seisgo.stacking

The stacking module provides a unified interface to multiple waveform stacking algorithms, from simple linear averaging to advanced phase-weighted and cluster-based methods.

Function Summary

Stacking Methods Reference

All methods are accessible via the unified seisgo.stacking.stack() wrapper.

Method name	Description	Key parameter(s)
`linear`	Arithmetic mean along the stack axis.	`axis`
`pws`	Phase-weighted stack. Applies a phase coherence weight in the time domain.	`p` (phase power, default 2)
`tf-pws`	Time-frequency phase-weighted stack using the Stockwell transform.	`p`
`tfpws-dost`	TF-PWS using the discrete orthonormal Stockwell transform (DOST).	`p`
`robust`	Iterative robust stack (Pavlis & Vernon 2010). Downweights incoherent traces.	`epsilon`, `maxstep`, `win`, `ref`
`acf`	Adaptive covariance filter (Nakata et al. 2015). Enhances coherent signal.	`g`
`nroot`	N-th root stack. Enhances coherence at the cost of amplitude fidelity.	`p`
`selective`	Correlation-based selective stack. Only includes traces above a coherence threshold.	`cc_min`, `epsilon`, `maxstep`, `win`, `ref`
`cluster`	Cluster-based stack. Groups traces by similarity and returns the dominant cluster.	`h`, `win`, `normalize`

Usage Examples

Unified wrapper

from seisgo import stacking
import numpy as np

# d: 2-D array, shape (n_traces, n_samples)
linear = stacking.stack(d, method="linear")
pws    = stacking.stack(d, method="pws",    par={"p": 2})
robust = stacking.stack(d, method="robust", par={"epsilon": 1e-5, "maxstep": 10})

Selective stack with custom reference

ref = np.median(d, axis=0)
sel = stacking.stack(d, method="selective",
                     par={"cc_min": 0.6, "ref": ref, "win": [100, 400]})

Returning stacking statistics

newstack, weights, n_iter = stacking.robust(d, stat=True)