“Quantitative risk assessment”, “hazards analysis”, and the concept of acceptable risk

There was a standing-room-only crowd at the Middletown Township Building for the pipeline safety presentation last night (Monday, July 23). If you were part of that crowd, it would be understandable if you left feeling confused and overwhelmed. Richard Kuprewicz, Middletown’s pipeline safety consultant, covered a variety of topics and made a number of controversial statements at the meeting; and sometimes he used terminology that bewildered his audience. If you asked five different people in the audience for a summary of the evening, you would probably get five strikingly different accounts of what was discussed.

For this post, I would like to focus on one important aspect of the discussion that I think left many observers confused. That is the distinction between a “hazards analysis” and a “quantitative risk assessment”, and the circumstances under which each is useful. Neither type of study has been done for the Dragonpipe (Mariner East 2 pipeline)—or at least, none that is available to the public.

This topic may seem pedantic, but I hope you’ll read to the end: this distinction is critically important to anyone who faces the prospect of living with this pipeline.

Kuprewicz began the session with a set of slides. One of the points he emphasized was his prior recommendation that Middletown Township did not need a “quantitative risk assessment” (QRA). He argued that a QRA was not justified because it would have no impact on the Township’s plans for emergency response to a pipeline accident. Kuprewicz did not mention at that point (but it was subsequently mentioned by Township Council Chair Mark Kirchgasser) that Kuprewicz is conducting a “hazards analysis” (HA) for the Township.

Separately from the Middletown situation, there are already two QRAs moving forward: one for Chester and Delaware counties, initiated by State Senator Dinniman (to be completed by the end of the summer), and a somewhat different one (Delaware County only, and including the Adelphia pipeline project) which the Delaware County council will be voting on tomorrow (July 25).

So: what is the difference between a QRA and an HA? They sound similar, and they are in some respects, but they take different approaches and have different purposes. Here is an over-simplified explanation.  (And to be clear, this is my way of distinguishing them. Kuprewicz would probably put it differently.)

Hazards Analysis (HA). An HA is a description of possible hazards resulting from pipeline operations. It is the basis for developing an emergency response plan for each type of hazard. It does not assign a probability to each hazard, and it does not attempt to quantify the potential damage caused by the hazard.

Quantitative Risk Assessment (QRA). A QRA is more ambitious and has a different purpose than an HA. It looks at each of those same hazards, but then attempts to attach a measure of the consequences (how many injuries? how many deaths?) and a measure of probability (once in 10 years? once in 100 years?) to each hazard. (For a more detailed explanation of QRAs, see “Is there a serious risk of death by pipeline explosion? Here’s how we can tell whether we need to worry.”)

Ultimately, the HA is a tool for emergency responders to use as a basis for response planning, and for zoning boards and planning departments to use as a basis for deciding where development should or should not be permitted.

A QRA, on the other hand, is a tool for the public, to help them understand the level of risk they face. They can then make decisions about where they want to live and where their children should go to school. A QRA is also a useful tool for developing policy around pipeline routing and permitting, although QRAs are not used that way in Pennsylvania.

Acceptable risk. So why did Kuprewicz say a QRA would not be useful? His position seems to be that the usefulness of a QRA depends on having a measure of “acceptable risk”. (Kuprewicz did not help his case by referring to this measure of acceptability as a “kill threshold”.) The concept of acceptable risk is used in some countries (including Canada, but not the US) in evaluating the risks associated with pipelines, compared with other risks we encounter in daily life. An example of such a risk might be the naturally-occurring gas radon, which is can be found in many basements in our area. It increases our risk of lung cancer compared to other regions of the country, but we live with it because we consider the risk to be minor, and therefore “acceptable”.

Detailed examples of the use of the “acceptable risk” approach in Denmark and the EU can be found in this report. The general idea is that a risk can be acceptable if either (1) it would be catastrophic but the probability is vanishingly small or (2) it is fairly likely but would not be too serious. Based on these considerations, a curve can be constructed, with probability on one axis and seriousness (for example, number of fatalities) on the other. Points on one side of the curve represent “acceptable” risks, and points on the other side are “unacceptable”.

What Kuprewicz seems to be saying, then, is that we do not have regulations that embody this kind of precise measure of “acceptability” and therefore even if we could measure the probability and consequences of a pipeline accident (which is what a QRA would do), it would not lead to a conclusion of “acceptable” or “not acceptable”.

That point of view may make some kind of theoretical sense, but it misses the practical point that each of us needs to know the level of risk that we face, and that our communities face. A QRA gives us that information. A QRA might also tell us whether or not a given pipeline would be legal in Europe or Canada; and that, too, would be useful information about the risk we face. As residents, we need these answers, even if Mr Kuprewicz thinks Middletown Township doesn’t need them.