SpecsVerification (version 0.4-1)

EnsCrps: Continuously ranked probability score (CRPS) for ensemble forecasts.

Description

Calculate the continuously ranked probability score (CRPS) for ensemble forecasts and their corresponding observations. The forecast distribution is approximated by a step function with equal weights on the ensemble members.

Usage

EnsCrps(ens, obs)

Arguments

ens

N*K matrix. N cases of a K-member ensemble forecast.

obs

vector of length N. The corresponding verifying observations.

Value

The function returns a vector of length N with the individual score values of the N ensemble forecasts.

References

Hersbach, H (2000). Decomposition of the Continuous Ranked Probability Score for Ensemble Prediction Systems. Weather and Forecasting. DOI: 10.1175/1520-0434(2000)015<0559:dotcrp>2.0.CO;2

Examples

  # Example:
  ens <- matrix(rnorm(500),100,5)
  obs <- rnorm(100)
  mean(EnsCrps(ens, obs))