By Bryan Norcross

Special to The Washington Post

“I wish I had a job where I only had to be right 50% of the time.” People in the weather-forecasting business have heard that a million times. Weather science is better than ever, and we are right most of the time. But the reality is much more complicated.

The rabbit hole begins with the app. Open any weather app, and you find a stew of good science and wishful thinking. We would all like modern science to tell us exactly when it is going to rain this afternoon and what the weather is going to be a week from now. In the end, many of those forecasts are pseudoscientific.

The sound meteorological science behind the app is being used to make bogus forecasts because graphical slots on the app have to be filled. It’s easy to design graphics that show hourly forecasts for the rest of the day, but it’s impossible to credibly fill those slots in all weather situations.

For example, in hot and humid climates, individual thunderstorm cells interact in chaotic ways, which defy down-to-the-ZIP-code predictions. Still, the app’s forecast engine chugs along, making forecasts anyway and forcing numbers into graphical holes, whether the weather is predictable or not.

Looming over every forecast system are also the intrinsic uncertainties in weather predictions seven or 10 or more days in the future.

If you think about every hour and every day part as a specific forecast an app user might rely on, you can see how people constantly bump into likely-to-fail predictions. A sour taste about the ability of meteorologists to predict the weather is a natural result.

In today’s app-driven world, everybody has to have one. Most local news stations market weather apps — some branded with their weather team or chief meteorologist — that offer the full set of hourly forecasts and predictions a week ahead, or more. You might think the local weather team is sitting around all day tuning the numbers to insert the best possible forecasts for their viewers. In general, however, the app is populated with forecasts from an out-of-town computer, or sometimes it’s a mix of computer-driven and locally modified information. And that mix changes through the day.

It would be impossible for any local weather team to monitor every hour and day for every location in its coverage areas — so most teams just throw up their hands and let the app be the app.

Local forecasters cover more news shows than ever, while monitoring the local weather to make the best possible forecasts. And top-notch scientists are involved in the most popular weather apps, tapping into cutting-edge research and pushing the forecasting envelope.

The breakdown comes when the product designers tell their forecasting computers to come up with forecasts to fill graphical slots, even when the meteorological situation is uncertain. And, of course, since other apps do it, everybody has to do things that are not scientifically sound to be competitive. And like sheep, they follow each other off the cliff into meteorological mumbo-jumbo.

Anybody who has ever forecast the weather knows some situations are highly uncertain, while other times, you’re pretty sure what’s going to happen. The further in the future you look, it’s even more often that two or more disparate possibilities appear possible. But the way we present weather forecasts today — an icon with a few words and numbers — makes it seem we are as certain about tomorrow as we are about the weather a week from now.

On television, a good weather forecaster can blunt the misleading simplicity of the seven-day forecast graphic with personality and explanations. Most apps, however, cannot.

The individual elements of those ubiquitous seven-day graphics, where the days are lined up vertically with icons and numbers conveying the forecast, are appropriately called tombstones. Tombstones indeed: They rest on the graves of meteorologists’ credibility.

24116748