Using Copernicus reanalysis data

[j-atkins]

After recent discussions, and based on the requirements of different courses at UU, now is a sensible time to move to using Copernicus Reanalysis products, rather than purely the Analysis & Forecast products which are used at present. This will mean that expeditions can be run back to 1993 and up to present day (defaulting to the analysis product when the scheduled time is within ~2 months of present day or 2 weeks into the future).

The consideration we will have to make here is that the reanalysis products are only at daily resolution, whereas the analysis product (physical variables at least) are available at (6) hourly.

My concern is just in case users ever want to use VirtualShip for testing intra-daily physical phenomena (note, BGC is daily in all products), diurnal cycles etc.

With that in mind I suppose we either make clear that we don't support intra-day functionality if we use the reanalysis products throughout (this is probably the simplest option but removes functionality), or indeed we keep the analysis product for as many years as possible and then state somewhere that intra-day physical measurements are only possible in the near-present (2022 onwards).

Keen to hear your thoughts on the issue @ammedd and @erikvansebille!

[erikvansebille]

I would keep it as simple as possible for now, and focus on support for the reanalysis products. I'd say that's more then sufficient for educational purposes

I don't think it's a major issue to not only daily data. But I can see use cases where users would want to use the forecasts (think "dads who rescue seal pups"); so at a later point we might provide an option to choose the forecast mode too?

Since the field names are (almost?) the same, swapping the dataset from reanalysis to forecast should be as simple(?) as changing the dataset_id in the copernusmarine tool? And hopefully, this will be easy with the Parcels v4 Fieldset.from_copernicusmarine() method too? Or am I missing something?