Flood prevention schemes: guidance for local authorities

SECTION 4: CHOOSING AND COMMUNICATING

Option selection

4.1 It is important to be clear about the reasons for a risk assessment, and what the assessment will be used for. It is of no value to identify and quantify a risk, and do nothing about it. Section 2 indicates how risks can be utilised in applying risk management. In project appraisal, risk assessment can be used to compare different solutions, and help define preferred options.

4.2 At each stage of a project, different options exist. Project appraisal is intended to assess those options and optimise decision-making. The 3 main factors influencing choice of any option are (i) technical viability and sustainability; (ii) environmental acceptability; and (iii) economic justification. In appraising options and making a choice, these factors have traditionally been considered in a relatively deterministic manner, without reference to the scale of the associated sensitivities or risks. Incorporation of risks into the assessment process enables a more informed choice.

4.3 The best option will be influenced by different factors, including the characteristics of the scheme, the acceptable level of individual risks, and their relative importance. It is not therefore possible to be prescriptive about the selection procedure. In some circumstances, the nature of the risk may override economic arguments. Flood defences are expected by the public to resist high water levels, wave activity, and river flows, with low probabilities of failure. In practice, the consequences of failure may influence the acceptable probability, and possibly the choice of option. Generally, a scheme with a failure mode that could have major consequences will be less acceptable than one involving a minor impact.

4.4 The relative risks between different options can be compared, using an option appraisal matrix, irrespective of how each risk is quantified. For example, some risks may be identified as High/Medium/Low, Serious/Not Serious, Likely/Unlikely, while others may have quantified probabilities. Application of this approach may be appropriate at all stages of appraisal and decision-making. As long as each risk is evaluated consistently for each option, direct comparison is possible. In addition, it can be advantageous to extend the matrix to rank the risks, both by severity or interest (those risks having the gravest consequences), and by scheme type. This can provide a clearer indication of the relative merits of different options and assist in selecting a preferred option.

4.5 An example of an option appraisal matrix and ranking of options is given in Example A. Example B discusses options for a flood prevention scheme.

Example A: Option selection

Option appraisal matrix

Ranking of risks

The top section of the matrix shows how each option A to E performs in terms of risks 1 to 6, in either quantitative (where available) or qualitative terms. In the bottom section, the risks are ranked in decreasing order of priority and the options are ranked according to how they address each risk. For example, from the upper section, option D is the lowest risk solution in relation to Risk 1, whereas option C is the highest, followed closely by option A. This is reflected in the lower section, where it can also be seen that Risk 1 is classified as minor. Similarly, for Risk 4, which is described as paramount, option A scores the lowest risk, whereas option D has the highest. Overall, an initial qualitative review would appear to show that option A offers the lowest risk solution.

Example B: Option selection within a flood prevention scheme

Following severe flooding from a stream running through the centre of a small town, an improved level of protection was considered. The realistic options were (i) enlarge the existing channel, (ii) provide temporary flood storage upstream, or (iii) provide a tunnel diversion for flood flows. Initial economic assessment suggested that all 3 options could have similar benefit-cost ratios, depending on the results of further studies.

Environmental assessment did not favour more than a limited increase in channel capacity through the town. A channel that would have provided a significant increase in capacity would have resulted in a 'concrete canyon' dividing the town.

From a risk viewpoint, a preliminary assessment of ground conditions, in the area where flood storage would need to be provided, suggested that the construction of any raised structure might be difficult. It was considered that costs could well rise through the design and construction phases. There were also risks associated with the provision of flood storage immediately upstream of an inhabited area.

The tunnel was expected to be driven for most of its length through competent sedimentary shales. However, it was acknowledged that tunnels always carry a risk of encountering adverse ground conditions, with a consequent increase in construction costs. With limited channel improvements through the town, there would be significant residual damages, and potential loss of life, when the design capacity was exceeded. However, design and construction costs could be estimated with reasonable accuracy.

It was decided that the risks of high residual damages, plus the environmental disbenefits of the channel, and the increased cost risks of the flood storage option, outweighed the tunnelling risks. In the event the tunnel option was chosen, but adverse ground conditions were encountered and costs increased significantly.

This does not mean that the decision was wrong, though with hindsight it might have been worth carrying out further site investigation to assess the tunnelling risks more fully, and possibly to investigate other tunnel routes.

Presentation of results

4.6 Risk-based approaches can provide more information than a deterministic analysis. This should enable a fuller and better informed assessment of the risks and the robustness of the decision. However, there is a limit to the information processing capabilities of any individual or organisation. Therefore, when communicating results, the emphasis should be on conciseness and clarity. To ensure that the results are accessible to the non-specialist, a summary report may be required, supported by technical annexes.

4.7 Many risk-based methods will provide graphical results in the form of probability distributions that give an immediate and concise impression of variability. Particular care should be taken to give the axes appropriate labels and scales.

4.8 As with any scientific or engineering analysis it is important to state the assumptions and sources of information. Results should be accompanied by a commentary on:

• the approach adopted;
• particular sources of uncertainty in the models used;
• the scope and results of any sensitivity testing
• residual risks;
• the potential impacts of phenomena which are not well understood, or are not represented by the models used, particularly where this may influence the choice of preferred option;
• which input parameters have been modelled as probability distributions, and which have been given deterministic values; and
• how dependency between input parameters has been dealt with and, in particular, where independence has been assumed and why.

4.9 Where there has been a significant input of expert judgment this should be stated, and a commentary provided on:

• the evidence that the expert judgment was based upon;
• the use, if any, made of peer review; and
• sensitivity to variability in expert judgment (including the expert's own estimates of potential variability).