Joint Determination of Event Location and Magnitude from Historical Seismic Damage Records

Large earthquakes have long recurrence intervals. It is crucial to consider long-time seismicity for a proper assessment of potential seismic hazards. It is required to use historical earthquake records to complement the long-time seismicity records. Historical earthquake records remain as in seismic damage description with limited accuracy in source parameters including event location and its size. It is important to determine epicenters and magnitudes of historical earthquakes accurately. A noble method to determine the event location and magnitude from historical seismic damage records is introduced. Seismic damage is typically proportional to the event magnitude, and is inversely proportional to the distance. This feature allows us to deduce the event magnitude and location from spatial distribution of seismic intensities. However, the magnitude and distance trade off each other, inhibiting unique determination of event magnitude and location. The Gutenberg-Richter frequency-magnitude relationship is additionally considered to constrain the source parameters. The Gutenberg-Richter frequency-magnitude relationship is assumed to be consistent between instrumental and historical seismicity. A set of event location and magnitude that satisfy the chance of event occurrence according to the Gutenberg-Richter frequency-magnitude relationship is selected. The accuracy of the method is tested for synthetic data sets, and the validity of the method is examined. The synthetic tests present high accuracy of the method. The method is applied to historical seismic damage records, which allows us to calibrate the source parameters of historical earthquakes.