TY - JOUR TI - Developing proxies to reconstruct the intensity and magnitude of prehistoric tropical cyclones and earthquakes DO - https://doi.org/doi:10.7282/t3-m85q-c607 PY - 2020 AB - Coastal risk assessments using observational records and instrumental measurements are insufficient in identifying the potential magnitude and recurrence of tropical cyclones (TC), earthquakes, and tsunamis. Coastal sediments from low-energy depositional environments preserve evidence of prehistoric events as changes in litho- and biostratigraphy. Chapter 1 investigated the sediments deposited by TC Pam at two sites (Manuro and Port Resolution Bay) to provide a modern analogue for reconstructing the long-term variability in tropical cyclone frequency and intensity. TC Pam was a Category 5 intensity tropical cyclone with 10-minute sustained wind speeds as high as 270 km/h that made landfall on Vanuatu. Manuro is a mixed-carbonate embayment with a measured TC Pam flow depth of 4.2 m above mean sea level (MSL) and a minimum inland extent of 400 m. The overwash sediment from TC Pam at Manuro is a coarse- to medium-grained sand that displays a coarsening upward sequence. Port Resolution Bay, a volcaniclastic beach had a measured TC Pam flow depth of 3.3 m above MSL and an inundation distance of 320 m. The overwash sediment at Port Resolution Bay is a coarse- to medium-grained sand that displayed a fining upward sequence in Trench PRB2. The application of laboratory derived settling velocities from the TC Pam overwash sediments were promising. I applied a sediment transport model to reconstruct a field-measured estimate of flow depth at Port Resolution Bay within 11% of the measured value. But future applications of the sediment transport model in mixed-sediment environments will continue to require site-specific calibrations of settling velocities. Chapter 2 examined how the distribution of modern diatoms across tidal wetlands can be used to reconstruct RSL rise caused by Cascadia subduction zone earthquakes of the coast of Washington state. I described and quantified inter- and intra-site variability in diatom distributions and related environmental variables (elevation, grain-size, total organic carbon, porewater salinity, and nutrients) across four shore-perpendicular transects from three tidal marshes of Willapa Bay, Washington. The modern training set from the three marshes consisted of 367 diatoms species across 94 genera that captured the range of diatoms found in marine, brackish, and freshwater environments. Multivariate statistical techniques identified floral zones at each tidal marsh with differing controls by the environmental variables. Two transects (Bone River Transect 1 and the Niawiakum River Transect) showed vertical zonation of diatoms with elevation as the major environmental control (35% and 28% of variance explained, respectively). Vertical zonation was absent at the Naselle River Transect with salinity identified as the major environmental control (21% of variance explained). The combined regional training set and an intra-site comparison of two transects at Bone River also did not display vertical zonation despite overlapping elevations. In the absence of a well-defined relationship with elevation at all marshes due to the localized nature of assemblages, my results suggest that alternative methods must be sought to address limitations in diatom-based RSL reconstructions. Chapter 3 compared the reconstruction of coastal subsidence along the Cascadia subduction zone using an established and new Bayesian diatom-based transfer function. The application of an established transfer function approach (weighted-averaging partial least squares) to six, intertidal subsidence stratigraphic sequences (mud-over-peat contacts) of coseismic subsidence at the Redtail locality in Willapa Bay produced a poor modern analogue for all 77 fossil samples. Using a novel dimension reduction approach, I grouped diatoms based on similarities in abundance and distribution with elevation. A hierarchical clustering method created 12 species groups, which increased the good analogues from 0 to 21%. The groups displayed broad, unimodal, and bimodal response distributions of occurrence across elevation with three groups displaying a high probability of occurrence within a narrow elevation range, which suggested that the grouped approach is more suitable for reconstructing coseismic subsidence. My new Bayesian diatom transfer function integrated the grouped species and stratigraphic prior information (i.e., sediment lithology) to produce reliable subsidence estimates of 0.46 m ± 0.33 m (1700 CE earthquake) and 0.51 m ± 0.34 m (1130 to 1350 yrs CE earthquake) for the two uppermost mud-over-peat contacts. No matching analogues were found for four earlier mud-over-peat contacts, which suggests the modern intertidal environment from the training set does not capture the range of paleoenvironmental conditions that existed in Willapa Bay from ~1400 to 3500 yrs CE. KW - Oceanography KW - Geology, Stratigraphic -- Vanuatu KW - Geology, Stratigraphic -- Washington (State) LA - English ER -