Kinzua Viaduct (July 21, 2003)
Nathan C. Eck, BAE, Penn State, 2010


On July 21, 2003 the Kinzua Viaduct, centerpiece of the Kinzua Bridge State Park, was struck by an F-1 tornado causing it to collapse. The bridge was 2053 ft. long and at its highest point reached 301 ft. off of the ground. The bridge, originally constructed in 1882 of iron and then again in 1900 of steel, was in use until October 5, 1950. The railroad was then purchased for scrap by the Kovalchick Salvage Company in 1959. However, before salvage operations commenced, the company reconsidered and sold the railroad to the state and was declared a landmark in 1963. Since then the viaduct had suffered from considerable neglect which caused rusting of the anchor bolts and cracking of the concrete foundations. When the tornado struck the bridge was subjected to 94 mph winds which blew perpendicular to the span. The force from these winds resulted in the failure of the rusted anchor bolts and the uplifting of three support towers from their foundations. In passing, the tornado felled 11 of the 20 support towers with no injuries or loss of life.


  • Bridge Collapse
  • Kinzua Viaduct
  • Tornado
  • Structural Neglect
  • Design Oversight

Initial Construction

The first Kinzua Viaduct was constructed in response to the rapid growth of mineral resources during the late 19th century. The Commonwealth of Pennsylvania saw the development of a complex network of railroads traversing a variety of terrains in order to transport these minerals. Fossil fuels such as coal were especially needed in the Great Lakes cities and the favored method of transportation was by railroad. However, in order to utilize this method, the railroad would have to span many valleys, gullies, and gorges. (Leech, July 2005)

General Thomas L. Kane, owner of the New York, Lake Erie and Western Railroad and Coal Company, proposed the construction of the worlds tallest bridge. The reason for this proposal stemmed from the location of a potential railroad bridge that would provide access to the Great Lakes region of Pennsylvania. This location was the Kinzua Creek gorge which at its lowest point would provide the bridge with an elevation of roughly 300 feet.(Leech, July 2005)

Aldolphius Bozano of the Pheonix Bridge Company accepted the task, claiming that they would "build you a bridge a thousand feet high, if you provide the money...."(Leech, July 2005). They completed the task within 94 days with the resulting structure reaching a height of 301 feet at its highest point. The $275,000 viaduct spanned a total length of 2,053 feet divided into 41 spans and was constructed primarily of iron; particularly the patented wrought iron Pheonix Columns. At this time it was the tallest bridge in the world, but would only hold the record a mere two years before it was bested by the Garabit Viaduct over the Truyere River, Massif Central, France.(Leech, July 2005)


picture courtesy of DCNR

The first iteration of the Kinzua Viaduct would remain in use for twelve years after its completion. However, in that time railroad technology had vastly improved which allowed for heavier train loads. This lead to a necessary reexamination of the Kinzua Viaducts load carrying capacities and subsequent reconstruction. The second design of the viaduct used span and tower arrangements similar to the original design with the existing structure used as an erection platform for the new one.
The redesigned structure featured riveted, latticed, built-up steel columns and a Vierendeel lateral bracing system, omitting all diagonal sway braces(Leech, Nov. 2005) with the new structural towers mounted on the existing masonry pedestals.(Leech, July 2005)

In order to secure the new superstructure to the existing masonry pedestals, the designers opted to reuse the original 1-1/4" anchor bolts. The anchor bolts on the eastern face of towers 4 through 17 incorporated roller bearings to allow for lateral expansion of the the bridge. However the inclusion of roller bearings made the original anchor bolts to short to connect. This was remedied by the use of a steel collar coupling assembly which spliced the existing anchor bolts with the new steel anchor bolts(DCNR, Dec. 2003).

Operation: 1900 - 2002

As the 20th century progressed, nearby coal resources were depleted, and rail traffic diminished. The viaduct remained in service but activity was light. Finally, in 1959 the railroad company sold it for scrap to the Kovalchick Salvage Company. However, the owner of the salvage company, realizing the value of the Kinzua Viaduct, would abandon the salvage operation and instead sell the viaduct to the Commonwealth of Pennsylvania in 1963 for nearly the same price that the salvage company paid(Leech, Nov. 2005).

The commonwealth then established at the site a state park that featured the bridge as its centerpiece. Listed in the National Park Service’s National Register of Historic Places in 1977 and designated a civil engineering landmark by ASCE in 1982, the viaduct was used by a private railroad concessionaire from 1980 through early 2002(Leech, Nov. 2005). In 2002 an in-depth inspection for the Department of Conservation and Natural Resources determined that the structure was at risk to high winds which prompted the immediate closure of the structure to recreational pedestrian and railroad usage(DCNR, Dec. 2003) and a maintenance repair contract was issued in 2003.

Events Leading Up To The Collapse

Repair work began on the structure in mid-February of 2003 operating on a $3.9 million repair contract. The goal of the repairs was to return pedestrian and excursion-train traffic to the bridge. The contractor chosen to do the repair work assigned two full time crews to the task. The work entailed repairs to the bottom struts of the towers, the tower legs and their lacings, and the replacement of expansion bearings with a concentration on the bottom four members where corrosion was most severe(Gruber, April 2003). The repairs were carried out by W.M. Brode Co. up to the time of the collapse and continued on the remaining towers afterward.

picture courtesy of DCNR

On the afternoon of July 21, 2003, a light rain started to fall at Kinzua Bridge State Park, and by 3:00 p.m. the crew repairing the viaduct had left the work site and returned to the construction compound located adjacent to the bridge. A series of unfavorable weather conditions produced a severe weather event which would prove to be the largest severe weather outbreak in the Commonwealth of Pennsylvania for the 2003 severe weather season. A mesoscale convective system (MCS) crossed the Commonwealth that day which produced a series of spiral like cloud banks, which moved in counter clockwise direction as the entire MCS moved in an easterly tracking direction. At the leading edge of the front, the contribution of wind shear and moisture with afternoon instability initiated intense thunderstorms (DCNR, Dec. 2003).
Tornado activity appeared due to intense smaller vortices within the larger MCS system and at approximately 3:20 pm local time an F-1 tornado touched down. An F-1 tornado is recognized as a tornado event with associated wind speeds varying between 73 and 112 mph. The distinct, large-scale debris field at the site extended more than 1 mile in a northerly direction, revealing the counterclockwise, or cyclonic, winds within the vortex and associated inflow winds trailing the vortex(Leech, Nov. 2005).

courtesy of DCNR

The tornado’s counterclockwise vortex extended as much as 0.25 miles in diameter. At touchdown, the leading edges of the rotating vortex produced high, straight-line winds in the immediate vicinity of the structure. Because of its counterclockwise rotation, the tornado initially made contact the structure and the surrounding forest from the east. The vortex was fed by concentrated inflows of air channeled along discrete northerly tracts. These inflow winds put severe stresses on the structure while wholly bypassing the construction compound approximately 200 ft south of the bridge where construction workers and park attendants had assembled as they concluded their workday activities(Leech, Nov. 2005)


The complex, high velocity wind patterns produced by the tornado caused the collapse of 23 of the 41 spans and 11 of the support towers. The collapse was rapid but happened in three distinct but separate episodes as indicated below. The failure of the structure in three phases wast the result of the arrangement of the wind locks within the girder system. To control longitudinal thermal forces within the structural system, the 1900 design introduced expansion joints (and accompanying wind locks) at irregular locations within the structure such that individual continuous girders accompany the towers in either a triplet or doublet pattern. The breakup of the structure in three distinct episodes was controlled by the location of the wind locks. The locks permit limited longitudinal thermal movement and limited resistance to lateral forces(DCNR, Dec. 2003)).

picture courtesy of DNCR

Episode 1 (DCNR, Dec. 2003)

① Tornado touches down – easterly winds grow rapidly – local wind speeds (from the east) exceed 90 mph – as wind speeds grow, the towers oscillate laterally at their
natural frequency as a unit until the wind lock bolts shear and the rails pull apart.

② “Separation” failures occur within the “expansion” anchor bolt system of Towers 10, 11, 12, 13 and 14.

③ Small rotation about the “fixed” bearings of towers 10, 11, 12, 13 and 14 occurs followed rapidly by tensile failure of the “fixed” anchor bolts. The rails separate at Tower

④ Tower sections 10 and 11 and attaching girders separate at the wind locks (of Towers 9 and 12) and become airborne. Total collapse of this segment is rapid. Wooden
decking, rails and structural steel collapse as a unit. The wooden decking and rails come to rest in the immediate vicinity of the girders.

⑤ Towers 12, 13 and 14 initially become airborne and “jump” a small distance north and westward. Towers 12, 13 & 14 momentarily come to rest in the upright position on
the ground but do not initially catastrophically collapse. The rails momentarily hold the towers and momentarily prevent immediate catastrophic collapse of Towers 12,
13 & 14.

Episode 2 (DCNR, Dec. 2003)

① Tornado moves northward – easterly winds grow rapidly – local wind speeds (from the east) exceed 90 mph. As wind speeds grow, all towers vibrate at their respective natural lateral frequencies.

② Wooden decking and rails (spans 1 – 18) separate from the structure. (The rails, separated at Tower 12 during episode 1, ride down with Towers 10 and 11 pulling ties and decking from Towers 9 through 4.)

③ “Separation” failure occurs in sequence within the “expansion” bearings of Towers 9, 8, 7, 6, 5 and 4. Tower 9 fails, momentarily, followed by failure of Tower 8, etc.

④ Small rotation about the “fixed” bearings of Towers 9 thru 4 occurs in sequence followed rapidly by tensile failure of the “fixed” anchor bolts.

⑤ In rapid but distinct sequence, Towers 9, 8, 7, 6 and 5 individually become airborne, pivot approximately 90° clockwise about the base of the fixed bearings and strike the ground upon impact. Collapse is progressive from south to north.

Span 7 and Tower 4 are initially restrained by Tower 3; however, after fracture of the connection to Tower 3, rapid collapse and clockwise twist of the tower occurs. Tower 3, although standing, is visibly distorted.

Episode 3 (DCNR, Dec. 2003)

Tornado moves northward – rapid and confined inflow winds attack from the south.

② The wooden decking and rails (spans 24 – 29) separate from the structure. The wooden decking undergoes lift and falls in a northeasterly direction. The rails remain attached to Tower 15.

③ Spans 25 and 27 and Towers 12, 13 and 14, having separated from the bearings during Episode 1, now become airborne and are displaced to the north and east. The displacement induces torsional buckling of the columns of Tower 12. Collapse is total. Towers 12 & 13 twist 90° counterclockwise and collapse on top of previously collapsed Tower 11. Tower 14 twists and collapses separately.

④ Span 29 oscillates (laterally several times) at the Tower 15 connection, eventually separating and rotating upside down before impact. The rails remained attached and “hang” from Tower 12. (Subsequent to Board of Inquiry Investigation, the “hanging” rails were cut and allowed to drop to the ground.)

courtesy of DCNR


The structural forensics investigation for the Kinzua Viaduct was conducted by Gannett Flemming. It was determined that the failure of the bridge initiated at the "weak link" in the system. This "weak link" being the anchor bolts on the eastern face which were installed during the 1882 construction. This failure occurred as a separation within the anchor bolt system by one of two distinct modes which are described below.

picture courtesy of DCNR

Failure Mode 1: Separation at the Boundary of 1882 and 1900 Construction – Expansion Bearing/ Anchor Bolt Collar Coupling

This mode accounts for approximately 3/4 of observed separation failures. All collar couplings observed at the site exhibit a radial cracking pattern. The equiangular cracks completely penetrate the collar couplings. The observed couplings either remained engaged on the bearing assemblies or are found loose within the debris field. Coupling fractures show evidence of fatigue fracture with secondary fractures occurring by overload presumably during the collapse. Many of the couplings were fractured by fatigue prior to the time of the collapse incident(DCNR, Dec. 2003).

The 1900 construction provided washers “surrounding” the collar couplings. As a result, the collar couplings and associated cracking were “hidden” from view. The cracks within the collar couplings could only have been visually observed if the restraining bolts and washers were removed during an inspection cycle. Due to the complete penetration cracking of the couplings, a majority of the collar couplings were judged to be ineffective and could not transmit uplift forces to the substructure. Therefore, for analysis purposes, no uplift capacity was attributed at the locations where the 1882 anchor bolts remain in-situ, without attachment of collar couplings(DCNR, Dec. 2003)).

Cracked Collar Coupling (picture courtesy of DNCR)

Expansion Bearings (picture courtesy of DNCR)

Failure Mode 2: Ductile Failure (Separation) within the existing 1882 anchor bolts – Expansion Anchor Bolts

picture courtesy of DCNR

This mode accounts for approximately 1/4 of observed failures. Examination of fractured original 1882 anchor bolts showed that the fracture resulted from tensile overload and was a fully ductile fracture. The estimated tensile capacity of one 1882, 1-1/4 inch anchor bolts at failure, assuming a 20% corrosion loss, is 30 tons. Based on the observed, 3:1 ratio of collar coupling separation to ductile anchor bolt failures, an uplift capacity of 30 tons per tower can be estimated. This capacity establishes a lower bound, critical wind speed of 94 mph, which was sufficient to initiate failure which was sudden and catastrophic(DCNR, Dec. 2003).

The separation of the base of the tower from the pedestal left the structure vulnerable to winds from the east.

Lessons Learned

  • The design of tall, light structures with a large height to base aspect ratio will typically be governed by lateral loads and accordingly, these structures should receive additional attention with regard to failure in tension.
  • In tall, slender structures vibrations will prove a significant part of the design in that they can create inertial effects within the structure as should therefore be taken into account when designing structures such as the Kinzua Bridge.
  • Had the wind locks' design been more detailed to fully anticipate critical loads, more redundancy would have been introduced into the structure and the resulting failure may have been less catastrophic (Leech, Nov. 2005).


The Kinzua Bridge collapse occurred as a result of a lack of detail in design, the neglect of the structure throughout the years, and severe conditions brought on by in-climate weather. While the DCNR's report states that the reuse of the existing anchor bolts was acceptable in the redesign of the viaduct, more detail should have been paid the the design of the collar coupling linking the existing anchor bolts to the new anchor. Additionally, a structure such as this in which all of the load bearing members are exposed to natural elements, should be routinely inspected for structural deficiencies. Had the structure been properly monitored in the past, the catastrophic failure of Kinzua Viaduct may have been avoided or at the very least averted long enough for proper repairs. Finally, the natural elements are truly the greatest foes when it comes to prolonging the life of a structure and should never be underestimated during the design process.


“Emergency Repair Work Begins On Kinzua Bridge In Mckean County .” News Release. 27 Feb. 2003, Department of Conservation and Natural Resources. 28 Sept. 2010 \


News Article: This source is a news release provided by the Department of Conservation and Natural Resources which discusses the commencement of restoration
efforts for the Kinzua Viaduct.

Gruber, John. “Kinzua Bridge To Get Emergency Repairs.” Trains. April 2003: 76-78.

Magazine: In this article John Gruber gives a summary of the restoration efforts that were set to begin just before the collapse of the bridge.

“High Winds Topple Historic Railroad.” 22 July 2003. 27 Sept. 2010 <>

News Article: recounts the collapse of the Kinzua Viaduct and the condition of the bridge before the collapse.

“Kinzua Bridge Report.” Info. Dec 2003, Department of Conservation and Natural Resources. 28 Sept. 2010 <>

Report: This is the official report produced by the Department of Conservation and Natural Resources providing a wide range of information such as initiation of failure,
progression of collapse, and mechanics of collapse as well as an abundant amount of photos.

“Kinzua Railway Viaduct.” History & Heritage of Civil Engineering. ASCE. 27 Sept. 2010 <>

Website: This source provides general information on the Kinzua Viaduct.

Leech, Thomas G. “Lessons From The Kinzua.” Civil Engineering Nov. 2005: 56-61.

Trade Publication: This source provides an analysis of the Kinzua Viaduct collapse and recommendations for future structures that may be susceptible to high winds.

Leech Thomas. “The Collapse of the Kinzua Viaduct.” American Scientist. July/Aug. 2005: 348-353.

Trade Publication: This article, written for American Scientist, gives a general overview of the Kinzua Bridge collapse and the main factors contributing to the collapse.

“Officials fear popular Pa. tourist attraction is near collapse.” 22 Dec. 2002. 25 Sept. 2010


News Article: This periodical, published just over half a year before the Kinzua Viaduct collapsed, expresses concern about the structural stability of the bridge and
discusses potential costs for repairs.