Bias-correct a forecast using quantile-matching on computed distributions of retrospective forecasts and observations.

gev_forecast_correct(data, obs_location, obs_scale, obs_shape,
  retro_location, retro_scale, retro_shape, extreme_cutoff,
  when_dist_undefined)

Arguments

data

matrix representing forecast values

obs_location

location parameter from observed data

obs_scale

scale parameter from observed data

obs_shape

shape parameter from observed data

retro_location

location parameter from retrospective data

retro_scale

scale parameter from retrospective data

retro_shape

shape parameter from retrospective data

extreme_cutoff

clamping value for computed forecast quantiles (1/extreme_cutoff < quantile < 1 - 1/extreme_cutoff)

when_dist_undefined

assumed quantile of forecast when retrospective distribution is undefined

Value

a matrix with a corrected forecast