elements and failure modes associated with the type of dam. This is
achieved through special ‘Location, Cause, Indicator’ (LCI) charts which
consider the components of the dam and the mechanisms of deterioration
or failure. A numerical value is assigned to each of these in terms of con-
sequence, likelihood of occurrence, and degree of confidence in these
values. The outcome is the identification of specific problems together
with aggregate numerical criticality and risk scores. The latter can be
deployed to assist in the prioritization of resource allocation and in assess-
ing the need for more rigorous dam-break analysis and contingency plan-
ning.
The application of risk assessment methodologies in the manage-
ment of dam safety programmes has aroused considerable interest, with
co-ordinated research effort extending across several countries. A Cana-
dian perspective on the issue is presented in Hartford and Stewart (2002),
and it is further explored in ICOLD Bulletin 130 (ICOLD, 2005). Method-
ologies for quantitative risk assessment (QRA) have aroused particular
interest in a number of quarters in recent years, and QRA is addressed in
some depth in Hartford and Baecher (2004).
In the UK, a QRA methodology based on the concept of an integ-
rated approach to the assessment of overall risk to a dam from specific
agencies, e.g. extreme flood events etc., has recently been developed
(Brown and Gosden, 2004). This incorporates features of the FMECA
approach to risk assessment proposed in Hughes et al. (2000), which it was
intended to supersede.
Brown and Gosden (2004) defines a procedure for determination of
a screening level assessment of the absolute risk represented by an indi-
vidual dam and reservoir. Absolute risk is then compared with the stand-
ards for public safety deemed applicable in other contexts such as
high-hazard industries and acceptability determined. The estimated
absolute risk can also be ranked relative to that determined for other
reservoirs within a population. The procedure requires estimation of the
annual probability of failure associated with:
- an extreme rainfall event;
- presence of an upstream reservoir;
- loss of internal stability (embankment and appurtenant works (e.g.
outlet works)); - a matrix of other significant threats (e.g. seismicity etc.).
(The extreme rainfall event of 1 above should not be regarded as in con-
flict with Section 4.2 and Table 4.1).
The overall annual probability of failure is estimated, the individual
probabilities subsequently being reviewed to confirm that they in turn rep-
resent appropriate estimates.
The consequence of breaching and failure is estimated from