The DWMP blog – 33. How much runoff gets into the sewers?

Background

This episode of the blog was prompted by a short video posted by Peter Melville-Shreeve of runoff overflowing a roof gutter.?So how much runoff actually gets into sewerage systems and where does the rest go??I did a brief review of this for one of the water companies a few years ago and borrow heavily from that work here.?They were concerned at the predicted large increase in flooding from their sewerage systems with future climate change.

This episode of the blog is very much focussed on current modelling practice in the UK using InfoWorks ICM.?However, there may be lessons for of from other jurisdictions and other models.

Current UK approaches to modelling inflow into the sewerage system were originally set up to answer the question; how much water gets into that particular sewer??I think that the question that we now need to ask is; where does the water go?

Those modelling approaches assume that the volume of runoff getting into the sewerage system is roughly linear with the intensity of rainfall.?There is no limit set on the inlet capacity of the systems.?The models are verified against modest rainfall events and then with this linearity they extrapolate to much larger events.

Good practice is to check the verification against reported flooding in historical events, but this process involves a lot of uncertainty.?In part this is because a high intensity event may still cause flooding, but due to overland flow of surface water that cannot get into the sewerage system rather than flooding coming out of the sewerage system.?This “pluvial” flooding requires a different type of model and different engineering solutions.?The situation is complicated even further as it is not necessarily the responsibility of the sewerage undertaker.

Inlet capacity limits

The assumption of linearity is not correct in the real world.?Direct inflows to the systems from roads, roofs or other paved areas will have restricted inlet capacity at gully inlets and drains that will impose a cap on the inflows.?Let us look at the potential limiting inlet capacity for:

• Road drainage
• Roof runoff
• Other property runoff

Road drainage

Gullies are designed based on the width of runoff on the road surface rather than the limiting capacity of the gully itself.?Typical gully spacing gives one gully for every 120 m2 of road surface.

The gully grating has a large inlet capacity and the real limitation on the inflow to the drainage system is the capacity of the connecting pipework.?This typically has a capacity of about 8 l/s.

This translates to a limiting rainfall intensity of about 240 mm/hr.

This assumes that the gullies are regularly maintained so that they are not blocked with silt and that the gratings are not blocked with leaves or litter.?In practice many gullies will have a lower limit.

Roof runoff

Design guides for roof drainage consider both the capacity of the gutter and the capacity of the downpipe.

For gutters fixed at a significant slope, the limiting capacity is the downpipe at about 3.5 l/s.?However, for normal UK practice of gutters fixed with little or no fall, the limiting capacity is the gutter leading to the downpipe with a capacity of typically 1.5 l/s.

An average house is 100 m2 with typically 2 downpipes.

This translates to a limiting rainfall intensity of about 108 mm/hr.

Other property runoff

Paved areas within the property such as patios and driveways may also contribute flows to the sewerage system.?It is particularly common for patios to be connected to systems that were designed for only foul flows.

These would drain through a trapped back inlet gully.?These have a typical capacity of 1 l/s.

Assuming that this is draining an area of 20 m2, this would translate to a limiting rainfall intensity of 180 mm/hr.

The combined picture

For a partially combined system taking some roof and curtilage drainage, an average limit of 140 mm/hr would be appropriate.?For a fully combined system that is also taking road drainage then this limit could be higher at 200 mm/hr.

How to represent it

As we impose limits on the inlet capacity of the sewerage systems, we need to also consider what happens to the water that cannot get in, so that we have an integrated view of surface water drainage.?However, that does not necessarily mean that we need an integrated model.?We could use a model that represents the flow that gets into the sewerage systems and discards the excess and then have a separate model that represents overland flow of water that is unable to enter the sewerage systems.?

There are several ways to represent the limits on flows entering the sewerage systems, ranging from the crude to the detailed.

The simplest would be to not consider events where the peak rainfall exceeded the notional inlet limits as set out above.?We need to allow for some water being stored on the ground surface and entering the sewerage system after the peak rainfall.?So I suggest setting the limit as the average intensity over the worst 8 minutes of the rainfall event rather than the instantaneous peak.?As short duration events have higher peak intensities, we can estimate the minimum duration to be considered for each probability of storm occurrence.?I assessed this for an example catchment in south-east England (other parts of England are available).?This shows an estimate of the minimum duration to be used for each annual probability.?(It would need more work to prove that this crude approach was reliable).

A more complex method would require changes being made to the coding of the runoff model so that a limiting inlet capacity could be defined for each land use as a limiting flow per hectare of surface.?Ideally excess flow should be temporarily stored to give the extended period of flow into the sewers.?This would be easy to implement for new models but would require those changes to the coding being made.

Even more complex would be to add more detail to the model to apply a limiting entry capacity using an orifice between the subcatchments and the pipe network to represent the appropriate flow limit for the sub-catchment.?This adds some complexity to the model, but could be done using the existing model coding.

The most complex approach would be to use a rainfall on grid runoff model and represent each gully inlet to the sewerage system.?This would add a huge amount of complexity to the model particularly to represent individual roof drainage and other drained areas.?Also it would not necessarily give good accuracy as it would depend on an exact match between landform from the digital terrain model and the location of each gully entry.?This approach would be appropriate for design studies of new developments or for detailed assessment of local pluvial flooding problems, but is unlikely to be useful more generally.

Considering exceedance flow

For all of these methods it would also be necessary to consider what happens to the excess water that is unable to enter the sewerage system.?The simplest approach suggested above would not allow the use of a fully integrated model.?The other approaches would allow this; but that does not mean that it is necessarily the correct approach.?More on this in the next episode of the blog.