COLUMN: I’m afraid I can’t predict that, Dave: The fallacy of winter weather forecasts

By Tyler Friederich

Anywhere from six to twelve inches of snow was predicted for the Champaign area on Thursday night and Friday morning. Warnings of a severe winter storm swept across television screens as meteorologists displayed various computer models and showy graphics highlighting what should have been a treacherous snow storm. Most of the Champaign-Urbana area actually received less than one inch of snow on top of about a half inch of freezing rain. What gives?

My friend, who lives in my hometown of Belleville, Ill., near St. Louis, and I were discussing the impending winter storm Wednesday night and yesterday afternoon. We remained skeptical, and we both predicted that neither of us would receive the enormous amounts of snow that were predicted for both of our areas. Various weather sources, such as the Weather Channel and St. Louis media outlets, initially predicted up to a foot of snow in the St. Louis area. Similar to our situation, that snow never materialized.

With broad advances in computer technology, predicting the weather has become more of an art of analysis of computer models instead of a scientific art using statistics. Instead of relying on intuition, logic and independent thought in relation to weather statistics, weather forecasters have shifted to the indelible reliance on computer models.

While I admit that weather forecasts have improved dramatically in the past fifteen years, as indicated by the fact that forecasts for the next ten days are just as good now as forecasts for the next three days were just 20 years ago, computers are gradually replacing the need for independent analysis (or so it seems).

The 1996-97 National Collegiate Weather Forecasting Contest pitted 737 weather forecasters, of all ranges of expertise, against each other in a bid to correctly forecast the weather in various cities across 26 months. An automated computer model forecasting system based at Penn State beat 90 percent of all professional forecasters. Although advanced computer systems are integral to the forecasting process, one can’t help but wonder whether the art of using one’s own intuition has been lost.

Get The Daily Illini in your inbox!

  • Catch the latest on University of Illinois news, sports, and more. Delivered every weekday.
  • Stay up to date on all things Illini sports. Delivered every Monday.
Thank you for subscribing!

While I do not hold any expertise in meteorology, up until about four years ago I had planned on becoming a meteorologist. I became fascinated with the weather at a very young age, and ever since understanding it has become a hobby of mine. I probably spent about four hours on Wednesday and another five on Thursday anxiously sitting in front of my computer analyzing several weather forecasts, radars, and maps in anticipation of the winter storm. Yes, as my friend Marc indicated, I’m a nerd. On Wednesday night, my mom called me from home in Collinsville, near St. Louis, asking what I thought about the forecast. I confidently told her that they would definitely receive plenty of freezing rain and sleet, but not much snow at all. My friend came to the same conclusion. We were both right, and just about every meteorologist was wrong. But why?

Again, I am no expert, but it is my contention that the meteorologists, as crazy as it seems, are somewhat intimidated by the notion of using their own intuition when it comes to forecasting the weather, especially when it comes to winter storms, undoubtedly the most difficult type of weather situation to forecast. If you pay attention to their forecasts on television, you will often hear the words “the computer models indicate.”, or “I just received the computer models, and.”, but you will usually never hear “despite what the computer models say, I think.”, and so on.

Of course, the public rarely applaud meteorologists when they get the forecast correct, which is the vast majority of the time. Perhaps the perception that meteorologists “always get it wrong” stems from the fact that we only notice when they do. Nonetheless, the fact that two college students with negligible experience necessary to forecast the weather got this one right, then something must be wrong – computer models do not have the ability to reason, but we do.