Modelación estadística de eventos extremos para la predicción de terremotos en el Ecuador.

Abstract

The objective of this study was to model extreme events and propose a model that allows obtaining probabilities of occurrence of an earthquake, for which the Global Centroid Moment Tensor (CMT) seismic catalog was used, filtering the information from Ecuador between the years 1976-2018. And the seismic catalog of the Geophysical Institute of the National Polytechnic School (IGEPN) between the years 2013-2018. Unifying these catalogs, a georeferencing was carried out, observing that earthquakes with magnitudes between 2.0-6.0 on the Richter scale predominated in the country. For the modeling, from the threshold value 6.4 calculated in the Gutenberg-Richter law, it was identified that extreme events follow a power law since the shape parameter of the Generalized Pareto Distribution (GPD) is greater than zero and in the proposal of the model, the parameters of the Gutenberg-Richter law and the Omori-Utsu law were used, to estimate their parameters, the Maximum Likelihood Estimation method (MLE) was applied, resulting in the Gutenberg-Richter law a level regional seismicity, 11.94 and a proportion of large earthquakes versus small, 1.39, calculating 𝛽 using the Deluca and Corral method, while in the Omori-Utsu law taking the earthquake of April 16, 2016, the aftershock productivity obtained, 171.91, higher number of aftershocks after the main earthquake, 0.083 and an aftershock decay rate, 0.96. Finally, the proposed model was compared with the Epidemic-type Aftershock Sequence (ETAS) model using the mean square error and the root mean square, giving a lower error in the proposed model, concluding that it also helps to obtain probabilities of occurrence. It is recommended to carry out more studies applied to this line of research since the seismicity in the country is high due to its geographical location and these studies would help significantly.