The application of intertidal foraminifera to reconstruct coastal subsidence during the giant Cascadia earthquake of AD 1700 in Oregon, USA

Changes in species assemblages of intertidal foraminifera can be used to estimate the amount of earthquake-related subsidence during plate-boundary earthquakes at the Cascadia subduction zone. The accuracy and precision of foraminiferal methods in paleoenvironmental reconstruction is underpinned by the relations between contemporary taxa and their environment, which are used to calibrate fossil foraminiferal assemblages in sediment sequences. A contemporary training set of surface sediment samples from five intertidal marshes along the Oregon coast was used to determine foraminiferal distributions and prevailing environmental control(s) along elevational transects. Dominant taxa include Balticammina pseudomacrescens, Trochamminita irregularis, Haplophragmoides wilberti, Trochammina inflata, Jadammina macrescens and Miliammina fusca. Unconstrained cluster analysis and detrended correspondence analysis was used to identify two elevation-dependent faunal zones: Faunal Zone I (upland, high marsh, middle marsh) dominated by Balticammina pseudomacrescens, Haplophragmoides wilberti and Trochammina inflata, and Faunal Zone II (low marsh and tidal flat) dominated by Miliammina fusca. Site-specific differences in assemblages at three marshes enabled further subdivision of Faunal Zone I. Zone la is based on one or more of Balticammina pseudomacrescens, Trochammina inflata, Trochamminita irregularis and Haplophragrnoides sp., and Zone Ib on Jadammina macrescens, Haplophragmoides sp., Trochammina inflata and Miliammina fusca. Canonical correspondence analysis (CCA) and partial CCA of the training set from the Nehalem River marsh transect was used to infer that the zonation of foraminifera is elevation-dependent (39% of explained variance). A transfer function was developed to reconstruct sudden changes in relative sea-level during plate-boundary earthquakes in Oregon. The results indicate a robust performance of the transfer function (r(jack)(2) = 0.82) with the error estimate (RMSEPjack = 0.20 m) comparable to local and regional transfer functions from other temperate marshes. To illustrate the potential of the technique, the transfer function was applied to reconstruct subsidence during the AD 1700 earthquake using at Alsea Bay, Oregon. The reconstruction (0.18 +/- 0.20 m) is less than half the subsidence estimate of Nelson et al. [2008. Great-earthquake palogeoesy and tsunamis of the past 2000 years at Alsea Bay, central Oregon coast, USA. Quaternary Science Reviews, 27, 747-768] using their foraminiferal transfer function, perhaps because of differences in taxonomy and the species relationship to elevation. (C) 2009 Elsevier Ltd and INQUA. All rights reserved.