Groundhog Day is a day that creates a collective sigh for some meteorologists across the country. It is a day when a groundhog predicts whether the US will see winter continue (if it sees its shadow) or an early spring (if it does not). It's a tradition from the Dutch in Pennsylvania which continues to this day. However, how accurate is the Groundhog in its predictions? Does a groundhog know the future? (no...but it's still fun to see!). How about his track record in Omaha?
I put his skills to the test, using his past predictions since 1900, we can see how many times he was accurate! What constitutes an early spring or a late winter? While the definition differs for everyone, for this I used temperature and snowfall data for March. If temperatures were above average, or snowfall below average, that is an early spring. On the contrary, if temperatures are below average, or snowfall is above average, that is a late winter. If it's either or (i.e. warm March and above-average snow or cold March and below-average snow), it will be up to my discretion. So without further ado, let's jump into it!
As a baseline, the average temperature in March when you combine the highs and lows is 40.3 degrees based on the last 30 years of data. The average snowfall for March is 5.4" based on the last 30 years of data. For this experiment, I will be using the average at the time.
HOW ACCURATE IS PUNXSUTAWNEY PHIL IN OMAHA?
1900-1909
Phil predicted a late winter 9/10 times, the only early spring prediction was 1902. In terms of temperature, Omaha experienced above-average March's 8/10 times. For snowfall, it was flipped where Omaha typically experienced less snowy March's. Overall, results for the first decade of the 20th century are mixed with only half the predictions being accurate. 5/10
1910-1919
You might already start to notice a pattern from Phil, he likes to predict "late winters" for the US, and this theme will be common throughout the next century. The 1910s were relatively warm for March in Omaha, with most of our years ending up above average. Snowfall was interesting, generally, March was dry for snowfall in Omaha, with a notable exception being 1912 where it was the snowiest March on record with 29.2"! (Side note: March 23, 1913, was the day of the Easter Sunday Tornado, a prediction out of Phil's wheelhouse). 2/10
1920-1929
The 1920s continued the pattern of the 1910s with warm and less snowy March's, and with Phil's continuation of late winter predictions, it does not look good for his accuracy. 1923 was a notable exception, with the 2nd snowiest month on record with 28.8" of snowfall. Otherwise, the accuracy of Phil was low for the 1920s. 3/10
1930-1939
The 1930s were a rough time for weather in Omaha as the throes of the Dust Bowl gripped the state with dry weather, however, snow still fell in March. Phil also finally did not see his shadow, an early Spring, but too bad he was incorrect...4/10
1940-1949
The 1940s are perhaps the wildest decade for little Phil. The Second World War brought the tradition to a halt for two years. The entry for 1942 read "War Clouds have blocked out part of the shadow", not sure what that meant for his prediction, so 1942 will not be considered. Phil did not even show up for 1943, meaning two years do not have data. In terms of Omaha weather, the 1940s were a dry decade with colder temperatures. Therefore, Phil did his best since the 1900s despite his absence for two years. 6/8
1950-1959
The 1950s were generally cold for Omaha, with many of our March's below average for much of the decade. This decade was also the trickiest as some had both warm & snowy or cold & dry months, this was where looking at individual data came into play and where I made a personal decision. Nevertheless, this has been Phil's best decade. 7/10
1960-1969
The 1960s were full of weather whiplash for Omaha. 1960 was the coldest March on record, while it was also one of the snowiest. In terms of predictions by Phil, his track record has certainly improved over the last few decades. 6/10
1970-1979
The 1970s in Omaha were warm, dry months of March. This was bad news for Phil, whose string of late winters ran into the unfortunate reality of the warm temperatures. However, he did predict two early springs! 2/10
1980-1989
Beginning in the 1980s a trend began in Omaha snowfall, the snow becomes less during March. In 1980, the average snowfall in Omaha was 7.1", now in 2022 the snowfall average is 3.0". This is the reason that I am using the average at the time rather than the 2022 average, as it would skew the data. As for Phil, it was a good decade for the groundhog. 7/10
1990-1999
The last decade of the 1990s was not kind to snow in Omaha, many months were incredibly dry in terms of snow. Data for Eppley Airfield is missing for 1998 in terms of snowfall, so that prediction was not counted. As for Phil, finally not seeing his shadow with 4 early Spring predictions! 5/9
2000-2009
As the 21st century dawned, the trend of increasingly snowless March continued as many seasons were below average in terms of snowfall. For Phil, his hit-and-miss streak continues even into the digital age. 4/10
2010-2019
The final decade is measured here. March 2012 was the warmest March on record, and no snow fell. This has become a trend of the last decade where hardly any snow has fallen in Omaha in March. Phil seems to flip back and forth between late winter and early spring, but he was incorrect on much of them. 4/10
FINAL RESULTS
Now that we have looked at each decade...or you scrolled down to this part, it's time to look at the final results. We looked at 117 years of data from 1900 to 2019, except 1942, 1943, and 1998. Of these 117 years, Phil got 55 of those years right.
Doing a bit of math, 55/117 = .47 or 47%.
This means that Phil got 47% of his forecasts correct, or just under half! This might seem somewhat impressive, but as we have seen Phil almost always saw his shadow meaning a late winter was almost correct. As he only chooses one of two options, there is a 50/50 chance he will get it correct. So, the simple probability would put the number close to 50% as it has, so does Phil have predictive skills? Probably not.
Now Unadilla Bill on the other hand...that rodent knows how to do a forecast. We meteorologists, however, do not just flip a coin, we can make forecasts and predict specific weather. So trust your local meteorologists!
