the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 03 Sep 2024
A Multiagent Socio-hydrologic Framework for Integrated Green Infrastructures and Water Resource Management at Various Spatial Scales
Abstract. Green infrastructures have been widely used to manage urban stormwater, especially in water-stressed regions. They also pose new challenges to urban and watershed water resources management. This paper focuses on the green infrastructure-induced dynamics of water sharing in a watershed from three spatial scales. A multiagent socio-hydrologic model framework is developed to provide an optimization-simulation method for city-, inter-city-and watershed-scale integrated green infrastructures and water resource management (IGWM) that comprehensively considers the watershed circumstances-, the urban water managers-and the watershed manager-urban water managers interactions. We apply the framework to conduct three simulating experiments in the Upper Mississippi River Basin, the US. Four patterns in city-scale IGWM are classified and two dynamics of cost and equity in inter-city- and watershed-scale IGWM are characterized through various sensitivity, scenario, and comparative analyses. The modeling results could advance our understanding of the role of green infrastructures in urban and watershed water resources management and assist water managers in making associated decisions.
- Preprint
(5737 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on hess-2024-232', Cyndi Vail Castro, 23 Sep 2024
The overall paper is an important contribution to the science of decision-making with GIs across spatial scales. There are a few conceptual questions I had that I was not able to follow in the paper, but it appears the work is there. A bit further clarification would be helpful, and in that case, I would accept the manuscript for publication.
Please see below a few comments:
- The grammar and writing style is excellent. I do not have any major comments on the technical writing. However, while I understand the rationale for doing so, the many use of acronyms reads to me as confusing. Several of these acronyms mean essentially the same thing at the decision-making scale. Perhaps consider condensing the number of acronyms if feasible? Not necessary, it just was hard for me to follow. I see the graphic in Figure 1, which does help explain this concept a bit across the 3 spatial scales, but it is still hard to follow when reading the introduction. Perhaps re-state the acronyms in Fig. 1 caption and use all in the graphic? (e.g., UWB, SM are missing).
- Very minor corrections noted here:
- Line 116: "a agent-based framework" should be "an agent-based framework"
- Line 159: "dynamic(s) of (the) watershed"?
- Figure 1: Should the text near WM Agent state "Bi-level multiagent system"?
- Line 578: Period instead of ;?
- The socio-hydrological application at various spatial scales applies to GIs in any urban community, not just alongside rivers. Perhaps consider re-phrasing the references to how GIs interact with hydrology near river networks.
- Most of the mentionings of the "socio" part of GIs being part of a socio-hydrological system in the introduction refer to housing types or anthropogenic activities in upstream portions of the drainage basin. While this is true, the socio component of GI systems extends far beyond these considerations and might be worth a mention. For example, how GIs become community recreational meeting spots, have been shown to reduce human health issues, improve mental well-being, provide urban sources of food, improve heat island effects, reduce noise, etc. Or, are you mostly referring to the "socio" component being the complex decision-making required? This was a bit confusing to me.
- Conceptually, while I agree that the planning paradigm described at the city-level in Figure 1 is ideally how city-scale GIs should be constructed, in practice, I do not think this is happening. Instead, due to the long timeframe associated with stormwater funding and construction, GIs tend to be developed sporatically on a project-by-project basis, not in real-time, tightly coupled with the water use and demand, as depicted here. Although I understand that in order to model this as an ABM, you had to make such a coupling decision, perhaps consider a time delay in the model, or mention this limitation in reflecting real-world decision-making patterns. This also applies to the hydrologic connections in Fig 1B, between each urban area linked by real-time riverine flows between them. The city-scale decision-making of GI does not align temporally with the hydrologic flows of connecting river systems, even though the time steps appear to be monthly.
- Another conceptual question I have - the overall model framework depends on GIs being used at-large for water storage and demand. Perhaps this is common in some parts of the world, but in the US, where the case study is conducted, most GIs are used for runoff abatement, which is then linked back into the greywater infrastructure system and sent offsite to reduce flooding issues. Some rainwater harvesting systems are used for on-site capture and use for irrigation, but these are very small-scale in nature (like someone's personal lawn) and not designed systematically to be a major contributor to widespread irrigation needs. At least I am not aware of this being common practice in the US.
- It seems to me that the concept being simulated is actually systematic decision-making for detention-pond and reservoir storage at both the city watershed scales, and how inter-city feedbacks can impact the overall cycle. Which is not necessarily "Green Infrastructure" as I understand it to be used in the literature and community.
- Perhaps this is me not understanding the Markov property, but it seems counter-intuitive to apply this property, essentially stating the decision-making process is stochastic and its future evolution is completely independent of its history, as the opposite paradigm is key to a system operating "socio-hydrologically".
- Case Study:
- This is an extremely large watershed area for designing with GI. Was this selected at random, or is there a decision-making entity that manages this trans-state watershed systematically? I looked up the UMRBA association, and they don't seem to actively manage water supply and use in this basin.
- I looked at the 1 citation mentioned for using GI in this geographical region (Askey-Merwin, 2020), and this publication addresses mitigating flooding, not storing rainwater via GIs and re-use in widespread irrigation projects. Moreoever, the Mississippi River is one of the highest-flow rivers in the US and has a complex network of laws and regulations regarding water extraction for municipal or industrial use. I am not aware of water quantity being an issue here for irrigation purposes, so I am confused why there is a study suggesting GI is being actively proposed, managed, and constructed at the watershed scale and the city scales in real-time to ensure water availability here. If it is purely for a conceptual purpose of explaining complexities of GI planning in general, that is fine, but in that instance, I wouldn't necessarily limit the model to connecting urban communities along a main river stem, as this limits the application substantially. However, the model is already built under these assumptions, so I do not recommend re-designing. I am just pointing out it conceptually is difficult for me to see the application outside of this theoretical explorative study.
- Where are you getting water demand data for irrigation, and how does this change over time? What factors drive this in the model? I see that you simulated "urban" water demand via population and urban layouts, but as mentioned, this is a tiny percentage of the overall water resources in the basin, and is likely to be impacted significantly by urban-scale GI units. I see some mentioning of irrigation demand in Eq A10, but it is unclear to me what these equations mean or how the underlying data were gathered. Is the irrigation a basis of cropland type?
- I am not qualified to review the set-up of the ABM model, particle optimization schemes, or economic theory choices. Please ensure one of the other reviewers has this expertise and can comment on the methodology.
- I do not see where the channel geometry is used for the Muskingum-Cunge routing method.
- I am not following how USGS stations for the Mississippi river could be used to calibrate urban water use.
- The discussion of results is very convoluted. Perhaps consider a concluding paragraph with bullet-points of the main take-aways that can be widely applied. For example, key insights about water costs and conflicts among adjacent communities, the importance of communication of watershed-scale managers and city-scale planners, the impact of assuming agent rationality / Markov property / Stackelberg in understanding the overall socio-hydro dynamics, what this means on overall water policy as GIs become more popular, interaction between urban and irrigation water use and demand, what this study informed us about social equity in water decisions, etc.
- The Conclusions section is too convoluted for me. I recommend removing a lot of the technical jargon and focusing on the bigger picture here.
Citation: https://doi.org/10.5194/hess-2024-232-RC1 -
AC1: 'Reply on RC1', Mengxiang Zhang, 29 Oct 2024
Dear Reviewer,
Thank you for your valuable comments and suggestions, which will undoubtedly strengthen our study and enhance the readability and comprehensibility of our paper. We will address your comments and those of the other reviewers in detail and will make the necessary revisions to our manuscript.
Please find our detailed responses to your comments item by item in the attached supplement document. Should you have any further comments or suggestions, please do not hesitate to share them with us via the HESS discussion system.Thank you once again for your insightful feedback.
Sincerely,
Ting Fong May Chui and Mengxiang Zhang
-
RC2: 'Comment on hess-2024-232', Anonymous Referee #2, 28 Oct 2024
This manuscript discusses an interesting question in urban water management where green infrastructure is integrated into multiple cities’ water supply operations. The authors approach this problem by developing an agent-based modeling (ABM) framework to discuss cross-scale interactions among city and watershed water managers (i.e., agents) and explore water equity and policy implication through imposing a penalty for overdraft.
I appreciate the authors ambition to take on the challenge of solving the complex urban water problem and efforts in developing an integrated modeling tool. The introduction effectively highlights the importance of integrated water management at watershed scale and the need for integrated modeling approaches. I believe the scope of this study will be of great interest to the HESS community. That being said, the current manuscript suffers several major flaws that make it difficult to follow and obscure its contributions and intellectual merits. I am fully committed to helping elevate the quality of this manuscript. If any comments arise from my lack of knowledge on specific points, please accept my apologies in advance. Below are the summaries of my comments and suggestions followed by the specific comments.
My first comment is about the writing. The current manuscript is difficult to follow due to the excessive technical terms and ambiguous language. For example, IGWM (short for integrated green infrastructure and water resources management) was applied in describing models, agents, and agents’ decisions, which gave me a headache. Other examples include WM (water manager), UWM (urban water manager), and HUWS (hybrid urban water system). Some technical terms are not well-defined. For example, ABM (agent-based model) and MAS (multi-agent system) are often used interchangeably in the literature, but it was presented in this manuscript as two distinct modeling approaches applied for building two models (i.e., city-scale and inter-city). Similarly, rainwater and stormwater are the same thing to me, and yet they are listed as two water sources (lines 183-184). In the results section, the authors discuss the usage of the four water sources (surface water, groundwater, stormwater, and rainwater), so the model must have simulated the water sources. However, I could not find their definitions nor how the water supply portfolio is simulated except for surface water. I will recommend a more rigorous quality control and assurance to improve the flow and readability.
My second major comment is about the framing of the model. After reading the method section a few times, the modeling components become clear to me. The modeling framework includes three models coupling together. However, it is essentially one agent-based model with two agent types (city agents, UWM, and a watershed agent, WM) and a hydrologic model (including UWB-SM and M-C) representing the spatial connections among the city agents and the watershed environment. I can understand the authors’ intention in examining the interactions among agents across multiple sales; however, framing the models separately at different scales has had the opposite effect for me, leaving me confused and obscuring my understanding of the study. The suggestion here may be somehow subjective, but I am hoping that the manuscript could benefit from my perspective.
Generally, I found it difficult to follow the results and discussion, partly attributed to not fully understanding the modeling components. Since I did not go through all the details in the Appendices, I was not sure whether I could agree or disagree with the findings and discussion. I will suggest presenting the key components of the models in the main text. For example, in lines 297 – 298, the authors mentioned an assessment index (Gini coefficient) set by the WM agent but did not go any further to explain how it was incorporated into WM’s decision-making nor describe what the Gini coefficient means and how it is calculated. I feel the results and discussion can be condensed to focus on key findings as a long discussion would lose its audience. Another suggestion is to provide more details of the urban water balance model (UWB-SM) in the main text as the model of the physical environment since it is where the water partition is determined. Contrarily, the texts related to the routing model (M-C method) and solution approach (S-APSO) can be moved to the Appendix. Is the S-APSO approach the original creation of this work? If so, I think the solution approach as well as the UWM model could be a separate paper.
Overall, this manuscript has the potential to be a high-quality paper (by the modeling framework itself) if the authors can improve the clarity in the methodology, experiment designs, and discussion and highlight its contributions.
Specific Comments
- Lines 27-34: The introduction highlights the need for multi-scale green infrastructure frameworks in urban water management. The introduction needs to provide a detailed positioning within recent literature and how the current research contributes to the body of knowledge. Integrating findings from recent studies on similar frameworks could help contextualize their suggested approach within the broader field and clarify its unique contributions. Suggestion: Expand the literature review to include recent ABM applications in socio-hydrology and water resources.
- Line 15–20: Add brief mention of the specific experimental scenarios (e.g., “streamflow penalty” and GI adoption) to give readers a clearer picture of the paper’s approach and key findings at the outset. This will make the abstract more informative for readers skimming the content.
- Line 35–40: The statement on the importance of GIs could be made more impactful by adding specific challenges (e.g., “urban flooding, groundwater depletion, and inter-city water conflicts”) that this framework aims to address. This would help sharpen the focus on the practical problems the model intends to resolve.
- Ensure that acronyms such as “GI” (for Green Infrastructure) and “UWM” (for Urban Water Manager) are consistently defined and used throughout the text. For instance, Line 42 introduces GI without explicitly defining it, which may confuse readers unfamiliar with the abbreviation.
- Specific terms, such as "hydrologic regime" and "multiagent system," are used inconsistently. A brief definition of these terms early in the manuscript (in the Introduction or Methods) would improve consistency.
- The methodology section presents a layered framework with urban and watershed scales involving socio-economic and hydrologic variables. However, the description of how these scales is integrated within a complex system would benefit from additional detail and clarity.
- Lines 170-182: The agent-based modeling (ABM) setup could be explained more systematically. Clarifying the assumptions behind each agent's decision-making process, especially for UWMs and watershed managers, would make the model's structure more understandable. Additionally, line 175 references the "Markov property," but a brief explanation or contextualization within the model would benefit readers unfamiliar with this concept.
- Line 250 briefly mentions historical hydrologic data without indicating the data sources, calibration metrics, or validation techniques. Include a clear explanation of the calibration and validation processes. A summary table with parameter ranges, calibration techniques, and validation outcomes would strengthen the model's reliability and replicability.
- Line 280–285: The hydrologic and socio-economic data sources description is somewhat broad. Including a brief list of specific datasets used, such as U.S. Geological Survey data or climate records, and their date ranges would clarify the model's foundation. The manuscript presents three spatial scales (city, inter-city, and watershed) for experimental analysis, focusing on GI policies. However, these scenarios are presented with minimal contextual detail.
- Lines 315-327: Discussing the experimental conditions would help explain why specific scenarios were chosen, such as the "streamflow penalty" policy in line 319. A description of how this penalty reflects real-world practices would better convey the practical relevance of this scenario.
- Lines 390-420: This section would be more accessible if the results for each spatial scale (city, inter-city, watershed) were divided into distinct subsections rather than being presented together. This would help readers understand the unique impacts observed at each scale.
- Lines 460-475: While the discussion briefly mentions the potential impacts of GI policies, it could provide more concrete suggestions for policymakers, especially regarding implementing penalty-based policies. For instance, specifying how such policies could be enforced across jurisdictions or considering potential limitations would strengthen the section.
- Lines 490-500: This discussion could explore the model's adaptability to other similar regions or hydroclimatic conditions and cross-case comparisons. The authors could broaden the study's relevance by highlighting how it might apply to other areas. Expand the discussion on the policy implications of GIs, considering practical challenges and enforcement strategies. Including recommendations for policymakers, such as adaptive management guidelines or climate-resilient infrastructure planning, would enhance the study's applicability.
- Lines 520-530: The conclusion summarizes the key findings well but could further emphasize the study's contributions and the potential for broader application. Highlight how the study advances the field of socio-hydrologic modeling, specifically regarding multi-agent frameworks for GI integration. A concluding sentence on how this framework could guide future studies in water management would leave a stronger impression.
Technical corrections
- Minor grammatical issues and ambiguous phrases appear throughout the text. For example, line 175, "up-and downstream imbalances," could be clarified as "upstream-downstream imbalances." A thorough proofreading would enhance readability.
- Figure 1 provides a schematic of the model, but its components and interconnections need to be labeled clearly. A legend or detailed figure description indicating each component's function within the model would enhance interpretability. Improvement Suggestion: Include a step-by-step description or flowchart illustrating the interactions between socio-economic factors, hydrologic processes, and policy influences. This would help in clarifying the multi-agent interactions and coupling between scales.
- Figures 2 and 5: These figures would benefit from concise captions that specify what variables or trends they are intended to show. For instance, state whether they display policy impacts, flow distributions, or demand-supply imbalances explicitly.
- Figure 3: While Figure 3 illustrates IGWM patterns for UWMs, more context on the visualized policy implications and the decision-making dynamics among UWMs would make the figures more impactful. Additionally, increasing the color contrast between scenarios in this figure would improve readability. Improvement Suggestion: Provide a rationale for each experimental scenario, focusing on its real-world applications. Consider adding a flowchart or table summarizing the experimental setups and their objectives to help readers follow the study's design.
- Figure 4: The information presented in Figure 4 lacks sufficient labeling to identify different variables. Clear labels or a more detailed caption would clarify how the results vary by scale and policy. Improvement Suggestion: Reorganize the results section by scale and add clear interpretations of findings about policy scenarios. Additional labels and color coding in figures, especially Figures 3 and 4, would improve clarity. Why are the lines zigzagging?
- Figures 3 and 4: Improve color contrast and include labels for specific variables to make visualizations easier to interpret. Including a note explaining the data or variables displayed in each figure would enhance accessibility.
- Line 367: Typo ‘experiment’
- Line 381: What are rimax, rrmax, and rsmax ?
- Line 430: References?
Citation: https://doi.org/10.5194/hess-2024-232-RC2 -
AC2: 'Reply on RC2', Mengxiang Zhang, 27 Nov 2024
Dear Reviewer,
Thank you for your valuable comments and suggestions, which will undoubtedly strengthen our study and enhance the readability and comprehensibility of our paper. We will address your comments and those of the other reviewers in detail and will make the necessary revisions to our manuscript.
Please find our detailed responses to your comments item by item in the attached document. Should you have any further comments or suggestions, please do not hesitate to share them with us via the HESS discussion system.Thank you once again for your insightful feedback.
Sincerely,
Ting Fong May Chui and Mengxiang Zhang
Status: closed
-
RC1: 'Comment on hess-2024-232', Cyndi Vail Castro, 23 Sep 2024
The overall paper is an important contribution to the science of decision-making with GIs across spatial scales. There are a few conceptual questions I had that I was not able to follow in the paper, but it appears the work is there. A bit further clarification would be helpful, and in that case, I would accept the manuscript for publication.
Please see below a few comments:
- The grammar and writing style is excellent. I do not have any major comments on the technical writing. However, while I understand the rationale for doing so, the many use of acronyms reads to me as confusing. Several of these acronyms mean essentially the same thing at the decision-making scale. Perhaps consider condensing the number of acronyms if feasible? Not necessary, it just was hard for me to follow. I see the graphic in Figure 1, which does help explain this concept a bit across the 3 spatial scales, but it is still hard to follow when reading the introduction. Perhaps re-state the acronyms in Fig. 1 caption and use all in the graphic? (e.g., UWB, SM are missing).
- Very minor corrections noted here:
- Line 116: "a agent-based framework" should be "an agent-based framework"
- Line 159: "dynamic(s) of (the) watershed"?
- Figure 1: Should the text near WM Agent state "Bi-level multiagent system"?
- Line 578: Period instead of ;?
- The socio-hydrological application at various spatial scales applies to GIs in any urban community, not just alongside rivers. Perhaps consider re-phrasing the references to how GIs interact with hydrology near river networks.
- Most of the mentionings of the "socio" part of GIs being part of a socio-hydrological system in the introduction refer to housing types or anthropogenic activities in upstream portions of the drainage basin. While this is true, the socio component of GI systems extends far beyond these considerations and might be worth a mention. For example, how GIs become community recreational meeting spots, have been shown to reduce human health issues, improve mental well-being, provide urban sources of food, improve heat island effects, reduce noise, etc. Or, are you mostly referring to the "socio" component being the complex decision-making required? This was a bit confusing to me.
- Conceptually, while I agree that the planning paradigm described at the city-level in Figure 1 is ideally how city-scale GIs should be constructed, in practice, I do not think this is happening. Instead, due to the long timeframe associated with stormwater funding and construction, GIs tend to be developed sporatically on a project-by-project basis, not in real-time, tightly coupled with the water use and demand, as depicted here. Although I understand that in order to model this as an ABM, you had to make such a coupling decision, perhaps consider a time delay in the model, or mention this limitation in reflecting real-world decision-making patterns. This also applies to the hydrologic connections in Fig 1B, between each urban area linked by real-time riverine flows between them. The city-scale decision-making of GI does not align temporally with the hydrologic flows of connecting river systems, even though the time steps appear to be monthly.
- Another conceptual question I have - the overall model framework depends on GIs being used at-large for water storage and demand. Perhaps this is common in some parts of the world, but in the US, where the case study is conducted, most GIs are used for runoff abatement, which is then linked back into the greywater infrastructure system and sent offsite to reduce flooding issues. Some rainwater harvesting systems are used for on-site capture and use for irrigation, but these are very small-scale in nature (like someone's personal lawn) and not designed systematically to be a major contributor to widespread irrigation needs. At least I am not aware of this being common practice in the US.
- It seems to me that the concept being simulated is actually systematic decision-making for detention-pond and reservoir storage at both the city watershed scales, and how inter-city feedbacks can impact the overall cycle. Which is not necessarily "Green Infrastructure" as I understand it to be used in the literature and community.
- Perhaps this is me not understanding the Markov property, but it seems counter-intuitive to apply this property, essentially stating the decision-making process is stochastic and its future evolution is completely independent of its history, as the opposite paradigm is key to a system operating "socio-hydrologically".
- Case Study:
- This is an extremely large watershed area for designing with GI. Was this selected at random, or is there a decision-making entity that manages this trans-state watershed systematically? I looked up the UMRBA association, and they don't seem to actively manage water supply and use in this basin.
- I looked at the 1 citation mentioned for using GI in this geographical region (Askey-Merwin, 2020), and this publication addresses mitigating flooding, not storing rainwater via GIs and re-use in widespread irrigation projects. Moreoever, the Mississippi River is one of the highest-flow rivers in the US and has a complex network of laws and regulations regarding water extraction for municipal or industrial use. I am not aware of water quantity being an issue here for irrigation purposes, so I am confused why there is a study suggesting GI is being actively proposed, managed, and constructed at the watershed scale and the city scales in real-time to ensure water availability here. If it is purely for a conceptual purpose of explaining complexities of GI planning in general, that is fine, but in that instance, I wouldn't necessarily limit the model to connecting urban communities along a main river stem, as this limits the application substantially. However, the model is already built under these assumptions, so I do not recommend re-designing. I am just pointing out it conceptually is difficult for me to see the application outside of this theoretical explorative study.
- Where are you getting water demand data for irrigation, and how does this change over time? What factors drive this in the model? I see that you simulated "urban" water demand via population and urban layouts, but as mentioned, this is a tiny percentage of the overall water resources in the basin, and is likely to be impacted significantly by urban-scale GI units. I see some mentioning of irrigation demand in Eq A10, but it is unclear to me what these equations mean or how the underlying data were gathered. Is the irrigation a basis of cropland type?
- I am not qualified to review the set-up of the ABM model, particle optimization schemes, or economic theory choices. Please ensure one of the other reviewers has this expertise and can comment on the methodology.
- I do not see where the channel geometry is used for the Muskingum-Cunge routing method.
- I am not following how USGS stations for the Mississippi river could be used to calibrate urban water use.
- The discussion of results is very convoluted. Perhaps consider a concluding paragraph with bullet-points of the main take-aways that can be widely applied. For example, key insights about water costs and conflicts among adjacent communities, the importance of communication of watershed-scale managers and city-scale planners, the impact of assuming agent rationality / Markov property / Stackelberg in understanding the overall socio-hydro dynamics, what this means on overall water policy as GIs become more popular, interaction between urban and irrigation water use and demand, what this study informed us about social equity in water decisions, etc.
- The Conclusions section is too convoluted for me. I recommend removing a lot of the technical jargon and focusing on the bigger picture here.
Citation: https://doi.org/10.5194/hess-2024-232-RC1 -
AC1: 'Reply on RC1', Mengxiang Zhang, 29 Oct 2024
Dear Reviewer,
Thank you for your valuable comments and suggestions, which will undoubtedly strengthen our study and enhance the readability and comprehensibility of our paper. We will address your comments and those of the other reviewers in detail and will make the necessary revisions to our manuscript.
Please find our detailed responses to your comments item by item in the attached supplement document. Should you have any further comments or suggestions, please do not hesitate to share them with us via the HESS discussion system.Thank you once again for your insightful feedback.
Sincerely,
Ting Fong May Chui and Mengxiang Zhang
-
RC2: 'Comment on hess-2024-232', Anonymous Referee #2, 28 Oct 2024
This manuscript discusses an interesting question in urban water management where green infrastructure is integrated into multiple cities’ water supply operations. The authors approach this problem by developing an agent-based modeling (ABM) framework to discuss cross-scale interactions among city and watershed water managers (i.e., agents) and explore water equity and policy implication through imposing a penalty for overdraft.
I appreciate the authors ambition to take on the challenge of solving the complex urban water problem and efforts in developing an integrated modeling tool. The introduction effectively highlights the importance of integrated water management at watershed scale and the need for integrated modeling approaches. I believe the scope of this study will be of great interest to the HESS community. That being said, the current manuscript suffers several major flaws that make it difficult to follow and obscure its contributions and intellectual merits. I am fully committed to helping elevate the quality of this manuscript. If any comments arise from my lack of knowledge on specific points, please accept my apologies in advance. Below are the summaries of my comments and suggestions followed by the specific comments.
My first comment is about the writing. The current manuscript is difficult to follow due to the excessive technical terms and ambiguous language. For example, IGWM (short for integrated green infrastructure and water resources management) was applied in describing models, agents, and agents’ decisions, which gave me a headache. Other examples include WM (water manager), UWM (urban water manager), and HUWS (hybrid urban water system). Some technical terms are not well-defined. For example, ABM (agent-based model) and MAS (multi-agent system) are often used interchangeably in the literature, but it was presented in this manuscript as two distinct modeling approaches applied for building two models (i.e., city-scale and inter-city). Similarly, rainwater and stormwater are the same thing to me, and yet they are listed as two water sources (lines 183-184). In the results section, the authors discuss the usage of the four water sources (surface water, groundwater, stormwater, and rainwater), so the model must have simulated the water sources. However, I could not find their definitions nor how the water supply portfolio is simulated except for surface water. I will recommend a more rigorous quality control and assurance to improve the flow and readability.
My second major comment is about the framing of the model. After reading the method section a few times, the modeling components become clear to me. The modeling framework includes three models coupling together. However, it is essentially one agent-based model with two agent types (city agents, UWM, and a watershed agent, WM) and a hydrologic model (including UWB-SM and M-C) representing the spatial connections among the city agents and the watershed environment. I can understand the authors’ intention in examining the interactions among agents across multiple sales; however, framing the models separately at different scales has had the opposite effect for me, leaving me confused and obscuring my understanding of the study. The suggestion here may be somehow subjective, but I am hoping that the manuscript could benefit from my perspective.
Generally, I found it difficult to follow the results and discussion, partly attributed to not fully understanding the modeling components. Since I did not go through all the details in the Appendices, I was not sure whether I could agree or disagree with the findings and discussion. I will suggest presenting the key components of the models in the main text. For example, in lines 297 – 298, the authors mentioned an assessment index (Gini coefficient) set by the WM agent but did not go any further to explain how it was incorporated into WM’s decision-making nor describe what the Gini coefficient means and how it is calculated. I feel the results and discussion can be condensed to focus on key findings as a long discussion would lose its audience. Another suggestion is to provide more details of the urban water balance model (UWB-SM) in the main text as the model of the physical environment since it is where the water partition is determined. Contrarily, the texts related to the routing model (M-C method) and solution approach (S-APSO) can be moved to the Appendix. Is the S-APSO approach the original creation of this work? If so, I think the solution approach as well as the UWM model could be a separate paper.
Overall, this manuscript has the potential to be a high-quality paper (by the modeling framework itself) if the authors can improve the clarity in the methodology, experiment designs, and discussion and highlight its contributions.
Specific Comments
- Lines 27-34: The introduction highlights the need for multi-scale green infrastructure frameworks in urban water management. The introduction needs to provide a detailed positioning within recent literature and how the current research contributes to the body of knowledge. Integrating findings from recent studies on similar frameworks could help contextualize their suggested approach within the broader field and clarify its unique contributions. Suggestion: Expand the literature review to include recent ABM applications in socio-hydrology and water resources.
- Line 15–20: Add brief mention of the specific experimental scenarios (e.g., “streamflow penalty” and GI adoption) to give readers a clearer picture of the paper’s approach and key findings at the outset. This will make the abstract more informative for readers skimming the content.
- Line 35–40: The statement on the importance of GIs could be made more impactful by adding specific challenges (e.g., “urban flooding, groundwater depletion, and inter-city water conflicts”) that this framework aims to address. This would help sharpen the focus on the practical problems the model intends to resolve.
- Ensure that acronyms such as “GI” (for Green Infrastructure) and “UWM” (for Urban Water Manager) are consistently defined and used throughout the text. For instance, Line 42 introduces GI without explicitly defining it, which may confuse readers unfamiliar with the abbreviation.
- Specific terms, such as "hydrologic regime" and "multiagent system," are used inconsistently. A brief definition of these terms early in the manuscript (in the Introduction or Methods) would improve consistency.
- The methodology section presents a layered framework with urban and watershed scales involving socio-economic and hydrologic variables. However, the description of how these scales is integrated within a complex system would benefit from additional detail and clarity.
- Lines 170-182: The agent-based modeling (ABM) setup could be explained more systematically. Clarifying the assumptions behind each agent's decision-making process, especially for UWMs and watershed managers, would make the model's structure more understandable. Additionally, line 175 references the "Markov property," but a brief explanation or contextualization within the model would benefit readers unfamiliar with this concept.
- Line 250 briefly mentions historical hydrologic data without indicating the data sources, calibration metrics, or validation techniques. Include a clear explanation of the calibration and validation processes. A summary table with parameter ranges, calibration techniques, and validation outcomes would strengthen the model's reliability and replicability.
- Line 280–285: The hydrologic and socio-economic data sources description is somewhat broad. Including a brief list of specific datasets used, such as U.S. Geological Survey data or climate records, and their date ranges would clarify the model's foundation. The manuscript presents three spatial scales (city, inter-city, and watershed) for experimental analysis, focusing on GI policies. However, these scenarios are presented with minimal contextual detail.
- Lines 315-327: Discussing the experimental conditions would help explain why specific scenarios were chosen, such as the "streamflow penalty" policy in line 319. A description of how this penalty reflects real-world practices would better convey the practical relevance of this scenario.
- Lines 390-420: This section would be more accessible if the results for each spatial scale (city, inter-city, watershed) were divided into distinct subsections rather than being presented together. This would help readers understand the unique impacts observed at each scale.
- Lines 460-475: While the discussion briefly mentions the potential impacts of GI policies, it could provide more concrete suggestions for policymakers, especially regarding implementing penalty-based policies. For instance, specifying how such policies could be enforced across jurisdictions or considering potential limitations would strengthen the section.
- Lines 490-500: This discussion could explore the model's adaptability to other similar regions or hydroclimatic conditions and cross-case comparisons. The authors could broaden the study's relevance by highlighting how it might apply to other areas. Expand the discussion on the policy implications of GIs, considering practical challenges and enforcement strategies. Including recommendations for policymakers, such as adaptive management guidelines or climate-resilient infrastructure planning, would enhance the study's applicability.
- Lines 520-530: The conclusion summarizes the key findings well but could further emphasize the study's contributions and the potential for broader application. Highlight how the study advances the field of socio-hydrologic modeling, specifically regarding multi-agent frameworks for GI integration. A concluding sentence on how this framework could guide future studies in water management would leave a stronger impression.
Technical corrections
- Minor grammatical issues and ambiguous phrases appear throughout the text. For example, line 175, "up-and downstream imbalances," could be clarified as "upstream-downstream imbalances." A thorough proofreading would enhance readability.
- Figure 1 provides a schematic of the model, but its components and interconnections need to be labeled clearly. A legend or detailed figure description indicating each component's function within the model would enhance interpretability. Improvement Suggestion: Include a step-by-step description or flowchart illustrating the interactions between socio-economic factors, hydrologic processes, and policy influences. This would help in clarifying the multi-agent interactions and coupling between scales.
- Figures 2 and 5: These figures would benefit from concise captions that specify what variables or trends they are intended to show. For instance, state whether they display policy impacts, flow distributions, or demand-supply imbalances explicitly.
- Figure 3: While Figure 3 illustrates IGWM patterns for UWMs, more context on the visualized policy implications and the decision-making dynamics among UWMs would make the figures more impactful. Additionally, increasing the color contrast between scenarios in this figure would improve readability. Improvement Suggestion: Provide a rationale for each experimental scenario, focusing on its real-world applications. Consider adding a flowchart or table summarizing the experimental setups and their objectives to help readers follow the study's design.
- Figure 4: The information presented in Figure 4 lacks sufficient labeling to identify different variables. Clear labels or a more detailed caption would clarify how the results vary by scale and policy. Improvement Suggestion: Reorganize the results section by scale and add clear interpretations of findings about policy scenarios. Additional labels and color coding in figures, especially Figures 3 and 4, would improve clarity. Why are the lines zigzagging?
- Figures 3 and 4: Improve color contrast and include labels for specific variables to make visualizations easier to interpret. Including a note explaining the data or variables displayed in each figure would enhance accessibility.
- Line 367: Typo ‘experiment’
- Line 381: What are rimax, rrmax, and rsmax ?
- Line 430: References?
Citation: https://doi.org/10.5194/hess-2024-232-RC2 -
AC2: 'Reply on RC2', Mengxiang Zhang, 27 Nov 2024
Dear Reviewer,
Thank you for your valuable comments and suggestions, which will undoubtedly strengthen our study and enhance the readability and comprehensibility of our paper. We will address your comments and those of the other reviewers in detail and will make the necessary revisions to our manuscript.
Please find our detailed responses to your comments item by item in the attached document. Should you have any further comments or suggestions, please do not hesitate to share them with us via the HESS discussion system.Thank you once again for your insightful feedback.
Sincerely,
Ting Fong May Chui and Mengxiang Zhang
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 244 | 114 | 221 | 579 | 15 | 13 |
- HTML: 244
- PDF: 114
- XML: 221
- Total: 579
- BibTeX: 15
- EndNote: 13
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 204 | 35 |
| China | 2 | 46 | 8 |
| Germany | 3 | 30 | 5 |
| Netherlands | 4 | 23 | 4 |
| Romania | 5 | 21 | 3 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 204