LISFLOOD routine in reservoirs with small degree of regulation

[casadoj]

The LISFLOOD routine causes strange reservoir behaviour in some cases where the degree of regulation is small.

In the case below (grand_id=42), the average inflow is 112 m3/s, which accounts for 9.7 hm3 of water per day.The calibration stablished a flood and normal-adjusted flood storage of 105.8 and 97.3 hm3, i.e., a flood zone of 8.5 hm3. The fact that the routine updates storage with the current inflow at the beginning of each time step causes that the computation of the outflow often supposes that the reservoir is in a flood simulation. When at the end of the timestep outflow is deducted from storage, the reservoir comes back to a normal situation. That's what causes that "backward" line in the results.

This effect happens in reservoirs where the inflow accounts for a good deal of the reservoir storage, i.e., reservoirs with low degree of regulation. For instance, the reservoir in the plot below has a degree of regulation of 0.042 according to GRanD. For reference, Shrestha et al. (2024) discards reservoirs with a degree of regulation smaller than 0.08.

[casadoj]

I have added a condition on the degree of regulation to the selection of reservoirs to be calibrated. See issue #12.