Skip to contents

Polyphase stress inversion via spectral clustering on \(S^5\) (Stage 5).

Source: R/stress_inversion_wissi.R
slip_inversion_wissi_polyphase.Rd

Identifies k stress phases automatically using the eigengap heuristic, then runs slip_inversion_wissi() on each phase subset.

Usage

slip_inversion_wissi_polyphase(
  x,
  weights = NULL,
  k_max = 4L,
  sigma_K_deg = 30,
  seed = NULL,
  ...
)

Arguments

x

"Fault" object where the rows are the observations, and the columns the coordinates. Object must be complete, i.e. no NA values. For Michael's, Angelier's, and Yamaji-Sato's methods, at least 4 rows of fault measurements are required, while Hansen's method requires at least 7.

weights

Optional length-N numeric vector of non-negative data quality weights. NA values are replaced with the observed mean before scaling (neutral imputation). The vector is normalised internally so that mean(weights) = 1, meaning only the ratios between weights matter. Use scale_weights() to construct weights from field quality ranks, measurement errors, and prior RUP values. Default: uniform weights.

k_max

Maximum number of phases to consider. Default 4.

sigma_K_deg

Affinity bandwidth in degrees of angular stress distance. Faults within this distance are considered similar. Default 30. Increase to merge nearby phases, decrease to split them.

seed

Optional RNG seed for k-means reproducibility.

...

Additional arguments passed to slip_inversion_wissi() for each phase.

Value

A named list with:

assignment

integer vector of length N (phase label 1..k per fault)

k_opt

number of phases identified

gaps

Laplacian eigenvalue gaps (eigengap criterion)

phase_results

list of k slip_inversion_wissi() results, one per phase

D_mat

N x N pairwise ASD matrix between fault poles (degrees)

See also

Other wissi: slip_inversion_wissi(), slip_inversion_wissi_boot()

Examples

res <- slip_inversion_wissi_polyphase(angelier1990$KAM, sigma_K_deg = 15)

# check amount of clusters detected
res$k_opt
#> [1] 3


# plot the results in lambert projection:
cols <- assign_col_d(seq_len(res$k_opt))#' 
par(mfrow = c(1, 2))

stereoplot()
angelier(angelier1990$KAM, col = assign_col_d(res$assignment))
legend('topleft', title = 'Phases', fill = cols, legend = 1:3)


stereoplot()
angelier(angelier1990$KAM, col = 'grey')
for(k in seq_len(res$k_opt)){
   res_k <- res$phase_results[[k]]
   points(res_k$principal_axes, col = cols[k], pch = 16:18, cex = 2) 
 }
legend('topright', title = "Principal axes", pch = 16:18, col = 'black', 
   legend = c('S1', 'S2', 'S3'))

 
 dev.off()
#> null device 
#>           1