Previous biostratigraphic studies based on diatoms, planktonic foraminifera, and calcareous nannofossils provide a temporal framework for Site 32 (Barron, 1992). Samples for palynology were collected from a nearly complete sequence from about 4.5 Ma to about 3 Ma (younger Pliocene sediments were not recovered). Age estimates for these samples were determined by extrapolating from the existing biostratigraphic framework based on other microfossils.
Pollen of Taxodiaceae-Cupressaceae-Taxaceae (TCT) accounts for more than 50% of the pollen sum in most palynological assemblages from Site 32, although some assemblages are dominated by Pinus (e.g., at ~3.75 Ma and ~3. 00 Ma; see Figure 2). Pinus in Pliocene sediments of Site 32 is less abundant than amounts reported by Heusser and Balsam (1977) for modern sites from the northeast Pacific Ocean. Adam et al. (1990) developed a model for the Tule Lake, California, core record in which increased Pinus and decreased TCT indicate cooler conditions in Pliocene time, though perhaps still warmer than today. Other elements of the modern California flora present in Site 32 samples include Quercus, Artemisia, Chenopodiaceae, and Asteraceae. Exotic taxa that may be relictual from the California Miocene include Carpinus, Carya, Ilex, Taxodium, Tilia, Pterocarya, and Ulmus.
Sequoia pollen is significantly reduced in all Site 32 samples, relative to today. One of the most important factors controlling the distribution of Sequoia is fog. Low amounts of Sequoia pollen at Site 32 may reflect climatic conditions less favorable for the formation of fog. The presence of Taxodium suggests that the summers may have been wetter than today.
Results of DECORANA (Detrended Correspondence Analysis) reveal patterns in the overall data matrix from Site 32 that may relate to overall paleoclimatic variability with fluctuations apparently occurring on a 200- 400 k cycle. Inspection of the data matrix reveals that these fluctuations are in part related to changes in the relative abundances of mixed conifers (Picea, Tsuga heterophylla, and Abies). Because these taxa along the Pacific Northwest coast are controlled to a large degree by moisture, the fluctuations may reflect variations in precipitation during the Pliocene.
Preliminary analysis of the palynological data from Site 32 reveals patterns that suggest departures from modern climatic conditions during the Pliocene. Some taxa have increased or decreased relative abundances when compared to their relative abundance in modern marine core top samples. Quantification of these differences in terms of temperature and precipitation awaits acquisition of reliable modern pollen data sets from marine core tops in the northeast Pacific.
Previous workers have developed a temporal framework for the Centerville Beach section based on microfossil biostratigraphy, paleomagnetism, and radiometric dating of volcanic ashes (Barron, 1992; McCrory, 1990). Additional paleomagnetic studies are underway to refine the magnetostratigraphy in the Gauss paleomagnetic chron. Palynological samples from the Centerville Beach section have proven to be productive; additional sampling may be necessary to obtain a complete time series and to verify preliminary results.
Palynological results from the Centerville Beach section are only preliminary but suggest that the variability observed in DSDP Site 32 may also be present at Centerville Beach. Plots of the mixed conifer group (Picea, Tsuga heterophylla, Abies) display fluctuations on about the same frequency as at Site 32 and appear to be approximately in phase with these. Coincidence of these patterns between Centerville Beach and Site 32 suggests that the variability may reflect a climatic, as opposed to a taphonomic, signal. Around 2.5 Ma, the mixed conifer group shows an increase in relative abundance, possibly reflecting a southern shift of the mixed conifer forests of the Pacific Northwest. This is perhaps related to the pre-Pleistocene glaciation at approximately this time.
Palynology samples were collected throughout the 2-4.5 Ma interval, with closely spaced samples above and below the 3 Ma horizon. Unfortunately, samples processed from about 3 Ma are barren or only marginally productive. However, some samples from about 3.5 Ma and older are productive. Although palynological recovery from the Anza-Borrego section is apparently inadequate for evaluating paleoclimate at 3 Ma, reworked fossils in the section may provide important insight into climatic conditions of the Colorado Plateau during the Pliocene.
Pollen assemblages recovered from the lower part of the Anza-Borrego section contain several species of Cretaceous pollen and spores. Some of these species are distinctive and have restricted biostratigraphic and paleobiogeographic ranges in the Western Interior of North America. Proteacidites spp. first appear in the Coniacian and range through the Maastrichtian; they are present in sedimentary rocks throughout the Western Interior. Aquilapollenites spp. and Mancicorpus spp. first appear in the Campanian and also range through the Maastrichtian; they are present in the northern part but are absent from rocks in the southernmost part of the Western Interior. The southernmost extent of the latter taxa forms a line that approximately parallels the Utah-Arizona and Colorado- New Mexico boundaries.
The stratigraphic distribution of these Cretaceous taxa in Pliocene sediments of Anza-Borrego suggests that erosion of Cretaceous rocks containing Proteacidites, but lacking Aquilapollenites and Mancicorpus, had begun by the early Pliocene. Erosion into Cretaceous rocks containing Aquilapollenites and Mancicorpus did not begin until about 3.7 Ma and supports the hypothesis that the cutting of the Grand Canyon occurred during the Pliocene, accompanied by extensive and rapid erosion of Cretaceous and older rocks. Cretaceous pollen and spores were transported as detritus down the Colorado River to the ancestral Gulf of California. Rapid erosion and transport of this amount of material requires significantly increased precipitation on the Colorado Plateau during the Pliocene. This is consistent with climate models involving wetter summers and wetter winters induced by relatively recent uplift of the Colorado Plateau.