This calculates the observed Ar and delta Phi (amplitude ratio and phase shift) using all training dates, not only those from the PRMS simulation. (which greatly increases the number of observations) This requires the assumption that the air temperature signal properties (phase and amplitude) from the PRMS period (WY 1981 - 2016) apply to the later training period (WY 2017-2020). For now we're assuming they are constant during the PRMS period, so it seems reasonable to also assume that they don't change much during the next couple years, but this could be something to look into in the future.
This calculates the observed Ar and delta Phi (amplitude ratio and phase shift) using all training dates, not only those from the PRMS simulation. (which greatly increases the number of observations) This requires the assumption that the air temperature signal properties (phase and amplitude) from the PRMS period (WY 1981 - 2016) apply to the later training period (WY 2017-2020). For now we're assuming they are constant during the PRMS period, so it seems reasonable to also assume that they don't change much during the next couple years, but this could be something to look into in the future.