Electronic Supplement to
Holocene Behavior of the Brigham City Segment: Implications for Forecasting the Next Large-Magnitude Earthquake on the Wasatch Fault Zone, Utah

by Stephen F. Personius, Christopher B. DuRoss, and Anthony J. Crone

Section 3. Evaluation of Box Elder delta earthquake chronology of McCalpin and Nishenko (1996) and McCalpin and Forman (2002).

McCalpin and Forman (2002) excavated 14 trenches across a ~300-m-wide zone of multiple fault scarps developed on the surface of a Provo-phase (16.0 ± 1.0 ka) delta complex at the mouth of Box Elder Canyon (Fig. 1). They inferred evidence of 7 or 8 earthquakes (T through Z) since development of the delta, and calculated earthquake times and recurrence intervals from a suite of radiocarbon (mostly "MRT" ages on bulk soil organics), and thermoluminescence (TL) and infrared-stimulated luminescence (IRSL) ages. As with any paleoseismic study of multiple fault traces, conclusions about the number and ages of prehistoric earthquakes were inhibited by uncertain correlations between multiple fault exposures and limited chronological data. Our analysis of their chronology, starting with the most recent earthquake, is presented below. Unit and trench numbers used in the following discussion are from original figures in McCalpin and Forman (2002).

Earthquake Z

The most recent earthquake (MRE) is the best-documented earthquake in the McCalpin and Forman (2002) trenches. Consistent radiocarbon and luminescence ages of 2-2.5 ka from 4 trenches on three fault strands strongly support a younger age for the most recent earthquake than documented in previous studies at Bowden Canyon (Personius, 1991a), and are similar to ages constraining the MRE at Kotter Canyon (event KC1; DuRoss et al., 2012).

Earthquake Y

Three radiocarbon ages from three trenches on two fault strands strongly support identification of earthquake Y at the delta site. The delta age constraints of McCalpin and Forman (2002) are consistent with those of earthquake 3 at Bowden Canyon (Personius, 1991a) and earthquake KC2 at Kotter Canyon (DuRoss et al., 2012).

Earthquake X

Earthquake X was not identified at the delta site, but McCalpin and Forman (2002) inferred that the documentation of this event at Bowden Canyon (Personius, 1991a) and the Pole Patch (Personius, 1991b) was strong enough to include earthquake X in their delta chronology. Given that at least five of McCalpin and Forman's (2002) trenches exposed undated colluvial deposits that conceivably could correlate with earthquake X, we agree that the event likely ruptured one or more scarps on the delta. Our analysis of the delta chronologic data yield a broadly defined age of earthquake X consistent with earthquake 2 at the Bowden Canyon site and possibly earthquake 3 at the Pole Patch site.

Earthquake W

Earthquake W is only constrained by a single radiocarbon age, but the stratigraphic relations in the exposure (trench 14) of McCalpin and Forman (2002) are straightforward and the chronologic data from the trench yield a broadly defined age of Earthquake W consistent with Earthquake 1 at the Bowden Canyon trench site (Personius, 1991a).

Events U and V

Our analysis indicates that the evidence for events U and V is too fragmentary and inconsistent to support the dating of these two earthquakes. For instance, event V is based on a single TL age of 7.5 ± 1.0 ka from a loess-rich soil on deltaic deposits (unit 1) buried by scarp-colluvial deposits (unit 2) in one trench (trench 2) on the 250-m-long, 3.5-m- high scarp D. In contrast to trench 2, a nearly identical stratigraphic sequence in trench 12, 100 m away on the same scarp yielded TL and IRSL ages of 10 ± 1.0 ka and 11 ± 1.0 ka, respectively, from a soil formed on colluvial deposits (unit 2) overlying a presumably correlative loess-rich soil on deltaic deposits (unit 1). McCalpin and Forman (2002) acknowledged this age discrepancy but provided little explanation for their choice to accept the younger age as the close-maximum constraining age for event V in trench 2, while at the same time interpreting the older ages from a similar stratigraphic sequence in nearby trench 12 as possible evidence for older earthquake U (their Figure 32).

Similar age discrepancies are apparent in the dating of event U in trenches across scarp G. McCalpin and Forman (2002) used two nearly identical total-bleach TL ages from trench 3 (8.5 ± 1.5 ka) and trench 6 (8.5 ± 1.0 ka) to date this event, despite voicing misgivings about the significant overlap of the 7.5 ± 1.0 ka and 8.5 ± 1.0-1.5 ka ages that were used as sole justification for delineating V and U as separate events. Additional dating complications include a partial-bleach TL age of 12 ± 1.5 ka and a radiocarbon age of 15.6 ± 1.5 ka on detrital charcoal from the same soil that yielded the 8.5 ±1.0 ka age in trench 6, and discordant TL (12 ± 1.5 ka) and IRSL (9 ± 1.0 ka) ages from the same soil in trench 13 across scarp G less than 200 m south of trench 3. Some of these discrepancies could be caused by multiple surface-faulting earthquakes, but in our opinion, these ages are simply too imprecise to be used alone to identify and date individual earthquakes U and V in the Box Elder delta trenches.

Event T

Identification and dating of event T involves similar chronological inconsistencies, poor exposures, and complications raised in recent studies of the Provo phase of Lake Bonneville (Godsey et al., 2005; Godsey et al., 2011; Benson et al., 2011; Janecke and Oaks, 2011). McCalpin and Forman (2002) use a single radiocarbon age (15.6 ± 1.5 ka) on detrital charcoal from trench 6 to date event T. This sample was collected from a soil formed on scarp colluvium (unit 10) deposited after the earthquake and thus conventionally would be considered a not-so-close minimum-limiting age (e.g. McCalpin, 1996). As previously discussed, this same soil yielded TL ages of 8.5-12 ka that were used to date event U. We found the use of these conflicting ages to closely constrain two different events to be confusing. Additionally, despite the fact that the dated colluvium was apparently deposited subaerially and thus post-dated abandonment of the Provo delta, McCalpin and Forman (2002) concluded that event T occurred while the delta surface was still active. This discrepancy may account for a possible additional, unlettered earthquake between their events T and U mentioned in the abstract but not discussed elsewhere in their report. Our analysis supports a multi-earthquake scenario: abundant evidence (subaqueous? colluvium (trench 10), interbedded colluvium and deltaic deposits (trench 4), folded and liquefied deltaic sediment (trench 3), and a shear-zone termination in deltaic sediment (trench 13) indicates at least one earthquake occurred during active delta deposition, and the subaerial colluvium from which the detrital charcoal sample was obtained indicates a later event that postdates the retreat of the lake from the Provo shoreline about 15 ka. Alternatively, the charcoal could be an accurate minimum age of an earthquake that occurred during a temporary regression from the Provo shoreline (e.g. Benson et al., 2011) sometime prior to delta abandonment. Regardless, the sparse and often conflicting chronological and stratigraphic record presented in McCalpin and Forman (2002) indicates to us that the early part of the paleoseismic record (pre-event W) at the Box Elder delta is poorly constrained and likely incomplete and thus should not be used to calculate recurrence intervals and probabilities of future large earthquakes.

[ Back ]