J. Kade Carlson’s research while affiliated with Utah Valley University and other places

How structural segment boundaries modulate earthquake behavior is an important scientific and societal question, especially for the Wasatch fault zone (WFZ) where urban areas lie along multiple fault segments. The extent to which segment boundaries arrest ruptures, host moderate magnitude earthquakes, or transmit ruptures to adjacent fault segments is critical for understanding seismic hazard. To help address this outstanding issue, we conducted a paleoseismic investigation at the Traverse Ridge paleoseismic site (TR site) along the ∼7-km-long Fort Canyon segment boundary, which links the Provo (59 km) and Salt Lake City (40 km) segments of the WFZ. At the TR site, we logged two trenches which were cut across sub-parallel traces of the fault, separated by ∼175 m. Evidence from these exposures leads us to infer that at least 3 to 4 earthquakes have ruptured across the segment boundary in the Holocene. Radiocarbon dating of soil material developed below and above fault scarp colluvial packages and within a filled fissure constrains the age of the events. The most recent event ruptured the southern fault trace between 0.2 and 0.4 ka, the penultimate event ruptured the northern fault trace between 0.6 and 3.4 ka, and two prior events occurred between 1.4 and 6.2 ka (on the southern fault trace) and 7.2 and 8.1 ka (northern fault trace). Colluvial wedge heights of these events ranged from 0.7 to 1.2 m, indicating the segment boundary experiences surface ruptures with more than 1 m of vertical displacement. Given these estimates, we infer that these events were greater than Mw 6.7, with rupture extending across the entire segment boundary and portions of one or both adjacent fault segments. The Holocene recurrence of events at the TR site is lower than the closest paleoseismic sites at the adjacent fault segment endpoints. The contrasts in recurrence rates observed within 15 km of the Fort Canyon fault segment boundary may be explained conceptually by a leaky segment boundary model which permits spillover events, ruptures centered on the segment boundary, and segmented ruptures. The TR site demonstrates the utility of paleoseismology within segment boundaries which, through corroboration of displacement data, can demonstrate rupture connectivity between fault segments and test the validity of rupture models.

The Denali fault in south-central Alaska is a major right-lateral strike-slip fault that parallels the Alaska Range for much of its length and represents the largest seismogenic source for interior Alaska. The fault system is over 1200 km in length, and identification of paleoseismic sites that preserve more than 2–3 paleoearthquakes has proven challenging due to its remote location and difficulty of access. In 2012 and 2015, we developed the Dead Mouse site, which provides the first long paleoearthquake record west of the 2002 Mw 7.9 Denali fault earthquake sequence rupture extent. This site is located on the west-central segment of the Denali fault near the southernmost intersection of the Parks Highway and the Nenana River. We hand-excavated three fault-perpendicular trenches and documented new evidence for seven surface-rupturing paleoearthquakes from deformation in the upper 2.5 m of stratigraphy. Evidence for these events includes offset units, filled fissures, upward fault terminations, and an angular unconformity. Chronological constraints from Bayesian sequence modeling of radiocarbon ages and one tentative tephra correlation indicate these seven earthquakes occurred at 388 cal B.P. (442–319; E1), 807 cal B.P. (853–764; E2), 1282 cal B.P. (1392–1160; E3), 2652 cal B.P. (2805–2460; E4), 3402 cal B.P. (3790–3010; E5*), 5673 cal B.P. (6676–4632; E6*), and 6987 cal B.P. (7281–6668; E7*). Although there are likely missing earthquakes in our chronology prior to E4, the intervals between E1 and E4 suggest significant variability in recurrence period at the Dead Mouse site. Additional paleoearthquake chronologies at neighboring sites are required to make reliable estimates of the spatial and temporal rupture history for the west-central Denali fault, but our data demonstrate the potential for recurrence periods as short as 300–600 yrs, well within range of the current open interval for the Denali fault at the Nenana River.

Recent acquisitions of light detection and ranging (lidar) data along the Wasatch Front, and the proliferation of Structure from Motion (SfM) software present exceptional opportunities for detailed analysis of the Wasatch fault zone (WFZ). Between the Salt Lake City (SLCS) and Provo segments (PS), the WFZ takes a 7-km left bend along the Fort Canyon fault over Traverse Ridge. The Fort Canyon fault at the segment boundary is important because it bridges the most active and populated portions of the Wasatch fault system; the SLCS and PS rupturing together could produce an earthquake approaching Mw 7.5. As part of an effort to identify paleoseismic sites to constrain the timing of ground ruptures and importance of fault segmentation in rupture propagation through the Fort Canyon fault, we used lidar-based digital elevation models and derived products in concert with field and remote investigations to construct a 21-km long fault strip map documenting more than 159 topographic surface breaks that are inferred to be Holocene fault scarps. The segment boundary is more complex than neighboring parts of the SLCS and PS. Within the segment boundary the fault zone is wider, fault scarp lengths are shorter, the fault is more discontinuous, fault orientation is more variable, and scarp density is greater. Additionally, right-steps in the fault trace are associated with negative topography and left-steps are associated with positive topography, which can be explained by a component of right-lateral slip within the segment boundary. Mapping and fault outcrops in natural exposures and paleoseismic trenches demonstrate that Mw 6.6 and greater earthquakes have ruptured with meter-scale displacements through Traverse Ridge multiple times during the Holocene. This study also demonstrates that SfM derived high resolution digital surface models can be utilized to accurately locate even minor fault traces through sparsely-vegetated, semi-arid Basin and Range piedmonts, but also highlights the importance of lidar for mapping faults where they cut through higher, wetter, and more vegetated terrain.

The Fort Canyon fault (FCF) trends east-west across Traverse Ridge, linking the Salt Lake City and Provo segments of the Wasatch fault, which dips directly beneath the densely-populated Salt Lake-Provo metropolitan area. Understanding the history of surface-rupturing earthquakes on the FCF is important because the extent to which segment boundary faults influence the propagation of large normal-faulting earthquakes is a significant source of uncertainty in seismic hazard assessments. To address this issue, we conducted a paleoseismic investigation at the Traverse Ridge site (40.492°, -111.805°). Two trenches were excavated across parallel traces of the FCF. The north trench exposed a fault zone with Tertiary volcanic rock in the footwall and three scarp-derived colluvial wedges. The south trench exposed a fault zone with Tertiary alluvial fan units in the footwall and one distinct colluvial wedge overlying a massive deposit of fault-derived colluvium in the hanging wall. Three of these colluvial wedges have dark soil A-horizons preserved within them and are found beneath steep scarps that are similar to those along adjacent parts of the fault with established Holocene records of rupture. Thus, we infer that 2-3 earthquakes have ruptured through this site during the Holocene. Colluvial wedge heights, ranging from 0.5-1 m, provide a lower limit for vertical displacement in these events. Given this displacement minimum, and considering empirical relationships among rupture parameters and magnitude, we infer that these events were greater than M 6.5. Forthcoming radiocarbon age results will be used to constrain the timing of earthquakes at this site and compare the FCF earthquake chronology to records for the Salt Lake City and Provo segments. This will enable us to assess whether recent ruptures on the FCF were (1) spill-overs from earthquakes on an adjacent segment, (2) floating ruptures centered on the segment boundary, or (3) larger multisegment ruptures.

Segmentation along normal faults is often expressed as a mountain salient that separates fault-bounded valleys to either side. Segment boundaries are areas of lower cumulative slip and thus their faults typically cut through bedrock units making it challenging to determine the age of recent earthquakes. However, bracketing the age of such earthquakes and their kinematics is necessary to establish linkages with earthquake histories along adjacent primary faults. This type of data is important for constraining models of fault interaction and earthquake hazard scenarios. As a part of the 2013 Utah Valley University geology field camp we produced a 1:6K geomorphic fault strip map across the salient separating the Salt Lake City and Provo segments of the Wasatch Fault. Our mapping was based upon 2 m LiDAR data, 0.25 m aerial photographs, and field validation. We identified multiple sub-parallel surface traces across the 7 km segment boundary. Individual fault traces range from 0.2 – 2 km in length and trend roughly east to west. Scarps are generally south facing with less than 10 m of relief. Based upon this mapping, we reoccupied the westernmost of seven large consultant trenches which were left open following a 2006 geotechnical investigation. In this trench we documented three faulted exposures. The central exposure coincided with the most prominent fault scarp. This exposure revealed four colluvial wedges. From oldest to youngest the heights of the wedges are 0.4, 1.2, 0.8, and 1.3 m with uncertainties of about 25%. The northern exposure revealed evidence for one earthquake with 0.5 m of displacement. This event evidence appears to be older than those of the central exposure. The southern exposure revealed evidence for two surface ruptures with 0.4 and 1.0 meters of displacement. We do not know the age of these events relative to those of the central exposure. Each of these wedge deposits consisted of dark, reworked soil a-horizons. We collected bulk samples from each wedge deposit for subsequent radiocarbon analyses. In summary, we documented evidence for at least five different surface rupturing earthquakes with 0.3-1.5 meters of vertical displacement that have broken through the 7 km long boundary between the Provo and Salt Lake City segments of the Wasatch Fault.

Observations of major subduction zone earthquakes and results from numerical modeling of faults have demonstrated that the creeping sections of plate boundary faults can participate in large ruptures. It is therefore important that the geological record of slip along the central creeping section of the San Andreas Fault (SAF) be documented for better characterization of earthquake hazard throughout California and for understanding the range of behaviors of faults with potentially mixed seismic and aseismic strain release. We opened nine trenches in 2012 and 2013 near the center of the creeping section of the SAF at the Dry Lake Valley paleoseismic site (DLV site; 36.46791, -121.05564). Here, the surface trace steps several times, forming sag ponds and pressure ridges within a low-relief alluvial fan complex sloping to the southwest. We trenched both sag pond and fan stratigraphy, revealing a record of deformation extending back more than 1300 years. This record was documented by constructing 3D surface models and orthorectified photomosiacs of the trench walls using structure from motion as implemented in Agisoft Photoscan. The resulting geometric models and high resolution texture maps provide valuable base information for documenting trench exposures. The primary structures observed in these trenches are discontinuous, left-stepping, en-echelon shear bands that are oriented 5-45 degrees NE from the nearby trend of the SAF. These oblique shear bands are localized within multiple 0.2-0.5 m fault zones spanning a 6-14 m width beneath each geomorphic expression of the fault. Across these oblique shear structures, we observed small to modest apparent vertical displacements. Some structures were continuous into the surface stratigraphy whereas other structures terminated beneath the soil A-horizon. Within some of these fault zones, we also observed small cracks, less than 5 cm wide, filled with fine sediment. Although the DLV site does not preserve any unequivocal evidence of large ground rupturing earthquakes within the upper 2-3 meters of fault zone stratigraphy, it does preserve a spectacular record of aseismic fault creep. Filled ground cracks likely indicate that this part of the fault has experienced multiple periods of accelerated fault creep perhaps due to moderate magnitude earthquakes along the creeping section of the fault or due to larger earthquakes along adjacent locked sections of the fault.