Register the 360Day Year as a Calendar Temporal Reference System

[chris-little]

Many climate simulations and predictions used a simplified calendar of 360 days per year. This was to make calculations easier, in particular, the calculation of monthly or seasonal average properties, as all months, including "February", had 30 days, and most seasons had 90 days. Many of these simulations are valuable archived inforamtion and there is no concise formal method of indicating this provenance and usage metadata. This issue was raised recently by NASA in an OGC Meeting.

There is no simple mapping from simulation days to real days. The missing days, and fractions of a day, could be added "at the end" or distributed evenly throughout the simulation year, or distributed unevenly. Actually, preferred practice is not to map at the level of a day, but only at the level of a season, or perhaps month.

Some newer climate models may have adopted the 365Day Year, without the complications of leap days.

In the simulations, various constants, such as solar radiation or the Coriolis affect, are adjusted to account for the different days/years. Such adjustments are relatively small compared to others terms in the equations and considered not significant. The validation of the simulations, such as predicting the past climate, which is relatively well measured, demonstrate that the day/year approximations are reasonable.

It is proposed to register both the 360Day and 365Day years as calendars in the OGC Naming Authority registers to enable sicentists to create meaningful and useful provenance and usage metadata.

In the longer term, it seems sensible to propose both calendars to ISO if and when they establish their list of calendars (ISO NP34300)

@cmheazel @ronaldtse @ghobona for your info.