Clouds header.

Please stop using basic weather apps to make decisions

The sun is shining on a beautiful summer day. Temperatures are in the 90’s, the sky is hazy and the air is humid. You’re enjoying spending time outside with family and friends near the pool. In a matter of 5 minutes, all of this changes. Torrential rain is falling, lightning flashes and thunder cracks, debris flies as high winds gusty. This time, everyone stays safe as you hurry indoors. The conversation once everyone gets inside is always the same:

“That storm came out of nowhere, I had no idea it was supposed to rain like that today!”

Your weather apps probably didn’t tell you so, either.

Stories like these are all too common. While they can generally be taken lightly in the “kids at the pool” scenario, the story has a different ending when the thunderstorms are more severe; or if you are a contractor, or if your landscaping company has employees caught out in the storm, or if you make decisions for a sports team — and just put thousands of people at risk.

The idea of using an algorithm-based weather app to make decisions regarding the weather is crazy.

The idea of using an algorithm-based weather app to make decisions that affect thousands of people, or may cost your company thousands of dollars, is bordering on insanity.

Here’s why:

These weather apps use algorithms, or in some cases simply raw data, from a single weather model and display them on your phone, tablet, or computer as a forecast. Yes, that’s right: You’re using a single computer simulation of the atmosphere to plan your day, your business, or your event. There are hundreds of weather models, run multiple times per day, available on different platforms, producing different outputs.

There are many different weather models, all of which are a critical part of the forecast. (Image of PSU Ewall).

There are many different weather models, all of which are a critical part of the forecast. (Image of PSU Ewall).

None of the weather models are worth much without a meteorologist to interpret what they are “saying”. Think of it this way: All of the weather model data that is available to the public can help us. But not all of it is easy to understand. Weighing different models into a forecast based on the atmospheric evolution, and known biases of individual weather models, is a critical part of the forecast.

In winter, forecasters know not to weigh GFS (Global Forecast System) data, past 84 hours in advance too heavily into their forecast. With Eastern United States storms in particular, the GFS has a history of being too “progressive”, or fast and flat, with potential storms. So while other forecast models may be showing the storm (like the ECMWF, for example), your weather app which pulls data from the GFS is telling you it will be partly sunny with a high near 40 in three days.

The unfortunate truth is that weather apps typically don’t include the critical input, analysis, and expertise of a meteorologist. This leaves many making decisions that are based off of data that isn’t being correctly interpreted. And this leads to communication problems that can become exponentially detrimental to you, your business, or your employees.

Here’s an example of the data you’re actually using to make decisions. The NAM model (North American Mesoscale) produces freely available forecast data that shows temperature, sky cover, precipitation chances, wind direction and wind speed. This data is, quite often, used in your handy weather app.

 

NAM model forecast data. Take it and run. Except don't.

NAM model forecast data. Take it and run. Except don’t.

But what happens when the NAM is wrong? As we talked about earlier, the NAM is wrong quite a bit — it has a bias to overdo thunderstorm chances, run too high with dew points, and produce too much precipitation. But your weather app won’t tell you that. Your decisions will be based on one weather model, which a meteorologist will tell you is probably mishandling the forecast.

Meteorologists use many different weather models each day — but they are simply used a guidance tools to help the forecast process. They aren’t meant to be used for their raw data as a take-away forecast. A meteorologists job is to understand the atmospheric process, use these models as guidance, and create the best forecast moving forward.

One of the biggest problems we are currently facing in this industry is a lack of communication between meteorologists and the public. Many don’t understand the risks and uncertainties associated with the weather forecast on any given day. Weather apps make that communication even more difficult.

In the end, it is a meteorologist who will provide you with the best forecast — and put you in the best position to make the correct decision. Meteorologists aren’t going to be right 100% of the time. But at the very least, meteorologists provide experience and knowledge that can help communicate risk and uncertainty, while helping to grow everyones understanding of exactly how the atmosphere works.

Next time, make the right decision.

Comments

comments