Speaker
Description
Machine Learning for Travel Time emulation (MaLTTe) is a deep learning method and computer code for emulating seismic-phase travel times that are based on a three-dimensional (3-D) Earth model. Greater accuracy of travel time predictions using a 3-D Earth model are known to reduce the bias of event location estimates and improve the process of associating detections to events. However, practical use of 3-D models is challenged by slow computational speed and the unwieldiness of pre-computed lookup tables. MaLTTe uses the XGboost method and trains on pre-computed travel times, resulting in a compact and computationally fast way to approximate travel times based on a 3-D Earth model. MaLTTe is trained using approximately 850 million P-wave travel times based on the LLNL-G3D-JPS model from randomly sampled event locations to 10,393 global seismic stations. After training, the MaLTTe code is approximately 10 Mbyes in size and travel times are computed in approximately ten of micro-seconds on a single CPU. Currently achieved prediction accuracy is approximately ~0.2 second, which is significantly smaller than the inherent accuracy of the 3-D model. With additional development, MaLTTe will enable easy use of 3-D models in routine seismological processing and analysis.
Promotional text
3-D Earth models can improve seismic travel-time prediction accuracy, which leads directly to more accurate event locations. Machine learning efficiently emulates travel time calculations, opening the possibility of using state-of-the-art Earth models in the operational system.
| [email protected] | |
| Oral preference format | in-person |
