Abstract
Fission-track (FT) thermochronology was applied to the Atotsugawa Fault in the Hida Metamorphic Belt, central Japan, to detect any ancient thermal anomaly associated with fault displacements. Apatite and zircon grains from gouges (c. 2 cm wide) and fractured rocks (c. 10 cm) at six fracture zones within a 15 m-wide fault zone were dated. Most of the zircon (120–150 Ma) and apatite (44–60 Ma) ages agree well with emplacement ages for the granites that intrude the Hida Belt. The discordance in zircon and apatite FT ages is interpreted to reflect cooling due to regional uplift and associated erosion. A thermal anomaly was identified at one of the gouge zones that showed an exceptionally young apatite age (c. 32 Ma) with an unimodal FT length distribution. It presumably indicates secondary heating induced by frictional slip during an earthquake, possibly giving a younger limit of the initiation of the activity in the Hida Belt.
In order to investigate a long-term landscape evolution of a certain region, it is important to reveal the activity history of the faults that exist in the region. Abundant frictional heat during a fault displacement is implied by the occurrence of pseudotachylyte, a melt rock in the fault zone (e.g. Sibson 1975). The thermal footprint within rocks in fault zones can possibly be detected by low-temperature thermochronometry (e.g. Scholz et al. 1979). Fission-track (FT) thermochronology has recently been applied to detect the thermal anomaly in fault zones, such as the Nojima Fault in central Japan (Murakami et al. 2002; Yamada et al. 2007a), and the San Gabriel fault zone in southern California (d'Alessio et al. 2003). The FT method is suitable for detecting thermal anomaly associated with fault activity because: (1) the minerals commonly used for analysis (apatite, zircon) are relatively resistant to weathering so that they are likely to survive under the hydrothermal conditions in some fault zones; (2) closure temperatures for FT methods are relatively low (c. 110 and 240 °C for apatite and zircon, respectively; e.g. Gallagher et al. 1998), so they can serve as sensitive indicators for thermal events in the upper crust; and (3) well-documented annealing kinetics (e.g. Laslett & Galbraith 1996; Yamada et al. 2007b) allow quantitative modelling of a rock's thermal history (e.g. Ketcham et al. 1999). From the view point of seismology, the amount of heat produced along a fault is essential information when investigating fault strength, which controls the earthquake generation process. The significance of this kind of approach is to assess directly the parameters related to fault strength with natural samples that are marked by ancient earthquakes, independent from laboratory friction experiments.
In this study, we applied FT thermochronology to reveal an activity history of the Atotsugawa Fault, central Japan, by detecting palaeo-thermal anomalies induced by fault activities. This fault, 60 km in length, lies within the Hida Belt, which cuts through Palaeozoic metamorphics and Mesozoic granites. The most marked characteristic of this fault is the coexistence of creeping and locking sections along the same fault system (Geographical Survey Institute, Japan 1997). It is important to know when and how the activity started in order to understand the mechanism of earthquake occurrence, as well as to reveal the regional tectonic history. We sampled gouges and fractured rocks at six fracture zones within a 15 m-wide fault zone in the eastern segment of the fault zone, where possible creeping has been reported (Geographical Survey Institute, Japan 1997). The merit of using samples from small fracture zones is that the dimensions and alignment of samples can directly be measured for characterizing the thermal anomaly in detail. To identify an ancient thermal anomaly related to the fault, we compared FT ages of the fracture zone with those of the granitic rocks that intrude the metamorphic country rock (Matsuda et al. 1997). The present analysis seems to have identified an ancient thermal anomaly in a sample of the fault gouge. Also, the regional uplift and erosion history in the Hida Belt will be discussed based on the discordance in apatite and zircon FT ages found in this analysis.
Outline of the Atotsugawa Fault
The Atotsugawa Fault is a strike-slip fault with a right-lateral displacement along 60 km surface trace (Matsuda 1966). The strike of the fault trace is approximately N60°E, and the fault plane is almost vertical (90°±10°) near the surface. There have been large historical earthquakes along the fault; one of the largest events was the 1858 Ansei Hietsu earthquake with M 7.0 (Mikumo et al. 1988). A spatial heterogeneity of the seismicity along the Atotsugawa Fault is suggested by the distribution of epicentres beneath the trace of the fault (e.g. Mikumo et al. 1988). The seismicity in the central segment of the fault is rather low compared with the eastern and western segments. This segment seems to correspond to a possible creep zone implied by the geodetic observation of the baseline length across the fault trace. The Geographical Survey Institute, Japan (1997) found evidence of a creep-like movement, with a rate of approximately 1.5 mm year−1 in the central segment, and no observable baseline change in the western segment.
A geological map of the Hida Belt around the Atotsugawa Fault is shown in Figure 1a. The basement consists of the Hida metamorphic rocks of Triassic or earlier age, and two types of granites, namely Funatsu and Shimonomoto types, ranging in age from Triassic to Jurassic. Marine and non-marine sedimentary rocks of the Tetori group, ranging in age from Jurassic to Cretaceous, unconformably overlie the metamorphic and granitic units. Several geochronological studies have been undertaken on the metamorphic and granitic units in the Hida Belt. The reported ages cluster around three periods of approximately 170–190, 210–230 and 330 Ma (Ota & Itaya 1989). Sohma & Kunugiza (1993) argued that these periods reflect the main metamorphism of granulite facies at 330 Ma, regional metamorphism of amphibolite facies at 250–200 Ma and contact metamorphism resulting from the intrusion of Funatsu-type granites at 170–200 Ma, respectively. In order to reveal the cooling history and the effect of regional uplift, Matsuda et al. (1997) measured the ages of the Shimonomoto-type granites with various dating methods including FT data (Fig. 2b). They concluded that: (1) Shimonomoto-type granites intruded at 200–180 Ma and cooled to about 300 °C at 150 Ma; and (2) since 150 Ma the granites have slowly cooled to the surface due to uplift and erosion at a uniform rate. The initiation of fault activity is presumed to be during the late Tertiary based on geological and geomorphological studies (Matsuda 1966).
(a) A simplified geological map around the Atotsugawa Fault (modified after Yamada et al. 1989; Kano et al. 1999). The location of ATGR2 sample is also shown. (b) Sketches of the fault outcrop and sampling locations. Distances were measured to the tentative reference point, shown as ‘Base’ in the pictures.
(a) Distribution of apatite and zircon FT ages in the fault outcrop samples. Open and solid circles (as shown in Fig. 1b) indicate fault gouge and fractured rocks, respectively. Length distribution of apatite FTs for sample ATG1G is also shown. Shaded bands behind the plot indicate the apatite and zircon FT age distributions of the Shimonomoto-type granites (Matsuda et al. 1997). (b) Cooling histories of the Shimonomoto-type granites revealed by various radiometric dating methods, modified after Matsuda et al. (1997). Different symbols indicate samples from different rock bodies. Error bars are 2σ.
Samples and experiments
Six fracture zones are within an outcrop 15 m long near the portal of the Kamioka Mine prospect tunnel, located on the right bank 1.5 km upstream from the confluence of the Atotsugawa and the Takahara rivers (c. 350 m altitude: Fig. 1b). Deformation and alteration of the Hida metamorphic rocks in the fracture zones were more advanced than those in the surrounding area. Each fracture zone consists of a gouge zone 1–3 cm wide, with fractured rocks in a surrounding zone 10–15 cm wide on both sides of the fault gouge. FT samples were collected from both the gouge and fractured rocks (200–400 g each) about 10 cm apart in each of the six fracture zones. These were called samples ATG1G-6G (gouges) and ATG1F-6F (fractured rocks) (Table 1). The conditions of these samples were not good in general because many grains had cracks on grain surfaces probably due to deformation in the fault zone. To investigate the effect of the fracture zones, two reference samples were also sampled where no fractures were observed, one from the vicinity of the fault outcrop (ATGR1) and another from the Miyagawa area, that is presumed to be a locked segment of the fault (ATGR2) located about 15 km west of the outcrop. Consequently, 14 samples were analysed in total. The locations of the samples are identified by the distance to the tentative reference point, shown in Figure 1b.
Results of FT analysis
FT analysis was carried out using the external detector method (ED1: Gleadow 1981). Ages were calculated following the zeta approach (Hurford 1990). A detailed description of the experimental procedure followed and the system calibration are documented in Danhara et al. (2003). Thirty grains with good shapes and homogeneous spontaneous track distributions were randomly chosen from each sample, except for the apatites from ATG1C and 3C. Sufficient numbers of grains were not obtained from these two samples because of the poor condition of the grains.
Results and interpretation
Analytical results are listed in Table 1. For some samples that did not pass the χ2 test (Galbraith 1981) at the 5% significance level, the mean ρs/ρi ratios were quoted (Wagner & Van den Haute 1992). Lower P(χ2) values for zircon samples with the ED1 method can be explained by non-Poissonian variation that is attributed to the difference in uranium contents above and below the observed internal surfaces due to the zoned distribution of uranium (Danhara et al. 1991). Lower P(χ2) values for apatites indicate a broad range of single-grain ages within each sample, probably due to partial resetting or uneven thermal effects by secondary heating events.
FT ages are plotted against sample distance to the tentative reference point (Fig. 2a). Open and solid circles (as shown in Fig. 1b) indicate fault gouge and fractured rocks, respectively. For comparison, the age distribution for the apatite and zircon samples of the Shimonomoto-type granitic rocks (Fig. 2b; modified after Matsuda et al. 1997) are also shown as shaded bands behind the plot. Zircon FT ages are concordant among samples within 2σ errors, and their distribution also agrees with that of the granitic rocks. The effect of fracture zones is negligible with respect to the zircon ages, as indicated by the insignificant difference in mean age of samples from fracture zones (131±6 Ma, 1σ) and that of reference samples (142±6 Ma), where the mean ages are weighted by the square of the error of the ages. Apatite FT ages are concordant among samples within 2σ error, including those with lower P(χ2) values, and their distribution also agrees with that of the granitic rocks; except for ATG1G, a gouge sample from a fracture zone. Compared with zircon FT ages, apatite FT ages are exceedingly young for each rock samples analysed. With the exclusion of ATG1G apatite, the difference in mean age of samples from fracture zones (47.2±2.6 Ma, 1σ) and that of reference samples (48.7±3.1 Ma) is negligible.
The age of ATG1G apatite was 32.1±3.2 Ma, a significantly younger age at the 2σ level, compared with other apatites samples within a fault outcrop 15 m long. In order to characterize the thermal history of this sample, the distribution of track lengths was analysed, as shown in Figure 2a. A typical narrow-length distribution of confined tracks (Gallagher et al. 1998) indicates that this sample cooled rapidly after it underwent the secondary heating up to the temperature range over the partial annealing zone. Because many of the grains in this sample have cracks on their surfaces, there were not enough confined tracks to enable numerical modelling of the thermal evolution of this sample.
Discussion
Regional uplift and erosion of the Hida Belt
The discordance in apatite and zircon FT ages, which have different closure temperatures, is considered to reflect thermotectonic evolution due to slow cooling of the host rocks. Matsuda et al. (1997) revealed the slow cooling process of the Shimonomoto-type granitic rocks by means of thermochronological analyses including the FT method, and suggested that the slow cooling process may be attributed to regional uplift and erosion of the Hida Belt. Both apatite and zircon FT ages of the Hida metamorphic rocks obtained in this study are mostly concordant with those of the Shimonomoto-type granitic rocks (Matsuda et al. 1997), except the apatite age from one of the gouge samples. This indicates that both the Hida metamorphic rock and the Shimonomoto-type granites in the Hida Belt were uplifted concurrently, at least to a level corresponding to temperatures cooler than the closure temperature of the zircon FT system (c. 240 °C). Because samples were collected only from a small faulted outcrop, it is not adequate to use them for detailed landscape evolution. The ranges of zircon and apatite FT ages were, however, concordant with those of granites that intrude this region, and the sampling area was confirmed not to have heterogeneous exhumation history.
Skarn deposits in the Kamioka Mine, distributed widely in the middle of the Hida Belt, were formed by chemical metasomatism of limestone in the Hida metamorphic rocks. The time of the formation of these deposits could be marked by thermal impact related to hydrothermal metamorphism. The formation of the deposits has been dated as Late Cretaceous–Early Paleogene (hornblende K–Ar ages of 63.8–67.5 Ma: Sato and Uchiumi 1990). These reported ages are, however, considerably younger than zircon FT ages of the metamorphic rocks obtained in this study, although the zircon FT system is more thermally sensitive than the hornblende K–Ar system. This might be because the deposits were formed by very local hydrothermal activities in each metallogenetic provinces and so did not overprint the zircon FT ages in the studied area, reflecting the regional uplift history.
Thermal anomaly in a fracture zone
Secondary heating is presumed for ATG1G because the discordance in apatite ages was found at a very narrow range in a fracture zone. This indicates the existence of a thermal anomaly in the gouge in the fracture zone, and a relationship between the secondary heating and local geological phenomena at a gouge. Local change in apatite FT age around a fault plane was also found for the San Gabriel Fault (d'Alessio et al. 2003) and the Nojima Fault (Murakami & Tagami 2004; Yamada et al. 2007a).
Possible sources for the secondary heating are: (1) frictional heat during a fault displacement; and (2) heat transfer or dispersion via geothermal fluids at the interior of the crust (e.g. Murakami et al. 2002). The circulation of hot fluids within the fracture zone is a possible heat source associated with faults, because a fracture zone can be a channel of high-temperature fluid from the deep crust when an earthquake occurs. This is, however, implausible in this case because in the investigated area only a small, approximately 2 cm-wide, gouge zone is present, and if the fluid conveys the heat in such a narrow channel it must be higher than 500 °C according to the numerical assessment of heat diffusion (Mizoguchi et al. 2009), which may also have affected the apatite ages of fractured rocks beside ATG1A. Therefore, frictional heat generated during a fault displacement that formed the gouge in the fracture zone may have caused the thermal anomaly. Although this fault gouge zone could have been activated several times, it may safely be assumed that the secondary heating was caused by a single displacement at the first fracturing of the gouge zone because of following reasons: (1) it is expected that the fault strength may be maximum and the amount of the frictional heat is the largest at the first fracturing of a particular gouge zone; (2) the effect of accumulated heat is not significant because frictional heat generated by each activation should diffuse into the vicinity of rocks with normal thermal diffusivity during the recurrence interval of this fault system (c. 3000 years: Takeuchi et al. 2003); and (3) FT annealing is more sensitive to a change in temperature than to duration, as indicated by the decrease in temperatures in the partial annealing zone of approximately 40 °C (apatite) and 20 °C (zircon) for a magnitude of longer annealing duration, estimated by using the published kinetics (e.g. Laslett & Galbraith 1996; Yamada et al. 2007b).
If the apatite FT age of sample ATG1G was affected by the frictional heat during a fault displacement at the gouge zone, it can be presumed that this age may approximate a younger limit of the initiation of the activity of the Atotsugawa Fault, to which this gouge zone belongs. This is concordant with Matsuda (1966), who suggested that the activity of the Atotsugawa Fault started in the late Tertiary, based on the displacement of landforms and the slip rate of the recent activity of the Atotsugawa Fault.
The numerical modelling of the heat generation by frictional slip and the subsequent heat conduction with respect to the geometry and alignment of the samples can provide the constraints on estimates of fault strength. If a single displacement on the fracture zone is assumed to cause the secondary heating of the ATG1G sample, the amount of the heat can be constrained by the fact that the zircon FT age was not reset to the apatite age.
Conclusions
Both apatite and zircon FT ages of the Hida metamorphic rocks obtained from a fault outcrop are mostly concordant with those of the Shimonomoto-type granitic rocks that intrude the Hida Belt (Matsuda et al. 1997). The discordance in these ages suggests that both the metamorphics and the granites in the Hida Belt were uplifted concurrently, at least to a level corresponding to cooler temperatures than the closure temperatures of the zircon FT system (c. 240 °C).
An apatite FT age from one of the gouge zones was significantly younger than other samples at the 2σ level with an unimodal FT length distribution, indicating the secondary heating up to the temperature range over the partial annealing zone. This thermal anomaly in a small, approximately 2 cm-wide gouge zone may have been caused by frictional heat during a fault displacement at the first fracturing that formed the gouge in the fracture zone. Therefore, the apatite FT age of sample ATG1G (32.1± 3.2 Ma) may approximate a younger limit of the activity of the Atotsugawa Fault. This is consistent with a previous estimate of the initiation during the late Tertiary, based on geological and geomorphological studies (Matsuda 1966).
Acknowledgments
We thank Drs E. Fukuyama, A. Kubo and F. Yamashita for their help with sample collection, and Dr K. Mizoguchi for his discussion on thermal history analyses. We appreciate detailed reviews by Prof. P. Kamp and Dr F. Lisker.
- © Geological Society of London 2009
References
Fission-track analysis of the Atotsugawa Fault (Hida Metamorphic Belt, central Japan): fault-related thermal anomaly and activation history
