Rodent Forecasters

It was a typical early February day last week Monday morning. Cloud cover obscured the sun as it came over the horizon. At this time, the city of Sun Prairie pulled out of hibernation a groundhog. The purpose of the ground hog was for it to make its annual forecast for when winter will end. Jimmy did not see his shadow, which legend portends a shorter winter season. Several years ago Jimmy, who has been at this since 1969 (albeit different animals), experienced 15 minutes of fame when the rodent bit the Mayor's ear. This fame was sandwiched between  the fame the city received when Christopher Hovel last played football at Sun Prairie High and the notorious July 2018 gas pipe explosion that took out part of the downtown. Jimmy may even have had more press that year, due to the ear bite, than Punxsutawney Phil, the oldest rodent winter weather predictor in the nation, which was started in 1887. Every year it seems Phil gets the press. Phil had a different prediction than Jimmy in the binary selection process.

Not sure who this guy is, or where he got the 30% for Phil
Oh, what AI can do!

At the same time, there seems to be a contest of who is the most accurate predictor of an early spring, as compared to six more weeks of Winter. The web site for Jimmy does not give his percent success rate, but reads: "Jimmy has an extremely high accuracy rate on his annual predictions, making him an outstanding leader in meteorological circles." So, I wondered, what metrics are used to determine if spring arrived early. Many negativists say it really does not matter, since by the calendar spring arrives in about six weeks anyway. This year, calendar spring will arrive on March 20. Some years it is March 21. This year it is six weeks and four days past Groundhog Day. Perhaps sometime I will delve into why it is Feb 2, and not a couple days later putting it right at the halfway point through the calendar winter. Meteorologically, winter is said to be Dec, Jan and Feb. Hence, March, April and May make spring.

Phil's recent scorecard, which got me wondering how this was calculated

Regardless of calendar or meteorological springs, as people we generally look to certain signs of spring. Flowers, snow has melted, temperatures warming. It is hard to define, but we seem to know it when it comes early. Same thing for winter, which sometimes starts, as this past year, right after Thanksgiving. In 2019 we had several inches of snow at the end of October (with other early storms)  followed by very cold weather for many weeks; it then warmed before Christmas and the snow had mostly melted. The ground froze, crops remained in the field and the sewer district had months of sludge that had not yet been worked into the ground. The district had to find former farm storage tanks to store it until spring. Mother nature sure knows how to throw curve balls. As a shoulder months, March and April can be very unpredictable. Although seldom is weather consistent with you would expect. A few years ago camping at Cunard Lake I talked a man who was there camping, and he had come up for the opening of fishing season on May 1, only to find that the lakes were all frozen and covered with several inches of snow. Here it was less than two months later and I was swimming in that lake. The latest known frost date in Boulder Junction is June 30, which probably explains why their claim to fame is muskies and not groundhogs. I would think ice and snow on lakes counts as a late spring. Does frost on the last day of June count as a late spring? The day my youngest son was born (Apr 29) we went from unseasonable warm, the wife who was very pregnant that year would say hot, to having seven inches of snow on his date of birth. Luckily, the warm pavement and ground allowed much to melt quickly. Without metrics, how does on acknowledge if the rodent is correct?

Jimmy about to pounce on Mayor's ear
Apparently, it needs a muzzle

My wife thought I was being too analytic in attempting to figure out the metrics used for the rodent handlers to say if they were correct or not. Her point was this is all in jest.  But, when they start giving out stats, it is not just jest. I get it, they are trying to prove who has the best prognosticating rodent. I look at my desire for a sophisticated metric to be very Germanic, as who else but the Germans would desire detailed calculus of how the groundhogs perform? Germans demand precision, just like in that former car commercial. 

NOAA top ten, see link in paragraph below for the whole ranking

As items of rodent prognostication showed up on my Facebook page I realized two main points. First, there are many more rodent prognosticators present than I thought, and, two, that the National Atmospheric and Oceanic Administration (NOAA) has rated the 18 groundhogs and one turtle. To meet their criteria, they had to be doing this at least 20 years as of 2 Feb 2024. What they found was rather striking, with success varying from 85% to 30%, with Phil and one other woodchuck at the low end for groundhogs. The outlier was the poor turtle who had a lowly success rate of 20%. Jimmy came in fifth at 65%. The problem is that they do not say what metric they used to calculate. Was it based on how much a location was above or below average, or did it use some level of standard deviation from the norm? After all, there is variability in weather. As Josh and Jase (the Brits who visited Michigan this past January) have said, in the Midwest all you have to do is turnaround and the weather may have changed. It must have been a down week at NOAA for them to do find the weather and do the calculations on groundhog forecast proficiency. Perhaps Donald Trump saw this and it led to his agency cuts. Yet, with such a low rate of success, should Phil be the determining factor for early verse non-early spring?  Most news outlets seem to think so. Phil predicted a long winter, while Jimmy has predicted an early spring. Now, weather can vary by region in the US, so who is to say Phil is correct for Wisconsin, much less California?

The 18 groundhogs are in 8 different states, with Pennsylvania leading the way with four, followed by NY with three. Unfortunately, a Friendly Illinois Brethren rodent tied with Jimmy for fifth place. Communities desire some claim to fame, and sad to say Sun Prairie had the chance at other aspects, but chose a groundhog. A bit quirky, and I guess it beats Fitchburg's Recycling Capital of Wisconsin, which no longer applies since most of Dane County all recycle the same stuff (although McFarland, through a local club here recycles stretchable plastic, with bins at the library). Actually, Pennsylvania and New York make sense for Groundhog Day did not start with Bill Murray, but with German, and more specifically Dutch Germans in the 17th and 18th centuries. In the old country a hedgehog was used. It is an outgrowth of Candlemas where, according to lore, if an animal like a badger or hedgehog, saw its shadow there would be another winter. Lacking hedgehogs, and probably finding badgers a bit unruly, the Pennsylvania Dutch-Germans used the groundhog, with the first observance by Phil in 1887. For those not aware, Wisconsin, the Badger State, did not get the moniker from actual badgers, but from lead miners in Southwest Wisconsin who were referred to as Badgers because they lived in caves in the 1820's to 1830's as temporary shelter, and of course dug for lead. Actually, this whole thing of groundhog prognosticators involving German and Dutch-Germans is rather hard to believe for such  heritages are not known for frivolity. Germans, I am sure would be pleased to have measuring metrics as it goes against the whole mirthfulness of the occasion. Being measurable it brings precision and math to an event that is otherwise very questionable, particularly in Punxsutawney. Looking up the stereotype of Germans as precision oriented, I found even the business school in Munich recognizes such. 

This has me thinking there is a market for a prognosticator on the flip side, as we move from autumn to winter. at around Halloween time, for perhaps a chipmunk or a opossum to let us know when winter really will begin. The opossum may well play dead and hence fail to give a forecast when most needed. Is a sign of spring when chipmunks come out from their hibernation behind my rock walls?

I had thoughts of developing a metric, but thought it could get out of control, so now I am willing to let NOAA, although not sure of their metric, to be the judge over what groundhog is the best forecaster. My wife can now rest easier knowing that a higher authority than me has already developed a metric to rate the accuracy of the rodent forecasters.