Multivariate analysis of a suite of sediment characteristics provides a novel, inexpensive, and time efficient approach to characterizing estuarine depositional subenvironments, especially in the absence of diagnostic microfossils and macrofossils. The development, calibration, and implementation of a statistical model are discussed using data collected from the San Francisco Bay estuary, California.
Forty-seven surface sediment samples were collected from three marshes within the San Francisco Bay estuary and from six different subenvironments ranging from tidal salt marsh below mean higher high water (MHHW) to tidal freshwater marsh sediments located above MHHW. Sediment samples were analyzed for percent organic content, percent water content, percent clay content, sediment density, and iron concentration. Standard statistical analysis of the data indicates that changes in sediment characteristics exist within and between the individual marshes; however, individually these characteristics are insufficient to distinguish among marshes and subenvironments. Multivariate analysis therefore is employed to consider the array of sediment characteristics simultaneously. The individual marsh research sites are treated as idealized reference types. The results indicate a strong degree of, and statistically significant, predictive capability. These surface data and analyses are used in the interpretation of stratigraphic cores from three sites in San Francisco Bay. The results using this method were broadly consistent with interpretations using traditional methods; however, this method determined more subtle variability than was previously recognized.