Why meteorologists get a bad reputation

Latest ensemble forecast for YVR. Red box highlights the uncertainy with precipitation amounts. But I assure you, it will rain this month.

Latest ensemble forecast for YVR. Red box highlights the uncertainy with precipitation amounts. But I assure you, it will rain this month.

To be fair, I don't have extensive knowledge on social media, but I've noticed a growing trend that can't be ignored.

The bottom line for many social media and publishing agencies is simply the amount of engagement and buzz they can create.

They need a certain level of engagement to survive, but more importantly grow and attract advertisers to fund their projects and to hopefully turn a profit.

It's about clicks.

Because without them, advertisers will be less inclined to invest and support these ventures.

For example, take a look at the montage below of numerous headlines in 2012 by the Daily Express in the United Kingdom:

Photo by BBC Meteorologist Paul Hudson. Original article can be found here

Photo by BBC Meteorologist Paul Hudson. Original article can be found here

The tabloids in London are notorious for sensationalizing weather predictions, and more often than not when these hyperbolic titles and articles fail to verify, government agencies and private weather prediction firms take a big credibility hit. 

"Why don't they just look outside"

"Yes, meteorology, the only job where you can be wrong 50% the time and still keep your job"

"They said it was going to snow tomorrow. I'll believe it when I see it!"

Today, Brad Panovich (Chief Meteorologist of NBC Charlotte), posted a refreshing idea about of the major problems of weather forecasting: 

"First let address a fundamental aspect of being a good weather forecaster. Job #1 of any good and experienced forecaster is to forecast what is ”most” likely to happen not what “could” happen. I see entirely too much time spent on the “could”forecast and not the most likely forecast. While it can be fun to talk about the what ifs it doesn’t make for a very good forecast."

Certain media agencies and blogs love to focus on the 'could' forecast and discount the most likely scenario:

Image by: The Oatmeal

Image by: The Oatmeal

Why?

It's painfully obvious: It's much more fun, eye catching, plus the extreme solution is much more likely to be shared on social media. 

Wishcasting is becoming a painful epidemic, spreading through social media like wildfire. 

Vancouver media is not immune. The well known Vancouver blog Vancity Buzz has been guilty of wishcasting a couple times this fall/winter, including their post back in October claiming a snowfall was imminent when in reality there was close to zero chance of snow for Metro Vancouver.  

Exactly one day ago, another article by Vancity Buzz garnered massive social media buzz:

VANCOUVER FORECAST CALLS FOR 7-DAYS OF BLISSFUL SUNSHINE

  • 6,600 shares on social media
  • One Facebook post received over 1500 likes and dozens and dozens of comments

The more truthful title: 

VANCOUVER FORECAST CALLS FOR MORNING CLOUD/EXTENSIVE FOG, BUT MAYBE A CHANCE OF SUNSHINE...

This would not have collected over 6000 shares, even though it's the most likely scenario of verifying. 

But can you blame Vancity Buzz?

They simply showed one deterministic long range solution from The Weather Network, and if you're curious their long range trend now shows rain for next weekend.

Low cloud and fog present one of the most unique and complex problems to weather forecasting, eloquently highlighted by Professor Cliff Mass, who teaches atmospheric science at the University of Washington. 

At best, those 14 day trends should be used with extreme caution and as a rough guide, or for a quick glance at a possible state of the atmosphere in the future. Ideally, a disclaimer could be posted beside these long range outlooks.

It's clear these false headlines have the potential to damage the weather prediction industry. Forecasts in the 0-5 day range are very, very good. Much better than they were even 10 or 15 years ago. Trends are spotted between 7-14 days with improving accuracy, but there's still much to be desired.

Accuweather's recent implementation of a 45 day forecast?

Results: Worse than climatology.

Graphic By: Penn State Meteorology Students

Graphic By: Penn State Meteorology Students

So, in other words, if you just predicted what the weather ought to be for the certain day, the skill was significantly better than what Accuweather's new 45 day forecast was suggesting...

It's one thing to talk about a problem, but it's another to offer a solution. Comment below if you have a suggestion, so we try to create some positive discourse here on a very real problem. 

I haven't forgotten about talking about a some teleconnections, including the East Pacific Oscillation. I'll go into a bit more detail about a return to a more usual pattern, as the first signs in the long range weather models are hinting that our ridiculous resilient ridge may finally crumble. 

In short, there's hope.

#50shadesofVan