Journal of Geophysical Research: Solid Earth

Evaluation of Wasatch fault segmentation and slip rates using Lake Bonneville shorelines

Authors


Corresponding author: P. W. Jewell, Department of Geology and Geophysics, 115S. 1460 East, Rm. 383, University of Utah, Salt Lake City, UT 84112, USA. (paul.jewell@utah.edu)

ABSTRACT

[1] Analysis of Lake Bonneville shorelines using lidar digital elevation data challenges accepted models of Wasatch fault deformation since the late Pleistocene. While footwall deformation of the Weber segment of the Wasatch fault is consistent with back-rotation of the footwall block and greatest displacement rate toward the center of the segment, shorelines along the footwall of the Salt Lake City segment decrease in elevation toward the interior and are highest at the segment boundaries, an opposite pattern of footwall deformation than predicted for boundaries arresting or strongly inhibiting displacement during earthquakes. The spatial pattern of footwall rebound implies that some of the proposed persistent fault segment boundaries do not stop earthquake ruptures that originate on adjacent fault segments, nor constrain ruptures initiated within the Salt Lake City segment. Net vertical fault displacement at the boundary between the Salt Lake and Provo segments is 16–20 m over the past 16.3–18.5 ka, corresponding to a vertical displacement rate of 0.8–1.2 mm/yr, a net fault slip rate of 2.0–2.8 mm/yr, and horizontal extension rate of 1.8–2.6 mm/yr on the 25° west-southwest dipping fault that forms the southern Salt Lake City segment boundary. Shoreline analysis suggests isostatic rebound caused by a drop in lake level was concentrated during a relatively short (~2000 year) time period following the Bonneville flood at ~16 ka. Lidar-derived topography in conjunction with robust geomorphic datums improves our ability to map deformation associated with lithospheric flexure and faulting while demonstrating the limitation of lacustrine shorelines in this type of analysis.

1 Introduction

[2] The relationship between lateral segmentation of faults and earthquake mechanics is of primary importance when evaluating the fault displacement and ground motion hazards created by large earthquakes [Schwartz and Coppersmith, 1984; Manighetti et al., 2007]. Geometrical and structural “intersegment” boundaries are important in controlling the spatial distribution of fault slip and rupture length during earthquakes. Recent studies of global populations of faults that generate earthquakes Mw ≥ 6.0 suggest that the influence of these “intersegment boundaries” on rupture propagation evolves with the structural maturity of the fault zone [Manighetti et al., 2007]. That is, as fault displacement or “cumulative slip” increases over time, intersegment boundaries become less resistant to rupture propagation, and hence, the probability of coseismic rupturing through intersegment boundaries and into adjacent segments increases. This evolutionary behavior is apparently a fundamental feature of faulting that applies equally to faults of different slip sense, whether normal, strike-slip, or reverse [Manighetti et al., 2007]. Current knowledge of the mechanical process suggests that as displacement accumulates the intersegment boundaries transition from a zone of macroscopically plastic deformation marked by intense and broadly distributed secondary faulting to a zone of narrowly focused faulting that is less resistant to through-going rupturing [King, 1983; Manighetti et al., 2007, 2009].

[3] The evolution of intersegment boundaries from robust to weak rupture barriers as faults accumulate displacement, and hence, “structural maturity” is a critical factor to consider when evaluating the earthquake behavior of large fault zones [Manighetti et al., 2007]. In particular, jogs and bends that segment a fault zone geometrically may or may not be robust barriers to the spreading of ruptures between adjacent segments.

[4] Normal fault zones become structurally segmented, as populations of faults grow and link within the crust [e.g., Crider and Pollard, 1998; Cowie, 1998]. This relationship between the temporal evolution of normal fault zones and structural segmentation applies throughout the world. Structural configurations of this type are found in extensional tectonic terranes worldwide including Greece [Roberts and Jackson, 2002], the Appenines of Italy [Roberts et al., 2004; Chiaraluce et al., 2011], and the Wasatch Fault Zone (WFZ) which marks the boundary between the Rocky Mountain and Basin and Range tectonic provinces of western North America [Bruhn et al., 1987, 2005].

[5] High-resolution elevation data provided by light detection and ranging (lidar) swath mapping has revolutionized the study of the landscape because of its ability to image the Earth's surface beneath vegetation and to quantitatively analyze landforms in great detail over relatively large areas [e.g., Haugerud et al., 2003]. Lidar surveying has proven effective in active tectonics studies where the ability to map beneath vegetation has revealed previously concealed fault scarps [Haugerud et al., 2003] and provided the ability to map fault displacements and determine slip rate over much larger regions than was previously possible [e.g., Oskin et al., 2007; Arrowsmith and Zielke, 2009]. Recent developments include integrating lidar into structural analysis of both bedrock and fault scarps to produce neotectonic maps of active structures [Pavlis and Bruhn, 2011] and establishing the relative ages and offsets of faulted alluvial fans based on the detailed mapping of surface roughness [Frankel and Dolan, 2007; Frankel et al., 2007].

[6] In this study, we consider new evidence for deformation within and adjacent to the Wasatch normal fault zone in Utah that is based on interpretation of a high-resolution lidar elevation survey. This normal zone fault extends for over 400 km along the eastern edge of the Basin and Range Province and extends through the most populated and urbanized part of the state of Utah (Figure 1) [Schwartz and Coppersmith, 1984; Bruhn et al., 2005]. The central part of the Wasatch fault accumulated more than 11 km of vertical displacement [Parry and Bruhn, 1987], marking it as a “structurally mature” fault zone within the global population of normal faults [Manighetti et al., 2007, 2009]. Given the several hundred kilometer length and significant displacement in the central part of the fault zone, we consider the following question: Are the intersegment boundaries robust and persistent barriers to rupturing [Schwartz and Coppersmith, 1984; Wheeler and Krystinik, 1992] or are these relic barriers that have allowed rupturing spread between segments during the Holocene? [Chang et al., 2006; Duross, 2008] Furthermore, we are able to quantitatively address slip rates at some localities as well as the flexural versus rotational behavior of the WFZ since the Late Pleistocene.

Figure 1.

(a) Location of the Wasatch fault zone in central Utah. (b) Details of Wasatch fault segmentation. Lines represent the Weber, Salt Lake City, and Provo segments of the Wasatch Fault zone. Asterisks and numbers show the location of Figures 3, 6, and 7.

2 Geologic and Tectonic Setting

2.1 Wasatch Fault Zone

[7] The Wasatch fault zone (WFZ) cuts through the urban corridor of northern Utah where it poses a significant earthquake hazard (Figure 1). The normal fault zone is ~400 km long from end to end but has a cumulative trace length of ~566 km because of its sinuosity. The fault zone is divided into 10 segments that vary in length from ~30 km to more than 60 km. The largest paleoearthquakes are estimated to have been magnitude Mw 7.5–7.7 based on surface rupture lengths and fault displacements [Schwartz and Coppersmith, 1984]. Intersegment boundaries are located at jogs and bends in the fault zone, and where bedrock salients and ridges extend outward from the mountain front into the valleys [Schwartz and Coppersmith, 1984; Mabey, 1992; Wheeler and Krystinik, 1992].

[8] Vertical tectonic displacements vary along the length of the fault zone, with several kilometers of vertical offset inferred by gravity modeling in several segments [Mabey, 1992]. The Salt Lake City segment originated ~17 Ma, and the fault zone subsequently grew to its present length as normal faults initiated, propagated laterally, and ultimately linked together to form the eastern margin of the Basin and Range extensional province in northern Utah [Parry and Bruhn, 1987; Cowie and Scholz, 1992].

[9] Early paleoseismic investigations along the WFZ suggested that the geometrical segments ruptured independently of one another and that similar fault displacements occurred repeatedly during earthquakes at each trenching site. These observations led Schwartz and Coppersmith [1984] to cite the Wasatch fault zone as an example when defining the characteristic model of earthquake-faulting behavior. Subsequent paleoseismic work significantly increased the number of trenches and reviews of the fault displacement, and age data called into question the original concept that each segment ruptured independently [e.g., Chang et al., 2006; DuRoss, 2008].

[10] Holocene vertical tectonic slip rates between 1 mm/yr and 2 mm/yr have been determined by trenching in the central five segments of the WFZ [DuRoss, 2008]. Longer term rates of vertical tectonic offset are based primarily on apatite (U-Th)/He data collected from the footwall in the central part of the fault zone. These data indicate spatially uniform exhumation along the length of the fault zone at rates of 0.2–0.4 mm/yr during the last ≈ 5 Ma, with the exception of the footwall near the southern end of the Salt Lake City segment where exhumation rates are about double that amount [Armstrong et al., 2004]. According to Armstrong et al. [2004], there is no evidence of reduced long-term exhumation at intersegment boundaries other than at the southern end of the Salt Lake City segment. The intersegment boundary between the Salt Lake City and Provo segments is a down-to-south normal fault that extends from the Wasatch fault zone into the footwall block where it accommodates differential displacement and exhumation of the footwall between the Salt Lake City and Provo segments. The “Fort Canyon” cross fault creates a 7 km wide jog in the WFZ at Corner Creek where it links the southern end of the Salt Lake City segment to the Provo segment [Gilbert, 1928; Bruhn et al., 1987].

[11] The lidar survey used in this project covers the Salt Lake City and part of the Weber segment of the WFZ (Figure 2). The Salt Lake City segment extends for ~ 35 km between the Weber and Provo segments in the central part of the fault zone. The intersegment boundary at the northern end of the Salt Lake City segment is a large bedrock spur (Salt Lake Salient; Figure 2) that extends several kilometers west from the mountain front and is partially skirted by Holocene fault scarps. The southern end of the segment is located at the 7 km wide jog, where the east-trending cross fault connects the Salt Lake City and Provo segments.

Figure 2.

(a) Location and character of the Provo, Salt Lake City, and Weber segments of the Wasatch fault zone showing prominent structural elements of the Wasatch Front. Lines represent the location of known faults. (b) The same area of Figure 2a, but showing locations of selected faults and specific locations discussed in the text. (c) Gravity survey of the Salt Lake City segment, Wasatch fault zone. Shaded contours west of the mountain front represent thicknesses of Quaternary sediments (1 contour = 200 m) [from Radkins, 1990]. Squares are location of shoreline profiles (Figure 5).

[12] The relatively short length and complex fault trace (Figure 1) distinguish the Salt Lake City segment from other segments of the WFZ. There is also a significant difference in the structure and geomorphology between the northern and southern parts of this segment (Figure 2). In the southern portion, Quaternary fault scarps skirt along the base of the Wasatch Mountains with no antithetic fault zone to the west. In the northern part, the Quaternary scarps angle away from the range front starting at the northern end of Mount Olympus to a position roughly 7 km west of the range front where they curve eastward back toward the mountains (Figure 2b). A broad pediment buried by lacustrine deposits lies between the Holocene fault scarps and the fault-bounded mountain front in this area, and a 14 km long zone of mostly east-facing Holocene fault scarps marks the West Valley antithetic fault system in the middle of the valley (Figure 2b).

[13] The Weber segment is considerably longer (60 km), more linear, and less structurally complex than the Salt Lake City segment (Figure 1). The Holocene scarps are located at the base of the Wasatch Mountains throughout the segment and according to Nelson and Personius [1993] and DuRoss [2008] Holocene vertical displacements decrease near the intersegment boundaries. There are no antithetic fault scarps mapped in Quaternary deposits to the west of the WFZ in this segment.

2.2 Footwall Structural Geology

[14] Spatial variations in the structure and tilting of footwall blocks also provide important data when evaluating the segmentation of normal fault zones [Roberts and Jackson, 2002]. The footwall in the southern half of the Weber segment is a rotated fault block, where Tertiary and younger deposits are tilted eastward toward the Morgan Valley normal fault in the interior of the Wasatch Mountains (Figure 3) [Hopkins, 1982; Yonkee, 1992]. The amount of eastward rotation decreases up-section from a maximum of ≈ 35° in the oldest Tertiary deposits to, as we shall show in a subsequent section, a degree or less measured on late Pleistocene shorelines. The dip of the Morgan Valley fault decreases with depth, and the fault may merge into the WFZ beneath the western side of the Wasatch Mountains. Cumulative throw across the Morgan Valley fault is about 4 km [Royse et al., 1975; Hopkins, 1982; Yonkee, 1992]. The fault is divided into three sections, with scarps with the most recent movement < 15 ka along the central section, and < 750 ka along the northern and southern sections (Figure 3) [Black et al., 2004].

Figure 3.

Location of the Weber segment, Wasatch fault zone, and Morgan Valley faults. The thick line of the Morgan Valley faults represents a fault with the most recent movement dated at less than 15,000 years; other Morgan Valley faults range in age from 0.75 to 1.6 Ma. Squares are location of shoreline profiles (Figure 5).

[15] The Morgan Valley fault dies out northward at an interbasin ridge that separates Morgan Valley to the south from Huntsville Basin to the north (Figure 3). The Morgan Valley fault extends southward toward the latitude of the Weber-Salt Lake City intersegment boundary, where the fault either terminates or splays into several smaller normal faults that are distributed within the footwall [Hopkins, 1982]. Farther south, within the Salt Lake City segment, deformation of the footwall is dominated by crustal flexure rather than block-rotation [Zandt and Owens, 1980].

[16] Flexural deformation within the footwall of the Salt Lake City segment is concentrated within the southern part of the segment, where both geological and thermochronology data indicate ≈ 20° of eastward tilt [Parry and Bruhn, 1987; Armstrong et al., 2004; Stock et al., 2009]. The amount of crustal flexure dies out toward the northern end of the segment, where an unconformity between Oligocene volcanic and older rocks marks the mid-Tertiary land surface that appears to have undergone little if any eastward tilting. This partially exhumed paleoland surface provides a datum for comparing relative amounts of uplift and erosion within the footwall between the northern and southern parts of the Salt Lake City segment, which is at least several kilometers in magnitude [Parry and Bruhn, 1987]. The southward increases in uplift and eastward rotation of the mountain range coincides with an east-trending fault that intersects the WFZ on the north side of Mount Olympus, where the Holocene scarps angle away from the mountain front. This fault also aligns toward the west with a buried ridge that extends across the valley [Mabey, 1992; Radkins, 1990] and marks the southern end of the West Valley antithetic fault system (Figures 2a and 2b). Flexural uplift and eastward rotation of the footwall is truncated abruptly southward by the east-striking fault that extends into the footwall at the intersegment boundary between the Salt Lake City and Provo segments (Figure 2a). The footwall of the Provo segment shows no evidence for significant rotation toward the east by either flexure or rotation by faulting within the interior of the Wasatch Mountains [Bruhn et al., 2005].

2.3 Lake Bonneville

[17] Lake Bonneville existed in the eastern Great Basin from ≈ 10–32 ka and formed the prominent Stansbury shoreline at ≈ 20–22 ka [Oviatt et al., 1990], Bonneville high stand shoreline at ≈ 16.3–18.5 ka [Oviatt and Miller, 2006], and Provo shoreline at ≈ 13.5–16.3 ka [Godsey et al., 2005]. The Provo shoreline formed following the catastrophic Bonneville flood that lowered the elevation of the lake more than 100 m over a period of a few months [O'Conner, 1993]. These shorelines, as well as a number of more poorly documented shore zone features [e.g., Gilbert, 1890] are prominent landforms of the Wasatch Front and much of the eastern Basin and Range province (Figure 4a). Holocene erosion partially covered the shorelines with colluvium and degraded the outer portions of the original shorelines (Figure 4b).

Figure 4.

(a) Field photograph showing Lake Bonneville shorelines taken from Corner Canyon looking northward along the Salt Lake City segment. (b) Generalized features of a typical Lake Bonneville shoreline. (c) Hillshade of lidar digital elevation models (DEMs) from the Corner Canyon area, Wasatch Fault Zone showing possible obscuring of shorelines by alluvial fans.

[18] Shoreline elevations vary throughout the Bonneville basin as a result of isostatic rebound caused by changes in lake level [e.g., Gilbert, 1890] and by post-Bonneville normal faulting. Rebound as great as 70 m in the central portion of the Bonneville basin decreases to displacements of a few meters close to the central WFZ [Currey, 1982; Bills et al., 2002]. Published contours of isostatic rebound in response to withdrawal of Lake Bonneville trend at an acute angle to the Wasatch Mountain front and predict a systematic increase of a few meters elevation from the northern end of the Weber to the southern portion of the Salt Lake City segment [Bills et al., 2002]. We document departures of the elevation of Lake Bonneville high stand shorelines from this simple isostatic rebound pattern and analyze the record of deformation caused by displacements on the Wasatch Fault during the last ≈ 18 ka.

3 Geomorphic Analysis

[19] Data from the 2006–2007 lidar survey of the central Wasatch Front is posted at 1 to 2 m horizontally and a vertical accuracy of 20 cm (http://agrc.its.state.ut.us/). Postprocessing removed vegetation and buildings to produce a “bare-earth” DEM showing landscape features that we interpret as remnants of lacustrine shorelines, glacial and fluvial deposits, and fault scarps.

[20] Extensive urban development along the Wasatch Front has destroyed or modified many natural geomorphic features in the landscape. The lacustrine shorelines are preserved in sections that vary from several 10 s of meters to several kilometers in length. Shorelines that can be mapped in both hanging wall and footwall across the Wasatch Fault are located only at the southern end of the Salt Lake City fault segment where the Traverse Mountains are faulted against the Little Cottonwood stock at Corner Creek (Figure 2b) [Bruhn et al., 2005]. Elsewhere within the segment shorelines are confined to the footwall east of the fault zone. Provo and Bonneville shorelines along the Weber segment are present only in the footwall of the WFZ.

[21] Lacustrine shorelines are part of a shore zone encompassing a 10–15 m vertical elevation range over which erosion and/or deposition acted to produce a gently sloping wave-cut platform as the result of wave action and currents (Figure 4b). Once abandoned due to a drop in lake level, the shore zones were modified by deposition of colluvium, mass wasting, and dissection by gullies (Figure 4c). A typical Bonneville shore zone has an S-shaped profile reflecting shoreward accumulation of colluvium following retreat of the lake [e.g., Keller and Pinter, 1996] and erosional degradation at the foot of the shore zone (Figure 4b) [e.g., Pelletier et al., 2006].

[22] We extracted DEM profiles across the Lake Bonneville shorelines along the entire Salt Lake City segment and most of the Weber segment by examining shorelines at ~50–100 m intervals along strike using ArcGIS software. We decreased the sampling interval where it was possible and critical to have extra data (e.g., along the Draper bench where the only footwall shoreline locations are found; Figure 2). We excluded profiles with irregular slopes or obvious anthropogenic alterations from further analysis. We determined that only Bonneville shorelines are suitable for analysis along the Salt Lake City segment and that both the Bonneville shoreline and small stretches of the Provo shoreline are suitable along the Weber segment. Although multiple Provo shorelines have been documented [e.g., Godsey et al., 2005], we analyzed only the most prominent, uppermost Provo shoreline in this study. A total of 129 profiles extending over ~100 km of shoreline were identified (Figure 5).

Figure 5.

(a) Location maps for the shoreline profiles identified in this study (Figure 2), Draper bench. Universal time meridian (UTM) coordinates and details of the shoreline profiles can be found in the supporting information for this article. (b) Location maps for the shoreline profiles identified in this study (Figure 2), Corner Canyon. UTM coordinates and details of the shoreline profiles can be found in the supporting information for this article. (c) Location maps for the shoreline profiles identified in this study (Figure 2), North Big Cottonwood. UTM coordinates and details of the shoreline profiles can be found in the supporting information for this article. (d) Location maps for the shoreline profiles identified in this study (Figure 2), South Olympus. UTM coordinates and details of the shoreline profiles can be found in the supporting information for this article. (e) Location maps for the shoreline profiles identified in this study (Figure 2), Parleys. UTM coordinates and details of the shoreline profiles can be found in the supporting information for this article. (f) Location maps for the shoreline profiles identified in this study (Figure 2), Red Butte. UTM coordinates and details of the shoreline profiles can be found in the supporting information for this article. (g) Location maps for the shoreline profiles identified in this study (Figure 2), City Creek. UTM coordinates and details of the shoreline profiles can be found in the supporting information for this article. (h) Location maps for the shoreline profiles identified in this study (Figure 2), Warm Springs. UTM coordinates and details of the shoreline profiles can be found in the supporting information for this article. (i) Location maps for the shoreline profiles identified in this study (Figure 2), Bountiful-Centerville. UTM coordinates and details of the shoreline profiles can be found in the supporting information for this article. (j) Location maps for the shoreline profiles identified in this study (Figure 2), Farmington-Kaysville. UTM coordinates and details of the shoreline profiles can be found in the supporting information for this article. (k) Location maps for the shoreline profiles identified in this study (Figure 3), Layton-South Weber. UTM coordinates and details of the shoreline profiles can be found in the supporting information for this article. (l) Location maps for the shoreline profiles identified in this study (Figure 3), Weber. UTM coordinates and details of the shoreline profiles can be found in the supporting information for this article. (m) Location maps for the shoreline profiles identified in this study (Figure 3), Morgan. UTM coordinates and details of the shoreline profiles can be found in the supporting information for this article.

Figure 5.

(continued)

Figure 5.

(continued)

Figure 5.

(continued)

[23] Once a suitable cross-shore zone profile was extracted, we applied a systematic procedure to determine a well-defined, reproducible datum within the shore zone suitable for comparing spatial changes in uplift along the mountain front. We utilized a method to determine ancient shoreline datums that has been successfully applied to other Lake Bonneville shorelines [McCalpin et al., 1992; McCalpin, 1994] as well as Holocene Yellowstone Lake shorelines [Meyer and Locke, 1986; Locke and Meyer, 1994]. The shoreline platform angle is projected beneath the colluvium to an intersection point with the projected wave-cut cliff angle (Figure 4b). We first created a cross-shore profile across the entire shore zone from lidar-derived DEM points (Figure 6a). Note that standard 10 m DEMs derived from photogrammetry do not provide sufficient resolution for rigorous statistical characterization of the shore zone (Figure 6b). We calculated a first-order derivative (slope) of the shore zone lidar DEM points using a moving nine-point set of DEM elevations. We then computed the second-order derivative (slope of the slope) and assumed that very small values (<0.005) indicate planar surfaces of the original shoreline platform (Figure 6c) and the wave-cut cliff (Figure 6d). Each data set used to characterize the shoreline platform and wave-cut cliff contains at least three data points and has a correlation coefficient > 0.99; profiles not meeting these criteria were discarded. We then used the intersection of the two surfaces (Figure 6a) as the datum point for the overall shore zone [McCalpin, 1994; Locke and Meyer, 1994]. Correlation coefficients > 0.99 mean that the errors associated with intersection of the two lines extrapolated beneath the colluvium are much less than 1 m.

Figure 6.

(a) Example Bonneville shoreline profile extracted 2 m lidar-derived DEMs from the southern portion of the Weber segment (location shown in Figure 1). Location of profiles in Figures 6c and 6d are also shown. (b) The same location as Figure 6a, but with 10 m DEMs from the National Elevation Database. (c) Detailed elevation profile of lower, linear portion of the profile in Figure 6a. (d) Detailed elevation profile of upper, linear portion of the profile in Figure 6a.

[24] It is important to emphasize that locating a shoreline datum using this method does not imply that shorelines can be correlated along strike to meter-scale accuracy. Variability in wave run-up, bedrock lithology, and other factors produce variability in the shore zone profile [e.g., Keller and Pinter, 1996]. Indeed, similar data from adjacent modern shorelines on the Pacific coast show scatter of ~2–5 m for reasons that are not entirely clear [Bradley and Griggs, 1976]. Similar scatter are seen in projections of Holocene shorelines of Yellowstone Lake [Locke and Meyer, 1994].

[25] Results of this analysis revealed considerable variation in projected datums in adjacent profiles (Figure 7). While the precise cause of this variation is puzzling, it does not appear to be related to the relatively small errors associated with projecting the shoreline platform and wave-cut cliffs beneath the colluvium. One possibility is that, over 18,000+ years of tectonic activity, small fault offsets or bending and tilting of the shoreline perpendicular to the WFZ have developed and are reflected in the variability of the datum elevations.

Figure 7.

Location of profile shown in Figure 6 (BC-B), an additional Bonneville shoreline profile (BC-C), and two Provo shoreline profiles (BC-P3, BC-P4) showing the highly variable nature of shoreline datum elevations from adjacent profiles. The solid line represents the approximate location of the Bonneville shoreline; the dashed line is the approximate location of the Provo shoreline.

[26] While considerable scatter exists in the shoreline datums of the Weber and Salt Lake City segments, several features stand out. Shoreline elevation gradually increases from the northern end of the Salt Lake City segment to the northern end of the Weber segment with little indication of a distinct boundary between the two segments (Figure 8a). Shore line elevations in the Salt Lake City segment have considerable scatter, but trend from high at both ends of the segment to a minimum elevation at the base of Mount Olympus in the central portion of the segment (Figure 8b).

Figure 8.

(a) Summary of all Bonneville shoreline datums for the Weber and Salt Lake City segments of the WFZ. Filled diamonds represent results of this study; open squares are datums from Currey [1982]. There is no vertical exaggeration. The horizontal scale of both panels is the same. (b) Details of the Salt Lake City segment datums showing the very irregular nature of the segment and offset across the Fort Canyon fault of the WFZ. Vertical exaggeration is ~1.5×.

[27] The spatial pattern of shoreline elevations provides the opportunity to compute fault slip rates in the one locality where the shorelines cross the Wasatch Fault zone, and to map the pattern of footwall deformation over several 10 s of kilometers. In the first case, we explore the slip rate at the intersegment boundary between the Salt Lake City and Provo segments, where there is a 7 km wide jog in the fault zone. This boundary is marked by a narrow zone of faulting that dips between 25° and 30° into the subsurface along the base of the Wasatch Mountain front and directly links the Salt Lake City and Provo segments [Gilbert, 1928; Bruhn et al., 1987, 1992, 2005]. The Traverse Mountains extend west of the mountain front and across the hanging wall valley of the WFZ in this area, providing the opportunity to measure vertical tectonic offset of the Bonneville shoreline across the WFZ at the intersegment boundary (Figure 2).

4 Neotectonic Analysis

4.1 Vertical and Net Fault Slip Rates

[28] The elevation datums in the Traverse Mountains cluster within a spatially limited shoreline segment (~500 m) (Figures 5a and 9d); so no spatial extrapolation was attempted for these data; rather, we calculate an average elevation of 1568.4 ± 1.0 m. We then determined the vertical displacement of the Bonneville shoreline across the Wasatch Fault between this hanging wall datum and the intercept of the trend of the southern portion of the Salt Lake City segment datums (line labeled y = 0.5902× + 1586.1 in Figure 9d) with the Fort Canyon fault near Corner Canyon (1585 ± 0.9 m). Net vertical displacement is thus 16–20 m. The time interval over which this displacement took place is taken to be 16.3–18.5 ka on the basis of the best estimate for the time period during which the Bonneville shoreline existed [Oviatt and Miller, 2006]. Most studies indicate that occupation of this shoreline was for a shorter period of time [e.g., Oviatt et al., 1992] but the 2200 year interval is used in this analysis to provide a conservative range of displacement rates.

Figure 9.

(a) Topographic profile of the mountain crest east of the Salt Lake City segment, Wasatch Front. (b) Three-dimensional perspective of topography along the Salt Lake City segment showing location of faults (solid white) and Bonneville level shorelines (dashed white). (c) Diagrammatic representation of the amount of footwall uplift as a function of distance from the active fault surface. (d) Shoreline datums for selected sections of the Salt Lake City segment, WFZ.

[29] We calculate vertical displacement rates of 0.8–1.2 mm/yr from the time of the Bonneville shoreline to the present. The net fault slip rate and horizontal extension rate are found by resolving the vertical displacement rate onto the 25° dipping fault surface at this locality, yielding a fault slip rate of 2.0–2.8 mm/yr and extension rate of 1.8–2.6 mm/yr. The horizontal extension rate is of particular interest because it can be compared with contemporary geodetic displacement rates determined by GPS surveys across the Wasatch Fault. Chang et al. [2006] determined an extension rate of 1.6 ± 0.4 mm/yr averaged across a 65 km wide band centered on the Wasatch fault, a value somewhat less than the extension rate estimated from Corner Canyon shoreline data. The significance of this observation is that the southern boundary of the Salt Lake City fault segment does not stand out as a site of low horizontal extension rate.

[30] The vertical displacement rate at Corner Canyon is less than that estimated from paleoseismology trenches near the mouth of Little Cottonwood Canyon, ≈10 km to the north. Vertical displacement of 8–10 m dating back to 6 Ma is reported, requiring a vertical displacement rate of 1.3–1.7 mm/yr [Swan et al., 1980; Black et al., 1996; McCalpin, 2002]. This rate is significantly higher than the 0.8–1.2 mm/yr rate at Corner Canyon but may reflect the presence of a more steeply dipping fault rather than a faster slip rate than at the intersegment boundary. Bruhn et al. [1987, 1992] model a fault dip of ≈ 55° at the trenching sites based on an analysis of the strike of the fault at that area, a geometrical construction of fault dips based upon the known strike and dip at the southern segment boundary, and trends of fault section intersection lines. Bruhn et al. [1992] discuss the method and structural data in detail. If fault dips of 55° are accepted for purposes of discussion, the net fault slip rate is 1.2 mm/yr, and the horizontal extension rate is 0.9 mm/yr. These values are significantly lower than at Corner Canyon, suggesting either that the fault may dip less than is modeled by Bruhn et al. [1987] or that there is no deficit in the fault slip rate at the intersegment boundary when compared to the interior of the fault segment.

4.2 Footwall Deformation of the Bonneville Shoreline

4.2.1 Salt Lake City Segment

[31] Remnants of the Bonneville shorelines suitable for mapping deformation along the length of the footwall are located at the northern end of the Salt Lake City segment at City Creek Canyon on the Salt Lake salient, south of the salient at Red Butte Canyon, in the middle of the segment at Parleys Canyon and the base of Mount Olympus, and at Corner Canyon at the southern end of the segment (Figure 8). The shorelines at Parleys Canyon and Mount Olympus are at approximately the same elevations (1570–1580 m asl), but both sites are 10–12 m lower than the shorelines at City Creek Canyon and Corner Canyon (Figures 8b and 9). This pattern of footwall deformation, where the elevation of the shorelines is greater at the intersegment boundaries rather than in the interior of the presumed earthquake rupture segment is opposite to that expected even if the slip distribution is asymmetrically distributed along the length of the segment, with greatest slip localized near one of the segment's ends [e.g., McCalpin, 2009].

4.2.2 Weber Segment

[32] In contrast to the Salt Lake City segment, the Weber segment has a relatively uniform slope in shoreline elevations (Figure 8a). As mentioned previously, the Weber segment is characterized by block fault rotation, which is clearly evidenced by lower Bonneville shoreline elevations that slope gently eastward between the Wasatch and Morgan Valley faults (Figure 10). We also established a small number of datums for the Provo shoreline along the Weber segment (Figure 11). Interestingly, the Bonneville shoreline rises slightly to the north, and the Provo shoreline decreases slightly to the north. Possible implications of this observation are discussed below.

Figure 10.

East-west profile of Bonneville shoreline datums in the Morgan Valley area of the Weber segment, Wasatch fault. Black squares represent profile locations (Figure 5m).

Figure 11.

North-south profile of Bonneville and Provo shorelines, Weber segment, Wasatch fault. The scale of the shoreline datums is the same as the shaded relief view.

5 Discussion

5.1 Footwall Flexure Scales With Fault Displacement

[33] Flexural deformation and back-rotation of the footwall of a normal fault is both observed in nature and predicted by theory. Geodetic measurements following surface rupturing earthquakes and geological mapping of coastlines along the footwall of normal faults suggest that footwall uplift adjacent to the master fault is 10–25% of the coseismic fault displacement [McCalpin, 2009]. These field observations are consistent with geodynamic models of normal faulting [e.g., Bott, 1976], including those created for the Salt Lake City fault segment that explain the distribution of seismicity east of the Wasatch Mountain Front as a result of tectonic flexure and eastward tilting of the footwall [Zandt and Owens, 1980].

[34] The spatial pattern of footwall exhumation and tilting, together with the topography of the Wasatch Mountain crest, mimics that of the Bonneville shoreline elevations (Figure 9). This suggests that the late Pleistocene to Recent style of footwall deformation has been progressing over several million years and is not a transient phenomenon. Parry and Bruhn [1987] noted that the footwall of the Salt Lake City segment was tilted approximately 20° east at its southern end, where cumulative vertical tectonic offset across the fault is ≈ 11 km. In that area, the footwall is composed of uplifted and exhumed Eocene and Oligocene igneous intrusions injected into previously deformed Paleozoic and Precambrian rocks. Footwall tilting and exhumation tapers off significantly toward the northern part of the segment, where the volcanic rocks that originated from the Oligocene stock in the southern part of the footwall are preserved above an angular unconformity east of the range crest. The south to north transition from deeply exhumed to little exhumed footwall terrain is located in the middle of the segment, where the Wasatch Mountain crest forms a broad topographic saddle on the skyline (Figure 9a).

5.2 Fault Structure, Footwall Deformation, and Nonpersistent Segment Boundaries

[35] The structural style and geometry of the Salt Lake City fault segment is distinctive when compared to the other segments in the WFZ. The segment is relatively short (total length ≈ 35 km), varies markedly in strike, and is divided into two sections at a prominent gravity anomaly that extends across the hanging wall basin in the middle of the segment (Figure 2c). This gravity anomaly (1) coincides with the location where the fault zone bifurcates and the Quaternary scarps step outward several kilometers from the mountain front forming a midsegment boundary that may impede rupturing (Figure 2c) [Bruhn et al., 1987], (2) marks the southern end of the West Valley antithetic fault zone (Figure 2a), (3) separates deeper sedimentary fill in the southern part of the hanging wall basin from thinner fill to the north [Radkins, 1990], and (4) abuts the Wasatch Range Front at the low point of the Bonneville shorelines, and where west-striking and steeply north dipping faults extend outward from the mountain front into the hanging wall of the Wasatch Fault zone (Figure 2a).

[36] Does this midsegment boundary mark the termination of multisegment earthquake ruptures? That is, does the northern part of the Salt Lake City segment rupture from time to time with the Weber segment to the north, and the southern part with the Provo segment to the south? This scenario is certainly plausible given observed spatial patterns of Bonneville shoreline elevations in the Salt Lake City fault segment.

[37] We propose therefore that the rupture segmentation model for the Salt Lake City segment should be critically reevaluated. The intersegment boundary between the Provo and Salt Lake City segments is of particular interest because on the one hand it is marked by wide zone of secondary faulting that is distributed within the Traverse Mountains in the hanging wall while also located where at the point of maximum vertical displacement on the Wasatch fault zone [Parry and Bruhn, 1987]. This amount of displacement and the relatively narrow zone of localized faulting that can be mapped continuously between the Salt Lake City and Provo segments suggest in this area the fault zone is “mature” and that although the 7 km wide jog is a profound geometrical feature, it may no longer be a robust barrier to rupture propagation during earthquakes.

[38] The footwall deformation reported herein adds to the body of information from fault dislocation modeling [Chang and Smith, 2002] and paleoseismic geochronology [Chang and Smith, 2002; DuRoss, 2008; DuRoss and Hylland, 2012] suggesting that the intersegment boundaries may not persistently arrest rupture, as was previously assumed [e.g., Schwartz and Coppersmith, 1984]. Probabilistic models of strong ground motion and fault displacement hazard will be significantly impacted if this is the case because of the potential for increased earthquake magnitude during multisegment rupturing and unanticipated temporal behavior of earthquake recurrence along the fault zone. This is the type of fault behavior noted by Manighetti et al. [2007] during their study of earthquake rupturing and fault zone segmentation from a variety of sites around the globe.

5.3 Rotational Behavior of the Weber Segment

[39] Bonneville shoreline datums reveal eastward rotation of the footwall in the central portion of the Weber segment, where the Weber River canyon provided a natural pathway for lake waters to enter Morgan Valley. Total displacement between the Wasatch and Morgan Valley faults is ~24 m, or approximately 1.4 m/km. This is equivalent to a block rotation rate of ≈ 5 × 10−3 degrees per year (Figure 10). The ability to resolve this small but significant eastward rotation of the footwall block using lidar data has at least three significant consequences: (1) recognition that further paleoseismology studies of the Morgan Valley fault and its relationship to the history of rupturing on the Weber segment are warranted, (2) that the Morgan fault breaks the upper crust and must therefore dampen out flexural response of the footwall along part of the Weber fault segment, and (3) that the late Pleistocene and Holocene deformation by block rotation of the footwall is similar to that revealed by the pattern of Late Tertiary sedimentation in Morgan Valley, implying that eastward rotation of the footwall has occurred during for at least 10 Ma, and probably longer.

5.4 Significance of Lidar Data Set

[40] A number of studies have attempted to deal with pre-Holocene fault slip rates along the Wasatch fault using a variety of methods [e.g., Friedrich et al., 2003] and also attempted to test the fault segmentation model over periods of several million years [Armstrong et al., 2004]. The Bonneville shoreline data extracted from the lidar survey complement this other work but also stand apart in terms of the overall spatial distribution, vertical resolution, and temporal control on a well-defined geomorphic feature. Completing this type of work with portable GPS systems in the field would be extremely time-consuming and would lack the robustness provided by the spatial continuity of the airborne lidar data set. This only serves to illustrate that even in regions of limited vegetation, high-resolution airborne lidar surveying is extremely useful for geomorphology and paleoseismic studies. This study also demonstrates the necessity of using fine-scale DEMs produced by lidar to analyze the subtle shore zone features and that even then, considerable scatter in shore zone data are the norm (Figure 8).

5.5 Role of Isostatic Rebound

[41] While the general pattern of isostatic rebound of Lake Bonneville has been known for more than a century [Gilbert, 1890], relatively less is known about the relative roles of rebound and tectonic displacement on shoreline elevations. Currey [1982] published a data base of shoreline elevations that was used as the basis to calculate generalized contours of isostatic rebound for the entire basin [Bills et al., 2002]. Currey's [1982] analysis ignored shorelines per se, concentrating on geomorphic features such as spits, tombolos, beach ridges, and barriers the tops of which he felt more accurately reflected actual maximum water elevations. It is noteworthy that the small number of Currey's data points within the domain of this study fall within the general trend of our shoreline datums (Figure 8a) suggesting a general agreement for establishing paleowater elevations using a variety of geomorphic features and methods.

[42] Published contours of isostatic rebound in response to withdrawal of Lake Bonneville predict a systematic decrease of a few meters elevation from the northern end of the Weber to the southern portion of the Salt Lake City segments [Bills et al., 2002]. This is in general agreement with the trend of Bonneville shoreline datums, but not the Provo datums (Figure 11). One interpretation would be that a significant portion of the lithospheric rebound was complete by the time of the early Provo shoreline occupation, although this idea remains speculative.

6 Conclusions

  1. [43] Lidar digital elevation models (DEM) of Lake Bonneville shorelines challenge the model of Wasatch fault segmentation by both spatial pattern of footwall deformation of shorelines, and our slip rate estimates within the southern Salt Lake City segment boundary.

  2. [44] Bonneville shorelines preserved in the footwall of the Salt Lake City segment of the Wasatch normal fault systematically decrease in elevation toward the middle of the segment. This implies that the presently defined boundaries of the segment, which were previously assumed to rupture independently during Holocene earthquakes, may not persistently arrest ruptures from earthquakes that arise on adjacent segments, nor which initiate within the segment.

  3. [45] The average late Pleistocene to Recent fault slip rate at Corner Creek is 2.0–2.8 mm/yr. This is the first robust estimate of fault slip rate at an intersegment boundary of the Wasatch Fault and implies a horizontal extension rate that is somewhat larger than the contemporary rate measured by GPS geodetic surveying across the entire Wasatch Front. The southern boundary does not appear to be a location of slip rate deficit when the average extension rate for the last ≈ 18 ka is compared with the contemporary rate of extension across the Wasatch Fault.

  4. [46] The southern half of the Weber segment is marked by rotation of the footwall block toward the east based on shoreline elevations changes between the Wasatch Front and the Morgan Valley Fault.

Acknowledgments

[47] The manuscript was improved with the suggestions of Chris DuRoss and James McCalpin, an anonymous associate editor, and two anonymous reviewers.

Ancillary