Climate Service users find the dates provided in "pre-industrial" simulations confusing. In these simulations there is a discrepancy between the model time and the "validation time" (i.e. the calendar date in the real world which the model simulation is aligned with). The validation time is something like a continuously repeating 1850. In CF, standard_name = "time" is generally interpreted as validation time .. but changing the metadata of piControl simulations such that time can be safely interpreted as validation time would be disruptive. There is currently no standard name which would adequately describe the elapsed model time in these simulation. The forecast_period name is specifically defined for NWP as a difference between the start of the forecast and the validation time: it does not allow for the de-coupling between model and real time that happens when doing BC forced simulations.
Climate Service users find the dates provided in "pre-industrial" simulations confusing. In these simulations there is a discrepancy between the model time and the "validation time" (i.e. the calendar date in the real world which the model simulation is aligned with). The validation time is something like a continuously repeating 1850. In CF,
standard_name = "time"is generally interpreted as validation time .. but changing the metadata ofpiControlsimulations such thattimecan be safely interpreted as validation time would be disruptive. There is currently no standard name which would adequately describe the elapsed model time in these simulation. Theforecast_periodname is specifically defined for NWP as a difference between the start of the forecast and the validation time: it does not allow for the de-coupling between model and real time that happens when doing BC forced simulations.