Jitter a forecast to account for observation errors.

To account for observation errors in ensemble verification the forecast values can be perturbed by sampling from a specified error distribution. Jittering the forecast is only likely to have an effect on the ensemble spread and rank histograms.

Usage

jitter_fcst(.fcst, jitter_func, obs_col = NULL, ...)

Arguments

.fcst

An object of class 'harp_fcst'.

jitter_func

The function to be applied to the forecast values.

obs_col

The observations column if it is to be used in the jitter function. Can be the column name, quoted, or unquoted. If a variable it should be embraced - i.e. wrapped in {{}}.

...

Other arguments to jitter_func.

Value

An object of the same class as .fcst with jittered forecast values.

Details

Note that the jitter function should be a function that works on a vector of forecast values and it should have only one or two arguments. The first argument is for the forecast data, and if there is a second argument, that is for the observations.