Site investigation is often criticised for being slightly archaic and slow to adopt new developments, but when a new bridge was proposed across Oakland Bay in California, USA, it was anything but.
A vast array of sophisticated marine drilling equipment, geophysical techniques and a continually updated GIS database was used to integrate investigations into a tight design programme to form a detailed picture of the ground conditions.
The San Francisco-Oakland Bay Bridge is the main link between the two cities, carrying up to 280,000 vehicles every day. The 10 lane, double decked bridge has two sections, the east and west spans, that meet at Yerba Buena Island in the middle of the bay.
The area is one of the most seismically active regions in California. The bridge is only 18km from the San Andreas fault and 9km from the Hayward fault. In 1989, the 3.5km long east span suffered extensive damage when a section of the upper deck collapsed during the Loma Prieta earthquake.
Given the cost of making the east span meet seismic safety standards, it was decided to replace it with a new bridge to the north. The California Department of Transportation (Caltrans) commissioned geotechnical contractor Fugro-Earth Mechanics (a joint venture of Fugro West and Earth Mechanics) to carry out the site investigation for the new structure in January 1998. Work started soon after, comprising an extensive geophysical survey and two phases of land and over-water exploration.
Bridge design is being carried out by a joint venture of TY Lin International and Moffatt & Nichol Engineers. A 30% design was submitted in May 1998 and in June 1998 a single tower, self-anchored suspension main span and concrete skyway structure was selected as the final design. Present requirement is to produce a 65% design.
Concerns over the seismic safety of the existing bridge mean work is on a tight programme. It was decided that the best way to meet this was to run the site investigation, foundation design, earthquake engineering and preliminary structural analysis alongside one another. Key to this was the use of a GIS database.
Investigations were initially planned using data from the original bridge design, a proposed seismic retrofit of the bridge and adjacent investigations.
This information - lithology, borehole logs, laboratory test data, seismic data (P and S wave velocities) and historical aerial photographs - was put into the GIS database to create a subsurface model that could be continually updated as investigations were under way.
Marine geophysical surveys were carried out in January and February 1998 along a 3.75km long and 1.7km wide corridor centred over the proposed bridge alignment.
Bathymetry, side scan sonar, sea floor profiling, 2D seismic reflection and high-resolution 3D seismic reflection surveys were used to obtain an accurate picture of the topography, the nature of seafloor material and to locate any likely obstructions to construction.
Using 3D surveys is relatively unusual for coastal infrastructure projects. These surveys, which used multiple seismic cables and closely spaced vessel tracklines, built up a volume of closely spaced data that was invaluable in creating the geological model, especially where the bedrock sloped steeply and where stratigraphy was complex. One feature revealed by the surveys was a nested set of buried channels on the proposed route.
Results from these surveys, and data from land-based geophysics on Yerba Buena Island, were integrated into the database to refine the subsurface model and to decide locations for the geotechnical investigations.
The first phase of the geotechnical work comprised over-water drilling, sampling and extensive insitu and laboratory testing at 14 locations.
As most of the historical data was from boreholes to the south of the proposed route, it aimed to obtain high quality, relatively continuous (vertically) data to further characterise the geological units in the bridge area and to establish preliminary parameters for foundation design and earthquake site response analysis.
Equipment used for the over-water work is more commonly associated with offshore investigations and is rarely used for coastal ones. Logistically it is more complicated to use and obviously is far more expensive and sophisticated than conventional methods.
Boreholes were put down from two drilling barges into bedrock, up to 150m deep. Insitu CPT and vane shear tests were carried out between sampling as drilling proceeded. A full suite of downhole geophysics was carried out in each borehole. Three boreholes and three CPTs were put down on land at the bridge approaches.
Testing was carried out in a laboratory set up on one of the drilling barges and the data from this, along with the drilling and borehole logs, was emailed daily to the office and entered into the database.
This produced 2D and 3D cross-sections, maps and engineering design parameter plots, allowing engineers to refine the model and decide the locations of the second phase of geotechnical investigations.
Subsurface conditions were interpreted by correlating seismic reflectors imaged in the geophysics, velocity discontinuities measured in the downhole geophysical logs and stratigraphic boundaries observed in the boreholes.
Because the geophysics provided relatively continuous information over the area, it was possible to extrapolate and interpolate geological boundaries between the relatively widespread borehole locations.
The preliminary earthquake site response analyses were carried out for the 30% design of the bridge. These were based on probabilistic seismic hazard analysis to evaluate the likelihood of exceeding various levels of earthquake derived ground motion over a range of frequencies.
Comprehensive pier by pier earthquake site response analysis and dynamic soil-structure interaction analyses was carried out during final design. Site response analysis at critical points along the structure was conducted to evaluate the amplification and attenuation of ground motion at various depths within the ground profile.
After the bridge alignment and foundation locations were chosen, a second phase of marine site investigation was carried out towards the end of 1998.
Thirty boreholes were put down to depths between 100m and 120m at pier locations, with data to further refine the subsurface model, to finalise foundation design and make earthquake engineering evaluations for the pier locations. Thirty boreholes and 14 CPTs were also carried out on the bridge approaches during this phase.
The model revealed that bedrock slopes steeply eastwards from Yeuna Buena Island. Further east, the rock surface slopes relatively gently under most of the skyway. The bedrock is overlain by marine clays and alluvial sand and gravel. The soil thickens eastward to a maximum of 135m at the Oakland end of the route.
The main span pylon is in an area of shallow bedrock and is likely to be supported on large diameter drilled piles socketed into the rock. However, the main span east piers and skyway are underlain by over 85m of sediment and because of the relatively high seismic loads, large diameter (between 2m and 5m), up to 100m long raked driven steel pipe piles were chosen here.
These piles are typically used for offshore structures to cope with large cyclic environmental loads. The piles will be driven open ended and partially filled with concrete.