The DWMP blog – Episode 13. How much flood risk is too much?

If you haven’t already seen the earlier episodes in this series, they are all here (https://tinyurl.com/MartinOsborneArticles) I suggest that you start from Episode 1 (https://tinyurl.com/DWMP-blog).

For people who suffer the trauma of being flooded, any risk of flooding is too much.?But how should we convert that into performance targets?

The National Audit Office report of 2004 discussed the then key metrics of sewer flooding of those at risk with an annual probability of more than 1:10.?Water companies were then also reporting properties at risk at 1:20 and had a design target of 1:30.?The new flood risk metric introduced alongside the DWMPs is to report 1:50.?This shows the change in attitudes to flood risk and the progress that we have made, but where should we stop??Is the long-term target 1:100, 1:200, 0?

Causes of flooding

There are two main causes of flooding from sewerage systems: blockages and excess flow due to intense rainfall.?(Sewer collapses and pumping station failures are much less of a problem.) ?Most sewer flooding is now due to blockages, partly because of investment to increase sewer capacity to cope with rainfall and partly because of the increasing use of wet wipes that do not break down in the sewers and contribute to blockages.?However, climate change is increasing rainfall intensities and the flooding that it causes.

Both of these causes of flooding are random in time and to some extent in location.?Properties in low spots will always be at some risk of flooding from either cause and will tend to suffer repeat flooding.?If they are in areas with limited sewer capacity or an area with a lot of wet wipes or fats discharged to the sewer (for example just downstream of a badly run chip shop) then that flood risk will be greater.?

Performance measures

There are two measures of flood performance: how many properties have actually been flooded each year, and what predictive modelling shows as the probability of future flooding.?Actual performance will vary from year to year depending on rainfall and blockages.?The predictive modelling will give the average probability over many years.

Normal practice is to use measurements of actual flooding for both blockages and lack of capacity but to use predictive modelling only for lack of capacity.?This means that we are not trying to predict the future for the biggest cause of flooding.?Using analysis of past incidents and sewer flow conditions we can predict where blockages are likely to be a problem; and most companies do this to help plan their sewer cleansing programmes.?I think that we should try to map which properties have a 1:50 probability of flooding from both causes to give the full picture of flood risk.?It is difficult to predict how the risk from blockage will change in the future as this depends on changing behaviour with fat and wet wipes and with changing cleansing programmes.

What is flood risk?

By flood risk we mean the combination of the probability of an event happening and the damage or impact that it will cause.?The damage could be measured as the financial cost or as the number of properties affected.?A graph of flood probability against impact is something like the graph below.

At the left-hand end is a catastrophic event that would flood many properties, but may never happen.?At the right-hand end is an event that will happen frequently but will not cause any impact.?The total flood risk is given by adding every bit of the graph to get the area under the curve.?This gives the number of properties expected to flood on average each year.

It is useful to redraw the graph as the risk (that is probability times impact) against probability.?This now shows that the catastrophic event poses little risk as it will probably never happen.?The biggest contribution to the overall risk is the band of events in the middle.?

Investments that we make in the sewerage system will push the right-hand end of the graph inwards and push the middle part of the graph down.?Improving the sewerage system will have little effect on the left-hand end of the graph.?This risk in extreme events will only be significantly reduced by protection measures for the individual properties that are at risk or modifying the ground slope to keep overland flow away from properties.

Summary

So, measuring risk to very extreme events is not useful to assess or plan sewerage system investments.?It is the middle range events of about 1:30 or 1:50 that we should be targeting and where we should measure performance.?However, we should also assess 1:100 or 1:200 probability events to plan the wider drainage system including flood protection and overland flood pathways.