LOC-FLOW: An End-to-End Machine Learning-Based High-Precision Earthquake Location Workflow
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abstract The ever-increasing networks and quantity of seismic data drive the need for seamless and automatic workflows for rapid and accurate earthquake detection and location. In recent years, machine learning (ML)-based pickers have achieved remarkable accuracy and efficiency with generalization, and thus can significantly improve the earthquake location accuracy of previously developed sequential location methods. However, the inconsistent input or output (I/O) formats between multiple packages often limit their cross application. To reduce format barriers, we incorporated a widely used ML phase picker—PhaseNet—with several popular earthquake location methods and developed a “hands-free” end-to-end ML-based location workflow (named LOC-FLOW), which can be applied directly to continuous waveforms and build high-precision earthquake catalogs at local and regional scales. The renovated open-source package assembles several sequential algorithms including seismic first-arrival picking (PhaseNet and STA/LTA), phase association (REAL), absolute location (VELEST and HYPOINVERSE), and double-difference relative location (hypoDD and GrowClust). We provided different location strategies and I/O interfaces for format conversion to form a seamless earthquake location workflow. Different algorithms can be flexibly selected and/or combined. As an example, we apply LOC-FLOW to the 28 September 2004 Mw 6.0 Parkfield, California, earthquake sequence. LOC-FLOW accomplished seismic phase picking, association, velocity model updating, station correction, absolute location, and double-difference relocation for 16-day continuous seismic data. We detected and located 3.7 times (i.e., 4357) as many as earthquakes with cross-correlation double-difference locations from the Northern California Earthquake Data Center. Our study demonstrates that LOC-FLOW is capable of building high-precision earthquake catalogs efficiently and seamlessly from continuous seismic data.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.002 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.000 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it