Page 8 - gt1606

Figure 4. DAN3D runout simulation
GSA TODAY | JUNE 2016 At this water level, Sentinel Lake would have covered an area of ~3
km2, extending >7 km upstream (Fig. 2C).
Deposition of sediment behind the rock avalanche dam began
to fill Sentinel Lake. Using modern sediment yields measured at
the nearby East Fork of the Virgin River (Andrews, 2000), rescaled
for the (nearly identical) North Fork drainage area, we calculate
that Sentinel Lake would have filled with sediment in ~600–800
yr. This value matches a previous estimate by Hamilton (1976),
which was generated from different data. Thus, from its formation
at 4.8 ka, we estimate that Sentinel Lake occupied Zion Canyon
for ~700 yr until ca. 4.1 ka. This timing is consistent with
Hamilton’s (2014) OSL age of 4.3 ka, as well as Hamilton’s (1976)
median radiocarbon age of 4.0 cal. k.y. B.P. for post-lake deposi-
tion of sand. Our results are not consistent with UGS radiocarbon
ages, measured from extremely small charcoal samples in cored
lake sediments, which placed the date of lake formation at ca. 8 ka.
Continued incision of the rock avalanche dam has resulted in
erosion of alluvial and lacustrine sediments deposited up to 7 km
upstream (Fig. 3A). In the area of Zion Lodge, the river now lies
~25 m below the highest alluvial sediments associated with the
slide. As local base level continues to drop, this material will ulti- avalanche, with an estimated volume in the range of 50 million m3
mately be lost, and the canyon will return to the steep and narrow
form seen in nearby tributaries. Based on observed rates of post-
slide erosion, this process is expected to require several millennia.
The transient disturbance to the Virgin River drainage basin
created by the Sentinel rock avalanche may thus control upstream
base-level and prohibit fluvial incision of bedrock for >10 ka (cf.
Hewitt et al., 2011).
CONCLUSIONS
The flat valley floor of modern Zion Canyon owes its origin to a
286 million m3 rock avalanche at 4.8 ka arising from the cata-
strophic collapse of a nearly 900-m-high wall of predominantly
Navajo Sandstone. Aided by weak layers in the underlying Kayenta
8
of the Sentinel rock avalanche. Panels
(A–E) show snapshots of the runout at
indicated times. Thickness at 200 s
represents final modeled values and
has been bulked by 28% for compari-
son with mapped and reconstructed
values (F). See the GSA Supplement
Data Repository (see footnote 1) for
animation. Shaded relief map and 30 m
contours show recreated sliding sur-
face topography. Thickness values <2 m
have been omitted for clarity.
Formation, the failure sent material across the canyon at speeds
possibly reaching 90 m/s; some deposits retain portions of their
original stratigraphy although in a highly deformed state. Field
mapping, runout modeling, and cosmogenic nuclide dating
support our hypothesis of single-event, catastrophic emplacement
of rock avalanche debris. Zion Canyon was blocked over a
distance of 3.3 km, damming the Virgin River and creating a lake.
At its high-stand, Sentinel Lake may have covered more than 6
km2, including all of Zion Canyon to The Narrows, but this water
level was short-lived as incision of slide debris lowered the breach
elevation to ~1345 m. There Sentinel Lake stood relatively stable
for ~700 yr until filling with sediment. Today, incision of the rock
avalanche dam continues, lowering the base-level of the Virgin
River and causing erosion of lacustrine and alluvial materials
upstream in Zion Canyon.
Large rock avalanches represent an infrequent but extreme-
magnitude hazard in Zion National Park. Our study helps address
the dynamics and timing of one such event in the densely utilized
area of Zion Canyon; however, deposits of several other large
valley-blocking landslides have been documented in the park
(Hamilton, 2014). One similar event is the Hop Valley rock
and minimum age constrained by radiocarbon dating as ~2.6 k.y.
B.P. The consequences of a similar event occurring today, espe-
cially within the narrow confines of Zion Canyon, could be disas-
trous. Moreover, smaller recent landslides have caused notable
damage; for example, the 1995 Virgin River slide (Figs. 1A and 3A)
affected Zion Canyon’s access road, requiring extensive repairs,
while the 1990 Middle Fork Taylor Creek slide dammed a remote
canyon, which drained suddenly three years later, creating a debris
flow that impacted vehicles on a nearby interstate (Lund et al.,
2010). Recognition of these hazards, combined with detailed field
investigation, dating, and runout analysis, are the first steps
toward mitigating landslide risks.