I. Introduction On September 24, 2015 a tragic crash occurred on the Aurora Bridge on State Route 99 in Seattle that caused 5 deaths. The collision involved a tour vehicle operated by Ride the Ducks and a charter bus. At the time of the writing of this report, the investigation [led by the National Transportation Safety Board) is still in progress. To date, all indications point to a mechanical failure of the Ride the Ducks vehicle. This failure caused the vehicle to cross over the centerline and strike the charter bus. In the wake of this incident, the Seattle Department of Transportation (SDOT) and the Washington Department of Transportation decided to launch a review of the bridge. This review will analyze the operation of the roadway along SR99 from Bridge Way to Halladay St and provide possible alternatives to the existing operation. These alternatives have been analyzed for feasibility and would require further design and analysis prior to implementation. SDOT's Transportation Operations Division performed this review and developed design alternatives in conjunction with WSDOT, who looked at operational and structural feasibility. During the review process there were multiple community meetings held with community groups that represent areas neighboring the bridge. Furthermore, policy experts in the City of Seattle Mayor?s Office, Seattle City Council, Washington State Governor?s Office, and Washington State Legislature have been consulted on this review. II. Background Information A. Bridge Information The George Washington Memorial Bridge, more commonly referred to as the Aurora Bridge, was built in 1932. Because there was no interstate freeway system at the time, this corridor served as the main north-south connection to downtown. The bridge is 2,945 feet long and about 6? feet wide. B. Existing Operational Conditions SR 99 is a primary arterial street. It is part of the state highway network. Operationally, the City of Seattle is responsible for operating and maintaining the roadway, however, the State owns the roadway and bridge and is responsible for any permanent changes. The details of this arrangement are laid out in the (SM-20 agreement between WSDOT and SDOT. SEA000581 1 The bridge currently carries roughly 70,000 vehicles per day. Peak hour volume is approximately 4,000 vehicles in each direction. In the northbound direction, volume is relatively steady with a peak in the PM rush hour. In the southbound direction, there are both AM and PM peaks during rush hour period. 4,500 4,000 3.500 3,000 2,500 NB Volume 2,000 1,500 1,000 500 0?909909?0?0?'v'olume This section of SR 99 has one of the highest vehicles volumes in Seattle. This corridor also sees 492 daily bus trips carrying roughly 28,000 daily bus riders. The RapidRide Line, which is the highest ridership route in the King County Metro system, accounts for a large portion of these bus riders. There are six lanes on the bridge; each lane is approximately 9.5 feet wide. There are also two 5 foot sidewalks. The lane widths are narrower than the typical principal arterial roadway. There is no median barrier on the bridge. SEA0005812 ETFT 9_ 5FTTHEU SB THRU SB THHU NE THHU NB THRU NB THRU LANE LANE LANE LANE LANE LANE Fl .g1@c E?gap EXISTING CROSS SECTION At the north end of the bridge there is an off-ramp and on-ramp at Bridge Way N. The off-ramp is signalized at 38?? St. The on?ramp has a transit queue jump. At the south end of the bridge, there are at grade intersections both Raye St and Halladay St. The speed limit on the roadway is 40 miles per hour (mph). The off-peak median speed varies depending on the lane of travel. They range from 33 in the exterior lanes to 50 in the interior lanes. The 35th percentile speed also has a similar distribution, with speeds of 43 in the exterior lanes and 55 in the interior lanes. The overall 85th percentile speed is 49 mph. SEA0005813 Aurora Bridge Speeds 30,000 0 1: 25,000 20.000 15.000 '3 10.000 5000 soo?oowapo gaseoeowee NWW 0: a (04.4.4. Speed [mph] C. Collision History This review analyzed the 10 year collision history on the Aurora Bridge prior to the incident on September 24, 2015. The analysis period was January 1, 2005 to September 23, 2015. During this 10 year span, 124 collisions occurred. [If those collisions, 31% [102) resulted in an injury. Of those, 3% were serious. A serious injury is any injury that results in at least a temporary impairment, e.g. a broken limb. It does not mean the collision resulted in a permanent disability. There were no fatal collisions during the time period analyzed. [Graphic] Rear-end collisions and side swipefmerge collisions represented the largest percentage of the total collisions at 40% and 39%, respectively. The only other type of collision greater than 1% of the total was head-on collisions at [Graphic] With regard to injury collisions, rear-end collisions represented the largest percentage at 51%, followed by side swipe/merge collisions and head-on collisions In terms of proportionality, however, the percentage of head-on collisions resulting in an injury was 88%, followed by rear-end collisions and side swipefmerge collisions [Graphic] SEA0005814 Looking at the collisions more granularly, the majority of rear end collisions resulting in an injury occurred when the vehicle being struck was stopped. These collisions mostly happened in the exterior lanes on both ends of the bridge as a result of traffic being stopped at Bridge Way and Raye St. [Map] To better understand how the collision history along this corridor rates compare to the rest of the City, we analyzed the 3 year collision history of the bridge in relation to other Transportation Safety Corridors in the City as well as parallel bridges. A three year analysis is standard for our Transportation Safety Corridors. Our analysis showed that the Aurora Bridge had the lowest collision and serious/fatal collision rate of the corridors we measured. The rate is comparable to that of the Fremont Bridge, which is roughly half that of the Ballard Bridge. In comparison to our other Traffic Safety Corridors, the Aurora Bridge has a fraction of the collision rate. Location Average Total of Length Collision Rate Total ll of Injury Bi Fatal Daily Collisions {ml} (per million InjuryfFatal Rate Traffic {2012-2014} veh. mllesl Collisions (per million veh. (vehicles) miles) Ballard Bridge 63,800 39 0.6 0.93 35 0.83 Fremont Bridge 31,000 12 0.2 1.7? 3 044 Lake City Way {120th to 135th} 3?,100 133 0.8 4.09 81 2.49 Rainier Ave 5 (Letitia to Seward Park 29,900 911 4 8.06 493 4.36 Ave 51 Aurora Bridge 20,000 35 0.6 0.?6 19 0.41 D. 2003 Route Development Study In 2003, WSDOT partook in a route development study that analyzed the entire north end of the SR 99 corridor from the Battery St Tunnel to the city limit at 1-451h St. This study recommended the Aurora Bridge be widened to 70 feet, which would widen the lanes to 11.5 feet and provide a median barrier and pedestrian walkways underneath the bridge. This recommendation was estimated to cost $29 million in 2002 dollars. SEA0005815 :l'T 11f1 RAIL HE IFHU 5H THFU PH NH TlmU RAIL WE LANE IANE LANE LA ?Nu saw; 4mm nHv_. ?um any, .mm arm PHOPOS ED FUTURE CROSS In the current review of the bridge there were multiple aspects of this proposal that were more highly examined. First, the Route Development Plan provided a mid-span cross-section that did not account for transitions back to the at-grade sidewalks on either side of the bridge. Providing an ADA accessible pathway would require 4 independent structures at each corner of the bridge. At the south end of the bridge the existing vertical clearance for 6 Ave is 14 feet 6 inches. Providing an 8 foot pedestrian pathway under the bridge deck would not be feasible in this location, therefore the pathway would need to touch down on the north end of 6th Ave N, which would provide challenges in maintaining an accessible descent from bridge deck and sight distance issues for an at-grade crossing. There would be a similar issue on the north end at Troll Ave N, although we believe the descent would not be as challenging to design or construct. Next, the pedestrian pathway under the bridge has been scrutinized for security issues. The lack of visibility from the roadway or either side of the bridge would attract unwanted behavior under the bridge. With the length of the bridge, it would be very challenging to provide adequate timeliness to police response to requests for immediate service. Finally, given the ADA concerns, the cost estimate provided is severely under estimated. We feel that a project of this type would cost over $100 million in 2016 dollars. Proposed Alternatives In the review of the bridge operation, the team looked at different alternatives to try to address some of the issues, mainly speed and collision history, while also considering congestion and transit demand. A. 5 Lane Section This alternative would remove one of the northbound lanes on the bridge and provide wider travel lanes with a median barrier. The median barrier would help prevent future head-on collisions on the bridge while the wider lanes could help reduce the number of side swipesfmerge collisions. SEA0005816 M-l T. 6-2; .aF'Jri- :ux-?hwl This alternative could increase the number of stopped rear-end collision currently experienced at the Bridge Way and Raye St intersections. This alternative could also increase speeds during the off- peak times, clue to the comfortability associated with the wider lanes. We performed a basic capacity analysis to better understand what type of impact the removal of a lane would cause. The analysis determined that the existing 3 lanes in each direction provide the capacity necessary to carry existing traffic volumes; however, the reduction of the 1 lane in either direction would cause congestion. Since the SB direction has peaks in both the AM and PM rush hours, it was determined that the NB direction would be more appropriate to reduce. 5,000 4,500 4,000 3,500 3,000 NB Volume 2,000 1,500 - 1,000 500 ?Existing Capacity ?Proposed 2 Lane Capacity Proposed 3 Lane Capacity 0 Ill I I This alternative has been analyzed for structural feasibility. We determined that the amount of weight on the bridge associated with the lane being removed would compensate for the additional weight of a median barrier. In reviewing this analysis, we compared costs of similar projects and determined this would be a $2 to $4 million project. This project would need additional structural and traffic analysis. We would also want to look more in depth at congestion mitigation for reallocation of volumes and looking further into existing congestion at Bridge Way and Raye St. SEA0005817 B. 5 Lane Section with Reversible Lane This alternative would remove one of the lanes on the bridge and provide a reversible middle lane. This would allow for wider travel lanes with a median barrier. This barrier would be a movable barrier commonly referred to as a "zipper" barrier. It could help prevent future head-on collisions on the bridge, while wider lanes could reduce the number of side swipes/merge collisions associated with the existing narrow lanes. h: 5 ll Iclaw Tim-es? Sula-.2 - This alternative could increase the number of side swipe/merge collisions due to merging zones caused by the lane reduction in the off-peak direction. This alternative could also increase speeds due to the comfortability associated with the wider lanes. This alternative is not expected to cause additional congestion on the bridge because the capacity of the 2 lane section is greater than the volumes during the off-peak period. The only congestion that is anticipated would be associated with the merging zones caused by the lane reduction in the off-peak direction. This alternative has been analyzed for structural feasibility and it was determined that the amount of weight on the bridge associated with the lane that is being removed would compensate for the additional weight of the median barrier. In the review of this analysis, we compared costs of similar projects and determined this would be a $6 to $8 million project. This project would need further and more specific structural analysis and further operational analysis to look more in depth at the costs and scheduling associated with operating the zipper barrier. C. Urban Corridor This alternative proposes reevaluating the corridor similar to the 2003 Route Development Plan. The proposal would look at changing the character of the corridor from the ?fast link? to downtown that it was designed to be in the 1930?s to a more modern urban corridor. This would entail dynamic lane assignment, new traffic signals from Battery Street Tunnel to Greenlake, transit only lanes, and center running transit with median stops. This alternative would reduce speeds on the corridor and reduce the severity of collision along the corridor. This alternative would remove a lane during off-peak periods on the bridge and provide a larger clear space on the bridge which would reduce head-on collisions. SEA0005818 This alternative would increase the number of angle collisions on the corridor due to the introduction of at-grade crossing at traffic signals. This alternative would cause slower travel times to general purpose traffic, but may decrease travel times for transit. Given the high percentage of transit ridership on the corridor, this may increase person throughput. Due to the slower general purpose travel times, this alternative may cause demand shift to other corridors. This alternative has not been complexly analyzed structurally or operationally. This proposal would need vigorous traffic and environmental analysis during the route development process. This analysis would inform the full scope of the project. Since the sc0pe of this project is not as well defined as the other alternatives provided, no specific estimate can be provided, however the magnitude is expected to exceed $100 million for construction. The analysis necessary to develop the scope would cost roughly 52 million. D. 5 Lane Section with Peak Hour Lane This alternative would maintain the lane cross section, however, the northbound interior lane would be treated as a shoulder that is only available for use during the PM peak period. This treatment would allow for a greater shy distance between vehicles moving in opposite directions during times when speeds are higher. Dynamic speed management signage would be installed southbound to reduce on- coming speeds during the PM peak. [insert cross-section] This alternative would reduce head-on collisions and severity of those collisions. This alternative would need to be further analyzed to ensure that capacity reductions met the demand models. Further structural analysis is not necessary because there would be no changes to loading. The analysis necessary to develop a scope and model would cost roughly $200,000. This project would cost roughly 52 million. E. Short Term Supplemental Treatments The alternatives provided above are long-term capital projects that require a significant investment. Given the low rate of serious injury and fatal collision in comparison to other safety corridors in the City SEA0005819 this investment may not be feasible in the near-term. With this in mind, multiple short-term, low cost treatments could make strides towards reducing collisions and speeds on this corridor. First, the existing markings on the bridge are typical roadway paint with dashed lines. This allows drivers to change lanes on the bridge. A treatment to consider is using solid lines with profiled MMA. The solid lines would prohibit drivers from switching lanes on the bridge which would reduce side swipe/merge collisions. In addition, the profiled MMA which is a common treatment on Washington freeways would provide drivers feedback when they veered out of their lanes. [insert photo of profiled While profiled MMA is a common treatment on state highways it is not a standard treatment for SDOT. Currently, SDOT is not capable of maintaining these markings internally. This is not currently an issue because these markings have a longer life than typical paint; however, when they begin to show wear in 7 to 10 years they will need to be touched up by an external contract. Second, since the bridge is 3 lanes in each direction it can be difficult for drivers to see side mounted signs on the edge of the roadway that indicate the speed limit. A treatment to consider is marking speed limit legends on the roadway in each lane to remind drivers what the speed limit is on the bridge. [insert photo of speed limit legends] Speed limit legends are a common roadway treatment however they require a significant amount of maintenance. The useful life is 3 to 5 years depending on where they are installed in the lane and the roadway volumes. With the volumes on the Aurora Bridge the expectation for wear would be about 3 years. Third, standard speed enforcement is a common practice on the Aurora Bridge especially for northbound traffic. During the 2009 SDOT Traffic Safety Corridor project there was heavy speed enforcement all along the SR 99 corridor which showed a reduction of collision attributed to speeding of 20%. SDOT would propose to engage SPD increasing speed enforcement in both directions on and near the Aurora Bridge. Finally, we would propose that automated speed enforcement be considered at this location. Standard speed enforcement can be very difficult on the bridge because there are limited locations where police can perform the enforcement safely and also limited locations where they can pull over violators safely. Furthermore, automated speed detection has shown that when police are performing enforcement on the corridor motorists generally reduce their speeds to meet the speed limit while when police are not enforcing the speeds they general travel well beyond the speed limit. Currently, automated speed enforcement is only legal under Washington State Law in certain locations as a pilot program. Seattle Police Department performed pilots in two of the allowed locations in 2011. This study showed reductions in average speed up to ranging from 6.8% to 15.6%. The study also showed a reduction in violation rates of 64% to 67%. We would recommend that special consideration be given to this location by the State Legislature and City Council to allow for automated speed enforcement given the specific challenges. IV. Conclusion SEA0005821