Thesis Aplicación del método bayesiano y Abic para obtener incertidumbre asociada la inversión de tsunami en el evento de Iquique 2014
Loading...
Date
2018
Journal Title
Journal ISSN
Volume Title
Program
Ingeniería Civil
Departament
Campus
Campus Casa Central Valparaíso
Abstract
En este trabajo se presenta el método bayesiano y ABIC para calcular incertidumbre en la inversión de tsunamis basado en las observaciones de los mareógrafos y boyas DART. De esta forma se calcula retrospectivamente la deformación inicial de la superficie del mar y su propagación hacia la costa, obteniendo estimaciones de altura de ola y tiempos de arribo, del evento de Iquique 2014 y sus correspondientes incertidumbres. Para este cálculo se aplica el método de Inversión de la forma de la onda de Tsunami [Tsushima et al., 2009] que consiste en una serie de funciones de Green, las cuales, al ser modeladas en el programa de simulación de tsunamis COMCOT, generan variaciones del nivel del mar en los mareógrafos y boya DART que son almacenadas en una base de datos. Al generarse un tsunami estos mareógrafos, que en el caso del evento de Iquique son en especial Arica, Iquique, Patache, Mejillones y la DART 32401, captan las variaciones de nivel, las cuales son el dato de entrada al algoritmo. De las series de tiempo de la base de datos, se plantean combinaciones lineales de parámetros desconocidos, calculándolos mediante mínimos cuadrados y Akaike’s Bayesian Information Criterion (ABIC) [Akaike, 1980] para obtener de forma estimada las variaciones que captan los mareógrafos. Este método, además de entregar un resultado promedio de la superficie inicial, también arroja la incertidumbre asociada a cada parámetro de la combinación lineal, la cual disminuye a medida que se obtienen más datos.
This study presents the bayesian method and ABIC to calculate uncertainty in tsunami inversion based on tide and DART’s data observations. By this way it’s calculated sea’s initial deformation and its propagation to the coast in real time, obtaining seawave heights and arriving times estimates, applied to Iquique 2014 event. For this calculation it’s implemented the tsunami waveform inversion method [Tsushima et al., 2009] consisting of a set of unit Gaussian sources, modeled by the software of tsunami simulations COMCOT, that produce variations of the ocean’s level in the tides and Dart buoy that are stored in a database. When a tsunami occurs, these tides, hich in the case of Iquique’s event are Arica, Iquique, Patache, Mejillones and DART 32401 buoy, they record these variations that are the data input for the algorithm. From the time series of the database, it is proposed lineal combinations of unknown parameters, calculating them by least squares method and Akaike’s Bayesian Information Criterion (ABIC) [Akaike, 1980] to obtain estimates of the variations recorded by the tides. This method, in addtion of delivering an average result of thinitial surface, it also shows the uncertainty asociated to each parameter of the lineal combination, which decline when more information in real time is available.
This study presents the bayesian method and ABIC to calculate uncertainty in tsunami inversion based on tide and DART’s data observations. By this way it’s calculated sea’s initial deformation and its propagation to the coast in real time, obtaining seawave heights and arriving times estimates, applied to Iquique 2014 event. For this calculation it’s implemented the tsunami waveform inversion method [Tsushima et al., 2009] consisting of a set of unit Gaussian sources, modeled by the software of tsunami simulations COMCOT, that produce variations of the ocean’s level in the tides and Dart buoy that are stored in a database. When a tsunami occurs, these tides, hich in the case of Iquique’s event are Arica, Iquique, Patache, Mejillones and DART 32401 buoy, they record these variations that are the data input for the algorithm. From the time series of the database, it is proposed lineal combinations of unknown parameters, calculating them by least squares method and Akaike’s Bayesian Information Criterion (ABIC) [Akaike, 1980] to obtain estimates of the variations recorded by the tides. This method, in addtion of delivering an average result of thinitial surface, it also shows the uncertainty asociated to each parameter of the lineal combination, which decline when more information in real time is available.
Description
Catalogado desde la version PDF de la tesis.
Keywords
Akaike, Inversión de tsunami, Iquique 2014, Método Bayesiano
