GFS model from yesterday morning shows storm near New Orleans on August 28. (

It’s all over the internet.  Florida and (or) the Gulf Coast could get hit by a hurricane next week.  Of course, they may not, too.  The storm may never form, might hook out to sea, or could get shredded apart passing over the Caribbean islands.

But what might not happen isn’t the word hitting the street and that’s giving some in the weather community heart palpitations.

GFS model from yesterday afternoon model shows storm near Lake Charles, Louisiana on August 28. (

Many weather communicators want the public to understand the full range of possibilities.  We want the public to appreciate that model simulations for storms more than five days into the future aren’t realistic.  We lament that armchair meteorologists (amateurs, students, novices, etc.) post unreliable model simulations on social media – without any context – of a storm obliterating a coastal city.

We cringe when these suspect forecasts are shared thousands of times, misleading an unknowing public. While some us are secretly envious of the attention, we ultimately worry about a loss of public trust in weather forecasting when they are wrong (most of the time).

Forecast skill for hurricane intensity is "negative" five or more days into the future (NOAA)
Forecast skill for hurricane intensity is “negative” four or more days into the future (NOAA)

The concerns are legitimate, but there’s little an individual weather communicator can do to stop these bad practices.  Computer model forecasts are everywhere and anyone can post them.   It’s pointless to expose and shame those who we feel are putting out potentially misleading information.   It’s a never-ending and unwinnable game of whack-a-mole.  (I’ve tried to do this, with little effect.)

Complicating matters, a number of professional meteorologists are just as guilty as amateurs in posting model simulations without the needed qualifiers.  If weather communicators publicly call out colleagues who they feel are committing meteorological sin, it could be construed as professionally discourteous and devolve into ugly conflict.

GFS model from yesterday night shows storm near Houston, Texas on August 31. (

So how should the weather community move forward?  Every meteorologist and weather communicator who cares about providing useful and credible information needs to focus on educating their readers and viewers about the limitations of weather forecasts.  They should also re-double their efforts to effectively discuss what is known and not known when a hazardous weather event is showing up in the long-range.

Yes – this can include showing some of the “sexy” model simulations that attract eyeballs and allow 15-year-old high school students to get thousands of likes and shares on their Facebook pages, as long all other alternatives are adequately explained.

If we do this and do this consistently, we solidify our position as both entertaining and trusted sources. We can also share good examples of colleagues doing this the right way.

GFS model from overnight Mobile on August 29. (

Education is the only weapon we have in this fight against social media misinformation.  The more we do it and the more of us that do it, the smaller percentage of the public that is left to be misled.

Constructively, the National Hurricane Center posted this message to Facebook this afternoon about the limitations of long-range hurricane forecasts:

Nearly every [tropical] disturbance poses some potential to become a dangerous hurricane. Folks in hurricane-prone areas should always be keeping an eye on the tropics and be prepared to respond when a true threat develops, but also remember the limitations in the science. NHC’s forecasts of tropical cyclone formation and track extend out only to 5 days – because the science hasn’t advanced enough to reliably forecast beyond that time frame. (We’re working on internal forecasts out to 7 days, but we’re a ways away from feeling comfortable making them public because the errors can be quite large.)

The bottom line really is: be alert, be prepared, but also be wary of long-range projections that go beyond what the science can offer.

That’s a good start.

GFS model from this morning shows storm near Miami on August 26. (

(Sprinkled within this post are the last five simulations of the GFS model which show a tropical storm or hurricane making landfall anywhere from Florida to Texas between August 26 and 31. Not shown are other models with entirely different solutions.)

Thoughtful related reading, from The Vane at Gawker: A Hurricane Is Coming

Video: See my discussion of hype and social media on The Weather Channel’s WxGeeks program (links embedded below)…