the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 06 Jan 2025
Enhancing Urban Pluvial Flood Modelling through Graph Reconstruction of Incomplete Sewer Networks
Abstract. This work presents an efficient graph reconstruction-based approach for generating physical sewer models from incomplete information, addressing the challenge of representing sewer drainage effect in urban pluvial flood simulation. The approach utilizes graph-based topological analysis and hydraulic design constraints to derive gravitational flow directions and nodal invert elevations in decentralized sewer networks with multiple outfalls. By incorporating linearized programming formulation to solve reconstruction problems, this approach can achieve high computational efficiency, enabling application to city-scale sewer networks with thousands of nodes and links. Tested in Yinchuan, China, the approach integrates with a 1D/2D coupled hydrologic-hydrodynamic model and accurately reproduces maximum inundation depths (R2 = 0.95) when the complete network layout and regulated facilities are available. Simplifications, such as adopting road-based layouts and omitting regulation facilities, can degrade simulation performance for extreme rainfall events compared to calibrated equifinal methods. However, design rainfall analysis demonstrates that the physical reconstruction approach can reliably outperform equifinal methods, achieving reduced variation and higher accuracy in simulating inundation areas. However, proper configuration of regulated facilities and network connectivity remains crucial, particularly for simulating local inundation during extreme rainfall. Thus, it is recommended to integrate the proposed algorithm with targeted field investigations to further improve urban pluvial flood simulation performance in data-scarce regions.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Hydrology and Earth System Sciences. The peer-review process was guided by an independent editor, and the authors also have no other competing interests to declare.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(7250 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2024-3780', Anonymous Referee #1, 23 Jan 2025
- AC1: 'Reply on RC3', Ruidong Li, 01 Jun 2025
-
AC2: 'Revision of egusphere-2024-3780 with tracked changes', Ruidong Li, 01 Jun 2025
Publisher’s note: the content of this comment was removed on 4 June 2025 since the comment was posted by mistake.
Citation: https://doi.org/10.5194/egusphere-2024-3780-AC2
-
RC2: 'Comment on egusphere-2024-3780', Anonymous Referee #2, 11 Feb 2025
The manuscript is relevant as it aligns with the research on sewer network data scarcity. The authors propose a methodology for sewer network reconstruction using a graph-based approach. The results appear comprehensive and more accurate than other methodologies, such as generating a virtual network based on road layouts, as proposed in previous studies. Some minor comments could help improve the manuscript:
- In Section 3.2, the manuscript explains how water is drained from the TSU to the network, but it does not clarify whether flow is generated from the network to the TSU when manholes are overloaded.
- It is also recommended to explain whether the 2D-1D coupled model maintains mass balance.
- In line 292, it is recommended to describe the specifications of the computer used for the analysis.
- Clarify whether the methodology can be applied without information on pumps and WTPs. What alternatives exist if data on their location or technical specifications are unavailable?
Citation: https://doi.org/10.5194/egusphere-2024-3780-RC2 - AC1: 'Reply on RC3', Ruidong Li, 01 Jun 2025
-
AC2: 'Revision of egusphere-2024-3780 with tracked changes', Ruidong Li, 01 Jun 2025
Publisher’s note: the content of this comment was removed on 4 June 2025 since the comment was posted by mistake.
Citation: https://doi.org/10.5194/egusphere-2024-3780-AC2
-
RC3: 'Comment on egusphere-2024-3780', Anonymous Referee #3, 12 Mar 2025
This paper presents a graph reconstruction-based approach for generating physically realistic sewer models from incomplete information, for addressing a significant challenge in urban flood modeling. The authors develop a methodology that leverages graph theory and linearized programming formulations to derive gravitational flow directions and nodal invert elevations in decentralized sewer networks with multiple outfalls. They conduct a case study in Yinchuan, China, by performing a 1D/2D coupled hydrologic-hydrodynamic model and reproduces maximum inundation depths with high accuracy (R² = 0.95) when complete network information is available. The paper's strengths lie in its mathematical formulation that transforms complex nonlinear problems into computationally efficient linear equivalents, and its comprehensive comparison against alternative approaches across varying rainfall scenarios. However, there are some major issues that authors need to address before publication.
Major Comments
- The paper lacks technical discussion on the implementation, i.e., the solver used for the linearized formulations. There are several existing solvers, each with their own strengths, weaknesses, and parameters requiring tuning. Sometimes solvers require citation too, so the authors need to credit the developers of packages and software used to develop their tool.
- While authors discuss briefly that their approach is fast, they do not provide any supporting evidence, especially since the source code is not available for review. The authors mention they will open source their code, but it's common practice nowadays for authors to provide access to the code that was used to generate the results during the review process for reproducibility and validation.
- The authors do not provide any discussion on the sampling approach used for getting node elevations from the DEM. DEMs, especially in urban areas, can contain artifacts that if not addressed through DEM processing operations such as conditioning can lead to propagation of uncertainties in raw DEMs into the generated network. While the algorithm provided by the authors tries to address the hydraulic feasibility of the network, if the source DEM contains major artifacts, especially for high-resolution DEMs in built areas, it can adversely impact the quality of the generated network.
- Pages 9-11 are allocated to providing mathematical details of the AUTOSHED model. It's not clear from the paper whether the authors actually made any developments in the coupled 1D/2D model, or they are just simply using the model as an off-the-shelf tool. If they are simply using the model, there's really no need for this detailed discussion, and can even be misleading, as the authors can mention some relevant high-level information since the focus of this study seems to be on generating a hydraulically feasible network.
- The discussion on the outfall is not very clear. Do the authors assume that the outfall locations are known? If that's the case, this can be a major limitation as they are often not readily available. If the authors have some way of identifying the outfall locations, they must explicitly discuss the details; otherwise, this should be explicitly mentioned as one of the limitations as the authors state that one of their major contributions is deriving hydraulically feasible decentralized sewer networks with multiple outfalls.
- It appears the algorithm has only one tunable parameter: pipe segmentation, with a default value of 50 m. However, the authors did not provide a discussion or perform a formal sensitivity analysis on this parameter. It's important to quantify how this parameter impacts the model performance.
- The authors do not provide a direct validation of their approach. They do so indirectly by measuring simulation results with inundation depths. While I understand that there aren't many public sewer data with actual flow directions, but since the contribution of authors is mainly related to enforcing correct flow directions throughout the network, and they perform pipe-level simulations, lack of validation with existing pipe flow directions is one of the major limitations that needs to be explicitly mentioned. If authors have access to such data, even partially, direct validation of flow direction can be helpful for quantifying the performance.
- In the introduction, I suggest providing a discussion on how this work compares to more recent relevant work in the literature on this topic, such as https://doi.org/10.1016/j.envsoft.2025.106358.
Minor Comments
- For all figures containing a basemap, make sure the attribution follows the specific language required by Esri.
- Throughout the text, some citations are not correct. For example, on line 144 the citation is "(Wagner (1959); Giangrande et al. (2013)" which not only misses a parenthesis but also is not in the correct format which is "(Wagner, 1959; Giangrande et al., 2013)". This should be a LaTeX-related issue. Please carefully go over the text to catch citations with similar issues, as I found many instances, e.g., on pages 10 and 11.
- There are some awkward sentences throughout the text. Please carefully proofread the paper.
- What does the "equifinal method/approach" mean? Does it mean conceptual models? While equifinality is commonly used in the hydrology literature, I am not familiar with using equifinal method in the context of different modeling approaches. Please clarify this term or replace it with a more common jargon.
- Fig 8: Use a different legend, it's very difficult to see what's going on.
Â
Citation: https://doi.org/10.5194/egusphere-2024-3780-RC3 -
AC1: 'Reply on RC3', Ruidong Li, 01 Jun 2025
Dear reviewers,
Please see attached our response to your comments.
Many thanks for your evaluation of our work!
Best,
Ruidong on behalf of all co-authors
-
AC2: 'Revision of egusphere-2024-3780 with tracked changes', Ruidong Li, 01 Jun 2025
Publisher’s note: the content of this comment was removed on 4 June 2025 since the comment was posted by mistake.
Citation: https://doi.org/10.5194/egusphere-2024-3780-AC2
-
AC2: 'Revision of egusphere-2024-3780 with tracked changes', Ruidong Li, 01 Jun 2025
Status: closed
-
RC1: 'Comment on egusphere-2024-3780', Anonymous Referee #1, 23 Jan 2025
Comment 1: In Sect. 2.1, the authors introduce the definition of "source" (Line 93) but never reuse it again in the following analysis. Additionally, they also mention terms related to "inlet" like "Non-inlet nodes" (Line 118). Do they share the same meaning? If so, please revise the related sentences for consistency between definition and analysis.
Â
Comment 2: For computational efficiency, the authors only use rough words like "a few minutes" without the information of specific time statistics and device specification (Line 292). If possible, please add more related details to offer readers/users clearer insights into the efficiency of algorithm execution.
Â
Comment 3: In Sect. 4.3, the comparison results show that regulation facilities has significant effects on the results of inundation simulation. Thus, compared to the current comparison between FSN and RSN, it might be more valuable to consider the possible effects of road-base layout simplification with controls, given that the location of regulation facilities can be easily identified by field investigation and then incorporated into the flood model.
Â
Comment 4: For Line 144, there might to be an extra ")"
Citation: https://doi.org/10.5194/egusphere-2024-3780-RC1 - AC1: 'Reply on RC3', Ruidong Li, 01 Jun 2025
-
AC2: 'Revision of egusphere-2024-3780 with tracked changes', Ruidong Li, 01 Jun 2025
Publisher’s note: the content of this comment was removed on 4 June 2025 since the comment was posted by mistake.
Citation: https://doi.org/10.5194/egusphere-2024-3780-AC2
-
RC2: 'Comment on egusphere-2024-3780', Anonymous Referee #2, 11 Feb 2025
The manuscript is relevant as it aligns with the research on sewer network data scarcity. The authors propose a methodology for sewer network reconstruction using a graph-based approach. The results appear comprehensive and more accurate than other methodologies, such as generating a virtual network based on road layouts, as proposed in previous studies. Some minor comments could help improve the manuscript:
- In Section 3.2, the manuscript explains how water is drained from the TSU to the network, but it does not clarify whether flow is generated from the network to the TSU when manholes are overloaded.
- It is also recommended to explain whether the 2D-1D coupled model maintains mass balance.
- In line 292, it is recommended to describe the specifications of the computer used for the analysis.
- Clarify whether the methodology can be applied without information on pumps and WTPs. What alternatives exist if data on their location or technical specifications are unavailable?
Citation: https://doi.org/10.5194/egusphere-2024-3780-RC2 - AC1: 'Reply on RC3', Ruidong Li, 01 Jun 2025
-
AC2: 'Revision of egusphere-2024-3780 with tracked changes', Ruidong Li, 01 Jun 2025
Publisher’s note: the content of this comment was removed on 4 June 2025 since the comment was posted by mistake.
Citation: https://doi.org/10.5194/egusphere-2024-3780-AC2
-
RC3: 'Comment on egusphere-2024-3780', Anonymous Referee #3, 12 Mar 2025
This paper presents a graph reconstruction-based approach for generating physically realistic sewer models from incomplete information, for addressing a significant challenge in urban flood modeling. The authors develop a methodology that leverages graph theory and linearized programming formulations to derive gravitational flow directions and nodal invert elevations in decentralized sewer networks with multiple outfalls. They conduct a case study in Yinchuan, China, by performing a 1D/2D coupled hydrologic-hydrodynamic model and reproduces maximum inundation depths with high accuracy (R² = 0.95) when complete network information is available. The paper's strengths lie in its mathematical formulation that transforms complex nonlinear problems into computationally efficient linear equivalents, and its comprehensive comparison against alternative approaches across varying rainfall scenarios. However, there are some major issues that authors need to address before publication.
Major Comments
- The paper lacks technical discussion on the implementation, i.e., the solver used for the linearized formulations. There are several existing solvers, each with their own strengths, weaknesses, and parameters requiring tuning. Sometimes solvers require citation too, so the authors need to credit the developers of packages and software used to develop their tool.
- While authors discuss briefly that their approach is fast, they do not provide any supporting evidence, especially since the source code is not available for review. The authors mention they will open source their code, but it's common practice nowadays for authors to provide access to the code that was used to generate the results during the review process for reproducibility and validation.
- The authors do not provide any discussion on the sampling approach used for getting node elevations from the DEM. DEMs, especially in urban areas, can contain artifacts that if not addressed through DEM processing operations such as conditioning can lead to propagation of uncertainties in raw DEMs into the generated network. While the algorithm provided by the authors tries to address the hydraulic feasibility of the network, if the source DEM contains major artifacts, especially for high-resolution DEMs in built areas, it can adversely impact the quality of the generated network.
- Pages 9-11 are allocated to providing mathematical details of the AUTOSHED model. It's not clear from the paper whether the authors actually made any developments in the coupled 1D/2D model, or they are just simply using the model as an off-the-shelf tool. If they are simply using the model, there's really no need for this detailed discussion, and can even be misleading, as the authors can mention some relevant high-level information since the focus of this study seems to be on generating a hydraulically feasible network.
- The discussion on the outfall is not very clear. Do the authors assume that the outfall locations are known? If that's the case, this can be a major limitation as they are often not readily available. If the authors have some way of identifying the outfall locations, they must explicitly discuss the details; otherwise, this should be explicitly mentioned as one of the limitations as the authors state that one of their major contributions is deriving hydraulically feasible decentralized sewer networks with multiple outfalls.
- It appears the algorithm has only one tunable parameter: pipe segmentation, with a default value of 50 m. However, the authors did not provide a discussion or perform a formal sensitivity analysis on this parameter. It's important to quantify how this parameter impacts the model performance.
- The authors do not provide a direct validation of their approach. They do so indirectly by measuring simulation results with inundation depths. While I understand that there aren't many public sewer data with actual flow directions, but since the contribution of authors is mainly related to enforcing correct flow directions throughout the network, and they perform pipe-level simulations, lack of validation with existing pipe flow directions is one of the major limitations that needs to be explicitly mentioned. If authors have access to such data, even partially, direct validation of flow direction can be helpful for quantifying the performance.
- In the introduction, I suggest providing a discussion on how this work compares to more recent relevant work in the literature on this topic, such as https://doi.org/10.1016/j.envsoft.2025.106358.
Minor Comments
- For all figures containing a basemap, make sure the attribution follows the specific language required by Esri.
- Throughout the text, some citations are not correct. For example, on line 144 the citation is "(Wagner (1959); Giangrande et al. (2013)" which not only misses a parenthesis but also is not in the correct format which is "(Wagner, 1959; Giangrande et al., 2013)". This should be a LaTeX-related issue. Please carefully go over the text to catch citations with similar issues, as I found many instances, e.g., on pages 10 and 11.
- There are some awkward sentences throughout the text. Please carefully proofread the paper.
- What does the "equifinal method/approach" mean? Does it mean conceptual models? While equifinality is commonly used in the hydrology literature, I am not familiar with using equifinal method in the context of different modeling approaches. Please clarify this term or replace it with a more common jargon.
- Fig 8: Use a different legend, it's very difficult to see what's going on.
Â
Citation: https://doi.org/10.5194/egusphere-2024-3780-RC3 -
AC1: 'Reply on RC3', Ruidong Li, 01 Jun 2025
Dear reviewers,
Please see attached our response to your comments.
Many thanks for your evaluation of our work!
Best,
Ruidong on behalf of all co-authors
-
AC2: 'Revision of egusphere-2024-3780 with tracked changes', Ruidong Li, 01 Jun 2025
Publisher’s note: the content of this comment was removed on 4 June 2025 since the comment was posted by mistake.
Citation: https://doi.org/10.5194/egusphere-2024-3780-AC2
-
AC2: 'Revision of egusphere-2024-3780 with tracked changes', Ruidong Li, 01 Jun 2025
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 714 | 144 | 25 | 883 | 16 | 38 |
- HTML: 714
- PDF: 144
- XML: 25
- Total: 883
- BibTeX: 16
- EndNote: 38
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Comment 1: In Sect. 2.1, the authors introduce the definition of "source" (Line 93) but never reuse it again in the following analysis. Additionally, they also mention terms related to "inlet" like "Non-inlet nodes" (Line 118). Do they share the same meaning? If so, please revise the related sentences for consistency between definition and analysis.
Â
Comment 2: For computational efficiency, the authors only use rough words like "a few minutes" without the information of specific time statistics and device specification (Line 292). If possible, please add more related details to offer readers/users clearer insights into the efficiency of algorithm execution.
Â
Comment 3: In Sect. 4.3, the comparison results show that regulation facilities has significant effects on the results of inundation simulation. Thus, compared to the current comparison between FSN and RSN, it might be more valuable to consider the possible effects of road-base layout simplification with controls, given that the location of regulation facilities can be easily identified by field investigation and then incorporated into the flood model.
Â
Comment 4: For Line 144, there might to be an extra ")"