Environmental Inequalities in Ecosystem Services Benefits of Green Infrastructure: A Case Study from China

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Review of forests-2743696

This manuscript describes the results of a geospatial analysis of four ecosystem services in Chongqing China and the extent to which those services are correlated (or anticorrelated) with different socio-economic variables (income, housing, age, educational attainment). The objective (to determine whether ecosystem service availability differs for groups with different socio-economic status) is topical and likely to be of interest to readers. The methods, however, do not provide enough detail for the results to be interpreted with confidence. It is not always clear why or how certain analyses have been done. Furthermore, there are a variety of studies conducted in recent years that have addressed the inequitable distribution of green infrastructure resources in urban areas. The discussion section of this paper needs to frame the results of this study in that broader context so it is clear how it builds and expands upon what is presently known. Given this, I recommend rejecting the manuscript in its present form, with an opportunity to resubmit should the methods and discussion sections be improved. I have provided detail level comments below in the event that this manuscript is revised and resubmitted.

 

Abstract

It would be nice to see some of the results described in the abstract expressed numerically (how much difference in wellbeing was there between groups with different socioeconomic status? Were these results significant?)

 

Introduction

General:

I think that this study would benefit from looking at some of the work that has come out of ARIES (Artificial Intelligence for Ecosystem Services). There are quite a few ARIES studies that focus on mapping the distribution of GI benefits (particularly if GI is broadly interpreted as it is in this study). Given that this is a stated advance of this work, other studies that have trialed similar approaches should be reviewed in the introduction.

See for instance:

-          Martinez-Lopez, J. et al. 2019. Towards globally customizable ecosystem service models. Science of the Total Environment, 650, part 2, 2325-2336

-          Zank et al., 2016. Modeling the effects of urban expansion on natural capital stocks and ecosystem service flows: a case study in the Puget Sound, Washington, USA. Landscape and Urban Planning, 149, 31-42.

Detail level:

Line 32: As the sentence beginning on line 32 is presently written its not clear what the paradox is. I think this point is important and needs to be explained more thoroughly

Line 55: What kind of drawbacks? Provide an example here

Line 56: Is the last sentence of this paragraph included to show an example of a win-win situation? If so, make this explicit. Perhaps also include an example of a lose-lose situation for completeness.

Line 69: I’m not familiar with the term accessibility equity perspective. Do you mean factoring equity into accessibility assessment?

Line 74: Even though SES is defined in the abstract, you’ll want to spell it out at its first time of use in the introduction as well

Line 87-89: please provide references for this statement.

Line 89 – I’m not sure that the sentence that begins on line 89 is an example of there being a dearth of approaches to achieve multiple sustainable development goals. This probably needs to be rephrased

Line 100 – is this a review? I’m not sure that review is the word you are looking for here

 

Methods:

General:

The methods are lacking in detail. In some instances, equations are provided, but the reason behind performing the calculation is not explained in the context of the study itself (i.e., how will it be used to inform our understanding of GI accessibility or ecosystem services?). In other instances equations are not provided making it unclear how or why the analysis is being performed. Not all methods subsections are appropriately referenced

I’m concerned that some of the correlations you see occur simply because the independent variables you consider (housing, age, income, education) are themselves correlated. Did you control for this at all, perhaps pre-screening the variables using the variance inflation factor and omitting variables with a VIF >5. When this isn’t done model error is inflated and fits may be significant but not interpretable.

 

Detail level comments

Line 123: This sentence does not make sense as written. Consider rephrasing

Line 126-128 – This line indicates that images were classified as trees, grass, built environments or water and then references Fig. 1c, which does not have colors for all of these categories. What is green infrastructure  vs Recreational GI’s in Fig. 1c? How does this relate back to the tree and grass categories indicated in Line 126-128?

Line 135-138. I’m not sure what is meant by this sentence. How does the wellbeing of different SES groups factor in? Ecosystem service benefits are not the same thing as wellbeing and its not immediately clear to me how you get from mitigated UHI, recreational functions, etc to the wellbeing of different SES groups

Table 2: Please explain how evapotranspiration is being equated to managed stormwater. Evapotranspiration is a component of the urban water balance and is therefore important for stormwater management, but its not the only component that matters. Note: I see that later on in the methods (section 2.5.1), that ET is part of the managing stormwater model. This begs the question of where the other terms in the bucket model were sourced from (i.e., if ET comes from the source indicated in table 2, where did P and SF come from?)

Line 166: From what I’m reading here it seems like accessibility is being defined not by whether or not green infrastructure can be accessed (e.g., by public transit lines etc) but by whether or not green infrastructure supply is proximal enough to demand. If true, it would be helpful to make this clearer when you talk about accessibility in your introduction. I tend to think of the value you have defined as availability, not accessibility – accessibility necessitates both availability and infrastructure to facilitate access (e.g., public transit, lack of parking fees, walkable, etc.).

Equations 5 and 6 appear prior to a description of the variables in equation 4. The variables in equations 5 and 6 are presently not defined so it is not clear what these equations are for.

In section 2.4.3 it would help to have more context for why you have chosen the different models you chose. I understand that one includes a lag in y and that the other includes a spatial error term. What isn’t clear is why you feel these two corrections are likely to be important for Ecosystem service analysis. Please explain what these two model forms help us understand from an ecosystem services assessment standpoint.

In section 2.4.4 it would be helpful to contrast local regression with global regression in the context of ecosystem services (it is clear how they differ mathematically, but not why this difference matters for the kinds of variables you are interested in). How does this help us understand GI accessibility better?

In section 2.5.1 you need to provide more detail about where the terms in this model are sourced from. In the description you mention the importance of subsurface runoff and groundwater, but neither of these terms are included in the model.

Section 2.5.2. More detail is required here. No references are provided and the description is extremely minimal

Section 2.5.3. A reference to Li [14] is provided, but it would be helpful to indicate how these indices are calculated and why they should be evaluated using PCA to get a composite pattern score. What does such a composite indicate? Perhaps give an example of how the first principal component from such an analysis is typically interpreted…

 

Results:

General:

There is quite a bit of discussion and interpretation in the results. Given that there is a separate discussion section I’d suggest moving your interpretations to this section so that the results becomes more about what was found and less about what they night mean.

The figures are very hard to read (some of the font is extremely small)

If the goal is to talk about multi-benefit spaces, it would be useful to see a plot that speaks directly to multifunctionality (i.e., perhaps create a composite multifunctionality index across the 4 services you evaluate…this could be as simple as a weighted score) and then compare the spatial distribution of the index to see where the more multi-benefit systems are). You get at this some by visually comparing the different maps, but it would be easier to see if you could include a fifth set of panels that directly visualize multifunctionality.

Line 259: what is meant by benefiting the cities ecological matrix? Do you mean it benefits the environment?

Line 262: The wording of this sentence makes it difficult to understand. Perhaps rephrase.

Line 279: this is interpretation and should be moved to the discussion

The legends on Figs 2 and 3 are too small to read (please enlarge).

Table 3: are the models shown in this table the ones with the lowest AIC? Were all model types run for all variables with just these best options being shown? Please clarify in the table that these are best-fit regressions or (alternatively) provide the results of all three regressions for each ecosystem service so that readers can compare them.

Line 314: is it really older houses that mitigate UHI effects or is it just that older houses tend to have larger yards/more trees which mitigate UHI effects? This seems like a reasonable correlation, but what do you think is the actual cause? I recommend addressing this later on in the discussion.

Line 336: I’d delete the word prefer, which is subjective. All you can say is that they are more likely to live in areas with high landscape connectivity. I would also delete the sentence that says such residents may be more cognizant of the benefits (this is purely speculative).

Line 341-344 – is the implication here that stormwater management is better in low income areas? This seems counterintuitive to me.

Line 348-352: this is interpretation and belongs in the discussion

Table 4: Are the OLS models in this table the same models in Table 3? Are they really a mix of SLM and OLS models?

Line 372: please don’t use the word confirm here. We can only refute things, not confirm them. I also don’t think that you can say people prefer to reside in the suburbs or that its due to the increased ES benefits they receive. All you can say is that education is positively associated with ES benefits in the suburbs. Its not appropriate to speculate as to why unless you conducted human subjects surveys that actually reveal people’s stated preferences.

Lines 374-379 belong in the discussion and would require supporting references if you choose to keep them.

Figure 4: legends are too small to read. Its unclear if the red to blue scale has the same interpretation on each plot (i.e., are the colors strictly comparable or is the color map scaled differently in each plot)

 

Discussion

General:

I’m seeing multiple instances where correlations between SES variables and ecosystem services are being used to infer people’s preferences. People don’t always live where they would prefer to live (even high-income individuals – maybe they selected their house based on a school district and it has nothing to do with ES at all), so these associations seem suspect. I’d prefer to see more emphasis on what people have access to and how access might be higher or lower for different SES groups. That is something that can be readily inferred from the data. Preferences, not so much.

The discussion is very brief. I’d like to see it expanded, focusing on comparing the insights gained here to insights from other regions. What has been seen in other countries (e.g., how do your results compare to the studies from Table 1?). What are the pros and cons of implementing different green infrastructure forms (trees, grasses, pocket parks, etc) given your results? Do you have any recommendations for how to foster urban greenspaces that are more multifunctional? Recommendations for ensuring ES supply matches up with demand better in urban areas? Addressing all of these is not needed, but expanding the discussion to address one or two of these areas would make the implications of your findings clearer.

It would also be nice to see a limitations section incorporated into your discussion.

 

Detail level comments:

Line 393. Again, statistically speaking you cannot confirm a hypothesis….you can only reject one. Please use different words here.

Line 395: Delete the s at the end of determinants.

Line 406: I don’t think you show that ES benefits are prioritized by higher income groups. Its impossible to say for sure why people live where they live….it might not be about ES at all…

Lines 402-412: much of this is repeated from your discussion. Delete it from the results. I’m also still not clear how you’ve determined that more educated individuals recognize the latent benefits of urban vegetation (proximity and access are not the same as preference).

Comments on the Quality of English Language

I've indicated the sections where I thought that sentence structure was confusing in the comments provided above

Author Response

Thank you for giving us an opportunity to revise our manuscript, we appreciate editor and reviewers very much for their positive and constructive comments and suggestions on our manuscript. We have studied the comments carefully and have made revision in this paper. Please refer to the attachment for detailed response and modification information.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript This study tested a set of social-environmental benefits to evaluate green infrastructure’s ecosystem services such as mitigation of the urban heat island (UHI) effect, recreational functions, enhanced landscape connectivity, and efficient stormwater management in Chongqing, China. It also analyzed the correlations between GI's ecosystem services and socioeconomic factors. The study helps us better understand this issue in developing countries and the findings can be used for diverse stakeholders. Overall, the study is well-designed and the findings are well-presented. However, the manuscript can be improved by doing the following: 

• Title: as this is a case study in Chongqing, the authors might want to clarify that in the title to make it clear.  

• Abstract: The authors need to add more details to the findings (e.g., numbers/facts).  

• Keywords: “Green infrastructure” and “Ecosystem services” were already mentioned in the title, the authors might want to replace them with other keywords.  

• 1. Introduction: Some of the text (e.g., Ls82-112) should be moved to “Materials and Methods” section; more citations about environmental justice should be added; the hypotheses of the study should be added.  

• 2. Materials and Methods: Statistical analysis methods should be added.  

• 4 Discussion: The authors might want to compare the findings with other cities/similar studies to deep a bit deeper.

Comments on the Quality of English Language

Might want to ask a native English speaker to proofread the manuscript. 

Author Response

Thank you for giving us an opportunity to revise our manuscript, we appreciate editor and reviewers very much for their positive and constructive comments and suggestions on our manuscript. We have studied the comments carefully and have made revision in this paper. Please refer to the attachment for detailed response and modification information.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors Very interesting topic of research and very well done. The work is well conceived, the methods are well explained, the literature used corresponds to the topic of the work.

Author Response

Thank you very much for your valuable feedback on our research. We are delighted to hear that you found the topic interesting and the work well done. Your positive comments on the conception, methodology, and alignment with the relevant literature are truly encouraging.

We appreciate your time and thoughtful evaluation, and we will carefully consider any additional suggestions or recommendations you may have to further enhance the quality of our work.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors responses have addressed my previous concerns about the manuscript's methods and results

Author Response

I appreciate your feedback and I'm glad to hear that the authors' responses have adequately addressed your concerns regarding the manuscript's methods and results.

Thank you for your time and consideration.

Reviewer 2 Report

Comments and Suggestions for Authors

Just some minor comments/suggestions for consideration:

-Title: capitalization needs to be consistent, so it should be "Environmental Inequalities in Ecosystem Services Benefits of Green Infrastructure: A Case Study from China". 

-Figures 2 and 3: some text in the figures is not showing properly, please fix them. 

-Green infrastructure (GI): The abbreviation of green infrastructure has been used several times throughout the manuscript. The authors only needs to do that for the first time of mentioning, please check and fix it. 

Author Response

Thank you for giving us an opportunity to revise our manuscript, we appreciate editor and reviewers very much for their positive and constructive comments and suggestions on our manuscript. We have studied the comments carefully and have made revision in this paper. Please refer to the attachment for detailed response and modification information.

Author Response File: Author Response.pdf