CFE 2019: Start Registration
View Submission - CMStatistics
Title: Functional data analysis application in TEC disturbance caused by Tsunami Authors:  Ryuichi Kanai - University College London (United Kingdom) [presenting]
Serge Guillas - University College London (United Kingdom)
Abstract: The ionospheric plasma disturbance caused by a large tsunami after subduction earthquake can be detected by measurement of the total electron content (TEC) between global positioning system (GPS) satellites and their receivers. TEC depression which is termed as tsunami ionospheric hole (TIH) lasting for several tens of minutes is formed above the tsunami source area. However, even if the network of GPS receiving stations is dense, it is impossible to detect all data around tsunami source. Firstly, we propose a method using bivariate spline fitting to interpolate the detected TEC data. By this method, it becomes possible to estimate tsunami source using this fitted data. In addition, using this fitting method, less GPS receivers can show a sufficient estimation of TEC disturbance. Secondly, with this fitted data, we succeeded in evaluating the TEC dip quantitatively using Functional Principal Component Analysis (FPCA) score on a fixed longitude. Based on the FPCA scores, we can estimate when the TEC disturbance becomes the largest. Thirdly, applying the functional linear regression to these data, we can estimate the TEC disturbance data at least 30 seconds before.