Discussion: Flood plans in England Free

I would add that there is one key principle to the Reservoirs Act that could be added to Table 1 in the paper, namely that regulation should be proportionate to risk. Section 2A of the Act allowed for the risk designation of reservoirs to provide more targeted regulation of large raised reservoirs with the ‘high risk’ and ‘not-high risk’ categories.

An unintended consequence of requiring flood plans for not-high risk large raised reservoirs is to require undertakers to appoint a supervising engineer (SE), if they had not already done so for other reasons, to certify the plan. Although this appointment does not have to ‘be at all times’ it does require, in my view, a site visit in order to certify the plan on first drafting and when major changes are made. If a not-high risk reservoir does not require an SE for day-to-day safety reasons, why does it (a) require a flood plan and (b) certification of the flood plan by a Qualified Civil Engineer?

In the final paragraph of the paper, I am not sure section 12AA(3) certificate will provide much useful information on plan maintenance on its own as the pro-forma certificate does not include any information on the reason or nature of the update.

We agree that regulation should be proportionate to risk, and this is a slightly different expression of principle 3 in Table 1 of the paper. This is also confirmed in the recent ICOLD (International Committee on Large Dams) Bulletin 191 (ICOLD, 2021), which states in section 7.1 that:

It is therefore disappointing that the Defra guidance  does not provide estimates of costs or the carbon footprint of preparing and exercising the plans, especially ‘full site simulations’, to justify the frequencies suggested in the Defra guidance. Nor does the guidance allow the undertaker/appointed engineer to determine an appropriate frequency of testing for which the cost is proportionate to the risks at each specific reservoir. This is discussed further below.

We agree that it appears illogical and inconsistent with the Reservoirs Act to require an appointed engineer to certify flood plans for not high risk (NHR) reservoirs, when the Reservoirs Act states that a SE is not required for NHR reservoirs.

ICOLD Bulletin 191 states in section 3.3: Dam safety assurance:

Safety assurance involves demonstrating that the desired level of safety management has been achieved. This is done by regularly verifying the effectiveness of every stage of the safety management process.

Governments and dam owners both have essential roles in providing assurance of the safety of dams under their control.

We note that the act and direction do not cover how maintenance of flood plans will be assured. We agree that a 12AA(3) certificate provides evidence that a plan has been updated but not what other maintenance of the plan has been carried out. The authors suggest that for high-risk reservoirs the guidance for the content of a SEs annual statement in section D5.2 of the Guide to the Reservoirs Act (ICE, 2014) is replaced by the text in Table 1 of this discussion note, below.

If the requirement for NHR to have flood plans is retained, then a section 12AA(3) certificate should be issued on first preparation to confirm the content of the plan meets the direction. As there is no SE at NHR reservoirs the simplest way for the regulator to monitor the update of plans is for the regulator to periodically ask for a copy.

The Defra guidance suggests full-scale simulation at one dam every ten years for undertakers with 15 or fewer reservoirs, or every five years for undertakers who own 16 or more dams. This does not take into account the consequence category of the dam, and in many cases the cost is likely to be disproportionate. This is illustrated by the simple sensitivity test in Figure 1, considering two scenarios of an undertaker with a single dam, and one with a portfolio of 15 dams, with other assumptions as shown in Table 2. If a value to prevent a fatality (VPF) of £1.7M, and a disproportionality factor of 10 is used, as per section 10.3 of risk assessment for reservoir safety (RARS) (Environment Agency, 2013), then the cost of a full-scale simulation is disproportionate when the average societal loss of life (ASLL) is less than 50 for a single dam, or 5 for a large (>15 dam) portfolio.

An alternative description is that the cost of carrying out a full-scale simulation on a category D or NHR dams with ASLL of 0.001 or less is equivalent to a minimum of £800 million per life saved, which is clearly grossly disproportionate. At a time of climate crisis, it is also questionable whether the carbon footprint of travel and mobilising equipment for a full simulation is appropriate for these low consequence dams.

Cleary these values would change significantly (e.g. by an order of magnitude) for base values varying from those in Table 2 – for example, if the existing probability of failure was higher or lower than that used in the sensitivity. However, it establishes the principle that full site simulation is unlikely to be proportionate for lower hazard and NHR dams. It also confirms that exercising is worthwhile for dams with an ASLL in excess of 100 lives.

Based on Figure 14 of Balmforth (2021) full-scale exercising could be disproportionate for up to 50% of UK dams. Hence from this analysis, the authors suggest that a risk-based approach to exercising flood plans could be as set out in Table 3.

In addition to those in the paper the authors suggest that:

• In implementing and updating the direction and associated documents the regulator takes cognisance of the guidance in ICOLD Bulletin 191.

• The reservoir safety regulator, as part of risk-based regulation, should ensure that lessons learnt from full-scale simulation at high consequence dams are promulgated to undertakers who own low hazard dams.