Pollen is abundant and well preserved in late Neogene sediments deposited on the northern California continental margin, with more pollen at the site closest to California vegetation than at sites farther offshore. Mean pollen abundance at Site 1018 (12,000/gdws) is two and three times larger than at Sites 1022 and 1020, respectively (Fig. 2, Fig. 3, Fig. 4). These differences in the amounts of pollen reflect differential pollen productivity of plant assemblages from which the pollen is derived, as well as pollen sedimentation (e.g., initial pollen dispersal and subsequent fluviomarine transport; Heusser and Balsam, 1977; Heusser, 1983, 1988). At Site 1018, the high concentrations of pine and juniper/cedar types, both from heavy pollen producers, contribute to the large amount of pollen in sediments deposited at this nearshore site (Fig. 2). The hydrodynamic efficiency of buoyant pine pollen is seen in the usual offshore increase in the relative abundance of pine (Traverse, 1988). Mean pine percentages from Leg 167 drill sites off Northern California increase from 37% at ~76 km to 74% at 364 km. The overall increase in pollen concentration at ~4 Ma in Hole 1022C sediments may reflect tectonic movement of the core site, changing the distance between the paleodeposition site and source vegetation (Lyle, Koizumi, Richter, et al., 1997), as well as changes in the vegetation onshore (e.g., increased cedar/juniper in regional vegetation during the last ~3.8 m.y.).
Pliocene and Pleistocene pollen spectra are composed of taxa that are morphologically similar to taxa that presently grow on the north coast of California (Fig. 5, Fig. 6, Fig. 7). Miocene pollen assemblages from Hole 1022C (samples from Core 167-1022C-33X to Section 41X-1) include exotics such as Liquidambar, Pterocarya, Carya, and Tilia (Fig. 6), extant taxa that became extinct in the north coast forest (Axelrod, 1977). Paleoecologic interpretations of these pollen spectra are based on several assumptions: (1) the diagnostic components of pollen assemblages from marine cores, like those from terrestrial cores, reflect the composition of onshore vegetation formations (C.J. Heusser, 1978; L.E. Heusser, 1978a, 1978b, 1983, 1988; Heusser and Balsam, 1977); (2) modern climatic tolerances of vegetation with pollen spectra similar to fossil pollen spectra provide a reliable foundation for reconstructing past vegetation and climate; and (3) changes in late Neogene vegetation of California reflect regional and global climatic change (precipitation and temperature) (Axelrod, 1977; Huntley and Webb, 1988; Raven and Axelrod, 1978).
Pollen assemblages from Holes 1018A, 1020C, and 1022C deposited during the last ~5 m.y. reflect a latitudinal difference in vegetation, not unlike that of the present. In the southernmost site, open oak woodlands with abundant shrubs and herbs (including chaparral and sage types) are better represented than north coast forest elements such as hemlock, spruce, and redwood (excluding two samples) (Fig. 5). In the northernmost site (Fig. 7), these relationships are reversed, and hemlock, spruce, and redwood are more prominent than oak and shrub/herbs. Other taxa associated with north coast rain forests, alder and ferns, are also better represented in the north than in the south. Oscillations in the general south-north gradient in effective precipitation and temperature implied by these data are inferred from the high variability in oak, redwood, juniper/cypress, and pine pollen concentration and relative abundance (Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 7). Oak and redwood peaks presumably correspond to temperature maxima in Central California; in Northern California, hemlock, alder, and redwood maxima are associated with temperature and precipitation maxima (Adam et al., 1981a; Adam and West, 1983; C.J. Heusser, 1985; L. Heusser, 1988; L.E. Heusser, 1998).
A change in vegetation and/or climate in coastal California during the late Pliocene is inferred from the greater and/or more persistent abundance of redwood, hemlock, and spruce after ~2 Ma, possibly reflecting reduced evapotranspiration associated with the general global cooling trend (Sites 1018 and 1020). This does not seem to be associated with an obvious event in global climate records, such as ice volume. The early Pliocene decrease in oak and redwood and increase in shrubs and herbs between ~5.0 and ~3.8 Ma are interpreted as evidence of cooling and decrease in effective summer precipitation. The subsequent rise in oak suggests that temperatures were comparatively higher during the mid-Pliocene than during the early and late Pliocene. The high-frequency variations in the pollen data imply climate oscillations; however, comparison with late Pliocene climate oscillations in the northern California interior is precluded by the imprecise chronology of both data sets.
Late Miocene pollen assemblages from Hole 1022C that are distinguished by the presence of taxa no longer extant on the northwest coast of North America, accompanied by high percentages of oak (Fig. 6), obviously have no modern analog on the California coast. Climatic parameters of living taxa similar to the Miocene relicts (those of the mesophytic forests of subtropical China, for example) suggest that late Miocene climate on the north coast of California was more equable than that of the Pliocene and Pleistocene (Wang, 1961). Reduced seasonality in precipitation and temperature during the Mio-cene has also been inferred from the presence of "Arcto-Tertiary" taxa in California and Oregon Miocene floras (Axelrod, 1977; Axelrod and Bailey, 1969; Chaney, 1951; Raven and Axelrod, 1978).
To synthesize the climatic evolution of north coastal California for the last 6 m.y., the ratio of oak pollen percentages to oak + pine pollen percentages for Sites 1018 and 1022 is shown in Figure 8. The relationship between these taxa provides an index of vegetation and associated climate, with higher values indicating warm, equable environments (Adam and West, 1983; Axelrod, 1977). The late Neogene increase of pine in California was ascribed to displacement of mesic Tertiary coastal forests in response to cooling and drying—associated, at least in part, with late Pliocene elevation of California mountain ranges (Raven and Axelrod, 1978). Although precise interpretation of the nature and timing of events in this composite record from Hole 1022C is constrained by the subjective nature of the climate index, broad sampling intervals, and limited age control, the climate index provides a starting point for comparing terrestrial and oceanic signals from the California margin.
Climates of Northern California reflect broad late Neogene global climate trends such as an overall cooling from a late Miocene climatic optimum (~6-5 Ma), a Pliocene warming (between ~4.0 and ~3.5 Ma), and late Pliocene-Pleistocene high-frequency temperature oscillations. If the age model for Site 1022 is valid, then timing of Pliocene warming in coastal California inferred from pollen data deposited on the continental margin and Northern California (Axelrod, 1944) apparently precedes the mid-Pliocene warming event (~3.15 and ~2.85 Ma) identified in various geologic records elsewhere (Poore and Sloan, 1996). However, the warming event in coastal California may coincide with warming at ~4 Ma in Japan and offshore waters, when relict Tertiary taxa made their final appearance (Heusser and Morley, 1996), and with early Pliocene warming events reported elsewhere (e.g., Ciesielski and Grinstead, 1986; Zubakov and Borzenkova, 1988).