You are here

Assessment of risk

Disastrous Floods as Indicators of Systemic Risk Neglect (Part #5)

[Parts: First | Prev | Next | Last | All ] [Links: To-K | From-K | From-Kx | Refs ]

One widely-cited excuse by authorities for the damage in Queensland was the exceptional nature of the event -- a "200-year" event, necessarily beyond any reasonable government mandate. Whether or not this figure is statistically accurate or the consequence of faulty modelling, a more correct understanding is that there is then a probability that such an event will occur once in ever 200 years. But, as noted by a citizen in one community at risk, it is then just as likely (statistically) to be repeated in a few years time -- since the "200 years" is but a statistical average over a much more extended period.

A month later, at the time of writing, the impact of a new storm is anticipated with trepidation, as reported by Graham Lloyd and Andrew Fraser (Queensland's cycles of havoc, The Australian, 2 February 2011). Their report notes

According to a paper by BOM Queensland weather forecaster Jeff Callaghan, the frequency of severe land-falling tropical cyclones had declined to low levels in recent decades in line with the El Niño weather patterns. Callaghan's analysis shows that landfalls occurred almost twice as often in La Niña years as they did in El Nino years and that more than one cyclone only ever hit land during La Niña years. Callaghan says it would be imprudent to suppose the low number of tropical cyclones crossing the coast in recent decades would continue and planning should reflect the possibility of a rapid return to higher landfall rates. Callaghan's research confirms [Jonathan] Nott's analysis that tropical Australia is overdue for a dramatic intensification of cyclonic activity, regardless of whether there is a climate change signal in what is happening now or not....

What the longer term records show, however, is that the frequency of extreme cyclones follow a predictable long-scale pattern. "What the record shows is we go through extended periods, hundreds of years, of high activity and extended periods of little activity," Nott says. "The past 100 to 150 years has been very quiet in Queensland in terms of what happened in the past. The couple of hundred years prior to that were very active." According to shorter term decadal scale-that uses a 10-year cycle- Queensland can also expect a big increase in the number of severe cyclones. The Interdecadal Pacific Oscillation indicates the tropical north is due to emerge from a three-decade period of low cyclonic activity and return to the conditions of the 50s, 60s and 70s.

It is questionable whether such potentially embarrassing issues will fall within the remit of the official commission of inquiry into the efficacy of the emergency response services -- which may well terminate its report in the midst of the next "200-year flood".

The appropriate approach to collective learning as a consequence of disasters past, or in anticipation of those to come, is more fruitfully illustrated by:

Especially striking, in the case of the toxic asset "flood" that triggered the financial crisis, was the dependence in risk analysis on the Gaussian copula, as separately discussed (Uncritical Strategic Dependence on Little-known Metrics: the Gaussian Copula, the Kaya Identity, and what else?, 2009). This dependence is admirably described by Felix Salmon (Recipe for Disaster: the formula that killed Wall Street, Wired, 17.03, March 2009) -- or on the title page of the issue as The Secret Formula that Destroyed Wall Street. As David Li had indicated by the discoverer of the innovative formula, David X. Li in 2005: Very few people understand the essence of the model (Mark Whitehouse, Slices of Risk, The Wall Street Journal, 12 September 2005). A second description is offered by Kevin Drum (The Gaussian Copula, Mother Jones, 24 February 2009).

As noted by Salmon:

The damage was foreseeable and, in fact, foreseen. In 1998, before Li had even invented his copula function, Paul Wilmott wrote that "the correlations between financial quantities are notoriously unstable."... During the boom years, everybody could reel off reasons why the Gaussian copula function wasn't perfect. Li's approach made no allowance for unpredictability: It assumed that correlation was a constant rather than something mercurial. ... In hindsight, ignoring those warnings looks foolhardy. But at the time, it was easy. Banks dismissed them, partly because the managers empowered to apply the brakes didn't understand the arguments between various arms of the quant universe. Besides, they were making too much money to stop. In finance, you can never reduce risk outright; you can only try to set up a market in which people who don't want risk sell it to those who do. But in the CDO market, people used the Gaussian copula model to convince themselves they didn't have any risk at all, when in fact they just didn't have any risk 99 percent of the time. The other 1 percent of the time they blew up. Those explosions may have been rare, but they could destroy all previous gains, and then some. [emphasis added]

In the case of flooding rivers, it is precisely this kind of thinking which drives the false confidence in the "normality", to which authorities and vested interests aspire. This is valid "99 percent of the time", but is proven to be false "1 percent of the time" (by "200-year floods") when such predictions fall apart.

An earlier example was provided by the 2004 Indian Ocean earthquake and tsunami about which warnings had previously been provided by the head of the Thai meteorological office -- then forced to retire, accused of scaremongering and jeopardising the tourist industry.

Another example is provided by severe flooding occurred in March 2010 as a result of the Xynthia storm on the Atlantic in the Vendée region. French President Nicolas Sarkozy, declared: We have to find out how families in France in the 21st-century can be surprised in their sleep and drowned in their own houses. Mr. Sarkozy added, We have to shed light as urgently as possible on this unacceptable and incomprehensible drama.

As noted in response by Claire Le Guern Lytle (The Memory of Risks, Coastal Care, April 2010)

As much as this drama is utterly unacceptable, it is all too comprehensible and sadly, previously announced by warnings from many scientists, locals, and even more relevantly by an official 2008 report from the Vendée Equipment Department, DDE. The risks of marine submersion were known to the Vendée DDE, which strongly addressed and questioned coastal safety, citing in particular the fragile sea walls in L'Aiguillon-sur-Mer and La Faute-sur-Mer, as well as their existing location and development in flood-prone areas

Ironically, with respect to the "200-year" figure in Queensland, one interviewed respondent in France (regarding the flooding there in 2010) noted that it was unfortunate that the dikes constructed 200 years ago in the Vendée area by Napoleon had not been maintained since then.

[Parts: First | Prev | Next | Last | All ] [Links: To-K | From-K | From-Kx | Refs ]