Flood prevention schemes: guidance for local authorities

SECTION 2: USING RISK ASSESSMENT WITHIN PROJECT APPRAISAL

Introduction

2.1 There is wide acceptance that significant benefits can be gained from a more systematic approach to the assessment and management of risks. Risk assessment has often been applied instinctively; risks have remained implicit, and have been managed by judgment informed by experience. The assessment should, instead, make the risks explicit, formally describing them and making them easier to manage.

2.2 Risk assessment should be a 'live' process intimately connected with the study or project. It should be reviewed and reconsidered as the project proceeds, to account for any changes in design or intention. Associated documents, such as risk registers, should be available to those responsible for any subsequent stages of project development.

2.3 It is vital that the appraisal is linked to reality. An assessment that predicts severe flooding every year when no floods have occurred for 30 years is probably wrong. Similarly, input data should be carefully reviewed and cross-checked with similar data for nearby areas. This should provide some basis for estimating the likely uncertainty in the inputs.

2.4 Bounds have been established on what constitutes an acceptable health and safety risk (references 13 to 16). Acceptability is generally considered to be dependent on (i) the degree of control an individual has over the risk; (ii) the familiarity of the risk; and (iii) the number of individuals who may be killed or injured in any one incident. While these issues have not generally featured in appraisals of flood prevention schemes, they could be a legitimate consideration at some sites, and should be addressed in risk assessments in such cases.

Risk assessment framework

2.5 Figure 2.1 shows a flow chart framework for risk assessment within project appraisal.

Figure 2.1 Framework for risk assessment and management within project appraisal

2.6 Risks need to be considered as an integral part of the project appraisal process. That process, and hence risk assessment, should be carried out for all stages of a project. These may include large-scale planning and strategic studies (Chapter 3), in addition to the scheme development and design stage.

2.7 For the particular stage under consideration, it is important to define the objectives and aims of the programme or project; identify constraints; and determine options. The risk assessment should, for example, identify all significant hazards or threats to the objectives; estimate the likelihood and consequences of outcomes; and evaluate the importance, or significance of risks. Risks should be assessed for all options. Different options may involve different hazards and threats. They may also be affected by uncertainties to differing extents.

Approaches

2.8 There are various approaches to risk assessment, depending on the specific application. Guidance on this aspect is given in Table 2.1. The initial approach should be at a broad-brush level to identify, without complication, the key risks, issues and solutions. From the table, the most appropriate initial method is 'risk screening', using checklists and risk registers. This is a qualitative method. A ranking of risks using qualitative scales provides sufficient information for many decisions. This approach can highlight areas where detailed quantitative analyses are necessary. Such analyses require more specialised techniques and these are discussed separately in section 3.

Table 2.1 Risk assessment approaches

Approach	Risk screening	Qualitative risk estimation	Quantitative risk assessment
Tools	Check lists, risk registers	Scoring and weighting (also known as multi- criteria approaches and multi-attribute techniques)	Fault trees, decision trees, Monte Carlo modelling
Scope	Broad	Specific	Specific
Nature	Qualitative	Qualitative	Quantitative
Applications for flood defence	All, for audit purposes Generally, check lists will assist with hazard identification Construction Risks	Catchment management plans Strategic plans (with 'tangible'and 'intangible impacts) 1	Strategic plans (with mainly 'tangible' impacts) 1 Schemes (design and appraisal)

¹ 'Intangible' impacts are impacts beyond those associated with damage to property, infrastructure, etc (ie 'tangible' impacts). They include damage to areas of environmental significance, impacts on recreational usage, high levels of stress, etc (Chapter 5).

2.9 A tiered approach is recommended for the structuring of risk assessments. This begins with the broadest view of the problem or issue. Next, all risks that could have a bearing on the issue should be identified. Some of these may prove to be insignificant and may be eliminated. The next logical step is to identify the most important risks among those remaining, usually by assessing the associated likelihoods and consequences.

Risk registers

2.10 Risk registers help to identify hazards, to record information about risks, and to document decisions taken. They also help to manage the risks, and are suitable for a wide range of projects, from large-scale planning to scheme implementation. The structure of the register should reflect the overall risk assessment and management process. An example format is given in Table 2.2. Issues affecting compilation of the register are discussed in paragraphs 2.11-2.25.

Table 2.2. Example of a format for a risk register

Risk identification (column 1, Table 2.2)

2.11 Guidance on the identification of risks is given in Annex A, which lists a number of risk categories as follows:-

• Funding
• Statutory framework
• Socio-economic framework
• Management of project development
• Strategic
• Impact of natural processes
• Performance of existing works
• Ecology, heritage and amenity
• Human intervention
• Design parameters
• Knowledge of principles/methods
• Scheme performance and response
• Procurement and construction
• Operation and maintenance

2.12 Some of the categories are non-project specific, a number are concerned with project development, and others are related to project implementation. The categories, and examples of associated risk areas, are considered in more detail in Annex A. It should be noted that the risk areas given in the annex are indicative. They will not be relevant to all projects. Further, the extent to which each needs to be considered will vary from project to project. Specific risks within each area are not given due to their potentially large number, and the fact that many would be project specific. Approaches such as 'brainstorming' should be used to ensure that all relevant areas are covered.

2.13 Risk assessment in large-scale planning and strategic studies needs to be commensurate with the decisions being taken, and the level of information available. All major risk areas should be identified, but it is likely that many will remain to be considered and resolved at the more detailed design stage.

Probability assessment (column 2, Table 2.2)

2.14 The probability of each risk needs to be identified. As a first stage, this may reflect the judgment of a knowledgeable individual or panel, based on best available information. The probability may be rated in one of a number of 'bands'. At its simplest, these may be ranked in High/Medium/Low categories. Alternatively, probability may be given on a scale of 1 to 5. Table 2.3 shows a suggested 5 point scale (see also reference 2).

Table 2.3 Suggested 5 point scale for assessing probability

Likelihood	Guidance	Frequency of occurrence
Frequent	Likely to occur many times during the period of concern	100/N
Probable	Several times during the period of concern	10/N
Occasional	Some time during the period of concern	1/N
Remote	Unlikely but possible (eg one in 10 times the duration of concern)	1/10N
Improbable	Can be assumed, for most purposes, that it will not occur	1/100N

NB: In the table, N is the period of concern. For a flood prevention scheme design, N may represent 100 years, while for a storm during construction its value may be one year. There is, of course, no reason not to use actual estimated probabilities where they are available, even in the initial risk register, rather than a scale or category. Use of probabilities may be more transparent, and will be an eventual goal for most significant risk areas in the later stages of assessment.

2.15 Where a scheme involves a number of individual lengths or phases, the failure probability of the system will depend on the failure probabilities for each part. It will also depend on the relationship between the parts; that is, whether they are dependent or independent. The assessment of probabilities in such situations may require expert judgment, modelling or other specialised techniques as described in section 3. This too may be required when assessing breach probabilities and their change over time.

Evaluation of consequences (columns 3,4, Table 2.2)

2.16 The consequences associated with flood defence include those in the economic, environmental, and social categories. They may not be confined to the area of the engineering works. For example, there could be significant impacts in distant downstream locations. Both remote and local impacts need to be identified.

2.17 Consequences and impacts also need to be evaluated. As with probabilities, it may be appropriate to use actual consequences (eg monetary values) where available, rather than a scaled version (eg very high, high, medium, and low).

2.18 Scales for probabilities and consequences should be similar, and designed so that their combination reflects the desired weighting. An example of how risks may be characterised and ranked in terms of probability and consequence (when these are quantified) is given in Table 2.4.

Table 2.4 Example of risk ranking in terms of probability and consequence

Risk	Probability of occurrence	Consequence (anticipated cost if risk arises) (£K)	Risk value
			(£K)	(Rank)
R1 Contaminated ground	5%	10	0.5	(4)
R2 Dewatering - property damage	25%	400	100	(1)
R3 Mine shaft collapse	5%	500	25	(2)
R4 Severe flooding	1%	100	1	(3)

This approach can be used to assess the acceptability and response to the risk; for example, it may be judged that risk R2 is unacceptable, and must be removed by changing the scheme concept or method. Risk R3 may be judged undesirable, and risk mitigation investigated and implemented as appropriate. Risk R4 may be accepted on the basis of low probability, and the cost of reducing flooding, but managed by detection and warning. Risk R1 may be accepted and dealt with if it arises.

2.19 The importance of particular risks will vary according to the stage of the appraisal. For example, strategic risks will command a very high priority during strategic studies, but only a medium priority in the context of scheme implementation, since major issues should already have been considered. On the other hand, procurement risks will have a very low priority during any large-scale planning stage, whereas they will be high priority at scheme stage.

2.20 Probabilistic analysis of consequences can add another dimension to appraisal. In certain situations, this may provide useful insights. However, it will often contribute little to achieving the overall project objectives. Under most circumstances, it is therefore preferable to adopt conventional methods of assessing consequences, and avoid quantitative analysis.

Mitigation measures and residual risks (columns 5,6,7, Table 2.2)

2.21 Mitigation measures are generally designed to reduce either the probability or consequence of a risk. There are many types of such measures and it is not possible to give a definitive list. However, the risk register and associated analysis should help to identify those that are most appropriate.

2.22 Some mitigation measures may remove the risk entirely; others may reduce it, leaving a residual probability and consequence. The importance of this must be assessed. If the residual risk is too high to be accepted, other mitigation measures should be considered.

2.23 Mitigation measures should be justified by their impact on the risk. In some cases, this will be directly measurable in economic or financial terms. Their effectiveness can be investigated by comparing the cost (ie expected value) of the risk without the mitigation measure, against the cost of the residual risk plus the mitigation measure.

2.24 Where risks and costs cannot be easily quantified, mitigation action may be justified on the basis of judgment. As well as reducing the mean or expected cost of risk, mitigation measures may also reduce the associated uncertainty. This may be important, particularly for complex risks with uncertain or unpredictable consequences. In these cases, it may be better to implement measures to control the risk and reduce the uncertainty. The register enables an auditable record of responsibility for mitigation actions.

2.25 It may seem difficult to estimate probabilities and consequences in the risk register, even with a relatively coarse scoring system. However, even where there is great uncertainty, the register is still a useful tool to establish and monitor appropriate mitigation measures, although judgment will be needed to decide the appropriate level of resources to mitigate unquantified risk

Dealing with climate change

2.26 There is, currently, reasonable agreement that, due to climate change, relative sea levels will rise over time, though the rate of rise is less certain. There is much less certainty regarding other potential impacts, such as the frequency, duration or intensity of storms and the consequences for extreme wave activity or river run-off. The indications are that such storm activity may well increase, but quantification of the impacts is difficult.

2.27 The predicted increases in sea levels and peak flows discussed in Chapter 4 are based on a recent study by the Babtie Group for the Scottish Executive and have high levels of uncertainty (reference 25). In addition, there is considerable variation in the magnitude of the increases depending on factors such as geographical location, the time frame under consideration, and the nature of the flooding, that is whether it is fluvial or coastal. It is therefore difficult to recommend robust climate change scenarios for the purpose of risk assessment.

2.28 In the future, it may be possible to make recommendations for the adjustment of input rainfall to flood event or continuous simulation models to take account of climate change, but this is not currently feasible. As an interim measure, for proposed flood prevention schemes in east and southwest Scotland, it would be reasonable to include in a sensitivity analysis an assessment of the impact of progressively increasing, over the next 50 years, the flow estimates in the flood frequency curve by up to 15% due to climate change, along with other considerations of uncertainty. In north and northwest Scotland, the predicted increases in peak flows are relatively small and do not appear to justify a detailed risk analysis. However, if a proposed scheme in that area, or in other parts of the country, is likely to be affected by an increase in sea levels, it may be appropriate to conduct a similar sensitivity analysis, consistent with the predicted increases and timescales referred to in Chapter 4.