Issue with max_hours Calculation for Hydropower Plants in Cascaded Systems

When setting the flag hydro_max_hours: "energy_capacity_totals_by_country" the residual volume (e_missing) is used to scale the reservoirs that don't have max_hours defined. In code here. For Norway, this leads to max_hours being more than 8760 h for some hydropower plants. The country's total storage capacity then becomes correct, but some of the plants will have much larger storage capacity than actual.

This issue is connected to how duration is computed for JRC data in powerplantmatching (link). Where the column StorageCapacity_MWhis used, but I think it makes more sense to use avg_annual_generation_GWh instead as there could be more reservoirs upstream. This, does, however, lead to the total storage capacity being significantly higher for most countries (water can be used multiple times in cascaded systems):

The problem as I see it is that most Norwegian hydro reservoirs are in cascade, see this for instance, but PyPSA-Eur considers only a single reservoir connected to a power station.

I'd appreciate a discussion on this issue.

Start with the max_hours handling in scripts/add_electricity.py around line 517, then read the JRC duration calculation in powerplantmatching/data.py around line 439. Compare StorageCapacity_MWh with avg_annual_generation_GWh for cascaded Norwegian hydropower and establish which interpretation should be used. Done requires an agreed resolution for reservoir representation and its effect on country totals and plant-level max_hours.