Track a smooth sedimentation-rate ridge through an eCOCO score map

Within-window min-max normalised score, quadratic jump penalty (penalty 0.35, jump scale max(2, 4*sr step)), Viterbi back-tracking (trackEcocoRidge in Acycle ecoco.m).

Usage

ecoco_track_ridge(
  score,
  sr,
  depth,
  rho,
  p,
  norbit,
  fallback = NULL,
  jump_penalty = 0.35
)

Arguments

score

Score map through which the ridge is tracked (numeric matrix, rates in rows, windows in columns), e.g. element score of the output of ecoco_adaptive; it is min-max normalised within each window before tracking. Required, no default.

sr

Sedimentation-rate grid of the score map, in cm/kyr (numeric vector, one value per row of score). Required, no default.

depth

Depths of the window centres, in metres (numeric vector, one value per column of score). Required, no default.

rho

Correlation coefficients (numeric matrix of the same size as score); the values along the ridge are returned. Required, no default.

p

Global Monte Carlo p-values (numeric matrix of the same size as score); the values along the ridge are returned. Required, no default.

norbit

Number of contributing orbital periods (numeric matrix of the same size as score); the values along the ridge are returned. Required, no default.

fallback

Result returned for windows the ridge does not reach, the output of ecoco_best_per_window (data frame); NULL computes it from the other arguments. Default=NULL.

jump_penalty

Weight of the quadratic penalty on rate jumps between neighbouring windows (numeric, >= 0): each step costs jump_penalty * (|sr jump| / max(2, 4 * sr step))^2. Larger values give a smoother ridge, 0 picks the best rate in every window independently. Default=0.35.

Value

data.frame with depth, sr, rho, p_global, norbit, score, sr_low, sr_high