I have attended a meeting and was listening to talk of some leading scientist in the field of radiological risk assessment who presented cases of Fukishima and the associated cancer risk. The numbers they gave was 10X additional cancer cases. Is this really significant? I had seen a similar case in Kuwait during the 1991 war when the Harward School of Public Health submitted a report that there can be two additional cases of cancer due to depleted uranium. But all these willl be insignificant in statistical terms. Should there be a different approach or actually are there really insignificant figures?
The numbers are not simply insignificant. The numbers have nothing to do with reality. it is irresponsible to claim numbers that are smaller than the uncertainty of the estimate. We do not need a different approach to performing the estimates. We need honest estimators. We cannot expect that to happen any time soon. There is too much to gain by claiming that danger lurks under every rock.
Thanks four your thoughts Joseph. I dont mean that these numbers or any single number for that matter is insignificant. But when these are presented in stastistical terms it will be 1/1000 million and honestly that will look insignificant at any confidence level. So to avoid this should some additional approaches be looked at?
The current practice in risk models is to ignore the uncertainty as a metric for the application of the model. The uncertainty is used to indicate a range of effect with the implication that the model is true for the effect, only the range of the effect is uncertain. There is no acknowledgement that the model is so uncertain that it has lost all ability to predict the effect. The precautionary principle is applied. Once applied the effect is real. No minimum credible effect is reported. There is no acknowledgement that the model is useless for controlling risk.
Yes. A new approach is needed for assessing risk.
From a practitioner standpoint, I would share that we always struggle with risk/hazard assessments. Metrics are but one factor and we must include subjective inputs as well. In our region, we experience seasonal flooding quite regularly (yes, even in the Nevada mountains!) There is a commonly used reference of a "100-year Flood", that is the 1% probability expressed in a time continuum. Needless to say, the general public hears this and wonders what our scientists are drinking, as these "100-Year" floods are occurring every 10 years or so. (Climate change models are showing a marked increase of more extreme cyclical weather patterns).
From my perspective, when assessing the risks of radiological hazards, there remains many questions as to the long-term impacts of contamination. Considering the environmental consequences of Chernobyl and recent studies of wildlife flourishing in the region, it is uncertain what those long-term impacts are on our species. Needless to say, we can't fully test impacts on humans for ethical reasons.
Shifting to Kuwait and the residual effects of depleted Uranium, the metrics may be statistically insignificant, however our instincts suggest we cannot ignore the risk. Herein lies the challenge of the outcome of assessments. One of the purposes is to provide decision-makers with a tool. We cannot ignore information, simply because it is "statistically-insignificant" for the risk of a Black Swan consequence. The consumer of these assessments (decision-makers) must be provided both the metric and the opinion in order to make informed decisions.
I hope this minor contribution is of benefit to your evaluation.
Respectfully...
Risk managers deal with the impact of extreme natural and man-made (accidental or malevolent intentional) hazards through a risk-based approach, where risk is a function of the likelihood of an event occurring and the resulting consequences. Black Swan events are not foreseeable by the usual statistics of correlation, regression, standard deviation or return periods. Expert opinions are also of minimal use since their experiences are tainted by biases and constrained by finite human life span. This inability to estimate the likelihood of occurrence for Black Swan events precludes the application of risk management. The precautionary principle involves possible, if improbable outcomes with no model basis for predicting the possible improbable. The Black Swan events are totally unforeseeable. It would be wonderful if we could protect against the unforeseeable and the improbable and avoid the ensuing litigation. Perhaps the practitioner standpoint demands that every eventuality be covered and no expense be spared. Nevertheless, I suspect no-expense-be-spared and strict avoidance of all improbable and unknown hazards have financial limitations that must be managed with competent risk analysis.
Mr. Alvarez brings up some excellent points, but I will take some exception to the statement that Black Swan events are totally unforeseeable. To use Nassim Taleb's definition of a Black Swan event, particularly in the context of the financial markets meltdown in 2008, there were a number of indicators that pointed to looming fragility.
Circling back on the metrics of risk assessment, the pragmatic use of the tools requires both analytic, empirical , and experiential approaches. Food for thought.
The financial meltdown of 2008 violates Nassim Taleb's definition of a Black Swan event because there were a number of indicators.
Risk assessment must be competent and evidence-based, which means analytic, empirical , and experiential.
Health impact assessment could improve EIA in such subject as long as it includes toxicology
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication