Late Quaternary slip rates on the Sierra Madre fault zone and paleoseismic evidence on the size and frequency of past ruptures
Published:May 18, 2020
- PDF LinkChapter PDF
R. Burgette*, K. Scharer*, S. Lindvall*, 2020. "Late Quaternary slip rates on the Sierra Madre fault zone and paleoseismic evidence on the size and frequency of past ruptures", From the Islands to the Mountains: A 2020 View of Geologic Excursions in Southern California, Richard V. Heermance, Joshua J. Schwartz
Download citation file:
The Sierra Madre fault zone is a south-vergent, active reverse fault that accommodates shortening between basins on the northern margin of the Los Angeles region and the San Gabriel Mountains. The preservation of late Quaternary alluvial fill and fan surfaces in the hanging wall of the fault provides evidence of long-term uplift. Surface rupture from the 1971 Mw 6.6 San Fernando earthquake and evidence of large prehistoric displacements from trenching investigations emphasize the ongoing hazard posed by the fault system to the region. This one-day field trip visits some of the key locations near Pasadena and San Fernando, California, where slip rates have been determined from cosmogenic and luminescence dating of abandoned surfaces dating to 50–70, ca. 30, and ca. 12 ka and surface offsets measured from lidar and pre-development topographic maps. Another stop is the site of a paleoseismic trench, which provided key evidence on the timing and displacement of past ruptures on the fault. In combination, results from these field investigations converge on a slip rate for the eastern ~100 km of the fault zone of 1–2 mm/yr, which matches or exceeds the rates for other reverse faults in southern California. This rate, in combination with trenching data that show no evidence of post–mid Holocene ruptures along the central and eastern portions of the fault, indicate the fault zone poses a significant seismic hazard to the region.
The Sierra Madre fault zone (SMFZ) of southern California extends 140 km from the San Jacinto fault in the east to the faults of the Ventura basin in the west and accommodates compression associated with a broad left bend in the trace of the San Andreas fault (Fig. 1). At the regional scale, the SMFZ delineates a geomorphic boundary between deeply incised Proterozoic and Mesozoic crystalline bedrock of the San Gabriel Mountains to the north and the smooth alluvial fill of the San Bernardino, San Gabriel, and San Fernando basins to the south (Fig. 1). The SMFZ is split into sections based on historic earthquakes, geomorphic expression, and geometric segmentation by left lateral splay faults. This field trip focuses on recent work on three sections: the Cucamonga fault (CF), the Central Sierra Madre fault (CSMF), and the San Fernando fault (SFF). In 1971 the Mw 6.6 San Fernando (or Sylmar) rupture broke the western ~20 km of the SMFZ. Concurrent rupture on the CSMF and SFF could produce an ~Mw 7.3 earthquake (Fig. 2), or a larger event if neighboring fault sections are involved. Ruptures such as these would produce severe to violent shaking over much of the northern portion of the basins (Fig. 2), impacting millions of people and the water, power, and transportation infrastructure that is critical for the region.
This field trip summarizes locations where late Quaternary slip rates and paleoseismic trenching investigations have been completed to understand the hazard this fault poses to the Los Angeles region. The goal of these investigations is to understand the strain accommodated by the fault and the sizes of past earthquakes that release that strain. The emphasis of the field guide is on the dating methods and displacement estimates for alluvial surfaces uplifted by the fault and it also provides a summary of trenching results along the fault trace. The first stop is designed to set the SMFZ within the regional context of other thrust and oblique reverse faults south of the San Andreas fault.
Drive to Stop 1
The field trip starts at an overlook on the Verdugo Mountains within Cherry Canyon Park, permitting views of the compressional structures between the San Gabriel Mountains and the Los Angeles basin. From the 2020 meeting’s conference hotel (Westin Pasadena, 191 N. Los Robles Avenue, Pasadena, California), drive north on Los Robles Avenue. After 200 ft, turn left onto E. Walnut Street, and then after 0.8 mi turn right onto I-210 West toward San Fernando. After 2.3 mi, take the Arroyo Boulevard exit. Turn right on Windsor Avenue. Take the first left onto Oak Grove Drive and cross the Devil’s Gate reservoir spillway. At the first stop sign, turn left to cross the freeway overpass, and then right onto Highland Drive. After 1 mi, keep left onto Highland Drive, then turn left onto Chevy Chase Drive. After 700 ft, turn right onto Inverness Drive. Take the first left onto Berwick Drive and then right onto Hampstead Drive. After 0.5 mi, turn left onto Sugar Loaf Drive (La Canada Flintridge, California). Drive to the end and park on the side of road near fire road entrance (34.1876°N, 118.2071°W). Pass through the gate and walk a short distance uphill on gravel roads, curving left (south) to stand at base of radio towers for best view.
Stop 1. Overlook (34.1874°, −118.2077°)
At this stop, we will review the general geography, geologic history, and current configuration of the San Gabriel Mountains and Los Angeles basin in order to introduce the question of how shortening is accommodated. Geodetic studies resolve 6–7 mm/yr of shortening between the San Gabriel Mountains and the Palos Verdes Peninsula (Fig. 1A, Argus et al., 2005; Daout et al., 2016), but discriminating which fault(s) have the highest slip rate and accumulate the largest moment depends on modeling assumptions of the crustal structure and fault geometry (e.g., Marshall et al., 2009; Rollins et al., 2018; Walls et al., 1998). Geologic studies establish that shortening is accommodated by thrust and oblique-reverse faults under each of the mountains visible from the Stop 1 vantage point and a few structures that are buried by high sedimentation rates in the basins. From south to north (Fig. 3), the faults (and slip rates) include the Wilmington (0.14–0.18 mm/yr; Wolfe et al., 2019) and Compton blind thrusts (0.9–1.7 mm/yr; Leon et al., 2009); the Puente Hills blind thrust fault system under downtown Los Angeles (0.2–1.3 mm/yr, Bergen et al., 2017); the Elysian Park anticline and associated thrust faults (shortening of 0.6–1.1 mm/yr; Oskin et al., 2000); the Santa Monica–Hollywood fault (~0.6 mm/yr, Dolan et al., 2000) and the Verdugo fault (no Quaternary slip rate available, estimate is ~0.4 mm/yr). Slip rates for the SMFZ range from 0.6 to 3 mm/yr; we emphasize recent dating by Burgette et al. (2020) that estimate a median rate of ~1 mm/yr.
Background on the Sierra Madre Fault Zone
Early mapping of the Proterozoic and Mesozoic bedrock and structures of the San Gabriel Mountains was completed by Ehlig (1975) (and references therein) and the long-term tectonic history is detailed in Matti and Morton (1993). Paleogeographic reconstructions of the San Gabriel Mountains indicate they have been a source for of sediment in the Los Angeles basin since the Miocene (see Blythe et al., 2002, and references therein for summary). Fission-track dating shows a pulse of uplift initiated ca. 12 Ma, likely reflecting initiation of the San Gabriel fault (Fig. 1, Blythe et al., 2002). Low-temperature apatite He ages of 7 Ma from the Sierra Madre block are interpreted as reactivation of reverse faulting on the Sierra Madre fault (Blythe et al., 2002). Two detailed studies by Crook et al. (1987) and Morton and Matti (1987) establish through geomorphic mapping and trenching investigations that crystalline bedrock through late Quaternary sediments are thrust over younger deposits along the entire range front. Both long-term thermochronology (Spotila et al., 2002) and shorter-term catchment average erosion rates show higher erosion rates (>0.5 mm/yr) in the east, along the CF and eastern half of the CSMF, reducing to <0.1 mm/yr along the western CSMF (Spotila et al., 2002; DiBiase et al., 2010). The SMFZ is multi-stranded and each strand exhibits differing histories (e.g., Treiman, 2013; McPhillips and Scharer, 2018; Crook et al., 1987). Limited geologic data for each strand (Fig. 1B) contributes to the challenge of sorting out the along-strike slip rate of this fault system. Investigations of the Holocene earthquake activity along the full SMFZ show an irregular pattern of earthquake recurrence (Crook et al., 1987; Tucker and Dolan, 2001; Rubin et al., 1998).
Brief Summary of Neotectonic Investigations on the Cucamonga Fault
The Cucamonga fault (CF; Fig. 1B) differs from the rest of the SMFZ because the majority of its surface trace is south of the range front, crossing more continuous fan surfaces. Older studies of the late Quaternary slip rate estimates (e.g., Morton and Matti, 1987) are high (~5 mm/yr) compared to the rest of the SMFZ. These higher rates were based on fan ages determined from regional soil chronosequences. More recent estimates combine analysis of cosmogenic nuclide dating (Homer, 2006; Lindvall and Rubin, 2008) with detailed study of the fault strand activity and offsets, and produce a lower slip rate (1–2 mm/yr) since ca. 38 ka (McPhillips and Scharer, 2018). There are no published trench logs from the Cucamonga fault that provide details on the earthquake history, but an abstract by Dolan et al. (1996) suggested several earthquakes with 1–3 m of slip have occurred since ca. 7 ka (Walls, 2001).
Drive to Arroyo Seco, Stops 2–4
Drive to Stop 2
Drive southwest down Sugar Loaf Road for 0.1 mi; turn right onto Hampstead Road. In 0.5 mi turn right onto Inverness Drive and take the next left onto Chevy Chase Road. After 0.1 mi turn right onto Highland Drive and continue for 1.3 mi to Linda Vista Road. At this location you have exited the Verdugo Mountains and are driving on a Q3 surface on the west side of Arroyo Seco. Cross the freeway and turn right onto Oak Grove Drive then continue straight onto Woodbury Road a total of 0.9 mi. Turn left on Lincoln Ave, heading north for 1.4 mi. Turn left onto Canyon Crest Road, drive 1.2 mi and take the left on Rising Hill Road (also called Cloverhill Road). Take next right onto El Prieto Road.
This section of the drive goes up the mouth of Millard Canyon (Figs. 4 and 5). The route passes through a thick fill deposit of Q4 and ends on a south sloping terrace surface of Q3.3. The bus will drop off at 4338 El Prieto Road Altadena (34.2121°, −118.1649°). A confluence of drainages intersects here; the largest is the 46 km2 Arroyo Seco drainage, which is joined by the 2.6 km2 El Prieto tributary. To the east, Millard Canyon drains 7.2 km2 of the San Gabriel Mountains.
NOTE: Stops 2, 3, and 4 are located within the broader Arroyo Seco drainage (Fig. 5); the guide is designed to walk between these stops following the green line in Figure 5, a total of 3 miles. If unable to shuttle vehicles, park at the Stop 4 pickup location (34.2001°, −118.1647°), 898 W Altadena Drive, and walk to the stops, although this involves more backtracking.
Walk to Stop 2
From parking on 4338 El Prieto Road, walk northwest through the gate at end of residential driveway and continue northward down the path cut into terrace riser.
Stop 2. The Meadows (34.2136°N, 118.1653°W)
The purpose of this stop is to introduce the late Quaternary deposits and abandoned surfaces and provide a basic introduction to the dating techniques used in recent studies of the SMFZ. This stop is located on a Q3.2 surface dated by Burgette et al. (2020) (Fig. 5). As the terrace remnant is narrow and colluvium from housing construction produced a thick apron on the surface, the depth profile for this terrace deposit (AS-4, 33.0 +7.3/–7.9 ka) is a composite of four samples collected from a 1.8 m soil pit dug into the surface, and two deeper samples collected from the exposure in the canyon wall (Fig. 6; Hanson, 2018). An infrared stimulated luminescence (IRSL) sample collected from the colluvium (SM-3.1L) produced an age of 2.1 ± 0.4 ka. Look to north for view of Q3.3 surface indicated in Figure 4.
Late Quaternary Deposits along the Central Sierra Madre Fault
The Pleistocene to Holocene units along the CSMF consist of fine to coarse sand, silty sand and gravel layers from alluvial sheetflood and fluvial sources, interbedded with debris flow and mud flow deposits. North of the SMFZ, the deposits are preserved as fill and fill-cut terraces and south of the fault the geomorphic features include fill-cut terraces and alluvial fans (Fig. 7). Crook et al. (1987) split late Quaternary alluvial deposits of the CSMF into four broad age categories (Qal1–Qal4) based on soil development, relative geomorphic position and extent. They differentiated abandoned surfaces with an f (e.g., Qal3f) where soil development indicated a stable surface. Burgette et al. (2020) simplified the labeling to units Q1–Q4 and further subdivided each unit (e.g., Q3.3, Q3.2) based on analysis of the surface elevations from lidar and field mapping.
This field guide largely uses the designations of Burgette et al. (2020), as the focus is on the dating of these surfaces; previous soils studies used different naming conventions (e.g., McFadden, 1982; Walls, 2001). Distinguishing features of each unit are summarized below.
Q4 (Qal4) alluvium is red to reddish-yellow (2.5 YR to 5 YR) in hue, strongest in surfaces with preserved soils, and has higher clay content and weathered clasts. Q4 deposits fill the mouth of Arroyo Seco to a thickness of over 90 m on the hanging wall and can be seen in high outliers along the range front. (Visible from Stop 1 are the La Vina Fan, Gould Mesa, and ancient La Canada Fan of Crook et al., 1987.) Boreholes through the footwall alluvium at Arroyo Seco show a similar red, clay-rich layer preserved with a throw of 230 ± 10 m across the CSMF (Crook et al., 1987; Burgette et al., 2017). Crook et al. (1987) assigned the upper surface of Q4 an age of 100–300 k.y. based on correlation with regional soil profiles in the San Joaquin valley, noting a Pleistocene mammoth fossil of was found in a correlative soil in Monrovia. These estimates are supported by Burgette et al. (2017) who analyzed 10Be isotope concentrations in the clasts buried at least 8 m below the top of Q4. Assuming no inheritance and a muon production rate, they calculated a mean exposure age of 170.6 ± 31.5 ka.
Q3 (Qal3 and Qal3f) deposits are yellow to yellowish-brown to pale brown in hue (7.5 YR to 10 YR) and contain moderately weathered clasts and little to no clay (Crook et al., 1987). This unit is characterized by an abundance of large-boulders that can be observed in gullies close to the mountain front. Q3 fan surfaces continue as far as 3 km south of the fault and underlie what is now largely residential developements. Crook et al. (1987) argued based on soil development that the Q3 fan surfaces are older at the southern limits (e.g., at the 210 freeway exposures in Pasadena) and that younger Q3 deposits are preserved at the mouths of the canyons. Mapping by Burgette et al. (2020) established several subunits (Q3.3, Q3.2, Q3.1) of Q3 based on elevation in Arroyo Seco (Figs. 5, 7).
Q2 (Qal2f) is preserved as discrete alluvial fans deposited on Q3 deposits at the mouths of smaller drainages (e.g., east of Arroyo Seco) and as preserved terrace surfaces within larger drainages (e.g., west side of Arroyo Seco south of the Jet Propulsion Laboratory [JPL]). The deposits are unconsolidated and olive to pale brown in color (5 YR to 10 YR) with a poorly developed A horizon. Crook et al. (1987) summarize several 14C dating studies that estimate the start of unit 2 deposition occurred at the Pleistocene-Holocene boundary based on radiocarbon dating. Burgette et al. (2020) provide dates for three samples from Q2 deposits using IRSL and terrestrial cosmogenic nuclide (TCN). The ages support an older range for Q2; from 28 to 10 ka. (See Stop 3.)
Q1 (Qal1) is present in the modern channels and flood plains. Historic floods produced extensive deposits of unit 1 in 1934 and 1938 in the vicinity of La Canada–Flintridge, leading to development of a series of debris dam structures along the range front. (See McPhee, 1990, for a popular review of this topic.)
General Summary of Dating Methods
The chronology of terrace and fan formation in Burgette et al. (2020) is based on TCN surface exposure, infrared-stimulated luminescence (IRSL), and radiocarbon dating (14C). Each of these methods offers different strengths and complications. TCN dating provides ages of the surfaces used to estimate offset across the fault and IRSL dating yields ages of burial of sediment. Burgette et al. (2020) used IRSL to date sediment below geomorphic surfaces as well as sediment covering fan surfaces to understand the history of aggradation of cover sediment near TCN sample sites.
Cosmogenic nuclides are produced through interactions between high energy particles derived from cosmic rays and nuclei of atoms in target minerals (e.g., Gosse and Phillips, 2001). Production rates of cosmogenic nuclides vary spatially due to latitudinal differences in cosmic ray flux controlled by the Earth’s magnetic field as well as surface elevation, as the cosmic radiation is progressively attenuated through the thickness of the atmosphere. The accumulation of cosmogenic nuclides is concentrated near Earth’s surface, and consequently, TCN dating provides an effective means to measure the duration of exposure of geomorphic surfaces following abandonment by down-cutting streams (e.g., von Blanckenburg and Willenbring, 2014). Complications of TCN dating include post-abandonment deposition or erosion and nuclides inherited during erosion and transport. These issues can be mitigated by selection of sample sites on stable surfaces and measurement of inherited nuclide inventories through sampling over a depth profile.
Luminescence dating measures the time since mineral grains were last exposed to sunlight or heat (e.g., Aitken, 1998). The principle of luminescence dating is that radioactive decay in surrounding minerals causes electrons to become trapped in metastable, higher energy states within crystal lattice defects. With the addition of energy through exposure to light, the trapped electrons are released and the mineral grain luminesces with an intensity that is proportional to the total dose of radiation received during the interval of burial. Measurement of the local radiation dose rate, and the laboratory derived sensitivity of the observed aliquots of minerals yields the age since the last exposure to sunlight. Complications of luminescence dating include insufficient exposure to light during transport and deposition (“partial bleaching”), sensitivity of mineral grains to radiation, and in the case of feldspar, “anomalous fading,” a process in which some trapped electrons are lost without an external stimulus. As luminescence dating of quartz is unreliable in some cases in southern California (Lawson et al., 2012), Burgette et al. (2020) used infrared luminescence dating (IRSL) of feldspar. To mitigate the effects of anomalous fading, they employed a post-IR IRSL protocol, where aliquots are pretreated with a limited dose of IR light at elevated temperature before measurement of the signal, as well as applying a correction for fading based on behavior in the laboratory (Burgette et al., 2020).
Walk to Stop 3 (1.1 mi)
Continue down path, turning sharp left at junction with the main trail in El Prieto. (Listen and look for mountain bikers coming quickly down path!) As you walk downstream, notice the thick gravel deposits on hornblende–biotite–quartz diorite (Cretaceous Wilson Diorite of Miller, 1934) in steep canyon walls. After ~0.5 miles arrive at the trail junction and head right (uphill) on this wider trail. Note the pale red color of cobbly deposits on the left and the change to finer-grained alluvium as you near asphalt pad with fenced utility structures. Cross the low fence to east side of terrace to overlook El Prieto drainage.
Stop 3. Jet Propulsion Lab Tower Site, on Q3.3 (34.213646°N, 118.167458°W)
This location on a Q3.3 surface provides a view of the Q3 and Q4 deposits as well as an ability to see depositional relationship between cobbly gravel terrace deposits and finer-grained capping deposits used to test stability of the surface. To the east, across the El Prieto tributary, observe the lower Q3.2 surface (Stop 2) and its continuation to southwest (where metal roofing of horse stables is visible). To the west, across the main Arroyo Seco drainage, a 90 m fill deposit topped by red soil (Q4) is visible. To the south, the Arroyo Seco itself (Q1) is seen at the base of a Q3 riser. The Raymond fault crops out just south of the tall buildings of Pasadena.
Review of Arroyo Seco Dates and Displacements
Burgette et al. (2020) dated four locations in Arroyo Seco (Figs. 5, 6); in the hanging wall there are two on a Q3.3 surface and one on Q3.2 (Stop 2). On the footwall they dated a younger fan (Q2.2). Due to the narrow extent of preserved surfaces in the hanging wall, displacements are possible to measure on only the Q3.2 and Q3.3 surfaces.
Based on TCN dates in two locations, the Q3.3 surface in the Arroyo Seco area (Figs. 5, 6) was developed in the late Pleistocene, prior to the global Last Glacial Maximum. Burgette et al. (2020) dated the Q3.3 surface both at this location and another site to the east from a cut exposure along Canyon Crest Road. The TCN sample site at Stop 3 was selected near the end of a deposit of fine-grained sediment that caps the fluvial gravel deposits and thickens toward the hillslope to the north. Two IRSL samples collected from the capping sediment show stratigraphic consistency and indicate that the cover sediment aggraded after abandonment of the terrace surface (Fig. 6B). The TCN age from this site is 46.7 ± 11.2 ka, based on a depth profile that showed the exponential decay of 10Be concentration accumulated in stable near-surface sediment (Burgette et al., 2020). This age is calculated assuming the 25 cm of fine-grained cover sediment has been present since incision of the surface. Neglecting the aggradation of the cover sediment over millennia biases the TCN age to an older value; however, potential more recent erosion would impart a bias toward younger age. Consequently, they regard the TCN model ages to be accurate within limits to resolve and model the history of the sampled surfaces. An age for the Q3.3 surface to the east along Canyon Crest Road is older at 62.1 ± 15.0 ka (Fig. 6A). That TCN depth profile showed more evidence of surface disturbance, although it is located closer to the site where they measured vertical separation of the geomorphic surface. Given such tradeoffs and the overlap of model age distributions at the 2-σ level, Burgette et al. (2020) used a summed age probability distribution that preserves the full range of both Q3.3 age estimates (53.4 +21.3/–14.5 ka) when calculating slip rate for Q3.3.
The Q3.3 surface is offset across the two southerly strands of the CSMF east of Arroyo Seco (Fig. 6F). Vertical separation across the northern strand (Stop 5) is 6.1 +1.0/–0.9 m, and there is 21.8 +1.7/–2.2 m of vertical separation across the southern strand (Stop 4). The preserved surface on the footwall appears to be the Q3.2 surface, thus the total vertical separation of 27.9 +2.1/−2.5 m may be a minimum (Burgette et al., 2020). However, the distal portion of Q3.2 used as the marker appears to have received less deposition, supported by the interpretation of older soil ages down-fan (Crook et al., 1987), so the potential bias may be minimal. Estimated dip slip across the two strands using fault dip information is 58.5 +46.3/−14.4 m since the abandonment of the Q3.3 surface, producing a median slip rate of 1.1 +1.0/−0.4 mm/yr, assuming a dip range of 10–55°.
The Q3.2 surface is faulted by the southernmost strand of the fault (Stop 4) with a vertical separation of 13.8 +2.6/–2.8 m (Fig. 6E). Combined with a dip range of 10–55° and the Q3.2 age (Fig. 6C), the post–Q3.2 incremental dip-slip rate is 0.8 +1.0/–0.3 mm/yr.
Other CSMF Slip Rates
Burgette et al. (2020) also dated late Quaternary surfaces offset by the CSMF in Pickens Fan and Dunsmore Canyon (Fig. 1B). The Pickens Fan was mapped as Qal3f by Crook et al. (1987) and produced overlapping IRSL and TCN ages that are consistent with the Q3.2 surface at Arroyo Seco (37.4 +8.4/–9.1 ka). The fan at Pickens Canyon is offset by three strands of the CSMF. Assuming a dip between 14 and 45° the total slip is 30.9 +20.2/–8.7 m, producing a slip rate of 0.8 +0.6/–0.3 mm/yr. At Dunsmore Canyon, two strands of the CSMF cut a young Q2.1 fan surface, measured by Crook et al. (1987) at ~6 m of vertical (a large debris basin is now excavated where the scarps traversed the fan). IRSL dating (12.1 ± 2.5 ka) of alluvium below a Q2.1 terrace preserved upstream provides a maximum age for the surface that when combined with a dip slip displacement of 7.5 +5.4/–3.1 m produces a minimum slip rate of 0.6 +0.7/–0.3 mm/yr. All of the incremental rates calculated by Burgette et al. (2020) overlap at ~1 mm/yr, but examined individually, suggest modest slowing over the latest Quaternary (Fig. 8).
Walk to Stop 4
The next stop is the JPL “Bridge Fault” of Crook et al. (1987), 1.2 miles from Stop 3. From Stop 3, return back down the trail. Note the inset Q3.1 terraces upstream as you walk down, and the views of Q4 surface above. At the junction with El Prieto trail stay right to continue downstream. Cross the El Prieto stream itself and at junction with paved road turn left (downstream). Cross two bridges then look for the access bridge and road to JPL below and to your right. Take the footpath to base of JPL bridge on upstream side.
Stop 4. JPL Bridge (34.2030°N, 118.1657°W)
The southernmost mapped stand of the CSMF crosses Arroyo Seco at this location. Crook et al. (1987) called this strand the “Bridge Fault” because it crosses a bridge used to access JPL. Although the fault is obscured by colluvium, note that crystalline bedrock transitions to alluvial deposits at the elevation of the road over just a dozen or so meters. A similar relationship can be observed in the bottom of the channel itself. In the winter of 2019, a drainage pipe failed and scoured out a fresh exposure of the fault, revealing a green-white clay gouge above Quaternary gravel (Fig. 9). This relationship establishes that fault movement has occurred since deposition of the gravel, which are likely Q3 units as they crop out in the road cut above this stop (Fig. 5).
Walk to Stop 4 Pickup Location (34.2001°N, 118.1647°W)
The bus will be waiting near 898 W. Altadena Drive, on the Q3.2 surface just south of and above Stop 4. To get there, continue along the main paved road for ¼ mi, walk up the trail on left (east) that cuts behind the Pasadena Water Treatment building (see small sign labeled “Gabrielino Trail”).
Drive to Stop 5
To get to Loma Alta Park, head east on W. Altadena Drive for 0.4 mi, then turn left onto N. Lincoln Ave. The park is on the right after 0.4 mi; park in lot west of swimming pool or the larger lot on the east side of community recreation buildings, accessed by turning right on Loma Alta Drive and left onto Sunset Ridge Road.
Stop 5. Loma Alta Park (34.2038°N, 118.1590°W)
Loma Alta Trench
A paleoseismic trench in Loma Alta Park on strand C of the CSMF where it crosses a Q3.3 surface documented two large-displacement earthquakes in the past ~15 k.y. (Fig. 5; Rubin et al., 1998). A minimum of ~4 m of reverse displacement was produced in the most recent event and ~10.5 m of cumulative displacement is documented for the past two ruptures, yielding an average slip of ~5 m per event. The trench was located directly north of the parking lot in the broad south-facing scarp (Fig. 10). Results of the Rubin et al. (1998) trench study at Loma Alta Park are summarized below.
The Sierra Madre fault appears in the trench wall as a single primary trace with minor faulting in a zone ~0.5 m wide within the hanging wall of coarse alluvial gravels (Fig. 11). Multiple lines of evidence were used to identify prehistoric earthquakes. Truncation of stratigraphic units is the clearest evidence for past events. Rubin et al. (1998) also interpret wedge-shaped colluvial deposits to represent material shed off the fault scarp immediately after surface-rupturing earthquakes. Trench-wall exposures show two colluvial wedges (Fig. 11, units 3 and 4a). The younger colluvial wedge contains southward thinning coarse gravels (unit 4a), lies directly above the fault zone, and is the primary evidence for the most recent earthquake. The wedge-shaped geometry and composition of the unit imply that it is a colluvial deposit derived from unit 1, exposed on the fault scarp shortly after an earthquake. Unit 3, also a massive pebbly sand deposit sits on top of a well-developed soil horizon (unit 2) and is interpreted as an older colluvial wedge, formed by the penultimate earthquake (Fig. 12).
Restoration of the dip-slip component of motion on the fault reveals the approximate geometry of the alluvial and colluvial deposits and buried soil before the most recent earthquake and gives a minimum of ~3.8–4.0 m of dip slip from the most recent earthquake. Restoration of the upper tip of unit 1 to below unit 2 (the buried soil) yields a cumulative minimum slip of ~10.5 m for the last two events.
The earthquake ages are estimated from 14 radiocarbon dates of detrital charcoal samples collected from the trench (Fig. 11). The wide range of detrital radiocarbon ages is interpreted to be the result of recycling detrital charcoal from older alluvial deposits exposed uphill and north of the fault scarp. Units 2 through 4a are late to latest Pleistocene in age. The youngest detrital charcoal age from unit 3 of 18,500 cal yr B.P. provides a maximum age for the colluvial units and the penultimate earthquake. Four fragments of detrital charcoal from unit 4a yielded ages of 10,600–18,200 cal yr B.P. Thus, the maximum age about the most recent earthquake is 10,600 cal yr B.P.
A minimum age for the most recent event is less certain because of extensive bioturbation that typically occurs in A horizons (topsoil). Two detrital charcoal fragments recovered from a bulk sample yielded radiocarbon ages of 1107 and 3350 cal yr B.P. These young radiocarbon ages were interpreted by Rubin et al. (1998) as minimum ages because it is likely that the charcoal was incorporated biogenically after deposition of unit 4b.
Other Paleoseismic Studies along the Sierra Madre Fault
The eastern end of the Central Sierra Madre fault in the town of San Dimas was investigated by Tucker and Dolan (2001) (Fig. 1B). In this work, they combined large-diameter borings and a 62-m-long paleoseismic trench along the eastern margin of a small drainage incised across the range-front fault. Geologists were lowered into three of the borings, and other borings were logged by cuttings taken at sub–50 cm intervals. The trench and borings exposed fault gouge dipping at 22° to the north that places Miocene sedimentary rocks on Quaternary gravels. Younger alluvial units are deposited unconformably across the top of fault. Radiocarbon dating places the most recent motion of this fault strand between ca. 20 and 8 ka. Tucker and Dolan (2001) also use the displacement of the Miocene unit (>14 m) and the date of last motion to estimate ~4.6 m of accumulated strain on this section of the fault.
Crook et al. (1987) report on several trenches completed at Arroyo Seco and Dunsmore Canyon. A trench at Arroyo Seco, excavated on a lower (Qal1) terrace on the east edge of the channel revealed Cretaceous monzonite thrust over Qal3 alluvium. Three additional trenches on the west side of the channel (within JPL property) were excavated to determine the youngest age of faulting; in these trenches they found no evidence faulting in their “Unit 2” deposits. In a trench log from Pasadena Glen located ~9 km east of JPL, Crook et al. (1987) show Qal2 colluvium resting on a fault that places Cretaceous diorite over Qal4 alluvium, in turn overlain by unfaulted Qal2 alluvium (Crook et al., 1987; their fig. 2.10). Finally, at Dunsmore Canyon, 9 km west of JPL, trenches show ~5 m of displacement where a Qal2 alluvial deposit with capping soils preserved is thrust over undifferentiated Qal2 alluvium (Crook et al., 1987; their fig. 2.5). Unfortunately, dating techniques at the time precluded dating of these Q2 deposits, so the timing of last motion at these sites is unknown.
There are a few reports of trenching investigations along the San Fernando section of the Sierra Madre fault. Farthest west is the Oak Hill trench of Bonilla (1973), where there is evidence that the 1971 earthquake faulted an older colluvial wedge that contained a piece of wood dated to 100–300 years old. Bonilla (1973) argue displacements are similar (~1 m) in the historic and penultimate event. Similar information was determined by Midttun et al. (2015) in their analysis of unpublished trench photographs and logging by Fumal and colleagues on a trench at the Middle Ranch site completed in 1995. They report two events (the 1971 and penultimate) occurred in the past ~400 yr, each with a total of 0.7–1 m of dip-slip motion.
Summary of SMFZ Trenching Results
Although dating at many of the trench sites is hindered by lack of material given the dating techniques available at the time, results show some similarities within each fault section (Fig. 1B). Dates from the eastern and central portions of the CSMF suggest the most recent earthquake occurred since ca. 8–10 ka and displacement was large (~3–5 m) (Rubin et al., 1998; Tucker and Dolan, 2001; Morton and Matti, 1987). The long open interval is supported by the lack of faulting observed in the latest Pleistocene to Holocene Q2 deposits ubiquitous along the rangefront. Taken together, these observations suggest that the CSMF behavior in the Holocene and latest Pleistocene may be characterized by infrequent, large events.
A ca. 8 ka quiescence on the CSMF differs from results from the west, where at two trench sites on the San Fernando fault (SFF) there is evidence of an earthquake in the past 400 years that preceded the 1971 Sylmar earthquake. Though the records are temporally limited to the past ~500 years, this may indicate the SFF ruptures separately from the CSMF, producing more frequent, moderate earthquakes. Although the records are too short to establish a hard segment boundary between the SFF and the CSMF, this is an important question to test, as a rupture of both the CSMF and SFF simultaneously would impact a much larger region, with potential magnitudes as large as Mw7.3 for a 75-km-long rupture (Fig. 2). Similarly, large ruptures that link the SFF and the Santa Susana fault to the west (Fig. 1A) are suggested by the pattern of fault scarps (Figs. 13, 14) that diverge northwest from the 1971 primary surface rupture at Pacoima Wash (see below).
Drive to Stop 6
From Stop 5, drive down Lincoln Avenue, turn right on Figueroa Drive and then left on N. Windsor Avenue to enter the westbound I-210 Freeway. Continue west on 210 Freeway for 17.5 miles and exit on Maclay Street. Turn right at the bottom of the ramp and proceed northeast on Maclay Street toward the 1971 surface rupture trace and scarp visible in the road near the first intersection (Gladstone Avenue). Continue on Maclay Street and make a left turn on Fenton Avenue to climb from the Qt3 terrace surface onto the higher Qt4 surface. The highest point is the uplifted “Hubbard Hills” hosting Qt6 deposits which are visible northwest down Fenton Avenue. Make a right on Harding Street and proceed along the relatively flat Qt4 surface. The road begins to cut into the Qt4 terrace surface beyond the Eldridge Avenue intersection. At the base of the slope, turn left onto Maclay Street and then make an immediate left into the driveway of the Sylmar Independent Baseball and Softball League. Stop 6 is located in the dirt shoulder on the north side of the paved surface ~150 ft north of the driveway entrance.
Stop 6. Pacoima Wash (34.3101°N, 118.4104°W)
This stop will discuss the San Fernando fault (SFF), 1971 M6.6 earthquake rupture, geomorphic mapping, dating of Quaternary surfaces, and slip rate estimates.
9 February 1971 Mw 6.6 San Fernando Earthquake
The Mw 6.6 San Fernando earthquake occurred on 9 February 1971, producing surface rupture on the SFF (Fig. 1B). The focal mechanism indicates an almost purely dip-slip reverse sense of slip (Mori et al., 1995), which differs from the earlier, more complex, two-fault rupture model by Heaton (1982). Modeling by Mori et al. (1995) imaged the fault plane by relocating aftershocks of the 1971 earthquake, which suggests a single fault plane extending to 15 km depth and dipping ~40° north. The earthquake produced a complex surface rupture ~19 km in length with a maximum displacement of 2.5 m and a significant left-lateral component of slip (Sharp et al., 1975). The net slip distribution curve (Fig. 13) illustrates the variability of slip along strike and an average slip of 0.95 m for the 1971 earthquake (Wesnousky, 2008). This earthquake prompted building code changes and the State of California’s Alquist-Priolo Earthquake Fault Zoning Act of 1972, which focused on regulating development near active faults and addressing the hazard of surface fault rupture by prohibiting the location of human-occupied structures on active fault traces.
Slip Rates on the San Fernando Fault at Pacoima Wash
Lindvall and Rubin (2008) report on three slip rate studies using Quaternary surfaces (Pacoima Wash, Wilson fan, and Lopez fan) completed for the San Fernando fault section of the SMFZ that build on the work of Lindvall et al. (1995), Horner (2006), and Horner et al. (2007). This stop focuses on the results from Pacoima Wash, a large stream that runs along the western edge of a 4 km reentrant in the surface trace of the SFF. At this location, Lindvall and Rubin (2008) mapped six levels of abandoned terraces along the west side of the wash, and completed dating of the Qt4 level, which sits ~19 m above the modern drainage and is offset by three primary scarps and other minor offsets from the 1971 rupture (Fig. 14). Lindvall and Rubin (2008) use contours from a 1935 topographic map with 5-ft contours to estimate 27 ± 1 m of vertical separation in the Qt4 surface (Fig. 15). This vertical separation is considered a minimum since the separations were measured from the uplifted terrace surface on the hanging wall to the modern alluvial surface on the footwall.
Dating focused on Qt4, the most well preserved and largest of the alluvial surfaces mapped adjacent to Pacoima Wash (Figs. 14, 16, 17). Cosmogenic nuclide dating of three cobbles collected from the modern wash (Fig. 14) yielded modern ages, indicating negligible inheritance from the source catchment. An age estimate for abandonment of the Qt4 surface is provided by five 10Be cobble ages (Figs. 14, 16, 17). Of these, two produced a mean age of 66 ka and are inferred to be inherited from older Quaternary deposits in the area, and three produced a mean surface exposure age of 31.5 ka (±0.73 ka) that is more consistent with soil age estimates of ca. 20–30 ka for the Qt4 surface (Lindvall et al., 1995). Using the 27 m displacement since 31.5 ka yields a vertical uplift rate of 0.86 ± 0.4 mm/yr. A dip-slip rate of 1.2 ± 0.2 mm/yr was derived by Lindvall and Rubin (2008) using a 45° dipping fault plane, as this dip was observed at several nearby locations in mapping of the 1971 rupture completed by Kamb et al. (1971). If a horizontal to vertical slip ratio of 1:1 is assumed (which is similar to observations made following the 1971 rupture), then an ~2 mm/yr oblique slip rate is obtained across this >600-m-wide (>2000 ft) zone of faulting (not all of which ruptured in 1971).
The geomorphic mapping in the Pacoima Wash study also highlights a difference between the west-striking 1971 scarps and older scarps (scarps 1, 2, 3 in Fig. 14) which diverge northwestward. These older scarps are well expressed in uplifted deposits of the Hubbard Hills (Fig. 14) but farther northwest are buried by younger alluvium. These features suggest that the 1971 rupture may not be characteristic and that past ruptures may have extended northwest to step or connect with the Santa Susana fault along the older range front at the northern margin of the San Fernando Valley.
Cosmogenic nuclide dating and topographic profiling along a Wilson Canyon fan remnant, located ~3 km west of the mouth of Pacoima Canyon on the older range front (Fig. 1B), yield an additional ~1 mm/yr dip slip rate (Homer, 2006; Lindvall and Rubin, 2008). Summing slip from both the older range front and the younger uplifted surface in the “Hubbard Hills” adjacent to Pacoima Wash suggests ~2 mm/yr of dip slip on the SMFZ in the Sylmar area.
There is strong geologic evidence of late Quaternary to recent motion of the Sierra Madre fault zone (SMFZ), and recent advances in cosmogenic and luminescence dating provide a coherent set of slip rates and suggest similar timing of fan formation at some locations. Offsets of 30–40 ka terraces at four locations along the SMFZ—Pacoima Wash on the San Fernando section, Pickens Fan and Arroyo Seco on the Central Sierra Madre fault, and Day Canyon on the Cucamonga fault—produce concordent slip rates of 1–2 mm/yr along an 85 km span of the fault (Fig. 1; Lindvall and Rubin, 2008; Burgette et al., 2020; McPhillips and Scharer, 2018). This rate is similar to that determined from an older (50–74 ka) fan at Arroyo Seco, while younger, 9–14 ka deposits on the Cucamonga fault and at Dunsmore on the Central Sierra Madre fault may support a slower minimum rate of >0.6 mm/yr (Burgette et al., 2020; Tucker and Dolan, 2001). At face value, the lower Holocene rates suggest a slowing on the fault, however, given >5 k.y. recurrence intervals documented along these sections of the SMFZ (Rubin et al., 1998; Tucker and Dolan, 2001), the lower Holocene rates may reflect variability intrinsic to averaging across only one to three seismic cycles.
Paleoseismic data from the SMFZ are sparse and hint at patterns which deserve more research. Trenching investigations on the Central Sierra Madre fault and the Cucamonga fault indicate long recurrence intervals of >5 k.y. and ruptures with 2 to >5 m of slip per event (Fig. 1). In contrast, the 1971 San Fernando earthquake and paleoseismic trenching at two locations along San Fernando fault suggest that moderate events occur on this section (Bonilla, 1973; Midttun et al., 2015). No similar observations of moderate-sized, recent events have been made on the Central Sierra Madre fault (Crook et al., 1987). Additional trenching investigations are needed to assess if these patterns are persistent over longer timescales. Improved radiocarbon dating and advances in luminescence dating make further study on the size and frequency of SMFZ earthquakes a tractable problem.
This research was funded by internal and external funding from the U.S. Geological Survey under grant nos. G15AP00039, 03HQGR0084, the Southern California Earthquake Center award 15179, and the U.S. Geological Survey–National Association of Geoscience Teachers Cooperative Field Training Program. We thank Dick Heermance, Steve DeLong, and Ryan Gold for reviews of the work and Mike Reed for producing the GIS map of the Sylmar rupture. This research was supported by the Southern California Earthquake Center (Contribution No. 10051). The Southern California Earthquake Center is funded by National Science Foundation Cooperative Agreement EAR-1600087 and U.S. Geological Survey Cooperative Agreement G17AC00047.
Figures & Tables
From the Islands to the Mountains: A 2020 View of Geologic Excursions in Southern California
This volume includes five geologic field-trip guides in the Los Angeles region associated with the 2020 GSA Cordilleran Section Meeting that was scheduled for May 2020, in Pasadena, California. The guides are organized in a generally counterclockwise order around the Los Angeles Basin. The first guide by Burgette et al. provides new slip rates, age constraints, and observations of the active Sierra Madre fault zone that borders the northern side of the San Gabriel and San Fernando Valleys. The Nourse et al. guide takes a new look at the San Gabriel Mountains from a basement and geomorphologic perspective. Further west, Keller et al. provide one of the first published field-trip guides focused on the 9 January 2018 Montecito debris flows that caused 23 deaths. The volume then moves south to Santa Cruz Island, where Davis et al. provide an updated review of the island’s geology within the California borderlands. The final guide returns to the east, where Platt et al. present the unique geology of Santa Catalina Island with a focus on the subduction-related Catalina Schist.