From the present data set, 17 biohorizons that include LOs, HOs, LCOs, and a HCO were discriminated for potential biostratigraphic markers (Table T2; Fig. F3). They include possible biohorizons based only on relatively sporadic occurrences of the taxa. Seven acmes (abundance peaks) are also listed.
In the lower upper Miocene to middle Miocene section of Hole 1151A, a number of dinoflagellate cyst biohorizons were recognized, whereas the assemblages from the uppermost lower Pliocene to upper upper Miocene section, being generally dominated by Brigantedinium spp. and Xandarodinium sp. A, do not display so many appearances and disappearances of species. This contrast may be, partly at least, affected by the contrast in sedimentation rate.
When compared to studies by Matsuoka et al. (1987) and Obuse and Kurita (1999), which are based on composite onshore sections in northern Japan (Fig. F1), parallels in the general biostratigraphic trends are clearly seen.
First, comparison is made with the biostratigraphy and biozones proposed by Matsuoka et al. (1987), who studied material from the Oga and Niigata areas in the backarc side of the Honshu Island (Fig. F1). The distinctive contrast between the assemblages from the Diphyes latiusculum Zone (late early to early middle Miocene) and those from the higher horizons, as stated by Matsuoka et al. (1987), is easily recognized in the present data set as a series of HOs of species including Diphyes latiusculum in the lower middle Miocene. However, it should be noted that in the present section this contrast is accentuated by the hiatus between the lower middle Miocene and the upper Miocene.
The general trend of stratigraphic changes in species composition in the upper Miocene to lower Pliocene interval is expressed by Matsuoka et al. (1987) as the disappearances of Achomosphaera spongiosa (as Achomosphaera sp. A in Matsuoka et al., 1987), Spiniferites hexatypicus, and Spiniferites ellipsoideus, which are followed by abundant occurrences of Brigantedinium simplex. These stratigraphic changes were used by Matsuoka et al. (1987) to establish the boundary between the Capillicysta fusca Zone and the Achomosphaera callosa Zone. Similar stratigraphic changes were recognized in the present study as the successive HOs of Spiniferites hexatypicus, Achomosphaera spongiosa, and Spiniferites ellipsoideus in the upper Miocene, which were followed by the base of abundant occurrences of Brigantedinium spp., and then by the LO of Operculodinium centrocarpum sensu Wall and Dale, 1966 (as Operculodinium centrocarpum in Matsuoka et al., 1987). This major biostratigraphic trend is therefore observed both in the present section and in the backarc location studied by Matsuoka et al. (1987), although there are slight discrepancies in the ages of the biohorizons.
Differences between the results of the present study and those of Matsuoka et al. (1987) include the Miocene biohorizons of Evittosphaerula sp., Pyxidinopsis spp., and Operculodinium giganteum, which were not reported by Matsuoka et al. (1987). In contrast, the Pliocene assemblages of Matsuoka et al. (1987) are apparently more diverse than those of the present study. This may be due to the predominance of protoperidiniacean species in the Pacific assemblages that might have masked out other species. These differences may be attributed to the paleoecological differences between the forearc and the backarc.
Initial work by Matsuoka et al. (1987) was extended by Obuse and Kurita (1999) in their study of onshore sections in northern Japan from more widespread localities including the Joban, Niigata, Akita, Oga, and Ishikari areas (Fig. F1). Obuse and Kurita (1999) also provided more detailed age constraints, based on diatom biostratigraphy, in their range chart of species from the upper lower Miocene to upper Pliocene interval. Their results were generally similar to those of Matsuoka et al. (1987), although they were able to refine the biostratigraphy and identify several new biostratigraphically useful taxa.
A comparison of the present data set with that of Obuse and Kurita (1999) shows a number of mutually recognized biohorizons. Those include the LO of Operculodinium centrocarpum sensu Wall and Dale, 1966, the HO of Spiniferites hexatypicus, the HO of Achomosphaera spongiosa (including cf. A. spongiosa), the HO of Heteraulacacysta campanula, the LO of Spiniferites firmus, the LO of Melitasphaeridium choanophorum, the HO of Operculodinium giganteum, the HO of Cleistosphaeridium placacanthum, the HO of Diphyes latiusculum, the HO of Dinocyst B (as "penitabular dinocyst" in Obuse and Kurita, 1999), the HO of Pyxidinopsis spp. (as Pyxidinopsis cf. tuberculata in Obuse and Kurita, 1999), and the HO of Evittosphaerula sp. (as E. paratabulata in Obuse and Kurita, 1999). These biohorizons have little differences in the order of stratigraphic appearances, although the ages of the biohorizons may vary slightly in some cases.
In addition to these biohorizons, some of the acme events in the present section are also correlatable to the records from the onshore localities. For example, the acmes of Tuberculodinium vancampoae and Impagidinium patulum are recorded in the lower Pliocene both in the present study and in Obuse and Kurita (1999). If these acme events are proven to be isochronous, they should be useful for biostratigraphic correlation.
The overall similarity between the present offshore data set from the forearc setting and earlier studies based on onshore material from various localities including the backarc region indicates a clear potential for regional stratigraphic correlation. Further study is needed, however, because there are still discrepancies in the ages of some biohorizons. These may be explained by contrasting paleoenvironments between the forearc side facing the Pacific and the more isolated backarc side facing the Sea of Japan. Or they may be due to effects of sampling interval, inadequate lithology, or other geologic and nongeologic biases.
The dinoflagellate cyst assemblages from the uppermost lower Pliocene-upper upper Miocene interval of the present section are marked by the predominance of protoperidiniacean species of Brigantedinium spp. and Xandarodinium sp. A, which contrasts with those from the underlying interval. This may be interpreted in the context of paleoceanographic changes in the forearc basin, such as those in temperature, salinity, and/or nutrient supply.
A similar biostratigraphic contrast was noted by Bujak (1984), who studied Neogene material from DSDP Leg 19 in the Bering Sea and the northern North Pacific. He reported that the late Miocene to Pleistocene assemblages are dominated by protoperidiniacean genera such as Brigantedinium, Lejeunecysta, Selenopemphix, and Xandarodinium, whereas the older assemblages lack this characteristic. Bujak (1984) concluded that the onset of protoperidiniacean cyst dominance could signal the establishment of nutrient-rich water because almost all protoperidiniacean dinoflagellates, as he stated, are heterotrophic and require organic nutrients.
Recently, Matsuoka (1999) discussed the role of protoperidiniacean dinoflagellates in the marine ecosystem, particularly in shallow depths, and proposed a food web among the algal community that places the heterotrophic dinoflagellates including protoperidiniaceans as the consumer (predator) of smaller algae including diatoms and other photosynthetic microorganisms. Emphasizing the predating nature of protoperidiniacean dinoflagellates, this model indicates that increase of protoperidiniaceans in cyst assemblages can be a record of eutrophication.
It is notable that both the present data and that of Bujak (1984) show the onset of protoperidiniacean cyst dominance occurring in the late Miocene, both in off the Sanriku Coast and in the Bering Sea and the northern North Pacific. Following the suggestion of Bujak (1984) and Matsuoka (1999), the dominance in these regions would be a direct consequence of the establishment of diatom-rich algal community, or nutrient-rich water, in the late Miocene in the northwestern Pacific-Bering Sea region. This possible paleoceanographic change may be a result of global cooling that might have caused enhancement of bottom-water circulation as well as higher concentrations of silica in high latitudes.
Remaining questions about this issue include that the onset of protoperidiniacean cyst dominance in the late Miocene appears not to be very clear in the assemblages from the backarc location of northern Japan according to the studies by Matsuoka (1983), Matsuoka et al. (1987), and Obuse and Kurita (1999). This may be related to the paleoenvironmental differences between the areas, as a backarc location might be more isolated. In addition, it may be important in further study to relate the Miocene diatomaceous successions in the backarc area with dinoflagellate cyst records as well as with other paleoenvironmental indicators.