Assessing Low Probability Catastrophic Events

For many catastrophic events such as earthquakes, cyclones, floods, bushfires etc. the assessment of their consequences is very complex.

Not only are the consequences of these events significant in economic terms, they will vary substantially with the intensity of the event itself.

For example, in assessing the consequences of a flood mitigation scheme one needs to consider the intensity of the rainfall patterns. In most cases the scheme will be designed to withstand a certain storm intensity or recurrence interval, e.g. 1 in 100 years.

To determine the level of investment that is warranted today, organizations need to assess the net present value (NPV) of the consequences of failure if nothing is done.

The most appropriate method is to determine the consequences of failure for the probability of each storm event. We need to determine the economic value of the consequences of the different probability floods and plot these on a probability graph as shown below.

The consequences of failure will rise substantially with the low probability/high intensity flooding and unless the catchments have a high retardation capacity, then the consequences will rise substantially.

Using a timescale related to the probability we now have the average annual business risk exposure, which can be expressed as an average annual annuity of the area under the graph.

The risk reduction achieved by different flood mitigation options can then be assessed by removing the portion of the average annual annuity attributable to the probability on which the treatment option has been designed (e.g. 1:250 years). This is shown by the following graph.

The risk reduction is taken horizontally through the area based on the design probability. The area above this point is then assessed and brought forward as an annual annuity.

This has been done because, if flood mitigation works are installed to eliminate risk up to that probability, then even should floods occur in excess of this design intensity, their impact will be mitigated substantially by the works in place and will offer a good basis on which emergency mitigation works such as sandbags can be implemented to reduce even higher risk.

This assumption is questionable. It could be assumed that if the event was so catastrophic that the protection devices (levees) failed, then the consequences would be as great. Therefore the consequences brought forward as an annuity should include the area beyond the probability.

An alternative to this progressive annual approach would be to look at the single event consequences occurring half way through their probability cycle. This approach is not considered valid, as we cannot predict when these catastrophic events would occur.

A further analysis could look at the true economics of these assumptions, based on actual events. However it must be remembered that the key to the economic evaluation is whether or not a discounted cash flow should be used. If it is, then these long-term events will be significantly discounted or only lesser works will be able to be justified.

A similar type of assessment can be made for the overtopping of dams and other events such as cyclonic wind and associated rainfall. In this instance the probability would be related to the intensity of the storm matched to the probability of it actually passing directly over the point in question.

More detailed analysis is now available from the Bureau of Meteorology on these issues including cyclonic storm or surge assessments.

An objective comparison of the different risk cost exposures involves a more sophisticated ranking system that converts our perceived risks into economic values. The different impacts of failure are weighted according to the relative importance and these are multiplied into the consequential costs to derive the weighted risk exposures.

Risk analysis programs allow us to rank the assets in terms of business risk. It is useful for identifying the relative criticality of the vast majority of assets in an organization. An asset with a high consequence of failure may not be too critical if it is operating under favorable conditions and hence has a lower probability of failure. However, at the extreme ends of the scale, organizations should pay attention to those assets which:

• Have extraordinarily high consequences of failure even if the probability of failure is very remote
• Are sure to fail in the near future, where the impact of the failure would lead to inconveniences in business operation even if the consequences of failure are not critical.