Final Program - Society for Risk Analysis

separately, and asked to write down words that they associate with them. The secondgroup was also given the words uncertainty and risk, separately, but asked todraw images associated with them. The last group was given a group of images, andasked to check whether it is associated with risk or uncertainty. The study aims tounderstand how the public visually ties risk and uncertainty to concrete objects. Theresults suggest that individuals tie abstract concepts to concrete objects that are eithergeographically available or experienced.P.112 Erraguntla NK, Sielken RL, Valdez-Flores C, Grant RL; neeraja.erraguntla@tceq.texas.govTexas Commission on Environmental QualityAN UPDATED INHALATION UNIT RISK FACTOR FOR ARSENIC ANDINORGANIC ARSENIC COMPOUNDS BASED ON A META-ANALYSISOF EPIDEMIOLOGY STUDIESThe United States Environmental Protection Agency (USEPA) developed aninhalation unit risk factor (URF) of 4.3 x 10-3 per μg/m3 for arsenic in 1984 for excesslung cancer mortality based on epidemiological studies of workers at two smelters:the Asarco smelter in Tacoma, Washington and the Anaconda smelter in Montana.Since the USEPA assessment, new studies have been published and exposureestimates were updated at the Asarco and Anaconda smelters and additional years offollow-up evaluated. The Texas Commission on Environmental Quality (TCEQ) hasdeveloped an inhalation URF for lung cancer mortality from exposures to arsenicand inorganic arsenic compounds based on a newer epidemiology study of Swedishworkers and the updates of the Asarco and Anaconda epidemiology studies. Usingmeta-analysis on the URFs from these three epidemiology cohort studies, the finalinhalation URF is 1.5×10-4 per μg/m3. The de minimis concentration level (i.e., airconcentration at 1 in 100,000 excess lung cancer mortality) is 0.067 µg/m3. This valuewill be used to evaluate ambient air monitoring data so the general public in Texas isprotected against adverse health effects from chronic exposure to arsenic.W3-H.1 Ertem M, Bier VM; bier@engr.wisc.eduUniversity of Wisconsin-MadisonDEFENDER-ATTACKER MODEL FOR COMPUTER NETWORK SE-CURITYWe propose a general defender-attacker model for security of computer networks,using attack graphs to represent the possible attacker strategies and defenderoptions. The defender’s objective is to maximize the security of the network undera limited budget. In the literature, most network-interdiction models allow the attackeronly one attempt (assuming that the attacker is captured and disabled aftera single failure); other models allow multiple attempts, but assume that any subsequentattempt begins at the point in the network where the previous attempt failed.These models are not appropriate for computer security, where the attacker could be94operating from the safety of a foreign country, and the cost of starting over with acompletely different attack strategy may be quite low. To represent the ability of theattacker to launch multiple attempts, we represent the attacker’s success or failure onany one arc of the attack graph probabilistically, and formulate the resulting securityproblem as a multiple-stage stochastic network-interdiction problem. In the resultinggame, a non-myopic defender anticipates both the attacker’s strategy choices, andthe probability of success or failure, and chooses a single defensive strategy (i.e., aset of arcs in the attack graph to be protected) by which to defend against multipleattempted attacks. The attacker then launches an optimal attack against the system,based on knowledge of which arcs have been protected. If the attacker fails at thefirst attempt, a second-stage optimal attack strategy is chosen, based on a revised attackgraph showing which arcs have already been successfully traversed (now assumedto have success probabilities of 1), and which arc led to failure of the first-stage attack(now assumed to have a success probability of 0). We solve the resulting stochasticoptimizationproblem using two-stage stochastic optimization with recourse.M2-J.3 Evans AM, Rice G, Teuschler LK, Wright JM; evans.amandam@epa.govAssociation of Schools of Public Health, National Center for Environmental Assessment, Officeof Research and Development, US Environmental Protection AgencyA CUMULATIVE EXPOSURE ASSESSMENT OF NOISE AND VOLA-TILE ORGANIC COMPOUNDSAlthough humans are exposed daily to multiple chemical, biological, and physicalstressors, epidemiological studies and risk assessments typically examine them individually.Occupational studies show that concurrent exposure to elevated levels ofnoise and organic solvents (e.g. toluene, ethyl benzene, styrene, xylenes) have a synergisticeffect on hearing loss. In this research, we characterize concurrent exposuresof volatile organic compounds (VOCs) and noise among population subsets in theurban area, San Francisco County, CA. Demographic variables, personal VOC measurementsand exposure factors were extracted from the 1999-2000 National Healthand Nutrition Examination Survey dataset. 2000 census block group population demographics(aggregated by race- gender-educational attainment) were extracted fromthe National Historical Geographic Information System for all census block groupswithin the county. Personal exposure measurements of 10 VOCs (benzene, toluene,ethylbenzene, m,p-xylene, and o-xylene, 1,4-dichlorobenzene, chloroform, trichlorethylene,tetrachloroethylene, and methyl tert-butyl ether) were available for 851 individualsfrom passive samplers worn for 48-72 hours. Census block group level noiseexposures were estimated using ArcGIS from a noise map of San Francisco (Seto etal., 2007) based on traffic-induced noise levels estimated from the Federal HighwayAdministration’s Traffic Noise Model 2.5 model. Exposure differences were detectedby races gender, and education for noise and VOCs. Hazard Indices (based on permissibleexposure levels in occupational settings and inhalation reference concentra-