CRPS {s2dv}

Compute the Continuous Ranked Probability Score

Description

The Continuous Ranked Probability Score (CRPS; Wilks, 2011) is the continuous version of the Ranked Probability Score (RPS; Wilks, 2011). It is a skill metric to evaluate the full distribution of probabilistic forecasts. It has a negative orientation (i.e., the higher-quality forecast the smaller CRPS) and it rewards the forecast that has probability concentration around the observed value. In case of a deterministic forecast, the CRPS is reduced to the mean absolute error. It has the same units as the data. The function is based on enscrps_cpp from SpecsVerification. If there is more than one dataset, CRPS will be computed for each pair of exp and obs data.

Usage

CRPS(
  exp,
  obs,
  time_dim = "sdate",
  memb_dim = "member",
  dat_dim = NULL,
  Fair = FALSE,
  ncores = NULL
)

Arguments

exp

A named numerical array of the forecast with at least time dimension.

obs

A named numerical array of the observation with at least time dimension. The dimensions must be the same as 'exp' except 'memb_dim' and 'dat_dim'.

time_dim

A character string indicating the name of the time dimension. The default value is 'sdate'.

memb_dim

A character string indicating the name of the member dimension to compute the probabilities of the forecast. The default value is 'member'.

dat_dim

A character string indicating the name of dataset dimension. The length of this dimension can be different between 'exp' and 'obs'. The default value is NULL.

Fair

A logical indicating whether to compute the FairCRPS (the potential CRPS that the forecast would have with an infinite ensemble size). The default value is FALSE.

ncores

An integer indicating the number of cores to use for parallel computation. The default value is NULL.

Value

A numerical array of CRPS with dimensions c(nexp, nobs, the rest dimensions of 'exp' except 'time_dim' and 'memb_dim' dimensions). nexp is the number of experiment (i.e., dat_dim in exp), and nobs is the number of observation (i.e., dat_dim in obs). If dat_dim is NULL, nexp and nobs are omitted.

References

Wilks, 2011; https://doi.org/10.1016/B978-0-12-385022-5.00008-7

Examples

exp <- array(rnorm(1000), dim = c(lat = 3, lon = 2, member = 10, sdate = 50))
obs <- array(rnorm(1000), dim = c(lat = 3, lon = 2, sdate = 50))
res <- CRPS(exp = exp, obs = obs)