Option to crop uncertainty intervals at zero for RC and Q time series?

[benRenard]

BaRatinAGE generates total uncertainty intervals that can go negative, due to the structural component. Several users have suggested that such intervals should be cropped at zero. I believe this should not be the default behavior, because this simply results from BaRatin assumptions, and I'd rather show the direct result of these assumptions. But we could provide an option to do that if the user decides so. Questions:

1. Should we do it, or rather let the user take care of that by post-processing BaRatinAGE exports?
2. If we implement the option, should it be a general option that always remains activated, or an option that should be activated for each case study?

@IvanHeriver, @JeromeLeCoz, what do you think?

[JeromeLeCoz]

1. I'm not sure, but it may be useful at stations with precise controls for discharges going to zero. Not the default behavior anyway, agreed. If it's easy to implement, why not (that would save arguments with users)
2. For each case study, I'd say. As you said earlier, I also believe it is a useful warning when it happens (except for very small flows).