THe spatiotemporal evolutions of long-period(LP), hybrid and volcanic tectonic (VT) seismicity are important for tracking the evolution of underlying volcanic proces. Here, by defining a spectral dissimilarity metric, we perform cluster analysis in seismic events in Kilauea volcano.
The data are HV.OTLD, From Jan., 2011 to Oct., 2018．The window length of each event is 10s before the p wave arrival time and 10s after S wave arriival time
The length can be shorter than this example, but sholud be more than 5 sec. The name of the events is the original time of eathquakes (The catalog time).
"data_dir": The path of sac files. "snr": 3 "win_len": 1.28 (We select the 12.8s before and after P wave arrival time to calculate the spectrum) "events_catalog": "events_catalog" (events catalog) "n_cls": 20 (The number of clusters)
You can run the test code as:
python volcano_signals_classification.py (-P) config_json
out\png : clsi.pdf (100 spectra of each clusters); mean_spectra; fre_energy; med_dis.pdf; hist.png and dendrogram.pdf
out\text: clusteri.dat (the catalog of 20 clusters); new_catalog (add FI and cluster Num. to the events_catalog)
peak_amp_size: used to plot fre_energy.pdf