Landslides risk assessment

Jun 13th, 2018

Natural and man-made slopes portfolio require Landslides risk assessment in order to ensure sustainable management. Slopes present hazardous geo-morphological processes such as landslides which can occur discretely (some deformation, cracking followed by a large movement) or “continuously” (deformations ranging between a few millimeters and several centimeters per years with occasional accelerations followed by a return to “average velocity”).

The Lidong Village landslide, in Yaxi Township in Lishui, Zhejiang Province in East China occurred at about 1050 pm on 13th November.

Shall we manage hazards or risks?

Any slope, anywhere in the world, has a probability of occurrence (of “first” failure). Continuously sliding phenomena have a probability of sudden acceleration (multiple “failures” over a long period of time).

Technically, Reliability is the chance of not having a failure, so it is the numeric complement to one of the probability of failure, or, in the probability the hazard will not occur..

Enhancing a single or multiple slopes reliability means reducing their probability of failure by the means of mitigation. However, managing portfolios of slopes, like for any other hazards portfolio, has to make sense from both the economic and safety point of view.

Since various failures within a portfolio will generally cause different consequences (to users, public, environment, traffic, infrastructure, etc.), enhancing the reliability alone will be fallacious from the economic, safety and social point of view.

To optimize the reliability enhancement approach one has to combine the probability of failure and failure consequences, hence look at the risks that each slope can generate. Thus and efficient Reliability Enhancement effort should be seen as a “enterprise risk management” action, even if the “enterprise” is a government.

Operational, tactical and strategic risk management, planning require a well defined glossary.

Below we summarize some of the technical terms applied to Landslides risk assessment.

Failure

Landslides failures are generally characterized by rather high frequencies and extremely variable consequences, hence highly variable risks.

We will note that “failure” should be carefully defined, as it can occur:

• suddenly (one time “fragile” failure of a dam, for example) or
• slowly, under the form of a continuous movement with discrete bursts of acceleration, paroxysms (continuous slide of large “Alpine” landslides).

The latest edition of the risk and crises glossary proposed in this website will be strictly used through this paper to ensure clarity.

Reliability

Technically, Reliability is the chance of not having a failure, so it is the numeric complement to one of the probability of failure.

The probability of failure evaluation can follow many different approaches depending on the level of knowledge available. That include geotechnical data, climate data, design construction and maintenance, etc.

Frequency and probability

These terms require clear definitions:

• frequency -f_q– is the average number of events (within a population) during a given unit of time (failures per year in the Andes, for example).
• Probability -p_f– is the chance that a failure may occur next year (for example 5% for a specific slope).
• When the frequency is small, the number that represents it looks very similar to the number that represents a probability. Thus many people confuse one for the other. However, for f_q>0.1 the values of p start differing significantly.

Figure 1 shows the f_q=f(p_f) or p_f= f( f_q) graph.

Fig. 1 Frequency vs. Probability. Vertical axis probabilities (limited by definition to 1), horizontal axis corresponding value of 1/frequency. Frequency expressed in events per year.

Mathematical rules allow to evaluate f_q=f(p) or p= f( f_q) for any value of f_q or p given a few conditions.

Consequences

Consequences C are generally multidimensional, because lives, infrastructures, environmental, cultural assets, etc. are generally at stake.

Any risk assessment that looks at a partial list of dimensions is biased and censored.

Risk

Risk is the combination (multiplication) of the probability of failure (a number between 0 and 1) and the consequences (multidimensional consequences) of that failure. Thus it has the dimensions of the consequences. Annual probabilities are oftentimes used, however, when looking at a facility with a certain expected life span we can look at at the probability over that life span, which is more understandable by the public simply because the numbers are larger, more familiar.

Risk Tolerance

Risks (from any hazard) can be tolerable or intolerable from a economic, social or ethical point of view. Intolerable risks can be manageable or unmanageable.

Manageable means that the risk can become tolerable within the realm of credibility by allotting sustainable mitigative funds. Unmanageable means the risk remains intolerable, no matter the mitigative funds allotment.

It is obvious that this type of repeatable and rational definition requires other means than the classic factor of safety against sliding and arbitrary classifications normally performed based on intuition or experience.

Once risks (i.e. p_f,C) are evaluated for an entire hazard portfolio, then a graph as the one in Figure 2 can be drawn. The orange curve is the risk tolerance (can be societal or corporate) threshold. Case, locale, culture, society are among the driving parameters to its definition.

In Figure 2 we can actually see the three different classes of risk, namely the blue, yellow and red classes.

The blue are tolerable risks. Therefore, by definition, in a risk management approach, one can put them aside until there are no more yellow and red risks. They are generally operational. If they have high frequencies, low consequences they actually simply impact routine maintenance and operations.

The yellow risks are intolerable but manageable. Mitigation can reduce their probabilities (brought below tolerance in the realm of credibility).

And the last class, the red risks, are unmanageable. Mitigation cannot make them tolerable within the realm of credibility unless we change the system.

 

Fig. 2 a pf-C graph with various risks scenarios and risk tolerance threshold. Three classes of risks are visible. Blue=tolerable, yellow=intolerable but manageable, red=intolerable and unmanageable.

Landslides risk assessment

Risks that are amenable to sustainable and economic mitigation below tolerance by hazard probability reduction are tactical risks. Instead, intolerable and unmanageable risks are strategic risks. They require system’s alterations. That is mitigation to reduce consequences and get the risk under tolerance.

A strategic shifts would mean for example, to permanently evacuate a region. The system would be altered. With this in mind and as an other example, buttressing a dam to reduce its breach probability is a tactical mitigation whereas changing the dam location, breaching it to end its service life, or not building it in a certain area represent strategic mitigation.

In fact tactical risks are under Management responsibility, whereas Strategic risks, might require upper management or governments to shift their objectives, incentives, planning, etc…

Tagged with: Hazards, Landslide Risk Tolerance, Landslides risk assessment, risks

Category: Hazard, Mitigations, Risk analysis, Risk management, Tolerance/Acceptability