1 Introduction

Forearc basins develop between volcanic arcs and trenches, and they are one of the basic components of a subduction system (Dickinson 1995). Their uplift and subsidence generally reflect long-term (at least 10–100 ky) depositional and tectonic processes due to forearc accretion and erosion (Clift and Vannucchi 2004; Takano et al. 2013; Noda 2016). It is known that forearc basin segmentation occurs because of the growth of zones of uplift and subsidence along the forearc wedge when subduction of the upper surface of a slab with a convex topography (e.g., seamounts; Tréhu et al. 2012) and/or oblique subduction take place (Wells et al. 2003; Okamura and Shishikura 2020). These processes affect the underlying accretionary wedge structure, which is closely related to the partitioning of the rupture zones of megathrust earthquakes (Sugiyama 1992; Wells et al. 2003; Fuller et al. 2006). Such segmentation is expected to be recorded as stratigraphic discontinuities and structural differences, and this might be recognized in ancient forearc basin fills that are exposed on land. However, very few cases of such stratigraphic discontinuities have been recognized, probably because the temporal accuracy is often not high enough to track major stratigraphic discontinuities across an entire basin in outcrops.

Forearc basins such as the Kumano Basin were developed along the Pacific side of the Japanese island arc, and their development has been affected by interactions of subduction tectonics and sedimentation (Underwood and Moore 2012). Using seismic facies and borehole data, the long-term evolution of these forearc basins has been discussed in terms of depositional processes (e.g., development of submarine fans) and the tectonic uplift/subsidence that produces unconformities (Takano et al. 2013; Moore et al. 2015; Kimura et al. 2018). In terms of forearc basin development owing to subduction of the Philippine Sea Plate, the Miocene–Pleistocene forearc basin fills that make up the basement of the Tokyo metropolitan area provide a good on-land analog of a modern forearc basin. The forearc basin fills are well exposed around Tokyo Bay, especially on the Miura and Boso peninsulas (Fig. 1). The deposits provide the type stratigraphy for other equivalent sedimentary sequences on the Japanese islands and in the northwest Pacific (Kazaoka et al. 2015; Ito et al. 2016), largely because they contain many marker tephra beds and well-preserved planktic microfossils that allow bed-by-bed correlations and estimates of sedimentary age.

Fig. 1
figure 1

Maps illustrating the geotectonic setting of the Miura and Kazusa groups in the southern Kanto region on the Pacific side of central Japan. a Present-day tectonic setting of central Japan. b Surface distributions of the Miura and Kazusa groups on the Miura and Boso peninsulas, based on Suzuki et al. (1995). Thinner lines within the yellow and green areas represent boundaries between formations. HMUB = Hayama − Mineoka uplift belt

There is a well-known stratigraphic discontinuity called the Kurotaki unconformity (Koike 1951) that divides the basin fill into the Miura (Awa) and Kazusa groups. The Kurotaki unconformity has received considerable attention because it was thought to represent a tectonic transition owing to a change in the direction of movement of the Philippine Sea Plate (Takahashi 2006), and also because other unconformities of similar age can be observed in the Japanese islands (Tamura and Yamazaki 2010; Oda et al. 2011). Meanwhile, the chronostratigraphy of the Miura Peninsula, as reported by Utsunomiya et al. (2017), showed that continuous sedimentation occurred in the western part of the basin at the same time when the Kurotaki unconformity was forming in the eastern part, and this indicated the need for a better stratigraphic model to explain these regional differences. The purpose of our study was to construct a new stratigraphic model based on tephra bed correlations and calcareous nannofossil biostratigraphy. We revealed a previously overlooked major erosional surface and coeval mass-transport deposits (MTDs) in the forearc basin fills, and we discussed their significance for forearc basin segmentation.

2 Geologic setting

2.1 Forearc basin fills: the Miura and Kazusa groups

The central part of the Honshu arc is located near a trench–trench–trench-type triple junction (Fig. 1a) where the Philippine Sea Plate is subducting beneath the North American Plate and the Pacific Plate is subducting beneath the North American and Philippine Sea plates (McKenzie and Morgan 1969). Deformation of the Neogene deposits exposed on the Miura and Boso peninsulas records compressional convergence owing to the interaction of the Philippine Sea and Eurasian plates near the triple junction after the Miocene (Ogawa et al. 1989). In this tectonic setting, the Hayama–Mineoka uplift belt consists of an ophiolitic complex that acted as a trench–slope break (Takahashi 2008) that created space for the deposition of the forearc basin fill (Fig. 1b). The forearc basin fills on the Miura and Boso peninsulas are called the Miura (middle Miocene–Pliocene) and Kazusa (Pliocene–middle Pleistocene) groups, which are approximately 2000 and 3000 m thick, respectively (Suzuki et al. 1995). These groups consist of shallow- and deep-marine deposits (Mitsunashi and Kikuchi 1982; Ito and Katsura 1992), and they contain numerous marker tephra beds and abundant microfossils that facilitate lateral stratigraphic correlations and reliable determinations of depositional age (e.g., Sato and Takayama 1988; Pickering et al. 1999). Moreover, the fact that the tephra beds act as stratigraphic markers and can be traced laterally has allowed sedimentologists to recognize spatial variations in the depositional sequences (Ito 1992a, b, 1995), the sandstone beds in submarine fans (Hirayama and Nakajima 1977), and large-scale mass-transport deposits (Utsunomiya 2018; Utsunomiya et al. 2018; Utsunomiya and Yamamoto 2019). Previous studies on stratigraphic correlations between the Boso and Miura peninsulas (Ida et al. 1956; Mitsunashi 1973; Mitsunashi and Kikuchi 1982; Suzuki et al. 1995) have been based mainly on similarities of the macroscopic and microscopic characteristics of the marker tephra beds. A combination of microfossil biostratigraphy and the geochemical compositions of volcanic glass in the tephra beds allowed significant improvements in correlating the stratigraphic units of Pleistocene sequences around Tokyo Bay, such as on the Miura, Boso, and Choshi peninsulas (marine deposits) and in the Tama hills (terrestrial and shallow marine deposits) (Fujioka et al. 2003; Fujioka and Kameo 2004; Takahashi et al. 2005; Tamura et al. 2010, 2019; Mizuno and Naya 2011; Suzuki and Murata 2011; Utsunomiya et al. 2019). The stratigraphy established for the Miura and Boso peninsulas has played a significant role in interpretations of seismic profiles (Chiba Prefecture 2004; Asao and Ito 2011) and deep-borehole sequences around Tokyo Bay (Suzuki and Horiuchi 2002; Yanagisawa et al. 2006; Chiyonobu et al. 2007; Naya et al. 2013). Moreover, the tephra beds of the Kazusa Group recently recognized at depth (> 1000 m) around Tokyo Bay (Tamura et al. 2010; Utsunomiya et al. 2020) have allowed the integration of seismic-scale structures with the high-resolution stratigraphy established from outcrop-based studies.

The boundary between the Kazusa and Miura groups is called the Kurotaki unconformity, with the lower Kazusa Group lapping onto the underlying Miura Group as the depocenter of the Kazusa Group migrated northward (Mitsunashi and Yamauchi 1988). On the Boso Peninsula, the unconformity between the Kazusa and Miura groups is distinctly angular, representing a temporal gap from ca. 3 to 2.4 Ma (Kameo and Sekine 2013), but on the Miura Peninsula there is no significant angular relationship between the two groups (Koike 1951; Akamine et al. 1956). No significant temporal gap corresponding to the Kurotaki unconformity has been identified in the Inubo Group, exposed on the Choshi Peninsula, based on biostratigraphy and tephrostratigraphy (Sakai 1990; Tamura and Yamazaki 2010; Tamura et al. 2014). Utsunomiya et al. (2017) demonstrated that the uppermost parts of the Miura Group and the lowermost parts of the Kazusa Group on the Miura Peninsula are time-equivalent deposits, based on the lateral tracing of tephra beds. Gradual variations in the rate of sedimentation indicated that the Kazusa Group of the Miura Peninsula was deposited conformably on the Miura Group during 3.2–2.4 Ma, based on calcareous nannofossil biostratigraphy and magnetostratigraphy. The stratigraphic flamework is therefore being reconstructed with a focus on Pliocene stratigraphy as the key interval.

2.2 Miura peninsula

The geologic map of the northern Miura Peninsula in Fig. 2a has been modified from Eto (1986a), Eto et al. (1998) and Utsunomiya et al. (2017) based on the distribution of tephra beds shown by Inagaki et al. (2007), Utsunomiya et al. (2017), and this paper. The Miura and Kazusa groups generally dip 5–30°NE, but the Miura Group and the lowermost parts of the Kazusa Group have been deformed by a NE–SW trending syncline. The Miura Group consists of the Miocene–Pliocene Zushi and the Pliocene Ikego formations, in ascending stratigraphic order. A basal conglomerate called the Tagoegawa sandstone and conglomerate Member (15–50 m in thickness) is overlain by mudstone-dominated alternating sandstones and mudstones (1000–1500 m in thickness) of the Zushi Formation (Eto et al. 1998). The lower part of the Ikego Formation consists of the Takatoriyama volcaniclastic Member (0–210 m in thickness) and the Jimmuji volcaniclastic and mudstone Member (0–60 m in thickness), and the main upper part of the Ikego Formation consists of alternating beds of sandstone and tuffaceous sandy mudstone 150–400 m in thickness (Eto et al. 1998). In detail, the Takatoriyama Member consists of massive and cross-bedded tuffaceous sandstones and conglomerates that have been interpreted as canyon-fill deposits, given the convex-down shape inferred from their geographic distribution (Soh et al. 1991). The Jimmuji Member consists of mass-transport deposits (Fig. 3b; Eto 1993; Yokohama Defense Facilities Administration Bureau 1993) that contain chaotic deposits with blocks of sandstone and mudstone, likely generated by submarine landslides of both the canyon wall and the canyon fill (Taira et al. 1993). The upper Ikego Formation consists mainly of alternating beds of sandstone and tuffaceous sandy mudstone, which cover an irregular erosional surface of a slide block of sandy mudstone in the Jimmuji Member (Fig. 3; Eto 1993; Yokohama Defense Facilities Administration Bureau 1993; Utsunomiya et al. 2017). The lowermost part of the Kazusa Group, the Urago Formation, consists mainly of tuffaceous medium- to coarse-grained sandstones and tuffaceous muddy sandstones (Utsunomiya and Majima 2012). Trough- and tabular-cross bedding is common in the tuffaceous sandstones, and this has been attributed to the migration of dunes under the influence of bottom currents that varied in direction between northward and eastward (Utsunomiya et al. 2015).

Fig. 2
figure 2

Geologic maps of a the northern Miura Peninsula (Eto 1986a; Eto et al. 1998; Utsunomiya et al. 2017), b the western Boso Peninsula (Nakajima and Watanabe 2005), and c the central Boso Peninsula (Tokuhashi and Ishihara 2008)

Fig. 3
figure 3

Distribution and outcrop occurrences of the Zushi Formation, the Jimmuji Member, and the upper part of the Ikego Formation. a Geologic map of the Hisagi and Ikego areas, Zushi City, showing the locations of the sketched outcrops reported by previous researchers, and the locations of the studied sites. Gray-colored strike and dip symbols are from Eto (1993). b Sketches of outcrops exposed by residential construction illustrate stratigraphic levels of the Nt and Ng tephra beds and the internal structure of the Jimmuji Member (after Eto 1993, and Yokohama Defense Facilities Administration Bureau 1993). c Route maps in the Ikego Forest Natural Park showing location of outcrops of the Zushi and Ikego formations. d Geologic columns along traverses Z1 and Z2 where the uppermost part of the Zushi Formation and the overlying Ikego Formation are exposed. Note that the Jimmuji Member becomes thinner and pinches out westward where a tens of cm-thick pebbly sandstone bed covers the erosional surface. The lithologic legend is shown in Fig. 4

The paleobathymetry of the upper Ikego Formation and the Urago Formation has been estimated at between 500 and 2000 m based on benthic foraminiferal assemblages (Eto et al. 1987) and between 400 and 600 m based on molluscan asemblages (Utsunomiya and Majima 2012). The upper part of the Ikego Formation has been assigned to the CN12a subzone of Okada and Bukry (1980) by Okada (1993) and to the RN12 Zone of Kamikuri et al. (2009) by Suzuki and Kanie (2012).

Although the top of the Ikego Formation has been interpreted as the westward extension of the Kurotaki unconformity (Mitsunashi 1973; Eto 1986b; Eto et al. 1998), the bedding planes of the Ikego Formation are almost parallel to those of the overlying Kazusa Group (Koike 1951; Akamine et al. 1959). Utsunomiya et al. (2017) demonstrated that the Urago Formation of the Kazusa Group conformably overlies the Ikego Formation, based on tephra bed correlation, then identified the following stratigraphic markers in ascending order (Fig. 4): top (3.21 Ma) of the Mammoth subchronozone (C2An.2r); base (3.13 Ma) and top (3.05 Ma) of the Kaena subchronozone (C2An.1r); last appearance of Discoaster tamalis (MIS G7–G8, 2.76 Ma; Kameo and Okada 2016); the widespread tephra bed KGP (ca. 2.5 Ma); and the last appearance of Discoaster pentaradiatus (MIS 95, 2.41 Ma; Kameo and Okada 2016). The KGP tephra bed occurs within the lowermost Matuyama chronozone (Tamura et al. 2010; Ueki et al. 2013), which suggests the Gauss–Matuyama boundary (2.61 Ma) is in the middle of the Urago Formation.

Fig. 4
figure 4

Geologic columns of the Ikego and Urago formations, Miura Peninsula (after Utsunomiya et al. 2017), showing magnetic polarity and calcareous nannofossil biozones. Note that tephra beds IkT16 and 19 are intercalated immediately above the top of the Mammoth subchronozone. Ahn is within CN12b–c. Ahn is between the top of Discoaster tamalis (2.76 Ma) and tephra bed KGP (ca. 2.5 Ma), which is consistent with the estimated age of the widespread tephra UN-MD2 (2.7–2.6 Ma). The calcareous nannofossil zonal scheme follows that of Okada and Bukry (1980). The notation of the paleomagnetic polarity zones is from Ogg (2020)

2.3 Boso peninsula

The geologic maps of the western (Fig. 2b) and eastern (Fig. 2c) Boso Peninsula are from Nakajima and Watanabe (2005) and this study, respectively. Although the Miura Group on the Boso Peninsula is often called the Awa Group (Nakajima et al. 1981; Nakajima and Watanabe 2005), for convenience we use the name Miura Group following Suzuki et al. (1995) in order to identify time equivalent strata on both peninsulas. The Miura Group on the western Boso Peninsula generally strikes E–W, and it has been deformed by ENE–WSW trending folds (the Seri Syncline and Terao Anticline: Fig. 2b). On the central Boso Peninsula, the fold axes trend ESE–WNW, and the folds are called the Kiyosumi syncline and Kiyosumi anticline. The Miura Group on the Boso Peninsula consists of the Fukawa, Kanigawa, Kinone, Amatsu, Kiyosumi, and Anno formations in ascending stratigraphic order (Nakajima et al. 1981), among which the Kiyosumi and Anno formations were the subject of our study.

The Kiyosumi Formation consists mainly of alternating beds of hemipelagic mudstone and sandstone, and it varies in thickness from 76 to 196 m on the western side (Nakajima 2005) to ~ 800 m in the center of the peninsula (Tokuhashi and Ishihara 2008). The Kiyosumi Formation represents sand-dominated submarine fan deposits with a feeder system from north to south (Tokuhashi 1989). The spatial–temporal distribution of turbidite deposits shows the development of syn-sedimentary E–W trending folds, which resulted in the strata being thicker along the synclinal axis (Tokuhashi 1976a, b; Tokuhashi 1989). The paleobathymetry of the Kiyosumi Formation has been estimated to be 1000–1800 m (Hatta and Tokuhashi 1984), based on benthic foraminiferal assemblages.

The Anno Formation conformably overlies the Kiyosumi Formation and is unconformably overlain by the Kazusa Group across the Kurotaki unconformity. The thickness of the Anno Formation is 125–386 m in the western part of the Boso Peninsula (Nakajima 2005). Previous studies numbered the marker tephra beds in the Anno Formation in ascending stratigraphic order with numbers prefixed by “An” (Nakajima et al. 1981; Natural History Museum and Institute, Chiba, 1995, 1996; Nakajima 2005). In the western part of the peninsula, the Anno Formation generally shows an upward-coarsening succession from mudstone, sandy mudstone, and muddy sandstone to sandstone (Nakajima 2005). The Anno Formation represents submarine fan deposits (Nakajima et al. 1981; Ishihara and Tokuhashi 2005) or a slope base system (Saito and Ito 2002). Ishihara and Tokuhashi (2005) recorded the lateral distribution and successive changes in the turbidite facies and paleocurrent directions. They identified six sedimentary facies associations: channel fill deposits of pebbly sandstone and conglomerates; natural levee deposits of alternating thin turbidites and hemipelagic mudstones; thick-bedded lobe deposits; marginal deposits of alternating thin turbidites and hemipelagic mudstones; marginal deposits of alternating hemipelagic mudstones and tephra beds; and extensive slump deposits. The upper part of the Anno Formation consists of marginal deposits of alternating hemipelagic mudstones and tephra beds that coarsen upward, suggesting a shallower depositional environment or the winnowing of fine-grained sediment owing to bottom currents (Ishihara and Tokuhashi 2005). There are many mass-transport deposits in the Anno Formation, of which “Ta slump” below An22 and “Sak slump” below An157 can be traced laterally (Nakajima 2005). There is a hiatus due to submarine landslide(s) and residual mass-transport deposits called “Sak slump” between An155 and An157 (Fig. 5). Sak slump ranges in thickness from 0.2 to 5 m, and it caused the erosion of 8–15 m-thickness of underlying strata, including tephra beds An155-2–An156-4 (Nakajima 2005). The paleobathymetry of the Anno Formation has been estimated to be 1200–2800 m (Hatta and Tokuhashi 1984), based on benthic foraminiferal assemblages.

Fig. 5
figure 5

Geologic cross section, route map, and geologic columns of the Anno Formation on the western Boso Peninsula. a Stratigraphic correlation of the uppermost part of the Anno Formation showing east–west (Kobata–Iwamoto–Kunawa) and south–north (Kunawa–Terao, Nagasaki) variations in horizons underlying the Sak slump and the equivalent erosional surface (blue line). b, c Route maps in Nagasaki b and Seki c with outcrops of the upper part of the Anno Formation. d Geologic columns of part of the Anno Formation that shows lateral facies changes, based on tephra bed correlation. In a similar manner to the Ikego area (Fig. 3), a mass-transport deposit named “Sak slump” becomes thinner and pinches out eastward where a tens of cm-thick pebbly sandstone bed covers the erosional surface. The lithologic legend is shown in Fig. 4

The depositional ages of the Kiyosumi and Anno formations have been estimated using calcareous nannofossil biostratigraphy (Kameo et al. 2010; Kameo and Sekine 2013), fission-track analyses of the tephra beds (Kasuya 1987; Tokuhashi et al. 2000), and magnetostratigraphy (Niitsuma 1976; Haneda and Okada 2019, 2022). The Kiyosumi Formation includes C3n.4n and C3n.3n normal polarity subchronozones, and the top of this formation belongs to the C3n.2n normal polarity subchronozone (Niitsuma 1976; Takahashi 2008). Haneda and Okada (2019) recognized the period in the Anno Formation that extends from the top of the Nunivak normal polarity subchronozone (4.49 Ma) to the top of the Mammoth reverse polarity subchronozone (3.21 Ma). The Sak slump is intercalated within the Mammoth subchronozone, and it may have caused the erosion of strata over a period of approximately 90 kyr (Haneda and Okada 2019).

The Kazusa Group on the Boso Peninsula is made up of a shallow- to deep-marine succession that is ~ 3 km thick (e.g., Ito and Katsura 1992; Suzuki et al. 1995) in the middle of the peninsula (Nakajima and Watanabe 2005; Utsunomiya and Ooi 2019). The Kazusa Group is divided into 13 formations based on lithofacies and stratigraphic positions. The lower and middle parts of the Kazusa Group consist of deep-sea (submarine fan and basin plain), upper slope, and outer shelf deposits, whereas the upper parts of the Kazusa Group comprise shallow marine deposits (Ito and Katsura 1992). The Kazusa Group laps onto the Miura Group, and this resulted in the base of the Kazusa Group younging toward the central part of the peninsula. The basal sandstone and conglomerate succession called the Kurotaki Formation is thought to be contemporaneous with the laterally adjacent turbidite successions of the lower part of the Kazusa Group on the eastern and western sides of the peninsula (Kawabe et al. 1981; Ito et al. 1992; Nakajima and Watanabe 2005). The paleoslope dips southeastward in the lower part of the Kazusa Group and northeastward in the middle and upper parts, as inferred from paleocurrent analyses of the turbidites (Hirayama and Nakajima 1977; Tokuhashi 1992).

2.4 Widespread tephra beds used for inter- and intra-basinal correlation

There are many widespread vitric ash beds intercalated in the Miura Group and the lower parts of the Kazusa Group on the Boso Peninsula, and these tephra beds are also distributed around the Tokai, Kinki, and Chubu districts in the western Japanese islands. In the Miura Group, tephra beds An51, An53, An77, An85, An112, An129, and An130 are regarded as widespread tephra beds named Trb1-Ya4, Sk-Ya5, Ksg-An77, Znp-Ohta, Ymp-SF8.3, Hgs-An129, and Sr-Ity, respectively (Kurokawa and Higuchi 2004; Satoguchi et al. 2005; Tamura et al. 2008; Tamura and Yamazaki 2010; Satoguchi and Nagahashi 2012). In the lower parts of the Kazusa Group, KW2, KH1, MY, KB, HS C, Kd44, Kd39, Kd25, and Kd24 are regarded as widespread tephra beds named Fup-KW2, Obr-Bnd1, OM1-OKIII, Bnd2-O1, Tmg-R4, Kd44-Nk, Ho-Kd39, Eb-Fukuda, Om-SK110, and Srt-SK100, respectively (Nagahashi et al. 2000; Suzuki and Nakayama 2007; Tamura et al. 2019). Moreover, some thick (several meters) crystal-rich tephra beds with a “salt and pepper” appearance have been used for intra-basinal correlation. For example, Am78, Ky21, and Ky26 on the Boso Peninsula have been correlated with Ok, Hk, and Nt on the Miura Peninsula (Urabe et al. 1990; Urabe 1992; Suzuki et al. 1995).

3 Methods

We used visual observations to describe the lithologies and thicknesses of tephra beds, and we classified the tephra beds according to the dominant grain size as fine ash (< 63 μm), coarse ash (63 μm–2 mm), and lapilli (> 2 mm), following the classification scheme of Heiken and Wohletz (1985). Lapilli-dominated beds were classified into pumice-rich lapilli beds and scoria-rich lapilli beds, based on the dominant grain component. Tephra samples (10–50 cc) were washed using a disposable nylon mesh sieve to remove fine-grained particles (< 63 μm), and they were then cleaned using gentle ultrasonication to remove clay minerals. After that, the samples were dried at less than 60 °C for half a day in a constant-temperature drying oven. The very-fine sand-sized particles (63–125 μm) were collected using a disposable nylon mesh sieve with a pore size of 125 μm, and these particles were then fixed onto a glass slide using the optical adhesive Nd = 1.545, produced by Furusawa Geological Survey, Co. Ltd., and then cured under ultraviolet light. These slides were then used to examine the shapes of glass shards and to identify the dominant heavy minerals using a polarizing light microscope (BX-53, Olympus Co., Tokyo, Japan) with a magnification of 100 × under both cross-polarized and plane-polarized light. The glass shard classification used follows Kishi and Miyawaki (1996). Refractive index measurements of glass shards were performed using the refractometer system MAIOT (Furusawa Geological Survey, Co. Ltd.), as described by Furusawa (1995). The major and trace element compositions of the glass shards (10–15 grains) were determined using EDX (energy dispersive X-ray spectrometry) and LA–ICP–MS (laser ablation inductively coupled plasma mass spectrometry), following Furusawa (2017). Major element compositions were determined using EDX (EMAX Evolution EX-270, HORIBA, Ltd.) by scanning 4 × 4 μm areas of the polished surfaces of glass shards with an electron beam (90 nm beam size) for 50 s under an SEM (scanning electron microscope SU1510, HITACHI High-Tech Corp). Trace element compositions were measured bringing particles generated from a laser ablation system (LSX-213 G2 + [Nd: YAG 213 nm], TELECYNE, Ltd.) into ICP–MS (iCAP Qc, Thermo Fisher Scientific, Inc.) using helium as the carrier gas. GSE-1G from the United States Geological Survey was used as a reference glass.

The sediment samples used for analyses of calcareous nannofossils were collected from fresh outcrop surfaces using a chisel or portable drill. Smear slide samples were prepared by fixing sediment samples onto a glass slide using the optical adhesive GL-4006 (Gluelabo Ltd., Mie, Japan) and then cured under ultraviolet light. These slides were then examined using a polarizing light microscope (BX-53, Olympus Co., Tokyo, Japan) with a magnification of 1500 × under cross-polarized and plane-polarized light. We used the zonal schemes of Okada and Bukry (1980). The ages of the datum planes follow Raffi et al. (2006) and Raffi et al. (2020).

4 Results

4.1 Lithostratigraphy and sedimentary structures

In this section, we describe the local stratigraphic and sedimentary characteristics for deposits around the major erosional surface in the following three regions: the Miura Peninsula (Figs. 3, 4), the western Boso Peninsula (Fig. 5), and the eastern Boso Peninsula (Fig. 6).

Fig. 6
figure 6

From left to right, cross section, geologic map, geologic columns, and stratigraphic distribution of calcareous nannofossils in Onaura, Katsuura City, eastern Boso Peninsula. The lithologic legend is shown in Fig. 4

4.1.1 Miura peninsula

The uppermost part of the Zushi Formation, the Jimmuji Member, and the upper part of the Ikego Formation are exposed in Zushi City (Fig. 3a). The Zushi Formation consists mainly of mudstones in which thin (< 20 cm) sheet-like sandstones and two vitric tephra beds Nt (Urabe 1998) and the newly named Nagoe (Ng) are intercalated. Ng is a synonym of the Nt-mid of Eto (1993). Ng is intercalated 8–9 m below Nt (Additional file 1: Fig. A1) at a location 100 m east of the outcrop described by Urabe (1998). Both these tephra beds can also be recognized in the C1 outcrop that was exposed during residential constructions (Fig. 3b). The two basal slide planes of the mass-transport deposits of the Jimmuji Member are located 37 and 38 m above the top of Nt (Fig. 3b; Eto et al. 1993; Yokohama Defense Facilities Administration Bureau 1993). Chaotic deposits and slide blocks of sandstone and mudstone in the Jimmuji Member are well exposed in the A1 outcrop (Fig. 3b). The Jimmuji Member becomes thinner and pinches out westward, where a tens of cm-thick pebbly sandstone bed covers the erosional surface (Figs. 3b, d, 7e). The upper part of the Ikego Formation (90 m thick) is exposed continuously along traverses 1 and 2 of Utsunomiya et al. (2017) (Figs. 4, Additional file 1: A1, A2) and along traverses Z1 and Z2 in the Ikego Forest Natural Park (Fig. 3c, d). Tephra beds were numbered in ascending stratigraphic order and prefixed with “IkT”, and some of these were traced laterally by Utsunomiya et al. (2017) (Fig. 4). The stratigraphic levels of all IkT tephra beds are shown in Additional file 1: Fig. A2.

Fig. 7
figure 7

Outcrops of ae pebbly sandstone beds covering the major erosional surface and f, g coeval MTDs (mass-transport deposits) formed at around 3.2 Ma. ae A poorly-sorted pebbly sandstone with mud clasts on the major erosional surface exposed at section F3 in Seki a, b on the western Boso Peninsula, c, d at section K1 on the eastern Boso Peninsula, and e section Z1 on the Miura Peninsula. See Figs. 3, 5c, 6 for the localities. f, g Mass-transport deposits exposed f in Nagasaki on the western Boso Peninsula and g along traverse 3 on the Miura Peninsula. See Fig. 5 and Utsunomiya et al. (2017) for the localities. Blue arrows in b, d and e indicate the major erosional surface. White arrows in d and e indicate the mud clasts

Here we provide details of the lithologies and chemical compositions of IkT16 and IkT19 in the Ikego Formation and Ahn in the Urago Formation. IkT19 was first reported by Tamura et al. (2014) as Ikego 1 (Ikg1), which was correlated with a tephra bed named In1 in the Inubo Group on the Choshi Peninsula, based on petrography and the chemical compositions of glass shards. Here, we use the tephra name IkT19 instead of Ikg1 for consistency with the numbering used by Utsunomiya et al. (2017). Ahn in the Urago Formation was reported by Tamura and Yamazaki (2010), and this bed is the same as AhG named by Tamura et al. (2010). This tephra bed has been correlated with the widespread tephra UN-MD2 (2.6–2.7 Ma), based on petrography and the chemical compositions of glass shards (Tamura and Yamazaki 2010).

4.1.2 Western Boso peninsula

The uppermost part of the Anno Formation is well exposed along the Minato River and the branch streams in Futtsu City, on the western side of the Boso Peninsula (Figs. 2b, 5). Stratigraphic cross-sections (Fig. 5a) from east to west (Seki–Iwamoto–Kunawa) and from south to north (Kunawa–Terao, Nagasaki) show that the thickness between An154 and An157 varies laterally owing to erosion by submarine landslide(s) that resulted in a mass-transport deposit (MTD) (Figs. 5d, 7f). This MTD is named the “Sak slump” (Nakajima 2005), and the equivalent erosional surface cuts the underlying strata (blue line in Fig. 5a). At outcrop, the MTD pinches out eastward, and pebbly sandstones cover the erosional surface (Figs. 5d, 7a, b). Undulating parallel or low-angle cross stratification is weakly developed in these sandstones. We have not determined the paleocurrent directions of the pebbly sandstone beds, so it remains unclear whether the low-angle cross lamination is foreset or backset.

We constructed geologic columns along traverses F1and F2 in Nagasaki (Fig. 5) where a 5 m-thick Sak slump deposit is intercalated between tephra beds An156-4 and An157. In these sections, tuffaceous sandy mudstones (13 m thick) are successively exposed, and they overlie the Sak slump that consists of folded blocks of alternating beds of sandstone and sandy mudstone (Fig. 5). Most of the tephra beds in the uppermost part of the Anno Formation are coarse-ash-dominated beds and lapilli-dominated beds. The 40 cm-thick scoria-rich lapilli bed An158 (Nakajima 2005) is intercalated 11 m above the top of the sandstone that covers the Sak slump (Fig. 5d). Although Nakajima (2005) in a broad sense named a set of tephra beds that included An158 as An158 for the purposes of stratigraphic correlation, here, in a stricter sense, we consider just the 40 cm-thick scoria-rich lapilli bed as An158. Two fine vitric ash beds are intercalated below and above An158; herein, the lower vitric tephra bed is An157.5, and the upper bed is An158.5 (Fig. 5d).

The Kiyosumi and the lower part of the Anno formations are successively exposed along a branch stream of the Koito River in Okugome, Kimitsu City, in the central part of the Boso Peninsula (Additional file 1: Fig. A1), and this is where the marker tephra beds Ky25 and Ky26 are intercalated (Nakajima et al. 1981; Natural History Museum and Institute, Chiba, 1994). Ky26 is intercalated 6 m above Ky25 in the sandstone-dominated alternating beds of sandstone and mudstone (Additional file 1: Fig. A1).

4.1.3 Eastern Boso peninsula

The Kiyosumi Formation and the overlying Kurotaki and Katsuura formations are exposed at Onaura in Katsuura City (Fig. 6). These formations are cut by NE–SW-trending normal faults. The Kiyosumi Formation dips gently to the northeast and is overlain by the Kurotaki Formation, with the subparallel erosional surface displayed at section K1 (Fig. 7c, d).

At locality K1, a 50-cm-thick pebbly sandstone overlies the turbidite deposits of the Kiyosumi Formation on a subparallel erosional surface (Fig. 6). Unlike the submarine canyon fills with terrestrial links in the Kazusa Group (Yamauchi et al. 1990; Ito and Saito 2006), the sand and gravel of the scar fill is characterized by mud clasts that consist mainly of volcanic rocks and scoria, similar to the pyroclastic material in the Anno Formation. Rather than being massive sand and gravel, the sedimentary rocks display undulating parallel or low-angle cross stratification and are poorly sorted (Figs. 6, 7c, d). Calcareous nannofossil biozones in the underlying Kiyosumi and overlying Kurotaki formations can be assigned to CN10-CN11 and CN12a, respectively (Fig. 6). Above this, the muddy sandstone correlated with the Kurotaki Formation shows an upward-fining succession into the Katsuura Formation, in which angular unconformities can be observed (Figs. 6, 8a, b, d). The surface of the unconformity between geologic columns K2 and K3 (Fig. 6) is covered by a thin sand layer with burrow structures (Fig. 8d). This upward-fining succession has been interpreted as a transgression from shallow water into the sedimentary facies of a distal submarine fan of the Kazusa Group (Ito 1992b). Despite these angular unconformities, the temporal gap is within the CN12b subzone (2.8–2.5 Ma), and this gap is much smaller than that represented by the erosional base of the pebbly sandstone bed. Tephra bed Onr was first reported from the Kurotaki Formation (Fig. 6). We could not identify other marker tephra beds, most likely because of the coarser facies.

Fig. 8
figure 8

Outcrops of the angular unconformity a, b, d and channel fill deposits c, e in the Kurotaki Formation, which are dated between 2.8–2.5 and 2.5–2.0 Ma, younger than the major erosional event. a Angular unconformity exposed in the quarry cliff at Loc. K3. b Angular unconformity exposed in the sea cliff. c Channel fill deposits of conglomerate of the Kurotaki Formation overlying the Kiyosumi Formation at Loc. K4. d Close-up photograph of a. The angular unconformity is covered by a thin, volcaniclastic-rich fine-grained sand bed, which is laterally obscured due to bioturbation. A burrow cutting the underlying strata can be observed. e Close-up photograph of basal conglomerate of the channel fill deposits. Yellow dashed lines in a and b indicate the bedding plane. The white arrows in a, b, and d indicate the angular unconformity. Red arrows in c and e indicate the erosional surface underlying the channel-fill deposits

Channel-fill conglomerate (Kawabe et al. 1980; Ito 1992b; Ito et al. 1992) cuts all the sedimentary sequences described above at section K4 (Fig. 6). The erosional base of this conglomerate shows a staircase-like geometry (Fig. 8c, e). Calcareous nannofossil assemblages in the mudstone immediately above the channel-fill conglomerate indicate the CN12d subzone (2.5–2.0 Ma) at section K4, which is considerably younger than the erosional surface at section K1.

4.2 Descriptions and remarks on the chemical compositions of the tephra beds

Here we describe ten tephra beds, and for eight of these we determined the refractive indices of the glass shards and their chemical compositions.

4.2.1 Nt

Tephra bed Nt consists of a 37 cm-thick crystal-vitric lower unit (medium-sand to coarse-sand size), a 33 cm-thick crystal-vitric middle unit (coarse-sand size), and a 57 cm-thick crystal-vitric upper unit (silt size). These units would correspond to the Nt-1, Nt-2, and Nt-3 units described by Urabe (1998). The tephra sample from 10 cm above the base of Nt is a coarse crystal-vitric ash characterized by orthopyroxene as the dominant heavy mineral with a little hornblende and clinopyroxene. We did not determine the refractive indices or chemical compositions of the glass shards in this tephra bed.

4.2.2 Ng

Tephra bed Ng is a 15 cm-thick vitric ash bed consisting of a white-colored 8 cm-thick lower unit (very-fine-sand- to silt-sized) and a light-gray-colored 7 cm-thick upper unit (silt size) (Fig. 9). The boundary between the upper unit and the overlying tuffaceous mudstone is gradual. The heavy minerals are dominantly orthopyroxene along with hornblende and a little clinopyroxene and biotite (Table 1). Glass shards of bubble-wall, small bubble, stripe, sponge, and fiber types were observed (Table 1), and the refractive indices of the glass shards (n) range from 1.502 to 1.505 with a mode at 1.503–1.504 and with 1.498 as an outlier (Table 1). The glass shards contain 3.41 wt.% K2O and have relatively high (583 ppm) values of Ba (Table 2).

Fig. 9
figure 9

Geologic columns showing lithologies and sampling horizons, and photographs of the correlated tephra beds

Table 1 List showing location, glass type and refractive index of the tephra layers
Table 2 Major and trace element compositions of the volcanic glass shards

4.2.3 IkT16

Tephra bed IkT16 is a white-colored 10–12 cm-thick vitric ash bed with very-fine-sand- to silt-sized particles (Fig. 9). Glass shards of small-bubble, stripe, and sponge types dominate over the bubble-wall type. Heavy minerals are rare in this tephra bed. The refractive indices of the glass shards (n) range from 1.507 to 1.511 with a mode at 1.509–1.510 (Table 1). The glass shards have relatively low SiO2 (76.95 wt.%) and K2O (1.15 wt.%) contents compared with the other analyzed tephra beds but relatively high FeO (2.51 wt.%), CaO (2.44 wt.%), and Al2O3 (12.59 wt.%) contents (Table 2).

4.2.4 IkT19

Tephra bed IkT19 is a white-colored 6–7 cm-thick vitric ash bed with a very-fine sand to silt grain size (Fig. 9). Coarse-sand- to granule-sized pumice grains are scattered immediately below this tephra bed. Glass shards of bubble-wall type dominate over the other types. Heavy minerals are rare in this tephra bed. The refractive indices of the glass shards (n) range from 1.500 to 1.506 with a mode at 1.503–1.504 (Table 1). The glass shards have relatively low SiO2 (78.43 wt.%) and K2O (1.47 wt.%) contents and slightly high contents of Y (44.7 ppm) (Table 2).

4.2.5 Ahn

Tephra bed Ahn is a white-colored 6 cm-thick vitric ash bed with a silt grain size (Fig. 9). A fine-grained sandstone bed is intercalated immediately below this tephra. Glass shards of bubble-wall and stripe types dominate over the other types. The heavy minerals are dominated by biotite, with a small amount of orthopyroxene (Table 1). The refractive indices of the glass shards (n) range from 1.498 to 1.500 (Table 1). The glass shards have relatively low contents of SiO2 (77.31 wt.%), FeO (0.98 wt.%), and CaO (0.78 wt.%) and slightly high contents of K2O (4.80 wt.%) (Table 2).

4.2.6 Ky26

Tephra bed Ky26 consists of a 22 cm-thick crystal-vitric lower unit (medium-sand to coarse-sand size), a 29 cm-thick crystal-vitric middle unit (coarse-sand size), and a 100 cm-thick crystal-vitric upper unit that shows convolute lamination (silt size). These units correspond to the Nt-1, Nt-2, and Nt-3 units described by Urabe (1998). The tephra sample obtained 10 cm above the base of Nt consists of a coarse crystal-vitric ash with orthopyroxene as the dominant heavy mineral and minor hornblende and clinopyroxene. We did not determine the refractive indices and chemical compositions of the glass shards in this tephra bed.

4.2.7 Ky25

Tephra bed Ky25 is a 13 cm-thick vitric ash bed that consists of a white-colored 6 cm-thick lower unit (very-fine-sand- to silt-sized) and a light-gray-colored 7 cm-thick upper unit (silt size) (Fig. 9). The boundary between the upper unit and the overlying tuffaceous mudstone is gradual. The heavy minerals are dominated by orthopyroxene with some hornblende and clinopyroxene and a little biotite (Table 1). Glass shards of bubble-wall, small bubble, stripe, sponge, and fiber types were observed (Table 1). The refractive indices of the glass shards (n) range from 1.503 to 1.505 with a mode at 1.503–1.504 (Table 1). The glass shards contain 3.46 wt.% K2O and have relatively high (542 ppm) contents of Ba (Table 2).

4.2.8 An157.5

Tephra bed An157.5 is a white-colored 5–6 cm-thick vitric ash bed of very-fine-sand- to silt-sized volcanic ash (Fig. 9). Glass shards of small-bubble, stripe, and sponge types dominate over the bubble-wall type. Heavy minerals are rare in this tephra bed. The refractive indices of the glass shards (n) range from 1.508 to 1.510, with a mode at 1.508–1.509 and with 1.513 as an outlier (Table 1). The glass shards have lower contents of SiO2 (76.86 wt.%) and K2O (1.15 wt.%) than the other analyzed tephra beds (Table 2) but higher contents of FeO (2.59 wt.%), CaO (2.47 wt.%), Al2O3 (12.56 wt.%), and Sr (136.2 ppm).

4.2.9 An158.5

Tephra bed An158.5 is a white-colored 2–4 cm-thick vitric ash bed of very-fine-sand- to silt-sized volcanic ash, and it is overlain by a 1–2 cm-thick fine sandstone bed and a 5 cm-thick white-colored fine ash bed in ascending order (Fig. 9). Coarse-sand- to granule-sized pumice grains are scattered immediately below this tephra bed. Glass shards of bubble-wall type dominate over the other types. Heavy minerals are rare. The refractive indices of the glass shards (n) range from 1.500 to 1.507 with a mode at 1.503–1.504 (Table 1). The glass shards have relatively low contents of SiO2 (78.51 wt.%) and K2O (1.53 wt.%) and slightly high contents of Y (47.3 ppm) (Table 2).

4.2.10 Onr

Tephra bed Onr is a white-colored 10 cm-thick vitric ash bed with a silt grain size (Fig. 9). A fine-grained sandstone bed is intercalated immediately below this tephra bed. Glass shards of bubble-wall and stripe types are dominant. The heavy minerals are dominantly biotite, along with a little orthopyroxene and hornblende (Table 1). The refractive indices of the glass shards (n) range from 1.497 to 1.499 (Table 1). The glass shards have relatively low contents of SiO2 (77.34 wt.%), FeO (0.95 wt.%), and CaO (0.80 wt.%) but relatively high contents of K2O (4.76 wt.%) (Table 2).

5 Discussion

5.1 Tephra correlation

Based on previous calcareous nannofossil biostratigraphic and magnetostratigraphic studies, the upper part of the Ikego Formation above the Jimmuji Member on the Miura Peninsula may be the time-equivalent of the uppermost part of the Anno Formation on the Boso Peninsula. In addition, the uppermost part of the Zushi Formation on the Miura Peninsula can be correlated with the uppermost part of the Kiyosumi Formation. The thicknesses, lithologic features, and compositions of minerals in Nt, Ng, IkT16, IkT19, and Ahn on the Miura Peninsula and Ky26, Ky25, An157.5, An158.5, and Onr on the Boso Peninsula are in good agreement with each other, as are the shapes, refractive indices, and major-element compositions of their glass shards (Figs. 9, 10, Tables 1, 2). The minor-element compositions of these tephra beds are also similar (Table 2). The depositional ages of these tephra beds will be discussed below.

Fig. 10
figure 10

Major element compositions of volcanic glass shards in the tephra beds

The mudstone immediately below Nt was assigned to the CN10c subzone of Okada and Bukry (1980), which is defined as the interval between the LAD (last appearance of datum) of Ceratolithus acutus (5.3 Ma) and the LAD of Amaurolithus spp. (4.5 Ma) (Kanie et al. 1991). Kameo et al. (2010) determined the LAD of Amaurolithus spp. (4.5 Ma) to be between the Ky29 and Ky31 tephra beds in the studied traverse along the branch stream of the Koito River in Okugome. They also demonstrated stratigraphic variations in the sizes of the coccoliths Reticulofenestra spp., and they assigned Ky25 and Ky26 to the intervals IIa of Kameo and Bralower (2000) and A1 of Takayama (1993). They interpreted Ky25 and Ky26 to be above the C3n.3n normal subchronozone (4.7–4.9 Ma), based on the chronostratigraphy established in ODP Hole 806B. Therefore, Ng and Ky25 can be correlated with each other on the basis of the calcareous nannofossil biostratigraphy. This supports the correlation of Nt and Ky26, as suggested by previous researchers (Urabe et al. 1990; Suzuki et al. 1995).

IkT16 and IkT19 are located around the top (3.21 Ma) of the Mammoth subchronozone (C2An.2r) (Utsunomiya et al. 2017), and An158 is located above the top of the Mammoth subchronozone (Haneda and Okada 2019). Utsunomiya et al. (2017) showed that IkT19 is located in the normal subchronozone immediately above the Mammoth subchronozone (Additional file 1: Fig. A2). Normal polarity was detected immediately above IkT16, and reverse polarity was detected 2 m below IkT16 (Additional file 1: Fig. A2). Although An157.5 and An158.5 have not been reported from the Shikoma River (Haneda and Okada 2019), the top of the Mammoth subchronozone was estimated to be between 3.0 and 3.6 m below An158 (Haneda and Okada 2019). Therefore, IkT16 and IkT19 can be correlated with An157.5 and An158.5, respectively, which is consistent with the calcareous nannofossil biostratigraphy and magnetostratigraphy. The two tephra beds IkT16-An157.5 and IkT19-An158.5 are significant stratigraphic indicators for the top of the Mammoth subchronozone.

An158, a 40 cm-thick scoria-rich lapilli bed, can be traced laterally, whereas both the fine ash beds An157.5 and An158.5 (found in Nagasaki) are rare in Seki (Fig. 5d). The absence of fine ash beds is a probable result of strong bottom currents, which are considered to have been dominant when the coarser facies was deposited (Utsunomiya et al. 2015, 2017), and this can explain the absence of An157.5 and An158.5 in other areas.

Ahn is within the CN12b or CN12c subzones (Fig. 4), whereas Onr is within the CN12b subzone (Fig. 6). Ahn has been correlated with the widespread tephra UN-MD2, which has been assigned an age of 2.6–2.7 Ma (Tamura and Yamazaki 2010). The depositional ages are therefore consistent for both these tephra beds.

5.2 Major erosion at around 3.2 Ma and the resultant deposits

We have shown that the stratigraphic horizon between Nt-Ky26 and IkT16-An157.5 on the western side of the Boso Peninsula becomes remarkably thinner on the Miura Peninsula. Moreover, we have not found the widespread tephra beds of the Anno Formation (An51, An53, An77, An85, An112, An129, and An130) on the Miura Peninsula (Fig. 11a). To explain this, we will discuss here whether sedimentation rates were significantly different in the rocks of the Miura and Boso peninsulas. The thickness of the strata between Hk and Nt is ~ 200 m, based on the stratigraphy of Suzuki et al. (1995) and Eto (1986b, a), and the thickness between Ky21 and Ky26 is ~ 180 m, based on Tokuhashi and Ishihara (2008). Assuming the ages of Hk-Ky21 and Nt-Ky26 are ca. 4.9 Ma (near the base of C3n.3n: Niitsuma 1976 and Takahashi 2008) and ca. 4.7 Ma (above the top of C3n.3n), respectively, the rates of sedimentation between the two tephra beds are estimated to be ~ 100 cm/ky on the Miura Peninsula and ~ 90 cm/ky on the Boso Peninsula. The rates of sedimentation in the Mammoth subchronozone of the Anno Formation range from 9 to 66 cm/kyr (Haneda and Okada 2019), whereas the rates for the interval between the Mammoth and Kaena subchronozones in the upper part of the Ikego Formation range from 29 to 36 cm/kyr (Utsunomiya et al. 2017). Therefore, the thicknesses and sedimentation rates of the correlated stratigraphic intervals on the Miura Peninsula and the western Boso Peninsula do not differ markedly.

Fig. 11
figure 11

Stratigraphic correlations and chronostratigraphic model with interpretation of regional relationships of subsidence and uplift that contributed to segmentation of the Miura–Boso forearc basin. a Stratigraphic correlations of the Miura, western Boso, and eastern Boso peninsulas. Stratigraphic divisions, calcareous nannofossil biostratigraphy, and magnetostratigraphy are based on our data sets and those of previous researchers (Urabe et al. 1990; Eto 1993; Eto et al. 1998; Kameo et al. 2010; Tamura and Yamazaki 2010; Kameo and Sekine 2013; Utsunomiya et al. 2017; Haneda and Okada 2019). b Stratigraphic model against a chronological axis. c Our interpretation of the regional relationships of subsidence and uplift that contributed to segmentation of the Miura–Boso forearc basin. The Kurotaki unconformity can be understood in terms of the major erosional event at around 3.2 Ma and several subsequent erosional events and overall onlapping deposition

As an alternative idea, we can consider whether the Jimmuji Member and the Sak slump, which are mass-transport deposits, show that submarine landslide(s) contributed to the stratigraphic discontinuity. The age interval of the missing horizons is at least 1.3 Myr, if the top of the Zushi Formation and base of the upper part of the Ikego Formation are assigned ages of 4.5 Ma (top of Amaurolithus spp.) and 3.2 Ma (top of Mammoth subchronozone), respectively (Fig. 11b). A linear extrapolation using the sedimentation rate of the upper part of the Ikego Fomation (29–36 cm/kyr) suggests that at least 370-m-thick strata was eroded owing to the submarine landslide(s) that resulted in the formation of the Jimmuji Member.

The top of the Jimmuji Member and the top of the Sak slump are located 11 and 10 m below the IkT16-An157.5 tephra bed, respectively (Fig. 5). A similar stratigraphic level at the top of these deposits as well as residual blocks suggest that submarine landslide events resulted in the formation of both the Jimmuji Member and the Sak slump at the same time around 3.2 Ma.

In the eastern Boso Peninsula, a conglomerate in the K4 outcrop was considered by Ito et al. (1992) to have filled a channel that cut the Anno Formation during a sea-level lowstand during the early stage of formation of the Kurotaki unconformity. Some previous researchers have emphasized the role of submarine landslides as an essential erosional process, based on observations of the outcrop near section K4 (Kawabe et al. 1980; Otsubo et al. 2011). However, our geologic survey and biostratigraphic studies indicate that the Anno Formation had already been eroded away before the channel cut the strata sometime between 2.5 and 2.0 Ma. Rather than the channel deposits at section K4, the base of the pebbly sandstone bed at section K1 actually marks most of the major erosion in this area (Fig. 6). We could not identify the marker tephra bed immediately above the major erosional surface at section K1, and this resulted in a lack of an accurate age constraint (the CN12a zone lies between 3.5 and 2.76 Ma) for the major erosional surface in this region. Despite the uncertainties in the ages, we interpret the base of the Kurotaki Formation in the eastern Boso region to have been deposited immediately after the 3.2 Ma event (Fig. 11), based on the lithologic similarities of the pebbly sandstones and the bedding patterns and the consistency in nannofossil ages.

The basal pebbly sandstone pervasively covers a major erosional surface (Fig. 7) and consists of a poorly sorted pebbly sandstone that contains mud clasts and shows evidence of bed-load traction. These pebbly sandstone beds vary laterally into thick (up to 60 m) mass-transport deposits and into the overlying turbidite sandstones. Similar turbidite sandstones to those overlying the irregular upper surface of the MTD in this area have been reported from other MTDs (Yamamoto et al. 2007; Dykstra et al. 2011; Utsunomiya and Yamamoto 2019), and they become pebbly sandstone beds with sharp, flat erosional surfaces. The deposits resting on the basal erosional surfaces include a wide range of grain sizes, but they are dominated by granule- to pebble-sized pyroclastic fragments in medium–coarse-grained sandstones. Such coarse deposits are unusual in the lower part of the Kazusa Group, which is generally characterized by alternating beds of sheet-like turbidites and hemipelagic mudstones (Ito and Katsura 1992; Tokuhashi 1992; Utsunomiya 2019), suggesting a supply from lapilli beds in the collapsed section of the Anno Formation owing to a submarine landslide. Another candidate for a sedimentary source would be ocean-current-induced pyroclastic-rich sand-ridge deposits that formed on an upper slope, as seen in the Urago Formation (Utsunomiya et al. 2015), because it is possible to assume sand-ridge deposits were developed on the excavated upper slope. In either case, the pebbly sandstone would have been deposited rapidly, resulting in poor sorting.

5.3 Kurotaki unconformity and forearc basin segmentation

The stratigraphic context and tectonic significance of the Kurotaki unconformity between the Miura and Kazusa groups has long been debated following the work of Koike (1951). However, no previous researchers have discussed the differential erosional features described here, and we suggest that these features provide key information for explaining regional differences in the temporal gap. Here, we propose a new stratigraphic model that assumes the segmentation of the forearc basin by a residual topographic high into two distinct sub-basins (Fig. 11a). An important point in this model is that there is a time gap between ca. 3 and 2 Ma (Kameo and Sekine 2013) that compensates for the smaller amount of erosion at 3.2 Ma in the western Boso region (duration of 90 kyr, Haneda and Okada 2019) (Fig. 11b). This suggests the presence of a residual topographic high in the western Boso region, where sedimentation would have been punctuated, whereas the depressions created by erosional event(s) in the Miura and eastern Boso regions remained for the accommodation of subsequent deposits.

The Kazusa Group probably filled the field of subsidence in the eastern Boso region, as indicated by seismic surveys that showed strata more than 1000 m thick underlying the Katsuura Formation in the eastern offshore region of the Boso Peninsula (Okubo et al. 1990; Ozaki et al. 2019). In addition, there is no difference in the consolidation yield stresses of the Miura and Kazusa groups in the eastern Boso region, which suggests that this region was undergoing continuous subsidence while the western Boso region was being uplifted by tectonic movement (Kamiya et al. 2017, 2020). The thick lower part of the Kazusa Group offshore from the Boso Peninsula is presumed to thin and be draped on the major erosional surface toward section K1 on the eastern shore of the Boso Peninsula (Fig. 11a), and this situation is similar to the thin-bedded fine-grained turbidites that are draped over the slope (an aggradational onlap pattern; Smith and Joseph 2004). At section K1, we can therefore observe such layers that have been draped on the scarps of basin-wide submarine landslide(s), but they are generally too thin to be imaged using seismic data (Strasser et al. 2011). Similarly, syn-depositional subsidence would have occurred in the Miura region, because several hundred meters of sediment was deposited at paleobathymetries of 500–2000 m or 400–600 m (Eto et al. 1987; Utsunomiya and Majima 2012) after the major erosional event(s) that resulted in the formation of the Jimmuji Member. Our model therefore assumes uplift in the central to western Boso Peninsula but subsidence in the Miura Peninsula and eastern Boso Peninsula, thus inducing forearc basin segmentation (Fig. 11c).

Forearc basin segmentation is also observed in the basins that developed on the Pacific side of Japan, and these basins are divided by N–S trending topographic highs that were uplifted from the Pleistocene (Okamura 1990; Sugiyama 1992; Okamura and Blume 1993). The passage of undulations in the upper surface of the Philippine Sea Plate during its oblique subduction was considered to create alternating zones of uplift and subsidence in the forearc regions (Okamura and Shishikura 2020). The presence of middle Miocene igneous basement rocks around the Kii Peninsula would also enhance the segmentation of the forearc basin (Kimura et al. 2022). Oblique subduction is closely related to the change in direction of the Philippine Sea Plate from N to NW, but whether it occurred during the Pliocene (e.g., Seno and Maruyama 1984; Sugiyama 1992; Takahashi 2006, 2017) or Pleistocene (Nakamura et al. 1984) is subject to debate. The spatial switch of erosion and deposition in the forearc basins along the Pacific side of Japan would have occurred during this tectonic reconfiguration. Although tectonic reconstructions are beyond the scope of this paper because of the complicated and frequent changes in stress regime in this region (e.g., Angelier and Huchon 1987; Yamaji 2000; Otsubo et al. 2017; Utsunomiya and Otsubo 2022), additional data on the structure and orientation of folds and faults would be required, which together with our new stratigraphic framework, would help us reach a better understanding of the tectonic contribution to forearc basin segmentation.

6 Conclusions

We correlated the tephra beds that are intercalated in the forearc basin fill on the Miura and Boso peninsulas using stratigraphic levels, lithologic characteristics, and the chemical compositions of glass shards. The tephra correlations are consistent with the biostratigraphy and magnetostratigraphy established by our new work and that of previous researchers. These tephra beds are important chronological indicators, and they contribute to the construction of composite sections through the Pliocene in the forearc basin that can be used for studies of the regional paleoenvironment and basin evolution. We identified previously overlooked major erosional surfaces, which represent basin-scale mass-wasting event(s) that are associated with a disconformity in the upper Mammoth subchronozone (Mammoth event: 3.2 Ma). We present a new stratigraphic model that takes into account the geometry and lateral variation of the Kurotaki unconformity throughout the basin, and this is the first work on this unconformity since it was described 70 years ago by Koike (1951). The Mammoth event contributed to preservation of a continuous sedimentary record during the period 3.2 to 2.4 Ma on the eastern and western sides of the basin, whereas later local erosional event(s) occurred on the eastern side of the basin. The interaction between sedimentary systems and local relief owing to erosion and residual blocks in the mass transport deposits (MTDs), as well as bottom-current-induced bypassing of fine-grained sediments, had a major influence on the spatial–temporal variations in erosional thickness and sedimentation after the major erosional event at 3.2 Ma. This paper is the first to provide a detailed stratigraphic framework for the Miura–Boso forearc basin that is based on the concept of basin segmentation that resulted in differential erosion. We hope this work will further enhance our understanding of the evolution of forearc basins.