Each year we wait on edge to watch Punxsutawney Phil wake up, leave his comfortably-warm home and go predict the rest of the winter season. But how accurate are these seasonal forecasts?
First, let’s take a look at the history books. Punxsutawney Phil’s forecasts began in 1887 when his first prediction was for a longer winter when he saw his shadow. According to NOAA/NCEI, that prediction was only partly true. Monthly weather summaries from that time showed the (correct prediction) Northeast and Great Lakes had colder-than-average temperature in February and March of that year. However, many other summaries showed that the (incorrect prediction) South and West had warmer weather, meaning an earlier Spring.
It goes to show that it depends where you live matters when verifying Phil’s forecasts. But for the sake of conducting research on each region, we’ll look at the Lower 48 as a whole.
When looking at the Contiguous United States, we observe the temperature departure in February and March and compare that to Phil’s forecast. If Phil doesn’t see his shadow and the temperature those two months is warmer-than-average, then he was correct! According to NOAA/NCEI between 2010 and 2019, Phil’s predictions were correct 4 out of 10 times, thus 40%. That number is in line with the entirety of Phil’s forecasts since this event began.
Groundhog Day is this Sunday, February 2nd at sunrise!