Loading...
Loading...
Contributors:
Dataset URI:
10.5281/zenodo.23122686Reproducibility release for a separate locked, fixed-window study of submarine DAS covariance learning in Cook Inlet, Alaska. The experiment uses 93 development earthquakes and 30 previously unused validation earthquakes on two cable routes. The archive contains derived development caches, fitted model weights, predictions, nested case scores, event-level results, frozen protocol and provenance, environment lock, data-quality tables and scientific figure assets. The source-code ZIP snapshots GitHub commit 2ab6ef0e2cc07f0334d4e54187bc2f35c186c9c8 (release v2026.10.03). The results show a dense-versus-sparse supervision benefit, but do not establish neural superiority over both non-neural baselines. This is a short-horizon covariance-processing evaluation on known earthquake archive records, not pre-rupture earthquake prediction, continuous-record detection or an operational early-warning system. Automated replay and tests do not substitute for independent human scientific review. Original source code is BSD-3-Clause. Derived data, model and figure assets are CC BY-NC-SA 4.0 with upstream Cook Inlet/UW Fiber Lab attribution retained; third-party data are not relicensed. Raw waveform HDF5, manuscripts, cover letters and submission files are excluded. Earlier historical scientific assets remain available on the preceding GitHub releases. This research received no external funding. Scientific archive SHA-256: 68751d499e45edc735db68a27fd2d7fb458aac8294965c32004190c45b8df9b0. Source archive SHA-256: 89bb5bf9773de134bd710ab4e2d0332420bc7865b30648db2d0f5a77ac497d90. Installation, cached-data replay and raw-data requirements are documented in docs/FIXED_WINDOW_VALIDATION.md. Upload packaging note: the scientific ZIP is distributed as 71 byte-preserving parts: 55 parts no larger than 16 MiB and sixteen 1-MiB pieces replacing part049, because larger browser transfers repeatedly stalled. Download all parts together with UPLOAD_PARTS_MANIFEST.json and reassemble_archive.py, then run python reassemble_archive.py. Every part and the reconstructed ZIP are SHA-256 verified; the reconstructed ZIP is identical to the GitHub scientific release. This packaging changes no research data or results.
Keyword(s):
MeSH Term(s):
Grants and Funding:
No Grants or Funding Data Available
Associated Publication(s):
No Associated Publication Data Available
Supplementary Data:
No Supplementary Data Available
Related Dataset(s):
No Related Dataset Data Available
The National Library of Medicine (NLM) provides information about biomedical datasets. Inclusion on an NLM Web page does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: NLM Web Policies (nih.gov)