Abstract
SUMMARYIn a recent study, we showed that convolutional neural networks (CNNs) applied to network seismic traces can be used for rapid prediction of earthquake peak ground motion intensity measures (IMs) at distant stations using only recordings from stations near the epicentre. The predictions are made without any previous knowledge concerning the earthquake location and magnitude. This approach differs significantly from the standard procedure adopted by earthquake early warning systems that rely on location and magnitude information. In the previous study, we used 10 s, raw, multistation (39 stations) waveforms for the 2016 earthquake sequence in central Italy for 915 M ≥ 3.0 events (CI data set). The CI data set has a large number of spatially concentrated earthquakes and a dense network of stations. In this work, we applied the same CNN model to an area of central western Italy. In our initial application of the technique, we used a data set consisting of 266 M ≥ 3.0 earthquakes recorded by 39 stations. We found that the CNN model trained using this smaller-sized data set performed worse compared to the results presented in the previously published study. To counter the lack of data, we explored the adoption of ‘transfer learning’ (TL) methodologies using two approaches: first, by using a pre-trained model built on the CI data set and, next, by using a pre-trained model built on a different (seismological) problem that has a larger data set available for training. We show that the use of TL improves the results in terms of outliers, bias and variability of the residuals between predicted and true IM values. We also demonstrate that adding knowledge of station relative positions as an additional layer in the neural network improves the results. The improvements achieved through the experiments were demonstrated by the reduction of the number of outliers by 5 per cent, the residuals R median by 39 per cent and their standard deviation by 11 per cent.
Highlights
Having information about earthquake generated ground motions in the shortest time possible (Minson et al, 2018) is of great importance for earthquake monitoring
In a recent study (Jozinović et al, 2020) we showed that convolutional neural networks (CNNs) applied to network seismic traces can be used for rapid prediction of earthquake peak ground motion intensity measures (IMs) at distant stations using only recordings from stations near the epicenter
For experiments 1-5, we report the results of residuals on all the data here, and the results are split into residuals on observed target labels and ShakeMap-derived target labels in the electronic supplement
Summary
Having information about earthquake generated ground motions in the shortest time possible (Minson et al, 2018) is of great importance for earthquake monitoring. The outputs of the ML model were the intensity measures (IMs; i.e., PGA, PGV and SA at 0.3 s, 1 s, 3 s periods) on the selected stations. This configuration entails that the strongest shaking would be recorded on the traces of the stations closest to the epicenter, while the model would give predictions for the stations farther from the epicenter. It was found that the ML model was able to predict with useful accuracy the IMs at the stations which had no input data available (and were replaced with a window of zeros)
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
