The most common way to define the start of winter is to appeal to the solstice, the day on which the noontime sun is positioned at its farthest southern point of the year. This is the astronomical start of winter — Dec. 22 this year. On that day, the noontime sun will be directly overhead at 23.5 S latitude.
Of course, by the time Dec. 22 comes around those of us who live in southern Wisconsin have usually already experienced a reasonable dose of winter.
So, is there an alternative way to define the beginning of winter that is more in line with our real experience of the season?
It turns out that there is such an alternative and we can base it on the likelihood of falling snow, perhaps the distinguishing characteristic of winter weather.
Imagine you have 100 years of weather data for Madison. You might reasonably ask what kind of precipitation has historically fallen on Aug. 1 at Madison. In fact, it has never snowed on Aug. 1 in any of those 100 years at Madison. As we move into the fall toward the winter solstice, however, snow begins to show up in the 100 years of precipitation data with greater regularity. By the time we get to Nov. 15, precipitation that falls on that date has been equally split between rain and snow.
This would seem to be a pretty good way to determine the start of meteorological winter, since after this date the most likely form of precipitation is snow. So, though it hasn’t really felt like it yet (but it will by Wednesday), we are already in our meteorological winter season.