinds of ecosystem services.