Currently the parameters for the reference dataframe are taken from the first 4 years of observation.
It might be more consistent to create reference parameters by setting the GMT-anomaly gmt_scaled = 0and use the trained pymc3 model to sample the corresponding parameters
Currently the parameters for the reference dataframe are taken from the first 4 years of observation. It might be more consistent to create reference parameters by setting the GMT-anomaly
gmt_scaled = 0
and use the trained pymc3 model to sample the corresponding parameters