11:00 AM - 11:15 AM
[SCG39-26] Land and marine long-term magnetotelluric observations around the western Nankai slow slip area
Keywords:Slow earthquake, Bungo Channel, magnetotelluric, OBEM, Network-MT, 3-D resistivity
Imaging of resistivity distribution gives us fundamental information about fluid and lithological feature around focal area of slow earthquakes because it depends on the pore-fluid distribution, salinity, connectivity of fluid-filled rock pores, and clay minerals the crust and upper-most mantle. Resistivity distribution of the crust-mantle scale can be estimated based on long period magnetotelluric (MT) observations. In this study, we introduce results of long period MT observations around the slow earthquake zones in the western Nankai trough mainly conducted by the JSPS KAKENHI project "Science of Slow Earthquakes". The observations were carried out by the two research teams for marine and land MT surveys.
Marine MT surveys were conducted between 2017 and 2020 at fifteen sites around the off-Miyazaki and off-Kochi area, Japan, based on Ocean Bottom Electro-Magnetometers (OBEMs). The observation periods of each OBEM were between 5 and 12 months. The training ship “Fukae-maru” and research vessel “Kairei” owned by Kobe University and JAMSTEC, respectively, cooperated for deployment and recovery of OBEMs. The OBEMs also acquired absolute pressure in three sites to detect SSEs. Based on the observed electromagnetic data, we estimated magnetotelluric impedances using BIRRP code (Chave and Thomson, 2004). The estimated MT impedances showed strong coastal effect due to complicated bathymetry in the western Nanakai Trough area. Based on the impedances at the twelve OBEM site in middle and eastern side of study area, we conducted forward resistivity modeling to evaluate the coastal effect using the 3-D modelling code (Tada et al., 2012). A simple optimum model consisting of seawater, seabed conductive layer and underneath background area can explain the coastal effect. In addition, the forward modeling implied heterogenous resistivity structure in the background area. Thus we applied 3-D inversion using the forward model as initial model. The inverted model mostly explained observed MT impedances. The model showed resistive anomaly around the ruptured area of the 1968 Hyuga-nada earthquake (M7.5) (Yagi et al., 1998). The resistivity model also shows conductive area in the western area part of the study area. It may relate to structural heterogeneity dure to subducting Kyushu-Palao ridge.
The network-MT surveys have been carried out in the western part of Shikoku Island and eastern part of Kyushu Island since April, 2016 to discuss resistivity distribution and its time variation in the long-term slow slip events (L-SSE) area beneath the Bungo channel. We used metallic telephone line network of the Nippon Telegraph and Telephone (NTT) Corp. to measure the electrical potential difference with long baselines over several kilometers. The electrical potential differences measured in this way are known to be less affected by small scale near-surface lateral resistivity heterogeneities (e.g. Uyeshima, 2007). We also measure the geomagnetic field at two stations in the target region. We estimated the frequency-domain response functions in the western Shikoku area of good quality with the aid of the BIRRP code. After applying the 3-D DASOCC inversion code (Siripunvaraporn et al., 2004), which directly inverted the Network-MT response and obtain the 3-D resistivity structure in the western Shikoku area. Remarkable low resistivity anomalies were detected in the middle crust of the hanging wall above the subducting Philippine Sea slab. However, no remarkable conductive anomaly was shown around the area of L-SSE in the Bungo channel region. Thus we also examined how the Network-MT responses are sensitive to temporal resistivity change along the slab because the L-SSE repeatedly occurred about every 6 or 7 years. In a test model, 10 km thick conductive layer is assigned just beneath the top surface of the slab. When the resistivity is set to be 20 Ohm-m (thus 500 S anomaly), RMS misfit was significantly increased from the best fit model. It indicates that temporal change could be detected due to current Network-MT survey.
We successfully obtained long-period MT impedances and preliminary resistivity distributions around western Nankai slow earthquake area based on the land and marine MT observations. Thus we plan to integrate the marine and land MT data in this study and past network-MT data in Kyushu Island (e.g. Hata et al., 2015) and estimate reliable 3-D resistivity structure covering the slow EQ areas in the western Nankai subduction zone.
Figure: Locations of OBEM sites and baselines of network-MT
Marine MT surveys were conducted between 2017 and 2020 at fifteen sites around the off-Miyazaki and off-Kochi area, Japan, based on Ocean Bottom Electro-Magnetometers (OBEMs). The observation periods of each OBEM were between 5 and 12 months. The training ship “Fukae-maru” and research vessel “Kairei” owned by Kobe University and JAMSTEC, respectively, cooperated for deployment and recovery of OBEMs. The OBEMs also acquired absolute pressure in three sites to detect SSEs. Based on the observed electromagnetic data, we estimated magnetotelluric impedances using BIRRP code (Chave and Thomson, 2004). The estimated MT impedances showed strong coastal effect due to complicated bathymetry in the western Nanakai Trough area. Based on the impedances at the twelve OBEM site in middle and eastern side of study area, we conducted forward resistivity modeling to evaluate the coastal effect using the 3-D modelling code (Tada et al., 2012). A simple optimum model consisting of seawater, seabed conductive layer and underneath background area can explain the coastal effect. In addition, the forward modeling implied heterogenous resistivity structure in the background area. Thus we applied 3-D inversion using the forward model as initial model. The inverted model mostly explained observed MT impedances. The model showed resistive anomaly around the ruptured area of the 1968 Hyuga-nada earthquake (M7.5) (Yagi et al., 1998). The resistivity model also shows conductive area in the western area part of the study area. It may relate to structural heterogeneity dure to subducting Kyushu-Palao ridge.
The network-MT surveys have been carried out in the western part of Shikoku Island and eastern part of Kyushu Island since April, 2016 to discuss resistivity distribution and its time variation in the long-term slow slip events (L-SSE) area beneath the Bungo channel. We used metallic telephone line network of the Nippon Telegraph and Telephone (NTT) Corp. to measure the electrical potential difference with long baselines over several kilometers. The electrical potential differences measured in this way are known to be less affected by small scale near-surface lateral resistivity heterogeneities (e.g. Uyeshima, 2007). We also measure the geomagnetic field at two stations in the target region. We estimated the frequency-domain response functions in the western Shikoku area of good quality with the aid of the BIRRP code. After applying the 3-D DASOCC inversion code (Siripunvaraporn et al., 2004), which directly inverted the Network-MT response and obtain the 3-D resistivity structure in the western Shikoku area. Remarkable low resistivity anomalies were detected in the middle crust of the hanging wall above the subducting Philippine Sea slab. However, no remarkable conductive anomaly was shown around the area of L-SSE in the Bungo channel region. Thus we also examined how the Network-MT responses are sensitive to temporal resistivity change along the slab because the L-SSE repeatedly occurred about every 6 or 7 years. In a test model, 10 km thick conductive layer is assigned just beneath the top surface of the slab. When the resistivity is set to be 20 Ohm-m (thus 500 S anomaly), RMS misfit was significantly increased from the best fit model. It indicates that temporal change could be detected due to current Network-MT survey.
We successfully obtained long-period MT impedances and preliminary resistivity distributions around western Nankai slow earthquake area based on the land and marine MT observations. Thus we plan to integrate the marine and land MT data in this study and past network-MT data in Kyushu Island (e.g. Hata et al., 2015) and estimate reliable 3-D resistivity structure covering the slow EQ areas in the western Nankai subduction zone.
Figure: Locations of OBEM sites and baselines of network-MT