MANY ROADS TO TRAVEL: ALTERNATIVE APPROACHES TO ROUTE SELECTION FOR YUCCA MOUNTATION SHIPMENTS

Transcription

1 MANY ROADS TO TRAVEL: ALTERNATIVE APPROACHES TO ROUTE SELECTION FOR YUCCA MOUNTATION SHIPMENTS Fred Dilger Clark County Nuclear Waste Division Las Vegas, NV Robert J. Halstead State of Nevada Agency for Nuclear Projects Carson City, NV ABSTRACT Tis paper examines te routing implications of te potential sipment of spent nuclear fuel and ig-level radioactive waste to te proposed repository at Yucca Mountain, Nevada. Te autors describe a model wic can be used to evaluate te impacts of transporting spent nuclear fuel nationally and in te State of Nevada. Te autors used tis model to evaluate te cross-country igway routes identified in te DOE s Final Environmental Impact Statement and tree alternative routing scenarios. Te autors evaluate tese routing scenarios based on four attributes: counties, exposed populations, total county populations, and sipment miles. Te autors furter evaluate te impacts of tese alternative routing scenarios on two selected corridor states. Te autors conclude tat routing decisions ave significantly different impacts on different state and local governments. INTRODUCTION Tis paper argues tat route selection decisions for te sipment of radioactive waste to Yucca Mountain would ave differential effects nationally and terefore must be evaluated at a national level. Tere are several reasons wy te program to sip nuclear waste to Yucca Mountain would be uniquely appropriate for a national route selection scenario. First, spent nuclear fuel and ig level waste are potentially more dangerous tan oter kinds of radioactive materials. Altoug te probability of an accident resulting in a release is low, te uman ealt and economic consequences could be severe. Even during routine transportation, tere exists te potential for occupational and public ealt effects due to routine radiation. Second, te Department of Energy s (DOE) program would ave an extremely long duration and intensity- 24 to 38 years wit up to 40 times more waste sipped tan te total sipped in te US between 1964 and 1998 (1). Terefore, training first responders and oter considerations cannot be once-off events, but will ave to be sustained. Knowing were effects are most severe will be important. Tird, unlike oter HAZMAT sipments, carriers of nuclear waste will ave little flexibility, except in emergency situations, once routes are selected. Finally, tere ave been many stakeolder discussions of desirable route attributes (e.g. low population exposure, low accident rate), but te national effect of tese attributes as not been systematically evaluated. Tis paper reports te results of modeling four alternate routing scenarios to Yucca Mt. Te scenarios are compared on four attributes: County population, number of counties, exposed population, and sipment miles. Tese attributes were cosen to illustrate ow tis approac to modeling route selection can generate meaningful results. Te paper concludes by providing an overview of alternate route evaluation attributes and describing ow tese attributes can be applied at a national level. Te results of tis modeling effort support tree conclusions: 1. Modeling HLW transportation routes at a national scale is desirable because of te differential effects between various routing scenarios. Were effects take place and ow many people are affected vary. 2. Routing scenarios, alternative metods for route coice, can be tested to evaluate route attributes. 1

2 3. Te profusion of modern analytical tools and publicly available data enable sopisticated analysis tat will be employed in a policy arena. Tere is a difference between policy oriented analysis and operational analysis. Analytical tools ave to be tailored to fulfill teir particular role in tis area. Tis paper confines its discussion to truck sipments. Tere is currently no rail access to te Yucca Mountain site. DOE as not yet formally specified a preferred mode for repository sipments, nor as DOE demonstrated tat rail access is feasible. However, te same approac to modeling rail sipments would be desirable. A model of te alternate rail routes as been developed. It is necessary to draw a distinction between operational and policy analysis and to describe ow te modeling was accomplised. GIS AS A POLICY TOOL Te ability to prepare tis model igligts te maturation of Geograpic Information Systems (GIS) tecnology from being te solution in searc of a problem of te mid 1990 s to a critical tecnology tat offers great promise for policymakers. Altoug tis paper is not about GIS per se, it reports on four alternative routing policies tat are evaluated using a GIS-based model. Te orientation and development of te model are relevant to discuss because te approac to te problem is different tan previous analytical efforts. It is important to draw a distinction between different types of GIS. For te purposes of tis paper, we identify two types: operational and policy. Te systems tat were developed to support ongoing DOE operations-notably TRAGIS are operational systems-tey suit te day-to-day requirements for routing truck and rail sipments, and performing radiological assessments. Te policy model developed in tis paper was constructed to support ad oc requests made by decision-makers. Altoug tis model can provide te same service as an operational GIS, tis model was built to support policy activities. Tere are important differences between te typical capabilities of operational and policy GIS. Te operational GIS is typically built to exact specifications to solve a specific problem. For example, te TRAGIS model was constructed at Oak Ridge to support DOE operations. Terefore, it ad to fit witin te specific, predetermined requirements set for it by te clients. Tese requirements make TRAGIS an extensive and robust GIS. Te TRAGIS model creates outputs for RADTRAN and oter radiological assessment programs and does tis well. However, tese advantages also tend to make te operational GIS less flexible. Operational GIS typically fit into a specific organizational context and so te place of te GIS is predefined. For example, land ownersip data is used te same way for te same purposes across te nation-it fulfils a clear function in te organization. Te alternative model reported on ere was originally built only to determine te sortest pats from te origin sites to te Yucca Mountain destination. However, wen te notion of te model was first considered, it was not clear wat te ultimate data requirements would be. Unlike te operational model, te policy model must be built for flexibility. Wat is important to decision makers is seldom predetermined. It is also necessary to be prepared to incorporate new data in a way tat will permit exploration of new policy alternatives or even to generate new alternatives. In practical terms, tis means te model must incorporate new data, work at any geograpic scale and ave superior display capabilities. Because policy is time-sensitive, te model must also be able to adapt rapidly. To implement te model it was necessary to assemble te software and data on a suitable ardware platform. Te software cosen was te TransCAD software prepared by Caliper Corporation. Tis software is te most widely used transportation modeling software in te US because. It is a full-featured GIS wit integrated transportation planning modules tat support regional transportation demand modeling, advanced supply-cain management, and geospatial analysis. Te software is most commonly used for regional travel demand analysis. Te algoritms to solve Hickson s problem (minimizing some route attribute) are already built into TransCAD. Te next step was to select te data for te model. Te data cosen came from tree sources: te data contained in te Bureau of Transportation Statistics National Transportation Atlas provided 1:100,000 rail and igway layers, census data and oter common information. Te Department of Energy (DOE) provided data about te sources of te waste and number of sipments. Te DOE s Yucca Mountain Project office provided information about te rail and igway 2

3 routes troug Nevada. Te end result is tat using freely available data it was possible to construct an extremely sopisticated model tat effectively describes te national impacts of transporting tese materials. Te source data was verified by comparison wit te data in te FEIS. Te results of te routing model were validated by comparing more tan alf of tese route results wit route results from te DOE s TRAGIS model. ADVANTAGES OF A MODERN GIS Te primary advantage of te TransCAD software is tat it is fully topological, tat is, te data structures used by te software contain information tat relate spatial location among different data types. Tis permits enanced data analysis and geograpic operations. An example of te way tis topological advantage assists modeling is in te calculation of exposed population. An important policy goal in comparing alternative routing scenarios is to minimize te population exposed during te safe, routine transportation of tis waste. Terefore te most accurate metod for calculating exposed population is optimal. Te callenge is to determine ow many people live witin 800 meters of eiter side of te routes on wic te casks are transported. Tis must also be done across te entire nation. In tis case, te routes were calculated by minimizing a route caracteristic defined by te routing scenario. Next, te census tracts troug wic te routes passed were selected and a buffer of 1600 meters was calculated around eac link of te entire route-tis created a 1600 meters buffer around all of te routes. Finally, te buffer was overlaid on te census tract layer to calculate exposed population. Te steps described above are routine geograpic operations. One way in wic te differences between topological and non-topological mapping systems are manifested is in te way in wic population is calculated. Non-topological operations rely on centroids tat contain data to determine population. Wen te centroids fall outside te buffer, te data is not included. Tis problem becomes acute in te Western US were census tracts are relatively large and te centroids furter from te route. Te problem is illustrated below: Centroid wit excluded data Centroid wit included data Buffer around waste route Radioactive Waste Route Figure 1 Problems wit calculating population from centroids For tis model, a more sopisticated approac was taken. Exposed population was calculated by taking te percentage of te census tract s population included by te buffer. For example, if 1000 people live wit te census tract and te buffer covered 20 percent of te tract, ten te population calculated as exposed would be 200. Tis process is illustrated below. 3

4 1000 resdents living in census tract Buffer around waste route Radioactive Waste Route Buffer covers 20% of census tract terefore 200 people are calculated as exposed Figure 2 Buffer calculations of exposed population Tere are two sortcomings wit tis approac. First, it includes only residential population. A previous analysis of a 16 mile long, potential route segment troug Las Vegas estimated tat total exposed population witin one-alf mile, would be tree times greater tan resident population wen otel/casino workers and guests were included (2). As te model develops, it will be possible to include time of day and to adjust exposed population numbers to fit to employment population data. Second, te distribution of population witin te census tract is typically uneven. Tis problem is acute in te western US were development typically occurs close to te interstate routes. A percentage approac, taken in tis paper is better tan te centroid approac; owever, it is well witin current capabilities to determine were in a census tract te population resides. It will be relatively simple to incorporate aerial potograpy and pattern recognition to determine wat percentage of a tract is developed and were te population resides. Despite tese sortcomings, tis metod creates a better analysis number tan using centroids. One last comment on ardware is appropriate. Te rapid improvements in ardware made all of tis modeling possible on an obsolete laptop computer. METHOD Te examination of te routes was deliberately constructed to support a decision science approac. By identifying and isolating specific route attributes tat can be measured and compared, tis kind of modeling supports a quantitative assessment of routes. Te problem of radioactive materials route selection as already been famously treated by expert decision analysts (3). Te rail routing modeling was used to construct a forecast of DOE s likeliest coice for te rail route to Yucca Mountain. However, tis study demonstrates ow tis kind of modeling can be used to produce te inputs for structured decision-making. However, te autors did not present a formal decision analysis because te necessary value judgmentstat are te key part of any structured decision analysis-would ave to be inferred for a decision maker. Evaluating routes nationally really means comparing alternative sets of routes. Tese sets are defined by some caracteristic. Te route evaluation began by creating four routing scenarios for comparison. A routing scenario is defined by te attribute cosen to be minimized from witin te network of routes. Te algoritm is applied to a network to determine te lowest value of some attribute from te origin to te destination. Te TransCAD software by Caliper Corp. implemented te sortest pat algoritm. TransCAD is also a full-featured Geograpic Information System and so it allows a wide range of data to be incorporated into te model and complete presentation ability. Te software uses two kinds of data types relevant for te analysis, first is a base network. Te second is a route system tat calculates te sipment routes and can ten ave sipment numbers assigned to eac route for subsequent aggregation. 4

5 Tere is a wealt of oter data tat could ave been included in tis analysis tat was omitted for simplicity: Indian reservations, metropolitan statistical areas, urban areas, environmentally sensitive areas, and accident rate data. All of tis oter data is readily available and can be used to improve analysis or to generate new alternatives. One of te critical benefits of tis approac to modeling is tat new alternatives (new routing scenarios) can be identified. DESCRIPTION OF ROUTING SCENARIOS AND ATTRIBUTES Te coice of routing scenario is important because it ultimately sets limits on te analysis. In tis case, eac scenario was cosen based on previously publised work or legally allowable action. Tis analysis examines questions of interest not only to Nevada, but also to dozens of potentially affected states and Indian tribes, and undreds of local government jurisdictions. Coosing between route attributes is important. Tere a number of different sets of attributes already at and, for example, in te NWPAA and Te Guidelines for Selecting Routes for Hig Level Radioactive Waste Routes (4, 5). In eac case, te coice of route attributes appears to make a positive contribution to safety. For tis analysis, te attributes were cosen based on available data related to near route populations, effected jurisdictions, and sipment miles (wic are a surrogate for a variety of oter attributes suc as system costs). Te route attributes used are: County Population: te year 2000 census population living in counties troug wic te routes pass. Number of counties: te number of counties troug wic te routes pass. Tis is a useful measure of organizational complexity facing te routing scenarios. Exposed population: te year 2000 census population living witin 1600 meters on eiter side of te centerline of te route. Sipment : Te cumulative number of miles traveled by te cumulative number of sipments. Tese route attributes are intended to demonstrate tat national modeling of tese routes is possible. Tey are not presented as a conclusive set of routes. However, eac variable was included for a specific reason. County Population is included to indicate te total number of people wo may be affected by transporting tese materials. Te number of counties is presented as proxy attribute to indicate te degree of organizational complexity attributable to eac of te routing scenarios. Counties are a relevant level of organization to consider because tey are so important in emergency response. Te exposed population is included for te obvious reason tat it allows comparison of te number of people wo will involuntarily receive doses of radiation. Sipment miles are a proxy indicator for oter transportation system attributes suc as costs and accidents. It is reasonable to expect tat te more sipment miles, te greater te cost and te more likely an accident. Te remaining question is to sow te national effects of different routing scenarios. Te results of te analysis are displayed quantitatively and grapically using scaled symbol maps of te potential routes to Yucca Mountain tat depict were te sipments will concentrate. SCENARIO 1: SHORTEST PATH INTERSTATE ROUTES TO YUCCA MT. Te first map depicts ow truck sipments of spent fuel would aggregate using te sortest, primarily interstate routes from te point of origin at te reactor sites (and DOE facilities) to te destination at Yucca Mountain. Tese are ypotetical routes only, since tey ignore state designated preferred routes. Tis metod was cosen because it depicts te most straigtforward metod of route selection. 5

6 Yucca Mt Power Plants SortSip Figure 3 Potential interstate igway routes tat minimize distance Te route attributes for tis scenario are: County Population: 125,889,614 Number of Counties 706 Exposed Population: 15,180,685 Sipment miles: 88,210,781 SCENARIO 2: REPRESENTATIVE ROUTES EVALUATED IN THE FINAL EIS Te second map depicts wat DOE s FEIS called representative routes from te point of origin at te reactor sites (and DOE facilities) to te destination at Yucca Mountain. Tis set of routes conforms closely to te HM 164 exemptions tat currently exist across te United States. Tat is, tese routes take into consideration te state-designated preferred routes across te country. 6

7 Yucca Mt Power Plants FEIS Routes Figure 4 Potential Interstate routes identified in te Final EIS Te route attributes for tis scenario are: Population: 124,144,342 Number of Counties 687 Exposed Population: 13,878,181 Sipment miles: 92,212,284 SCENARIO 3: CONSOLIDATED SOUTHERN SHIPPING ROUTES Te tird map depicts one potential alternative approac to cross-country routing identified in a 1996 contractor report for te State of Nevada Agency for Nuclear Projects (6). Tis scenario, developed by Planning Information Corporation, assumed tat cross-country truck sipments migt be consolidated on I- 40 to avoid winter weater disruptions on I-80 troug Nebraska and Wyoming, and to avoid conflict wit te State of Colorado s intention to proibit use of I-70 west of Denver to avoid sipments troug te Eisenower and Glenwood tunnels.. 7

8 Yucca Mt Power Plants CSSip Figure 5 Potential Interstate routes tat minimize use of I-70 and I-80 Te route attributes for tis scenario are: Population: 128,693,569 Number of Counties 707 Exposed Population: 12,252,469 Sipment miles: 97,941,505 SCENARIO 4: HYPOTHETICAL NEVADA ROUTE DESIGNATION TO AVOID CLARK COUNTY Te fourt map depicts one combination of potential alternative routes identified in a contractor report prepared for te Nevada Department of Transportation (7). Under tis scenario, tere would be no sipments troug Clark County. Most sipments from te east would enter Nevada from Uta on I-80. Sipments from California and te soutwest would enter Nevada from California on SR 127. Tis scenario was prepared to test a State of Nevada designation, under HM 164, of routes tat avoided te state s most populous county. It does not indicate any policy position taken by te State of Nevada. In te past, te State of California, and San Bernardino and Inyo counties ave opposed te use of SR 127 for radioactive materials sipments. 8

9 Power Plants Nevada Hypotetical Yucca Mt Figure 6 Potential Interstate routes resulting from ypotetical Nevada route designation Te route attributes for tis scenario are: Population: 123,251,080 Number of Counties 669 Exposed Population: 11,393,646 Sipment miles: 91,764,522 RESULTS OF THE APPLICATION OF THE ROUTE ASSIGNMENTS Te ultimate objective of tis paper is to establis tat evaluating routes on a national basis makes sense and to demonstrate tat it is readily possible. Te results of tis evaluation are: Table I Summary Table of Alternative Route Caracteristics Population witin 1600 meters Sipment Population of Counties Traversed Number of Counties Involved Sortest Pat 15,180,685 88,210, ,889, Final EIS 13,878,181 92,212, ,144, Consolidated 12,252,469 97,941, , Soutern Nevada Hypotetical 11,393,646 91,764, ,251, One of te main advantages of tis analytical approac is tat it quantifies te implications of various routing strategies on specific corridor states and counties. To our knowledge, suc an effort examining te national impacts of realistic alternative routing scenarios as not been previously conducted. DIFFERENTIAL EFFECTS ON SELECTED CORRIDOR STATES A key conclusion of tis paper is tat te alternate routing scenarios ave differential effects wic vary from state to state. Tese effects vary considerably from state to state. To illustrate tese differences, we 9

10 ave evaluated tese differential effects for te state of Uta and Illinois. Te number of counties troug wic HLW will pass in Uta and Illinois are compared for eac of te routing scenarios. Figures 7 and 8 compare only te representative truck routes in te FEIS wit routes tat avoid Clark County. Uta County (Hig Res):1 Uta Higways Higways:1 FEISSip Nevada Hypotetical Figure 7 Alternative Routing Scenarios troug Uta 10

11 Illinois County (Hig Res):2 Higways:1 FEIS Higways:1 Outside FEISSip Uta Population witin 1600 meters Figure 8 Alternative Routing Scenarios troug Illinois Table II Route scenarios compared for two states Illinois Population witin 1600 meters Uta Sipment Illinois Sipment Uta Population of Counties Traversed Illinois Population of Counties Traversed Uta Number of Counties Affected Illinois Number of Counties Affected Sortest Pat 211, ,619 12,524,447 4,679,737 1,495,949 9,315, Final EIS 142,846 2,112,776 12,848,862 3,593,633 1,480,809 10,743, Consolidated 103, , ,612 3,416,512 1,463,566 8,313, Soutern Nevada Hypotetical 119,344 2,112,776 4,719,434 3,593,633 1,152,556 10,743, Te conclusion to be drawn from tis comparison is tat individual states and communities will be significantly affected by routing decisions and will terefore ave an interest in ow te routes are selected and optimized. CONCLUSIONS Te autors conclude tat it is necessary to examine route selection for HLW at a national level because of te differential effects tose routes will ave. Readily available analytical tools make it relatively simple to conduct suc analysis and will be used to engage policy makers. Te relative simplicity of constructing tis model wit accurate, readily available data suggests tat te route selection process will ave to be an open process. Te autors furter conclude tat future impacts of sipping HLW must be fully evaluated to elp policy makers refine emergency management preparations, make appropriate canges to urban plans, as well as oter canges tat can mitigate te impacts of transporting tese materials. Te model described ere can be enanced to provide more useful insigts into te effects of transporting tese materials. 11

12 Te most likely enancement to te model will be to add additional data suc as te establised lists of route attributes included in te USDOT Guidelines. Some examples include: link-specific accident rate data, commodity flow data, Indian reservations, and environmentally sensitive areas. Te purpose of adding tis new data will be to test te attributes and determine te contribution te attribute makes to safety. Oter improvements will be to identify new routing scenarios. Te ypotetical example of Nevada s route selection suggests tat direction. It is possible to evaluate any combination of routes for any of te attributes cited ere as well as additional data for any affected state. REFERENCES 1. U.S. Department of Energy, Final Environmental Impact Statement for a Geologic Repository for te Disposal of Spent Nuclear Fuel and Hig-Level Radioactive Waste at Yucca Mountain, Nye County, Nevada, DOE/EIS-0250, February 2002, (Cited ereinafter as FEIS; available at ttp:// 2. A Mountain of Trouble: A Nation at Risk, Report on Impacts of te Proposed Yucca Mountain Hig- Level Nuclear Waste Program, Prepared by Nevada Agency for Nuclear Projects, Office of te Governor, Volume I, February Tis and oter reports prepared for te Agency for Nuclear Projects, can be accessed on te web at ttp:// 3. R.L. Keeney, Value-Focused Tinking, Cambridge, MA: Harvard University Press, Public Law , Nuclear Waste Policy Act as Amended 1987, Wasington, DC: U.S. Government Printing Office, December U.S. Department of Transportation, Guidelines for Selecting Preferred Higway Routes for Higway Route Controlled Quantity Sipments of Radioactive Materials, Wasington, D.C.: US Department of Transportation, US DOT RSPA Report, DOT/RSPA/MTB-84/ Planning Information Corporation, Te Transportation of Spent Nuclear Fuel and Hig-Level Radioactive Waste: A Systematic Basis for Planning and Management at National, Regional and Community Levels, Report Prepared for State of Nevada, Agency for Nuclear Projects, September, M.V. Ardila-Coulson, Te Statewide Radioactive Materials Transportation Plan, Pase II, Prepared for Nevada Department of Transportation, Prepared by College of Engineering, University of Nevada, Reno, December,