Ministry of Controversy, Unusually Clever People

Calling long range forecasting on its BS

weatherHe had me at “silly on weather.” A retired, Buffalo-based, 30-year broadcaster and longtime meteorologist,  Don Paul still contributes occasionally to the local paper with smart articles like this one.

Reading them is so much more interesting than looking at some guy (or gal) standing in front of an animated map. There simply isn’t time to say anything in the couple minutes they allot for weather during the average half hour of local news—which isn’t really a half hour and you don’t get much news in it either. That’s why Walter Cronkite always wanted to end every broadcast with “for more details, see your local newspaper.”

I liked Paul’s recent column on the silliness of long range forecasting because I think we’ve all been seduced by the idea that we can look into this particular crystal ball. That must be why people buy and believe the predictions is such publications as the Farmers’ Almanac and the Old Farmer’s Almanac, on which Paul has this to say:

Both issue essentially worthless detailed forecasts for a year out based on either nothing at all or junk science. Some people take them seriously, but most don’t because they know better. Or, if they don’t, they SHOULD know better. There actually have been a number of statistical studies done on their forecasts proving their worthlessness, which was no surprise to meteorologists. Studying silly forecasts is probably grant money that could have been better spent elsewhere.

And then there are the month-long forecasts you can find elsewhere (he does not name them):

The concern is these few forecasting firms, a couple of which are very large and filled with well-educated meteorologists, have managers who KNOW better than to issue 45-day or 30-day daily forecasts. But these executives also know a lot of customers are willing to believe such forecasts are viable, coming as they do from well-known and seemingly respectable firms.

Personally, I try not to take anything too seriously beyond 2 days out, which is actually more conservative than Paul gets. Recently, we had a two-day art festival here and all the local forecasts insisted—through the early morning hours—that the first day would be at least 80% likelihood of rain. There wasn’t a cloud in the sky all that gorgeous day, but many stayed home and I’m some vendors lost money. Them’s the breaks.

Gardeners, and many others, spend much of their days with their eyes on the sky, which seems to be as good a way as any to assess the weather. But I’m glad to have some interesting and enlightening weather reporting in the place I’d expect to find it: print journalism.

Posted by on June 29, 2016 at 7:26 am, in the category Ministry of Controversy, Unusually Clever People.
Comments are off for this post

6 responses to “Calling long range forecasting on its BS”

  1. kermit says:

    Professor, if i fill a bucket with ball bearings, one of them painted red, and pour it into another bucket, I can tell you two things for sure. 1. I know ahead of time exactly how much more that bucket will weigh, and 2. I won’t have any idea ahead of time where that red ball bearing will be.

    That specific location of one ball bearing is weather (e.g. will it rain on Dec 3rd this year?)

    The total weight of the receiving bucket is climate – we know how much extra heat we will be retaining over the next few decades, given the increase in heat-retaining greenhouse gases,

    And I remember the coming ice age predictions – they were proposed, examined, and rejected as inconsistent with the observations.

  2. Mr Bill says:

    Yes, we all get a kick out of long range weather forecasts. We live in S. Florida. This time of year the only real forecast is “it may rain and it may not rain”. Anything else is only a guess. Thanks for the post.

  3. Absolutely 100% with you (and with Don Paul). In my neck of the woods, I’ve known TV weatherman to tell me there was no rain in the forecast in the same moment when I can look west and see thunderclouds and lightning.

    Isn’t it interesting that most of us will agree that any forecast by weather experts beyond 2 days is suspect, but yet at the same time we will knowingly support experts who tell us what kind of climate change we’ll see in the next 50-100 years? Except, of course, those of us who aren’t old enough to remember the “coming ice age” forecasts of the 1970’s….

  4. Arthur says:

    A few points:
    1. My dad knows a meteorologist, and he says that it’s nearly impossible to accurately predict weather more than 2-3 days ahead. My dad asked him why, if that is true, his station provides 7-day forecasts. The meteorologist said “because if we only did a three day forecast, no one would watch our broadcast!” Unfortunately, at the end of the day, they have to give people what they want.

    2. We have to keep in mind that when it comes to precipitation, forecasters are always going to err on the side of pessimism. If they predict rain and it turns into a nice day, people think “great! we dodged a bullet!” If they predict blue skies and people get caught outside in the rain, they’ll be mad. (I should say MOST people, as I find it very frustrating when I hold off on watering due to forecasted precipitation and we don’t get anything.)

    3. I got curious about temperature forecasting earlier this year and was somewhat able to quantify the accuracy of long-range forecasts. I tracked the 14-day forecasts on and compared them to actual temperatures to come up with standard deviations. (I should note here this is a small sample-March and April, but I think it paints a clear picture nonetheless.)

    Day of forecast-2.8; 1 day ahead-3.6; 2 days-3.5; 3 days-3.6; 4 days-4.6; 5 days-4.9; 6 days-4.7; 7 days-4.9; 8 days; 6.3; 9 days-6.4; 10 days-7.1; 11 days-8.1; 12 days-8.5; 13 days-4.8; 14 days-8.5

    This is actually kind of frightening, since those stats mean that tomorrow’s forecast is only going to be correct within three and a half degrees about two thirds of the time. Further, 1 day ahead forecasts were off by 5 degrees or more about 20% of the time. That’s a significant risk if you’re watching for frost!

  5. marcia says:

    I agree with you on long range forecasting, but I think they’ve gotten a whole lot better at predicting things like hurricanes pretty far out and other new oddities like the effect of the polar vortex. Also, in 1940, the chance of an American being killed by lightning was about 1 in 400,000. Today it’s 1 in 11 million. More people work indoors, but computers have helped predict changes over a short, and sometimes longer, time period.

    One thing that seems to be predictable is the urban heat island effect on cities and areas to their south. I will often see, on radar, the rain clouds move up and over Washington D.C. and take the rain south as the heat rises over the city when a front moves west to east. Those of us to the east miss out on the rain.

    So, my water bill goes up.

  6. anne says:

    Our son got married last Labor Day, and the people at the outdoor wedding venue assured them that 99.99% of the time, it never rains on that weekend–no one had a memory of it having ever rained then, in fact. But in our world, we live or die by the weather, so when the last chance came along to pay for the “rain contingency” (a huge deposit to ensure tents would be put up in the event it did rain), we advised him to do it. Guess what? It poured! I have a lot of respect for what meteorologists do, predicting the weather, but I agree that specifics are hard to nail down more than 2-3 days out.