Summary

FEMALE implements an unprecedented and new approach for the analysis of volcano-seismic data aimed at Forecasting Volcanic Eruptions using Signal Processing and Machine Learning techniques. It will use artificial intelligence techniques to analyse continuous seismic signals from a suite of volcanoes rather than isolated events to create a new parametric database.

Correlating parameter evolution with volcanic processes will advance the analysis of seismo-volcanic signals, improve the management of volcanic crises, and provide reliable tools for forecasting volcanic eruptions.

Analysing a variety of volcanic processes at selected volcanoes, both that culminated and did not culminate in eruptions FEMALE will create an Inventory of Case Studies based on the analysis of past volcanic unrest to be used for future volcanic eruption forecasts. This novel methodology of comparing eruptive and non-eruptive cases is critical to obtaining a true statistical base from which to forecast future eruptions.

FEMALE will establish a new universally applicable, exportable, simple, and useful methodology to forecast volcanic eruptions using continuous and labelled volcano seismic signals. We will derive a new perspective on the relationship between seismic activity and volcano dynamics by focusing on producing successful volcanic eruption forecasts.

FEMALE propose to change the procedures currently used for seismo-volcanic analysis. For the first time this new approach will allow the inclusion of all relevant signal information and supersede the use of compulsory a-priori models. The final product of FEMALE is conceived as a set of useful tools for managing forecasts of volcanic eruptions.

Analysing, detecting and identifying seismo-volcanic patterns and comparing them with specific volcanic processes will make it possible to forecast future volcanic scenarios.

Objectives

The main objective of the FEMALE is to provide reliable tools for Early Forecast Volcanic Eruption by developing innovative seismo-volcanic signal analysis using Machine Learning techniques. This novel ground-breaking approach involves the following sub-objectives and procedures (Figure 3): (O1, WP1) to define the features that best represent the raw continuous seismic waveforms; (O2, WP2) to parametrize a set of long-term data from selected volcanoes to create a new parametric volcano seismic database; (O3, WP3) using ML to identify which of these features change or are invariable during the evolution of volcanic activity and to create an Inventory of Case Studies; (O4, WP4) to export the knowledge to other volcanoes forecasting a probability of similarity with the existing case studies; and (O5, WP5) to improve the procedures to identify and label relevant volcano seismic events to obtain complete seismic catalogues able to be compared with the raw data analysis.

Finally FEMALE introduces a joint discussion integrated in WP6 focused in the comparison between results in Forecasting Volcanic Eruptions using parametric continuous data and labelled seismic data. The challenge in volcanic forecasting is to associate seismic parameters and labelled seismic data with the evolution of the volcano’s activity.

HOW TO CITE THE PROJECT

The Project FEMALE is a project of the Spanish Government with the reference PID2019-106260GB-I00.

FEMALE

FEMALE implements an unprecedented and new approach for the analysis of volcano-seismic data aimed at Forecasting Volcanic Eruptions using Signal Processing and Machine Learning techniques. It will use artificial intelligence techniques to analyse continuous seismic signals from a suite of volcanoes rather than isolated events to create a new parametric database.

Correlating parameter evolution with volcanic processes will advance the analysis of seismo-volcanic signals, improve the management of volcanic crises, and provide reliable tools for forecasting volcanic eruptions.

© 2021 Female

Website designed and developed by Ramón Ruiz