From the Editors

This issue considers risks ranging from microbes in food to asteroid strikes from space. It begins with a “Current Topic” essay by Aven and Cox that discusses how the discipline of risk analysis might contribute to making national and global risk studies and reports more useful in characterizing and comparing risks to help societies manage them more effectively. They suggest that current practice might be greatly improved by incorporating methods and principles from risk analysis into such reports, especially in clarifying what is meant by risk and how it is described; in scrutinizing how survey data are collected, interpreted, and used; and in studying how recipients of such reports apply the results, and how they might do so better. The uses of risk-related terms – specifically, the nouns risk, safety, and security and the adjectives risky, safe and secure – in both academic and ordinary language are discussed in a paper later in this issue by Boholm et al., who contrast the quantitative meanings often ascribed to such terms in technical discussions with their non-quantitative and comparative ordinary meanings. Understanding and taking into account such distinctions may provide one way to improve the use of risk language in expressing and communicating potentially valuable information to inform decisions. The editors encourage further contributions to making national and global risk reports more meaningful and useful. Three papers in this issue deal with aspects of microbial safety risk analysis for food and drinking water. An initial perspective by the Source Attribution Task Force of the World Health Organization's (WHO's) Foodborne Epidemiology Reference Group reflects on experiences using expert elicitation to try to close data gaps as part of an effort to estimate the global burden of foodborne disease. The authors find that expert elicitation can be carried out, but note some important challenges, including defining what constitutes useful expertise in this setting and creating geographical structures for the elicitation. Petterson discusses a modified quantitative microbial risk assessment (QMRA) framework for choosing among household water treatment systems in developing countries to reduce water-borne disease, taking into account long-term compliance behaviors. Evers et al. use hypothetical worst-case modeling to suggest that transmission of infectious bacteria (drug-resistant E. coli and Campylobacter) from farms to humans via flies leaving the farms could be at least as great as transmission via chicken fillets. They suggest that further investigation of environmental routes of transmission would be worthwhile. Reconstructing occupational exposures from decades ago when conditions and equipment were quite different from today is a challenge in many industries. Boelter et al. demonstrate a constructive approach for past railroad worker exposures. From air samples collected while a retired railroad pipefitter performed various operations on a vintage 1950s locomotive, they conclude that, for these reconstructed occupational exposure scenarios, asbestos exposures were well beneath both historical and current permissible exposure levels. Five articles in this issue address aspects of natural disasters, from quantifying potential risks to humanity from asteroid strikes to better communicating flash flood risks and warnings. How large is the threat from asteroid strikes and what, if anything, can and should be done about it? Reinhardt et al. present historical data on the frequencies and severities of asteroid impacts and develop a probabilistic simulation model for asteroid impact risks to humanity over the next century. They illustrate the use of this model in estimating the effectiveness of civil defense strategies such as sheltering or evacuation in reducing loss of lives in the event of an asteroid strike. Yang et al. consider a much more frequent risk: the disruption by tropical cyclones of road networks in Hainan province, China. They use probabilistic risk assessment and recovery models to estimate that the Hunan highway network may suffer greater than 90% functional loss almost one year in four, based on historical data and estimated return periods, and exceedance probabilities. Spatial characteristics of the road network and the cyclone path (but not its direction) interact in determining both risks and the effects of risk mitigation measures such as planting trees and improving road infrastructure to prevent physical damage, which can potentially increase recovery times from hours to weeks. Hurlbert and Gupta consider how governments and policymakers can respond adaptively to uncertain risks from climate change, droughts and floods, using as source material policy responses to these conditions in relation to agricultural producers in Chile, Argentina, and Canada. They identify a policy gap in responding to unstructured and “wicked” problems, including presumed effects of climate change, and note that in some cases Latin America lacks policies to deal with the increased flooding projected by the Intergovernmental Panel on Climate Change. They propose that an expanded model of adaptive governance and co-management can help to address uncertainties in science and policy and overcome policy failures. Anderson et al. develop statistical models for the number and locations of fires caused by earthquakes and tsunamis in Japan, finding that it is possible to substantially improve upon the predictive power (as measured by out-of-sample predictive performance) of previous models from the literature. Finally, building on previous work in Risk Analysis on mental models of people's beliefs about flash flood risks and warning decisions, and their resulting behaviors, Lazrus et al. examine what residents of Boulder, Colorado already know about flash flood risks and what else they need to know to revise and inform their actionable beliefs to better manage such risks. This leads to practical recommendations ranging from instilling useful life-saving heuristics (“Climb to safety!”) to making it clearer where and how to find more information. As cities expand, new tunnels are dug for water, gas, telecommunications, and power pipelines, and this new tunneling threatens to disturb old pipelines. Zhang et al. develop and apply a fuzzy Bayesian network (FBN) for representing knowledge and uncertainty about risk mechanisms and causality; combining available measurements and judgments on influential variables describing neighboring pipelines, relevant geology, pipeline depth and age, and technical and management variables; inferring where risks are greatest and the most likely causes of failures; and supporting construction risk management decisions to increase the probability of successfully constructing new urban tunnels without unintentionally disrupting or destroying existing infrastructure. This article provides a clear and detailed exposition that will be useful to other practitioners, explaining both Bayesian Networks (BNs) and the contribution of fuzzy methods in allowing practical incorporation of judgments and uncertainties of causally important but hard-to-quantify risk factors. The authors demonstrate the practical value of this approach via a case study applying FBN to safety analysis of pipelines adjacent to the Wuhan Yangtze River. Wearing life jackets while boating is like wearing seatbelts while driving: most people know that doing so significantly reduces the actuarial probability of death, yet lives are needlessly lost each year because this knowledge is not always translated into practice. Viauroux and Grunger apply Poisson regression, multinomial logit, and other statistical modeling to data from the U.S. Coast Guard's Boating Accident Report Database from 2008 to 2011 to put numbers on the risk reductions from life jacket use in recreational boats and to estimate the contributions of various risk factors to boating fatalities. They estimate a roughly 80% average reduction in fatalities per vessel-year associated with life jacket use, increasing slightly with operator age and with colder water temperatures, with the greatest risk-reduction benefit occurring for shorter boats. They also provide quantitative estimates of the numbers of drowning deaths that would be prevented if life jacket use increased from usual to always, and show how boating fatality risks vary with interacting factors such as age and alcohol, where the highest risk impacts of alcohol use occur at relatively young ages. What can risk analysis and life cycle assessment (LCA) learn from each other? Cucurachi et al. propose that both depend on development of trustworthy models for assessing the impacts of human activities under uncertainties, and they advance this goal by showing how to use global sensitivity analysis to clarify model structure and the importance of different inputs and their interactions, using quantification of the impacts of noise on humans as a case study. They show how various techniques used in LCA for uncertainty evaluation and propagation, including uncertainty and error propagation, perturbation analysis, and key-issues analysis, can be brought within the framework of mainstream sensitivity analysis, and they propose a protocol for developing life cycle impact assessment (LCIA) models in parallel with supporting uncertainty analyses and global sensitivity analyses to identify important input factors and their interactions. Finally, Wang et al. provide an important advance in the risk analyst's technical toolkit by showing how to extend the idea of copulas for modeling dependencies among correlated continuous input variables to handle dependencies among correlated discrete inputs, or among mixed discrete and continuous inputs. The proposed technique greatly reduces the large number of conditional probabilities that would otherwise have to be assessed to fully specify a model with statistical dependences among many input values, and it sheds new light on the importance of not discretizing variables prematurely (before characterizing their interdependencies) in applied work. The issue closes with a review by David Berube of two books on synthetic biology:Life at the Speed of Light: From the Double Helix to the Dawn of Digital Life by J. Craig Venter and Regenesis: How Synthetic Biology Will Reinvent Nature and Ourselves by George Church and Ed Regis. The reviewer finds that both books indulge in engaging hyperbole and touch on a number of fascinating and controversial science-policy questions that might benefit from (and challenge) a range of disciplines, from science to ethics to risk analysis. Much more remains to be done, however, to fathom the risks and opportunities and to decide what constitutes wise restraint in applying the new capabilities – and perhaps chancing the new risks – that synthetic biology offers.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.