Literature Review: Vessel Collision with Bridges

This literature review was developed primarily to support the technical paper by Roebling Labs.

Appendix A - Literature Review of Vessel Collision Bridge Collapse with Life Loss in the USA

Appendix B - Literature Review and Timeline of Vessel Collision Warning Systems for Bridges in the USA

Appendix C - Literature Review of International Vessel Collision Warning Systems Using AIS

Safety Recommendation H-25-030 - “As part of your short-term bridge risk reduction and mitigation strategies to protect the traveling public, evaluate the need for and, if appropriate, incorporate motorist warning systems capable of activating when a threat is identified and immediately warn and stop motorists from entering onto the bridge.”

Summary of Bridge Owner Responses to Safety Recommendations H-25-3 and H-25-4

Site includes official correspondence between NTSB and 30 Bridge Owner Agencies.

Roebling Labs' vessel collision warning system responds to NTSB Safety Recommendation H-25-030. Specifically, our system provides real-time risk assessment information that allows bridge operators to contact high-risk vessels or close the bridge.

Site includes official correspondence between NTSB and FHWA. FHWA has not yet responded (29Dec25).

This recommendation was voted at 18 Nov 2025 NTSB Board Meeting (Youtube).

Site includes official correspondence between NTSB and AASHTO.

For reference, AASHTO took 11 years to codify updates following the Sunshine Skyway bridge collapse in 1980.

• guidance and assistance from the Federal Highway Administration, US Coast Guard, and US Army Corps of Engineers Interdisciplinary Team identified in Safety Recommendations H-25-1 and H-25-2, and
• short- and long-term strategies to reduce the probability of a potential bridge collapse from a vessel collision. (Urgent)

Definitive design standard that has been partially superseded. Much of the content has been incorporated into the AASHTO LRFD Bridge Design Specifications, 10th Edition (2024). However, this Guide Spec remains relevant and includes additional information and context not found in the more general LRFD specification.

Particularly when calculating the “Consequences of Collision” associated with our real-time threat ranking of vessels, Roebling Labs will follow the formulas and procedures established by AASHTO in its design standards.

The code also contains applicable information to inform our approach for estimating “Probabilities of Collision” in real-time.

High level overview of NTSB investigations into bridge and tunnel collapses over the last few decades, with open recommendations to industry and owners.

Includes recommendations related to the need for vessel collision warning systems for vulnerable bridges.

Also includes recommendations related to warning systems to alert the traveling public immediately after a bridge has collapsed, to avoid driving into the void.

MAR8103 - RAMMING OF THE SUNSHINE SKYWAY BRIDGE BY THE LIBERIAN BULK CARRIER SUMMIT VENTURE, TAMPA BAY, FLORIDA

Includes the following recommendations to FHWA:

• M-81-18: Develop standards for the design, performance, and installation of bridge span failure detection and warning systems.
• M-81-19: Establish criteria to evaluate the need for installing bridge span failure detection and warning systems on existing and proposed bridges.

Definitive investigation into the March 1980 collapse of Sunshine Skyway bridge. Sudden storm caused loss of visibility while blowing ship off course.

This incident spurred the modern era of considering vessel collision (strength and avoidance) during the design of new bridges with 1st Edition AASHTO Guide Specifications for Vessel Collision Design for Highway Bridges (superseded) that was published in 1991 or 11 years after the collapse of Sunshine Skyway.

Describes the possibility that “the most practical warning device is considered to be some type of electrical conductor attached to or part of the bridge which will activate warning systems and/or gates when continuity is disrupted (span collapse).”

And advises owners that: “Federal funds may be used to construct warning systems on existing bridges and on new construction…”

Appendix A provides descriptions of “Alternative Surveillance and Warning Systems for Bridge Subject to Ship Collisions.” Three basic categories of systems are identified: I. Prevention, II. Detection, and III. Warning. Basic descriptions with pros and cons of each type and sub-type of proposed systems are presented, along with sketches.

I. Design Category - Prevention

• A. System Element - Pier Markings
• B. Subsurface Attenuation Devices
• C. Fender System

II. Detection
with Probabilistic Determination

• A. Ship Channel Surveillance (Laser)
• B. Radar System
• C. Closed Circuit Television
• D. Radio Beam Guidance System

with Impact Occurrence

• E. Bridge Continuity
• F. Pier Vibration
• G. Roadway Delineation (Reflectors)
• H. Roadway Delineation (Rail Lighting)

III. Warning

• A. Passive Signing with Flashers
• B. Dynamic Signing
• C. Signals and Gates
• D. Lane Control Signs
• E. Audible Alarm and Signs
• F. Citizen Band Radio
• G. Traffic Signals

A table scoring various combinations of the above elements is also presented.

An authoritative report from a committee formed from the top experts from industry, government, and academics to provide a broad overview of the problem and potential solutions to ship collisions with bridges.

Among other recommendations, is the “Urgently needed … Criteria for motorist warning and restraint systems.”

A sub-chapter on Motorist Warning Systems provides a less detailed discussion of ideas presented in Appendix A of the 1983 FHWA memo above.

Furthermore, the report includes an entire chapter called “The Historical Record” that lists and summarizes 20 significant ship collisions with bridges between 1960 and 1982, including three separate collisions with the Chesapeake Bay Bridge-Tunnel in Virginia.

The annual occurrence of serious ship-bridge collisions worldwide was 1.5 for the period 1971 to 1982.

”The statistical base is also more anecdotal than complete but is nevertheless sufficient to indicate that the damages, economic effects, and, most importantly, the loss of more than 100 lives (56 in the United States [from 1962 to 1982]) from ship collisions with bridges far exceed those from earthquakes, winds, and waves.”

”The record of ship-bridge collisions reveals some preliminary considerations for design that would not normally occur to a bridge engineer. … For example, in 19 ship-bridge collisions worldwide … 13 were struck in approach piers and only 6 in the main piers. In three of the collisions with approach piers, the side superstructure of the bridge was also struck.”

Highway traffic continued to drive into the void created by the collapsed spans after being rammed by empty barges after the tug boat captain lost consciousness. This incident included 14 fatalities and 5 injuries.

Note that this incident occurred on an inland waterway, not an ocean channel that has been the recent focus in response to the FSK collapse in 2024.

”New Recommendations” included:
To FHWA: “Develop an effective motorist warning system to stop motor vehicle traffic in the event of a partial or total bridge collapse. (H-04-30)”
To AASHTO: “Once an effective motorist warning system has been developed, provide guidance to the States on its use. (H-04-31)”

”The Safety Board has addressed the installation of bridge motorist warning systems in previous accident investigations involving the Lake Pontchartrain Causeway, Sunshine Skyway Bridge, and Sidney Lanier Bridge.”

Between 1992 to 2001 the Coast Guard/AWO - Allision Work Group database included 2,692 bridge allisions involving inland towing vessels and barges — one bridge alone was struck 170 times. USACE calculated a rate of 6 bridge allisions for every 10,000 towing vessel trips. Of the 2,692 allisions studied, 61 were classified in “Severity Class 4”, the highest, involving damage over $500,000 or loss of life or injuries or missing persons or oil spilled. Three of the incidents had fatalities.

”Bridge allisions involving towing vessels continue to occur, often with catastrophic consequences. In the past 10 years, 70 people have died in such accidents, all third parties who happened to be crossing the bridge at the time of the accident. The Coast Guard-AWO Work Group reported that 68 percent of bridge allisions were caused by poor operator decision-making. Consequently, remedial action to prevent or to reduce the frequency of bridge allisions involving towboats must focus not only on the physical characteristics of waterways but also on measures to improve the ability of towing vessel operators to make sound decisions.”

”Although ODOT installed pier protection cells inside the navigation channel, the I-40 bridge accident occurred outside the navigation channel. Such occurrences demonstrate that most bridges over navigable water can be struck either within or outside the regular navigation channel by barge tows and individual commercial vessels, thus increasing the complexity of bridge protection.”

The report includes a detailed chapter on Motorist Warning Systems with examples from Lake Pontchartrain Causeway, Sunshine Skyway, and related bridge sensor (SHM) research underway at various universities and FHWA.

”The most vulnerable locations for major damage to a bridge over water are the piers. Our code does not consider carefully enough collisions with barges and other commercial or enemy craft.”

Appendix C - Previous Recommendations on … Motorist Warning Systems provides summaries in response to NTSB investigations and reports of the below incidents:

• Sidney Lanier Bridge collapse in 1972;
• Lake Pontchartrain Causeway in 1974;
• Benjamin Harrison Bridge in 1977;
• Sunshine Skyway Bridge in 1980;
• Poplar Street Bridge in 1983;
• Poplar Street Bridge in 1984;
• Sidney Lanier Bridge in 1987;
• Judge William Seeber Bridge in 1993;
• Big Bayou Canot Railroad Bridge in 1993;
• Million Dollar Bridge in 1996;

International committee of technical experts writing on:

• Existing design codes in USA, Sweden, Norway, France, and Germany, and Eurocode
• International database of ship-bridge collisions with statistics and concluding remarks: leading causes include “bad maneuvering, insufficient clearance, storm, and ship failure. 53% are attributable to human error, 19% technical failure, and 23% extreme circumstances. Near misses or small accidents at a site are a good predictor of a serious collision in the future.
• Probability models including AASHTO, IABSE, Eurocode, UK, German, and applied to specific case-study bridges around Europe.
• Safety Criteria

”There are no generally accepted safety criteria around the world. But for structural failure, the criteria found show a range of 10^-4 (0.0001) to 10^-5 for the design life time of the structure (~100 years). Somewhat contradictory with this are the criteria found specifically for ship collision with bridges: they range from 10^-3 to 10^-4 per year, i.e. 10^-1 to 10^-2 per life-time of 100 years! These probabilities of collapse are rather high, and will probably result from economic considerations.”

• General Conclusions

”It would appear that more dangerous situations, such as bad weather or bends in the navigation channel, lead to more careful behavior, this not necessarily leading to more ship bridge collisions. This is a very tentative conclusion though.”
”The most advanced practical code is the United States AASHTO code.”

Recommendations include “development of warning systems, aimed at both ship captains and bridge passengers;”

Definitive international undertaking in the aftermath of Sunshine Skyway Bridge collapse and during the design of Great Belt Bridge in Denmark. This report documents the best practices developed in response to those two major events.

Section 10 “Protection of the Public” includes an overview of devices to detect vessel/bridge collision hazards, including vibration detectors, continuity circuits, and VHF radio link. It also includes an overview of manual or automatic traffic control devices including variable message signs, flashing beacons, and movable gates.

The section “Ship Domain Analysis” (page 11) introduces the concept of “Bumper Area” around each vessel that should not overlap with other vessel or obstructions such as bridge piers.

The dimensions of the bumper areas are:

• Open water at service speed: 8.0 L by 3.2 L
• Narrow channels and harbors at reduced speed: 6 L by 1.6 L

NTSB Safety Recommendation M-94

Unable to find an online copy of NTSB Report HAR-94-03 U.S. Towboat Chris Collision with Judge William Seeber Bridge.

The NTSB HAR 04/05 includes discussion on this incident.

Formal Investigation into the Circumstances Surrounding the Allision Between the Barge Tow of the M/V Brown Water V and The Queen Isabella Causeway Bridge on September 15, 2001 in Port Isabel Texas, Resulting in Multiple Loss of Life

The NTSB HAR 04/05 includes discussion on this incident.

The 291-foot-long cargo ship failed to turn with the channel and struck a pier and superstructure of the concrete box-girder highway bridge, wedging under before coming to a halt against the adjacent railroad bridge superstructure while shearing the ship's wheelhouse in the process. The bridges received minimal structural damage and re-opened to traffic within hours. The life loss was 1 person, the ship's chief officer, who was later found to have blood alcohol concentration three times the legal limit and is believed to have fallen asleep at the wheel at least 10 minutes before the collision.

"The task of VTS Great Belt is to supervise the Great Belt [vessel] traffic to protect the bridges spanning the Great Belt and to alarm in time to stop rail and the road traffic in case of a risk of collision with the bridges."

Despite the VTS Great Belt having a full-time staff of three people assigned to tracking nearby vessels at all times, the VTS did not become aware of the incident until receiving the Mayday call from the ship personnel's radios 6 minutes after the collision had occurred. It took the VTS an additional 4 minutes (10 minutes after collision) to sound the alarm to close the roadway and an additional 12 minutes (18 minutes after the collision) to sound the alarm to stop the trains.

The automatic alarm function had been turned off 8 years earlier because of excessive false alarms by relying solely on a 10-minute straight line projection of each vessel's current course.

A project was underway to modernize the automatic alarm functions with improved vessel trajectory forecasts and a goal to activate "if a ship that could constitute a hazard to the bridge unintentionally approaches the bridge at a distance of less than 10 minutes sailing time." The system update will also alarm if a vessel exceeds a speed of 7 knots in designated areas around the bridge.