A tunnel project manager runs a job whose schedule is set by ground nobody has seen and whose money is set by a geotechnical baseline report. Hire for fluency in that baseline, discipline about differing site condition notices, and time on the excavation method you will use. The size of the last job matters less.
Picture a tunnel boring machine 3,000 feet into a sewer drive, stopped at two in the morning. Cutterhead torque has climbed all shift, the muck on the belt has turned from silty sand to something with cobbles in it, and the shift engineer has written one line in the log: possible boulders. The scene is invented. The hour that follows is familiar to anyone who has run a soft-ground drive, because someone now has to decide whether this ground is what the geotechnical baseline report described, and whether written notice goes to the owner before anyone disturbs it. That someone is the tunnel project manager. Get the call right and the delay is priced under the owner’s contract. Get it wrong and it comes out of your margin.
Many contractors staff a tunnel from the same bench as the rest of their heavy civil work, on the theory that a good builder is a good builder. The argument here runs the other way. Underground, the ground is the schedule and the baseline report is the contract, and the person who can read both is a narrower hire than the title suggests.
What does a tunnel project manager own that a surface job never asks for?
A tunnel project manager owns the link between ground and contract. That covers the geotechnical baseline report the bid relied on, the differing site conditions clause that pays when the ground departs from it, the excavation method’s advance rate, and a separate federal safety standard for underground work.
The Federal Highway Administration’s road tunnel manual states the problem in one line: “Most claims in tunnel construction are related to unforeseen ground conditions.” The US industry answered with the differing site condition clause and with three geotechnical documents: a data report of borings and lab results, design memoranda that stay inside the design team, and the baseline report, which turns interpretation into contract statements bidders price against and takes precedence over the other two.
A surface project manager can build a long career without opening any of those documents. A tunnel project manager opens the baseline before the estimate is final and keeps it open until final payment.
The safety side is just as specific. OSHA’s underground construction standard requires a count of everyone below ground, air monitoring, a gassy-operation classification and rescue teams: with 25 or more people underground at one time, two five-person teams, one within half an hour and the other within two hours. Changing cutting tools on a pressurized-face machine can mean sending workers into compressed air, and the compressed air rules add a medical lock and a retained physician experienced in decompression illness. A rescue arrangement or intervention plan that is not ready stops the drive.
All of that is manageable on paper. The ground is the part that is not.
Why does the ground set a tunnel’s schedule?
On a bored tunnel the machine’s advance rate is the critical path, and the ground sets the advance rate. Boulders, mixed faces, water and stiffer clay slow the machine and add stoppages. Every foot lost at the face moves the finish date, because most downstream work waits on the drive.
FHWA’s tunnels program says that “subsurface conditions significantly influence risk and uncertainty, which raises the costs.” The arithmetic is blunt. Take a single 12,000-foot drive on a sewer storage tunnel. If the ground matches the baseline and the machine averages 60 feet a working day, the drive takes 200 working days. If a long mixed-face reach pulls the average down to 40 feet a day, it takes 300. If a boulder zone forces repeated stops and the average falls to 25 feet a day, it takes 480, more than double the plan, with the crew, the machine and the site overhead running the whole time. The figures are our illustration, not a study.
That is why the monthly forecast on a tunnel is a different document from the one on a surface job. Ask a tunnel site team what is still outstanding that has to go into the forecast, and the honest answer is usually that they do not know yet. The ground that will cost the money is still ahead of the cutterhead and has not announced itself. So the number gets built from what was logged behind the machine, and the part that would change it is still in the ground.
Working harder does not fix that. Forecasting against the baseline does, foot by foot, so the gap between what the contract assumed and what the face is showing gets measured every shift instead of discovered at month end. Which raises the question of what, exactly, the baseline assumed.
How should a tunnel project manager read a geotechnical baseline report?
The baseline report is a contract, a risk-allocation instrument stating in definite numbers the ground the owner agreed to treat as expected. Ground materially worse is normally the owner’s risk if the notice steps are met; ground within the baseline belongs to the contractor.
The FHWA manual is explicit that the baseline is neither a warranty nor a prediction of actual conditions. It is written in definite language, “is” and “will” rather than “may” and “should”, and kept to about 30 pages for a typical project and 50 for a complex one, so it can be read in one sitting. The manual’s own example shows how much rides on one number. Limited borings suggest hard rock will make up 10 to 30 percent of a tunnel’s length. An owner chasing lower bids sets the baseline near the low end and carries more risk of extras. An owner who wants fewer extras sets it near the high end and pays for that in the bid.
This is where a generalist’s resume meets the baseline report, and the meeting is short. The resume lists an interchange, two bridges and a program value with a lot of zeros. The baseline report says how many boulders above a stated size the contractor should expect along each reach of the drive. Only one of those documents will be on the table when the dispute review board sits down.
A candidate worth hiring can tell you which baselines a past bid leaned on hardest, how the field team measured against them, and what happened the first time the face disagreed.
What does the differing site conditions clause ask of a tunnel project manager?
On federal work the differing site conditions clause requires written notice promptly and before the conditions are disturbed. The project manager has to spot a departure from the baseline at the face, send notice in time and prove the difference from records kept every shift.
The federal text is FAR 52.236-2. It covers subsurface or latent conditions that differ materially from those the contract indicates, and unknown conditions of an unusual nature that differ from what that kind of work ordinarily meets. It bars an adjustment requested after final payment. Its companion, FAR 52.236-3, has the contractor acknowledge that it investigated the site and satisfied itself about conditions affecting the work, and an owner can cite it against a claim. Federal-aid highway contracts carry standardized changed-condition clauses, including differing site conditions, under 23 CFR 635.109. Sewer authorities and transit agencies write their own versions, and the manager needs to know which one governs before the machine launches. Selah Talent Partners is a search firm and not a law firm, and nothing here is legal advice.
What the clause asks for in practice is evidence. On a machine drive that means the data the machine already records (thrust, torque, penetration, face pressure, muck weight against the theoretical) tied to chainage and compared with the baseline every shift. On a sequential excavation heading it means face maps and the support class installed against the class the design predicted. A manager who sets that system up before launch can file a notice in hours with the evidence attached. One who builds it after the first disagreement is arguing about ground that has already gone up the muck conveyor.
Many tunnel contracts add a dispute review board of three underground practitioners who follow the job and give opinions on disagreements as they arise, plus escrowed bid documents that show what the bid assumed. Which records a board will weigh depends on how the tunnel is being dug.
Which tunnel method is the hire for?
Hire for the method on the job. A manager from pressurized-face soft-ground drives, one from sequential excavation and one from hard-rock work read different ground, document it differently and fail in different ways. Treat the tunnel type as part of the job title, and test for it directly.
| Tunnel type | Typical owner | What the manager must defend | Interview question |
|---|---|---|---|
| Pressurized-face machine in soft ground | Sewer authorities on overflow storage tunnels; transit agencies on running tunnels | Face pressure, settlement under buildings, boulder baselines, hyperbaric interventions | ”Walk me through the last time face pressure and surface settlement disagreed.” |
| Hard-rock machine or drill-and-blast | Water utilities on conveyance tunnels | Rock classes against the baseline split, groundwater inflow, support by class | ”How did the actual rock split compare with the baseline you bid?” |
| Sequential excavation | Transit agencies on stations and cross passages | Support class installed against class predicted, face mapping, deformation monitoring | ”Who decided the support class at the face, and how was it recorded?” |
| Microtunneling and pipe jacking | Highway, rail and utility owners on crossings | Jacking forces, obstruction baselines, recovery of a stuck machine | ”What happens when a jacked machine stops under a live highway?” |
| Cut-and-cover | Transit agencies on stations and approaches | Excavation support, groundwater control, adjacent structures | ”How did the excavation support change once you saw the groundwater?” |
Four pieces of evidence separate a tunnel manager from one who has worked near tunnels:
- A differing site condition notice they wrote: what they saw, when, and how fast the notice went out.
- A monthly forecast from a drive, and how it carried ground not yet mined.
- A baseline they disagreed with at bid time, and what they did about it before signing.
- An emergency and rescue plan they owned under the underground standard.
The general method for getting evidence instead of a rehearsed answer is in our note on interview questions. Finding someone who can produce all four is the harder part, because most of the people who can are mid-drive for another contractor.
Tunnel programs run for decades, and their managers are already booked
Tunnel work arrives in public programs that run for decades, and the people able to manage it are a small, employed group. EPA counts combined sewer overflows as a major water pollution and public health concern for approximately 700 communities in the United States. The Clean Water Act requires those communities to put controls in place. Large storage tunnels are one of them. The 2022 needs survey put total reported clean water needs at $630.1 billion over 20 years, with $36.5 billion of it for combined sewer overflow correction. Transit tunnels draw on FTA’s Capital Investment Grants program, which funds major transit capital projects.
Heavy and civil engineering construction employed 1,210.8 thousand people in September 2026, up from 1,184.5 thousand a year earlier, in BLS employment data. Seasonally adjusted construction job openings stood at 251 thousand in August 2026, against 213 thousand a year before, per JOLTS. Only a sliver of that workforce has run a machine drive. The occupational wage data put the median construction manager at $114,990 in May 2025 and the upper quartile at $151,640, and civil engineers sit at a median of $100,840. Neither is a tunnel figure. Both describe the wider pools tunnel managers come out of, and the BLS handbook page describes the general role.
If you post the role, you tell an owner that your key person on a live drive may be changing, and you tell the incumbent before you have decided anything. A storage tunnel in Atlanta and a drainage tunnel in Dallas would draw on the same national pool, most of it employed on someone else’s drive. Qualifications-based tunnel pursuits can score the named manager before award, which puts the proposal staffing problem in front of the hire. Selah Talent Partners handles confidential executive search for mandates that cannot be publicly posted, for contractors across the United States. Our note on confidential search explains how the approach works, and the senior project manager page covers the general seat. Before any of that, check whether the job needs a tunnel specialist at all.
Which underground jobs do not need a tunnel specialist?
Open-cut sewer and water mains, and short trenchless crossings that a specialty subcontractor runs under your coordination. A strong utility or site project manager covers those. A tunnel specialist is worth the search when the drive and its ground sit on your critical path.
Two other cases fall outside this hire. An agency hiring its own resident engineer is recruiting from the owner side, a related pool with different incentives. A contractor with no tunnel pursuit in view gains little from hiring ahead of the work: the specialist ends up on surface jobs and leaves for the next machine that launches elsewhere.
What does the right candidate do when the machine stops?
Before the crew mines another ring, they check the shift log against the baseline, decide whether the ground differs materially, and get written notice out while the evidence is still in place. Then they re-forecast the drive. A candidate who describes that sequence from memory, with dates, has done it.
The interview question that settles the hire is the opening scene. Describe the stopped machine, the changed muck and the one-line log entry, then ask what happens in the next hour. A generalist talks about a schedule recovery plan. A tunnel project manager asks which baseline the ground is being compared against, who has the shift data, and when the notice has to go.
If nobody on your bench gives that answer, our note on the heavy civil bench covers where the search goes next.
The baseline was written long before launch, from widely spaced borings. The ground between them is already there, waiting for the cutterhead.