Flawed Winter Forecasts Issued Too Early

/Flawed Winter Forecasts Issued Too Early

Flawed Winter Forecasts Issued Too Early

Good afternoon bloggers,

Screen Shot 2017-10-04 at 7.43.43 PM

Here is a winter forecast issued by ACCU weather, respected by many, not so much by myself.  The cycling weather pattern that we will be experiencing over the winter and through next September is just now beginning.  Day 1 of this next years pattern is likely happening in these next few days. So, how can a forecast be made before the weather pattern sets up? Well, it can be done, but with very little chance of being right. There are things to look at such as the developing weak La Niña, which is still in question, past years that seem similar to now that I believe is 100% flawed, and many other factors that are still unknown such as will the AO and NAO be positive or negative this winter. So, yes, we can look at some of these features and make broad statements. Many winter forecasts have come out already and we just want to stress that the most important factor has yet to show its face. We will learn more soon.

Screen Shot 2017-10-04 at 7.44.56 PM

What is this statement? Really? Okay, let’s use La Niña.  How well did the El Niño/Southern Oscillation do the past two winters over the western United States? I can tell you. The answer is that the forecasts were a complete failure. The exact opposite to the forecasts happened.  In the strongest El Niño ever the forecasts came out for an ending of the drought, but the rain and snow never arrived in California two winters ago.  Then, last winter the forecast for the drought to worsen due to La Niña developing were put out. What happened? The exact opposite happened. The drought didn’t only not continue or expand, but it got obliterated and ended. Why? The answer lies within what we have been sharing with you for 15 years on this very blog.  The cycling pattern set up in such a way that counteracted the ENSO based forecasts.

So, hang on.  We have another few weeks and we will have your winter forecast put out at the right time in late November.  In the mean time let’s watch Tropical Storm Nate get caught in this changing pattern. We will discuss Nate tomorrow.

Rain and a few thunderstorms are beginning to form and there is a good chance of rain in KC overnight. I have just returned to KC and will be going back to work on KSHB-TV, 41 Action News, on Monday. Thank you for spending a few minutes reading the  Action Weather Blog featuring Weather2020 and the Cycling Pattern.


2017-10-09T09:21:43+00:00October 5th, 2017|General|20 Comments


  1. Snow Miser October 5, 2017 at 5:54 pm - Reply

    Here’s what I want: Either a little more than a 30-day cycle, or a little more than a 60-day cycle, and have this storm come through again just before Christmas, in the form of snow. 😀

  2. Richard October 5, 2017 at 6:29 pm - Reply

    I thought you said the first day of new lrc would be around the 5th. Now it is “likely” in the next few days ?
    I know you have to wait until Dec, to see the repeat patterns, in order to go back and pinpoint the first day.
    Or, if I am wrong, can you really pinpoint the very first day to that much accuracy ?

    • Gary October 5, 2017 at 7:01 pm - Reply

      The fifth, 6th, 7th, 8th, 9th, 3rd, and 4th, are all likely candidates for the first day. I just don’t know yet.


  3. Richard October 5, 2017 at 6:36 pm - Reply

    Oh and welcome back Gary.
    How was HH ?
    The Atlantic ocean can be chilly even in late summer time. But being at the seashore is great any time. Rejuvenating .

    • Gary October 5, 2017 at 7:01 pm - Reply

      The water was VERY WARM, around 81 degrees.

  4. Adam October 5, 2017 at 6:41 pm - Reply

    I always call Accuweather Not So Accuweather. Like someone will say ‘My phone says it’s going to rain tomorrow.’ I say ‘That’s probably Not So Accuweather. Don’t count on it.’

  5. Craig October 5, 2017 at 6:48 pm - Reply

    That might be the most utterly worthless “forecast” map I’ve ever seen.
    So, for the winter, it’s likely to be cold in the north and warm at times in the south. Got it. Thanks, Accuweather.

  6. KS Jones October 5, 2017 at 7:15 pm - Reply

    Have gotten 1.42″ of rain (25 miles north of Manhattan) since It began raining Tuesday afternoon, and the National Weather Service is predicting an additional 3+” will fall here before it is over Saturday morning.

  7. Brian October 5, 2017 at 8:28 pm - Reply

    In the other people who have made forecasts defense, the opposite also happened on your forecast the last two years on snowfall.

    • Gary October 5, 2017 at 8:39 pm - Reply

      Not true at all! Last year we predicted under 15″ for the third year in a row!

      • Brian October 5, 2017 at 9:27 pm - Reply

        You predicted 23 inches in 2015 and we got 5.9. 14.5 in 2016 and we got 6.

      • Brian October 5, 2017 at 9:30 pm - Reply

        2014 was fairly close on your snowfall estimate. That could be considered accurate as you were only a couple inches off. I don’t think being 17 and 8 inches can be considered close unless we lived in Buffalo and the amount you were estimating were far higher.

      • j-ox October 5, 2017 at 9:58 pm - Reply

        Predicting snow amounts/totals is seasonal, yet flaky. Get it? Would be better and more comprehensive to predict the Winter’s total liquid precip – meaning rain/freezing rain would count in addition to snow, sleet, snizzle, pellets, hail, etc… JMO

  8. Kurt October 5, 2017 at 9:07 pm - Reply

    Rain keeps avoiding me just drizzle all day. Hope this is still the obnoxious old pattern that is being stubborn at producing rain. I think we’re one of the driest areas (lowest year to date) rainfall in Missouri and was hoping to put a debt in the growing deficit

  9. Rockdoc October 6, 2017 at 12:34 am - Reply

    I just don’t think the LRC can be used for snowfall in the winter. Why? Because it has not fully developed unless it is a very short cycle such as 35-40 days. To really get a good handle on it you need 2 full cycles due to seasonal differences. It doesn’t mean Gary and team are wrong, the beginning of the LRC happens at same time as transition into winter so it makes it more difficult!

    Anyways, just my two cents, or 6 inches of snow 😉

  10. Rockdoc October 6, 2017 at 12:37 am - Reply

    I should qualify that Gary and team cannot be as accurate, bang on, early on in the season. As time goes by, in January they have a much better diagnosis!

    • Gary October 6, 2017 at 5:01 am - Reply

      We can take a shot at it, but predicting one number for seasonal snowfall is really difficult. We did get it right two out of the past three winters, on predicting the below average snow. Let’s see how it sets up. You are correct in how the accuracy jumps after around December 20th.

      • Anonymous October 6, 2017 at 8:20 am - Reply

        LOL Gary – the spin master. While it may be true you predicted below normal snowfall, and you get credit for that, you were still way off on the amounts.

  11. Snowflake October 6, 2017 at 8:07 am - Reply

    Frankly, I’ve never really understood why anyone endeavors to put forth a full-season snowfall forecast for a region such as KC. Where the average is something in the high-teens…but the range of historical snowfall totals is something like a low of 3″ to a high of what, 45 or 50″? Any effort at a seasonal forecast is a shot in the dark, with one storm making or breaking that forecast. (a great example is when a forecaster calls for, say, a seasonal total of, say, 14″, but then we only get 6″…that forecast usually gets defended as close enough, because “after all, just one 8″ event would have made it perfect!”).

    A couple of years ago when Gary called for a mega-snowstorm that never materialized was a great example…that seasonal forecast could have been made 2 days before the supposed storm and it still would have been wildly off target.

    If one attempts to make those seasonal forecasts in November or December, if they make enough of them, time will show historically there simply isn’t sufficient precision or accuracy to them. Sure, one may land close to the target (luck), but over time, the spread will demonstrate the lack of actual skill involved.

  12. Urbanity October 6, 2017 at 9:13 am - Reply

    I agree with the Accu weather forecast, for no other reason than what they are showing is the most “likely” weather pattern. It happens about 70-75% of the time where most of Kansas stays dry. We get precip when Spring ushers in and when Fall ushers in, the rest of the time mostly dry in most of Kansas. East of the turnpike is a little different story as they usually have the moisture to work with. Either way it’s easy to predict dry weather for most of Kansas because negatively tilted troughs arriving out of the Southwest just don’t happen very often anymore. Now this year we have had a more amplified pattern, Hurricanes are a good signal of that IMO, along with the impressive fluctuation in temps this year. So perhaps, just perhaps, this year will be different.

Leave A Comment