Native bunchgrass diversity patterns and phytolith deposits as indicators of fragmentation and change in a California Coast Range grassland

This project investigates how habitat fragmentation might have affected the diversity of native bunchgrasses in 20 isolated grassland patches in the East Bay hills of the San Francisco Bay Area, California. Thought to have been continuous grassland before the early 20th century, the East Bay hills have become a mosaic of urban development, woodland, shrubland, and grassland. Based on island biogeographic theory, the standard fragmentation model posits that when habitat is broken up into small, isolated patches, local populations of native species go extinct, and the smaller and more isolated a fragment, the greater the rate of extinction. The standard model predicts that species richness will positively correlate with fragment area, and negatively correlate with distance from fragment to nearest similar habitat area and with fragment edge-to-interior ratio, due to deleterious edge effects. In the East Bay hills grassland system, only interfragment distance showed the expected degree of correlation, due perhaps to the relatively poor dispersal mechanisms of native bunchgrasses. Possible reasons for the low explanatory power of fragment spatial characteristics include: differential effects of rabbit grazing, differential invasibility of patches, differing substrates, species-specific reactions to edges, differential matrix effects, and other species- and site-specific factors.; Moreover, laboratory work challenged the central assumption underlying my project: that fragments of grassland in the study area are the relictual remnants of formerly continuous perennial grassland. A soil-phytolith analysis, a paleoecological technique, casts serious doubt on this central assumption and, in general, should alert researchers and other ecological workers to be wary of history-based assumptions about past ecosystem conditions or dynamics.; Several lines of evidence are presented to support the proposition that Baccharis-dominated northern coastal scrub may have been the primary vegetation type in the East Bay hills prior to settlement by the Spanish, although tree-dominated types were likely also important.