Ability to handle data series with unequal number of data points
Currently, the code expects both reference and modeled data to contain the same number of data points during the reference period. An improvement would be to make the code more flexible because the number of data points could differ because of missing data, different time series frequencies, multiple model members etc.