Story inc. 121k imageSerious flood-related reservoir safety incidents - most notably at Ulley Reservoir in June 2007 - have at last stirred concerns about floods and reservoir safety "at home".
Nevertheless, the community of British dam engineers still seems happier to point the finger at technical specialists who have dared to deliver new procedures for rainfall depth-duration-frequency estimation (see discussion below) than acknowledge that 1975 estimates of Probable Maximum Precipitation are (and always were) - like the Ulley spillway channel - capable of being breached.
Some of the links from this page are broken. Hopefully, this is a temporary problem consequent on the offloading of some responsibilities for reservoir safety from Defra to (in England and Wales) the Environment Agency.
The remainder of this page (including the next paragraph) is all unchanged. Let us hope that attitudes are changing.
It is worrying that Government appears to have accepted the suggestion that estimates of PMP might be used to temper empirical estimates of T-year rainfall depths.
Over the last 20 or so years, many works have been undertaken to enhance flood safety at UK reservoirs. These initiatives have been prompted or supported by the Flood Studies Report, the ICE Floods and Reservoir Safety guide, and the Reservoirs Act.
Further, I have detected, or imagined, signs that some in the industry believe that UK flood estimation procedures have yielded designs that are over-safe.
With no recent near-misses reported, reservoir flood safety is not a hot topic. But one only has to look at events in other fields - where the public has lost sight of (or been unaware of) imposed risks - to know that one catastrophe (natural or otherwise) will change all. Perhaps this is why no one seems to have gone on the record to suggest that UK reservoir flood standards are over-safe and, hence, excessively costly.
Another take on attitudes is perhaps that existing provisions against floods are generally seen as adequate relative to other reservoir safety concerns. These include the structural integrity of dams, effective monitoring & maintenance, and (non-reservoir) embankments that might unintentionally act as dams.
Without public or political pressure to adopt higher flood standards, and with ever-present demands for financial prudency, it is perhaps understandable if the industry seeks to maintain the status quo.
So what's the problem?
Two developments perturb the status quo:
The typically higher 10000-year design rainfall estimates given by the FEH rainfall frequency procedure (in comparison to the FSR rainfall frequency procedure);
Concerns that climate change may lead to increased storminess.
Only the first of these factors is discussed here.
It is right that the profession is pausing before deciding whether, and how, to implement the FEH rainfall frequency procedure in reservoir safety assessments. Points to note are:
The higher estimates principally reflect the new methods devised, not the additional data analysed;
The FEH rainfall frequency procedure was not explicitly developed for use at the very long return periods relevant to reservoir safety (see footnote);
Acceptance of the new procedure will in many cases lead to contradictions with the FSR estimates of Probable Maximum Precipitation, which have not been reworked;
Although a somewhat limp statement, acceptance of the new procedure does not change the actual risk of an over-topping event occurring; it just changes the assessed risk.
An exploratory report, and Government response, are available at DEFRA floods and reservoir safety.
Any change of flood estimation procedure potentially presents a problem to those responsible for major structures. Estimates of extreme design floods - such as the 10000-year rainfall depth or the Probable Maximum Flood - are inevitably based on an extrapolation of some kind or other. Because of their typically confined, complicated and bespoke nature, reservoir spillways tend to be very costly to re-engineer.
The decision to switch from Procedure A to Procedure B needs to be carefully considered, weighing up factors such as:
The improved performance - in so far as this can be verified - of the new procedure;
The extent to which Procedure B has a stronger theoretical or conceptual basis than Procedure A, thus increasing confidence in extrapolations;
Whether the improvement (in performance and confidence) is large enough to warrant the disruption and expense of change;
Whether the new procedure is to be routinely adopted in reservoir safety inspections, or invoked only as and when spillway renovation is required.
The research hydrologist can help to inform those taking the decision to change procedures. However, the decision lies with those senior members of the dam engineering profession who make recommendations to the Institution of Civil Engineers and to Government.
If these decision-makers believe that "the flood numbers" are already big enough they will need to either:
Reject Procedure B as inappropriate or unproven;
Accept Procedure B but relax flood standards (in Table 1 of the ICE Guide) in compensation;
or:
Arrange for further research.
This latter approach appears to be the one being taken, see DEFRA floods and reservoir safety.
I have produced this webpage to guard against misconstruction in case a flood-induced dam failure occurs sooner rather than later.
The Government department with responsibility for reservoir safety (now DEFRA) was represented at planning meetings for the FEH research programme, at which it was determined that rainfall frequency research should be targeted principally at meeting river flood protection requirements. The initial contract for the rainfall frequency study specified an upper return-period limit of 200 years. On advice from the FEH Advisory Group, and in recognition of the capability of the emerging method, this limit was later raised to 2000 years. Back.