The composition and physiognomy of forest types are strongly linked to distance inland along the northern California coast

Several forest types co-exist at low elevations in the North Coast Range of California, forming a mosaic at the landscape scale. The major types are: coast redwood forest (dominated by Sequoia sempervirens), Pacific northwest Douglas-fir forest (dominated by Pseudotsuga menziesii), and a broadleaved forest (dominated by any or all of the following: Arbutus menziesii, Chrysolepis chrysophylla, Lithocarpus densiflora, Quercus agrifolia, Q. chrysolepis, and Umbellularia californica; conifers not dominant. Our hypothesis was that the driving forces for this patchy mosaic pattern had to do with a gradient of micro- and macro-attributes from the coast, inland. We focused on the first 60 km of the gradient and we limited any latitudinal effects (i.e., the flora was kept homogeneous) by locating 17 old-growth stands in only seven counties within the boundaries of the Bay Area, sensu lato. The dominance of conifers relative to broad-leaved evergreen trees and of Pseudotsuga relative to Sequoia were analyzed by non-metric multidimensional scaling (NMS), regression, and Canonical Analysis of Principal Coordinates (CAP). The abiotic traits tested for degree of correlation with vegetation were: latitude, distance inland from the sea, slope aspect and steepness, annual precipitation, annual solar radiation, mean January minimum and mean July maximum temperatures, site elevation, and the elevation of an intervening ridge between the site and the coast (relative to site elevation). Some of these traits were significantly correlated with vegetation, the strongest being distance inland, from site to sea. The most parsimonious regression model that explained conifer dominance was a cubic function based only on distance from site to sea, which had an adjusted R-2 of 68 %.