It sounds like a silly habit, but it also raises a good question: Why do weather forecasts vary depending on which app you use?
The forecast process starts with data and observations that come from weather stations around the world, satellites, radar, reports from volunteers, and weather balloons that collect information about the atmosphere, such as humidity, wind speed, and temperature.
All of these measurements are fed into supercomputers run by the United States government and other countries. The European Union is a major source of weather data, but the British, French, German, Japanese, Canadians, and Chinese all have their own models too.
Supercomputers take those initial conditions and then use mathematical equations to come up with a forecast. There's no perfect algorithm because the Earth is so big that's impossible to have observational data for every parcel of air.
"Algorithms make some assumptions about the atmosphere," said Chris Maier, a meteorologist at the National Weather Service. "The equations provide an estimate of where certain atmospheric conditions are going to be."
Since each computer model uses a different mathematical formula, each weather forecast may be slightly different.
The outputs, or "solutions" to the equations, are typically maps that show things like pressure, temperature, and precipitation in a certain geographic area. Supercomputers with more power will produce higher resolution maps, which means they are more accurate.
"The forecast that you see on weather apps is based off one or a compilation of the model information," according to James West, senior meteorologist at WeatherBug. "The accuracy of weather apps depends on how each organization uses the data they are given."
Weather data is provided free by the U.S. government through agencies like the National Oceanic and Atmospheric Organization, making it relatively easy for anyone to develop a weather app. But that doesn't mean it will be good.
Organizations that get data from the government, but also from their own network of weather stations and proprietary computer systems, can take these models to generate their own forecasts.
In a situation where the weather is changing rapidly, such as during a blizzard or hurricane, "the organizations that have invested in technology will generally offer more frequent updates and more accuracy," according to West.
The forecast also depends on the experience of the meteorologist and how he or she interprets the models. There are models that are better at handling certain weather events (i.e hurricanes or snowstorms), models that are better designed for long-term or short-term forecasting, and models that are known to be less accurate. It's the meteorologist's job to know the reputation of each model and adjust and improve the forecast based on this knowledge.
So basically, there's variation that comes from the different supercomputers, and then organizations have their own formulas and meteorologists that adjust those forecasts further.