Assesses the quality of decadal or monthly forecasts as done in Central Asian Hydrometeorological Agencies.

This function computes a per period (decade or month) forecast quality assessment for each forecast for the test set provided. Ideally, the user provides an out-of-sample test set.

Usage

assess_fc_qual(df, plot)

Arguments

df

tibble with `date`, `obs`, `pred` and `per` column where the `per` column is either the decade or the month specifier (normally given as factor).

plot

TRUE/FALSE if plot should be generated

Value

A list with a tibble of per period forecasts that are acceptable expressed as percentages of the total test set length and the average of these plus, optionally, the handle to the figure.

Details

`per` goes from 1 to 12 for monthly data and from 1 to 36 in decadal data.

See also

Other Post-processing: computeAnnualFlowDurationCurve(), plotAnnualFlowDurationCurve()