Why do forecasts change?

This is an archived article and the information in the article may be outdated. Please look at the time stamp on the story to see when it was last updated.

How did we go from expecting dry weather on Friday and Saturday to this cloudy, damp, cool mess we have right now?

It’s a legitimate question, but a lot of the criticism comes from not getting the whole story. I hear we live in the information age, but it seems more like the “knee-jerk-reaction-to-incomplete-information” age. I’m guilty of it as much as the next person. Technology has given us great ways to communicate weather information like never before, but fewer people know less about what the forecast really says these days.

IMG_1547Live Alert 19 is a great app. It does what it was designed to do: give you a complete resource for weather in North Alabama and Southern Tennessee. If all you ever use is the pictured screen, you will never really get a full grasp on real weather expectations and the potential for some changes to the forecast.

Each one of those lines has three numbers and a picture for the overview; however, when you touch the date, more detailed information comes to view telling when, how much, who may or may not have the best chance of rain, sun, severe weather, etc.

There’s even more real-world ideas about the forecast in our extended discussion that we update several times each day.

Plus, we have social media that gives you the opportunity to ask us anything you want.

Jason 1So why is it such a surprise that we didn’t clear and didn’t get rid of the rain?

That was the higher
probability outcome given the information we had earlier in the week back when we thought that Joaquin would be gobbled up by a strong upper-air low near the coast.

Forecasts are bound to be altered as new information corrects previous errors in estimating how huge features interact with each other.

All of those features are variables with knowns and unknowns. It hit me today that some of my son’s toys could explain this in a simple way. These are two “cars” with similar wheels for a Brio-style train track. One is an actual train car, and the other is a steam roller.

The predictable car runs just as you would expect for this toy:

The "Known"

The “Known”

The steamroller is a wild card. You probably expect that it won’t stay on the track as well, but it takes some trial and error to figure out exactly how it will behave. Will the back wheels be enough to keep it on the track, or is there too much variability without those front wheels?


I ran that steamroller several times, and it always ran off in the same place after it gained momentum and hit the curve. On a perfectly straight track, it stays on longer. That little curve creates a huge change.

Now imagine those cars moving on an invisible track thousands of feet above our network of weather observations. It’s really tough to find those little curves that cause things to go off-track. Our forecast models do an incredibly good job of estimating them, but they are far from perfect.

So what’s the lesson here? Unknown (and sometimes unseen) factors can play a huge role in big-time storms like Hurricane Joaquin and the strong upper-air low over the South. One little “curve” under the right conditions can make the forecast look like this:

(Sorry Bama fans, I’m not rubbing it in, but this is about as opposite result as you could get from a missed field goal. I’ll even give you a Roll Tide to smooth it over)

Connect with me!
Google+: Jason Simpson
Facebook: Jason Simpson’s Fan Page
Twitter: @simpsonwhnt

Notice: you are using an outdated browser. Microsoft does not recommend using IE as your default browser. Some features on this website, like video and images, might not work properly. For the best experience, please upgrade your browser.