Figure 3. (a) Geology along the Fairweather fault in Glacier Bay National Park (Wilson et al., 2015). A restraining bend between Lituya Bay and Icy Point coincides with the onshore transition from strike slip to oblique slip, which drives uplift west of the fault. (b) Cross section depicting 4–6 km of structural relief across the eastern Yakutat microplate (Plafker, 1971) and geologic contrast with North America (Loney and Himmelberg, 1983). Thermochron AHe data from Lease et al., (2021). Fm.—formation.
Plafker (1971) inferred a major thrust or reverse fault offshore of Icy Point on the basis of well-exposed, steeply overturned bedding in the Topsy and Yakataga Formations:
”The narrow belt of Tertiary rocks south of Lituya Bay is folded into a shallow syncline and a highly asymmetric faulted anticline [and] these folds pass to the southeast into a seaward-facing homocline which is nearly vertical or slightly overturned. The south limb of the anticline is believed to be cut by an unexposed thrust or reverse fault that strikes parallel to the coast.”
Plafker (1967; 1971; 1987) mapped the fault offshore and very near the coast to explain steeply dipping sedimentary rock as young as Pleistocene (Figure 3). This interpretation seems reasonable because such moderate-to-steep uniform dips occur in layered rocks on the hanging walls of reverse faults in other contractional settings, including oblique-slip tectonic environments (Avé Lallemant et al., 1987; Bartley et al., 1990; Namson and Davis, 1990; Suppe and Medwedeff, 1990). Although unable to verify the presence of an offshore fault, Plafker (1967) depicted the ”unexposed thrust or reverse fault” in his 1967 geologic map. He also drew a cross section (Plafker 1971; 1987) showing as much as 4-6 km of structural relief along the northwestern edge of the Yakutat block where it abuts the Fairweather fault (Figure 3).
The contractional deformation that Plafker (1971) inferred to exist offshore to the west of Icy Point is evident in seismic reflection data. Early marine seismic data collected along two northeast-trending track lines offshore 6 km south of Icy Point (Bruns, 1983; Carlson et al., 1985; 1988) show deformed strata in Miocene and younger sediment along strike to the south-southeast of Plaker’s inferred fault. Carlson et al. (1988) identified this inferred fault as the Icy Point-Lituya Bay fault. Further support for the existence of this structure are shallow seismic reflection profiles collected in 2015 and 2017 that show an east-side-up reverse fault offshore to the south-southwest of Icy Point (Balster-Gee et al., 2022a, b) (Figure 2). These seismic data constrain the near shore position of the Icy Point-Lituya Bay fault to within 3.5 km of the coast between Lituya Bay and Crillon Lake (Figure 2).
2.3 Marine Terraces Between Icy Point and Lituya Bay
Where a transpressional fault system occurs in a coastal setting, the ages of uplifted marine terraces can help to constrain rock-uplift rates at millennial time scales (Lajoie, 1986; Kelsey, 2015). Five marine terrace levels preserved along the 42 km of coast between Lituya Bay and Icy Point imply rapid uplift rates over the past 125,000 years (Figure 2) (Hudson et al., 2022; Ugolini and Mann, 1979; Mann and Ugolini, 1985; Mann, 1986). The terrace chronology is poorly resolved, and moraines of Pleistocene age indicate that glaciers advanced over the three oldest marine terraces southeast of Lituya Bay (Terraces C, D, and E) (Mann and Ugolini, 1985; Mann, 1986). The two youngest terraces (Terraces A and B) are largely free of glacial deposits other than moraines of Neoglacial age deposited along the margins of the North Finger and La Perouse Glaciers, and adjacent to Lituya Bay, which was most recently deglaciated shortly before 1700 CE (Mann and Ugolini, 1985). This section reviews the marine terrace geology and places our study into the context of prior work. Interpreting a dynamic glacial landscape on a rapidly uplifting coast requires the compilation of published maps of marine terraces between Lituya Bay and Icy Point, which we overlay onto the 5-m Alaska Digital Elevation Model topography (U.S. Geological Survey, 2016) to correlate terrace surfaces along shore using the naming scheme of Mann (1986) (Figure 2).
Terrace E, originally mapped by Mann (1986), is the oldest marine terrace between Lituya Bay and Icy Point. Deeply incised remnants of Terrace E, with a maximum shore-perpendicular width of ~1 km, extend ~5 km southeast of Lituya Bay (Figure 2). Based on its similar topographic position, Terrace E may have been coeval with the High Terrace present northwest of Lituya Bay (Figure 2b). A stratigraphic section in the High Terrace (Figure 2, Echo section of Mann (1986)) exposes a shore platform at an elevation of ~500 m overlain by several meters of beach sediment, above which are ~10 m of alluvial deposits, which in turn are overlain by 40 m of glacial outwash and then till (Mann, 1986). Wood in peat beds near the top of the alluvial unit yielded an enriched radiocarbon date of >72 ka (Mann, 1986; sample QL-1613). This non-finite 14C date and the fossil pollen flora within the peat led Mann (1986) to suggest that the High Terrace (and hence Terrace E southeast of Lituya Bay) records a RSL high stand during Marine Isotope Stage (MIS) 5e, the Last Interglacial, ca. 115 to 130 ka. Mann (1986) used these age constraints and the 500 m elevation of the High Terrace to estimate tectonic rates of uplift between 4–8 mm/yr.
Terrace D ranges in elevation from 60 to 100 m along the 20 km of coastline southeast of Lituya Bay (Figure 2b). At Icy Point it has a mean elevation of ~125 m and its width ranges from 0.6 to 2.5 km. The ~370-m vertical separation between marine Terrace D and older Terrace E indicates a substantial time passed between the formation of the two shore platforms. Although no chronologic data constrain the age of Terrace D, its much lower elevation, smoother geomorphic expression, and greater coastal extent compared to Terrace E suggest that it was cut during MIS 3 (65-30 ka), which saw several RSL highs stands between 65 and 40 ka (Siddall et al., 2008). Along the Pacific coast of California, several MIS 3 marine terraces date to 40–60 ka (Simms et al., 2015). Mann (1986) estimated the age of Terrace D to be about 60 ka based on the uplift rate inferred for the High Terrace.
Terrace C ranges in elevation from 15 to 40 m within 20 km Lituya Bay and 65 to 100 m at Icy Point. It varies in width from 1.4 to 2.1 km. The minimum-limiting date on this terrace comes from the14C age of wood obtained from an exposure in the glacial trough cut by the southernmost lobe of the Finger Glacier (Figure 2) (Mann, 1986). Here the shore platform of Terrace C crosscuts bedrock structures and is overlain first by several meters of marine sand and gravel. Above the beach deposits are some 50 m of till and outwash interbedded with multiple buried soils. A log found in the lowermost till unit ~10 m above the bedrock surface of the terrace indicates that ice advanced seaward onto Terrace C between 14,150-15,010 cal yr BP (sample Beta-10647; 12430 ± 10014C yr BP) (Mann, 1986). The superposition of the lower till over beach sand deposited on a shore platform implies that Terrace C formed prior to about 15 ka.
Moraines deposited on the surface of Terrace C during ancient fluctuations of the Finger Glacier system provide constraints on the maximum height reached by RSL during the Late Glacial (10–15 ka). Mann (1986) attributed the presence of arcuate “trains of massive, cavernously weathered, erratic boulders” on the surface of Terrace C to a marine transgression occurring sometime after the terrace was formed and either during or shortly after it was overridden by an extensive glacial advance. Today, at Icy Point, these erratic boulder trains on Terrace C are elevated as much as 80 m above modern sea level. Mann (1986) attributed the wave-washed modification of the moraines to erosion and scour by ocean waves when glacial isostatic depression raised RSL at some time after the Last Glacial Maximum. Stratigraphic evidence for ice advance over Terrace C at the Eurhythmic section described by Mann (1986) (Figure 2) and wave modification of recessional moraines suggest that RSL attained a height sufficient to modify moraines on Terrace C sometime after 14–15 ka.
Hudson et al. (2022) and Ugolini and Mann (1979) investigated the two youngest surfaces, Terraces A and B, southeast of Lituya Bay, but no past studies have described in detail the youngest two terraces at Icy Point. Terrace B, a time-transgressive surface (see section 4.2) that varies from 0.3 to 1.1 km wide, extends along 42 km of the coast between Lituya Bay and Icy Point at elevations ranging between 12 and 80 m. Hudson et al. (2022) estimated that Terrace B is 2–3 ka based on the14C ages of the deepest samples of peat accumulated on the terrace. However, Ugolini and Mann (1979) showed that the terrace peatlands form as a result of plant succession and soil development as terrace vegetation changes from beach meadows to forests to peat bogs on younger to older surfaces, respectively. Because accumulation of peat lags the time of terrace emergence, basal peat ages provide only minimum-limiting estimates on the timing of terrace emergence. The age of a “beach-worn” driftwood sampled by Don Miller (blue circle, Figure 2) at the base of a 3 m section of interbedded sand and gravel deposited on a shore platform at an elevation of 46 m indicates that Terrace B emerged before 2960–4060 cal yr BP (Rubin and Alexander, 1958, p. 127; sample W-405, 3250 ± 200 14C yr BP).
The lowest marine terrace southeast of Lituya Bay is Terrace A, a 150 to 675 m wide surface ranging in elevation up to 14 m that borders the present shoreline between Lituya Bay and Icy Point. At a site near the inland limit of Terrace A near Lituya Bay, Ugolini and Mann (1979) estimated the age of the surface at 400 years based on the numbers of annual rings in spruce trees. However, at Icy Point, Terrace A probably post-dates 1750 CE if, as proposed by Mann and Streveler (2008), abandoned beach ridges that define its backedge record a high-sea stand and subsequent glacial isostatic rebound after the Little Ice Age (LIA). During the LIA (1300–1900 CE), the expansion of glaciers in the Fairweather Range and Coast Mountains (McKenzie and Goldthwait, 1971) caused widespread isostatic depression across the region (Motyka, 2003; Larsen et al., 2005). LIA isostatic depression caused ~4 m of RSL rise on both sides of the Fairweather fault (Mann and Streveler, 2008), which between Lituya Bay and Icy Point submerged and eroded the lowest, formerly subaerial portions of Terrace B. Starting ~1750 CE, rapid downwasting of the regional glacier cover triggered rapid isostatic rebound and a fall in RSL that formed Terrace A, which continues today at rates approaching 30 mm/yr in Glacier Bay (Larsen et al., 2005) and 16–18 mm/yr near Icy Point (Elliott et al., 2010).
2.4 The 1958 Fairweather earthquake effects at Icy Point
Post-earthquake surveys of the ground rupture effects happened within days of the 1958 earthquake (Tocher and Miller, 1959; Tocher, 1960). Although most of the rupture occurred under ice or water, windows of ice-free terrain between Crillon Lake and Icy Point revealed evidence for surface deformation consisting of predominantly right-lateral displacement (Tocher and Miller, 1959). Right-lateral offset averaged ~3.5 m along the fault rupture between Crillon and La Perouse Glaciers (Figure 2), where surface displacement could be measured, but one site stood out: northeast of Crillon Lake Tocher (1960) measured 6.6 m of right-lateral and 1.1 m vertical offset where the fault cuts a late Holocene alluvial fan (Witter et al., 2021). However, Tocher noted that west-side-up coseismic motion in 1958 was atypical, “The dip-slip component was evident only at this one locality” (p. 289, Tocher, 1960).
During fieldwork one year later at Icy Point, Tocher (1960) measured ~3 m of right-lateral offset across a shear zone, including 2.4 m of right-lateral offset on a fault striking N38°W and dipping 90°exposed in a “narrow gulch cut into bedrock.” Here again, evidence for substantial coseismic vertical displacement during the 1958 fault rupture is equivocal based on Tocher’s report: “No ground breakage was detected on the terrace surface adjacent to the gulch, even directly in line with the 8-foot (2.4 m) offset.” Because Tocher did not report any vertical offset, we assume that if surface fault rupture at Icy Point included a vertical component, then it was too little to detect.
Here, we aim to reconcile an apparent contradiction between the geomorphic record of late Pleistocene Icy Point uplift and the dominantly horizontal slip observed on the Fairweather fault in 1958. Although the Fairweather earthquake induced little-to-no permanent vertical displacement at Icy Point in 1958, repeated episodes of coseismic vertical displacement are required, as indicated by the emergence of marine terraces that bevel the peninsula and are bounded on the east by the 25-m tall, west-side-up Fairweather fault scarp.