Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Catania, Osservatorio Etneo, Piazza Roma 2, 95125, Catania, Italy.
Istituto Nazionale di Geofisica e Vulcanologia, Sezione Roma2, Via di Vigna Murata 605, 00143, Roma, Italy.
Sci Rep. 2019 Apr 24;9(1):6506. doi: 10.1038/s41598-019-42930-3.
Early-warning assessment of a volcanic unrest requires that accurate information from monitoring is continuously gathered before volcanic activity starts. Seismic data are an optimal source of such information, overcoming safety problems due to dangerous conditions for field surveys or cloud cover that may hinder visibility. We designed a multi-station warning system based on the classification of patterns of the background seismic radiation, so-called volcanic tremor, by using Self-Organizing Maps (SOM) and fuzzy clustering. The classifier automatically detects patterns that are typical footprints of volcanic unrest. The issuance of the SOM colors on DEM allows their geographical visualization according to the stations of detection; this spatial location makes it possible to infer areas potentially impacted by eruptive phenomena. Tested at Mt. Etna (Italy), the classifier forecasted in hindsight patterns associated with fast-rising magma (typical of lava fountains) as well as a relatively long lead time of the outburst (lava flows from eruptive fractures). Receiver Operating Characteristics (ROC) curves gave an Area Under the Curve (AUC) ∼0.8 indicative of a good detection accuracy that cannot be achieved from a mere random choice.
火山活动前，需要不断收集监测到的准确信息，以实现火山活动的早期预警评估。地震数据是此类信息的最佳来源，可以克服因现场调查的危险条件或可能阻碍能见度的云层而导致的安全问题。我们设计了一种基于自组织映射（SOM）和模糊聚类的背景地震辐射（即火山震颤）模式分类的多站预警系统。分类器可自动检测到火山活动异常的典型特征。SOM 颜色在数字高程模型上的发布允许根据探测站进行地理可视化；这种空间位置使得推断出可能受到喷发现象影响的区域成为可能。在埃特纳火山（意大利）进行的测试中，分类器对与快速上升岩浆（典型的熔岩喷泉）相关的模式进行了回溯预测，以及爆发的相对较长的前置时间（从喷发裂缝流出的熔岩）。接收者操作特性（ROC）曲线的曲线下面积（AUC）约为 0.8，表明检测准确性较高，这是单纯随机选择无法达到的。
