[S-CG59] Oceanic plate as inputs to subduction zone: evolution process of the oceanic plate

[SCG59-12] Evaluation of heat flow of Hi-net boreholes in the Kii Peninsula and estimation of heat flow distribution by integrating geophysical datasets

Takehiro Kobira¹, *Kazuya Ishitsuka¹, Weiren Lin¹, Kentaro Omura², Masashi NAGAI², Takumi Matsumoto² (1.Kyoto University, 2.National Research Institute for Earth Science and Disaster Resilience)

Keywords:Heat flow, Kii Peninsula, Hi-net, Philippine Sea Plate, Slab-derived fluids

The Kii Peninsula shows distinctive heat flow associated with the subduction of the Philippine Sea Plate, and seismic activity and temperature anomalies in the region are related to slab-derived fluids. This study investigated heat flow distribution of the Kii Peninsula, Japan, by refining the heat flow at the High Sensitivity Seismograph Network (Hi-net) boreholes of NIED and compiling other available heat flow data. Hi-net stations are distributed at intervals of about 15–20 km throughout Japan, and some observation boreholes are located at the Kii Peninsula. The drilling cuttings and the deepest 25 m core of the Hi-net boreholes are stored in each observatory facility. To refine the heat flow at the Hi-net boreholes, we measured thermal conductivities of drill cutting in addition to cores, and temperature logs were corrected by considering climatic effects. The spatial distribution of heat flow of the Kii Peninsula was further estimated by training the relationship between heat flow and available geophysical datasets, such as slab depth, gravity anomaly, and Moho depth, at and around the target region, using extreme gradient boosting (XGBoost), a machine learning technique.
The measurements of drill cuttings and cores identified the thermal conductivities in different lithologies. Compiling the Hi-net and other available heat flow data, we found that two peaks of high heat flow may exist with respect to distance from the trough axis. One peak corresponds to the previously known high heat flow zone, which is located approximately 100 km from the trough axis in the direction of plate subduction, roughly consistent with the location where deep low-frequency earthquakes have been reported. The other is located at the Hi-net borehole locations, approximately 130 km from the trough axis. This newly estimated peak of high heat flow is likely to be substantial even accounting for the uncertainties in thermal conductivity and the effects of topography and sedimentation rate on temperature profiles. The heat flow distribution estimated by XGBoost further suggested the existence of the two zones of high heat flow roughly parallel to the trough axis. Furthermore, 3D temperature distribution was estimated from the heat flow and temperature data. Assuming that the cause of this characteristic temperature distribution is due to the rise of high-temperature fluids, this high heat flow could be explained by the rising crustal fluid at depth.