Preserving the Historic Keystone Bridge

To view the figures and tables associated with this article, please refer to the flipbook above.

The iconic Keystone Bridge in Elkader, Iowa, was built in 1889 to provide the city with a long-lasting river crossing solution (Fig. 1). Before its construction, several timber and iron bridges were built to cross the Turkey River, but they all deteriorated rather quickly and were found structurally unsafe within a decade of their construction. In 1888, the city decided to replace the latest iron truss bridge with a more durable stone arch bridge. The historic bridge, one of the largest and best constructed twin arched stone bridges in the Midwest United States, has seen numerous repair campaigns over time but continuing material deterioration and the appearance of longitudinal cracks in each of the two arches prompted a low load posting of 5 tons. In 2015, the City of Elkader initiated plans for what was set to be the most extensive rehabilitation of the bridge since its construction in 1889. The city identified the AASHTO HS20 truck and the Iowa DOT legal vehicles as the target live loads for the rehabilitation. Construction started in April 2022; the structural repair scope of work included a new cast-in-place reinforced concrete deck, a new cantilever sidewalk attached to one of the spandrel walls, pinning of the spandrel walls, and stone masonry repair and strengthening to increase the longevity and load carrying capacity of the bridge to meet current design loads.

The Keystone Bridge’s location was particularly suitable for the use of stone masonry due to the presence of shallow, competent bedrock and a nearby quarry to source native limestone. Bids were requested, however, for the construction of either an iron bridge or a stone bridge: it turned out that the construction of an iron bridge was more expensive than a stone bridge with the same width, and the construction of the stone bridge was awarded for $13,000. The design by Mr. Tschirgi called for a double arch structure, 204 feet in length and 30 feet wide with a 6-foot sidewalk on the interior of the bridge. Each arch is 3 feet in thickness at the crown section and spans approximately 84 feet with a clear height of 28 feet. When the Keystone Bridge was constructed, it was described as “the longest and best constructed highway bridge west of the Mississippi River and east of the Rocky Mountains” (Engineering News 1891).

Construction of the bridge started in September 1888. The foundation was laid directly upon the native bedrock and filled with hydraulic cement grout. The centering—the temporary wooden structure used to support the arch during construction—consisted of only five frames, likely spaced at 7.5 feet on center, covered with closely spaced boards or planks upon which the stone units were laid (Fig. 2). About 50,000 feet of lumber were used for the centering and about 4,160 cubic yards of materials were used for the construction of the bridge. The weight of the bridge was estimated at more than 18,000,000 pounds. Although the bridge was originally expected to be completed in December 1888, it took until August of 1889 to finish the work. The total construction cost of the Keystone Bridge was reportedly around $16,300, equivalent to approximately $550,000 in today’s value.

Details of the bridge design became available when Mr. Tschirgi presented a paper to the Iowa Civil Engineers and Surveyors’ Association in 1889. With the horizontal thrust at the arch crown section estimated at 65,000 pounds and an assumed allowable compressive strength of 150 psi, the depth of the keystone, which is the wedge-shaped stone located at the apex of the arch, was calculated at 3 feet. This value is consistent with typical empirical rules available at the time of construction to define the thickness of the keystone.

Notably, the old iron bridge being replaced was reportedly kept open during construction to provide a convenient river crossing even though the new bridge was being erected in the same place as the old bridge. Evidence of this construction sequence is found in a historic photo showing the Keystone Bridge under construction: the deck and piers of the old bridge are still visible in the background (Fig. 2). The simultaneous construction of both arches was not possible due to the presence of the old bridge piers. This resulted in some concerns about the stability of a single arch and whether there was enough frictional resistance under the center pier to resist the horizontal thrust in the arch, with calculations showing only a safety factor of 1.8. Fortunately, construction was completed without problems and the bridge has served the City of Elkader ever since. Because of its historical significance, the bridge was listed on the National Register of Historic Places in 1976.

The evaluation process for the bridge’s rehabilitation began with the review of existing documents related to its construction and past repair campaigns to gather information in support of the structural analysis. The design team conducted a visual condition survey and found the stone masonry to be in fair condition overall, with localized areas of cracks and deteriorated limestone. Surface cracking and spalling of the limestone were visible along the south face of both arches, most likely due to freeze-thaw cycles. Longitudinal cracks were observed on the underside of the arches, with the condition most severe near the arch midspan. The presence of water stains at the underside of the bridge was an indication of water migrating from the top of roadway surface down into the arches, providing an active source of moisture for damaging freeze-thaw cycles and material deterioration. The design team planned to conduct a thorough condition assessment of the bridge, including surface penetrating radar (SPR) investigation of the spandrel walls and arches to evaluate typical stone thickness and solidity of the stone masonry construction. Typically, the arch thickness, a key factor in the determination of the structural capacity of an arch bridge, can be measured at the spandrel walls. Radar scanning is used not only to confirm that the arch barrels have the same thickness away from the spandrels but also to evaluate any loss in cross-section due to material deterioration that could adversely impact the structural capacity of the arches. Results from such investigation provide valuable information to support the structural analysis of the bridge and reduce conservative assumptions. However, due to the low load posting of the bridge, the team faced significant limitations in accessing the bridge with an underbridge inspection vehicle and was unable to position the vehicle effectively to conduct a detailed evaluation of the bridge. Without a close evaluation, the full extent of the stone deterioration could not be fully captured. In one specific instance, the severity of deterioration that was uncovered during the stone repair work warranted the use of temporary shoring (Fig. 3). The shoring was put in place to provide support to the arch, allowing the team to safely replace six stone units that were structurally unsound and separated from the arch. The units were located at the south face of the arch, between the quarter point and the crown sections, and replaced with stone sourced locally. The shoring consisted of two triangular steel frames with the addition of a curved section on top to follow the arch profile. This solution closely resembled the wood centering used in the original construction of the bridge.

Stone arch bridges exhibit complex structural behavior under gravity and vehicle loads. Often, the inherent strength of the original materials and design is diminished by conservative assumptions made to simplify the structural analysis. Less conservative results and load ratings can be obtained with a 3D finite element analysis (FEA) capable of capturing more realistic load and stress distributions within the structure. The main drawback of FEA is the higher modeling and computational costs. Considering the historic significance of the bridge, a full 3D analysis of the structure was seen as the most appropriate analysis method.

Atkinson-Noland created the computer model of the bridge in Midas FEA (Fig. 4). Continuum elements were used to model the arches, piers, spandrel walls, soil, and concrete deck. The stone masonry was modeled as a homogeneous material, without distinction between mortar joints and stone units. To account for the composite action, material properties representative of the mortar-stone assemblage were assigned to the masonry and selected based on typical published values for similar materials. The load rating analysis was conducted following the AASHTO Manual for Bridge Evaluation and only dead and live loads were considered. The analysis incrementally moved live loads along each design lane in 3.5-foot increments to find the maximum internal forces at critical arch sections (i.e., crown, quarter points, and springlines) to be used in the rating calculations. Atkinson-Noland calculated rating factors based on an evaluation of maximum tensile and compressive stresses developing in the arches in the longitudinal and transverse directions. To account for uncertainty in the collection of as-built data and other unknowns, a knowledge factor, l, equal to 0.9 was used to adjust the bridge capacity. In calculating component capacities, a knowledge factor is used to express the confidence with which the properties of a structure or structural components and the presence and extent of damage are known. Engineering judgment is used to establish the knowledge factor value from the information obtained from original construction documents and/or condition assessments, including destructive and nondestructive testing of representative components.

Atkinson-Noland also used the finite element model to evaluate the construction sequence and construction loads for potential overloading of the arches as a result of an unbalanced load condition or localized stress concentrations from construction vehicles (Fig. 5). An analysis was also carried out to provide design loads for the temporary shoring needed to support the arches during the replacement of full stone units. Shoring an arch bridge requires support in both horizontal and vertical directions at the bracing points. The 3D model of the bridge was used to evaluate the shoring approach, provide shoring loads, and help optimize the shoring design to mitigate additional construction costs.

The rehabilitation work included repointing of all deteriorated mortar joints with compatible mortar. Damaged and deteriorated stone, such as spalled and cracked units, were repaired with dutchmen, where deteriorated material is removed and a new stone keyed into the existing masonry and secured with stainless steel pins designed to facilitate mechanical locking and prevent possible slippage. Because the bridge is listed on the National Register of Historic Places, the rehabilitation plan followed the principles of minimal intervention as defined in the Secretary of the Interior’s Standards for the Treatment of Historic Properties. The compatibility of the replacement stone was evaluated through axial compression and absorption tests carried out by the masonry contractor. Atkinson-Noland conducted mortar analysis to identify binder to aggregate ratio, aggregate color, and aggregate size gradation to provide a compatible replacement mortar formulation.

The longitudinal cracks in the arch rings below the spandrels were repaired by installing stainless steel anchors grouted in core holes drilled from the face of the arch ring and extending several feet into the arch. The end piece (or plug) from cores was retained and reinstalled with restoration mortar matching the color of the surrounding limestone to cover the anchor at the face of the wall and conceal the application. The anchors were also designed to enhance the flexural tensile capacity of the arches in the transverse direction.

The historic Keystone Bridge is an iconic structure at the heart of the City of Elkader’s historic district. After a successful rehabilitation project, the 1889 bridge was repaired and strengthened to increase its load carrying capacity to meet AASHTO and Iowa DOT load requirements and will continue to serve the city for many years to come. The bridge was rededicated in 2024—the 135th anniversary of its original construction—and the rehabilitation project was awarded the 2024 Excellence in Archaeology and Historic Preservation Award by the State Historical Society of Iowa Board of Trustees.

About the Author

Carlo Citto, PE, SE, is a Structural Engineer and Principal at Atkinson-Noland & Associates. He has over 15 years of experience in structural analysis and load rating of stone masonry arch bridges and has been involved in the rehabilitation of many historic arch bridges throughout the country. He can be reached at (ccitto@ana-usa.com).