The Big Bend portion of the San Andreas fault currently lacks the type of robust paleoseismic data (slip rate, slip/event, and timing of past ruptures) that have been collected both to the northwest in the Carrizo Plain at Wallace Creek and Bidart, and to the southeast along the Mojave segment at Pallett Creek and Wrightwood. Sites along this reach of the fault near Gorman could provide: 1) an event chronology midway between Pallet Creek and the Carrizo Plain records, 2) the only well-defined slip rate between Carrizo Plain and Cajon Pass, and 3) slip/event data to help test current rupture scenarios and recurrence models.
The purpose of this study was to find potential paleoseismic sites along the Big Bend region of the San Andreas fault, evaluate the merit and potential of each site, and then select the best site for a major trenching effort in 1998. We focused our efforts in the Big Bend region between Cuddy Valley and Lake Elizabeth and found several potential sites that have likely recorded and preserved paleoseismic data that would help us better understand the behavior of San Andreas fault. Of these sites, the Frazier Mountain site offered the most geologic potential, as well as a landowner who was willing to let us perform preliminary excavations in 1997 and continue with more extensive trenching in 1998. We presented the Frazier Mountain Site and a few other potential paleoseismic sites in this region as field trip stops during the 3 day San Andreas Field Workshop held this past September.
The Frazier Mountain Site
The Frazier Mountain site is located on the northern flank of Frazier Mountain between the towns of Frazier Park and Gorman. The site contains a closed depression that has formed in a minor right stepover and blocked drainage within the fault zone (Figure 1). Transfer faults, evidenced by subtle scarps and vegetation lineaments, cut through the sag which is dry in the summer months. This site has the potential to provide the timing of earthquakes, slip rate and possibly slip/event data.
A low scarp crosses through the middle of the depression that we interpreted as a fault. We excavated a pit on the southern side of this probable fault scarp to find water at a depth of about 1.4 m. North of the scarp, we excavated a short trench that extended to the northern margin of the depression and found water at over 3 m. Consequently, we interpret the scarp as a fault (and groundwater barrier) which we chose not to trench across at this time to avoid the shallow groundwater.
The exploratory trench on the north side of the fault exposed well-stratified sediments interrupted by a couple of very weakly developed, 10-20 cm-thick A horizons, or zones of bioturbation. Figure 2 is a rectified photomosaic of a portion of the trench which shows two strands of the fault cutting up to the same paleosurface at about 1.3 m depth below the present ground surface.
The section contained abundant detrital charcoal of which we collected 50 samples for future dating. There were also four thin (1-3 mm) discontinuous peaty horizons that we also collected for potential dating. Some of the charcoal was associated with in situ burn layers and was probably very locally derived. Considering the density of the vegetation where we trenched (we placed the trench in the middle of the dense brush towards the northern end of the sag), it is likely that this area burns periodically (as evidenced by the burn layers) and that much of the charcoal represents growth in the preceding decades prior to the burn. Thus, irrespective of the well-known problems with detrital charcoal ages, we believe that charcoal can provide reasonable ages for the stratigraphy at this site, complemented by the sparse peats and the use of Bayesian statistics.
A couple of initial observations can be made at this time. First, the faults that we exposed areprobably not related to the 1857 rupture because they clearly do not cut up to the surface but rather have an event horizon at about 1.3 m depth. The surface scarp almost certainly formed in 1857 so we know that there are at least two events recorded in the upper 3 m of section here. If there are only two events, then it is likely that the sedimentation rate is high at Frazier Mountain, which is favorable for resolving a relatively long record if we can go deep enough. It also supports our contention that the charcoal and limited peat dates can adequately resolve an event chronology.
Dewatering at this site will be a must. We have now confirmed permission with the landowners and will have their full cooperation for an extensive project if we receive continued funding for this project in 1998. They have given us permission to dig a deep, long drain trench out to the active channel SE of the fan that blocks the depression. We will attempt to lower the water table about 6-7 m during the first phase of the project. That should allow resolution of at least the past 3-5 surface ruptures along this stretch of the fault. In particular, we will be searching for evidence of the 1812 (Wrightwood and Pitman Canyon, and possibly Pallet Creek) or ca. 1700 (the other option for Pallet Creek (1720+/-) and another Wrightwood event (1690+/-)) and the ca. 1470 event that may be similar in extent to 1857. In addition, we hope to find evidence for earlier events, assuming we can get the water table down deep enough.
At the southeast end of the sag pond, a small alluvial fan may have recorded slip in past earthquakes. If the fan contains small channels, it is likely that 3-D trenching can resolve slip per event by matching buried channels within the fan to the source channel across the fault at the fan head. The fan stratigraphy interfingers with the sag sediments so we should be able to tie slip on buried channels to the event stratigraphy in the center of the sag.
A few hundred meters southeast of this fan is a deeply incised stream offset about 75 m across the fault. The drainage enters and exits the fault zone at a high angle and has the potential to yield a Holocene slip rate averaged over a dozen or so earthquake cycles, if the alluvial surfaces surrounding the incision can be dated. Thus, at the Frazier Mountain site, we have the potential of resolving an event chronology, the slip rate, and slip per event for this largely unstudied portion of the San Andreas fault.
In additon to investigations at the Frazier Mountain site, we have been trying to gain access to two sites on Tejon Ranch to conduct preliminary studies. If granted access, we plan to put in exploratory trenches at 2 sites on Tejon Ranch property along Oakdale Canyon Road if funded in 1998. One site is located between the mouths of Robinson and Horse Camp canyons and the other is adjacent to the mouth of Cow Springs Canyon. The site located between Robinson Canyon and Horse Camp Canyon consists of a large area of ponded alluvium against a very steep south facing scarp. The Cow Springs Canyon site consists of a currently dry pond that appears to have formed within a dilational step within the fault zone. These sites along Oakdale Canyon Road have the potential to provide an additional event chronology for the Big Bend section of the San Andreas fault.
