Project Details
Description
Large earthquakes, together with other hazards they trigger, are the deadliest of all natural disasters. Hence, developing earthquake forecast is not only an urgent task for the public but also an ultimate goal of the scientists. So far, pre-seismic macroscopic anomalies have often been applied for the study of earthquake prediction and forecast. Particularly, electromagnetic signal anomalies are widely adopted in the earthquake precursor study. In the previous studies, we analyzed geoelectric fields with continuous monitoring in Taiwan through statistical methods, such as skewness and kurtosis, examined the correlation between statistical anomalies and earthquakes, and then evaluate forecasting performance through retrospective earthquake forecasts. The successful rate is close to 60%. Besides, the computation for optimizing the model parameters took an amount of time because we evaluated the model performance by using grid search. In order to improve the shortcomings, in this project, we will use a diversity of time series analysis, including natural time analysis, detrendedfluctuation analysis, Fisher-Shannon information analysis, and network topology analysis. We will analyze the electromagnetic signals and extract some characteristics related to earthquakes through these above-mentioned methods. Next, we examine the correlation between the extracted characteristics and earthquakes and compare the model performance among those analysis methods. This way, we expect to get a better technique extracting the earthquake-related precursory signals from electromagnetic data. Moreover, in order to reduce the computation time for optimizing the model parameters, we attempt to use a deep neural network of machine learning and develop a robust algorithm for earthquake forecasting. Through this project, we look forward tosharpening our understanding of seismo-electromagnetics and, in practice, realizing the feasibility of the precursor-based earthquake forecasting.
Status | Finished |
---|---|
Effective start/end date | 1/08/21 → 31/07/22 |
UN Sustainable Development Goals
In 2015, UN member states agreed to 17 global Sustainable Development Goals (SDGs) to end poverty, protect the planet and ensure prosperity for all. This project contributes towards the following SDG(s):
Keywords
- natural time analysis
- detrended fluctuation analysis
- Fisher-Shannon information analysis
- network topology analysis
- electromagnetic signal anomaly
- machine learning
- deep neural network
- earthquake precursor
- earthquake forecast
Fingerprint
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.