Calibración y validación del algoritmo de Bakun and Wentworth para calcular localización y magnitud de Terremotos históricos a partir de datos macrosísmicos en Cuba orienTal

Abstract

We present the results of a calibration analysis performed on eastern Cuba seismicity datasets of the approach proposed by Bakun and Wentworth (1997) to bound earthquakes epicentral area and magnitude from MSK macroseismic data only. The calibration is required as the algorithm derives an intensity magnitude (equal in the mean to moment magnitude Mw) from macroseismic observations by using a regionally suitable attenuation relationship of intensity as a function of Mw and source distance. To this purpose, a training set of 8 Post-900 earthquakes occurred in eastern Cuba was selected, for which a large number of intensity observations and reliable instrumental determinations of Mw and epicentral location are available. The model is given by:

I

=(3.11±0.40)+(1.40±0.09)*Mw-(0.0017±0.00027)*R-(3.17±0.20)*LogR

MSK

where R is the hypocentral distance in kilometres, the model is calibrated for earthquakes in the range of 5.22≥Mw≥6.81, hypocentral distances less than 400km and fixed depth at 10km. The model is validated using the 1947, 1976 and 1998 earthquakes, which were not used in the calibration process and also have instrumental data. The validation of the algorithm demonstrates reliability and applicability of the method for the analysis of historical earthquakes that have no records of instrumental parameters. An overall agreement is found for both the location and magnitude of these events with magnitude uncertainties between 0.23 and 0.35 equivalent to two standard deviations. However the MDPs distributions are partially azimuthal for those 11 analyzed earthquakes, the model can reproduce offshore epicentres within acceptable uncertainties ranges. A first application is made to a historical earthquake occurred on June 6, 1766 causing severe damage to the city of Santiago de Cuba.