Review Reports

Round 1

Reviewer 1 Report (New Reviewer)

Comments and Suggestions for Authors

General comments

This manuscript presents a valuable and methodologically sound study on a topic of practical importance. The core workflow is replicable and the findings are informative. However, the simplifying assumptions regarding RWH operation and the incomplete justification of the HEC-HMS integration method are important scientific shortcomings that must be addressed.

 

Major comments

1. The RWH framework may be oversimplified.The tank emptying depends on household water use patterns between storms. The model likely overestimates flood mitigation benefits, especially for closely spaced storm events.

 

2. How did the RWH integrate into the HEC-HMS? The method of adding captured volume as an increase to the initial abstraction is unconventional. While conceptually it removes water from the runoff volume, it functionally reduces the effective rainfall. The authors must justify why this approach was chosen over directly reducing the impervious connected area or adjusting the CN for contributing rooftops. A clearer rationale or reference to this modeling technique is needed.

 

3. Further discussion on the practical feasibility is recommended.A brief discussion linking the technical findings to implementation barriers would strengthen the conclusion's relevance for planners.

 

Specific comments

 

1. P2, Section 1 (Introduction): The definition of urban flooding depths (3-10 cm) is useful. Consider citing a key source for this specific numerical definition.

 

 

2. P8, Equations for V_runoff and V_RWH: The logic is clear. However, note that V_runoff is the potential harvestable volume, while V_RWH is the actual captured volume limited by tank capacity. This distinction is critical.

 

3. P9, Table 1 & P10, P11: The presentation of hydrological parameters is good. For clarity, consider adding a column for the percentage of impervious area in each sub-basin, as this is the primary driver of runoff in the model.

 

4. P14, Figure 6: The graph is effective. The caption should explicitly note that these results are based on the 4 m³ fixed tank size.

 

5. P15, Figure 7: The 3D bar chart is visually complex. A grouped 2D bar chart for each AMC condition might be clearer. Ensure the legend accurately reflects the adoption scenarios (0%, 25%, 50%).

 

6. P16, Conclusions: The stated runoff reduction percentages for 1-year storms (e.g., 26.4% for 25% adoption) are useful. Emphasize that these are volume reductions at the watershed outlet, which may not translate linearly to reductions in street-level inundation depth or extent. This is a key limitation of the watershed-scale approach.

Author Response

Please see attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report (New Reviewer)

Comments and Suggestions for Authors

The study addresses an important and underexplored application of decentralized RWH in semi-arid urban settings and demonstrates a solid GIS–HEC-HMS framework. However, the core limitation in accurately modeling fixed-capacity storage in HEC-HMS must be addressed to ensure the quantitative results are reliable. With these revisions, the manuscript has strong potential for publication in Water.

Major Comments

1. The approach of simulating rainwater harvesting by increasing the initial abstraction (Ia) in the SCS-CN loss method is conceptually flawed for fixed-capacity storage systems. Ia represents interception, depression storage, and early infiltration—processes that occur before runoff begins—but it does not account for the finite storage volume of tanks (here, 4 m³ per household). For storms where rooftop runoff exceeds tank capacity, overflow should occur, yet the current method effectively assumes unlimited storage up to the adjusted Ia. This overestimates capture efficiency, particularly for larger return-period storms. A more rigorous approach would be to explicitly model storage tanks as reservoir elements in HEC-HMS or use a modified loss method that caps retention at the available tank volume. This limitation should be clearly acknowledged, and sensitivity to tank capacity should be more thoroughly explored.
2. The reported reductions (e.g., from 13.1 to 10.6 m³ and 31 to 26.1 m³ for 1- and 2-year storms) appear to refer to very small volumes at specific locations (possibly a single junction or low-point). For a neighborhood with ~9,000 parcels, these absolute volumes seem implausibly low for total street runoff. Clarify whether these values represent total neighborhood runoff, a specific subbasin, or peak ponding volumes at critical points. If they are localized, the broader neighborhood-scale impact remains unclear. Provide total runoff volumes for the entire AOI with and without RWH to better contextualize the percentage reductions.
3. The manuscript mentions using both 1-hour and 24-hour durations but primarily reports results based on 1-hour events to represent short, intense storms typical of urban flooding drivers. However, the SCS-CN method and Frequency Storm approach in HEC-HMS are more conventionally applied with 24-hour distributions. Justify why 1-hour duration was prioritized and explicitly state which results correspond to which duration. Additionally, provide a comparison or sensitivity analysis between the two durations to strengthen the claim that RWH is effective under realistic convective storm patterns in semi-arid El Paso.
4. The study appears to assume tanks are empty at the start of each event (common in design storm approaches) and uses average CN(II) conditions. In reality, antecedent moisture and prior tank usage (especially in a semi-arid climate with clustered summer rainfall) could significantly reduce available storage. Discuss or test sensitivity to wet antecedent conditions (CN(III)) and partially filled tanks, as this could substantially reduce the reported mitigation benefits.
5. While supervised classification with Maximum Likelihood is described, no quantitative accuracy assessment (e.g., confusion matrix, kappa coefficient, or overall accuracy) is provided. Given the critical importance of rooftop area delineation for RWH potential, include validation metrics and discuss potential misclassification errors (especially distinguishing roofs from other impervious surfaces).

Minor Comments

1. The sentence “Even higher-magnitude storms (5- and 10-year) experience measurable decreases, with reductions of 10–10.4% at 25% adoption and 20–22% at 50% adoption” is slightly confusing—clarify that the range reflects different storms or conditions.
2. The definition of urban flooding (3–10 cm depth) is reasonable but somewhat arbitrary. Consider citing additional sources beyond [2] and [3] to support this threshold.
3. The manuscript carefully distinguishes active vs. passive RWH, but some readers may still confuse “active” with pumped systems rather than rooftop tank storage. A brief table summarizing the terminology used here versus common alternative usages would help.
4. The flowchart is useful, but consider adding a schematic of the overall workflow (GIS → parameter extraction → HEC-HMS → RWH modification) in addition to the detailed HEC-HMS steps.
5. Equation (7) uses 25400/254 (mm), while Equation (4) uses 1000/10 (inches). Ensure consistency in units throughout (preferably metric). Also, check equation numbering—some appear duplicated or misaligned in the provided text.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report (New Reviewer)

Comments and Suggestions for Authors

Dear authors,

your manuscript is a very well written contribution to the developing of GIS and HEC-HMS methods applied to the design of RWH systems.

The description of your case study is very detailed and useful for general applications in other similar climatic conditions.

I found only few minor issues, as the lack of a map illustrating and locating your case study, a typing error in a figure, some text to be moved in different positions,  a mixed use of Si and different units, and maybe a better heading and sub-heading required.

In the attached pdf you can find all my suggestions related to these issues.

My opinion is that your good work could be published after a minor revision round

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report (New Reviewer)

Comments and Suggestions for Authors

This study proposes a modeling framework that combines GIS with HEC-HMS, and quantifies the runoff reduction effects under different RWH adoption rates using El Paso as a case study. It is of great significance. Some minor revisions are proposed in terms of method transparency, depth of result interpretation, and extrapolation discussion. Line 131: The specific questions of the survey questionnaire and the online link?

Lines 152-173, 292: Formula variables should be in italics.

Line 262: you mentioned “The simulation runs can begin once the basin model setup is complete with meteorological and control specifications”. Can the meteorological data be presented? So that readers can see the rainfall data.

Line 269: you mentioned to identify and classify (1) Roof, (2) 269 Grass Open space, and (3) Driveways. Could you please provide a diagram showing how to combine the on-site photos, satellite photos, and the operation using ArcGIS?

You emphasized "mitigate flooding", but the model mainly deals with runoff volume. How can the reduction in runoff be translated into a decrease in water depth/overflow? Explain the necessity of requiring a city-scale hydraulic model in the future?

How is the capacity of the water storage tank set? Is it based on the local household water usage habits? How are the roof runoff coefficient and loss parameters determined and calibrated? The calibration and verification of HEC-HMS?

Line 342: Figure 2 should preferably include latitude and longitude and a compass.

Line 364: Too many decimal places have been retained.

Line 392: This picture is very blurry, and the aspect ratio of this area seems different from the others. Is this because the projected image and the satellite image undergo some distortion? Your numbers A-11 and A19-21 to Junction-2, A-12, 14-16, 22, 23 to Junction-4, were not found for A-13. The numbering sequence for A and Junction is a bit chaotic. It might be a characteristic of the residential area itself.

Is there a reduction threshold? Why did the reduction decrease under the intense heavy rain?

You mentioned things like permeable pavement and bioswales, but it was only a qualitative discussion. A brief comparison of typical reduction rates in the literature is presented to illustrate the unique advantages of RWH in cities with limited land, such as the absence of the need for land occupation.

Here is the more ideal scenario, but will the residents be willing to install it? Have costs, maintenance and subsidies been taken into consideration? Could some explorations of real-life scenario analyses be provided?

For the references of Figures/Tables, it is recommended to clearly state "(Figure x)" instead of merely describing the results.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report (New Reviewer)

Comments and Suggestions for Authors

The author have response all the comments from the reviewers.  I think it can be publish now

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

General comments:

The manuscript addresses the important issue of minimising localised flooding events in a densely urbanised area of El Paso, Texas, USA. The authors aim to demonstrate that distributed household storage can effectively prevent stormwater runoff from impervious building roofs from contributing to flooding.

I believe the rationale behind the work presented in this manuscript is as follows: in highly urbanised, arid regions, the initial abstraction of permeable surfaces is high, especially in case of sudden storms. This means that pluvial flooding volumes are largely contributed by impervious surfaces, and building roofs account for a large part of these surfaces. The percentage of rainwater collected from impervious roofs relative to the total runoff volume produced by a highly urbanised catchment can be significant and decreases with the magnitude of the rainfall event.

However, even if all the runoff produced from roofs is collected and stored in properly sized tanks, runoff will still form from other surfaces (e.g. low permeability surfaces like roads and paved areas). Therefore, expressing the performance in terms of “percentage kept off the streets” can be misleading and gives the impression that runoff can be totally avoided by just avoiding that roof runoff reaches the ground.

It is not clear from the text if the hydrological model involves the full AOI, including the building roofs. If that is the case (which seems reasonable) the fact that roof runoff volume does exceed the catchment runoff volume can only derive from poor calibration or from errors in the model runs. In theoretical terms, the maximum effect of roof runoff storage would be that the total catchment runoff equals the runoff from other catchment areas, excluding roofs. This could have been simulated easily in the adopted hydrological model within a solid conceptual framework. Comparing potential rainwater storage from roofs calculated with the average rainfall depth (Eq. 4) with the event-based runoff simulated using a hydrological model is incorrect. Indeed, roof runoff storage cannot affect the runoff from other areas at the ground. The performance of excluding roofs could be easily estimated by calculating the ratio of the total surface area of building roofs to the total catchment area, both weighted using appropriate runoff coefficients for the two surfaces (a weighted average for the catchment area).

However, the size of the tanks needed to achieve full storage of rainwater from roofs in El-Paso (not presented in the manuscript but easily calculated from data available on the web) is in the order of 12 cubic metres for a roof surface area of 400 m² already in the case of a 1-Year return period event (with duration of 24 hours). It is about equal to the minimum room size for US dwellings or a room of 2.4x2.5 m having a height of 2 m. This increases to 34 cubic metres for an event with return period of 25 years. If applied to a different rainfall climatology e.g. in my home area, this would require a three times larger tank size. The feasibility of the proposed solution is therefore questionable even at the scale of individual buildings in the AOI, and transferability to other cities is not discussed.

Furthermore, using these tanks as part of real rainwater harvesting (RWH) systems would only reduce their expected performance, since there is a non-null probability that the tanks will be partially full when an intense event occurs, thus reducing the available storage (runoff from roofs is not fully blocked).

The work contains various conceptual flaws. Indeed, the conducted work is incomplete since it addresses the capabilities of rainwater harvesting systems to mitigate pluvial flooding in the urban environment without any analysis of the continuity equation of the reservoirs and without any proposal of the size of these reservoirs. Indeed, the authors mentioned for the first time the term “tank size” in the “discussion” section at line 482 in the paragraph dedicated to further developments, therefore not addressed in the present work.

The conducted work is limited to the assessment of the land use map and the calculation of SCS Curve Number (CN). The authors dedicated the 90 % of the section 2 “Material and Method” and section 3 “Results” to these two simple concepts that are at the basis of any hydrological model and spent limited portion of the manuscript to describe and investigate the possibility to introduce RWH systems to mitigate pluvial flooding.      

Rainwater Harvesting is not the subject of this work. The exercise proposed in the manuscript assumes that the whole rainwater volume collected by roofs is retained in what is termed a RWH system. This is conceptually incorrect and unfeasible in practice, since RWH systems are not at all designed to accommodate the volume generated by the total rain depth over any given collecting surface, even for the shorter return periods. The rationale assumed by the authors seems much more suitable for the case of designing distributed storage tanks than RWH systems. Indeed, the authors make no assumption about the usage of the harvested water, which would impact on the behaviour of the system and largely reduce its performance on mitigating pluvial floodings.

The authors declare that the proposed method aims to mitigate “sunny-day flooding” but, as correctly reported by the authors in line 51, “sunny-day” flooding commonly refers to tidal processes or storm surge with no rain involved. No tidal or coastal floodings are instead addressed in the manuscript but rather pluvial flooding scenarios associated to rainfall events of different return periods.  Therefore, the use of “sunny-day” flooding is not justified in this work and the statement in line 52, “sunny-day flooding can happen anywhere where high-intensity rainfall occurs”, is incorrect (note that “sunny-day” flooding is not described in reference [1] as erroneously mentioned in L 46 – 47). The authors should have termed the objective of their work “pluvial flooding”, maybe due to sudden storms, following the definition reported e.g. by Falconer et al. (2009) that pluvial flooding results from “rainfall-generated overland flow and ponding before the runoff enters any watercourse, drainage system or sewer, or cannot enter it because the network is full to capacity”. Note that pluvial flooding in the urban environment is associated with rainfall events characterised by the typical return periods used to design stormwater drainage networks (up to 10 years). Rainfall scenarios with relatively large return periods (limited to 25-30 years) can be included in the simulation to cover a range of events that could lead to critical conditions in an urban environment and that could benefit from the proposed mitigation solution. Nevertheless, addressing rainfall scenarios with return periods way larger than 30 years is not justified (in the proposed work the authors simulate rainfall events with return period up to 500 years).
Ref: Falconer, R.H., Cobby, D., Smyth, P., Astle, G., Dent, J. and B. Golding (2009). Pluvial flooding: new approaches in flood warning, mapping and risk management. Journal of Flood Risk Management, 2, 198-208. https://doi.org/10.1111/j.1753-318X.2009.01034.x

In L 80 – 83 active and passive RWH systems are mentioned but the collection of rainwater from roofs is attributed only to the fist type. This is incorrect because both systems can collect and store rainwater from roofs or other catchment surfaces for later use. The difference between the two systems is that active systems have dedicated devices implemented to control the filling and emptying of the tank. In an active system the water storage is actively managed through automated systems, ensuring that all properties can comply with surface water management objectives regardless of the balance between demand and supply. The term passive therefore refers to how the system operates without active intervention or control mechanisms.
The authors should explain why they assume the planned system as active and the type of automation involved. Is there any active system implemented to ensure that the storage volume is maintained empty at the onset of each rainfall event? Note that in such case a forecasting model would play a crucial role to avoid the failure of the mitigation solution.

The authors chose the temporal resolution of the rainfall event at daily scale and state in line 232-234 that “A 24-hour design storm was selected because it allows for the assessment of total runoff and RWH potential over a whole storm event and provides a conservative basis for evaluating flood mitigation effectiveness”. This is conceptually incorrect. Daily rainfall intensity is much lower by construction than the intensity producing pluvial flooding in each sub-basin. It may be conservative in terms of water volume but not in terms of flooding potential.

The title and the text along the whole manuscript are misleading because it seems that the authors would simply convert each roof into a reservoir. It is not clear how eq. 4 is used and how the comparison reported in Table 4 between “runoff volume obtained using HEC-HMS” and the “Storage Capacity of Roof” must be interpreted. Is the “Storage Capacity of Roof” calculated using eq. 4? If yes, the proposed approach is incorrect because eq. 4 provides the runoff generated over the surface of the roof and cannot be interpreted as a storage capacity of a harvesting system.

Tables and Figures report trivial and redundant information. For instance, Figure 3 and Figure 4 can be merged in a single figure with two panels as well as Figures 5, 6 and 7. In Table 5 the same information is reported ten times by reducing the values reported in column 2 by 10 % in each subsequent column. This representation of the results is not suitable for an international peer review scientific journal.

What are the measurement units of the ‘rainwater captured’ volume in Equation 4? A mix of SI and non-SI units are used in the right-hand part of the equation, and I could not check if the equation is dimensionally consistent (what is ‘7.48 gal/ft’?). Later in the manuscript the acre-foot (AC-FT) unit is used for volumes, which seems not the one used in Equation 4. Please use consistent SI units throughout the manuscript (cubic meters for volumes).

What is the ‘average’ rainfall used in Equation 4? Is this the total event depth (total is not average) for events with given return period and 24-h duration? In that case, the calculated volume largely exceeds the rainfall volume that contributes to pluvial flooding, since the occurrence of flooding in a catchment is controlled by rainfall intensity rather than total rainfall depth. Most importantly, this is not consistent with the calculations performed by the hydrological model. Is the hydrological model run at the daily scale?

Based on the above relevant flaws, I believe the manuscript is not suitable for publication and must be rejected.

Minor comments:

L 117 please use the complete name of USDA where is written for the first time.

For eq. 1 “Ai” should be defined along the text despite easy to guess.

In section 2.2 additional details about the training dataset and the classification method should be added to the poor description available in L 170 – 172. It is not clear if the training dataset is outside to the study area or at least a limited portion of the study area. 

L302-303 this sentence is trivial because it is intrinsic in the relationship between CN and S reported in eq. 2.

L250 – 251 “Flooding caused by rain during the monsoon and non-monsoon seasons can put communities at risk of flooding”. This sentence is trivial

L276 “For impervious areas, such as parking lots, driveways, or roofs, the CN was 98.” Is this an assumption or a result? Is there any sensitivity to the selected CN values, especially for roofs against the values assumed for pervious areas?

L 459 – 460 No validation is presented in the manuscript.

 

Author Response

Please see the attachment. Thank You!

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The paper seems to be extracted from some report and does not have any innovative ideas for the research community as such. The study involves the application of HEC-HMS for simulation of catchment runoff and computation of RWH potential based on the roof areas, grassed areas etc. and a measure of the quantum of water than can be trapped through RWH for preventing floods.

Comments for author File: Comments.pdf

Author Response

Please see the attachment. Thank You!

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript examines the concept of converting roofs into reservoirs. I find the consideration of different return periods for storm events particularly interesting, and the methodology is generally well designed. The literature review is clearly structured. This work can be considered a mix between a case study and the application of a simplified methodology to assess the limits and advantages of applying the potential of decentralized active rainwater harvesting by means of a simplified approach. I think that the work is worth publishing.

1. I have the feeling that sub-hourly flow controls the urban rainfall response more than daily extremes. Could the authors justify better the choice of 24-hour events?
2. I think, to improve readability, that Table 6 should be transformed into a figure.
3. 3.1 GIS-Based Approach: I believe that this title is not appropriate. Here we are in the results section, not the methodology section.
4. Could the authors add a case study section? The current description of the case study in the result section could be misleading.

Author Response

Please see the attachment. Thank You!

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The innovative aspect of the study (‘this method was used for the first time to model El Paso’) should be emphasised more clearly in the introduction.

More advanced classification algorithms (Random Forest, SVM) should be used in terms of material and method, and accuracy should be reported. The sample size, method, and analysis method of the survey data can be clearly presented.

Different scenarios (climate change projections, population growth) should also be taken into account in the results.

In the discussion section, a comparison of costs and effectiveness should be made with other green infrastructure applications (rain gardens, permeable paving). The integration of local government policies should be evaluated.

In the conclusions, concrete steps such as “pilot application, field observation, cost calculations” can be suggested for future studies.

Comments for author File: Comments.pdf

Author Response

Please see attachment. Thank You!

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors interpreted the criticisms I expressed in my first review as minor comments, despite the fact that I had identified significant and substantial flaws which deserved more in-depth consideration and relevant changes to the manuscript. The main novelty of the revised manuscript is the addition of a lengthy section to the discussion, which interprets the required changes as potential future developments.

The authors' main argument is that rainwater captured from roofs constitutes a significant proportion of the runoff volume calculated using the hydrological model. However, the method used to demonstrate this is flawed in many ways, as detailed below.

The adopted approach and the model used are oversimplified, based on very unrealistic hypotheses. Simply curtailing the contribution of roofs at the daily scale to reduce the runoff volume is not representative of the potential role of rainwater harvesting (RWH) systems in mitigating pluvial flooding. Assumptions such as a 72% adoption rate are not credible; a more realistic figure would be 10–20%. Similarly, the availability of the full tank volume is unrealistic.

It is not explicitly stated whether the hydrological simulation includes the response of roofs. As the hydrological model does not include roofs as a source of runoff ('post-loss direct runoff'; see the authors' response to the first review), it is inappropriate to compare the resulting volume with that potentially captured on roofs. Indeed, street runoff will not be available for capture on roofs. If roofs are included, there must be an error in the numbers reported in Table 4 (the data in column 2 cannot be lower than those in column 3).

The most important point is that pluvial flooding events are driven by rainfall intensity rather than volume. Water accumulates on the streets due to the temporary inefficiency of drainage systems when the runoff rate exceeds the capacity of storm drain inlets to handle the water. The runoff rate at a given location in an urban area is determined by the hydrological response of the upstream catchment. This response is a function of the rainfall rate occurring over a duration roughly equivalent to the time of concentration of the upstream catchment.

As is typical of urban areas, the catchments investigated in this study have limited surface area (between 0.22 and 0.8 km² – see Table 1), and the corresponding time of concentration is also very limited (between 0.20 and 0.43 hours – see Table 2). The critical rainfall duration is therefore in the order of 15 to 25 minutes. Using the 24-hours event duration in the hydrological simulation is simply inconsistent with the dynamics of the hydrological systems investigated. Although this could be assumed as acceptable in simplified technical applications developed for other purposes, it is incorrect in the present case where the 24-hour event associated with any given return period would correspond to a much lower rainfall intensity than the critical event for the area (although for a reduced duration). The resulting runoff rate would therefore be much lower than the one producing pluvial flooding. It means that the runoff volume collected from the roof during the event would be much lower than the volume accumulated in the streets.

The method appears to be flawed because the volume of water collected from roofs over a 24-hour period has little impact on the volume of pluvial flooding, given the very different rainfall periods involved.

Equations 4 and 5 do not consider rain duration, except in the form of a 24-hour rain depth, which is assumed a priori. In contrast, the hydrological model responds according to the size and characteristics of the upstream catchment, even though it is fed with constant rain for 24 hours. It therefore inherently selects the correct critical duration for the catchment, producing a runoff rate consistent with this duration. The runoff volume generated by the roofs cannot be meaningfully compared to that produced by the model unless the timeframe is consistent.

Furthermore, the hydrological/hydraulic model lacks relevant details, making it impossible to determine whether it is appropriate for simulating pluvial flooding events. A critical aspect is the connection between surface runoff and the stormwater drainage network, specifically the hydraulic behaviour of inlets. The efficiency of this connection depends on the runoff rate, rather than the runoff volume.

The temporal distribution of rainfall within the critical duration is also important. While the captured volume from the roofs would be insensitive to the temporal evolution of rainfall, the hydrological response of the catchment would not be. If most of the rain falls at the beginning of the event, the rainfall peak will occur when the soil can infiltrate most of the rain; conversely, if most of the rain falls at the end of the event, the rainfall peak will occur after the soil is saturated.

Finally, it is unclear how the antecedent dry period is considered (e.g. in the selection of CN values), as this would also impact the results.

The term 'recurrent' is unnecessary and redundant when used to indicate the high frequency of storm events, since the manuscript refers to the return period of the target flood event. As the return periods investigated range from 1 to 50 years, the flood events referred to in this manuscript cannot be considered frequent (at least for the larger return period). 'Recurrent' means 'happening again many times' (see the Cambridge Dictionary), which is an inherent feature of urban flooding events and therefore does not need to be specified. Furthermore, the GIS and hydrological models are not used to mitigate urban flooding, but rather to assess their mitigation potential. The title would therefore be better formulated as 'Mitigating Pluvial Flooding Through Residential Rainwater Harvesting'.

In conclusion, due to the above considerations, I still recommend that the manuscript cannot be accepted for publication in the Water journal.

Reviewer 2 Report

Comments and Suggestions for Authors

All the suggested improvements have been carried out by the authors.