R/calcGapCutoff.R

In bocinsky/guedesbocinsky2018: Research Compendium for d'Alpoim Guedes and Bocinsky (2018)

#' Calculates the empirical cumulative distribution function.
#'
#' This function calculates the empirical cumulative distribution function,
#' then uses a user-defined probability level to calculate the "normal" gap length.
#' That way, GHCN records (years) with unusually long gap lengths may be discarded.
#'
#' @param rleVector the RLE vector
#' @param pLevel a p level
#'
#' @return The cutoff value
#' @export
calcGapCutoff <- function(rleVector, pLevel=0.95) {
  f.all.rles <- stats::ecdf(rleVector)
  x <- seq(0, 31, length.out = 100000)
  out <- mean(x[which(abs(f.all.rles(x) - pLevel) == min(abs(f.all.rles(x) - pLevel)))])
  cutoff <- Hmisc::ceil(out)
  return(cutoff)
}