I took my seat at the University of Georgia and University of Oregon football game that weekend, and the person sitting to my left started small talk. After learning that I’m director of the Atmospheric Sciences Program at the University of Georgia and a meteorologist, she said, “Maybe you can tell me why today’s forecast was so wrong my app said…” It was an innocent question, but one that I (and many of my colleagues) are intimately familiar with. Since it didn’t rain in the stadium, she thought the prediction was wrong. Here are four mistakes I hear about a lot when people consume information from weather apps.
That First Mistake is the long-standing misconception of “percent chance of rain”. The National Weather Service – Atlanta on its website explains a 40 percent chance of rain this way – “The Probability of Precipitation (PoP) simply describes the probability that the predicted grid/point in question in question will receive at least 0.01 inches of rain . So in this example, there is a 40 percent chance of at least 0.01 inches of rain at the specific forecast point of interest!” More specifically, the National Weather Service website further explains, “(1) If the forecaster were 80 percent confident that If rain were to develop but only expected it to cover 50% of the predicted area, then the forecast would say “a 40% chance of rain” for a given location. (2) If the forecaster expected a large area approaches a precipitation area with 100% coverage but is only 40% certain it will reach the forecast area, this would also result in a “40% chance of rain” anywhere in the forecast area.” The PoP is calculated by the weather forecaster’s certainty (C) that precipitation will occur (form or move into the area) is multiplied by the expected area coverage (A). An adjustment factor is made by shifting the decimal point two places to the left.
That second Mistake many people make with apps is related to duration. As a colleague noted on my social media page, there seems to be an assumption that in many people’s minds an 80% chance of rain means it’s going to rain “all day”. If you consume the definition correctly, the 0.01 inch of rainfall can occur early in the morning, over the course of a few hours, or later in the evening. A third Error related to the first two is assuming that the forecast is the same for every single location in the region. Typical summer days here can be sunny at my home and rainy at my son’s school a few miles away. The combination of probability, evolving weather, and its distribution over an area is difficult to summarize into an icon on our smartwatch or phone.
The last error is related to the evolving weather itself. Many people look at their weather app’s icon and it can have a sunny icon or a rainy cloud with a lightning bolt. However, this static representation of the weather is not suitable for dynamically developing weather situations, in particular storms. In dangerous weather situations, it is important to understand the evolving threat. A strength of many apps is the weather radar. They are useful for identifying the location and evolution of precipitation or storms. Many of them even have lightning detectors. I also strongly recommend that your wireless emergency alerts are turned on during severe weather, or that you have a weather alert app enabled.
Look, I’m a meteorologist. I can assess the weather myself with various tools. However, I have a few weather apps on my phone. I’m not an anti-weather app. This short essay is intended to warn you about how to use it and not dismiss the forecast because you have not understood how to use the information presented to you. At the end of the day, your app’s weather information doesn’t come from the “app fairy”. It is based on weather observations, numerical weather models and professional analysis. It is also incumbent on the providers of weather information to overcome the “inertia of practice”. In other words, we have a lot of weather industry jargon and jargon that is utterly meaningless to the general public. Know your listeners.