Skip to main content
Earth, Planets and Space
Fig. 5 | Earth, Planets and Space

Fig. 5

From: Application of deep learning-based neural networks using theoretical seismograms as training data for locating earthquakes in the Hakone volcanic region, Japan

Fig. 5

Examples of hypocenter estimation in the verification phase for one earthquake (top left, a) Seismic wave propagation images starting at approximately 6 s after the origin time. The time evolution of the estimated hypocenter parameters is shown for magnitude (top right, b), longitude (middle left, c), latitude (middle right, d), depth (bottom left, e) and origin time (bottom right, f). The horizontal axis of each panel represents the time in seconds, and the estimation starts at approximately 6 s after the origin time when the absolute amplitude exceeds 1.0e−7 m/s. The dotted line in each figure represents the parameter to be estimated by each snapshot. The gray line in each figure shows the value estimated using each snapshot. The colored line shows the average of the estimated value up to each time instant. It should be noted that in the verification of the latitude, shown in d, the estimated value is in agreement with the actual value, and the overlap of the three lines indicates that no differences exist.

Back to article page