5 Discussion
Our findings suggest that the Icy Point, located at the southern end of a restraining double bend in the Fairweather fault, episodically emerged above sea level during repeated fault ruptures that coseismically raised the coast. The faults bounding the peninsula that accommodated the uplift include the Fairweather fault to the east and both the Icy Point-Lituya Bay reverse fault and the Finger Glacier reverse fault to the west. We attribute -46.4 ± 2.4 m of relative sea level fall over the past 7.4 ± 2.0 ka at Icy Point to tectonic vertical land-level change, which implies Holocene rock uplift rates of 4.6 to 9.0 mm/yr. This Holocene rate is consistent with a thermochronometric record that implies Quaternary rock exhumation rates of 5–10 km/m.y. and right-lateral strike-slip rates of 45-49 mm/yr along the Fairweather fault (Lease et al., 2021). Using the slip rate and vertical displacement rate observations we discuss the possible plate-boundary fault geometries and rupture kinematics consistent with our observations in the vicinity of Icy Point.
5.1 Mechanism of late Holocene rock uplift at Icy Point
The 30-to-60-m-tall paleo-sea cliff separating Terrace B from Terrace C and the broad beveled platform fronting the paleo sea cliff (Figure 9), both imply sustained erosion during rapid post-Last Glacial Maximum sea level rise driven by eustacy in the early Holocene. Rapid tectonic uplift at Icy Point may have kept pace with eustatic sea level rise, resulting in very little change in RSL for several millennia. Mann (1986; p. 247) inferred just such a scenario: “Furthermore, wide bedrock terraces on a tectonically rising coast are thought to be cut during periods when rising eustatic sea level keeps pace with uplift (Bradley and Griggs, 1976).” However, after the decay of the large ice sheets ~7 ka, global eustatic sea-level rise decelerated (Fleming et al., 1998) and tectonic uplift at Icy Point outpaced other effects, causing RSL to fall. Terrace emergence resulted from a prolonged interval of marine regression, punctuated by at least 9–12 episodes of coseismic uplift that have left a descending series of paleo-shorelines marked by erosional scarps and barrier beaches across the lower extent of Terrace B (Figure 11).
On tectonically active coasts, tectonic uplift, frequently coseismic, can raise shorelines so that they become stranded above sea level (Wellman, 1969, Ota et al., 1991; Nelson and Manley 1992, McSaveney et al., 2006; Berryman et al., 2018). For example, Turakirae Head on the south end of the North Island of New Zealand features four prominent beach ridges that are vertically separated from each other by 3.4–7.1 m (Aston, 1912; Wellman, 1969; McSaveney et al., 2006). Little et al. (2009) inferred that the elevation difference of adjacent raised ridges at Turakirae Head is a measure of coseismic uplift that caused the stranding of successively higher-elevation shorelines.
In a similar way, we infer that coseismic uplift stranded shorelines at Icy Point. Three observations support this inference. First, as discussed above, the RSL history differs markedly on either side of the Fairweather fault, and we infer that the difference reflects repeated coseismic uplift on the west side of the fault, while little or no tectonic uplift occurred over the same interval east of the fault. Second, the geologic structure of the Icy Point peninsula provides an uplift mechanism (Figure 13): the Icy Point peninsula is bounded on the east by the Fairweather fault, which has a component of west-side-up throw; and Icy Point is in the uplifted hanging wall of the Icy Point-Lituya Bay thrust fault, which bounds the peninsula on the west. Third, the relatively narrow width (<10 km) separating the two faults, which bound the uplift, suggest permanent strain on shallow crustal structures. A series of coseismic uplift events accommodated by these two bounding faults can account for the observed RSL fall over the past 5.5–9.4 ky at Icy Point.
The nearfield kinematics within the restraining double bend of the southern Fairweather fault resemble simple, uniformly uplifted, “pop-up” structures bound by steep (>70°) faults demonstrated in analog models of transpressive systems with 15°–30° convergence angles (Casas et al., 2001). The tilt of Terrace B does not exceed typical shoreline gradients and implies uniform or block-like uplift at Icy Point (Figure 2c). Uniform uplift of a block can be explained kinematically by 1) deformation related to a back-thrusts off the Icy Point-Lituya Bay fault (Pratt et al., 2015), or 2) a shallow (<1 km) depth of the buried tip of the Icy Point-Lityua Bay fault offshore. Better constraints on fault geometries may resolve the kinematics of deformation at Icy Point in greater detail, as both explanations above are consistent with the expression of the faulting in Figure 5a. Comparison with other transpressive, strike-slip fault systems and analog modeling also may lead to insights about the nearfield kinematics along the southern Fairweather fault (e.g., Cowgill et al., 2004; Mann, 2007; Toeneboehn et al., 2018; Benowitz et al., 2021).
5.2 A history of earthquakes that sustain extreme uplift rates at Icy Point
The descending series of 9–12 abandoned shorelines present on Terrace B records repeated episodes of vertical displacement (Figure 11). Field measurements of the vertical step in topography that marked each shoreline ranged from 3 to 5 m (Figure 12). The along shore continuity and step-like geometry indicate sudden vertical displacements, which we attribute to coseismic uplift during earthquakes. The position of the shorelines in the hanging walls of two active reverse faults, the Finger Glacier fault and the offshore Icy Point-Lituya Bay fault provide further support for a coseismic origin. The highest shoreline angle of Terrace B formed 5.5–9.4 ka. Therefore, if each of the 9-12 Terrace B shorelines on Icy Point record coseismic vertical displacement, then the peninsula jerked upwards during a major earthquake every 460–1040 years. This recurrence interval is a maximum estimate because evidence for coseismic uplift events with smaller vertical displacements may not be preserved (or detected) in the coastal geomorphology.
Based on measured elevations of these abandoned shorelines (Figure 12), individual instances of coseismic uplift varied from 3–5 m, with an average of 3.4 m. We infer that coseismic uplift preserves these abandoned shorelines, isolating the former active shoreline and therein triggering formation of a new active shoreline and beach through wave swash processes. The average vertical separation of 3.4 m likely is a maximum, because some earthquakes may have produced <1 m of uplift insufficient to isolate and preserve a shoreline (Pratt et al., 2015), or the shoreline was obscured by beach deposits and not evident in the terrace geomorphology. Further, these values exceed average displacements for reverse and oblique reverse mechanism earthquakes in the historic record (Moss et al., 2022), suggesting the uplift is either focused at Icy Point (a region of maximum uplift) or each shoreline represents one or more events.
Our results imply that past earthquakes accompanied by vertical displacement sufficient to raise and preserve marine shorelines involved a different faulting mechanism than the 1958 Mw7.8 Fairweather earthquake. The 1958 earthquake was predominantly strike-slip and caused no observable uplift at Icy Point. Moreover, the concordance of beach ridge crest elevations on Terrace A and LIA barrier beaches both east and west of the Fairweather fault requires a common history of RSL fall and rules out measurable post-LIA differential uplift. Our geomorphic mapping (Figure 4) shows that the Fairweather fault also delineates a structural eastern boundary of marine and fluvial terrace uplift at Icy Point. Because our analysis of Icy Point terrace slopes (Figure 9) argues for uplift with minimal accumulated tectonic tilt, we infer that movement on both the Fairweather fault and structurally linked reverse faults (e.g., the Icy Point-Lituya Bay thrust fault) (Figure 2 and 10) produces the morphology of Icy Point abandoned shorelines.
5.3 Strain partitioning at the plate-boundary restraining double bend
Icy Point is at the south end of a restraining double bend in Earth’s fastest-slipping, transform plate boundary fault system (Witter et al., 2021). Restraining bends in strike-slip fault systems focus areas of high rock uplift, and the fault geometries that drive uplift at restraining bends are diverse (Cowgill et al., 2004; Cunningham and Mann, 2007; Mann, 2007). Inherited structural complexity is a major determinant in the fault geometry; for example, at restraining bends along the San Andreas fault (Anderson, 1990; 1994; Burgmann et al., 1994; Cowgill et al., 2004; Spotila et al., 2001; Spotila et al., 2007). For the restraining bend along the Fairweather fault north of Icy Point, the contrast in rock types and in rheology across the plate boundary influences the resultant tectonic structures (Molnar and Dayem 2010; Fitzgerald et al., 2014; ten Brink et al., 2018). Amphibolite-facies gneiss and gabbro in the Fairweather Range, east of the restraining double bend, provide a buttress and focus upper-crustal deformation to the west side of the restraining fault bend in the weaker sedimentary rocks of the Cretaceous-Cenozoic Yakutat block (Lease et al., 2021).