One of the types of IoT data utilized in recent years is footage captured by cameras in stores and other facilities. However, it remains unclear to what extent videos of earthquakes can serve as a novel data source for seismic activity and earthquake-induced anomalies include damage. To address this question, we developed an archive system to gather and preserve earthquake-related video data, which would otherwise be lost or dispersed. We annotated 607 video recordings from nine distinct earthquakes in the earthquake video archive system, adding damage-related tags to the footage. Furthermore, we examined damage characteristics for each seismic intensity and proposed a store damage classification table.

In this study, we estimated the S-wave velocity structure of the deep subsurface using microtremor array data in the Amami Islands, the Tokara Islands, and Yakushima Island. The depth of the top of the seismic basement is about 2 km with the largest depth at Kikaishima Island, and the depths at the other islands are 0.2 to 0.6 km. And also shown are the mean Vs values for each layer and the layer boundary depths at each location for the four-layer model, which were 0.69, 1.02, 2.18, and 3.49 km/s, respectively. These mean values are close to those of the Yaeyama Islands, Miyako Islands and Okinawa Islands. These results can be used as basic data for improving the S-wave velocity structure model for the Nansei Islands.

The authors attempted to evaluate the vibration characteristics (natural period and damping constant of the first-order modes) and story parameters (story stiffness and story damping) of an equivalent shear-type model from experimental results of a full-scale four-story RC building affected by the rocking of a shake table. The identification was based on the method proposed by the authors to approximate the first-order mode response of the sub-structure model. The validity of the identifications was discussed by comparing the results with those of preceding studies using the ARX model. For the case of random wave excitation, which corresponds to microtremors, it was confirmed that the natural period and damping constant were evaluated to be larger without the rocking correction than with it. The effect of the rocking correction on the story stiffness and story damping was found to be similar to the effect of the correction on the natural period and damping constant. Furthermore, it was shown that the natural period during earthquakes may be estimated from the natural period during random wave excitation before and after the earthquake. Finally, it was found that when the control method of the shaking table is basic control, the natural period and damping constant are evaluated to be large due to the rocking of the shaking table, while in the case of input compensation control, they are affected differently for each excitation wave.