Alternative approach to dividing seasons

All,

This talk of setting the seasons to align with mean temperatures
recalls to mind something I've thought of before, particularly for
Texas. Here in Texas, I think it would be possible to divide the
seasons thus:

Summer: hot, low standard deviation
Winter: cold, low standard deviation
Spring: period from winter changing to summer, high standard deviation
Fall: period from summer changing to winter, high standard deviation

I don't have the statistics at hand to quantify these notions, but
it's something that would be interesting to explore.

Victor