Stormwater Detention Ponds in Urban Catchments—Analysis and Validation of Performance of Ponds in the Ouseburn Catchment, Newcastle upon Tyne, UK

Round 1

Reviewer 1 Report

Intro

The authors have analysed the effect of the Kingston Park residential development in Newcastle upon Tyne on the flow regime of the Ouseburn river – an area with stormwater drains feeding directly into the river -   and the impact of the Newcastle Great Park developments - an area with a stormwater drainage system including large detention ponds. The effect of (the lack of) detention ponds on peak flows, average (Q50) flows low flows and recession curves is analysed using a unique dataset of the Ouseburn river flow regime that started in 1976, supported by data from the Blyth and Wansbeck river basin. In addition to the numerical analysis of the data a Shetran hydrological model was calibrated and used to compare “what-if” scenarios for the flow regime of the river.    

This type of analysis is indeed rare, as stated in L47, largely due to the lack of good datasets, despite their importance for our understanding of the impact of urbanization on the flow regime of local rivers and of the effect of stormwater control measures to achieve low impact development.

The valuable analyses presented in this manuscript is unfortunately weakened by several important issues, requiring revisions before publication.

Major issues

In L 488-489 the authors conclude “Without their presence, for more intense rainfall events, there would be larger peak flows and so potentially more downstream flooding.” However, the opposite could be true as well, both in this case and in general! Due to the urban developments part of the basin starts draining fast. As a result, the slow flood peak of the basin is lowered. If the fast flood peak is relatively small as compared to the slow flood peak, this urban development has a positive effect on downstream flood risk, rather than a negative. The conclusions presented in the paper hold for the Crag Hall, as illustrated in Figure 3, but further downstream more slowly draining water comes in and starts dominating the rapid peak (line 5c); at that moment the flood levels benefit from the fact that part of the flood water came early. In case we would construct detention basins that perfectly mimic the natural (pre-development) flow regime we in fact aggravate the flood risk downstream by stacking the urban flood peak on top of the upstream flood peak. This complex consideration is missing in the discussion and conclusion.

Melbury Detention Pond

Related to this is the question about the size and discharge rate of the Melbury detention pond. The controlled (how?) discharge rate of 0,01 m3/s equals 0,42 l/s/ha, which is extremely low. Consequently, the characteristic response time of this reservoir is larger than the response time of the upstream basin The Melbury Detention pond will respond slower to rainfall than the upstream basin. Perfect, but would it have been wrong to drain at least part of the peak before the large peak of the upstream part of the basin passes the outlet of the pond? If the discharge rate is controlled shouldn’t we make use of this? This would also reduce the risk of the reservoir not being empty at the start of the next storm event. Strange is also the capacity of 30,000 m3. 65 mm of rainfall over 0,24 km2 produces 15,600 m3 of water, neglecting all losses on its way to the pond. Why is the size almost the double? Advantage of this redundancy is that the reservoir will hardly ever be overloaded and spill water into the Ouseburn river in an uncontrolled way.

The authors use the Melbury detention pond to illustrate the low impact of the Newcastle Great Park residential development on the Ouseburn river flow regime. Unclear remains however in how far this Melbury system is comparable to the other ponds for the other parts of the great Park development. How about the stormwater detention for Elmwood, East Moor Village, Warkworth Woods,…?

Incongruency in the analysis

The impact of the Kingson Park Developments is evaluated on three aspects, (1) the 30/6/2007 rainfall event, with an analysis of observed and simulated discharges, (2) the Base Flow Index, comparing the Ouseburn catchment with the Blyth and Wansbeck catchment and (3) the changes in the recession curve of the Ouseburn river over the period 1976-2019. The impact of the Newcastle Great Park Developments is evaluated on four aspects, (1) the 6/8/2011 rainfall event, with an analysis of observed and simulated discharges, (2) a trends in flows analysis using the annual Q10 and Q50, (3) an analysis of the largest events and (4) a peaks over threshold analysis. The authors provide no explanation in the Data and methods chapter why these aspects were selected and why this difference in evaluation method between the two residential development is necessary and useful for the purpose of their evaluation. A comparison of the results for the two is not at all discussed.

Data on Ouseburn catchment and the residential developments

Section 2.1 provides interesting information on the Ouseburn catchment, but no information on soil type (relevant for surface runoff from unpaved surfaces, nor data on the type of urbanization in the basin, the ratio roofs / impervious surfaces /permeable pavement/ unpaved land / disconnected paved surface, etc. This info is relevant to the generation of flood peaks in the urbanized areas, as well as to the way this area is to be modelled in a reliable way (e.g. for modelling infiltration losses). In this context it also remains unclear how the urban areas are modelled in the Shetran model. The way this modelling is described in section 2.3 makes me worry. Work-arounds like setting conductivity to zero, removing precipitation form areas and manipulating the Strickler roughness parameter could result in an NS efficiency of 0.86 for the calibration run, but what is the performance in the validation run?

1978

The authors seem to be convinced that 1978 was a tipping point in the flow regime, but is the evidence convincing? 1976 was extremely hot and dry and in 1977 the hydrological system still suffered the consequences. Couldn’t that explain why the BFI (Figure 4) and the recession curves (Figure 5) were different. The authors lack to discuss the fluctuations in the meteorological regime over the 45 year and do not mention this as a consideration for observed trends.

Trends in flows (section 4.2):

It can be questions whether an analysis of annual Q10 and Q50 is acceptable and provides reliable results. This doubt is of course compensated by the fact that the results are used only to compare the observed discharges at two stations in the same river, but the uncertainty introduced by taking such a small sample (one year of observed discharges for each station) makes the assessment of in particular Q10 uncertain; this could explain why the ratio of flows (Figure 8c) at Crag Hall and Woolsington < 1.0 for several years.

Largest event (section 4.3)

The ‘proven’ trend in the ratio presented in Figure 9 is questionable, not only because of the reasons mentioned by the authors in L 373-380, but also due to other developments in the basin between Woolsington and the Crag Hall.

 

Minor issues

L2: The title of the article is too generic. The article is an analysis and evaluation of one single case – to be mentioned in the title.

L21: remove “a valuable and unusual resource” from this sentence.

L 23-24: remove “and make recommendations for ensuring similar performance in designing future similar systems” as these recommendations are not provided.

L35: Stormwater detention can equally be achieved using wet detention ponds. Could that have been a better option in the area as for the landscape quality ?

L59: the research question is “How do stormwater detention ponds work …?”

L 69: Was Kingston Park indeed built in three years? Very fast.

L75, 77: Howden and Newbiggin Hall are not on the maps in Fig 1 and 2.

L91: 65 mm in 6 hours has a rturn period of 1/100 years. L309 however reports a return period of 1/136 years

L 112: Hartford Bridge and Mitfort are missing on the map in Fig 1.

L143-156: Stormwater pipe flow was not a separate element in the model? Is that why the roughness parameter was manipulated, to mimic fast stormwater flow in the storm drains?

L162: Missing here – and in the Methods section - is an intro into sections 3.1-3.3, their relevance and relationship. A similar intro is missing after line 298

L188: What was the size if this post KP pond (detention and outlet capacity)? How was it designed and why was this size chosen?

L192: What were the initial conditions for this simulated event?

L207: Where can I see that the response is “very similar to the measured discharge”?

L 274-280: the analysis uses absolute values of the change in discharge. Why not use relative values for this analysis.

L429: effect

L 441: remove “is powerful and unusual as it”

L452-468: Not a conclusion

L469-470: remove “dramatic”

L490: remove “powerful”

L491: remove “great”

L 493: please elaborate on the “many uncertainties and challenges” in your discussion

 

 

Author Response

Please see attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors addressed a timely topic of the effectiveness of SuDS (also known under other names) in stormwater runoff control and protection of the receiving waters against the risk of flooding caused by catchment development.

The study is technically sound, well referenced, and the conclusions are well supported by findings from field measurements and computer modelling. Thus, the manuscript fits well the mandate of the Water journal, and adds new data to the ongoing discussion in the literature concerning SuDS/BMP/LID flow control at the catchment scale.

Two minor comments presented below concern possible improvements in the manuscript clarity and ease of understanding.

1. Limited information is provided for the detention ponds studied with respect to their design, dimensions (surface area, volumes, depth, hydraulic operation - particularly the distribution of outflows) and design motivation (described as “The ponds were designed to withstand rainfall events with return periods of at least 100 years…”, or 65 mm of rainfall over 6 hours.”). What is meant by withstand? Any particular limit on discharges during those design events?
2. Modelling and calibration – any estimates of uncertainties in calibrated peak flows?

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Thank you for your answer to my comments and questions and for revising the manuscript. I have two final comments on the draft, requiring minor changes.

1. Line 155-158. Is it correct that all the stormwater runoff from areas with a combined sewer system was removed from the model? Heavy rainstorms will however produce combined sewer overflows. Is it correct that these CSO discharges were neglected in the model? And can you explain why this is acceptable, also when simulating and analysing peak discharges? A subject to be addressed in the Discussion?
2. Line 447-463, the intro to the Discussion. A large part of this information belongs in, or is already covered in section 2.1. Please reconsider what belongs in section 2.1 and what in this intro to section 5.

Author Response

Please see the attachment

Author Response File: Author Response.docx