[ToDo] Error propagation

[bnubald]

Using the sea ice forecast mean and standard deviation, computing the error propagation would be of use.

Since metrics include thresholding, more complex non-linear approaches are likely suited, such as Monte Carlo simulations, but outside my scope for this work currently.