In the hemipelagic
sediments from Core EW9504-17 and from Sites 1019 and 1020, pollen concentration
is high, with slightly more pollen in sediments deposited closest to land (Fig.
2). Mean pollen abundance in sediments deposited at Site 1019 (~59 km
west of the California coast) is 5500 grains/gdws compared with mean
concentrations of 4000 grains/gdws in sediments deposited at Core EW9504-17
(~120 km offshore) and 3600 grains/gdws in sediments deposited at Site 1020
(~167 km offshore). Rapid oscillations in pollen abundance that occur in pollen
records from each of the three localities display high-amplitude peaks that show
systematic variations in Core EW9504-17 and Site 1019. Age plots (Fig.
2) show that although the overall trends of pollen abundance in Core
EW9504-17 broadly correspond to interglacial-glacial climate change (e.g.,
maxima occur in warm intervals such as the Holocene and OIS 5, and minima are
associated with cold intervals such as OIS 2), correlation between pollen
concentration and climatic events is not consistent. At Site 1019, maxima in
total pollen abundance coincide with some (OISs 5 and 7) but not all major warm
intervals, whereas variations in mass accumulation rates of redwood at Site 1020
(like that of terrigenous minerals; see Hovan et al., Chap.
18, this volume) correspond fairly closely to global 
18O
variations. Pollen concentration on the California continental margin reflects
vegetation density, pollen sedimentation (including fluviomarine sedimentary
processes), and climatic processes (regional and global; Heusser and Balsam,
1977; Traverse, 1988; Fig. 2).
The pollen records from Sites 1019 and 1020 as well as Core EW9504-17 are composed of taxa that presently grow in northwest America. 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 in varying degrees the composition of vegetation formations from which they are derived (Heusser, 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 (Heusser et al., 1980; Whitlock and Bartlein, 1997); and (3) changes in the Quaternary vegetation of California and southern Oregon reflect regional and global climatic change (Axelrod, 1977; Huntley and Webb, 1988).
At all three localities, pollen spectra of the uppermost samples are marked by a succession of alder, oak, and redwood peaks accompanied by ferns and lesser amounts of western hemlock, spruce, and cedar (Fig. 3, Fig. 4, Fig. 5). These distinctive assemblages of pollen representative of natural north coast forest and oak woodland communities occur repeatedly downcore: between ~13 and ~15 m in Core EW9504-17 (Fig. 3); also, between ~38 and ~32 mcd and between ~46 and ~66 mcd at Site 1019 (Fig. 4). In the record from Site 1020, the alder-oak-redwood-fern assemblage is repeated five times (Fig. 5). Although the basic composition is essentially the same in the three records, differences exist in the sequence and relative abundance of taxa. In Holes 1019C and 1019E, the redwood maxima at 33.48 mcd precedes that of oak (32.68 mcd). At Site 1020, redwood maxima vary in amplitude from ~35% at 44 mcd to ~10% at 26-28 mcd, and alder maxima range from ~22% at ~55 mcd to ~7% at 14 mcd. Except for the peak at ~14 mcd, oak peaks display less variation.
The peaks of alder that initiate abrupt expansions of coastal lowland forest types reflect the pioneer role of alder in aggressively colonizing areas disturbed by catastrophic events (flooding and infrequent wildfires) or by reorganization of plant communities related directly or indirectly to climate change. The apparent lack of systematic variation in charcoal fragments found in pollen samples suggests that wildfires were probably not the primary cause of the high-amplitude alder events. Alder habitats are wet, along streams and in marshy places in redwood and mixed evergreen forests (Alnus oregona and A. sinuata), and moist places in the north Coastal Range (A. tenuifolia; Barbour and Billings, 1988; Munz, 1968). Increased alder in sediments deposited on the California margin thus may reflect expansion and/or recolonization of alder habitats related to increased precipitation and/or increased runoff from snowmelt during deglaciation (Heusser and Shackleton, 1979). Alder expansion in areas of revegetation following disturbance is well documented in the Pacific Northwest (Barbour and Billings, 1988; Grigg and Whitlock, 1998; Heusser, 1985).
Subsequent expansion of the two different north coast lowland habitats—mesic redwood forests along the narrow band adjacent to the Pacific Ocean and xeric oak woodland farther inland—may be related to changes in temperature, precipitation, and other atmospheric and oceanic processes. Higher temperatures and change in the amount and seasonal distribution of rainfall that enhanced summer drought (increased summer moisture stress related to decreased effective precipitation and increased summer temperatures, for example) would favor development of lowland oak communities. Such conditions would not be conducive to redwood expansion. Coastal redwood requires much lower diurnal and annual temperature fluctuations than oak and is now restricted to that part of the coast where temperature extremes and summer moisture stress are modified by prolonged cloudy periods and marine fog associated with offshore upwelling (Zinke, 1977). We suggest that redwood expansion was probably closely related to the development of maritime conditions (changes in the distribution and intensity of upwelling in the California Current, for example), which moderated north coast climate. Detailed statistical analyses of radiolarian species and pollen taxa from Core EW9504-17 show a very high degree of correlation between radiolarians associated with coastal upwelling and redwood (N.G. Pisias, pers. comm., 1998).
All the pollen records contain lengthy intervals dominated by seemingly uniform assemblages composed of pine, herbs, and cedar, along with spruce, hemlock, and mountain hemlock (T. mertensiana, a subalpine species; Barbour and Billings, 1988). To some extent, the apparent uniformity reflects limitations of pollen analysis; that is, the presence of genera and families that cannot be discriminated into more ecologically specific taxonomic taxa. The composition of these conifer-dominated assemblages alternates between coastal forest taxa like those now growing north of the present distribution of redwood (in northern Oregon, Washington, and British Columbia) and forests or open pine woodlands not unlike those that now occur at higher elevations (Barbour and Major, 1977). Oscillations in herbs that occur throughout indicate variable development of vegetative cover.
Similar sequences of temperate lowland conifer forest and oak woodland assemblages alternating with those of montane forest/woodland occur in upper Quaternary pollen records from sites in Northern California, western Oregon, and Washington as well as in Quaternary pollen data from other marine cores taken off Northern California, Oregon, and Washington (Adam and West, 1983; Grigg and Whitlock, 1998; Heusser, 1985, 1998; Heusser and Shackleton, 1979). Regional differences in vegetation are readily apparent in the composition of the warm temperate pollen assemblages (i.e., the prominence of redwood and/or oak in the south and of western hemlock and spruce in the north) and in the composition of the more homogeneous pollen assemblages in the intervening cooler intervals, which are characterized by the greater prominence of juniper and cedar types in cores from the south. The similarity between the systematic downcore variations in marine and continental records from the same geographic area implies that marine pollen records, like those on land, capture systematic variations in regional vegetation and climate.
When pollen data are
plotted against age (Fig. 6, Fig.
7, Fig. 8), it is evident that
in Core EW9504-17, alder, oak, redwood, and fern maxima correspond to benthic 18O
minima and that downcore variations in pollen stratigraphies of Sites 1019 and
1020 also reflect orbital-scale global climate fluctuations. Small differences
in the relative abundance of the mesophytic, temperate taxa between OISs 1, 5,
7a, 7c, and 9 suggest that the development and composition of interglacial
vegetation in Northern California was not always identical in each of the last
four interglacials. Because of the preliminary nature of age models for Sites
1019 and 1020 (Lyle et al., Chap.
32, this volume), we focus on the well-dated EW9504-17 time series.
The double beat of OIS 7 18O
at Site 1020 (Fig. 8) is
mirrored in muted alder, oak, and redwood peaks that are more robust (as are
ferns) in the preceding interglacial (OIS 9). The OIS 6/5e transition in Core
EW9504-17 and at Site 1019 is marked by an abrupt rise in alder that is rapidly
succeeded in OIS 5e by oak and redwood peaks in Core EW9504-17. At Site 1019,
the redwood maximum precedes that of oak. Multiple oscillations in these taxa
occur in OIS 5c. These large-scale patterns also occur in pollen assemblages in
the upper ~130 k.y. of the lower resolution Site 1020 pollen record (Fig.
8). It is worth noting that in all three records, pollen assemblages
from OIS 5e are not exact replicates of those in OIS 1. During the last
interglacial, oak was more abundant than in the Holocene; the converse is true
for alder and redwood in Core EW9504-17 and at Site 1020.
Pine, dominant in OIS 5d and 5b, became increasingly important during the last full glacial, as did sage and other herbs. In OIS 3, brief pine events occurred at ~14, ~16, ~35, ~38, and ~42 k.y. in Core EW9504-17 (Fig. 6). Two well-defined events coincide with major episodes of North Atlantic ice rafting (Heinrich Events H1 and H4). The low amplitude of redwood and oak oscillations during OIS 3, which partly reflects overrepresentation of pine (a common feature of pollen dispersal and sedimentation; Traverse, 1988), precludes correlation with interstadial events elsewhere.
In Holes 1020C and 1020D, the rhythmic pattern of downcore variation in the alder, oak, redwood, and fern maxima (Fig. 5) is evident through OIS 13 (M. Lyle, pers. comm., 1998). Between ~43 and ~45 mcd (OIS 11), the robust redwood acme is preceded by a substantial peak in alder and ferns. Oak percentages are comparable to those of previous interglacials (excluding OIS 5). The high-amplitude pulse of alder between ~51 and ~55 mcd (OIS 13) leads a lesser rise in redwood, oak, and ferns. As in the younger part of the pollen record, interglacial assemblages display individualistic variations.
At glacial-interglacial
transitions (OISs 8/7, 6/5e, 2/1, and most probably 10/9), the abrupt changes in
west coast vegetation (identified by the rapid expansion of the pioneer alder)
and global warming (higher benthic 18O)
are nearly synchronous (Fig. 6, Fig.
8). The initial increase in alder and the shift in 18O
at glacial terminations occurs in exactly the same sample depths in Core
EW9504-17 and at Site 1020. Directly correlative pollen and 18O
data from two other northeast Pacific cores (Core Y7211-1 taken at 43°15´N,
126°22´W;
Site 893 taken at 34°17.25´N,
120°02.19´W;
Fig. 1) showed similar
relationships at the OIS 6/5e transition (Heusser, 1995; Heusser and Shackleton,
1979). These data imply that glacial-interglacial variations in northwest North
American climate and vegetation over the last 350 k.y. were (within constraints
of sample resolution) apparently nearly synchronous with orbital-scale global
ice-volume variations. Directly correlative terrestrial/marine records from
piston cores taken in the northwest Pacific (Morley and Heusser, 1997) and from
Site 594 in the southwest Pacific (Heusser and van de Geer, 1994) showed that
large-scale variations in Japanese and New Zealand ecosystems over the last ~350
k.y. could be attributed to orbital forcing of global climate mechanisms.
To display pollen data in a form less affected by overrepresentation of pine, we use pollen ratios (Fig. 9). The redwood and western hemlock/spruce ratio can be regarded as a temperature indicator of mesophytic lowland forests because average July temperatures in areas now dominated by redwood and/or western hemlock are ~1° to 2°C higher than in areas dominated by spruce (Heusser, 1985; Heusser and Shackleton, 1979; Zinke, 1977). The oak/pine ratio serves as an indicator of temperature trends in the more arid northern California interior since temperatures in the lowland oak woodlands are several degrees higher than temperatures in montane pine forests (Adam and West, 1983). At ~42°N, maximum mean monthly temperatures at 332-m elevation in the nearby mountains are 6.2°C higher than on the coast (Barbour and Major, 1977). Although we describe these climate proxies as temperature indicators, we recognize that effective precipitation is a major factor in vegetative composition and cannot be effectively separated from temperature in our paleoclimatic proxies.
The close correspondence
between pollen ratio, insolation, and 18O
curves implies that north coast environmental fluctuations were broadly
synchronous with changes in global climate over the last ~150 k.y. (Fig.
9). As suggested earlier, the lag in the response of redwood/western
hemlock communities probably reflects the significant role of sea-surface
conditions in the development of north coast maritime vegetation (Lyle et al., Chap.
32, this volume).
