Date of Award

2019

Degree Type

Dissertation

Degree Name

Doctor of Philosophy in Biological and Environmental Sciences

Specialization

Earth & Environmental Sciences (EES)

Department

Geosciences

First Advisor

Simon Englehart

Abstract

This study aims to supplement the paleogeodetic database of past Cascadia subduction zone earthquakes and further our understanding of the Cascadia subduction zone seismic and tsunami hazards. I first address a previously identified spatial gap within the Cascadia paleogeodetic database in southern Cascadia by refining the timing and magnitude of Cascadia subduction zone earthquakes over the past 2000 years in northern Humboldt Bay, California (~44.8°N, -124.2°W). There, I mapped wetland stratigraphy consistent with past megathrust earthquakes across three marshes; Jacoby Creek, McDaniel Creek, and Mad River Slough. To improve the existing paleoseismic chronology at northern Humboldt Bay, I employed Bayesian age modeling based on 21 minimum and maximum limiting ages of short-lived plant macrofossils. These AMS ages found above and below subsidence contacts coupled to the construction of Bayesian age models provide the tightest age distributions for stratigraphic evidence of plate boundary earthquakes along the southern Cascadia coastline over the last 2 ka. These subduction zone earthquakes are dated to CE 1700, ~870 cal yrs BP, ~1125 cal yrs BP, and ~1600 cal yrs BP. I also provide estimates of coseismic subsidence of 0.90 ±0.46m for the 1700 earthquake, 0.39±0.33 m for the ~870 cal yr BP earthquake, 0.99±0.44 m ~1125 cal yr BP earthquake, and ≥0.86 m for the ~1600 cal yr BP earthquake using a validated foraminiferal-based Bayesian transfer function (BTF).

To further improve our confidence in the BTF analysis required an evaluation of the stratigraphic and biostratigraphic variability preserved within the wetland stratigraphy across northern Humboldt Bay. Therefore, I compiled a large stratigraphic and biostratigraphic dataset that allowed for inter- and intra-site variability and replicability assessments of foraminiferal BTF coseismic subsidence estimates. I analyzed 26 sediment cores containing the four mud-over-peat contacts; nine for the 1700 contact (average of 0.63 ±0.36 m subsidence), five for the ~870 cal yr BP earthquake (average of 0.39 ±0.35 m), six for the ~1125 cal yr BP earthquake (average of 0.7±0.39 m) and six for the ~1600 cal yr BP earthquake.(≥0.86 m). The estimate for the 1600 cal yr BP earthquake is a minimum because, across the estuary, the contact formed above the upper limit of foraminiferal habitation. Intra-site variability of coseismic subsidence estimates reached a maximum of 0.59 m for the 1700 earthquake at McDaniel Creek. Inter-site averaged coseismic subsidence variability reached a maximum of 0.47 ±0.48 m between McDaniel Creek (0.80 ±0.41 m) and Mad River Slough (0.33 ±0.25 m) for the 1700 earthquake. The maximum inter-site individual core site variability is 0.76 m from McDaniel Creek (1.04 ±0.42 m) to Mad River Slough (0.28 ±0.26 m). Based on the variability results, I recommend a minimum of two to ideally three RSL reconstruction across the same stratigraphic sequence to provide increased confidence in the subsidence estimate.

The lessons I learned from northern Humboldt Bay were then transferred to southwest Washington, where I focused on identifying the variability of subsidence during the CSZ 1700 earthquake at six new sites, spanning 75 km along-strike, within the CSZ 1700 BTF paleogeodetic database; Copalis River, Ocean Shores, Johns River, Smith Creek, Bone River, and Naselle River. I provide eight estimates (two are within-site replicates) of subsidence for the 1700 earthquake that range from 0.39 ±0.37 m at Johns River to 1.52 ±0.51 m at Smith Creek. The seven new estimates of the CSZ 1700 earthquake from northern Humboldt Bay (1) and southwest Washington (6) were integrated into a margin-wide BTF paleogeodetic database (14), which increases the number of measurements by 50% to a total of 20 estimates margin wide. The updated BTF paleogeodetic data are used to constrain new 3-D elastic dislocation models that inform seismic and tsunami hazard assessments. In this work, we develop three alternate 3-D elastic dislocation models that improve modeled coseismic that identify multiple non-unique solutions that fit to the coastal coseismic subsidence estimates. Even with the new coseismic subsidence constraints, a four-patch slip distribution still the best hypothetical solution. Our results highlight the need for additional high-quality subsidence estimates within the remaining spatial gaps, .e.g., between Nehalem-Netarts, such as at Tillamook Bay.

Available for download on Friday, August 13, 2021

Share

COinS