12:00 PM - 12:15 PM
[SCG44-23] Illuminating slow earthquakes: an approach from thermal modeling
★Invited Papers
Keywords:southwestern Japan, thermal modeling, dehydration gradient, long-term slow slip event, deep low-frequency earthquake, tectonic tremor
1. Introduction
Twenty years have passed since the discovery of the slow earthquakes. Since then, a variety of researches have been done and the picture of the slow earthquakes is being clarified, but the actual situation is still unknown. We would like to shed light on this issue in terms of thermal modeling associated with subduction of an oceanic plate. We will focus on slow earthquakes in the Nankai subduction zone in southwestern Japan and show some results obtained so far.
2. Features of our model
In order to construct a numerical model that takes account of arbitrary slab geometry and subduction history, we incorporated a prescribed guide into our numerical model (Torii and Yoshioka, 2007; Ji and Yoshioka, 2015). The guide is like a mold that imitates the geometry of a subducted oceanic plate, and is set along the upper and lower surfaces of the slab at present. By pouring a substance that gradually sinks into this guide, we realized subduction of the oceanic plate with an arbitrary geometry. Further, although the geometry of the guide itself does not change with time, its position and the direction and velocity of the subducting substance can be changed with time. This made it possible to handle the subduction history such as the trench retreat and advance and plate motion along the trench axis.
3. Slow earthquakes in terms of distributions of temperature and dehydration gradient
Once the temperature is obtained using the above model, we can obtain the maximum water content distribution, using the phase diagram of the hydrous minerals in each layer in the slab. Furthermore, the dehydration gradient along the subduction direction can be obtained, by which it is possible to understand where and what kind of dehydration takes place in the slab. We will focus on the occurrence of long-term slow slip events (L-SSEs), deep low-frequency earthquakes (LFEs), and tectonic tremors (TTs), and describe the relationship between them and the temperature and dehydration gradient.
The Tokai L-SSE and the Bungo Channel L-SSE are well-known L-SSEs that have occurred in southwestern Japan. The temperatures at the up-dip limit of the L-SSEs were estimated to be approximately 350 oC (Suenaga et al., 2016; Nakata et al., 2017).
The temperatures at the up-dip limit of the LFEs and TTs that occur in the Tokai and the eastern part of the Kii Peninsula were estimated to be approximately 450 oC and 480 oC, respectively (Suenaga et al., 2016; 2019). The LFEs and TTs in southwestern Japan occur in a form of belt-like shape, but there is their gap area in Ise Bay, and they rarely occur in Kyushu. Additionally, LFEs and TTs have mobility in the slab strike direction, indicating the involvement of water. In the vicinity of Ise Bay, the dip angle of the Philippine Sea (PHS) slab is rapidly lower than that in the surrounding region. As a result of our numerical simulation, the temperature gradient was found to be smaller in the gap area than in the surrounding LFE and TT active area, and the dehydration gradient was also smaller by about a half (Suenaga et al., 2019).
The PHS plate subducting beneath Shikoku to Chugoku and beneath Kyushu is bounded by the Kyushu-Palau Ridge: On the east side, the young PHS plate is subducting at a low dip angle and the volcanic activity of the backarc is low, while on the west side, the old PHS plate is subducting at a high dip angle and volcanic activity is high. As a result of our numerical simulation, in the former, dehydration from the PHS slab turns out to gradually take place, contributing to the occurrence of LFEs and TTs, dehydration is almost completed in the forearc, and almost no water is left in the back-arc volcanic field. On the other hand, in the latter, dehydration does not occur in the forearc, which prevents occurrence of LFEs and TTs, and dehydration occurs at once in the deep slab, forming an active volcanic array directly above it (Tatsumi et al., 2020).
We will also mention similarity between the Nankai and the Alaska subduction zones.
Twenty years have passed since the discovery of the slow earthquakes. Since then, a variety of researches have been done and the picture of the slow earthquakes is being clarified, but the actual situation is still unknown. We would like to shed light on this issue in terms of thermal modeling associated with subduction of an oceanic plate. We will focus on slow earthquakes in the Nankai subduction zone in southwestern Japan and show some results obtained so far.
2. Features of our model
In order to construct a numerical model that takes account of arbitrary slab geometry and subduction history, we incorporated a prescribed guide into our numerical model (Torii and Yoshioka, 2007; Ji and Yoshioka, 2015). The guide is like a mold that imitates the geometry of a subducted oceanic plate, and is set along the upper and lower surfaces of the slab at present. By pouring a substance that gradually sinks into this guide, we realized subduction of the oceanic plate with an arbitrary geometry. Further, although the geometry of the guide itself does not change with time, its position and the direction and velocity of the subducting substance can be changed with time. This made it possible to handle the subduction history such as the trench retreat and advance and plate motion along the trench axis.
3. Slow earthquakes in terms of distributions of temperature and dehydration gradient
Once the temperature is obtained using the above model, we can obtain the maximum water content distribution, using the phase diagram of the hydrous minerals in each layer in the slab. Furthermore, the dehydration gradient along the subduction direction can be obtained, by which it is possible to understand where and what kind of dehydration takes place in the slab. We will focus on the occurrence of long-term slow slip events (L-SSEs), deep low-frequency earthquakes (LFEs), and tectonic tremors (TTs), and describe the relationship between them and the temperature and dehydration gradient.
The Tokai L-SSE and the Bungo Channel L-SSE are well-known L-SSEs that have occurred in southwestern Japan. The temperatures at the up-dip limit of the L-SSEs were estimated to be approximately 350 oC (Suenaga et al., 2016; Nakata et al., 2017).
The temperatures at the up-dip limit of the LFEs and TTs that occur in the Tokai and the eastern part of the Kii Peninsula were estimated to be approximately 450 oC and 480 oC, respectively (Suenaga et al., 2016; 2019). The LFEs and TTs in southwestern Japan occur in a form of belt-like shape, but there is their gap area in Ise Bay, and they rarely occur in Kyushu. Additionally, LFEs and TTs have mobility in the slab strike direction, indicating the involvement of water. In the vicinity of Ise Bay, the dip angle of the Philippine Sea (PHS) slab is rapidly lower than that in the surrounding region. As a result of our numerical simulation, the temperature gradient was found to be smaller in the gap area than in the surrounding LFE and TT active area, and the dehydration gradient was also smaller by about a half (Suenaga et al., 2019).
The PHS plate subducting beneath Shikoku to Chugoku and beneath Kyushu is bounded by the Kyushu-Palau Ridge: On the east side, the young PHS plate is subducting at a low dip angle and the volcanic activity of the backarc is low, while on the west side, the old PHS plate is subducting at a high dip angle and volcanic activity is high. As a result of our numerical simulation, in the former, dehydration from the PHS slab turns out to gradually take place, contributing to the occurrence of LFEs and TTs, dehydration is almost completed in the forearc, and almost no water is left in the back-arc volcanic field. On the other hand, in the latter, dehydration does not occur in the forearc, which prevents occurrence of LFEs and TTs, and dehydration occurs at once in the deep slab, forming an active volcanic array directly above it (Tatsumi et al., 2020).
We will also mention similarity between the Nankai and the Alaska subduction zones.