The DWMP blog – Episode 16. No detriment - or levelling sideways?

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).

In this episode I consider no-detriment policies.?That is, that when improving performance in one area we should not cause any deterioration in performance elsewhere.?Is this a barrier to levelling up the performance for worst served customers or most impacted areas of the environment?

This particular episode was written to try out ideas with no conclusion in mind.?My conclusion surprised me; it might surprise you.

Continuous or discrete measures

The first challenge is how we structure the performance measures that we use to define whether there has been any detriment.?Many of them are measured as yes-no discrete variables rather than continuous variables.

For example, flood risk is counted as the number of properties with an annual probability of flooding of more than 1:30 or 1:50.?For the new Ofwat 1:50 standard, if a property has a probability of flooding of 1:49 then it counts as a failure; if it has a probability of 1:51 then it counts as a pass.?A small detriment in performance could then tip a property just over the line.?However, a property that currently has a very small probability of flooding of 1:100 could have its probability almost doubled and still not show as a detriment against the measure

If instead we reported the total change in probability of property flooding for an area, then we would have a better representation of the detriment.

The situation for river quality is at first sight similar, with quality bands of poor, moderate, good and high.?Again, a small detriment could tip a waterbody from good to moderate or conversely a large detriment could not cause any change in banding depending on where in the band it currently sat.?In practice however, the EA will look at predicted changes in quality within the bands as well as the band boundaries.

Different causes of detriment?

In current application of no-detriment rules we often react differently depending on how the detriment is caused.

Climate change will cause an increase in flood risk for everyone.?There is no requirement that we must mitigate this detriment and reduce it to zero.?However, if new housing is to be built, then we would commonly require that the detriment on flood risk is reduced to zero through appropriate measures.?There is an implicit attitude here that climate change is inevitable and cannot be stopped whereas having adequate houses to live in is optional and we can choose not to have it.?Do we really think that?

Similarly, if we construct new sewerage to reduce high probability flooding for some customers, we generally would be required to mitigate any detriment to flood risk to other properties.?If we measure this as the number of properties greater than a threshold probability then this might be reasonable; but if we measured it as the real change in probability; perhaps from 1:100 to 1:90 then would this be reasonable.

For water quality we also have different causes of detriment.?As for flooding, climate change will cause detriment everywhere.?If that detriment is due to increase in average temperature and reduction in summer river flow with no change in sewerage discharges then it is not clear whether action would be taken to mitigate it straight away.?However, if climate change caused an increase in discharge from the sewerage system without any other changes in the system, then the water company would be expected to mitigate this as it is their climate change.

If the detriment was due to proposed development, then it would need to be mitigated or planning permission would be opposed.?Again, an assumption that development is somehow optional.

Should we allow detriment

In an ideal world we would not increase the probability of flooding, even if that probability would still be very low and we would not cause any deterioration in the natural environment, no matter how small.?However, it is difficult to see how that can be an absolute requirement.?We all already have an impact on flood risk and the environment just by existing and we cannot go back to a time when the human race did not have those impacts.?So, if we allow existing impact then we must be open to allowing some additional impact due to increasing population or due to the climate change that we have already caused.?But how much can we allow?

In assessing what level of detriment is acceptable we must bear in mind the errors and uncertainties in our predictions of what that detriment will be.?This is particularly the case for discrete step change standards where we do not really know how close we are to the standard.?The use of continuous measures is less susceptible to error as there will be similar errors for the before and after situation and so the magnitude of the change could be reasonably accurate.

But we probably still do need absolute targets that we should achieve and should defend against detriment.?However, these targets should not be set unreasonably high.

For river quality the EA typically restricts detriment to no more than a 10% change of pollutant concentrations within a river class and nothing that would prevent the achievement of Good quality.?That is a sensible approach, although some sensitivity testing is required to consider uncertainties in the prediction of the Good standard.

Perhaps we need something similar for flooding, with no more than a 10% change in total probability of flooding for the properties in an area and no property reaching an unacceptable probability of flooding.?But what is that unacceptable probability??The current standard is 1:30 annual probability but water companies are allowed to have a number of properties failing this before being penalised.?So, the level that is unacceptable is currently more than this, perhaps 1:20.?We should aim to improve this over time, but we will never reach 1:50 as an absolute standard that all properties must achieve.

Incremental detriment

There is a danger that if we allow a small detriment for a single cause – for example a new development – and then allow more for the next cause that we end up with an unacceptable incremental detriment.

DWMPs with their long-term planning horizon give us a mechanism to manage this issue.?The permitted detriment should be over the 25-year planning horizon taking into account all of the changes that need to be considered in the DWMP.?Any additional causes of change would have to have greater restrictions.