The project is located in Southern California where the average rainfall is only about 300mm per annum. Responding to rapid growth in Southern California, the client for the project constructed the Colorado River Aqueduct system in the 1930s. This was followed in succession by the State of California constructing the State Water Project and California Aqueduct system in the 1950s and 1960s. As supplies from the Colorado River have been increasingly reduced in recent years due to the growth of neighbouring states, the Metropolitan Water District of Southern California (MWD) planned to expand its capacity to deliver water to Southern California via the State Water Project.

The MWD serves 18 million people in Southern California over an area of 5,200 square miles and delivers up to 7.5 million cubic metres of water daily. The Arrowhead Tunnels project is a key component of the MWD’s larger Inland Feeder Program that when completed will more than triple the present water supply capacity from this source to 28 cubic metres per second; this also includes the recently completed Diamond Valley Reservoir which has a storage capacity of 1 billion m3 of water.

Construction of the two tunnels comprised 12.9km of 5.8m diameter tunnels bored through hard rock by twin TBMs near the base of the San Bernardino Mountains and passing within a kilometre of the San Andreas Fault and crossing numerous other significant faults, see Figure 1. Ground conditions range from highly fractured to massive, moderately jointed igneous and metamorphic rocks with fault and shear zones. Water pressures in the tunnel area reach as high as 20 bar.

Construction commenced in 2002 and is due to be complete in 2009, with an initial bid value of $242m. Construction Managers for the project are Hatch Mott MacDonald, with design support provided by Jacobs Associates and GeoPentech. The contractor is a joint venture of SheaKenny. Mike Bell, resident engineer for Hatch Mott MacDonald gave an overview of the project and Brian Fulcher, project director for Shea-Kenny JV presented some of the problems and challenges on the project.

Project History

The Arrowheads Tunnel project first started back in 1994, with a construction contract awarded to Shank Balfour-Beatty in 1996. “Junk” segments were used for the tunnel construction, these had no gaskets, but were grouted. Water inflows of up to 100l/s were encountered which affected the local water supplies from the aquifer. An injunction was served by the United States Forest Service (USFS) and construction stopped in 1999. The length of tunnel constructed during this first contract is shown as dotted on Figure 1.

The project then went through a re-design between 2000 and 2002, this time with a fully gasketed and sealed segmental lining. To re-start construction a USFS Special Use Permit was required; this limited groundwater in-flows into the tunnel, records of which had to be posted regularly on the project website. The second construction contract was awarded to Shea-Kenny in 2002 with tunnelling re-commencing in August of 2003.

Challenges and Milestones

Some of the major challenges encountered included the variable ground and water conditions, see Figure 2. Up to 270m groundwater pressure was predicted. In addition the alignment crossed active geological fault zones; 5 identified zones on the East Tunnel and 17 on the West Tunnel along with many smaller unidentified zones. The USFS also had review and oversight of the construction to ensure compliance with the groundwater inflow limits specified in the USFS Special Use Permit.

In addition to variable ground and high groundwater pressures there were two major fires in 2003 and 2007 in the Arrowhead area that shut down the site for days and following the fires in December 2003 a landslide caused by heavy rains dammed a watercourse which when overtopped flooded the Waterman Portal, see Figure 3, then fully submerged the TBM and shut down the job for four months.

As of the end of April 2008 the Arrowhead Tunnels are 96% complete; the East Tunnel is fully bored and the West Tunnel is currently 91% bored. Over 6,130 tonnes of cement grout has been injected for water control and ground improvement, this has allowed the works to comply with the stakeholder requirements. The tunnel production rates have increased year on year throughout the job as more was learnt about the ground conditions. The final 650m of tunnel left to bore has highly variable geology, with decreasing cover, groundwater issues and is on a 90 degree curve.

Water Inflows and Pre-excavation Grouting

Groundwater flows and pressures if not controlled could be substantial and can quickly curtail mining operations as shown in Figure 4. In practice up to 70% of the time spent in the tunnel was spent grouting, as to mine, the groundwater must be controlled. In good ground the probe length was approximately 150 feet (45.7m), in order to mine 130 feet (39.6m) and leave not less than a 20 feet (6.1m) grouted buffer. This buffer generated a hydraulic gradient that reduced the effective water pressure head acting on the cutterhead of the TBM.

In poor ground, the solution described above was revisited depending on the ground conditions encountered. Given the close proximity to the San Andreas and other faults, voids were often found and there was little consistency in the ground mass. It was often found to be highly altered. The drilling and grouting would be revisited with the solution found to be to combine drain holes for pressure relief, to decrease the pressure gradient, with pre-excavation grouting with between 6 and 17 splayed holes from the face, in small cycles, to improve both grouting and mining conditions. Overall a lot of grout went into the ground, to improve both the rock mass and to control water; grouting occupied a huge amount of the cycle time.

BASF assisted with field testing of their MP320T colloidal silica grout; designed to penetrate tight strata to control water inflows that could not be stopped with micro-fine cement grout. This grout added very little shear strength to the ground, but improved the impermeability. As a comparison, while the micro-fine grout was approximately 0.1mm diameter; the colloidal silica grout is 1000 times smaller. It was able to penetrate the ground, virtually like water, but could also be pumped with the same equipment as for the micro-fine grout.

A concept described in the Geotechnical Baseline Report (GBR) was to drill drain holes to effectively pull water and water pressure from the face of the TBM. This could be achieved by drilling through the segments of the tunnel to bring water in behind the TBM. This had merit in some cases, to relieve local pressures and to improve mining conditions, but if too much water were drained from the ground this would have been outside the requirements set in the USFS Special Permit.

TBM Design and Squeezing Ground

Both 5.8m diameter hybrid hard rock TBMs were built by Herrenknecht; these featured cutterheads that could articulate up to 6”, a shroud on the trailing edge of the cutterhead for improved ground control, two percussion drill decks, a casing advancing system for drilling probe and grout holes in poor ground conditions, screw auger wear plate improvements, an improved one-of-a-kind slurry handling and separation system, a vacuum segment erector (the first of its kind in the US) and a data-logger for ground assessment while drilling, see Figure 5. The TBMs were designed for up to 10 bar of groundwater pressure in a static mode, less than the full groundwater pressure.

One of the biggest improvements on the job was the data logging on the probe drill holes, which provided a ‘crystal ball’ view of what was ahead. But even still the TBM got caught in squeezing ground two times while driving the East Tunnel. This meant having to hand mine over the shield two times to free the TBM, see Figure 6, which was successfully completed on each occasion.

The Lining

The 4,880mm internal diameter segmental lining had a thickness of 330mm which was thick for this diameter, but this was due to the design groundwater pressure. Due to the squeezing ground up to four times the original design thrust was applied to the segments, so the thicker lining meant this was also possible. The single Phoenix (profile M 385-73) gaskets were tested to 40 bar pressure, although were designed for 27 bar pressure. The primary support of the segmental lining will have a final 12 foot (3,658mm) diameter RCCP tunnel lining installed inside, with the annulus backfilled with grout.


Mike McCall (retired, formerly Balfour Beatty) asked about the method of payment related to minimising inflows to the set requirements and secondly who paid for all the grout. Brian Fulcher noted that there was no penalty for not achieving the inflow rates. Mike Bell added that not all the grout that went into the ground had been paid for, there were some negotiations between the client and the contractor in 2006, but the commercial terms remain confidential, there is now a partnership between the client and contractor though.

Tim Healey (Capita Symonds) asked with the close proximity to the San Andreas Fault what allowance had been made for earthquake risk in the seismic design of the tunnel. Brian Fulcher remarked that if there were a rupture of the feed the water would be captured in distilling basins located at the tunnel portals. Mike Bell also noted that the lining pipes were formed of 10” (254mm) concrete, 0.5” (13mm) steel with then a grouted annulus to the segmental lining.

Phil Richardson (Natural Cement) asked whether there were any environmental concerns over grouting noting that San Manuel Indian Tribe were selling bottled extracted water. Mike Bell noted that both the San Manuel Indian Tribe and USFS had to approve all grouts before they were used.

Shani Wallis (Journalist) asked how much original SI had been allowed compared to the SI for the second contract and secondly how accurate was the GBR found. Brian Fulcher remarked that the SI for the east tunnel was comprehensive, but this was for a slightly different alignment, the SI was extended for the new alignment, but was not as comprehensive. Mike Bell added that the GBR did quite a good job of describing the conditions, it picked up the big faults, but it was the small unknown faults that caused a lot of the problems. The GBR did warn of the groundwater inflows.

Neville Harrison (Mott MacDonald) asked what proportion of time was used for drilling and grouting and secondly what production rates were achieved. Brian Fulcher reported that production rates were highly variable, on a good day 25m per day, on a bad day all day could be spent probing and grouting. Mike Bell added that on a good day ring excavation could occur in 30-35min, with rings built in 20-25min. Production rates were very hard to predict, but the probe drilling and data logging proved extremely useful.

Colin McKenzie (semi-retired) asked what rock strengths were being dealt with as he was surprised that very little cutter wear had been reported. Brian Fulcher agreed that there was very little cutter wear in both tunnels. The maximum rock strength was around 20,000psi (140MPa).

David Sharrocks (London Bridge Associates) asked what the build-up of the tunnelling crew was. Brian Fulcher quickly responded that he did not think this was a question that a contractor should answer. Paul Hoyland then thanked both speakers for an excellent honest presentation and wished the project success for the final few hundred metres of the drive.

Rapporteur: Nathan Wilmot

Fig1 – Location of the Inland Feeder Project Fig 2 – Geological profile of tunnels Flooding of the Waterman Portal Water inflows in the tunnels Hand mining over the shield