Geophysical Research Letters

The 1707 Mw8.7 Hoei earthquake triggered the largest historical eruption of Mt. Fuji



[1] Studies in magma-tectonics point to a spatiotemporal correlation between earthquakes and volcanic eruptions. Here, we examine the correlation between two great Japanese earthquakes, the 1703 Mw 8.2 Genroku and 1707 Mw 8.7 Hoei, and Mt. Fuji's explosive (VEI 5) Hoei eruption, 49 days after the 1707 earthquake. We model the static stress changes and dilatational strain imparted on the Mt. Fuji magmatic system due to each earthquake to determine if these mechanisms enhanced the potential for eruption. Our results show that both earthquakes clamped the dike from 8 km to the surface and compressed magma chambers at 8 km and 20 km depths. The 1707 earthquake decreased the normal stress on the dike at 20 km, the proposed depth of a basaltic magma chamber, by 1.06 bars (0.106 MPa). We hypothesize that the stress change and strain generated by the 1707 earthquake triggered the eruption of Mt. Fuji by permitting opening of the dike and ascent of basaltic magma from 20 km into andesitic and dacitic magma chambers located at 8 km depth. The injection of basaltic magma into the more evolved magmatic system induced magma mixing and a Plinian eruption ensued.

1. Eruption Triggering by Earthquakes

[2] There is a statistical correlation between large magnitude earthquakes and triggered magmatic activity [Linde and Sacks, 1998; Manga and Brodsky, 2006]. This correlation is the result of dynamic and static stresses generated by earthquakes that indirectly or directly trigger volcanic eruption. One triggering mechanism related to stress changes and strain is magma mixing [Sparks et al., 1977]. Geochemical and petrologic evidence indicates the mixing of magmas prior to triggered eruption (e.g., Santa Maria 1902 [Williams and Self, 1983]; Mt. Pinatubo 1991 [Bautista et al., 1996]). We investigate the static stress changes and crustal strain associated with the Mw 8.2 1703 Genroku and Mw 8.7 1707 Hoei earthquakes on Mt. Fuji, whose most explosive (VEI 5) historical eruption occurred 49 days after the Hoei earthquake and included mixed magmas. We also discuss the implications of static stress changes associated with the March 11, 2011 Mw9.0 Tohoku-Oki earthquake for future Mt. Fuji eruptions.

2. Tectonic Background

[3] Convergence between the Eurasian (EU), Philippine Sea (PS), Pacific (PA) and North American (NA) plates results in great earthquakes, volcanism, and complex upper plate tectonics in Japan (Figure 1). The 31 December 1703 Mw 8.2 Genroku earthquake ruptured two fault segments (Figure 1) of the Sagami Trough and was the largest historical earthquake on this fault system [Shishikura, 2003; Grunewald and Stein, 2006]. The earthquake parameters, including slip, rake, strike, length, and width (Table S1 in the auxiliary material), were estimated by Shishikura [2003]after modeling the vertical positions of paleo-shorelines and their distributions along the Boso and Miura Peninsulas. The 28 October 1707 Mw 8.7 Hoei earthquake ruptured five segments of the Nankai Trough (Figure 1). We used the earthquake source parameters from An'naka et al. [2003] and Furumura et al. [2011], who estimated the parameters through modeling the geographic distribution of tsunami deposits (Figure 1 and Table S1). This was the largest historical earthquake to shake Japan prior to the 2011 Tohoku-Oki earthquake.

Figure 1.

Topographic and bathymetric map of central Japan showing the location of Mt. Fuji (shaded triangle) and major tectonic plates. Modeled fault planes for this study are outlined by red rectangles, where A and B correspond to the 1703 Genroku earthquake [Shishikura, 2003] and N1–N5′ correspond to the 1707 Hoei earthquake [An'naka et al., 2003; Furumura et al., 2011]. White triangles are Holocene volcanoes.

[4] Mt. Fuji is a stratovolcano located at the triple junction of the NA, EU, and PS plates. The magmatic system is controlled by the regional maximum compressive stress, which is directed northwest–southeast as evidenced by the alignment of parasitic cones and exposed dikes [Nakamura, 1977]. On 16 December 1707, 49 days after the Hoei earthquake, Mt. Fuji erupted from three aligned (N323°) vents on its southeast flank [Tsuya, 1955]. The Plinian eruption produced 0.7 km3 of tephra [Miyaji and Koyama, 2007], and is noted for its emission of mixed and mingled andesitic and dacitic tephras followed by basaltic tephra [Tsuya, 1955; Yoshimoto et al., 2004]. Geophysical, geochemical, and geologic studies of Mt. Fuji indicate a layered, but dominantly basaltic magmatic system, with a deep (∼20 km) basaltic magma body [Lees and Ukawa, 1992; Aizawa et al., 2004; Nakamichi et al., 2007] and a shallow (∼8–9 km) magma chamber, in which differentiation of basaltic magma forms more felsic magmas (i.e., andesitic and dacitic magmas) [Kaneko et al., 2010]. We investigate the spatiotemporal correlation between the 1703 and 1707 great earthquakes and the 1707 Mt. Fuji eruption through modeling of the static stress changes and strain imparted on the Mt. Fuji magmatic system by the two earthquakes.

3. Modeling the Static Stress Change and Strain on Mt. Fuji

[5] A common method of studying earthquake-volcano interactions is to model the static stress change on a magmatic system produced by an earthquake [e.g.,Nostro et al., 1998; Diez et al., 2005; Walter and Amelung, 2006; Walter, 2007]. We used the Coulomb 3.2.01 program [Lin and Stein, 2004; Toda et al., 2005] to calculate the normal stress change and dilatational strain produced by the Genroku and Hoei earthquakes on the Mt. Fuji magmatic system. We used the parameters listed in Table S1 to model the 1703 and 1707 earthquakes. The Mt. Fuji magmatic system was modeled as a single, 13 km long [Acocella and Neri, 2009], N323° striking, vertical dike that connected the top of the basaltic magma chamber (20 km depth) to the andesitic and dacitic chambers (8 km depth), and from these magma chambers to the eruptive vents. This model assumes that the andesitic and dacitic magma bodies are nearly co-located. We calculated the normal stress change and dilatational strain generated by each earthquake on the dike from the surface to 30 km depth and the magma chambers, respectively, to assess the amount that the Mt. Fuji magmatic system was clamped (negative normal stress change) or unclamped (positive normal stress change), and dilated (positive dilatation) or compressed (negative dilatation) by the Genroku and Hoei earthquakes.

4. Modeling Results

[6] The Genroku earthquake caused a maximum change in normal stress of −3.14 bars (−0.314 MPa) from 8 km to the surface and −3.74 bars (−0.374 MPa) from 20 km to 8 km (Figure S1). Thus, the Genroku earthquake clamped the Mt. Fuji dike. The Hoei earthquake produced a maximum change in normal stress of −9.86 bars (−0.986 MPa) from 8 km to the surface (Figure 2). The normal stress change from 20 km to 8 km depth varied from −3.46 bars (−0.346 MPa) at the northwest end of the dike to 1.06 bars (0.106 MPa) at the southeast end of the dike (Figure 2). Therefore, the Hoei earthquake clamped the dike from 8 km to the surface, while it unclamped the southeastern section of dike (i.e., the eruption location) from 20 km to 15 km. We tested the sensitivity of our results to variation in the model parameters and found that for realistic strike, dip, rake and slip for the earthquakes and strike of the dike, our model results in the same conclusion.

Figure 2.

Normal stress change generated by the 1707 Hoei earthquake on the Mt. Fuji magmatic system. Contour interval is 0.5 bars (0.05 MPa). Depth slice of normal stress change on the dike at (a) 8 km and b) 20 km depths. Insets show the normal stress change detail at Mt. Fuji and location of cross-section shown in Figure 2c. (c) Cross-section of the normal stress change along the dike.

[7] We found that both earthquakes produced negative dilatational strain, compressing the magma chambers. The Genroku earthquake compressed the felsic magma chamber by a maximum volumetric strain of −3.34 × 10−6 (equivalent to 0.24 bars of pressure change), and the basaltic magma chamber by a maximum of −3.33 × 10−6 (0.24 bars) (Figures S2 and S3). The Hoei earthquake compressed these chambers by a maximum of −4.81 × 10−6 (0.4 bars) and −2.66 × 10−6 (0.2 bars), respectively (Figures 3 and S4).

Figure 3.

Dilatational strain produced by the 1707 Hoei earthquake. Contour interval is 3 × 10−6. Depth slice of strain at (a) 8 km and (b) 20 km depths. Insets show the dilatational strain detail at Mt. Fuji and location of cross-section shown in Figure 3c. (c) Cross-section of dilatational strain along the dike.

5. Discussion

[8] We modeled the static stress changes and dilatational strain arising from the 1703 Mw 8.2 Genroku and 1707 Mw 8.7 Hoei earthquakes on Mt. Fuji, which experienced its largest historical eruption 49 days after the Hoei earthquake. Our findings suggest that, while the 1703 Genroku earthquake did not directly trigger the Mt. Fuji eruption, the 1707 Hoei earthquake could have triggered the 1707 Hoei eruption by static stress transfer (i.e., a decrease in normal stress) and volumetric compression of the magma chambers. Figure 4 shows our model scenario. Nostro et al. [1998] found that increased compression by Apennines earthquakes on a magma body beneath Vesuvius could promote eruption. Bautista et al. [1996] also found that compression of the Mt. Pinatubo magma chamber following the Mw 7.2 Luzon earthquake might have pushed basaltic magma into the overlying dacitic reservoir or squeezed the dacitic magma towards the surface. However, if compression were the only requisite for triggering Mt. Fuji, then one might expect that Mt. Fuji would have erupted soon after the Genroku event.

Figure 4.

Model of the 1707 Mt. Fuji eruption triggered by the 1707 Hoei earthquake: (a) Mt. Fuji magmatic system prior to the earthquake. The dacitic (off white), andesitic (light gray), and basaltic (dark gray) magma chambers are shown at their approximate depths [Aizawa et al., 2004; Kaneko et al., 2010]. The dike connects the magma chambers to the surface. (b) Static stress from the 1707 earthquake compresses the magma chambers and changes the normal stress on the dike, clamping the dike's upper portion while unclamping part of it above the basaltic magma chamber. Basaltic magma migrates to the felsic chambers. (c) The basaltic magma triggers magma mixing. Magmatic overpressure, vesiculation and migration cause seismicity and the first phase of the eruption follows. (d) A plinian basaltic eruption occurs after the eruption of the dacitic and andesitic magmas.

[9] The 1703 Genroku earthquake clamped the dike system from 20 km to the surface (Figure S2). This suggests that the dilatational strain (Figure S3) and subsequent increase in magmatic overpressure were not great enough to exceed the strength of the conduit in 1703. Following the 1707 earthquake, clamping on the upper part of the dike temporarily trapped the andesitic and dacitic magmas in their respective chambers (Figure 2). However, the Hoei earthquake unclamped (normal stress change of 1.06 bars) the southeast section of dike leading from the basaltic magma chamber (Figure 4b). Diez et al. [2005] concluded that a reduction in normal stress of 0.10−1.0 bars caused by four local Mw 5.1 earthquakes could have triggered the 1999 eruption of Cerro Negro volcano, Nicaragua. De la Cruz-Reyna et al. [2010] similarly noted that eruptions of Popocatépetl in 1999 and Tungurahua in 2003 could be explained by decreases in normal stress of 0.01 bars and 0.003 bars, respectively, following earthquakes.

[10] In the case of Mt. Fuji, we propose that unclamping of the deeper dike and compression of the basaltic magma chamber opened the conduit and pushed the basaltic magma out of the chamber. We assume that magma compressibility is negligible for the basaltic magma chamber at 20 km depth [Rivalta and Segall, 2008]. The basaltic magma then migrated up the dike and into the more felsic magma chambers (Figure 4b). Following injection of the basaltic magma into the andesitic and dacitic magma chambers, magma mixing and rapid vesiculation occurred, resulting in magma migration from 8–9 km to the surface [Yoshimoto et al., 2004] (Figure 4c). The remnants of this mixing are preserved as a chemically zoned tephra deposit, including andesitic and dacitic banded (mingled) pumice [Yoshimoto et al., 2004]. Historical reports indicate that on 3 December 1707, seismic swarms began on the south flank of the volcano [Tsuya, 1955]. This suggests that magma migration from the ∼8–9 km deep magma chamber was initiated on or before 3 December 1707, 36 days after the Hoei earthquake (Figure 4c), and continued until eruption on 16 December 1707 (Figures 4c and 4d). Therefore, by decreasing the normal stress (1.06 bars) on the dike leading from the basaltic magma chamber beneath Mt. Fuji and by compression of the same magma chamber, the 1707 Hoei earthquake likely set in motion the processes that caused the most explosive, historical eruption of Mt. Fuji.

[11] Our study offers an example of how an earthquake might trigger a volcanic eruption through static stress transfer and strain. Although we only modeled the static stress changes and dilatational strain associated with these earthquakes, other processes may have played a role in triggering the eruption. Dynamic stresses associated with both earthquakes could have promoted the eruption through pressurization of the magmas [see Manga and Brodsky, 2006]. On the other hand, our study also investigated a powerful earthquake that did not trigger an eruption (i.e., the 1703 Genroku earthquake). This result underscores the importance of geometry between the source (i.e., earthquake fault planes) and receiver (i.e., Mt. Fuji magmatic system) systems [e.g., Nostro et al., 1998].

[12] Eruptions and their associated environmental effects can wreak havoc on modern societies. The Hoei eruption's tephra deposit covered Tokyo (<100 km away) with deposits ranging in thickness from 0.5 to 16 cm, demonstrating the eruption's long-range environmental effects [Miyaji and Koyama, 2007]. Given the recent 11 March 2011 Mw9.0 Tohoku-Oki earthquake, it is especially important to understand the relationship between great earthquakes along the NA-PA and EU-PS convergent margins, including the Tokai region, and the stressing of regional faults [e.g.,Toda et al., 2011] and magmatic systems (e.g., Mt. Fuji). We modeled the changes in stress and dilatational strain for the Tohoku-Oki earthquake on the Mt. Fuji magmatic system using the finite fault model ofHayes et al. [2011]. We note that the Tohoku-Oki event produced a positive normal stress change (unclamping) on the dike from 20 km depth to the surface with a maximum value of 1.05 bars (0.105 MPa) (Figure S5). This earthquake did not compress the magmatic system. Instead, it dilated the magmatic system by a maximum of 4.84 × 10−7 (−0.03 bars) (Figures S6 and S7). Ongoing geophysical monitoring at Mt. Fuji, will allow for testing of our model and assessing the role of static and dynamic stresses in triggering seismicity and magmatic activity.


[13] This research was made possible by the UNAVCO RESESS program and through NSF award EAR 0955560 to PL.

[14] The Editor thanks Michael Manga and an anonymous reviewer for their assistance in evaluating this paper.