Then why yellow line doesn't touch time axis? Function cannot always be bigger than its own mean. If there is point above mean, than there should be at least one point below mean. I'm assuming here mean is of temerature in that year.
Standard deviation is square root of sums of squared deviations divided by number of samples. Only complex numbers can result in negative values when squared. Negative amout of samples makes even less sense.
Deviation from mean is x - μ, standard deviation is this abomination:
Sounds mutually exclusive