Richard McLane and James Corcoran, members UCA, are project manager and project engineer, Traylor Bros., Inc.; Matt Neuner is senior geochemist, Golder Associates, now with Ecometrix Incorporate; Hugh Davies, member UCA, is project manager, Golder Associates, now with Newmont Corporation, and Joseph DeMello, member UCA, is deputy executive officer, Los Angeles County Metropolitan Transportation Authority, email email@example.com.
The highly anticipated Purple Line Extension (formerly Westside Extension, renamed D Line Extension in 2020) is an underground heavy rail segment of the Los Angeles County Metropolitan Transportation Authority (LA Metro) rail transit system being constructed to extend a high-speed link from downtown to the busy Westside of Los Angeles. The first of three sections, Section 1, is a $3.12 billion design-build project, connecting an existing station at Wilshire Blvd. and Western Ave. in the city of Los Angeles, extending 6.3 km (3.92 miles) to the west under Wilshire Blvd., and terminating approximately 167 m (550 ft) west of Wilshire and La Cienega Blvds in the city of Beverly Hills, CA. It was considered the backbone route for LA Metro’s rapid transit network as early as 1961 (Kaiser, 1961), but construction was delayed for decades due to challenges with difficult ground conditions and
The highly anticipated Purple Line Extension (formerly Westside Extension, renamed D Line Extension in 2020) is an underground heavy rail segment of the Los Angeles County Metropolitan Transportation Authority (LA Metro) rail transit system being constructed to extend a high-speed link from downtown to the busy Westside of Los Angeles. The first of three sections, Section 1, is a $3.12 billion design-build project, connecting an existing station at Wilshire Blvd. and Western Ave. in the city of Los Angeles, extending 6.3 km (3.92 miles) to the west under Wilshire Blvd., and terminating approximately 167 m (550 ft) west of Wilshire and La Cienega Blvds in the city of Beverly Hills, CA. It was considered the backbone route for LA Metro’s rapid transit network as early as 1961 (Kaiser, 1961), but construction was delayed for decades due to challenges with difficult ground conditions and other factors (city of Los Angeles, 1985; Cobarrubias, 1992).
Los Angeles was built over oil fields that comprise what was one of the world’s most productive basins in the early 20th century, the Los Angeles Basin. Long before development of the city, heavy oil seeping from the ground at what is now called the La Brea Tar Pits was used and traded by the Chumash indigenous people (Bilodeau et al., 2007). Drilling in the area began around the turn of the 20th century, and by the 1930s hundreds of wells had been completed in the Salt Lake Oil Field and the South Salt Lake Oil Field (Fig. 1; Crowder, 1961, Crowder and Johnson, 1963). During this period, urban areas expanded west from downtown and developers such as Gaylord Wilshire, J. Harvey McCarthy and Moses Sherman had started developing the Wilshire corridor; McCarthy had envisioned a subway under Wilshire Blvd. as early as the 1920s. After initial planning in the 1960s of a subway line along the Wilshire corridor, an extensive geotechnical program was completed in the late 1970s and early 1980s (CWDD, 1981; Kaiser and Gage-Babcock, 1983; Proctor, 1985).
On March 24, 1985, a methane (CH4) seep led to an explosion and fire at the Ross Dress for Less department store several blocks north of the La Brea Tar Pits (Fig. 1). While there was some controversy as to the source of the gas initially, the city of Los Angeles designated a methane high potential risk zone (Fig. 2; City of Los Angeles, 1985; Cobarrubias, 1992) that corresponded generally with mapped extents of the Salt Lake Oil Field, the 6th Street Fault, and the 3rd Street Fault (Fig. 1). In 1985, a moratorium was placed on federal funding for tunneling through the methane potential risk zones in Los Angeles by congressional order (Section 321 of U.S. Congress Public Law 99-190, 1985) and by resolution of Metro’s predecessor transit agency. Endres et al. (1991), Hamilton and Meehan (1992), and Chilingar and Endres (2005) concluded that the source of the gas that caused the explosion was the reinjection of gas and saltwater brine back into the underlying oil field along with the faulting in the area.
After the explosion, the Red Line (renamed B Line in 2020) was rerouted to the east and north of the methane potential risk zone in the Fairfax District (Fig. 2; SCRTD, 1987). LA Metro began operating the Red Line, the city’s first rapid transit line, in 1993.
In the 1990s, geotechnical investigations were completed for a reroute of the Western extension subway line south of the methane potential risk zone associated with the Salt Lake Oil Field (Elioff et al., 1995; Enviro-Rail, 1996; Metro, 1996). Following identification of gas in an unsaturated zone several blocks south of Wilshire Blvd. (Pico and San Vicente Blvds) with concentrations of hydrogen sulfide (H2S) gas as high as 20,000 parts per million (ppm), 200 times greater than the amount immediately hazardous to life or health (IDLH), construction was again delayed, due to funding constraints.
In 2005, a panel of experts engaged by the American Public Transportation Association (APTA, 2005) concluded, after cursory review, that tunneling along the Wilshire Blvd. corridor could be done safely due to advances in technologies for gas detection and pressure face tunnel boring, and experience tunneling in Los Angeles. Successful completion of the Gold Line (renamed L Line) Extension in East Los Angeles, constructed by Traylor Frontier Kemper JV, was cited as a basis of current tunneling technology. The congressional moratorium was repealed in 2007 (H.R.2764 Section 169 of U.S. Congress Public Law 110-161, U.S. Congress, 2007). LA Metro carried out additional geotechnical studies from 2009 to 2014, and a second APTA peer-reviewed report (APTA, 2012) concluded that tunneling using pressure face technology (EPB or slurry TBM) past the La Brea Tar Pits could be done safely. During these investigations, pockets of pressurized gas were identified (LA Metro, 2011) and gas in an unsaturated, sandy confined layer with approximately 95 percent by volume CH4 and 6,500 ppm H2S was identified at a single well (M-13; Fig. 1) along the alignment west of the La Brea Tar Pits. When the Geotechnical Baseline Report (GBR; Metro, 2014) was issued, the extent of the gas in this sandy confined layer, which would become known as the Crescent Heights Gas Zone (Fig. 1) was not well defined. In 2014, Metro awarded the design-build contract to the Skanska-Traylor-Shea Joint Venture (STS). STS hired the geotechnical consultants Golder Associates to further investigate risks associated with CH4 and H2S gases at this location as well as in the Fairfax Station area to the east, adjacent to the tar pits.
While tunneling through gassy ground is not new, particularly in California where detailed tunneling safety standards have been in place for decades, relatively few case studies of tunneling through gassy ground have been published. Proctor (2002) described the fatal San Fernando Tunnel explosion in June 1971, which occurred during tunneling in gassy ground associated with oil fields in northern Los Angeles (Sylmar Tunnel Explosion). This incident led to strengthening of the tunnel safety standards in California, which now form the state’s Tunnel Safety Orders (Cal/OSHA Title 8 subchapter 20, Tunnel Safety Orders). This incident and 10 others involving various tunnel types in eight countries are summarized by Copur et al. (2011). Most of these incidents involved tunneling through or near formations with oil and natural gas and inadequate ventilation, gas detection, automation and/or spark prevention.
For the Westside Purple Line Extension Section 1 (WPLE1), LA Metro and its consultants formed a proactive partnership with STS and Golder that facilitated building on the knowledge gained from decades of investigations while working collaboratively to develop a safe approach to tunneling past the La Brea Tar Pits and through the Crescent Heights Gas Zone. The main objectives of these efforts were to assess the risks and develop mitigations around two potential scenarios:
Ingress of explosive and acutely toxic gases (CH4 and H2S) into areas of the tunnel boring machines (TBMs) where workers were present and where ventilation exited the TBMs and would be emitted to the ambient air.
Migration of explosive and acutely toxic gases (CH4 and H2S) due to normal operation of EPB TBMs, from the pressure face to adjacent building basements or utility vaults — effectively inducing another Ross Dress for Less explosion by the TBM.
This article summarizes activities completed by the WPLE1 team over five years leading up to and including successfully and safely tunneling through exceptionally gassy ground.
The WPLE1 team carried out investigations from 2015 to 2019 to characterize the Crescent Heights Gas Zone, study potential gas sources and connections, develop the special mitigations of in situ oxidation and gas extraction, and test these mitigations in advance of tunneling. Prior to these investigations, several key uncertainties remained:
The extents of the gas zone around the M-13 well were not well defined, such that possibilities ranged from it potentially being a small gas pocket around the well (Getty 49) to it possibly being a large gas zone connected to the underlying oil fields.
It wasn’t clear whether there might be other gas zones like the one at the M-13 well that were possibly between the previous drill holes.
If the gas zone at the M-13 well was extensive, it would not be feasible to treat H2S at 6,500 ppm entering the ventilation system at the planned rates of excavation.
The implications of the potential for flowing gas from an extensive gas zone on both gas release inside the TBM and gas migration from the TBM were not well understood and presented a greater potential challenge than smaller gas bubbles that would release proportionally to the TBM advance rate.
While it was expected that the pressure face at the EPB TBM cutterhead would limit gas ingress to the working area inside each TBM (due to over- pressure), it was unclear whether the face pressure or injected foam required to maintain the pressure and condition the muck could cause gas to migrate to a building basement or utility vault.
The WPLE1 team proactively investigated each of these uncertainties prior to tunneling.
Potential gas source and connections
The extents of the gas zone around the M-13 observation well were likely related to the source of the CH4 and H2S gases. The WPLE1 team investigated the potential for connection of the gas zone to the underlying oil fields, with potential routes of connection including historical oil wells, exsolution from the tar, and gas flow up an adjacent fault.
Oil wells. Historical records (CalGEM Well Finder, formerly DOGGR, API 03715144) indicate that an oil well was drilled in 1907 near the intersection of Wilshire Blvd. and S. Crescent Heights Blvd. by a predecessor of the Getty Oil Co. (Arcturus Oil Co., Fig. 1). Referred to as the Getty-49 (Chevron 03715144) well, it was drilled into the South Salt Lake Oil Field and was immediately abandoned (stopped work) that same year with final plugging in 1913; the well record indicates abandonment involved explosives at 10 depths, pulling the three casings and filling [the] hole. The Getty-49 well was drilled into the upper limb of the main oil-bearing unit of the South Salt Lake Oil Field, such that it could potentially form a direct connection for gas migration from the oil field to the Crescent Heights Gas Zone (Fig. 3). The WPLE1 team attempted to locate the Getty-49 well using various methods from the ground surface and with magnetometer surveys from both vertical and horizontal borings drilled at the depth of the tunnel alignments. A magnetic anomaly was found near the intersection, which may have been the Getty-49 well. The alignments of the tunnels were modified so that the TBMs would tunnel to the south of the anomaly (Fig. 1) and avoid it by approximately 3 m (10 ft).
Several oil wells were drilled directionally from 1970 to 1974 from the Packard Drill Site into the upper limb of the South Salt Lake Oil Field (Fig. 3; Samuelian, 1990; CalGEM Well Finder). These wells were produced mainly from the 1970s to 1990s, and some of them remain active. From 1999 to early 2003, 1.2 billion cubic feet of gas was injected at an average pressure of 575 psi (40 bar) into one of these wells (P-70) to enhance oil production. It is possible that activities at these wells might have contributed to the formation of the Crescent Heights Gas Zone: for example, potentially by pressurizing the production zone in the oil field combined with migration up the Getty-49 well or a fault. Gas injection to P-70 and P-65 was suspended in early 2003 because gas seepage was discovered at the surface at the locations of old wells 0.5 mile southeast of the intersection of Wilshire Blvd. and South Crescent Heights Blvd. (Fig. 1; Chilingar and Endres 2005) and tracer gas (perfluorodimethylcyclobutane and perfluoromethylcyclcohexane) that was injected was not detected — essentially the injected gas migrated out of the oilfield (CalGEMs Well Finder, API 03721161 p. 81).
The WPLE1 team analyzed gas samples collected from shallow gas wells in the Crescent Heights Gas Zone (that is, tunnel alignment depth) for isotopes of carbon (tm13C in CH4, ethane [C2H6], propane [C3H8], and CO2) to aid with interpretation of gas sources. These analyses clearly identified the CH4 gas in the Crescent Heights Gas Zone as thermogenic gas: that is, the type of natural gas present in an oil field. The isotopic signature of the light hydrocarbon gases in the Crescent Heights Gas Zone was similar to gases reported by Jeffrey et al. (1991) that were sampled from the underlying oil fields (from oil wells shown in Fig. 3) and from the gas zone that caused the 1985 Ross Dress for Less explosion (Fig. 4). Further, the ratios of straight-chain n-butane (n-C4) to branched iso-butane (iso-C4) in gas samples from the Crescent Heights Gas Zone were indicative of nonbiodegraded thermogenic gases, similar to the n-C4 to iso-C4 ratios in samples from the South Salt Lake Oil Field (Jeffrey et al., 1991) and from the most active gas vent at the La Brea Tar Pits (Etiope et al., 2017).
Note the WPLE team did test for the tracer gases from P-65 and P-70 and did not detect either.
The WPLE1 team investigated the source of H2S in the Crescent Heights Gas Zone using isotopes of sulfur (tm34S) in H2S gas, in sulfate in groundwater, in tar, and in bulk soil samples. The team was able to rule out bacterial reduction of sulfate from anhydrite and thermochemical reduction of sulfur from anhydrite or organosulfur compounds in the tar. But the exact source of the H2S wasn’t determined, in part because information on H2S in the oil fields was not available. The source was interpreted to likely be due to biodegradation of the tar with associated release and reduction of sulfur. But other potential sources that could not be ruled out were similar biodegradation of oil in the oil field and mantle gases (Jung et al. 2015).
6th St. Fault. The oil, tar, and gas seeping at the ground surface at the La Brea Tar Pits reportedly migrate from the Salt Lake Oil Field along the 6th Street Fault (Hamilton and Meehan, 1992; Khilyuk et al., 2000; Chilingar and Endres, 2005; Bilodeau et al., 2007; Etiope et al., 2017). The 6th Street Fault (Fig. 3) is a seismically inactive fault that dips steeply to the northeast and forms a structural trap for oil and gas in the Salt Lake Oil Field (Lang and Dreessen, 1975; DOGGR, 1992). Studies of regional tectonics (Shaw and Quinn, 1986; Wright, 1987) indicate that the 6th Street Fault was likely seismically active until the late Pleistocene, after which time it became inactive and was concealed by deposition of younger alluvium. Continuous seepage of oil, tar, and gas at the site now known as the La Brea Tar Pits was reported as early as 1792 (Bilodeau et al., 2007) and in a 1906 USGS report (Arnold, 1906), and the rate of seepage of CH4 from the site makes it one of the most active hydrocarbon seeps in North America (Etiope et al., 2017).
The WPLE1 team drilled 43 wells into the Crescent Heights Gas Zone, and consistently measured positive gas pressures (that is, greater than atmospheric) that decreased from north to south. These measurements, together with development of a calibrated multiphase flow model (discussed below), indicate that gas naturally flows from north to south through the gas zone. Since this natural flow of gas within the Crescent Heights Gas Zone appears to originate to the north of and beyond the likely area of the Getty-49 well, the WPLE1 team interpreted the likely pathway of gas migration to be from the Salt Lake Oil Field along the 6th Street Fault to the gas zone (Figs. 1 and 3). It is noted that the location of the 6th Street fault as shown and projected to surface is the subject of interpretation and open to debate.
Gas extraction trials for the mid-city project, early 1990s
During investigations in the early 1990s for the re-routed tunnel alignment south of the methane potential high risk zone associated with the Salt Lake Oil Field, LA Metro encountered a gas zone beneath the intersection of Pico Blvd. and San Vicente Blvd. This gas zone had CH4 concentrations greater than 99 percent by volume and H2S concentrations as high as 20,000 ppm in unsaturated sandy sediments of the San Pedro Formation confined by clayey sediments of the relatively impermeable Lakewood Formation (Metro, 1996). Metro tested a method of in situ oxidation of H2S by extracting gas continuously while introducing atmospheric air into the gas zone (Elioff et al. 1995). Within two days, H2S concentrations within a radius of 30 m (100 ft) were lowered to approximately 100 ppm. More than a month after the gas extraction and air introduction was terminated, H2S concentrations were 300 ppm at the extraction well and <1 ppm at the air introduction wells. However, CH4 concentrations recovered to greater than 99 percent vol. within days of turning off the system. Due to the similarity of this gas zone to the Crescent Heights Gas Zone, this experience was valuable in understanding an effective mitigation of H2S concentrations.
Gas extraction trials, 2018 to 2019
The WPLE1 team carried out gas extraction trials in 2017, 2018 and 2019 to better characterize the Crescent Heights Gas Zone and to trial in situ oxidation and gas extraction systems prior to tunneling. A summary of the gas extraction trials is provided in Table 1.
Drilling and testing during the trials provided detailed characterization of the geologic conditions of the Crescent Heights Gas Zone under Wilshire Blvd., with 31 wells drilled by the end of the third phase of testing. Predominantly fine-grained sediments of the Lakewood Formation, typically dominated by silt and clay, were encountered from surface to a depth of approximately 18 m (60 ft) (Fig. 5). The contact between the Lakewood Formation and the top of the San Pedro Formation typically consisted of clayey ground underlain by sand or silty sand. The upper 1.5 to 6 m (5 to 20 ft) of the San Pedro Formation was coarse grained, unsaturated, forming the gas zone. Water-saturated sands formed the bottom of the central and western portions of the gas zone, and tar-saturated sands formed the bottom of the eastern portion of the gas zone (Fig. 5).
These drilling programs revealed that the Crescent Heights Gas Zone extended approximately 182 m (600 ft) from east to west and coincided with the depth of the tunnel alignments (Fig. 5). The north to south extents of the gas zone were confirmed to be at least 128 m (420 ft) but were thought to potentially extend approximately 457 m (1,500 ft) or more (Fig. 1).
Gas extraction Trial 1 (Phase 4.1) was carried out with two days of gas extraction using wells that were installed during baseline geotechnical programs and the initial gas investigation. Approximately 117,000 ft3 (Mcf) of gas were extracted from the M-13 observation well while 8 Mcf of air was passively introduced by opening two nearby gas probes once suction was confirmed. The CH4 concentrations in the extracted gas remained high (average of 91 percent vol.) during the testing, but the H2S concentrations decreased from 6,000 ppm to less than 1 ppm after extraction. This trial demonstrated that the gas zone was more extensive and that H2S could effectively be treated with in situ oxidation by passively introduced air.
Gas extraction Trial 2 (Phase 4.2) included installation of a network of six wells that were used for gas extraction and passive air introduction, and installation of three monitoring wells. Additional borings encountered water-saturated San Pedro sands, which identified the eastern and western boundaries of the gas zone (borings not shown on Fig. 5). Gas was extracted for six days (731 Mcf) by extracting from wells on one side of Wilshire Blvd. while allowing air to be passively introduced through wells on the other side of Wilshire Blvd. The configuration of gas extraction and passive air introduction was changed twice during the trial to maximize the footprint of in situ oxidation. Concentrations of H2S ranged from 88 to 8,000 ppm before the trial and were lowered to a maximum of 50 ppm in wells on Wilshire Blvd. 30 days after extraction was stopped, indicating successful in situ oxidation over a period relevant for tunneling. Concentrations of CH4, however, returned to levels near 88 to 95 percent vol. within 10 days after the end of extraction.
Gas extraction Trial 3 (Phase 4.3) was carried out over 28 days of extraction, with the goal of investigating whether a longer gas extraction and passive air introduction duration could result in complete depletion of CH4 and replacement with air. An additional 10 extraction/introduction wells and seven monitoring wells were installed with another drilling program. A total of 3,435 Mcf of gas was extracted over 28 days, while 1,489 Mcf of air was passively introduced. During gas extraction, suction was maintained throughout the Crescent Heights Gas Zone under the Wilshire Blvd/South Crescent Heights Blvd. intersection. Concentrations of H2S declined further and were approximately 2 ppm across much of the gas zone two weeks after the end of extraction. Concentrations of CH4, however, remained approximately 20 to 40 percent vol. during extraction. Within 10 days of the end of extraction, CH4 concentrations and gas pressures returned to baseline levels of approximately 95 percent vol. and positive pressure of up to 20 in. of equivalent water column. This trial confirmed that gas extraction and in situ oxidation could adequately mitigate the H2S hazard but not the CH4 hazard in advance of tunneling, and that the source was not a small pocket of gas, but a much larger, near-infinite source.
Planned mitigation, theory and expected outcome
Following the gas extraction trials, the WPLE1 team decided to mitigate the risks associated with tunneling through the Crescent Heights Gas Zone with the following measures:
- Ventilation within each TBM at a nominal rate of 100,000 cfm, which passed through specified activated carbon scrubbers to treat H2S.
- Gas detection systems, automated alarms and shut-off switches, and electrical equipment in accordance with Cal/OSHA requirements for gassy and extra-hazardous tunnels within each TBM.
- Gas extraction and passive air introduction in advance of tunneling using a network of wells to further remove H2S by in situ oxidation.
- Gas extraction during tunneling using a network of wells to apply suction and direct gas from the subsurface to systems to treat CH4 and H2S at surface — providing a guided preferential pathway to extraction wells.
- Monitoring of gas pressures in the ground within and above the gas zone with a network of vibrating wire piezometers (VWPs) and monitoring wells.
- Monitoring of CH4 and H2S concentrations in building basements and utility vaults adjacent to the gas zone using portable open-path infrared (Heath RMLD) and traditional gas detection equipment.
- Contingency plan prepared to respond to gas detections if needed, including triggers and an action plan for notification of responders, evacuation of affected area, supplemental ventilation systems, and other measures as might have been needed.
The WPLE1 team developed a conceptual model for how tunneling through the Crescent Heights Gas Zone could potentially result in gas migration and how gas extraction could mitigate this risk. As an EPB TBM advances, it applies mechanical stress (pressure) to the ground as thrust jacks push it forward. The mechanical stress is borne by grain-to-grain contacts, and soil deformation limits the influence of these forces to a radius on the order of tens of feet or less around the cutterhead, dependent on soil or rock strengths. Hydraulic pressure is applied to the ground around the cutterhead by the pressure bulkhead and the injection of fluids (water, surfactant, and air) to the excavation chamber and directly to the ground in front of the cutterhead.
Injection of water, surfactant, and compressed air to the excavation chamber results in a plastic muck with fluid pressure maintained at the desired levels, which are typically the hydrostatic pressure (that is, equivalent to a water column the height of the depth below ground surface) plus any load to support the face and a factor of safety. Planned depths of the bottom of the tunnels through the gas zone ranged from 20 to 23 m (65 to 78 ft, such that planned pressure to be maintained in the excavation chamber of each TBM was approximately 2 to 3 bar (800 to 1,200 in. H2O). Openings in the cutterhead allow the hydraulic pressure applied to the muck in the excavation chamber to also be applied to the ground at the cutting face. The equivalent pressure on both sides of the cutterhead define the EPB pressure, a highly successful mitigation to minimize ground settlement and groundwater inflows to a TBM. The EPB pressure also provides mitigation to ingress of gases to the TBM and tunnel (due to pressure gradient). But since the pressure maintained at the cutterhead (800 to 1,200 in. H2O) is much greater than the gas pressures in the Crescent Heights Gas Zone (up to 20 in. H2O), this pressure gradient causes increases in groundwater levels around the TBM that could potentially drive gas migration away from the TBM and potentially into receptors such as building basements or utility vaults. Additionally, compressed air injected for soil conditioning in the excavation chamber can release into the ground and potentially pressurize or displace ground gases, which can cause gas migration. Previously, the WPLE1 team had measured a zone of influence of 183 m (600 ft) when injecting compressed air in the chamber during a long stoppage.
Injection of foam directly to the ground in front of the cutterhead posed a greater risk of operation of the TBM potentially causing gas migration — thus creating another Ross Dress for Less explosion. Foam is injected to the ground through nozzles on the cutterhead to reduce permeability of the ground, which inhibits dissipation of the pressure in the chamber, and for conditioning of the muck so that it forms a viscous paste. Foam is produced in the TBM by mixing a surfactant (soap) into water and injecting compressed air; amounts of each of these can be varied to optimize performance of the foam. The amount of compressed air required, expressed as the foam expansion ratio (FER), is important for foam performance but also had potential to influence gas behavior in the ground. A necessary function of operating an EPB TBM, foam injection into the Crescent Heights Gas Zone posed a risk for gas migration.
A final potential mechanism for gas migration that could possibly be caused by tunneling was the mining of the clay directly overlying and capping the gas zone and/ or potential fracturing of clay overlying the gas zone. The WPLE1 team ranked the potential for this mechanism lower than other potential gas migration mechanisms due to the approximately 15 m (50 ft) of predominantly fine-grained sediments that would remain overlying the tunnels. However, the vertical thickness of fine-grained sediments between the tunnels and multilevel building basements ranged from approximately 1 to 5 m (3 to 16 ft) (Fig. 5).
The primary mitigation for H2S in the Crescent Heights Gas Zone was in situ oxidation treatment in advance of tunneling. The concept for in situ oxidation of H2S was that oxygen in air passively introduced to the ground during gas extraction would displace and rapidly oxidize H2S in the gas phase and then partition into the soil moisture and oxidize the aqueous sulfide dissolved in the soil moisture. The effect was long-lasting because as H2S flowed back into the gas zone with the CH4 and other natural gases that rapidly returned after gas extraction stopped, the dissolved oxygen in the soil moisture provided a reservoir of treatment capacity to continue oxidizing H2S.
The primary mitigation for CH4 in the Crescent Heights Gas Zone was gas extraction during tunneling to maintain suction in the gas zone and collect and treat/oxidize gases at surface. Based on the estimated gas permeabilities and transmissivities and measured radius of influence of gas extraction during the trials in 2018 and 2019, the WPLE1 team anticipated that suction could be applied across the gas zone with on the order of 10 gas extraction wells. To understand the potential interactions between the EPB TBMs and the gas zone, the WPLE1 team developed a three-dimensional (3D) multiphase flow model using the code PFLOTRAN (developed by the U.S. national laboratories; www.pflotran.org).
Modeling the Crescent Heights Gas Zone consisted of four steps: constructing a 3D geologic model using Leapfrog (www.seequent.com), constructing a 3D multiphase flow model in PFLOTRAN, calibrating the flow model to results from the gas extraction trials, and simulating EPB TBM tunneling to forecast gas flows around the TBMs. Key observations for setting up the flow model were that positive pressures existed in the gas zone, there is a north to south gas pressure gradient, and gas pressures returned to a steady state, positive levels soon after gas extraction stopped. To simulate these key observations, a constant gas source north of Wilshire Blvd. was required in the model, as shown in Fig. 6. Gas was simulated in PFLOTRAN to be flowing from the gas source, which represents natural gas flowing from the underlying oil fields north to south through the unsaturated San Pedro sands (that is, the gas zone), and then upward as a broad dispersed flow through the overlying fine-grained sediments of the Lakewood Formation to emit into the deep building basements and at the ground surface. Hence, gas migration to receptors was already occurring naturally at rates that were being managed with sufficient barrier walls and ventilation systems in many of the basements. Gas permeabilities were adopted from well test analyses done using the reservoir engineering software SAPHIR (Kappa Engineering), and the PFLOTRAN model was calibrated to the gas extraction rates, gas pressure responses, and groundwater level changes measured during the gas extraction trials in 2019 and baseline measurements of CH4 concentrations in basements. It should also be noted that no building owners reported issues with gas intrusion prior to tunneling.
To simulate interaction of the TBMs and tunnels with gas extraction from the gas zone, the WPLE1 team modeled the advancing TBM with injection rates of air (up to 5 cfm) and water (up to 2 cfm) to represent foam injection at the cutterhead and a no-flow condition at the walls of the tunnels. Foam injection rates were based on laboratory testing that the WPLE1 team completed on samples of San Pedro sands from boreholes. Operation of the gas extraction system was simulated in the model with the planned gas extraction rates from the extraction wells (discussed in the next section). Findings from the modeling completed prior to tunneling are illustrated in Fig. 6 and summarized as follows:
- Gas flows naturally through the gas zone, and baseline CH4 flows into four building basements adjacent to the tunnel alignments in the gas zone were on the order of 20 kg/d.
- Suction could be maintained across much of the gas zone and particularly across the more permeable central portion of the gas zone, even close to each TBM as it advanced through these areas in the model.
- In more confined areas in the eastern and western portions of the gas zone after the first tunnel is constructed, gas pressures were predicted to be high (on the order of 100 in. H2O).
- Gas migration into one basement was predicted to be increased on the order of 40 kg/d for a period of several days due to operation of the TBM, likely a gas ingress rate that could easily be managed to safe levels by the existing ventilation systems in the building basements.
- After construction, the tunnels could partially impede the natural north-to-south gas flow, potentially resulting in higher gas flow to basements and utilities on the north side of Wilshire Blvd.
The expected outcome of tunneling through the Crescent Heights Gas Zone, based on over three years of investigations and considerable prior experience, was that gas extraction would effectively mitigate the risk of gas migration. The residual risk was characterized as a moderate-to-high probability that the TBMs would cause minor, acceptable, temporary, additional gas migration to building basements and there was a very low probability that tunneling would result in catastrophic gas migration.
Gas extraction system
The gas extraction system consisted of 11 gas extraction wells, lateral piping in backfilled trenches, and three modular gas extraction and treatment systems. Monitoring points consisted of nested VWPs at six locations, six monitoring wells, and pressure and flow gauges on each gas extraction well. The layout of the system is shown in Fig. 7.
The wells and VWPs were installed by Golder and Gregg Drilling, lateral piping was installed by Golder and Lonestar West, and the gas extraction systems were assembled and operated by Envent Corp. with input on well hydraulics from Golder. Extraction wells were 4-in. and 6-in. diameter PVC. Nine of the 11 extraction wells were fitted to 2-in. diameter PVC pipes placed in backfilled tranches that terminated in well vaults in the middle lane of Wilshire Blvd. or in the median of McCarthy Vista Blvd. (Fig. 7). Gas extraction equipment was set up in the central three lanes of Wilshire Blvd. and in the median and central two lanes of McCarthy Vista Blvd.
The gas extraction equipment consisted of three systems, each having a positive displacement blower and a variable frequency drive, a moisture knockout tank, a dual liquid caustic scrubber to treat H2S, and a thermal oxidizer to combust CH4 and volatile organic compounds located in the middle of Wilshire Blvd. Each system was operated continuously, day and night, from the start of extraction on May 24, 2020, until an unplanned shutdown on May 30, 2020. Extraction resumed on June 15, 2020, and continued for 15 days during construction of the north (BR) tunnel. Equipment was demobilized and then re-mobilized, and extraction resumed on July 21, 2020, for 18 days during construction of the south (BL) tunnel. Between construction of the two tunnels, four wells in the tunnel alignment were abandoned.
Prior to tunneling, 155 Mcf of air was passively introduced to the gas zone over 10 days. Concentrations of H2S under Wilshire Blvd. were less than 7 ppm after the air introduction. But to the north and south of the footprint of the in situ oxidation, H2S concentrations typically ranged from 1,000 to 5,000 ppm. During construction of the BR tunnel (Fig. 7), the total gas extraction rate averaged 104 cfm, and a total of 2,653 Mcf of gas was extracted. Gas pressures were maintained at between approximately –40 and –120 in. H2O in extraction wells, which maintained gas pressures at monitoring locations in the gas zone between approximately –20 and –90 in. H2O with exceptions discussed in the next section. These suction ranges targeted optimization of gas flow to extraction wells while intending to avoid induced groundwater level changes that could interfere with gas extraction. During construction of the BL tunnel (Fig. 7), the total gas extraction rate averaged 50 cfm, and a total of 1,204 Mcf of gas was extracted. Gas pressures were maintained at between approximately –25 and –120 in. H2O in extraction wells, which maintained gas pressures below atmospheric levels in the gas zone except close to the TBM and in confined portions of the gas zone (described in the next section).
TBM operation and measured conditions
The TBM referred to as the Purple TBM (BR – right side looking west) mined through the 256-m (840-ft)long gas zone (included buffer zones within water- or tar-saturated sand) from June 19 to 29, 2020, to construct the BR Tunnel (Fig. 7) at an average rate of approximately 22 m/d (72 ft/d). STS operated the TBM continuously on three shifts per day. Pressure responses to tunneling measured at two of the monitoring locations (VWP-4 and VWP-5) in the highly permeable portion of the gas zone adjacent to 6245 Wilshire Blvd. (Fig. 7) are presented in Fig. 8. Operation of the gas extraction system resulted in gas pressures of approximately –70 in. H2O and a 0.5 m (1.5 ft) increase in groundwater levels in the days prior to the TBM passing 6245 Wilshire Blvd. As the TBM passed 6245 Wilshire Blvd, gas pressures increased by 30 in. H2O, but suction was maintained in the gas zone.
Gas pressures were approximately 10 in. H2O greater when the Purple TBM mined through the more permeable central and western portions of the gas zone than when it mined through the less permeable eastern portion (Fig. 8). Groundwater levels rose as much as two feet, but this response was more localized to the TBM with a radius of approximately 30 m (100 ft) in front and behind the cutterhead based on responses at VWP-4. Gas pressures responded to the Purple TBM as much as 91 m (300 ft) away, because gas pressures increased in response to TBM operation across the entire central and western portions of the gas zone. This widespread but low magnitude pressure response correlated with an increase in the FER, or amount of air in the soil conditioners that was required to maintain target face pressures in the permeable sands (Fig. 8).
Methane concentrations varied from 0 to 90 percent vol. and H2S concentrations were typically less than 1 ppm under Wilshire Blvd. as the Purple TBM mined through the gas zone. The WPLE1 team monitored gas concentrations three times daily in the basements of 13 buildings and 18 utility vaults during tunneling through the gas zone. Concentrations of CH4 and H2S remained low in the basements during construction of the BR Tunnel, and there were no measured increases in the basements or utility vaults correlated with TBM operation.
The TBM referred to as the Red TBM mined through the gas zone from July 27 to Aug. 8, 2020, to construct the BL Tunnel (Fig. 7) at an average rate of 24 m/d (79 ft/d), again mining continuously with three shifts per day. Gas pressure responses to TBM operation in the east and west portions of the gas zone, which were confined by the newly constructed BR Tunnel and the boundaries of the gas zone, are presented in Fig. 9. While the TBM passed 6222 Wilshire Blvd, gas pressures increased by up to 675 in. H2O, with substantial pressure increase in both the partially tar-saturated sands in that area of the San Pedro Formation and the overlying clayey silts of the Lakewood Formation. When the TBM mined through the more permeable central and western portions of the gas zone, gas pressures increased by up to 170 in. H2O. Gas pressures remained high as the TBM mined past 6330 Wilshire Blvd. Similarly, as for the BR Tunnel, gas pressure increases correlated with the higher FER required for the BL Tunnel in the clean sands of the central and western portions of the gas zone (Fig. 9).
Concentrations of CH4 under Wilshire Blvd. were lower for the BL Tunnel, apparently due to compartmentalization/barrier of the gas zone by the newly constructed BR Tunnel and gas extraction. But as the Red TBM passed 6222 Wilshire Blvd. and 6300 Wilshire Blvd., minor ingressions of CH4 were measured at the lowest level of each of these building basements. Each of these CH4 ingressions lasted for less than one day, correlating with the passing of the TBM, and were limited to small areas at basement sumps (that is, detectable in covered sumps but not in the basement atmosphere). There was also one instance of ingression of CH4 into the TBM during tunneling through the gas zone, but it caused only minor delay (less than an hour) until concentrations lowered by ventilation and authorization was given to resume mining.
Potential risks associated with tunneling through gassy ground adjacent to the La Brea Tar Pits in Los Angeles were key factors leading to decades of delays during planning stages for a critical portion of the city’s rapid transit system. During baseline characterization and further investigations for Section 1 of the Purple Line Extension, an extensive gas zone with 95 percent vol. CH4 and up to 6,500 ppm H2S, referred to as the Crescent Heights Gas Zone, was identified west of the La Brea Tar Pits. This gas zone at the depth of the tunnel alignments may be connected to the underlying oil fields by a fault or a historical oil and gas well.
To mitigate the risk of explosive and acutely toxic gases entering the TBMs or migrating away from the pressure-face TBMs into adjacent buildings or utility vaults, the WPLE1 team adopted several mitigations. A ventilation system and gas detection system were operated inside each TBM, and a gas extraction system was operated outside the TBMs and paired with monitoring of in situ ground conditions and building basements and utility vaults adjacent to the gas zone. Air in the soil conditioning system was turned off.
The performance of the gas extraction system and tunneling through the exceptionally gassy ground west of the La Brea Tar Pits are summarized as follows:
- In situ oxidation by passive introduction of air during gas extraction successfully lowered H2S concentrations by a factor of 1,000 such that the TBMs mined through ground with safe levels of H2S, and H2S was not detected inside the TBMs or adjacent building basements.
- Operation of a gas extraction system successfully maintained suction as the TBMs mined through the more permeable portions of the gas zone. While the TBMs mined through less permeable and more confined portions of the gas zone, particularly when further bounded by the newly constructed tunnel, TBM operation caused gas pressures to increase substantially and there were minor, short-lived ingressions of CH4 into the sumps at the deepest level of two building basements.
- Safely tunneling through the gassy ground was made possible by effective planning, investigations, operation of a gas extraction system, and coordination between the members of the WPLE1 team.
American Public Transportation Association (APTA). 2005. Peer Review Panel Report on the Wilshire Corridor Tunneling Project. Report prepared for Los Angeles County MTA, November 2005. 28p. http://libraryarchives.metro.net/DPGTL/peerreview/2005_apta_wilshire_corridor_tunneling_final_report.pdf.
American Public Transportation Association (APTA). 2012. Peer Review Report on the Wilshire Corridor Westside Extension Tunneling Project— Part-II. Report prepared for Los Angeles County MTA, November 1, 2012. 19p.
Arnold R. 1906. The Salt Lake Oil Field Near Los Angeles, Cal. Chapter in Eds. Arnold R, Clapp FG. Contributions to Economic Geology, 1905: Petroleum and Natural Gas. USGS Bulletin 285-G. https://pubs.er.usgs.gov/publication/b285G.
Bilodeau WL, Bilodeau SW, Gath EM, Oborne M, Proctor RJ. 2007. Geology of Los Angeles, California, United States of America. Environmental and Engineering Geoscience, XIII(2):99–160. https://www.aegweb.org/assets/docs/la.pdf.
California Geologic Energy Management Division (CalGEM). Well Finder. Online geographic information system. https://maps.conservation.ca.gov/doggr/wellfinder/#/. API 03715144: https://filerequest.conservation.ca.gov/WellRecord?api=03715144. API 03721161 2018 data file: https://filerequest.conservation.ca.gov/WellRecord?api=03721161.
Chilingar GV, Endres B. 2005. Environmental hazards posed by the Los Angeles Basin urban oilfields: an historical perspective of lessons learned. Environmental Geology, 47:302–317. https://link.springer.com/article/10.1007/s00254-004-1159-0.
City of Los Angeles. 1985. Task Force Report on the March 24, 1985 Methane Gas Explosion and Fire in the Fairfax Area, City of Los Angeles. Report prepared by spe- cial task force for the mayor and city council. http://libraryarchives.metro.net/DPGTL /losangelescity/1985_methane_gas_explosion_task_force_report_fairfax.pdf.
Cobarrubias JW. 1992. Mathane gas hazard within the Fairfax District, Los Angeles. Engineering Geology Practice in Southern California, Association of Engineering Geologists, Special Publication No. 4, 131–143.
Converse Ward Davis Dixon Earth Science Associates Geo/Resource Consultants (CWDD). 1981. “Geotechnical Investigation Report,” Volume 1, Southern California Rapid Transportation District Metro Rail Project. Available at UCLA library collection.
Copur H, Cinar M, Okten G, Bilgin N. 2011. A case study on the methane explosion in the excavation chamber of an EPB-TBM and lessons learnt including some recent accidents. Tunnelling and Underground Space Technology, 27:159–167. doi:10.1016 /j.tust.2011.06.009.
Crowder RE and Johnson RA. 1961. Los Angeles City Oil Field. Chapter in Summary of Operations California Oil Fields, Annual Report No. 47, Department of Natural Resources, Division of Oil and Gas. pp. 67–77.
Crowder RE. 1963. Recent Developments in Jade-Buttram Area of Salt Lake Oil Field. Chapter in Summary of Operations California Oil Fields, Annual Report No. 49, Department of Natural Resources, Division of Oil and Gas. pp. 53–58.
Division of Oil, Gas, and Geothermal Resources (DOGGR). 1992. California Oil and Gas Fields: Vol. II— Southern, Central Coastal, and Offshore California Oil and Gas Fields. DOGGR Publication TR12. 645p.
Elioff MA, Smirnoff TP, Ryan PF, Putnam JB, Ghadiali BM. 1995. Geotechnical inves- tigrations and design alternatives for tunneling in the presence of hydrogen sulfide gas— Los Angeles Metro. Proceedings of the 1995 Rapid Excavation and Tunneling Conference, Ch. 19, 299–318.
Endres B, Chilingarian GV, Yen TF. 1991. Environmental hazards of urban oilfield operations. Journal of Petroleum Science and Engineering, 6(2):95–106. https://doi.org/10.1016/0920-4105(91)90030-Q.
Enviro-Rail, 1996. Phase II Western Extension Reassessment Study. Report pre- pared for Los Angeles County Metropolitan Transportation Authority. March 1996. 574p. Figure 3-2.
Etiope G, Doezema LA, Pacheco C. 2017. Emission of methane and heavier alkanes from the La Brea Tar Pits seepage area, Los Angeles. Journal of Geophysical Letters: Atmospheres, 122:12,008-12,019. https://doi.org/10.1002/2017JD027675.
Hamilton DH, Meehan RL. 1992. Cause of the 1985 Ross Store explosion and other gas ventings, Fairfax District, Los Angeles. Engineering Geology Practice in Southern California, Association of Engineering Geologists, Special Publication No. 4.
Jeffrey AWA, Alimi HM, Jenden PD. 1991. Geochemistry of Los Angeles Basin Oil and Gas Systems. Chapter 6 in Ed. Biddle KT, Active Margin Basins. American Association of Petroleum Geologists. https://doi.org/10.1306/M52531C6.
Jung B, Garven G, Boles JR. 2015. The geodynamics of faults and petroleum migration in the Los Angeles Basin, California. American Journalof Science, 315:412–459. DOI 10.2475/05.2015.02.
Kaiser Industries Corporation (Kaiser). 1961. General Description of Rapid Transit System Backbone Route for Los Angeles Metropolitan Transit Authority. July 1961. http://libraryarchives.metro.net/DPGTL/lamta/1961_kaiser_general_description_rapid_transit_system_backbone_route.pdf
Kaiser Engineers California (Kaiser) and Gage-Babcock & Associates, Inc. (Gage-Babcock). 1983. Study of Methane and Other Combustible Gases Effect on Underground Operation of the Metro Rail Project. March 1983. http://libraryarchives .metro.net/DPGTL/scrtd/1983-study-of-methane-and-other-combustible-gases -effect-on-underground-operation-of-the-metro-rail-project.pdf
Khilyuk LF, Chilingar GV, Endres B, Robertson JO. 2000. Gas Migration: Events Preceding Earthquakes. Gulf Professional Publishing, Houston, TX. ISBN 0884154300, 9780884154303, 389p.
Lang HR and Dreessen RS. 1975. Subsurface structure of the northwestern Los Angeles Basin. California Department of Conservation Division of Oil and Gas Publication No. TP01. https://www.conservation.ca.gov/calgem/pubs_stats/Pages/ technical_reports.aspx.
Los Angeles County Metropolitan Transportation Authority (LA Metro). 2011. Westside Subway Extension Project, Wilshire/Fairfax Station Construction. Paleontological Resources Extraction. Attachment 3 to Appendix G Memorandum of Understanding for Paleontological Resources, Final environmental Impact Statement/Environmental Impact Report, Vol. 4.
Los Angeles County Metropolitan Transportation Authority (LA Metro). 2014. Westside Subway Extension Project, Section 1: Contract 1045 Geotechnical Baseline Report. Conformed November 3, 2014. Prepared by Parsons Brinkerhoff.
Proctor RJ and Monsees JE. 1985. Los Angeles Metro Rail Project: Design issues related to gassy ground. Proceedings of the 1985 Rapid Excavation and Tunneling Conference, vol 1, ch.30, 488–505.
Proctor RJ. 2002. The San Fernando Tunnel explosion, California. Engineering Geology, 67(1-2):1–3. https://doi.org/10.1016/S0013-7952(02)00042-X
Samuelian RH. 1990. South Salt Lake Oil Field. Chapter in California Department of Conservation, Division of Oil and Gas Publication No. TR32. Originally submitted 1984. https://www.conservation.ca.gov/calgem/pubs_stats/Pages/technical_reports.aspx
Shaw AC, Quinn JP. 1986. Rancho La Brea: A look at coastal southern California’s past. California Geology, 39:123–133.
US Congress. 1985. Public Law Statute 99-190. December 19, 1985. https://www.govinfo.gov/content/pkg/STATUTE-99/pdf/STATUTE-99-Pg1185.pdf
US Congress. 2007. Public Law Statute 110-161. December 26, 2007. https://www.congress.gov/110/plaws/publ161/PLAW-110publ161.pdf
Wright T. 1987. Geological setting of the Rancho La Brea Tar Pits. In AAPG Pacific Section 2009 — Petroleum Geology of Coastal Southern California (1987), 87–91.