You can model and produce forecasts with transformed data using the fable package in R very easily. The latest package has added the feature to extract either a median or mean point forecast output using the point_forecast() specification (in the previous package you set bias_adjust = F.)
My question is: How should I properly extract a variance or standard deviation from the forecast distribution if I am selecting the median point forecast or asking the bias adjustment to not occur? I am using median point forecast values with their associated standard deviation for different analyses.