地震地磁观测与研究2025，Vol.46Issue(6)：25-34,10.DOI:10.3969/j.issn.1003-3246.2025.06.003
基于BP神经网络的福建地震台网最远台站震中距的预测研究
Research on predicting the epicentral distance of the farthest station in Fujian Seismic Network based on BP neural network
摘要
Abstract
This article analyzes the relationship between the earthquake magnitudes in Fujian and Taiwan and the epicentral distance between the farthest station recorded by the Fujian Seismic Network(the epicentral distance of the farthest station)using BP neural network,focusingon the influences of earthquake magnitude,focal depth,and geographical location on the epicentral distance of the farthest station.We used multiple linear regression model and BP neural network model to obtain the quantitative relationship between earthquake magnitude and the farthest epicentral distance in the Fujian-Taiwan region,and evaluated the advantages and disadvantages of the model through four indicators.The comparison shows that the BP neural network model has a better fitting effect.This is because the BP neural network comprehensively considers the influence of earthquake longitude,latitude,depth,and magnitude on the epicentral distance of the farthest station,and possesses strong nonlinear fitting capability.The quantitative relationship between magnitude and the maximum epicentral distance that is defined by a quadratic polynomial regression model(D=27.27M 2+38.13M-34.63)and a BP neural network model(7 hidden layers with 9 nodes)provides a new approach to assist in detecting missed seismic phases in manual seismic catalogs.关键词
BP神经网络/非线性拟合/震级/最远台站震中距Key words
BP neural network/nonlinear fitting/earthquake magnitude/epicentral distance of the farthest station引用本文复制引用
林玮瑾,林彬华,杨源鸿,李军,韦永祥,丁炳火..基于BP神经网络的福建地震台网最远台站震中距的预测研究[J].地震地磁观测与研究,2025,46(6):25-34,10.基金项目
中国地震局监测预警任务(一般)(项目编号:CEA-JCYJ-202501031) （一般）
福建省地震局青年基金(项目编号:Q202401) （项目编号:Q202401）
国家自然科学基金青年项目(项目编号:42104062) （项目编号:42104062）
地震科技星火计划攻关项目(项目编号:XH23024A) （项目编号:XH23024A）
