4 December 2018 in Journal of Geophysical Research: Solid Earth
The prediction of volcanic eruptions and the evaluation of associated risks remain a timely and unresolved issue. This paper presents a method to automatically classify seismic events linked to volcanic activity. As increased seismic activity is an indicator of volcanic unrest, automatic classification of volcano‐seismic events is of major interest for volcano monitoring. The proposed architecture is based on supervised classification, whereby a prediction model is built from an extensive dataset of labeled observations. Relevant events should then be detected. Three steps are involved in the building of the prediction model: (i) signals pre‐processing ; (ii) represention of the signals in the feature space; (iii) use of an automatic classifier to train the model. Our main contribution lies in the feature space where the seismic observations are represented by 102 features gathered from both acoustic and seismic fields. Ideally, observations are separable in the feature space, depending on their class. The architecture is tested on 109,609 seismic events that were recorded between June 2006 and September 2011 at Ubinas Volcano, Peru. Six main classes of signals are considered: long period events, volcanic tremors, volcano‐tectonic events, explosions, hybrid events, and tornillos. Our model reaches 93.5% ± 0.50% accuracy, thereby validating the presented architecture and the features used. Furthermore, we illustrate the limited influence of the learning algorithm used (i.e., random forest, support vector machines) by showing that the results remain accurate regardless of the algorithm selected for the training stage. The model is then used to analyze 6 years of data.
27 March 2020 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
CFCarlos Alexandre Rolim Fernandes
Carlos Alexandre Rolim Fernandes
+1
This article proposes the design of an automatic classifier using the empirical mode decomposition (EMD) along with machine learning techniques for identifying the five most important types of events of the Ubinas volcano, the most active volcano in Peru. The proposed method uses attributes from temporal, spectral, and cepstral domains, extracted from the EMD of the signals, as well as a set of preprocessing and instrument correction techniques. Due to the fact that multichannel sensors are currently being installed in seismic networks worldwide, the proposed approach uses a multichannel sensor to perform the classification, contrary to the usual approach of the literature of using a single channel. The presented method is scalable to use data from multiple stations with one or more channels. The principal component analysis method is applied to reduce the dimensionality of the feature vector and the supervised classification is carried out by means of several machine learning algorithms, the support vector machine providing the best results. The presented investigation was tested with a large database that has a considerable number of explosion events, measured at the Ubinas volcano, located in Arequipa, Peru. The proposed classification system achieved a success rate of more than 90%.
19 April 2021 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
This article proposes a supervised tensor-based learning framework for classifying volcano-seismic events from signals recorded at the Ubinas volcano, in Peru, during a period of great activity in 2009. The proposed method is fully tensorial, as it integrates the three main steps of the automatic classification system (feature extraction, dimensionality reduction, and classifier) in a general multidimensional framework for tensor data, joining tensor learning techniques such as the multilinear principal component analysis (MPCA) and the support tensor machine (STM). By exploiting the use of multiple multichannel triaxial sensors, operating simultaneously in two seismic stations, the tensor patterns are constructed as stations × channels × features . The multidimensional structure of the data is then preserved, avoiding the tensor vectorization that often leads to a feature vector with a large dimension, which increases the number of parameters and may cause the “curse of dimensionality.”Moreover, the array vectorization breaks down the multidimensional structure of the data, which usually leads to performance degradation. The results showed a good performance of the proposed multilinear classification system, significantly outperforming its vectorial counterparts. The best result was obtained with the STuM classifier along with the MPCA.
25 March 2024 in Scientific Reports
The prediction of pathological changes on single cell behaviour is a challenging task for deep learning models. Indeed, in self-supervised learning methods, no prior labels are used for the training and all of the information for event predictions are extracted from the data themselves. We present here a novel self-supervised learning model for the detection of anomalies in a given cell population, StArDusTS. Cells are monitored over time, and analysed to extract time-series of dry mass values. We assessed its performances on different cell lines, showing a precision of 96% in the automatic detection of anomalies. Additionally, anomaly detection was also associated with cell measurement errors inherent to the acquisition or analysis pipelines, leading to an improvement of the upstream methods for feature extraction. Our results pave the way to novel architectures for the continuous monitoring of cell cultures in applied research or bioproduction applications, and for the prediction of pathological cellular changes.
24 March 2026 in Natural Hazards and Earth System Sciences
Wildfires are key ecological agents in the Gran Chaco, one of the world's largest tropical dry forest systems. We analyzed more than 100 000 fire patches across the Wet, Dry and Very Dry Chaco between 2001 and 2022, to quantify environmental and anthropogenic controls on fire size. Fire sizes were strongly right-skewed: more than 80 % were smaller than 5 km2, yet large and extreme fires dominated total burned area. Megafires (>100 km2) occurred in all subregions, while gigafires (>1000 km2) were rare but concentrated in the Dry Chaco. Fire Weather Index–burned area correlations exhibited strong spatial contrasts, reaching values of up to r=0.7 in the Wet Chaco and showing weaker, more heterogeneous relationships in drier regions. Meteorological conditions during fires, particularly persistent strong winds, were associated with larger and more elongated patches. Random Forest models showed that topography and land cover composition together accounted for about 60 % of total SHAP importance, whereas demographic variables had very low SHAP contributions in the models. Human pressures shape ignition timing but showed limited direct influence on fire size once landscape structure was included in the models. These results provide a quantitative basis for improving regional fire danger assessments in the Gran Chaco.