Paleoseismologic data from trenches excavated across the Calico fault in the Eastern California shear zone reveal evidence for four surface ruptures during the past ∼9000 yr. Twelve optically stimulated luminescence dates constrain the timing of these surface ruptures, which are defined by the geometry of growth strata, fissure fills, and upward fault terminations, to 0.6–2.0 ka, 5.0–5.6 ka, 5.6–6.1 (or possibly 7.3) ka, and 6.1 (or 7.3) to 8.4 ka. Geomorphologic mapping of the 8 km section of the fault extending southward from the trenches reveals two sets of displacements that record the slip from the past two or three surface ruptures. The slip caused by the most recent event was ∼2.0 m, while the cumulative slip during the penultimate (and possibly the antepenultimate) event was ∼4.5 m. The ages of the paleoearthquakes coincide with periods of clustered moment release identified previously on other faults in the Eastern California shear zone at 0–1.5 ka, 5–6 ka, and ca. 8–9.5 ka, with two Calico fault surface ruptures occurring during the 5–6 ka Eastern California shear zone cluster. These data strongly reinforce earlier suggestions that earthquake recurrence in the Eastern California shear zone is highly clustered in time and space. Such seismic clustering suggests that at least some regional fault networks undergo distinct periods of systemwide accelerated seismic moment release that may be driven by feedbacks between fault-loading rate and earthquake activity.
There is a growing body of evidence that earthquakes cluster at a wide variety of spatial and temporal scales over both individual faults and regional fault networks (e.g., Ambraseys, 1971; Marco et al., 1996; Dolan and Wald, 1998; Rockwell et al., 2000; Friedrich et al., 2003; Dawson et al., 2003; Weldon et al., 2004; Dolan et al., 2007). Documentation of the temporal and spatial scales over which clustering occurs is important for developing tectonic models for accurate seismic hazard assessment (e.g., probabilistic seismic hazard analysis [PSHA]). Moreover, an understanding of the mechanisms for clustering would contribute to a deeper understanding of earthquake physics and the geodynamics of the lithosphere.
The Calico-Blackwater fault system is the longest in the Mojave section of the Eastern California shear zone, a set of predominantly north- to northwest-striking dextral faults that trends northward from the San Andreas fault, across the Mojave Desert, and to the east of the Sierra Nevada (Fig. 1). Recent studies indicate a long-term slip rate for the Calico fault of 1.8 ± 0.3 mm/yr, averaged over the past 57 ± 9 k.y. (Oskin et al., 2007, 2008). Although this represents approximately one third of the total geologic slip rate across the six major faults (i.e., Helendale, Lenwood, Camp Rock, Calico-Blackwater, Pisgah, and Ludlow) that comprise the Mojave section of the Eastern California shear zone, the paleoearthquake history of the Calico fault has not previously been documented.
A compilation of paleoseismologic data from the majority of the faults in the Mojave section of the Eastern California shear zone (but not including the Calico fault) revealed that strain release has been highly episodic over the past 12,000 yr, with pronounced clusters of earthquakes at ca. 8–9.5 ka, 5–6 ka, and during the past ∼1000–1500 yr (Rockwell et al., 2000). The 1992 Mw 7.3 Landers and 1999 Mw 7.1 Hector Mine earthquakes are the two most recent earthquakes of the ongoing, latest Holocene cluster.
The lack of previous paleoseismic studies of the Calico fault thus leads to a simple test of the clustering hypothesis. Because the Calico fault slips faster than other nearby faults, larger or more frequent earthquakes should accommodate its slip. If regional clustering of earthquakes modulates fault-zone activity in the Mojave Desert section of the Eastern California shear zone, then these earthquakes should fall into the regional clustering time periods identified by Rockwell et al. (2000). In this paper, we describe the results of a paleoseismologic trench study and geomorphic analysis of offset features along the northern part of the Calico fault, and discuss the implications of our results for earthquake behavior and seismic hazard assessment.
Our study site is located in the town of Newberry Springs, California, ∼30 km east of the city of Barstow (Figs. 1–3). The ∼N30°W-striking fault trace at the site is well expressed on aerial photographs as a very low-relief (<15–20 cm high), predominantly east-facing scarp and a pronounced vegetation lineament that extend across the playa deposits (Fig. 2). Two parallel fault strands, a prominent eastern strand and a more subtle western strand, are recognizable on the surface. The sediment sources are alluvial fans to the west of the site originating from the Newberry Mountains.
We excavated two fault-perpendicular trenches in the playa where the fault trace was well expressed and where no vegetation was present to obstruct the excavation, or to disturb the youngest strata. The northern trench was 30 m long and 0.9 m wide, with a maximum depth of 3 m. The trench end points were located at 34°49.190′N, 116°39.697′W, and 34°49.182′N, 116°39.715′W. The length of the trench ensured that all fault strands were exposed. A second trench (the “southern trench”) was excavated ∼100 m to the south in order to assess the repeatability of the event chronology logged from the northern trench. The southern trench was 15 m long and 3.5 m deep. The western 11 m of this trench were 0.9 m wide, whereas the 3 m east of the main fault zone were 1.4 m wide due to wall-stability problems. The end points of this trench were located at 34°49.138′N, 116°39.668′W, and 34°49.132′N, 116°39.675′W.
Both trenches exposed fluvial, lacustrine, and playa sediments consisting of well-stratified and thinly bedded pebble gravel to coarse-grained sand, silt, and clay, respectively (Fig. 4). The fault zone exposed in the trenches consists of two strands—the eastern (main) strand and a less-pronounced western strand (Fig. 5). Although motion on the two strands is predominantly strike slip, subordinate vertical components of movement have created a down-dropped block between them. Total vertical separation across the eastern fault at the base of the northern trench is ≥2.5 m, and episodic thickening of growth strata across these faults into the central structural trough helped us to determine the rupture history of the fault. Both the eastern and western fault strands are exposed in the northern trench, but only the main, eastern strand of the fault and the eastern half of the interfault trough are exposed in the southern trench. Major stratigraphic units are numbered in increments of 10, increasing with depth, whereas subunits are labeled using a combination of the respective major stratigraphic unit number plus an alphabetical letter, descending with depth. For example, unit 90 is a pale yellowish-brown coarse-grained sand and gravel, whereas subunit 90A is a fine-grained layer that makes up the uppermost part of unit 90. Detailed descriptions of the stratigraphic units are included in the GSA Data Repository.1
One of the reasons we originally chose the study site was that playas tend to concentrate charcoal suitable for radiocarbon dating (the low-density charcoal fragments stay in suspension until deposited with the last remaining water to be evaporated on the playa). Unfortunately, we were unable to find charcoal in any of the deposits exposed in the trenches, despite intensive efforts to isolate even small quantities through “floating” the low-density charcoal in settling tubes. Fortunately, however, the sediments at the site were well suited to optically stimulated luminescence (OSL) dating, which is used to determine the time that has elapsed since a sediment sample was last exposed to daylight (Table 1) (Aitken, 1998). This technique has been successfully applied to dating deformed sediments from paleoseismic studies in the western United States (e.g., Machette et al., 1992; Crone et al., 1997; Rockwell et al., 2000; Lee et al., 2001; Kent et al., 2005; Wesnousky et al., 2005) and elsewhere in the world (e.g., Owen et al., 1999; Washburn et al., 2001; Wallinga, 2002; Rockwell et al., 2009).
We dated 12 samples using OSL methods. At least two samples were dated from each of the layers bounding each of the event horizons in order to replicate the ages. The sediment samples were collected by pounding 5-cm-diameter, 46-cm-long steel pipes with capped outer ends into the trench wall. The samples remained sealed in their tubes until processing in the safe-light conditions in the Luminescence Dating Laboratory at the University of Cincinnati. Sediment from at least 3 cm from both ends of the tube was removed, and the rest was dried in an oven at ∼90 °C. Approximately 20 g of the dried sediment were ground to a fine powder and sent to the U.S. Geological Survey (USGS) Reactor in Denver for instrumental neutron activation analysis (INAA) to determine the concentration of radioisotopes for dose-rate determination.
The ages of our samples range from 0.6 ka to 8.4 ka, indicating that deposition of all of the strata exposed in the trenches occurred during Holocene time (Fig. 6; Table 1). The oldest sample (081407A) was collected from the southern wall of the southern trench, whereas the rest of our samples came from the southern wall of the northern trench (Table 1). Although the OSL ages indicate that the sediment accumulation rate has averaged ∼0.4 mm/yr over the past 8.4 k.y., the rate has varied significantly over that time (Fig. 7). For example, between ca. 5.6 ka and ca. 5.0 ka, the rate of sediment accumulation increased to ∼2 mm/yr, but subsequently decreased to ∼0.2 mm/yr between ca. 5.0 ka and the present.
INTERPRETATION OF PALEO–SURFACE RUPTURES
The trench exposures revealed evidence for four Holocene surface ruptures on the Calico fault. We refer to these as events 1 (youngest) through 4 (oldest). The three most recent events are well expressed in the northern trench, whereas evidence for the oldest event is present only in the southern trench, in which we exposed deeper stratigraphy. As described in detail next, evidence for these paleoearthquake event horizons includes fissure fills, the geometry of growth strata (e.g., onlap of fault scarps), and upward fault terminations.
Event 1 is best expressed in the northern trench, where it is recorded by the geometry of growth strata and upward fault terminations within unit 30, a dense pinkish-brown clay layer. The best evidence for this surface rupture comes from thickening of unit 30 into the central trough across both the main (eastern) and western fault zones (Fig. 5; Figs. DR1 and DR2 [see footnote 1]). Specifically, unit 30 increases in thickness across the western fault strand from ∼10 cm west of the fault to 30–40 cm in the center of the northern trench. Unit 30 includes two subunits: unit 30A, a denser, more indurated clay, overlies unit 30B, which is slightly more friable. Unit 30B was most clearly defined on the southern wall of the northern trench, extending from meter 8, where it is fault contact with the uplifted older units to the east, across the midtrench trough to the western extent of the trench.
Unit 30B exhibits a relatively constant thickness of 10–15 cm, particularly west of meter 16. In contrast, unit 30A is restricted to the midbasin trough and onlaps unit 30B at meter 10 and meter 22 (Fig. 4). At meter 8, two strands of the main fault offset subunit 30B, but do not cut overlying unit 10. The eastern of these two strands juxtaposes unit 30B against unit 70 clay exposed in the uplifted block east of the main fault strand. Ten centimeters to the west, the second strand exhibits a 10-cm-high, east-side-up vertical separation of the lower part of subunit 30B.
Unit 30 thins markedly westward at the location of the western fault at meter 21.5. The exact stratigraphic level of the upward termination of this oblique reverse fault exposed is not clear; the fault cuts all strata below unit 40, but it cannot be traced confidently upward through the unit 40 soil overprint or into unit 30. However, the relatively constant thickness of subunit 30B from meter 16 to the west end of the trench across this feature argues that subunit 30B was deposited on a near-horizontal surface and has subsequently been folded by slip on the thrust fault. Subunit 30B is composed of playa clay, and such fine-grained strata can drape significant scarps, but even clay units would likely exhibit some thickening across such a pronounced topographic scarp if it had existed prior to deposition of subunit 30B. Taken together, these stratigraphic and structural observations indicate that the most recent surface rupture at the trench site occurred after deposition of subunit 30B and before deposition of subunit 30A, which is interpreted to have largely filled in the postevent structural trough created during event 1.
Thickening of unit 30 into the midtrench trough records sedimentary growth of at least 20–30 cm following event 1. However, the top of unit 30 adjacent to the western fault is not flat, suggesting that growth of unit 30A did not completely fill the trough to its pre-event, horizontal playa surface, leaving a subtle (∼10 cm high on southern wall, ∼30 cm high on northern wall) east-facing scarp. This remaining subtle event 1 scarp was subsequently completely buried during deposition of unit 10, which extends up to the current near-horizontal playa surface. Thus, total vertical separation in event 1 across both strands is probably at least ∼30 cm, and possibly more than 60 cm.
OSL sample 081507A from subunit 30B and samples 051707J and 051707B from subunit 30A constrain the age of event 1. The lower sample yielded an age of 2.0 ± 0.1 ka, whereas the upper samples yielded ages of 0.6 ± 0.1 ka and 1.1 ± 0.1 ka, respectively. Thus, the most recent surface rupture at the trench site occurred after 2.0 ± 0.1 and before 0.6 ± 0.1 (or 1.1 ± 0.1) ka. We note, however, that if there was any undetected partial bleaching of these (very young) samples, this would result in ages that are slightly too old. Whereas the general reproducibility of the ages of the other sample pairs from deeper in the section argues that partial bleaching is not a major problem in the trench, the ∼500 yr difference in the ages of samples 051707J and 051707B, which were collected from the same stratigraphic level only 30 cm from each other, suggests that there may be some minor partial bleaching in some samples. Although we suspect that this effect is relatively minor, it would be most pronounced in the youngest samples. Thus, the OSL ages constraining event 1 should probably be considered maxima.
The upward terminations of the secondary fault strands exposed at meter 9.5 and meter 11 require additional discussion. Both of these faults clearly extend upward to at least the top of unit 40, but they cannot be traced confidently upward into unit 30B. We suspect that this is because the shallowest few tens of centimeters of the trench exposures encompassing units 10–30 exhibited pervasive, irregular cracking. This cracking does not appear to be tectonic in origin, but rather is related to development of a weak, blocky soil texture in these units. Thus, although the stratigraphic contacts in this section could be mapped in detail, the exact upward terminations of the faults could not be unequivocally distinguished from nontectonic cracking within the lower part of unit 30. The possibility that the faults at meter 9.5 and meter 11 do terminate upward at the base of unit 30B suggests the possibility of another event horizon at this stratigraphic level.
If these upward fault terminations do record an additional, pre–event 1 surface rupture, this event must have occurred before deposition of unit 30B at 2.0 ka, and after the 3.4 ka age of unit 40. The 3.4 ka OSL date from unit 40, however, dates the deposition of this unit, and these sediments must have been exposed at the ground surface for a significant period of time prior to deposition of subunit 30B, as evidenced by the presence of the cambic and carbonate soils that developed within unit 40, but which are not evident in unit 30. The observation that the faults at meter 9.5 and meter 11 are clearly discernible to the top of unit 40, through the cambic and carbonate soil overprints, indicates that these faults must have ruptured after most, or all, pedogenesis in unit 40. Thus, if the poorly defined upward terminations of these faults do record a distinct event horizon, this earthquake must have occurred just before deposition of unit 30B at ca. 2 ka. Evidence against the occurrence of a separate event horizon at the base of unit 30B includes: (1) the inconclusive stratigraphic level of the upward terminations of these faults, which may have been obscured by pervasive soil-related cracking of unit 30; and (2) the complete absence of any stratigraphic or structural evidence for significant vertical separations of the unit 40–subunit 30B contact across the faults at meter 9.5 and meter 11. This is in marked contrast to the four surface ruptures that we documented, all of which exhibit significant vertical separations. Moreover, subunit 30B consists of clay that was deposited in a low-energy playa environment, thus making it unlikely that post–unit 40 erosion beveled off any scarps associated with the meter 9.5 and meter 11 faults strands. It is more likely, in our view, that these faults exhibited only minor slip in the most recent event, and are present, but difficult to discern within subunit 30B. Alternatively, many small-displacement faults may not reach all the way to the surface in earthquakes (e.g., Bonilla and Lienkaemper, 1990). Thus, although it is possible that another surface rupture occurred just before event 1, prior to deposition of unit 30B, we think this possibility is unlikely.
Weak evidence for a possible post–event 1 surface rupture comes from east of the main fault, where at least six cracks and small fissures terminate ∼0.1–0.2 m below the present ground surface at the base of, or within, unit 20, to the east of the main fault at meter 8 on the southern wall of the northern trench. Unit 20 is lithologically similar to unit 30, and these units may be correlative. However, unit 20 is only exposed on the eastern side of the eastern fault in our northern trench, so this correlation remains somewhat speculative, and the implication of these cracks (which do not appear to have accommodated any slip) remains unclear. If these cracks are younger than event 1, it is possible that they record cracking of the playa surface during strong ground shaking in other earthquakes (i.e., they did not record actual surface rupture on the Calico fault at our trench site).
The penultimate surface rupture (event 2) is defined in the northern trench by both upward fault terminations and sedimentary growth of units 40 and 50 across both the eastern and western fault strands into the central trough. As described in detail in the stratigraphic section of the Data Repository material (see footnote 1), units 40 and 50 are the same depositional unit. We distinguish unit 40 as a separate unit based on the development of a weak cambic soil that has been overprinted by a stage II carbonate soil. The depositional unit made up of units 40 and 50 consists predominantly of lacustrine, medium- to coarse-grained sand along with several pale-brown clay subunits that are continuous throughout the northern trench. Within unit 50, the presence of several thin, laterally continuous clay layers allows us to define the geometry of sedimentary growth following event 2, as well as the exact stratigraphic position of the event horizon. Unit 40/50, which onlaps eastward against the fault scarp of event 3 (discussed in the following section), thickens eastward across the western fault strand from a thickness of ∼20 cm at meter 26 to a thickness of ∼120 cm at meter 12. The parallelism of the lowermost clay subunits within unit 50 suggests that they were deposited on a relatively horizontal substrate. In contrast, fanning of bedding dips for the overlying clay interbeds at meter 10.5–12.0 records onlap during sedimentary growth above a west-facing scarp that developed along the main (eastern) fault during event 2 (Fig. 5). The geometry of these growth strata indicates that the event horizon for the penultimate surface rupture lies near the middle of unit 50. This observation is supported by minor reverse faulting that occurred during event 2 between meters 21 and 22 along the secondary, western fault strand (Fig. 5; Fig. DR1 [see footnote 1]). This strand displaces the layers below unit 50, as well as the lowest subunits within it, and terminates upward within the upper half of unit 50. Additional evidence in support of event 2 is provided by an upward fault termination at an interbedded clay subunit of unit 50 near meter 24.5 on the southern wall of the northern trench.
Thickening of unit 50 (from a thickness of 20 to a thickness of 120 cm) from meter 26 to meter 12 in the section of the midtrench trough records sedimentary growth of at least 1.0 m following event 2. This stratigraphic growth can be interpreted as a minimum because unit 50 is not exposed east of the main fault at meter 8, and therefore the upper part of unit 50 may have been eroded. Thus, the total vertical separation in event 2 is likely on the order of ∼1.0 m.
We dated one sample (081507H) from the bottom of unit 50 and two samples (081507G and 081507F) from the top of the unit, bracketing the occurrence of event 2. The lower sample yielded an age of 5.6 ± 0.4 ka, whereas the upper samples yielded an age of 5.0 ± 0.4 ka. These OSL dates indicate that event 2 occurred between 5.6 ± 0.4 ka and 5.0 ± 0.4 ka.
Event 3 is well expressed in the northern trench, where it is defined by the geometry of growth strata, upward fault terminations, and multiple fissure fills within unit 70, a pinkish-brown clay (Fig. 5; Figs. DR1 and DR2 [see footnote 1]). The sedimentary growth during event 3 is most pronounced across the main (eastern) fault zone, where unit 70 thickens westward from ∼50 cm at meters 8–11 to >1 m west of meter 11. East of meter 8, unit 70 thins and pinches out eastward by meter 3. The pronounced difference in thickness (0.6 m) of unit 70 between meter 7.5 (east of the main fault) and meter 8.5 (west of the fault) probably reflects erosion of the top of the layer east of the fault. Alternatively, it may reflect strike-slip juxtaposition of a layer of variable along-strike thickness. We think this is less likely, however, because of the near-horizontal morphology of the site, as reflected in the highly planar nature of most of the strata expressed in our trenches. Significant erosion of unit 70 east of the fault is also indicated by the fact that it is overlain directly by the much-younger unit 20. We did not fully expose the unit 70-80 contact across the entire length of the down-dropped, midtrench trough, but unit 70 thins westward from at least 90 cm at its thickest point adjacent to the main (eastern) fault to ∼60 cm west of meter 22 across the western fault strand.
Based on the geometry of several fissure fills and the depth of upward fault terminations, event horizon 3 lies within unit 70. Specifically, the event horizon is marked by the top of a fissure fill at meter 24. The fissure extends downward for ∼70 cm through unit 80 and is filled with clay from unit 70. Moreover, we were able to trace a weakly developed paleosurface for several meters to the east and west of the fissure on the southern wall of the trench at the same stratigraphic level as the fissure fill. A similar fissure fill was observed in the northern wall of the northern trench at meter 10 at the same stratigraphic level (Fig. DR1 [see footnote 1]). Moreover, the occurrence between meter 8 and meter 3 of multiple fissures within units 80C, 80D, 90, and 100 filled with clay from the lower part of unit 70 suggests that the lower part of unit 70 was at the ground surface during event 3. A fault between meter 11 and meter 12 also terminates upward at the same stratigraphic level as the top of the fissure fill. Unit 70 does not thicken appreciably across the western fault zone, implying that this fault strand experienced minor slip during event 3.
Two OSL samples (081507E and 081507D) from the bottom of unit 70 and two (081507C and 081507B) from the top of the unit constrain the age of event 3. The lower samples yielded ages of 6.1 ± 0.4 ka and 7.3 ± 0.5 ka, whereas the upper samples yielded ages of 5.4 ± 0.4 ka and 5.6 ± 0.4 ka. These ages indicate that event 3 occurred before 5.4 ± 0.4 ka (or 5.6 ± 0.4 ka) and probably after 6.1 ± 0.4 ka. We suspect that the 7.3 ± 0.5 ka age, which was collected from the same stratigraphic level as the 6.1 ± 0.4 ka age, may be less reliable because this sample exhibited the lowest dose rate of all the samples dated, perhaps resulting in an anomalously old age. Alternatively, if the 7.3 ± 0.5 ka age is valid, then event 3 occurred between 5.4 ± 0.4 ka and 7.3 ± 0.5 ka.
Event 4, the oldest event horizon we describe, was only observed in the southern trench, where slightly slower sediment accumulation rates allowed us to expose deeper stratigraphic levels (Fig. 4C; Fig. DR2 [see footnote 1]). The event is defined by geometry of growth strata and upward fault terminations. Specifically, if we assume that unit 70, which contains the event 3 horizon in the northern trench, was located near the present ground surface to the east of the main strand in the southern trench (i.e., above unit 90 on the east side of the fault), as seems likely, then the total event 4 vertical separation is ∼1.5 m across the main strand between meter 3 and meter 4 in the southern trench. The large minimum vertical separation across the top of unit 90 suggests that event 4 was a relatively large-displacement event. Event 4 is also defined by two fault strands that terminate upward within unit 80D (at meter 3 and meter 4). Moreover, unit 80D, which developed above unit 90, thins eastward against a preexisting scarp that ruptured the top of unit 90 or the lower part of unit 80D. We interpret this as possible evidence for an older event, although this scarp could also have formed during event 4. The overlying unit 80C pinches out against the west-facing scarp formed during event 4. Unit 80A does not change thickness within the trench exposure west of the fault, suggesting that post–event 4 deposition of unit 80C and the upper part of unit 80D had filled in the down-dropped area to the west of the fault and reestablished a near-horizontal surface. These observations indicate the occurrence of at least one surface rupture (i.e., event 4) either during or immediately after the latter stages of deposition of unit 80D.
One OSL sample (081407A) from the top of unit 90 in the southern trench, and two samples (081507E and 081507D) from the bottom of unit 70 in the northern trench, bracket the age of event 4. The sample from unit 90 yielded an age of 8.4 ± 0.7 ka, whereas the two unit 70 samples yielded ages of 6.1 ± 0.4 ka and 7.3 ± 0.5 ka. These three ages indicate that event 4 occurred before 6.1 ± 0.4 ka (or less likely 7.3 ± 0.5 ka) and after 8.4 ± 0.7 ka. As explained earlier, we suspect that the 7.3 ± 0.5 ka date may be less reliable than the 6.1 ± 0.4 ka date because the older sample exhibited the lowest dose rate of all the samples we dated.
MEASUREMENT OF SMALL-SCALE GEOMORPHIC OFFSETS
In addition to our trench study, we also measured displacements in Holocene alluvial deposits exposed over an 8 km length of the fault extending southward from the trench site. We focused on small-scale displacements (<10 m) in an effort to define slip during the past several surface ruptures of most relevance to our paleoseismologic results. Measured displacements at 16 different locations include horizontal and vertical offsets of channels and alluvial-fan surfaces that range from 1.1 ± 0.2 m to 8.7 +2.3/–4.1 m (Fig. 8; Table 2). Several of these offsets are preserved along a releasing bend of the Calico fault. In some cases, only one (horizontal or vertical) component of oblique slip could be measured, and total offset was inferred from the ratio of horizontal to vertical displacements at nearby sites (Fig. 8). Altogether the measured displacements fall into two sets: one set of six, ranging from 4.6 ± 0.7 m to 8.7 +2.3/–4.1 m, with an average of 6.7 ± 3.0 m, and a second set of 10, ranging from 1.1 ± 0.2 m to 3.0 ± 1.0 m, with an average of 2.0 ± 1.0 m. The presence of at least two different sets of displacements is further implied by the total oblique displacement of 2.1 ± 0.7 m at location 14 and 4.5 ± 0.7 m at location 13 (Fig. 8; Table 2). At these adjacent locations, a fault scarp (location 14) is present in an offset channel inset into a displaced alluvial fan (location 13). This observation suggests that the channel was incised after at least one earthquake that offset the alluvial-fan surface, and that a subsequent earthquake offset the channel further.
Although it is possible that multiple earthquake events are represented within each set of displacements, we did not observe evidence for more than two events in the field. We prefer an interpretation of two to three events, where the set of smaller-displacement measurements (1.1–3.0 m) records slip in a single earthquake, despite the variability of these offsets (∼1 to ∼3 m). Given the overlapping distribution of measurement errors of the smallest set of offsets, this interpretation seems likely, and such along-strike variations in slip at adjacent sites have been noted in surface ruptures on other faults (e.g., McGill and Rubin, 1994).
If we account for an average of 2.0 ± 1.0 m of displacement during the most recent event, the penultimate event had an average slip of ∼4.7 ± 2.0 m. Alternatively, the average displacement of 4.7 ± 2.0 m may record slip from two different surface ruptures (events 2 and 3), which our paleoseismic data indicate happened closely spaced in time. The potentially brief interevent time between events 2 and 3 suggests that new geomorphic features (e.g., gullies) may not have had sufficient time to develop between these earthquakes. In such a situation, the offsets in events 2 and 3 would appear as a single geomorphically defined offset. In summary, we interpret the two sets of displacements (2.0 ± 1.0 m and 6.7 ± 3.0 m) to record the most recent two, and possibly three, surface ruptures on the Calico fault south of Newberry Springs. However, because we lack age control for the features that are offset, these displacements should be considered maxima for the most recent two (and likely three) events revealed by our trenching study.
The paleoseismologic data described here demonstrate that the Calico fault has ruptured to the surface at least four times over the past ∼9 k.y. (Fig. 9). The ages of these surface ruptures strongly support the hypothesis that earthquake recurrence in the Mojave section of the Eastern California shear zone is temporally clustered, as originally proposed by Rockwell et al. (2000). Specifically, the most recent event on the Calico fault occurred after 2 ka, as part of an ongoing cluster of earthquakes in the Eastern California shear zone that has been occurring over the past ∼1000–1500 yr. Moreover, we see no evidence for any Calico fault surface ruptures during the pronounced seismic lull documented by Rockwell et al. (2000) between 2 and 5 ka, despite the continuous, well-dated stratigraphy exposed in our trenches for this time interval. The penultimate and antepenultimate events on the Calico fault, at 5.0–5.6 ka and 5.6–6.1 (or 5.6–7.3 ka), respectively, also occurred during a pronounced cluster at 5–6 ka documented for other faults in the Mojave part of the Eastern California shear zone by Rockwell et al. (2000). Thus, the Calico fault ruptured twice during the penultimate Eastern California shear zone cluster. Finally, although the timing of the antepenultimate cluster documented by Rockwell et al. (2000) is only broadly constrained between ca. 7 and 11 ka, with a peak in moment release at ca. 8–9.5 ka, it appears that our event 4 occurred toward the end of the cluster between 7.3 ± 0.5 and 8.4 ± 0.7 (or 6.1 ± 0.4 and 8.4 ± 0.7) ka (Fig. 9).
Evidence for two surface ruptures between 5 ka and 6 ka indicates that the Calico fault can rupture more than once during each cluster. Thus, although the occurrence of the most recent event after 2 ka suggests that the Calico fault has already ruptured as part of the latest, ongoing earthquake cluster in the Eastern California shear zone, the fault may still be at risk of a near-future earthquake. Furthermore, because no paleoseismic evidence exists for a surface rupture in the past ∼600 yr, it is unlikely that the Calico fault ruptured in 1887, as suggested by Bakun (2005). One explanation for these temporally close ruptures is that this section of the fault was an overlap zone between two ruptures on the southern and northern parts of the Calico fault system. Another possibility is that during the 5–6 ka seismic cluster, strain was released in two events on the Calico fault. At this time, we lack sufficient slip-per-event data to test these alternative rupture scenarios. However, the slip-per-event data gathered thus far suggest that ruptures of the Calico fault are similar in size to earthquakes on nearby shear-zone faults (∼3 m per earthquake; Rockwell et al., 2000). A higher earthquake frequency thus seems necessary to account for the higher slip rate of the Calico fault when compared to other nearby dextral faults in the Eastern California shear zone.
The vertical separations measured in the paleoseismic trenches may provide some insight into the displacements that occurred during the three most recent earthquakes on the Calico fault. If we assume that the slip vector has remained constant at our study site over the past three events, then the vertical components of the total offsets in each event reflect the relative size of each slip event. The most recent event shows a vertical separation of 0.5 m, suggestive of a small overall displacement. In contrast, the penultimate and antepenultimate surface ruptures exhibit larger vertical separations of ∼1 m each, consistent with larger overall displacement in each event. The small vertical separation during the most recent event observed in the trench is consistent with small (∼2.0 ± 1.0 m) displacements measured from nearby geomorphic offsets. The remaining geomorphically measured slip (4.7 ± 2.0 m) may be attributed to either one larger penultimate event, consistent with the larger vertical displacement observed in the trench, or it may represent the penultimate and antepenultimate events, which would imply a similar average slip of 2–3 m in each event. Geomorphic offsets on the order of ∼2–4 m along the Calico fault imply earthquakes ranging from Mw 7.0 to 7.3 (e.g., Wells and Coppersmith, 1994). The occurrence of larger-displacement events 2 and 3 of the penultimate cluster raises the possibility that the apparently small displacement during the most recent event may be indicative of an ongoing slip deficit.
Interestingly, in addition to the current seismic cluster, the Eastern California shear zone appears to be experiencing a strain transient characterized by geodetic rates (100–101 yr) that are faster than the long-term geologic rates (100–106 yr) (Peltzer et al., 2001; Oskin and Iriondo, 2004; Dolan et al., 2007; Oskin et al., 2008). Oskin et al. (2008) documented a cumulative slip rate across the Eastern California shear zone at 35°N of ≤6.2 ± 1.9 mm/yr. In contrast, the rate of right-lateral shear across the entire Mojave Eastern California shear zone measured by global positioning system (GPS) geodesy is 10–14 mm/yr (Savage et al., 1990; Sauber et al., 1994; Dixon et al., 1995; Gan et al., 2000). Thus, there appears to be a significant discrepancy between the short-term rate of elastic strain accumulation measured geodetically and the longer-term geologic slip rates. These observations suggest that the aseismically deforming lower crust and upper mantle beneath the Eastern California shear zone are shearing at a rate that is faster than the long-term average rate documented geologically. Mechanically, this implies that an elevated shear strain rate, likely on a localized, transiently weakened ductile shear zone or set of shear zones, is driving the current cluster of major earthquake activity here (Peltzer et al., 2001; Oskin and Iriondo, 2004; Dolan et al., 2007; Oskin et al., 2008). The mechanisms that permit variable loading rate, and the feedback relationships that may exist among shear-zone strength, fault loading, and earthquake activity, remain unresolved (Montesí and Hirth, 2003; Dolan et al., 2007; Oskin et al., 2008).
Regardless of the exact mechanism that controls the regional clustering of earthquake activity, or whether or not there is a correlation between earthquake clusters and transiently elevated rates of elastic strain accumulation, the paleoseismologic data from the Calico fault trenches strongly support the notion that seismic moment release within the Eastern California shear zone is highly episodic. Similar examples of spatially and/or temporally clustered seismic moment release have been observed elsewhere on both single faults (e.g., Marco et al., 1996; Stein et al., 1997; Hartleb et al., 2003, 2006; Friedrich et al., 2003; Rockwell et al., 2009b) and regional fault networks (e.g., Dolan and Wald, 1998; Rockwell et al., 2000; Dolan et al., 2007). These observations have fundamentally important implications for probabilistic seismic hazard analysis. Specifically, the increasing body of evidence for clustered earthquake occurrence for at least some, and perhaps many, faults and fault systems (including, for example, the Calico fault) suggests that seismic hazard may be strongly affected by temporally and spatially clustered moment release over epochs much longer than the typical decadal time scale of postseismic relaxation.
Paleoseismic trenches demonstrate that the past four surface ruptures on the Calico fault (event 1, 0.6 [or 1.1 ka] to 2.0 ka; event 2, 5.0 ka to 5.6 ka; event 3, 5.6 to 6.1 ka (or 7.3 ka); event 4, 6.1 to 7.3 ka [or 8.4 ka]) occurred during clusters of seismic moment release identified earlier on other faults in the Eastern California shear zone as documented by Rockwell et al. (2000). Two different sets of geomorphic displacements (2.0 ± 1.0 m and 4.7 ± 2.0 m) measured in the 8 km south of the trench site suggest that the past two (or three) of the Calico fault earthquakes were large-magnitude (Mw = 7.0–7.3) events similar in size to the 1992 Mw 7.3 Landers and 1999 Mw 7.1 Hector Mine earthquakes. Although the most recent event we documented in the trenches occurred after ca. 2 ka, during the ongoing Eastern California shear zone seismic cluster, evidence for two Calico fault surface ruptures during the 5–6 ka cluster suggests the possibility that the Calico fault may rupture more than once during each clustering time period. The evidence of seismic clustering on a regional network of faults provided in this paper adds to a growing body of data suggesting that episodic moment release is common on many faults and fault systems, possibly in response to enhanced loading. Future probabilistic hazard analyses need to account for the possibility of temporally and spatially clustered seismic moment release.
We thank Irene Crippen and Diana Williams for permission to trench on their property, Sandy Brittain for help with logistics, and all three for their much-appreciated hospitality and enthusiasm for this research. Eldon Gath, Chris Madden, Tom Rockwell, and Danielle Verdugo provided insightful comments and observations in the trenches. We wish to thank Austin Elliott, Kurt Frankel, Erik Frost, Ozgur Kozaci, Lorraine Leon, Miyu Sato, and Jason Yun, who helped us tremendously with the field work. In addition, we thank James Budahn for running the instrumental neutron activation analysis. An anonymous reviewer provided very thoughtful reviews that improved the manuscript. This research was supported by the Southern California Earthquake Center (SCEC). SCEC is funded by National Science Foundation (NSF) Cooperative Agreement EAR-0106924 and U.S. Geological Survey Cooperative Agreement 02HQAG0008. This is SCEC contribution 1320.