The Shimanto Group records a Cretaceous through Miocene history of subduction/accretion along the southwest margin of Japan. We used vitrinite reflectance on over 200 samples of shale and slate to determine regional trends in diagenesis and low-grade metamorphism on the Muroto Peninsula, Shikoku Island. Thermal structure throughout the section overprints all but the latest stage of a complicated polyphase structural evolution. Eocene strata of the Murotohanto subbelt display the highest levels of thermal maturity. Reflectance values range from 1.4 to 5.0%R m ; thermal maturity increases from north to south, and the %R m values correspond to estimates of paleotemperatures of 180 to 315°C. The Shiina-Narashi fault marks the boundary between Eocene rocks and Oligocene-Miocene strata of the Nabae subbelt. South of this fault, thermal maturity for the Nabae strata ranges from 0.9 to 3.7%R m , and these values correspond to estimates of peak paleotemperature of 140 to 280°C. Reflectance values increase in proximity to the Maruyama intrusive suite, and the maximum rock temperatures adjacent to the intrusions may have been as high as 500 to 550°C based on comparisons with laboratory heating experiments. Paleotemperatures within the Upper Shimanto Group were unusually high compared to shallow levels of other accretionary complexes. The geothermal gradient was at least 40°C/km on a regional scale and much higher locally. Subduction of juvenile oceanic crust can account for the anomalous levels of thermal maturity documented across the Muroto Peninsula.

The primary purpose of this special publication is to show how Eocene through middle Miocene strata of the Shimanto Belt were affected by the Cenozoic geothermal regime of southwest Japan. This introductory paper synthesizes the regional and local geologic framework, as it applies to detailed studies completed on the Muroto Peninsula of Shikoku. In particular, we focus on the temporal and geometric relations between peak heating events and discrete stages in the deformation history. Rocks of the Shimanto Belt display the effects of a complicated history of polyphase folding, faulting, and cleavage formation. Interpretations of the structural geology follow conventional models for subduction-accretion via offscraping, underplating, and out-of-sequence thrusting. The positions of plate boundaries during the Paleogene phase of plate convergence in southwest Japan are somewhat difficult to substantiate, but that episode of tectonism may have been punctuated by subduction of the Kula-Pacific Ridge near the end of the Eocene epoch. The Neogene, on the other hand, was definitely a time of ridge-trench interaction. Geologic events of the middle Miocene (∼15 Ma) are particularly noteworthy because they included anomalous near-trench magmatism, the cessation of backarc spreading in the Shikoku Basin, incipient collision between the Izu-Bonin and Honshu Arcs, the opening of the Sea of Japan, rapid rotation of crustal blocks in both northwest and southeast Japan, and the formation of high-rank coals within forearc-basin strata. The high ranks of organic metamorphism within the accretionary forearc, together with the anomalous basic and acidic volcanic and plutonic rocks, provide unambiguous and widespread evidence in favor of high geothermal gradients at relatively shallow depths. Furthermore, data from offshore extensions of the Shimanto Belt (Nankai accretionary prism) show that Holocene geothermal conditions are still unusually warm when compared to shallow levels of most “typical” subduction zones. Thus, much of the Cenozoic history of southwest Japan contradicts the paradigm of low-temperature, high-pressure metamorphism within subduction zones, and the exposed geology of the Muroto Peninsula provides some of the world’s best examples of the thermal and structural effects of ridge-trench interactions.

The Shimanto accretionary complex on the Muroto Peninsula of Shikoku comprises two major units of Tertiary strata the Murotohanto subbelt (Eocene-Oligocene) and the Nabae subbelt (Oligocene-Miocene). Field-based structural analyses and laboratory measurements of thermal maturity show that both subbelts were affected by thermal overprints long after the initial stages of accretion-related deformation. Paleotemperatures for the entire Tertiary section range from about 140 to 315°C, based upon mean vitrinite reflectance values of 0.9 to 5.0%R m . In general, older rocks of the Murotohanto subbelt display higher levels of thermal maturity and better cleavage than rocks of the Nabae subbelt. The Murotohanto subbelt also displays a spatial decrease in thermal maturity from south to north, and this pattern probably was caused by regional-scale differential uplift following peak heating. Conversely, the paleothermal structure exposed within the Nabae subbelt is fairly uniform, except for the local effects of mafic intrusions at the tip of Cape Muroto. The most important structural discontinuity to display a large component of post-metamorphic vertical offset is the Shiina-Narashi fault; this out-of-sequence thrust forms the boundary between hanging-wall rocks of the Murotohanto subbelt and the younger Nabae subbelt. Rapid heating of the Shimanto Belt evidently occurred immediately after a middle Miocene reorganization of the subduction boundary of southwest Japan. Whether or not all portions of the inherited (Eocene-Oligocene) paleothermal structure were overprinted during the middle Miocene remains controversial. Subduction prior to middle Miocene time probably involved the Kula, or fused Kula/Pacific, Plate. Hot oceanic lithosphere from the Shikoku Basin back-arc spreading ridge first entered the subduction zone at approximately 15 Ma; this event also coincided with the opening of the Sea of Japan and the rapid clockwise rotation of southwest Japan. Middle Miocene thermal overprints elsewhere in the Shimanto Belt are spatially related to and, therefore, probably synchronous with widespread acidic magmatism, local mafic intrusions, and formation of anthracite coals in the forearc. The Kii Peninsula of Honshu appears to be the hottest spot along the strike of the subduction zone; this locality may well correspond to the collision point between the Shimanto accretionary complex and the central ridge of the Shikoku Basin. However, the spreading ridge probably was unstable and disorganized at the time of the collision, which allowed widespread, strike-parallel flux of mantle heat and magma into the accretionary prism. Apatite fission-track data show that most of the Eocene-Oligocene Shimanto strata experienced differential uplift and cooling through the 100°C isotherm shortly after the ridge-trench collision (11 to 8 Ma). The Shimanto Belt provides an excellent example of an ancient accretionary complex that does not follow the paradigm of high-pressure, low-temperature metamorphism. Although the thermal history of southwest Japan was punctuated by an unusual burst of middle Miocene thermal and tectonic activity, the residual effects of interaction between the trench and a young subducting slab persist today, as evidenced by high heat flow within submerged portions of the Nankai accretionary prism.

Paleothermal data from the Cretaceous through Miocene Shimanto Belt reveal important variations in thermal history, both along and across strike within the ancient accretionary prism. Two types of variations are considered: those due to different tectonic settings (e.g., tectonic mélange versus coherent accreted rocks versus slope-basin deposits); and those due to different times of accretion. Within the Cretaceous subbelt on Shikoku, shale samples from mélange and coherent rocks show the same level of illite crystallinity, but samples from a probable slope-basin deposit are significantly less crystalline suggesting that the slope-basin strata experienced less heating than underlying accreted rocks. In contrast, within the Tertiary subbelt, mélange shales, coherent accreted strata, and slope-basin deposits all overlap with respect to illite crystallinity and vitrinite reflectance values; moreover, the grade tends to be higher than in the Cretaceous subbelt. This difference in grade may merely reflect different levels of exposure, but is more likely related to the fact that both the Eocene-Oligocene and Oligocene-Miocene subbelts experienced relatively high paleogeothermal gradients, apparently resulting from separate ridge subduction events in the late Eocene/early Oligocene and middle Miocene (DiTullio and others, this volume; Hibbard and others, this volume). In addition to overall differences in grade between the Cretaceous and Tertiary subbelts, marked variations in metamorphism and deformation style in the Shimanto Belt are apparent along strike between the Shikoku and the neighboring Kii Peninsula and Kyushu. In Kyushu, all the subbelts of the Shimanto Belt are characterized by low dips. In the relatively highly metamorphosed Cretaceous subbelt and in the Kitagawa Group of the Eocene-Oligocene subbelt, this may reflect deformation in an underplating regime. However, in the Eocene to lower Oligocene Hyuga Group, as well as probably originally in the Paleogene Otanashigawa Group and the Oligocene-Miocene Muro Group of Kii Peninsula, moderate- to low-dipping strata are associated with only diagenetic grades of thermal alteration. Although the Hyuga and Muro Groups may represent slope-basin deposits, clearly accreted Eocene rocks such as the Kitagawa and Otanashigawa Groups have experienced markedly less landward rotation than equivalently thermally altered accreted rocks in Shikoku. This difference in landward rotation may be related to landward-vergent structures observed in the Oligocene-Miocene subbelt. Landward-vergent structures may have resulted from a seaward-dipping backstop caused by the high paleogeothermal gradient resulting from subduction of the Shikoku basin spreading ridge beneath the Muroto Peninsula in the middle Miocene (Byrne and Hibbard, 1987). However formed, the landward-vergent structures probably contributed to the enhanced steepening of strata in the Shimanto Belt in Shikoku with respect to other areas along strike.

Tertiary rocks of the Shimanto Belt represent the youngest subaerial part of the accretionary margin of southwest Japan. Measurements of mean vitrinite reflectance (%R m ) from shales and cleaved metapelites show that the Eocene through early Miocene strata on the Muroto Peninsula of Shikoku Island were exposed to temperatures of approximately 140°C to 315°C. Analyses of inorganic phases corroborate these findings. Values of illite crystallinity index (CI) range from 0.87 Δ°2 θ to 0.21 Δ°2 θ . Most of the CI data are consistent with conditions of advanced diagenesis and anchimetamorphism (transition into greenschist facies), and a few CI values fall within the zone of epimetamorphism (lowermost greenschist facies). A significant statistical correlation exists between %R m and CI, with the best-fit curve corresponding to the following equation and correlation coefficient: %R m = 0.57 - 5.99 log (CI); r = 0.84. This curve conforms reasonably well with the boundaries of the anchizone, as established by independent compilations. Calibration of CI values with paleotemperature estimates (T, in °C), as derived from R% m data, results in the following correlation: CI = 1.197 - 0.0029(T). However, because of error propagation, uncertainties in the validity of extrapolation, and potential differences in the boundary conditions of low-grade metamorphism, this relation between CI and paleotemperature should be applied with caution to studies of other orogenic belts. Measured values of illite b o lattice spacings range from 9.001Å to 9.031Å; these data are consistent with moderate amounts of (Mg+Fe total ) in the illite unit cell. By analogy with other orogenic sequences, Shimanto metamorphism evidently was governed by intermediate pressure gradients. Maximum burial pressures were probably less than 2.5 kbar, and maximum burial depths were 9 km or less. Thus, at least when viewed within the blueschist-facies paradigm of subduction zones, the Tertiary Shimanto Belt must be regarded as somewhat unusual.

To place thermotectonic constraints on the evolution of the Cretaceous to Neogene Shimanto Belt, we carried out fission-track (FT) analyses of detrital apatite and zircon collected from both sandstone turbidites and blocks in mélanges on the Muroto Peninsula, Shikoku. Eight FT apatite ages show good agreement around 10 Ma, with all data except for one passing the χ 2 test. These results, in conjunction with depositional ages of Cretaceous to early Oligocene, demonstrate that the Shimanto Belt was heated hotter than ∼125°C (apatite total annealing temperature) and subsequently cooled below ∼100°C at ∼10 Ma. The cooling pattern is attributed to higher thermal gradients and/or uplift caused by rapid subduction of the newly formed Shikoku Basin at ∼1 5 Ma. On the other hand, 23 zircon samples show a large range of sample ages from 150 to 17 Ma, with all failing the χ 2 test except for one from a mélange. Hence, most parts of the Shimanto Belt have not been heated above the zircon partial annealing zone (ZPAZ; ∼190 to 260°C), whereas some parts of the mélanges have experienced temperatures above 260°C. FT age spectra of zircon single-grain data show that in the Northern Shimanto Belt (NSB) the youngest peak in each sample is younger than its depositional age, suggesting the maximum temperature was in the ZPAZ. In contrast, all age peaks from the Southern Shimanto Belt (SSB), are consistently older than depositional ages, providing no evidence of heating up to ZPAZ. These contrasting patterns probably reflect systematic differences in the maximum temperature reached during the evolution of the Shimanto Belt. The age-temperature paths estimated for individual tectonic units suggest successive accretion, growth, and uplift of offscraped imbricate packages as well as underplating of mélanges beneath them in an accretionary wedge.

ABSTRACT A 200-m-thick, near-vertical, middle Miocene (ca. 14 Ma), gabbroic sheeted intrusion in the Muroto area of the Shimanto accretionary complex of southwest Japan yields anisotropy of magnetic susceptibility (AMS) showing a magnetic foliation for the minimum axis (K min ) oblique (by ~70°) to the perpendicular of the intrusive contact. Assuming the K min axis represents the paleovertical axis, these data suggest that the gabbroic sheet was not intruded into the host sediments horizontally. Paleomagnetic measurements of the gabbroic intrusion show an in situ mean direction of reversed polarity (declination/inclination [Dec/Inc] = 287°/–65°, α 95 = 3°) that is considerably different from the expected, reversed-polarity dipole-field direction of this region (Dec/Inc = 0°/–56°). A structural analysis combining the paleomagnetic and AMS data led to the determination of a unique pole of rotation, around which the dike can be back-rotated to its initial orientation. The magnitude of rotation necessary for the in situ paleomagnetic direction to be back-rotated to the expected direction is ~60°, which is consistent with the rotation required for the K min axis to be vertical. This consistency can be regarded as independent support for our interpretation of the AMS results and the reliability of the paleomagnetic data. Consequently, we propose that the Muroto gabbro was intruded when the paleo–trench-fill sediments had been tilted landward by ~20°, presumably by accretion, and that the gabbro might have been intruded as a sill-like sheet along a structurally weak zone, possibly part of the frontal thrust plane in the Shimanto accretionary prism.