The existing Upper Pony Creek dam, located on the southern coast of Oregon, US, is typical of the region’s water storage dams built in the 1950s. It was constructed as a simple homogeneous embankment using local soils. However in the early 1980s two developments highlighted the need for a new dam.

First, the scientific community began to support and advance the theory that a mega-thrust earthquake might occur off the Oregon coast. It is predicted that this seismic event could have an approximate magnitude of 9.0 on the Richter scale, which is prompting a radical upgrade of the design criteria for all dams in the coastal Northwest US. Secondly, the area’s attractiveness as a community and coastal recreational destination is expected to result in higher water demands. In recent years, the Coos Bay-North Bend Water Board (CBNBWB) has found it necessary to request voluntary water conservation to avoid demand deficits.

The Coos Bay-North Bend watershed is located entirely within the city limits of Coos Bay, a town with a population of about 15,000. The neighbouring town of North Bend has a population of about 10,000. The watershed’s almost pristine nature has been preserved by limiting public access and conducting selective logging activities. The current pool area is about 52.6ha.

The existing dam is an earthfill structure with a height of approximately 13.7m. Local soils were placed to create a crest width of 4.6m and a crest length of 117.4m. The upstream and downstream slopes are 3H:1V and 2H:1V respectively, and the embankment volume is approximately 32,878m3.

The dam, which has a concrete lined spillway channel with a reported capacity of 39.6m3/sec, stores water from a 7.5km2 drainage area with a total average annual yield of approximately 6.7M m3. Only 2.7M m3 is retained behind the existing dam.

In 1997, CBNBWB decided to increase storage capacity and construct a new dam at Upper Pony Creek. The new structure will raise the pool elevation by 6.4m and provide a total of 7.7M m3 of storage. A 0.2m3/sec pump station and pipeline constructed under a separate contract will add extra water to this reservoir from an adjacent drainage.

New design

Work on the new dam began in January 1998 with an evaluation of possible designs. In recent years rockfill and roller compacted dams have been successfully constructed in the region. These alternatives were compared to the zoned-fill earthen embankment option.

Two independent seismic faults were analysed for their potential effects at this site in order to determine the appropriate seismic design criteria for the dam. The Barview Fault, approximately 2.4km from the site, could result in an earthquake with a magnitude of 6.5. The Cascadia subduction zone earthquake could occur about 37km off the coast and result in a magnitude 9 event with a peak ground acceleration of over 0.8g at the site.

Geotechnical investigations indicated that the foundation rock might not provide sufficient bearing resistance for a roller compacted dam. In addition, investigations revealed that suitable core and shell material are probably available onsite, so the comparison favoured an earthen embankment dam for economic reasons.

Six geological units were identified at the dam site, including the existing dam materials, waste materials (materials deposited from the original dam construction), topsoil, alluvium, marine terrace deposits and sandstone.

The investigation indicated that bedrock at the dam site ranges from very weak to friable, and fine to medium grained sandstone with thin shale interbeds. Tests showed it was deeply weathered on the right abutment to depths exceeding 16.8m. The sandstone, part of the Eocene Age Coaledo Formation, was found to be highly fractured in core samples with rock quality designations (RQD) of about 70% (fair), but often below 50% (poor to very poor). Unconfined compressive strength tests also indicated that the sandstone has very low compressive strengths (mean qu = 43.6kg/cm2).

The channel area consists of generally saturated waste materials and alluvium overlying the sandstone bedrock. Investigations indicated that the alluvial materials are approximately 6-9m thick in the centre of the channel.

The left abutment showed topsoil overlying marine terrace deposits, which in turn overlie the bedrock. The marine terrace deposits in the vicinity of the dam site appeared to be predominantly dune sand. Tests on the dune sand showed it was uncemented to weakly cemented, and poorly graded fine sand with a low fines content.

Groundwater was encountered close to the surface, especially in the channel area of the site. Water pressure tests indicated that most of the bedrock has a moderate hydraulic conductivity value in the range of 10-4 to 10-5cm/sec.

Dam design features

The new dam will have a height of 29.6m above the lowest point of the foundation. The crest width will be 7.6m and the crest length will be 344.4m. The slopes of the upstream and downstream faces will be 2.5H:1V and 2H: 1V respectively. In total, about 382,300m3 of excavation and 313,486m3 of embankment fill are required in this project.

The dam core was designed with a base width of 16.8m. The core will consist of silty clay soils and must accommodate a deflection of up to 0.6m which could result from the design peak ground accelerations.

A filter and drain system, consisting of well graded sand and gravel with a low percentage of fines, is included in the embankment downstream of the core. A 3m wide vertical chimney of filter materials will minimise migration of the core materials. Seepage from the filter will be collected in a parallel 1.5m wide chimney constructed of drain gravel. Similar filter and drain arrangements will collect any seepage from the foundation.

A 6m, 205.4m long concrete cutoff wall was designed under the dam axis centreline to intercept zones of potentially high hydraulic conductivity near the base of the dam, and to reduce uplift pressure on the downstream shell.

A 6m wide spillway will be constructed on the right abutment. A relatively narrow spillway width was selected to minimise the excavation of the steep right abutment. The spillway chute will be constructed with 1.2m high baffles to provide energy dissipation. The spillway will handle the peak routed outflow of 31.8m3/sec of the probable maximum flood.

The 1.2m diameter intake pipe will pass up to 11.3m3/sec. It is protected by a bar rack and a hydraulically operated sluice gate. The outlet structure houses a 0.3m jet flow gate to discharge normal flows. A 0.91m knife gate will be installed if a rapid pool drawdown is necessary. All controls can be hydraulically operated from an outlet structure located near the downstream face of the dam.

A separate dam will be constructed on a small drainage northwest of the new Upper Pony Creek dam. The smaller dam, referred to as the debris dike, will retain the poor quality alluvial materials which must be removed under the foundation of the new dam. In addition, the surface of this spoil site will be graded and capped with wetlands soil, carefully stockpiled at the start of the excavation. The planting of saplings will complete the 1.7ha of wetlands. Additional mitigation wetlands will be created in the fringes of the new, elevated pool under a separate contract.


The project construction contract was awarded in April 2000 for approximately US$9.8M. The contractor immediately began to refine plans for a concentrated, two-year construction period. Dam construction in the Pacific Northwest must be carefully scheduled due to the relatively short construction season. Rain often continues into June and usually returns by mid-October. The contract documents require the contractor to remove and replace the poor quality foundation materials at the downstream base of the existing dam by 31 October 2000. Meeting this deadline is critical to preserving the safety of the dam through the winter of 2001.

Preserving both the safety of the existing dam and the continued water supply for the Board has posed two conflicting challenges. Lowered pool elevations minimise construction risks and facilitate dewatering efforts at the cost of some valuable water reserves.

An interim water supply plan was developed to help prevent future water supply shortfalls until the new reservoir starts to fill in late 2001. The plan attempts to maximise the contributions from supplemental wellfields and inter-basin pumping and will be used to predict if and when conservation measures are needed.

To date, the excavation of the foundation has been completed and embankment fill, including core, filter drain and shell materials are being placed. Filter materials are about to be placed. The exposed foundation was similar to the conditions predicted by the geotechnical investigation. Only limited dental concrete was needed to prepare the foundation due to the highly weathered nature of the exposed sandstone.

Finding, processing and protecting the critical core materials has been a primary concern. Acceptable core material was found in a borrow area high above the dam site which is difficult to access. The in-place moisture of the core material is about 20% above optimum. To date, sufficient core material for this year’s construction has been transported down from the borrow site and dried by disking in thin lifts.

Intake and outlet structure construction is expected to continue when the first-year embankment efforts are complete. Coos Bay’s mild coastal climate allows concrete work to continue year-round, therefore spillway construction is planned for early next spring. It is expected that all construction will cease onsite in early December and resume in May 2001. Given the current progress and the availability of soils, all parties are optimistic that the project will be completed by the end of September 2001.