首页|2024年台湾花莲7.4级地震诱发地质灾害应急评价

2024年台湾花莲7.4级地震诱发地质灾害应急评价

扫码查看
2024年4月3日,中国台湾地区花莲县发生了7.4级地震,导致山区发生了大规模地质灾害,造成了严重的人员伤亡和经济损失.迅速准确地评估地震诱发地质灾害的空间分布概率,对震后紧急响应和安置决策具有重要意义.基于台湾同震滑坡数据库与人工智能神经网络算法,建立了一个近实时的地震诱发滑坡空间分布概率预测模型;在地震发生后1小时内成功实现了花莲地震诱发的滑坡空间分布概率预测.在震后5天内利用Sentinel-1A合成孔径雷达(SAR)和PlanetScope光学卫星影像,对地震核心影响区域进行了滑坡智能检测与目视解译.在无云影像覆盖区共解译876处同震滑坡,总面积为12.6 km2,主要分布于台湾中央山脉东侧高山峡谷区.通过已解译滑坡的验证,预测结果的曲线下面积(AUC)精度达到了90%,证明了在此次事件中预测模型的准确性和可靠性,为抗震救灾提供了及时有效的数据支持.
Assessment of geological hazards triggered by the 2024 Mw 7.4 earthquake in Hualien,Taiwan,China
On April 3,2024,an Mw 7.4 earthquake struck Hualien County in Tai wan,China,to trigger a number of extensive geological hazards in the mountainous areas of the region that led to a large number of casualties and significant economic losses.The rapid and accurate assessment of the spatial distribution of earthquake-induced geological hazards is crucial for post-earthquake emergency response and resettlement decisions.In this study,we leverage the Taiwan co-seismic landslide database and artificial intelligence algorithms to established a near-real-time model to predict the spatial distribution of earthquake-induced landslides.The proposed model can generate predictions of the occurrence of landslides within an hour of an earthquake.We mapped landslides by using automatic detection and visually interpreted them in an earthquake-affected region over five days based on synthetic aperture radar images from Sentinel-1A and optical images from the PlanetScope satellites.We used these data to interpret 876 co-seismic landslides over an area of 12.6 km2,primarily in high mountain gorges on the eastern side of the Central Mountain Range.The area under the curve(AUC)of predictions by the proposed model was 90%,which proves its accuracy and reliability.Our model can be used to provide timely and accurate data support for earthquake relief measures.

Hualien earthquakeco-seismic landslideprediction modelartificial intelligenceremote sensing-based detection

方成勇、范宣梅、王欣、戴岚欣、漆基孝

展开 >

地质灾害防治与地质环境保护国家重点实验室(成都理工大学),成都 610059

花莲地震 同震滑坡 预测模型 人工智能 遥感识别

国家杰出青年科学基金项目四川省自然科学基金项目四川省自然科学基金项目

421257022022NSFSC00032022NSFSC1083

2024

成都理工大学学报(自然科学版)
成都理工大学

成都理工大学学报(自然科学版)

CSTPCD北大核心
影响因子:1.596
ISSN:1671-9727
年,卷(期):2024.51(4)
  • 10