COCO analysis (Li et al. 2018) — coco

Correlation coefficient between the (AR(1)-corrected) data periodogram and an astronomical target over a grid of sedimentation rates, with the H0 (no orbital forcing) Monte Carlo test of Li et al. (2018). Translated from Acycle corrcoefslices.m / cyclecorr.m (Mingsong Li).

Usage

coco_classic(
  data,
  target,
  orbit = NULL,
  sr = seq(0.2, 20, by = 0.2),
  dt = NULL,
  pad = 5000,
  red = 1,
  nsim = 0,
  slices = 1,
  adjust = 0,
  method = c("Pearson", "Spearman"),
  seed = NULL,
  keep_null = FALSE,
  verbose = TRUE
)

Arguments

data: Record to analyse (matrix or data frame; the first two columns are used): depth in metres in column 1 and the proxy in column 2, evenly spaced (it is not interpolated; incomplete rows are dropped and the rows sorted by depth). Required, no default.
target: Astronomical target periodogram (two-column matrix or data frame): frequency in cycles/kyr in column 1 and power in column 2, e.g. the output of coco_target_from_periods, coco_target_from_series. Its peaks higher than half the mean power are rebuilt as Gaussian peaks with the Rayleigh width of the data at every tested rate. Required, no default.
orbit: Astronomical periods in kyr (numeric vector), used to count how many astronomical parameters contribute at each rate (and, with adjust=1, to define the target bands; then seven periods are needed). NULL uses the periods stored with the target (attribute "orbit", set by coco_target_from_periods), or else the seven periods of Li et al. (2018) (400, 120.5, 93.1, 38.6, 23.0, 21.8 and 18.5 kyr). Default=NULL.
sr: Sedimentation rates to test, in cm/kyr (numeric vector, > 0). Default=seq(0.2, 20, by = 0.2).
dt: Sampling interval of data, in metres (numeric, > 0); NULL uses the median spacing of the depth column. Default=NULL.
pad: Length (number of points) to which the data (or each slice) is zero-padded for the periodogram (integer, at least the number of samples); a larger value gives a finer frequency grid. Default=5000.
red: Red-noise treatment of the data periodogram (integer): 0 none; 1 the AR(1) background is subtracted (as in Li et al. 2018); 2 the periodogram is divided by the AR(1) background minus 1; 3 a robust AR(1) background (median-smoothed spectrum) is subtracted; a value from 50 to 99 divides by that AR(1) confidence level (in %) minus 1, so only power above it remains. Negative values are set to zero. Default=1.
nsim: Number of Monte Carlo simulations of the null hypothesis of no astronomical forcing (random spectra with as many peaks as the data spectrum), used for the p-value p_h0 and pCOCO (integer, >= 0; 2000 in Li et al. 2018); 0 skips them. Default=0.
slices: Number of equal-depth slices whose periodograms are averaged (integer, >= 1); 1 uses the whole record. Default=1.
adjust: Adjust the power of the target in four bands (long eccentricity, short eccentricity, obliquity, precession) to the power of the data in those bands at each rate (integer): 0 no, 1 yes (only for a seven-period target). Default=0.
method: Correlation coefficient between the data and target spectra (character): "Pearson" (as in Li et al. 2018) or "Spearman". Default=c("Pearson", "Spearman"), i.e. "Pearson".
seed: Seed of the random number generator (integer), so that the Monte Carlo results can be reproduced; NULL leaves the current random state unchanged. Default=NULL.
keep_null: Return the Monte Carlo null correlations (matrix, rates in rows, simulations in columns) as element null of the output (logical, TRUE or FALSE). Default=FALSE.
verbose: Print progress messages of the Monte Carlo simulations (logical, TRUE or FALSE). Default=TRUE.

Value

list with data.frame `result` (sr, rho, p_param, p_h0, norbit, pcoco)