Available trained models

gempa [9] provides many models trained on and applied to different data sets at different scales of epicentral distances, source depths and source types. They are listed in table deepc-models. Use these model names for configuration of modelWeights which exists for each sub-picker:

The corresponding DL models are installed with the dlmodels package. When installing deepc by gsm, dlmodels is installed automatically. A typical installation directory is /home/data/dlmodels.

Event classification models

Currently, gempa doesn’t provide a model for event classification.

model name

event types

applications

training data

pncx-if-3

earthquake, anthropogenic event, mining

signals from earthquakes and machinery in mining environment at epicentral distance of tens of meters up to 10 km

mine

pncx-if-10

earthquake, anthropogenic event, mining

signals from earthquakes and machinery in mining environment at epicentral distance of tens of meters up to 10 km

mine

Phase classification models

model name

specifics

applications

training data

pnc-cb-1_4

P, S, Noise

(re-) classify phase hints for picks

Combined (INSTANCE [3], SCEDC [6] and STEAD [7])

Joint time prediction and phase classification models

The Seismology Benchmark collection, SeisBench [23], has published many pre-trained picker models. gempa provides converted versions of some of these models and also in-house trained ones. These models predict probabilities for P, S, and Noise for each sample of a trace. DeepC post-processes their predictions to phase classification and onset refinement of a given trigger.

EqTransformer

The following models are trained on the SeisBench implementation of EqTransformer [22].

model name

specifics

applications

training data

eq-et

60 s, 100 Hz

regional data

ETHZ [1]

eq-ge

60 s, 100 Hz

teleseismic data

GEOFON [2]

eq-in

60 s, 100 Hz

regional data

INSTANCE [3]

eq-iq

60 s, 100 Hz

regional data

Iquique [4]

eq-ne

60 s, 100 Hz

regional data

NEIC [5]

eq-or

60 s, 100 Hz

regional data

original EqTransformer weights [22]

eq-sc

60 s, 100 Hz

regional data

SCEDC [6]

eq-st

60 s, 100 Hz

regional data

STEAD [7]

PhaseNet

The following models are trained on the SeisBench adaptions of PhaseNet [28].

model name

specifics

applications

training data

pn-et

30.01 s, 100 Hz

regional data

ETHZ [1]

pn-ge

30.01 s, 100 Hz

teleseismic data

GEOFON [2]

pn-in

30.01 s, 100 Hz

regional data

INSTANCE [3]

pn-iq

30.01 s, 100 Hz

regional data

Iquique [4]

pn-ne

30.01 s, 100 Hz

regional data

NEIC [5]

pn-or

30.01 s, 100 Hz

regional data

original PhaseNet weights [28]

pn-sc

30.01 s, 100 Hz

regional data

SCEDC [6]

pn-st

30.01 s, 100 Hz

regional data

STEAD [7]

pn-in_vl-74

30.01 s, 100 Hz

regional data

INSTANCE [3], Vogtland

pn-in_vl-97

30.01 s, 100 Hz, all-comps-normalized

regional data

INSTANCE [3], Vogtland

pn-vl-91

30.01 s, 100 Hz, all-comps-normalized

regional data

Vogtland

pn-vp

30.01 s, 100 Hz

volcanic data

VCSEIS [8]

pnv-bq-0

60 s, 100 Hz

volcanic data?

BQCombined

pnv-bq-10

30 s, 100 Hz

volcanic data?

BQCombined

pnv-il_vc-2

15 s, 100 Hz

volcanic data?

Iceland, VCSEIS_hawaii [8]

pnv_15s-et-0

15 s, 100 Hz

regional

ETHZ [1]

pnv_15s-il-2

15 s, 100 Hz

regional

Iceland

pnv_30s-et-sb

30 s, 100 Hz

regional

Iceland

pnv_60s-et-0

60 s, 100 Hz

regional

Iceland