3.8 Sanak section
The Sanak section (Figure 4) extends ~275 km and hosted the 1946 Mw 8.6 Unimak Island earthquake, which spawned a devastating local and trans-Pacific tsunami (Okal & Hébert, 2007). Shallow slip on the megathrust extending to near the deformation front in 1946 is inferred based on teleseismic, tsunami, and aftershock observations (Johnson & Satake, 1997; Okal & Hébert, 2007; Tape & Lomax, 2022). We follow Fournier and Freymueller (2007) in recognizing the Sanak section as distinct from the neighboring Shumagin section based on historical rupture and geodetic observations.
The geologic record of earthquakes is limited to reconnaissance studies on Sanak Island (Engelhart et al., 2015), which had stratigraphic evidence for the 1946 tsunami, a hiatus from 1946 to ~2ka, and then 4 tsunami sand sheets from ~4ka to ~2ka. This implies a mean tsunami recurrence of ~1,000 +/- 665 years spanning ~4ka to 1946 (assuming equal intervals between 4 ka and 2 ka) (Table 1), presuming that the older sand sheets are due to a proximal source and therefore analogous to the sand sheet associated with the 1946 rupture.
Geodetic observations in the Sanak section (Fournier & Freymueller, 2007; Freymueller & Beavan, 1999; Li & Freymueller, 2018) imply a nearly freely slipping plate interface. We assign 2% coupling extending ~50 km from the deformation front, with a locked depth on the interface of ~20 km (Figure 4). In the western part of the Sanak segment, geodetic data are restricted to the Alaska Peninsula and Unimak Island, as no islands are offshore in the forearc, and thus geodetic model resolution for the offshore region is very poor. It is likely that the high slip region of the 1946 earthquake (Lopez & Okal, 2006) occurred mainly in the western part of this section, where geodetic constraints on near-trench slip deficit are minimal. However, to date no source model for the 1946 earthquake has been presented that is also consistent with the lack of observed geodetic strain.