Methodologies to Estimate the Economic Impacts of Disruptions to the Goods Movement System (2012)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

43 4.1 Introduction Case studies were conducted to apply the concepts identified in the economic impact analy- sis methodology presented in Figure 3-1. The case studies encompass different freight modes, including rail, highway, ports, maritime, and inland waterways, and cover impacts at the local, regional, and national scales. Table 4-1 presents a number of disruption events, the principal modes of transportation impacted by each event, and the geographic extent (local, regional, national) of the economic impacts resulting from disruptions to the freight transportation network. The shaded cells indicate cases that were chosen for analysis. The cases were selected to provide an opportunity to study in some depth the various supply chain responses to disruptions and to generalize these findings to the extent possible in order to develop rules of thumb that can be used to estimate economic impacts within the high-level methodology. Although a number of the cases had impacts at all levels, all but one of the case studies was intended to focus on the economic impacts at a specific geographic scale. Thus, for example, although the impacts of 9/11 were felt both regionally and nationally, New York City experienced severe disruptions within its local freight transportation network, providing an opportunity to look at disruptions to the enormous volume of intra-regional truck deliveries that are the lifeline of the New York City economy. In contrast, the Los Angeles/Long Beach Port disruption was included in order to demonstrate the differences in the scope, magnitude, and incidence of economic impacts when viewed from a local, regional, and national perspective. 4.2 The Northridge Earthquake, Southern California, 1994 Step 1: Define the Direct Freight Transportation Network Impacts Nature of the Disruption On January 17, 1994, a magnitude 6.8 earthquake centered on the Northridge community in the San Fernando Valley caused widespread damage to buildings, gas supply lines, and four freeways in the Los Angeles area of Southern California. Transportation Facilities and Services Impacted The highway network to the north and west of the Los Angeles central business district suf- fered a number of bridge collapses and other severe structural damage. In particular, the highway interchange between I-5 (the Golden Gate Freeway) and State Route (SR)-14 collapsed. I-5 is the principal north-south truck route through the region. East-west traffic was also impacted. The I-10 (Santa Monica Freeway) bridges at Fairfax Avenue and Washington Boulevard collapsed; C h a p t e r 4 Case Studies

44 Methodologies to estimate the economic Impacts of Disruptions to the Goods Movement System and major damage also occurred to the Gothic Avenue and Bull Creek bridges in the western San Fernando Valley along SR-118 (see Figure 4-1). Modes Impacted The event disrupted highway travel, including the movement of significant volumes of truck freight traffic. Timeframe of the Disruption The highways were repaired within a few months of the disaster. On May 17 and 18, 4 months after the earthquake, I-5 through lanes were reopened. On July 8, less than 6 months after the earthquake occurred, the I-5/SR-14 interchange lanes were reopened, and the last connector ramps within the junction were reopened in November (Giuliano and Golob 1998; Wesemann et al. 1996). Step 2: Identify Current and Future Affected Network Flows by Facility and Link The I-5, I-10, SR-118, and SR-14 freeways all lost traffic due to closure. Parallel routes and local connectors in a five-county area were impacted and operated under congested conditions for a period lasting from 4 to 6 months. Traffic Flows During the Network Reconstruction Period The junction of I-5 and SR-14 is a complex set of mainline and connector lanes designed to accommodate large total traffic volumes and heavy truck traffic in very steep terrain. Also due to the terrain, there are few arterials in this area. The California Department of Transporta- tion (Caltrans) estimated the pre-earthquake corridor freeway traffic volume at around 260,000 Disruption Mode(s) National Regional Local Iceland Volcano Air C argo X 2011 Japan Earthquake Shipping X I - 4 0 Bridge Collapse Barge, Trucking X Hurricane Katrina Rail, Trucking X X I - 5 Floods (WA) Trucking X X Baltimore Rail Tunnel Fire Rail X LA/Long Beach Port Closure Shipping, Trucking, Rail X X X “ 9/ 11” Trucking X X Minneapolis Bridge Collapse Trucking X X Northridge Earthquake Trucking X X Economic Consequences to b e Examined X = Case Study Table 4-1. Major disruption events and their modal and geographic contexts.

Case Studies 45 vehicles per day. The traffic on the affected section of I-5 dropped 59 percent immediately due to lack of alternative routes (Zhu and Levinson 2008). According to Wesemann et al. (1996), traffic count data showed significant loss of traffic on I-5 leading up to the damaged section of highway, with the SR-14 truck bypass lanes, which offered a primary detour option, accepting 75 percent to 80 percent of the pre-quake trips. Arterial streets in the vicinity of the damaged I-5 and SR-14 interchange also experienced significantly higher traffic volumes during reconstruction, carrying as many as 30,000 vehicles (autos and trucks) per day more than they did before the earthquake occurred. A drop in traffic volumes on the order of 6 percent occurred. Long-haul regional truck traffic used to I-5 was sig- nificantly impacted by the network closures. Both Caltrans traffic counts and responses to a May 1994 truck intercept survey indicated that truck traffic dropped some 30 percent, with rerouted traffic having to travel significant additional distances in order to use alternative regional high- way corridors such as I-15 and US-101 (Wesemann et al. 1996). During highway reconstruction, the traffic volumes on the affected section of the eight-lane I-10 freeway dropped from some 310,000 to around 130,000 vehicles (average annual daily traffic) Source: Journal of Transportation and Statistics, Vol 1.2, May 1998. Figure 4-1. Northridge earthquake—highway damage sites.

46 Methodologies to estimate the economic Impacts of Disruptions to the Goods Movement System based on counts taken about 10 weeks into the recovery, with significant increases in parallel arterial street volumes in the vicinity of the damaged roadway. An Automated Traffic Sur- veillance and Control (ATSAC) system operated by the City of Los Angeles made it possible to improve throughput on these arterials through real-time signal adjustments. The detour included two designated alternate routes, with a shorter detour reserved for HOVs, while other traffic was diverted off the freeway for a longer distance and onto more distant arterials. On- street parking was removed from the detour arterials, providing additional travel lanes. Medi- ans were restriped to provide additional turning lanes at key intersections, and signal timing was adjusted to favor through traffic. The Freeway Service Patrol (a roving emergency response service) was expanded to cover the detour arterials. Extensive signage guided travelers along the detour routes. With minor modifications, the detour remained in operation until the freeway reopened on April 12 (Giuliano and Golob 1998). A drop in mixed traffic volumes on the order of 13 percent is estimated to have occurred during the reconstruction period (Wesemann et al., 1996). A limited number of travelers shifted to bus travel during the reconstruction phase, while ridership on Metrolink rail transit lines rose quickly at first then gradually declined back to pre-event levels as more throughput capacity was returned to the freeway corridors (Giuliano and Golob 1998). Traffic Flows after Re-Opening of the Freeways After restoring 70 percent of pre-earthquake capacity along I-5 by implementing a series of miti- gation projects, traffic volumes increased to 88 percent of pre-earthquake levels. After full capacity was restored in May 1994, total traffic increased quickly and went beyond the 1993 level in June by 1 percent. Once the detour had been completed, by the end of January, daily truck volumes on I-5 and SR-14 soon returned to near normal levels. Similarly rapid returns to pre-disruption traffic volumes occurred in the I-10 and SR-118 corridors. Step 3: Define Supply Chain Characteristics and Parameters by Flow Type The network disruptions caused by the earthquake were limited to the highway network and hence affected the costs of trucking materials into, out of, within, and through the Southern California region. Where delays in delivery times occurred, these could add further logistics costs at the pick-up and receiving ends of a shipment, whether from/to local customers or from/ to the region’s large seaports. Based on a post-event survey of businesses in the region, Gordon et al. (1998) reported that 11.2 percent of responding firms suffered from commuting interrup- tions, 4.2 percent from inhibited customer access, 7.4 percent from shipping disruption, and 4.6 percent from interrupted supplies. The limited duration of the event, with partial throughput capacity restored in a matter of 3 months and with alternative routes quickly upgraded and made available to truckers, suggests that most economic costs directly associated with the transportation network were trucking industry-related. Step 4: Supply Chain Response Modeling Based on survey data, most responses to the disruption by trucking firms were short term in nature, including (see Willson 1998, Table 6): • Rerouting (81 percent), • Rescheduling (69 percent), • Increases in driver overtime (55 percent), • Reductions in the frequency of deliveries/pick-ups (38 percent), and • Consolidation of loads (29 percent).

Case Studies 47 There was little use of alternative transportation modes (rail or air modes). The principal response was rerouting, especially among courier firms. Responding firms said that most (96 per- cent) of these actions were no longer being used by mid-May of 1994 (Willson 1998). Some 14.3 percent of firms surveyed said they changed their shipping practices, while 6.6 percent indicated that they had altered their supply arrangements since the network disruption (Gordon et al. 1998). Step 5: Economic Impact Modeling Timeframe The post-recovery reports available for this event suggest that the economic impacts were of a short-run nature, on the order of 6 months to a year, with the majority of costs associated with the diversion of trucks to much longer routes and to temporary increases in operational costs associated with logistics activities such as the need to consolidate less-than-truck-load (LTL) shipments and increases in truck driver overtime. No major modal shifts occurred, and no last- ing changes in industry supply chains, such as changes in the location of product sourcing or changes in product markets, appear to have taken place. Spatial Considerations The major economic impacts appear to have occurred within the state and mainly within the southern California region, mostly in the form of localized shifts in trucking activity. Most of the trucking firms impacted made intra-regional deliveries within Southern California. How- ever, business losses were spread much more widely. According to Gordon et al. (1998), of over $6.5 billion in lost business, 51 percent occurred within the immediate impact zone, most of it (47.7 percent) in direct economic costs. Some 15.8 percent of losses were sustained outside the Southern California region, including abroad. This implies some, if comparatively small, effects also on oceanic shipping and domestic rail shipments into and out of Southern California via the ports of Los Angeles and Long Beach. Direct Supply Chain Costs Wesemann et al. (1996) provide estimates of daily area-wide truck travel delay costs during reconstruction, broken down by delays on the I-5, I-10, and SR-14 freeways, on parallel arterial streets, and due to regional reroutings. These totaled 7,832 hours of daily delay, associated with some 63,000 daily truck trips. Evaluated at $19.20 per hour this is equal to $155,000 in truck- ing delay-related costs per day, as well as some $10,000 per day in excess truck fuel used due to additional en route congestion (as derived from data in Tables 2 and 4 of Wesemann et al. 1996). A post-event survey of firms in the region reported by Boarnet (1998) also indicated that transportation network damage played an important role in the business losses that occurred following the earthquake, with 43 percent of the responding firms indicating that some portion of their business losses was due to transportation damage, attributing some 39 percent of their losses directly to transportation service disruptions. When asked a series of questions about the biggest problems they faced as a result of the earthquake, the top four responses were all related to transportation system disruptions, in the form of customer and employee access to the busi- ness location and shipping delays to and from the worksite. Inter-Industry and General Equilibrium Impacts Gordon et al. (1996, 1998) used post-event survey data and an I-O model to estimate busi- ness losses in terms of lost person-years of industrial-sector-specific employment, which they translated into an estimated $6.5 billion loss of economic output, of which just over $1.5 billion is attributed to transportation service disruptions. Of this $1.5 billion dollars,

48 Methodologies to estimate the economic Impacts of Disruptions to the Goods Movement System almost $731 million was attributed to commuting time losses, with the rest assigned to busi- ness logistics issues. In round figures, the authors attributed costs associated with restrictions on customer access to $272 million ($229 million within the 5-county Southern California region), costs associated with shipping problems to over $484 million ($408 million within the region), and costs due to supply disruptions at $299 million ($252 million within the region). Case Study References Boarnet, M. G. (1996). “Business Losses, Transportation Damage, and the Northridge Earthquake.” Journal of Transportation and Statistics, 1.2:49–64. Gordon, P., et al. (1996). The Business Interruption Effects of the Northridge Earthquake, paper presented at the Western Regional Science Association, Napa, CA. Gordon, P., et al. (1998). “Transport-Related Impacts of the Northridge Earthquake.” Journal of Transportation and Statistics, 1.2:21–36. Giuliano, G. and Golob, J. (1998). “Impacts of the Northridge Earthquake on Transit and Highway Use.” Journal of Transportation and Statistics 1.2:1–20. Wesemann, L., et al. (1996). “Cost-of-Delay Studies for Freeway Closures Caused by Northridge Earthquake.” Transportation Research Record 1559: 67–75. Willson, R. (1998). “Impacts and Responses: Goods Movement after the Northridge Earthquake.” Journal of Transportation and Statistics, 1.2:37–48. Zhu, S. and Levinson, D. (2008). A Review of Research on Planned and Unplanned Disruptions to Transportation Networks, Department of Civil Engineering, University of Minnesota, Minneapolis. 4.3 Winter-Storm-Related Closures of I-5 and I-90 in Washington State, 2007–2008 Step 1: Define the Direct Freight Transport Network Impacts Nature of the Disruption In December 2007, heavy rainfall west of I-5 in the Willapa Hills, combined with melting snow from the mountains, created extremely high floodwaters in western Washington State. Twenty-four-hour rainfall intensities were 140 percent higher than the 100-year amounts for areas in Southwest Washington. High winds, heavy rains, mudslides, and falling trees made travel unsafe on highways across much of the western part of the state. Downed power lines blocked roads and, in many urban areas, rainwater overwhelmed drainage systems and water pooled on roadways. Dozens of residents were trapped by mudslides, some needing to be lifted to safety by U.S. Coast Guard helicopters. Transportation Facilities and Services Impacted A 20-mile section of I-5 near Cherhalis in western Washington was flooded and closed for 4 days, with 12 feet of flowing water on I-5 in some places. Flooding and a massive landslide at Mile Post 27 and various other smaller slides and flood debris also caused SR 6 between Centralia/ Chehalis and Raymond/Hoquiam/Aberdeen to be closed both ways to all traffic from Decem- ber 3 through December 29, 2007. In addition to the disruptions to highway traffic, the Curtis Industrial Park rail line and eight culverts, which connected the Port of Chehalis to the Curtis Industrial Park near Pe Ell, also experienced flood damages. The Chehalis-Centralia airport levee failed and water covered most of the airport during the 4 days of flooding along I-5. A section of I-90 through Snoqualmie Pass was closed for 4 days in February 2008 due to the threat of weather-induced heavy snowfall and avalanches. Modes Impacted Highway (truck) traffic along Interstates 5 and 90 were affected.

Case Studies 49 Timeframe of Disruption I-5 was closed for 4 consecutive days from December 3 through the morning of December 7 of 2007, with some limited capacity to handle heavier trucks late on December 6th. Beginning in late January, less than 2 months after the Chehalis area flooding closed I-5, the state experi- enced another major highway system disruption. Record snowfall and warm temperatures in the mountain passes created snowfall and avalanche threats that closed I-90 at Snoqualmie Pass for 89 hours from January 29 through February 2, 2008. Step 2: Identify Current and Future Affected Network Flows by Facility and Link Traffic Flows during the Network Reconstruction Period Prior to the flood damage, more than 54,000 vehicles traveled the affected stretch of I-5 daily, including almost 10,000 trucks. Washington State DOT (WSDOT) responded to the closure by providing e-mail updates to freight companies on the agency’s listserv and by frequently updat- ing WSDOT’s Website with the most current information on detour routes and road conditions. These updates provided companies and truckers with the best information available, enabling them to plan and strategize as much as possible. WSDOT also established several detour routes through southern Washington and Oregon (see Figure 4-2). Working with the Washington State Patrol and local communities, the state DOT also allowed trucks carrying emergency supplies and extremely perishable goods for local communities to use the nearby State Route 7, a much shorter alternate route (an extra 85 miles), on a case-by-case basis. Lack of capacity and safety issues, as well as noise and road damage, caused WSDOT to prevent trucks from using this route otherwise, based on prior experience with flooding in 1996. As shown in Figure 4-2, WSDOT established a primary truck detour along I-84 in Oregon and on I-82 and I-90 in Washington. This route added 440 miles and some 8.5 hours of driving to an otherwise approximately 200-mile trip from Portland to Seattle. (By the third day of the closure, trucking companies were at risk Source: WSDOT, 2008. Figure 4-2. I-5 closure detour map.

50 Methodologies to estimate the economic Impacts of Disruptions to the Goods Movement System of losing drivers due to federally mandated Hours of Service regulations, which require truck drivers to take a 10-hour rest period after 11 hours of driving.) Many trucks instead chose to take U.S. Route 97, which still added some 344 additional miles to the trip (WSDOT 2008, page 5). The closure of I-90 was similarly problematic. On a typical weekday, approximately 6,500 trucks travel over I-90 at Snoqualmie Pass. With most of the state’s east-west routes closed due to very heavy snowfalls, trucks were left with very few options. WSDOT’s recommended detour routes (see Figure 4-3) called for trucks to detour southward along I-82 to I-84 in Oregon. Severe weather also caused a closure of I-84 between Pendleton and La Grande for a time, cutting off all east-west detour routes. Highway 12 over White Pass also had to be closed for avalanche control and to clear accidents. Traffic Flows after Re-Opening of the Interstates No detailed, post-flooding assessments of traffic flows were undertaken. This is probably because truck traffic movements returned to normal soon after the interstate routes were reopened, as might be expected along routes whose pavements were not damaged sufficiently to prevent traffic from moving over them once the high water or snowfall had been removed from the road surface. Step 3: Define Supply Chain Characteristics and Parameters by Flow Type The network disruptions caused by the flooding and heavy snowfall were limited to the high- way network and hence affected the costs of trucking materials into, out of, within, and through Washington State. Where delays in delivery times occurred, these could add further logistics costs at the pick-up and receiving ends of a shipment. Step 4: Supply Chain Response Modeling According to the WSDOT economic impact analysis of the closures: “As a practical matter, there are no substitute routes to effectively transit I-5. The interstate is the West Coast’s major Source: WSDOT, 2008. Figure 4-3. I-90 pass closure detour map.

Case Studies 51 north-south corridor for both freight and auto traffic” and “I-90 is the longest Interstate highway in the United States, stretching from Seattle to Boston. It is the main highway route for east-west commerce in the state linking Puget Sound to Spokane in eastern Washington. The route con- nects eastern Washington agriculture businesses and other industries with urban markets in northwest Washington and Puget Sound, along with global markets via the Ports of Seattle and Tacoma.” (Ivanov, Hammond, and Reinmuth 2008). Step 5: Economic Impact Modeling Timeframe The post-recovery reports available for this event suggest that the economic impacts were limited to a few months. Spatial Considerations The major economic impacts appear to have occurred to businesses along the closed routes, but also with impacts on both north-south (I-5) and east-west (I-90) semitrailer and triple- trailer truck traffic carrying goods to/from Washington State seaports for both within and out- side the state. Direct Supply Chain Costs A WSDOT study of the total economic impact of these two weather-induced network disrup- tions to the state’s freight systems estimated combined costs at almost $75 million, of which some $47 million was associated with the I-5 disruption (Ivanov, Hammond, and Reinmuth 2008). The analysis was based on a survey that obtained responses from 2,758 affected trucking firms and freight-dependent businesses statewide and included queries about the following: • Revenue losses (losses incurred by firms that could not deliver products for their customers in time, including losses associated with perishable goods and losses incurred because firms did not receive the orders their customers had placed); • Additional business costs incurred by both the trucking industry and freight-dependent sectors (costs associated with delays, detours, and use of alternative modes of delivery, and other actions, which caused additional costs. Such costs encompass increased fuel charges, increased wages and overtime pay for drivers, additional communication costs, higher costs of using alternative methods for delivery of goods, and other operational costs). Trucking companies reported that the I-5 detours took a substantial toll on their businesses, requiring double the resources— including drivers, power units, and trailers—to make the longer trips—The additional cost of taking the detours was estimated to be between $500 and $850 per truckload; and • Future disruption prevention costs (i.e., costs that firms planned to expend in order to prevent additional future losses and to ensure retention of customers). Inter-Industry and General Equilibrium Impacts Estimation of the above listed direct impacts on the economic output of the trucking industry and freight-dependent sectors was subsequently expanded to include any indirect and induced economic impacts (as well as impacts on output, employment, personal income, and state tax receipts) using the IMPLAN-based Washington State input-output economic model. The state was estimated to have lost over $3.8 million in tax revenues, and 460 jobs (WSDOT 2008; Ivanov, Hammond, and Reinmuth 2008). Figure 4-4 shows the results of this analysis broken down by the two network disruption events. The $37.1 million in direct business losses attributed to the two highway closures combined were themselves constructed from three sources: losses in business sales (58 percent), additional freight costs such as fuel and overtime pay for drivers (39 percent), and future loss prevention costs reported as being taken by the freight carriers interviewed (about 3 percent). The analysis

52 Methodologies to Estimate the Economic Impacts of Disruptions to the Goods Movement System also broke down its economic impact estimates for the trucking industry and freight-dependent sectors within each of the state’s seven major geographical regions, noting similar impact across all regions other than the coastal region, where the I-5 closure impacts were noticeably more severe. The study also provides some interesting qualitative business firm specific case study reports. The state’s trucking industry, which depends on the state’s primary freight corridors for daily business, suffered the highest percentage of lost revenues. Trucking companies reported losing 0.51 percent of their total annual revenue because of the two highway closures. The state’s freight-dependent sectors reported an average loss of 0.05 percent of their total annual sales revenue, while freight-dependent companies that operated their own truck fleets lost 0.32 percent of their total revenues. All other economic sectors lost approximately 0.01 percent of the year’s sales rev- enues, a very small percentage, but multiplied by a large annual revenue number. In addition to the above estimated business losses, estimated highway damage from the winter storm was $18 million for state routes and another $39 million for city and county roads. Case Study References Ivanov, B., Hammond, P., and Reinmuth, S. (2008). Freight Transportation Economic Impact Assessment of the I-5 and I-90 Closures in 2007–2008. Presentation to the Puget Sound Regional Council Freight Mobility Roundtable, July 11, 2008. WSDOT (2008). Storm Related Closures of I-5 and I-90. Freight Transportation Economic Impact Assessment Report. Winter 2007–2008. Final Research Report WARD 708.1. Washington State Department of Trans- portation, Olympia, WA. 4.4 Local Impacts of the 9/11 World Trade Center Attack Step 1: Define the Direct Freight Transport Network Impacts Nature of the Disruption The events of September 11, 2001, are succinctly summarized in a report published by the Organisation for Economic Co-operation and Development (OECD 2002) as follows: On the morning of Tuesday, 11 September 2001, the United States was hit by a set of unprecedented ter- rorist attacks, calculated to inflict massive civilian casualties and damage. Four hijacked commercial jets crashed, two into the World Trade Center towers in Manhattan, which collapsed shortly thereafter, one on the Pentagon in Washington D.C., and the last one in Pennsylvania. Source: WSU/WSDOT Economic Impact Survey of I-5 and I-90 Winter Storm Closures, 2008. Figure 4-4. Impacts on economic output due to the I-5 and I-90 winter storm closures (in $ millions).

Case Studies 53 A September 27, 2002, Report for Congress on the subject starts with, “The tragedy of Septem- ber 11, 2001, was so sudden and devastating that it may be difficult at this point in time to write dispassionately and objectively about its effects on the U.S. economy” (Makinen 2002). More time has passed and, while the emotions associated with the event remain, many assessments have been made regarding the impacts. The focus of this case study is the impact of the attacks on the supply chain and, more specifically, on the movement of goods in the New York-New Jersey region. Step 2: Identify Current and Future Affected Network Flows by Facility and Link Transportation Facilities and Services Impacted The attacks immediately affected all aspects of the U.S. supply chain—roadways were closed, rail service slowed, and ports and airports were closed. Border traffic was halted, as follows: Immediately after the attack, the U.S. Customs Service moved to Alert Level One, which called for intensive anti-terrorism operations at all borders and points of entry. The U.S.-Mexico and U.S.-Canada borders were closed entirely . . . Railroads faced speed and service restrictions in the northeast United States immediately after the attacks . . . Out of all modes of transportation, airlines were the most adversely affected. All commercial air service was shut down, airports were emptied, and the Federal Aviation Administration (FAA) prohibited passenger planes from carrying cargo (Lee and Hancock 2005). Ports throughout the United States were immediately shut down. Step 3: Define Supply Chain Characteristics and Parameters by Flow Type Short-Term Impact on Supply Chain • On September 11, the U.S. Coast Guard shut down the port. However, recovery moved swiftly—by the afternoon of September 12, a plan to reopen the port was developed. On September 13, vessels began leaving the Port of New York and New Jersey. The port was fully opened and functioning on September 14. • All airports and U.S. air space were closed immediately after the attacks. An article noted, “A week after the attacks, most airlines were offering full services again, and the ban by FAA was lifted. The only contingency on carrying international cargo was that it had to be off-loaded at the first U.S. city at which the plane arrived, where it then had to be moved to its final destination by surface transportation.” (Lee and Hancock, p. 9) New York officials noted that airport opera- tions resumed in waves and that John F. Kennedy International Airport (JFK) was also affected by road closures in the New York City area. These conditions resulted in shippers and forward- ers trying alternative airports and having to think differently about their air cargo movements. • The bridges and tunnels between New York and New Jersey were similarly affected by the attack. The Holland Tunnel, located closest to Ground Zero, was most affected and was closed for the longest period of time. Utility work was required to return basic functionality. The tun- nel was also needed to remove debris from Ground Zero and for response efforts. New York City, at the time, had limited all traffic to the lower Manhattan area. Step 4: Supply Chain Response Modeling Even with the significant damage and closures, New York businesses and residents contin- ued to be served. Supplies and equipment to ramp-up alternative office facilities were quickly brought in. Response vehicles, construction equipment, and debris removal occurred.

54 Methodologies to estimate the economic Impacts of Disruptions to the Goods Movement System Assessments of the impact of 9/11 on the supply chain conclude that the greatest disruptions to the supply chain appeared to have resulted from the border closures. “The severe tightening of border controls following the September attacks resulted in long waiting times that disrupted the operations of manufacturing companies, especially at the U.S.-Canada border” (OECD 2002). Industries that had adopted just-in-time inventory strategies appeared to have been most affected. Lee and Hancock summarize the impact on the automotive industry as follows: Auto production was halted at more than 60 plants in the United States and Canada on 9/11 as manufac- turers were unable to get critical inventory . . . Ford closed all Canadian and U.S. manufacturing facili- ties, as did DaimlerChrysler Corp., Toyota Motor Manufacturing, and Mitsubishi Motor Manufacturing. Ford cut production 13 percent the week following the attacks because of blocked parts delivery and still ended up losing more than 16,000 units of production by the end of that same week. General Motors initially lost about 100 total hours of production at eight plants—six in the United States and two in Canada—and a week after the attacks reported 10,000 lost units of production (Bryce 2001). BMW reported 750 units of production were lost when it ceased production at its Spartanburg, South Carolina, plant, and Toyota and Honda had to shut down plants for a few days after the attacks. All automakers with North American facilities were affected, resulting in an estimated 52,636 units of production lost in the first week (Ward’s Auto World, 2001). The impacts on the New York-New Jersey roadways were related primarily to the rerouting of traffic. Many food services and JFK-bound trucks had used the Holland Tunnel, which was closed to truck traffic from September 11, 2001, until January 2011. Reduced access to JFK was a challenge for many trying to use the airport. Permissible truck heights at the Holland and Lincoln Tunnels were restricted prior to 9/11. It was noted during the interviews that prior to 9/11, the Port Authority was already considering restrictions at its tunnels for trucks larger than straight trucks for safety reasons—the tunnel had 10-foot lanes and sharp curves—and to be able to accommodate industry standard buses at the facilities. The overall impact on supply chain thinking is summarized in the following quote: The events of September 11, 2001, brought significant attention to several questions and motivated the need to think beyond lean and be able to deal with major supply chain disruptions. Before 9/11, supply chain security meant preventing things from being removed from the logistics cycle by unauthorized parties. In the post-9/11 world, concern has also focused on preventing disruptive elements from being inserted into supply chain operations to create violent havoc. Today we ask: How do we reduce the impact of new security initiatives on supply chain and freight transportation network productivity and how can we design competi- tively productive supply chains that exhibit “resiliency” (i.e., degrade gracefully and recover quickly) when a major disruption occurs? (White 2010) Resiliency considerations are now more often considered in both public- and private-sector planning, both for freight and passenger transportation. Step 5: Economic Impact Modeling The longer term impacts appear to be most related to discretionary cargo markets, implemen- tation of post-9/11 security initiatives, and resiliency in the supply chain. JFK, according to Port Authority officials, lost air cargo traffic because some shippers and forwarders stayed with the alternative routings explored immediately after 9/11. JFK had historically been used for interna- tional air cargo movements for a large portion of the North American market. By being forced to try alternative gateways and finding these alternatives better for serving certain markets, ship- pers and forwarders changed their long-held practices. JFK, the New York-New Jersey region’s predominant international airport, lost business and is working hard to regain market share. The Port of New York and New Jersey was not affected in the long term. All ports were equally affected; vessels had no place to go. The quick reopening of ports led to a resumption of maritime

Case Studies 55 traffic along pre-9/11 lines. It was noted by port officials that the 2002 West Coast Port Strike had a much greater impact on ports—in the aftermath of the strike, major shippers adopted a “port diversification” policy. Instead of relying on a single port, shippers increased the number of ports used so as not to be affected in the future by such region-specific disruptions. However, the significant changes and enhancements in seaport and airport security have required the shippers, terminal operators, agencies, and transportation providers to make equally significant investments in security equipment and personnel. Enhancements include perimeter fencing, new worker background checks and identification cards, radiation detectors, and scan- ning of cargo. Much has been written regarding the implications of enhanced security on the supply chain. However, it has also been noted that the enhanced security efforts also helped address long-standing theft issues. Case Study References Makinen, G. (2002). Report for Congress—The Economic Effects of 9/11: A Retrospective Assessment, Order Code RL 31617, Sept. 27, unnumbered page. Lee and Hancock (2005). “Disruption in Supply Chain Due to September 11, 2001,” Decision Line, March, pp. 8–9. Organisation for Economic Co-operation and Development (OECD) (2002). IV. Economic Consequences of Terrorism, Introduction and Summary, p. 117. “Terror Attacks Stall Industry; Long-Term Impact Unknown” (October 2001). Ward’s Auto World, 37(10), 25. White, C. (2010). “How the Post 9/11 World Has Made Supply Chain Design More Complex—Going Beyond Lean,” Public Policy & Sustainability, March 8, http://www.freightpublicpolicy.org/2010/03/how-the-post-911-world- has-made-supply-chain-design-more-complex-percentE2 percent80 percent93-going-beyond-lean/ 4.5 Los Angeles/Long Beach (LA/LB) Ports Shutdown, California, 2002 Step 1: Define the Direct Freight Transportation Network Impacts Nature of the Disruption In 2002, the labor contract between the International Longshoremen and Warehouse Union (ILWU) and the Pacific Maritime Authority (PMA) was expiring and came up for negotiation. The negotiations stalled, resulting in a work slowdown. The PMA responded by imposing lock- outs of dock workers in September and October (Farris 2008). Given that this potential disrup- tion was being played out in a public way, shippers and carriers could have anticipated the actual event, thus making this case study different than the others. Also, the importance of the Ports of LA/LB in the global supply chain was such that any shutdown of these ports would likely rever- berate throughout the world. Transportation Facilities and Services Impacted The Ports of LA/LB together represent one of the largest container port terminals in the world. The twin ports account for 32 percent of U.S. imports and 17 percent of exports (Park et al. 2002). Due to their size and importance, the impact of the shutdown spread nationally. Thus, not only were the port facilities and the surrounding support services affected, but the regional highway system, national railroads, and the corresponding highway/rail facilities serving those ports now receiving freight diverted from the LA basin were also affected. Modes Impacted The event disrupted cargo container transportation by water, rail, and highway. Timeframe of the Disruption The work slowdown and the lockouts spanned a period of several weeks starting mid-September 2002. On October 9, 2002, President Bush ordered the workers back to work by invoking the

56 Methodologies to Estimate the Economic Impacts of Disruptions to the Goods Movement System Taft-Hartley Act. However, ILWU members continued their systematic work slowdown and PMA reported a significant productivity decline. The slowdown ended in late November 2002, when PMA and ILWU reached an agreement (Farris 2008). Step 2: Identify Current and Future Affected Network Flows by Facility and Link Traffic Flows During the Network Disruption Period The Ports of LA/LB together rank as the sixth largest container port in the world. Together with Oakland, Seattle, Tacoma, and Portland, the six largest West Coast container ports handled 253 million tons of cargo and were responsible for more than half of all foreign containers passing through U.S. ports in 2001. The total worth was just over $300 billion, or 42 percent of waterborne trade in the United States (Farris 2008). Figures 4-5 and 4-6 show the before, during (Point 3), and after exports and imports for LA, San Francisco, and others, including San Diego, Columbia-Snake, and Seattle (Park et al. 2002). When President Bush stepped in on October 9, over 200 ships had lined up outside of the Ports of Los Angeles and Long Beach. It was estimated that it would take 6 to 7 weeks to clear this backlog. The backlog had grown to over 100 days by the time the contract between PMA and ILWU had been settled (Farris 2008). Traffic Flows after Re-Opening the Ports Figures 4-5 and 4-6 show that vessel imports and exports were restored to the pre-shutdown days quickly after the labor dispute was resolved. Cargo volume hit record levels after the new con- tract was signed and, in 2003, coast-wide container throughput was up 10 percent (Farris 2008). Figure 4-5. Foreign exports for vessel mode by customs district (Park et al. 2002).

Case Studies 57 Step 3: Define Supply Chain Characteristics and Parameters by Flow Type The port shutdown not only affected maritime traffic, but also had a spillover effect with respect to other modes. Truckers were unable to return containers to the port terminals or to pick up new cargo, railroads lost shipping business to and from ports, and air cargo saw a slight increase. In addition, due to the size and importance of the ports, not only was the state of Cali- fornia affected, but so were all coastal states in the West, as well as neighboring states and even the nation as a whole. The key variables with respect to the short-term supply chain response was the associated economic cost to the shippers and carriers that waited out the shutdown, or the additional costs associated with diverting traffic to other ports. For example, some carriers with cargo destined for the East Coast diverted ships through the Panama Canal to East Coast and Gulf ports. The Port of Savannah, Georgia, for example, experienced substantial growth in container movements during the shutdown and immediately after. Step 4: Supply Chain Response Modeling Although many vessels awaited clearance for the port, the major response to the incident in traffic terms was rerouting of vessels to other ports and the rescheduling of vessels. Park et al. (2002) indicated that there was a positive impact on port trade along the West Coast before the shutdown (due to the anticipation of some disruption) and that during the shutdown only LA and San Francisco experienced a negative impact, while other ports experienced growth. Ship- pers were trying to find other ports of entry in a way that satisfied supply chain constraints while minimizing costs. Park et al. (2002) also found that there were modest positive impacts on air cargo. Figure 4-6. Foreign imports for vessel mode by customs district (Park et al. 2002).

58 Methodologies to estimate the economic Impacts of Disruptions to the Goods Movement System Burlington Northern and Santa Fe Railways lost their container business serving the ports (nor- mally 4,000–7,000 units per day) during the shutdown, but claimed that much of that traffic was either shipped before or after the shutdown, resulting in minimal loss overall (Zuckerman 2002). In the longer term, there is some evidence to suggest that shippers and carriers viewed the port shutdown as an opportunity to re-examine their dependence on the ports for satisfying supply chain needs. For example, some carriers seriously examined the possibility of moving containers via the Suez Canal to eastern ports and, in fact, such services have been instituted (although it is not clear that they were done in direct response to the port shutdown). With the widening and deepening of the Panama Canal, it is likely that many shippers and carriers will be examining the use of the all-water route to the East Coast of the United States, primarily due to cost considerations, but in some cases, it is likely that dependence on a major port facility for satisfying supply chain movements (and the vulnerability associated with it) also is under consideration. Step 5: Economic Impact Modeling Timeframe The post-recovery reports of this event suggest that the economic impacts were limited to a few months, although they were significant to some in the supply chain. The majority of costs were associated with cargo delays and the temporary increases in operational costs associated with logistics activities. No major modal shifts have occurred, and no lasting changes in industry supply chains, such as changes in the location of product sourcing or changes in product mar- kets, appear to have taken place in any significant way, although the experience of the shutdown does seem to have influenced carrier and shipper thinking about the vulnerability of the supply chain with respect to the port. Spatial Considerations The major economic impacts appear to have occurred within the coastal states in the West (both positive and negative), as well as selected East Coast ports and in the nation as a whole. The ports export (and to a lesser degree import) products that come from all over the United States. Direct Supply Chain Costs Several studies have analyzed the economic impact of the port shutdown with varying results. Initial estimates indicated an economic cost to the U.S. economy of $1 billion per day. Mar- tin Associates conducted a study for the Pacific Maritime Association and estimated the total cost to be $21.4 billion, or $1.94 billion per day for the 11 days. They later adjusted that num- ber to $15.6 billion ($1.42 billion per day) after having interviewed over 200 operators, carriers, importers, exporters, etc. According to Martin Associates, $14.4 billion was lost by exporters and importers (Farris 2008). This $1–2 billion per day range was quickly adopted by the media, the maritime industry, and politicians. Other studies and several major newspapers, however, imply that this figure was an exaggeration. Anderson (2002) estimated that the total cost to the U.S. economy caused by the ports shutdown was only $1.67 billion, or $140 million per day. His estimates were based on his conclusion that while some food items may have spoiled, most cargo was simply delayed, not canceled. Hall (2004) criticized the method used by Martin Associates. He stated that while this method is useful for analyzing economic effects of cargo handling at ports and docks, there are many short-comings with respect to a broader economic impact study. Short-comings include a lim- ited short-term substitution behavior analysis (rerouting to other ports, changing production strategies, etc.) and the exclusion of both winners and losers of a disruption. Haveman and Shatz

Case Studies 59 (2006) point out that the lockout saw a significant (short-term) decrease in imports and exports at the Ports of LA/LB, but that the total imports at the national level only decreased by a few percentage points. Park et al. (2002) included rerouting, rescheduling, alternate modes, and “win- ners” in their analysis, and they concluded that the import-related economic impact was positive ($579 million) over a total of 5 months during and after the shutdown, while the economic losses for exports were $3 billion during those 5 months. Case Study References Anderson, P. L. (2002). Lost Earnings Due to the West Coast Port Shutdown—Preliminary Estimate (2002 working paper), Lansing, MI: Anderson Economic Group. Farris II, M. T. (2008). “Are You Prepared for a Devastating Port Strike in 2008?” Transportation Journal, Vol. 47, No. 1: 43–53. Hall P. V. (2004). “We’d Have to Sink the Ships: Impact Studies and the 2002 West Coast Port Lockout,” Economic Development Quarterly, Vol. 18, No. 4: 354–367. Haveman, J. D. and Shatz, H. J. (2006). Protecting the Nation’s Seaports: Balancing Security and Cost, Public Policy Institute of California, www.ppic.org/content/pubs/report/R_606JHR.pdf (viewed online on 10/18/11). Park, J., et al. (2002). “The State-by-State Economic Impacts of the 2002 Shutdown of the Los Angeles-Long Beach Ports,” Growth and Change, Vol. 39, No. 4: 548–572. Zuckerman, S. (2002). “Shutdown Not So Bad After All,” San Francisco Chronicle, Friday, October 18, 2002, http:// articles.sfgate.com/2002-10-18/bIOiness/17566615_1_west-coast-port-shutdown-economic-impact-cost- estimate (viewed online on 10/18/11). 4.6 The I-40 Arkansas River Bridge Collapse near Webbers Falls, Oklahoma, May 2002 Step 1: Define the Direct Freight Transportation Network Impacts Nature of the Disruption On May 26, 2002, the captain of the tugboat Robert Y. Love experienced a blackout that caused him to lose control of the vessel. As a result, the barge he was controlling collided with a bridge support, causing a 580-foot section of the I-40 bridge over the Arkansas River to collapse into the water (see Figure 4-7). Several automobiles and tractor trailers fell from the bridge, resulting in 14 deaths. Source: FHWA (2002). Figure 4-7. I-40 bridge collapse in 2002. See http.// xpda.com/i40bridge/for additional aerial pictures.

60 Methodologies to estimate the economic Impacts of Disruptions to the Goods Movement System Transportation Facilities and Services Impacted I-40 traffic had to be rerouted through local roads, and Arkansas River barge traffic on the McClellan-Kerr Arkansas River Navigation System into and out of Tulsa’s Port of Catoosa was stopped for about 2 weeks (Pant et al. 2011). The highway rebuilding cost was about $30 million, in part because of repairs to the alternate routes. Modes Impacted The event disrupted both highway and barge traffic through the region, including the move- ment of significant volumes of truck freight. Timeframe of the Disruption The I-40 highway bridge was re-opened to traffic on July 29, 2002, 65 days after the disaster occurred. Step 2: Identify Current and Future Affected Network Flows by Facility and Link Traffic Flows during the Network Reconstruction Period The I-40 bridge is on a major east-west transportation corridor connecting Memphis, Ten- nessee; Oklahoma City, Oklahoma; and Albuquerque, New Mexico. Prior to its collapse, the 1,988-foot-long, four-lane bridge carried an estimated 22,000 vehicles per day. During the 2-month closure, an estimated 20,000 vehicles per day had to be rerouted over other paths on the highway system (see Figure 4-8). This included an estimated 6,000 to 7,000 trucks. Eastbound delays were expected to be about 45 to 50 minutes while westbound traffic was expected to be delayed 15 to 20 minutes. Traffic Flows after Reopening of the Bridge Very little analysis of post-recovery traffic volumes has been conducted. Truck traffic appears to have returned to its pre-collapse routes soon after the bridge was reopened. Some barge traffic was able to move through the main navigable channel during the bridge reconstruction and appears to have returned to normal volumes soon after the incident. As Figures 4-9 and 4-10 show, the effects on port throughput, at least in total tonnage terms, appears to have been short lived. Step 3: Define Supply Chain Characteristics and Parameters by Flow Type Highway (truck) traffic was disrupted significantly by the bridge collapse. Simulation stud- ies of the potential highway networkwide impacts of the bridge collapse were carried out by Oak Ridge National Laboratory and Battelle Memorial Institute (see FHWA 2002 and Fig- ure 4-8) and also, more recently, by the Oklahoma Universities Transportation Center (Ingalls et al. 2009). Although the largest impact on truck movement volumes was on alternative routes in close proximity to the collapse site, both studies found that a significant number of trucks using I-40 originated, or were destined for, locations outside Oklahoma, with subsequent impacts on route-specific truck traffic volumes outside the immediate vicinity of the acci- dent (see Table 4-2). Figures 4-11 and 4-12 show the types and volumes of commodity flows impacted.

Case Studies 61 % Source: Estimates by Oak Ridge National Laboratory (Regional flows) and Battelle memorial Institute (National flows) for the FHWA Office of Operations Technical Services (FHWA, 2002) Figure 4-8. Projected percentage change in national and regional truck flows due to I-40 bridge collapse. Source: Tulsa Port of Catoosa, http://www.tulsaport.com/about_our_waterway.html 0 500,000 1,000,000 1,500,000 2,000,000 2,500,000 3,000,000 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 Figure 4-9. Annual tonnages through Tulsa Port of Catoosa, 1985–2010.

62 Methodologies to estimate the economic Impacts of Disruptions to the Goods Movement System 0 100,000 200,000 300,000 400,000 500,000 600,000 700,000 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 JUL JUN MAY Source: Tulsa Port of Catoosa, http://www.tulsaport.com/about_our_waterway.html Figure 4-10. May, June, and July tonnages through Tulsa Port of Catoosa, 1985–2010. By weight the three biggest commodity movements on I-40 through Webbers Falls are California to Georgia (4.5 million tons), Georgia to Oklahoma (4.1 million tons), and within Oklahoma (3.2 million tons). Source: Estimates by Oak Ridge National Laboratory for the FHWA Office of Operations Technical Services (FHWA 2002). Figure 4-11. Estimated interstate commodity flows on I-40 in 2002. Tons of Shipments Value of Shipments Assignable Truck Within Oklahoma 8% 1% 11% From or to Oklahoma 28% 14% 34% Through Oklahoma 64% 85% 55% Table 4-2. Estimated percentages of non-local commodity tonnages: using the I-40 bridge crossing at Webbers Falls, Oklahoma.

Case Studies 63 Waterway (inland barge) traffic was also disrupted for a short period of time. This included freight moving into and out of the nearby Tulsa land port of Catoosa. The following lists the main industries using the port in 2002 (Pant et al. 2011): • Food and beverage and tobacco products • Petroleum and coal products • Chemical products • Nonmetallic mineral products • Primary metals • Fabricated metal products • Machinery • Miscellaneous manufacturing Step 4: Supply Chain Response Modeling The major response to the incident in traffic terms was rerouting of truck volumes. Barge traffic was delayed temporarily, but appears for the most part to have awaited clearance of the navigable channel, which was open before final highway bridge reconstruction was completed. Step 5: Economic Impact Modeling Timeframe The post-recovery reports available for this event suggest that the economic impacts were limited to a few months. Spatial Considerations The major economic impacts appear to have occurred within the coastal states in the West (both positive and negative), as well as selected East Coast ports and in the nation as a whole. The ports export (and to a lesser degree import) products that come from all over the United States. Direct Supply Chain Costs The Oklahoma Department of Transportation (ODOT) estimated the daily cost to traffic using the detours to be about $212,000. Much of this traffic was directed onto winding, two-lane roads through small towns. ODOT estimated that trucks would incur an additional operating 0 5,000 10,000 15,000 20,000 25,000 1, 00 0’ s To ns Source: Estimates by Oak Ridge National Laboratory for the FHWA Office of Operations Technical Services (FHWA 2002). FARM PRODUCTS FOOD OR KINDRED PRODUCTS COAL NONMETALLIC MINERALS FOREST PRODUCTS TEXTILE MILL PRODUCTS METALLIC ORES FRESH FISH OR MARINE PRODUCTS LUMBER OR WOOD PRODUCTS CHEMICALS OR ALLIED PRODUCTS Figure 4-12. Top 10 commodities by weight crossing the I-40 bridge.

64 Methodologies to estimate the economic Impacts of Disruptions to the Goods Movement System cost of about $17.50 per delay hour, with eastbound delays on the order of 45 to 50 minutes and westbound delays on the order of 15 to 20 minutes. At the Port of Catoosa, some 50 firms that relied heavily on the waterway for supplies reported losses of some $300,000 a day in revenue when the waterway was closed. Inter-Industry and General Equilibrium Impacts A recent study by Pant et al. (2011) used a multi-regional inoperability input-output model (MRIIM) to simulate the cascading regional economic impacts of a 2-week closure of the Port of Catoosa, similar to the closure resulting from the I-40 bridge collapse. Catoosa is the largest land port in the United States, with some 2.2 million tons of freight received and moved by the port annually. Approximately 70 companies use the port, which covers a land area of approximately 2,500 acres and provides employment in one form or another for some 4,000 people (Tulsa Port of Catoosa 2011). The authors combined data on annual, monthly, and average daily arrival tonnages from a number of sources, including the 2007 U.S. Commodity Flow Survey, the U.S. Army Corps of Engineers, and the port itself to estimate daily commodity-specific import and export volumes. They used this data in a discrete time event queuing model to simulate the sup- ply chain characteristics associated with moving cargo through the land port over the course of a full year, under normal conditions versus a year containing the 2-week cargo supply conditions. This involved running 1,000 different simulations both with and without the network disrup- tion event, and each involving the estimated time delays and costs associated with cargo arrivals, temporary storage, crane assisted to/from vessel transfers, and departures. The simulation modeling, using an input-output model for 10 states and 62 industrial sectors, concluded that “the closure of the Port of Catoosa for 2 weeks causes losses to the industries that do commerce through it and is considerably cascaded to other interdependent industries that are not using the port” (Pant et al. 2011, p. 732). Direct industry demand losses were estimated at $48.6 million across all states impacted, with additional indirect output losses of $101.9 million. Case Study References FHWA (2002). Commodity Flows Affected by the I-40 Bridge Collapse at Webbers Falls, Oklahoma: A Preliminary Assessment (Draft). Federal Highway Administration, Washington, D.C., May 29. Ingalls, R. C., et al. (2009). Freight Movement Model Development for Oklahoma. Report prepared for the Oklahoma Department of Transportation by the Oklahoma Transportation Center. Pant, R., et al. (2011). Interdependent Impacts of Inoperability at Multi-Modal Transportation Container Terminals. Transportation Research Part E-47: 722–737. Tulsa Port of Catoosa (2011). http://www.tulsaport.com 4.7 CSX Howard Street Tunnel Fire in Baltimore, Maryland Step 1: Define the Direct Freight Transportation Network Impacts Nature of the Disruption On July 18, 2001, at about 3 p.m., an eastbound CSX freight train derailed while traveling through the Howard Street Tunnel in downtown Baltimore, Maryland (NTSB 2008, p. 1). The train, consisting of 60 cars and 3 locomotives, carried a mix of freight and hazardous materials (DHS 2001, p. 9). According to the NTSB report, 11 cars derailed, 4 of which contained hazard- ous materials—a tank car containing tripropylene, 2 tank cars with hydrochloric acid, and 1 tank car containing a plasticizer considered an environmentally hazardous substance. The tank car containing the tripropylene was punctured and the escaping tripropylene ignited. The fire, according to the NTSB, spread to several adjacent cars.

Case Studies 65 The result was a fire within the tunnel that lasted 5 days (Figure 4-13). The fire also resulted in a 40-inch water main break above the tunnel, disruptions to transit operations in Baltimore, the movement of freight to the Port of Baltimore, and Internet and telecommunications services due to damage to a cable running in the tunnel. The incident began during the evening rush hour in downtown Baltimore before a baseball game was to be played at Camden Yards. Public sirens were sounded and the Coast Guard closed the Inner Harbor. The intensity and location of the fire made it difficult to fight. On July 23, the incident commander declared the scene under control and authorized access to the tunnel without breathing apparatus for qualified personnel. According to CSX, the train derailment caused no structural damage to the tunnel. According to the FRA Report to Congress (2005, pp. 2–18), “the first freight train passed through the tunnel at 8:48 a.m. on July 24, 2001.” No lives were lost. The NTSB report noted five minor injuries to firefighters (Figure 4-14). Source: Baltimore Sun as shown in ITS Program Office report, Effects of Catastrophic Events on Transportation System Management and Operations: Baltimore, MD – Howard Street Tunnel Fire – July 18, 2001, p. 8. Figure 4-13. Howard Street Tunnel fire. Source: U.S. Department of Homeland Security, U.S. Fire Administration/ Technical Report Series, CSX Tunnel Fire: Baltimore Maryland, USFA-TR-140/ July 2001, p. 26. Figure 4-14. Inspecting damage in the Howard Street Tunnel.

66 Methodologies to estimate the economic Impacts of Disruptions to the Goods Movement System Step 2: Identify Current and Future Affected Network Flows by Facility and Link The Howard Street Tunnel is owned by CSX and used solely for the railroad’s freight trains. CSX describes the route as a “merchandise corridor”—predominately carload traffic, with some coal, some intermodal, and some auto traffic. It was a merchandise train that derailed and caught fire. The tunnel, according to the U.S. Fire Administration (USFA) report, was built over a 5-year period between 1890 and 1895. The one-track tunnel is nearly 2 miles long. The USFA analysis notes that while the tunnel was little used for several decades, it is now the longest active under- ground train route on the East Coast, handling 40 CSX rail freight trains daily—Currently, the tunnel, rarely seen by residents, facilitates the passage of tons of freight, everything from orange juice to automobiles, fine goods, and coal (DHS 2001, p. 4). Step 3: Define Supply Chain Characteristics and Parameters by Flow Type The tunnel is not used for passenger service nor does it serve other rail lines. The route is one of only two national Class I rail lines that runs from the Northeastern U.S. to the Southeastern U.S. The second line is operated by the Norfolk Southern (NS). The fire effectively severed one of these two major routes. In addition, access to the Port of Baltimore was hampered due to road closures. Step 4: Supply Chain Response Modeling According to CSX, the railroad rerouted traffic using alternative rail freight routings, rather than using passenger lines. The railroad made it a priority to notify businesses regarding changes and conditions affecting their shipments. The disruption caused by the tunnel fire affected both local movements in Baltimore and rail freight movements along the East Coast. Rail traffic was delayed or rerouted as far west as Ohio. The FRA report summarized the rail freight impacts as follows: To avoid the Howard Street Tunnel, CSX had to send freight trains west to Cleveland, Ohio, north to Albany, New York, and then south to Baltimore, incurring a 3- to 4-day delay. Some CSX trains were rerouted via the busy NS line through Manassas, Virginia, Hagerstown, Maryland, and Harrisburg, Penn- sylvania. At one point during the fire, eight CSX trains that would have used the tunnel were detouring through Cumberland, Massachusetts, and Youngstown, Ohio; five through Hagerstown and Harrisburg; five through Cleveland and Albany; and 12 trains were stopping in various yards (FRA 2005, p. 2–18). A 2002 report issued on the fire by the U.S. Department of Transportation ITS Joint Program Office noted the following impacts on the rail freight system (U.S.DOT 2002, p. 16–18): • CSX diverted or delayed a significant portion of the rail traffic along the Eastern Seaboard. • The railroad issued an advisory that freight moving to and from Chicago to Baltimore and Philadelphia had been rerouted through Selkirk and South Kearny, New Jersey, with expected delays of 18 to 24 hours. • Freight moving along the Eastern Seaboard from the Northeast to Florida and other southern states was advised to expect delays of 24 to 36 hours. • The Tropicana “Orange Blossom Special” was among the six trains rerouted by CSX over the NS track. • Three trains were canceled. • Freight destined for the Port of Baltimore was held either across the harbor from its destination or detoured via Philadelphia. The report notes that when the tunnel was reopened on July 24, much of the initial traffic consisted of crosstown traffic.

Case Studies 67 While the predominant impact was on the rail freight system on a multi-state level, the fire caused disruptions to businesses and surface transportation activities in the Baltimore area. The ITS Program Office report noted that, “The closing of Howard Street and the surrounding area in essence cut Baltimore’s central business district in half, closing off east-west traffic flows” (see Figure 4-15). The FRA report noted, “The fire and burst water main damaged power cables and left 1,200 Baltimore buildings without electricity. Severed fiber-optic lines backed up traffic regionally and nationally because the fiber-optic cable through the tunnel is a major line for the extremely busy Northeast corridor” (FRA 2005, pp. 2–18). Step 5: Economic Impact Modeling CSX indicated that there was no long-term impact on businesses because of the nature of the commodities being moved. The railroad noted that merchandise rail freight shipments are less time sensitive; the choice of rail is primarily based on cost rather than service profiles. It should be noted that the Tropicana train rerouted over NS track is an example of a more time-sensitive shipment. The train carries temperature-controlled Tropicana consumer products. Although the disruption may not have directly affected movements over the long term, the incident did lead to considerable discussions involving the movement and routing of hazard- ous materials, along with tank car designs and incident management. The USFA report noted, “The train derailment in Baltimore focused attention once again on the issue of transporting hazardous material, including radioactive and nuclear waste, through densely populated areas” (DHS 2001, p.10). Testifying before Congress in October of 2001, the Mayor of Baltimore noted, “One of the first things we realized—based on our experience in the CSX tunnel fire—was that rail yards and tracks, filled with chemical tankers and munitions cars, represent one of our most vulnerable targets” (DHS 2001, p. 22). Source: U.S.DOT (2002). Figure 4-15. Road closures in Baltimore due to Howard Street tunnel fire.

68 Methodologies to estimate the economic Impacts of Disruptions to the Goods Movement System CSX also worked to manage reputational damage. The railroad sent claims personnel out immediately. The USFA report noted the following: With the cause of the derailment still under investigation, CSX ran a full-page advertisement in the Bal- timore Sun, entitled “Thanks, Baltimore!” The advertisement, addressed to the citizens of Baltimore, thanked the mayor, fire chief, “the courageous professionals of the Baltimore City Fire Department,” and the emergency response personnel for their “tireless efforts, leadership and professionalism” following the derailment. It also thanked the community for its patience and support (DHS 2001, p.12). Perhaps the biggest impact of the Howard Street Tunnel fire was on the local community. The FRA report noted Beyond the adverse effects on railroad traffic, there was a massive effect on life and activities in downtown Baltimore. The incident forced the closing of streets and business over much of downtown for several days. Officials canceled three Baltimore Orioles game, resulting in a $5 million loss to the team. They also closed Howard Street, along with 14 other cross streets, for 5 days. A two-block stretch of Howard Street remained closed for 6 more weeks (DHS 2001, pp. 2–18). The USFA report listed the community impacts as including the following: • The changes in traffic routes (with roads closed or rerouted) caused rush-hour gridlock and affected the city’s light rail system. Bus service was substituted. • The train accident caused the worst congestion in cyberspace in the 3 years that a company monitoring Internet traffic had witnessed. • The Baltimore Department of Public Works estimated that approximately 60 million gallons of the city’s water supply was used to assist the fire suppression activities. Case Study References Federal Railroad Administration (2005). Report to Congress: Baltimore’s Railroad Network: Challenges and Alter- natives, U.S. Department of Transportation, Nov. National Transportation Safety Board (2008). Railroad Accident Brief, NTSB/RAB-04/08. U.S. Department of Homeland Security (2001). US Fire Administration/Technical Report Series, CSX Tunnel Fire: Baltimore Maryland, USFA-TR-140/July. Washington D.C. U.S. Department of Transportation, ITS Joint Program Office (2002). Effects of Catastrophic Events on Transpor- tation System Management and Operations: Baltimore, MD—Howard Street Tunnel Fire—July 18, 2001, July. Washington D.C.