Statistical analysis of modern seed assemblages from the San Francisco Bay: applications for the reconstruction of paleo-salinity and paleo-tidal inundation

Abstract

The feasibility of developing a seed calibration data set to infer environmental constraints was examined by studying modern seed and vegetation assemblages from 53 stations at three distinct tidal marshes within the San Francisco Bay (salt, brackish and freshwater). Pearson's correlation indicates autochthonous derivation of seeds at the marsh sites. Multivariate statistical analysis of the vegetation and seed data indicates discrete assemblages, which can be used as a proxy for modern salinity and tidal inundation within the estuary. Canonical discriminant analysis indicates that the vegetation and seed assemblages of the three marsh types are statistically distinct. Cluster analysis suggests that marsh vegetation is zoned at the mean higher high water mark at all the sites. The calibration set was then applied to stratigraphic seed data from the estuary to infer changes in paleo-salinity and -tidal inundation during the Holocene. The results of the calibration are compared to earlier reconstructions within the estuary. While the calibrated reconstruction broadly compares to previous reconstructions, a period of inferred lower salinity within the estuary between 3800–2000 cal yr B.P. is calibrated as brackish and not fresh by the calibration model.