Better handle variables with no data or all 0s

[guidocioni]

For some variables, like snow, it may be beneficial to avoid drawing/plotting/processing when they do not contain any information.
For example, if snow is less than a certain value (0.1) for all members/models, we can avoid drawing the trace.

• Do not draw snow trace if it is 0 everywhere in Ensemble plot
• Do not draw precipitation trace if it is 0 everywhere in Ensemble plot -> see what happens in case that there is no snow or precipitation forecast for the entire period
• Do not draw snow trace if it is 0 everywhere in Deterministic plot, also suppress the secondary axis labels creations!
• Do not draw precipitation trace if it is 0 everywhere in Deterministic plot -> see what happens in case that there is no snow or precipitation forecast for the entire period