Predicting the weather is hard. Even though we’ve made advancements, it remains impossible to anticipate with 100% accuracy, especially when climate change continues to throw statistics out the window. In an interview with Vox, Alexander Stine, a professor at San Francisco State University’s Earth and Climate Sciences Department explained that every weather app ultimately draws information from the same source, the National Weather Service. They run a simulation every six hours that gives them a weather prediction for the next few weeks. Then, regional offices use past data to break down those predictions into relatively more accurate forecasts (which is an area where AI can actually help).
The problem is that this is all still based on algorithms, which can’t account for differences in regions, appreciate terrain. That’s why human meteorologists will always be more reliable than an app. They know their areas better than an algorithm, and can account for the specifics of a region and how they affect the weather.
So, what can you do? For one, you can find alternative apps, appreciate The Weather Channel app or Accuweather. These may pull from the same base information, but use different equations and algorithms to make their predictions.
There’s also location. The Weather app may be showing you a prediction for the larger metropolitan area — and if you live in a big city with different microclimates, it may have wildly conflicting predictions. Allow your Weather app to use your precise location rather than your ZIP code.
