|
Abstract:
The seismic properties, longitude, latitude, depth, magnitude, and intensity, etc. of each earthquake are recorded in the public archive of the Central Weather Bureau (CWB) of Taiwan. Two types of earthquakes, labelled and unlabeled, are stored in the data base. The labelled ones have Richter magnitude scale larger than 4.0, and influence several cities and counties. On the other hand, the unlabeled ones only affect in the locality. Twenty-three years and five months - January 1995 to May 2018 are collected from CWB's archive in this study.
The purpose of this paper is to check the stationary of the monthly total (labelled plus unlabeled) and labelled number of earthquakes as well as their energy released time series in the 20 municipal cities and counties in Taiwan to avoid the spurious condition when performing regression analysis. The augmented Dickey-Fuller (ADF) method is used to check the stationary of monthly number of earthquakes and their energy released time series. After a careful hypothesis test at 5% significance, the following time series are stationary. They are: monthly total number of earthquakes; monthly labelled number of earthquakes; energy released by the monthly total number of earthquakes and energy released by the monthly labelled earthquakes. However, the time series of yearly number of earthquakes is no stationary. The stationary of the yearly labelled number of earthquakes is indecisive due to the lack of enough data. A positive relationship between earthquake numbers and energy released has been obtained. The energy is calculated in terms of the code-named Little Boy atomic bomb, which was dropped in Hiroshima, Japan. The coefficient of determination, which means the percentage of the independent variable can be explained through knowledge of the variability in the independent variable, which is as high as 25.57% for the energy released and labelled monthly number of earthquakes.
|
