ALBERT

Description: 〈span〉〈div〉ABSTRACT〈/div〉Effective early warning, emergency response, and information dissemination for earthquakes and tsunamis require rapid characterization of an earthquake’s location, size, and other parameters, usually provided by real‐time seismogram analysis using established, rule‐based, seismological procedures. Powerful, new machine learning (ML) tools analyze basic data using little or no rule‐based knowledge, and an ML deep convolutional neural network (CNN) can operate directly on seismogram waveforms with little preprocessing and without feature extraction. How a CNN will perform for rapid automated earthquake detection and characterization using short single‐station waveforms is an issue of fundamental importance for earthquake monitoring.For an initial investigation of this issue, we adapt an existing CNN for local earthquake detection and epicentral classification using single‐station waveforms (〈a href="https://pubs.geoscienceworld.org/srl#rf24"〉Perol 〈span〉et al.〈/span〉, 2018〈/a〉), to form a new CNN, ConvNetQuake_INGV, to characterize earthquakes at any distance (local to far‐teleseismic). ConvNetQuake_INGV operates directly on 50‐s three‐component broadband single‐station waveforms to detect seismic events and obtain binned probabilistic estimates of the distance, azimuth, depth, and magnitude of the event. The best performance of ConvNetQuake_INGV is obtained using a last convolutional layer with fewer nodes than the number of output classifications, a form of information bottleneck.We show that ConvNetQuake_INGV detects very well (accuracy 87%) and characterizes moderately well earthquakes over a broad range of distances and magnitudes, and we analyze outlier results and indications of overfitting of the CNN training data. We find weak evidence that the CNN is performing more than high‐dimensional regression and pattern recognition, and is generalizing information or learning, to provide useful characterization of new events not represented in the training data. We expect that real‐time ML procedures such as ConvNetQuake_INGV, perhaps incorporating rule‐based knowledge, will ultimately prove valuable for rapid detection and characterization of earthquakes for earthquake response and tsunami early warning.〈/span〉

Description: 〈span〉〈div〉ABSTRACT〈/div〉Effective early warning, emergency response, and information dissemination for earthquakes and tsunamis require rapid characterization of an earthquake’s location, size, and other parameters, usually provided by real‐time seismogram analysis using established, rule‐based, seismological procedures. Powerful, new machine learning (ML) tools analyze basic data using little or no rule‐based knowledge, and an ML deep convolutional neural network (CNN) can operate directly on seismogram waveforms with little preprocessing and without feature extraction. How a CNN will perform for rapid automated earthquake detection and characterization using short single‐station waveforms is an issue of fundamental importance for earthquake monitoring.For an initial investigation of this issue, we adapt an existing CNN for local earthquake detection and epicentral classification using single‐station waveforms (〈a href="https://pubs.geoscienceworld.org/srl#rf24"〉Perol 〈span〉et al.〈/span〉, 2018〈/a〉), to form a new CNN, ConvNetQuake_INGV, to characterize earthquakes at any distance (local to far‐teleseismic). ConvNetQuake_INGV operates directly on 50‐s three‐component broadband single‐station waveforms to detect seismic events and obtain binned probabilistic estimates of the distance, azimuth, depth, and magnitude of the event. The best performance of ConvNetQuake_INGV is obtained using a last convolutional layer with fewer nodes than the number of output classifications, a form of information bottleneck.We show that ConvNetQuake_INGV detects very well (accuracy 87%) and characterizes moderately well earthquakes over a broad range of distances and magnitudes, and we analyze outlier results and indications of overfitting of the CNN training data. We find weak evidence that the CNN is performing more than high‐dimensional regression and pattern recognition, and is generalizing information or learning, to provide useful characterization of new events not represented in the training data. We expect that real‐time ML procedures such as ConvNetQuake_INGV, perhaps incorporating rule‐based knowledge, will ultimately prove valuable for rapid detection and characterization of earthquakes for earthquake response and tsunami early warning.〈/span〉

Description: Seismic event detection and phase picking are the base of many seismological workflows. In recent years, several publications demonstrated that deep learning approaches significantly outperform classical approaches, achieving human-like performance under certain circumstances. However, as studies differ in the datasets and evaluation tasks, it is unclear how the different approaches compare to each other. Furthermore, there are no systematic studies about model performance in cross-domain scenarios, i.e., when applied to data with different characteristics. Here, we address these questions by conducting a large-scale benchmark. We compare six previously published deep learning models on eight datasets covering local to teleseismic distances and on three tasks: event detection, phase identification and onset time picking. Furthermore, we compare the results to a classical Baer-Kradolfer picker. Overall, we observe the best performance for EQTransformer, GPD and PhaseNet, with a small advantage for EQTransformer on teleseismic data. Furthermore, we conduct a cross-domain study, analyzing model performance on datasets they were not trained on. We show that trained models can be transferred between regions with only mild performance degradation, but models trained on regional data do not transfer well to teleseismic data. As deep learning for detection and picking is a rapidly evolving field, we ensured extensibility of our benchmark by building our code on standardized frameworks and making it openly accessible. This allows model developers to easily evaluate new models or performance on new datasets. Furthermore, we make all trained models available through the SeisBench framework, giving end-users an easy way to apply these models.

Description: Machine-learning (ML) methods have seen widespread adoption in seismology in recent years. The ability of these techniques to efficiently infer the statistical properties of large datasets often provides significant improvements over traditional techniques when the number of data are large (millions of examples). With the entire spectrum of seismological tasks, for example, seismic picking and detection, magnitude and source property estimation, ground-motion prediction, hypocenter determination, among others, now incorporating ML approaches, numerous models are emerging as these techniques are further adopted within seismology. To evaluate these algorithms, quality-controlled benchmark datasets that contain representative class distributions are vital. In addition to this, models require implementation through a common framework to facilitate comparison. Accessing these various benchmark datasets for training and implementing the standardization of models is currently a time-consuming process, hindering further advancement of ML techniques within seismology. These development bottlenecks also affect “practitioners” seeking to deploy the latest models on seismic data, without having to necessarily learn entirely new ML frameworks to perform this task. We present SeisBench as a software package to tackle these issues. SeisBench is an open-source framework for deploying ML in seismology—available via GitHub. SeisBench standardizes access to both models and datasets, while also providing a range of common processing and data augmentation operations through the API. Through SeisBench, users can access several seismological ML models and benchmark datasets available in the literature via a single interface. SeisBench is built to be extensible, with community involvement encouraged to expand the package. Having such frameworks available for accessing leading ML models forms an essential tool for seismologists seeking to iterate and apply the next generation of ML techniques to seismic data.

Description: Automated procedures for seismic arrival-time picking on large and real-time seismological waveform datasets are critical for many seismological tasks. Recent, high-performance, automated arrival-time pickers mainly use deep-neural-networks applied to nearly raw, seismogram waveforms as input data. However, there is a long history in earthquake seismology of rule-based, automated arrival detection and picking algorithms that efficiently exploit variations in amplitude, frequency and polarization of seismogram waveforms.Here we use this classical, seismological domain-knowledge to transform raw seismogram waveforms into input features for a deep-learning picker. We preprocess 3-component, broadband seismograms into 3-component characteristic functions of a multi-band picker (FilterPicker), plus the instantaneous modulus and inclination of the waveforms. We use these five time-series as input instead of the 3-component, raw seismograms to extend the deep-neural-network picker PhaseNet within the SeisBench platform. We compare the original, purely data-driven PhaseNet and our extended, domain-knowledge PhaseNet (DKPN), using identical training and validation datasets, with application to in- and cross-domain testing datasets.We find that the explicit information targeting arrival-time detection and picking introduced by the domain-knowledge processing enables DKPN to be trained with smaller datasets than PhaseNet. Relative to PhaseNet, DKPN shows improved performance and stability for P picking and slightly improved S picking, especially for cross-domain application. With increasing training dataset size PhaseNet performance generally improves and converges to that of DKPN, except for cross-domain P picking. The results suggest that DKPN primarily needs to learn pick characterization, while PhaseNet additionally requires learning the more difficult task of arrival detection.

Description: The Santiaguito Volcanic Complex is located in Guatemala’s Western Volcanic Highlands, along the westernmost section of the Central American Volcanic Arc. Santiaguito presents multiple hazards (including pyroclastic density flows, lava flows, lahars, explosions and debris avalanches) to the local population, which numbers 1.6 million people within the 30 km2 around the volcano. Lahars are mixtures of water and pyroclastic debris that includes gases which can rapidly initiate and flow at speeds of tens of meters per second down wide barrancas (canyons) making them highly destructive. Lahar occurrence strongly correlates with rainfall at the volcano and they are commonplace in the long rainy season. They pose a great hazard to local inhabitants who regularly cross the channels as they live and work on farms on or near the flanks of the barrancas. INSIVUMEH, the national seismic and volcano monitoring agency, has recently built a network of 10 seismic stations that can monitor these flows, in collaboration with external agencies. Using the seismic data from the 2022 rainy season we have built a Lahar catalogue. We supplement the dataset by adding synthetic Lahar waveforms to it, informed by the existing recorded events. We then use the hybrid catalogue to develop a Convolutional Neural Network Lahar detector, always testing on real events. We test using different input lengths to find an ideal timeliness/accuracy ratio for Lahar early warning. We plan to extend the method for Lahar location tracking. We expect the method to be portable to other volcanic areas.