You can now expect more accurate weather forecasts from Google
Google's improved weather training model for AI, WeatherNext 3, will be used on Search, Google Maps, Gemini, and the Pixel Weather app.
Weather forecasts from Google should improve now | Image by Google
How many weather apps do you have installed on your phone? On my iPhone 15 Pro Max I have downloaded six such apps and there is the native weather app for iOS. By the way, Siri AI told me how many weather apps I installed when I asked, a nice improvement for Siri. The legacy version of Siri would have opened a specific weather app or sent me to the App Store instead of getting into system data to count the number of weather apps I downloaded.
Gemini found six weather apps installed on my Pixel 6 Pro including the native Pixel Weather app. Why do I have so many weather apps on my phones? Well, there is almost a good reason behind this odd hoarder-like behavior. Some of them focus on what is called "hyperlocal weather" and there are apps that specialize in precipitation.
Apple iPhone users have a native iOS weather app that delivers a 10-day forecast that shows conditions for the next few hours and for the next 10 days. Users can also view 1-hour and 12-hour radar sweeps to give them early precipitation warnings.
Non-Pixel Android phones do not have a native weather app
On Android, things get a little confusing as Google 86'd its Google Weather app earlier this year forcing Android users to install a third-party app like The Weather Channel or
AccuWeather. Pixel users, on the other hand, have an outstanding app called "Pixel Weather" and there are some who will tell you that Google dropped the native Android Weather app so it could use Pixel Weather as a selling tool to generate more Pixel sales.
Which weather app do you find to be the most accurate?
4 Votes
The Pixel Weather app presents the same info as the Apple Weather app albeit with a more colorful presentation. They also don't agree on the current temperature or some of the other weather-related data. Google isn't admitting that it needs someone new to stick their head out the window at Mountain View in order to post current Weather data, but it has created a new AI-powered Weather model called WeatherNext 3 that it will use to produce weather-related content for Google Search, Google Maps, Gemini, and the Google Maps Platform Weather API that is the siurce of the weathjer data used for the Pixel Weather app.
Google introduces WeatherNext 3, the most accurate global weather model to date
The WeatherNext 3 is the most accurate and advanced global weather model to date according to independent live evaluations by Brightband. The model learns from observations made in real-time, which allows it to provide more timely predictions including localized forecasts. The model turns the raw weather data into a high-resolution forecast every hour. With a sharper resolution, and AI training, these forecasts are more accurate.
Here are some more great articles for you to read:
These forecasts are then read by those using Google products worldwide. WeatherNext 3 can show forecasters a look at the global weather picture that is five times sharper than Google's previous weather model, WeatherNext 2. The latter trained on data sourced from supercomputer-driven physics that had a six-hour data lag.
WeatherNext 3 is now employed on Google Search, Google Maps, and the Weather API used by Pixel Weather
Google is using WeatherNext 3 starting today for Google Search, Gemini app, Google Maps, Google Maps Platform Weather API, and Google Earth Engine. Longer term forecasts will be greatly improved and when planning ahead, you'll find that precipitation forecasts have improved by as much as 50%. The greatest improvements will take place in areas where previous forecasts were not so accurate.
WeatherNext 3 uses resolution 5 times sharper than the previous generation
If having an accurate weather forecast on your phone is important to you, this is great news. According to Google, "WeatherNext 3 brings us closer to a future where forecasts truly match what is happening on the ground." The bottom line is that instead of waiting every six hours for data from the atmosphere released by Uncle Sam, WeatherNext 3 looks at geostationary satellite images every hour.
Thanks to AI, no longer does there have to be a tradeoff between high-resolution local forecasts and low-resolution global forecasts. With WeatherNext 3, temperature and moisture can be visualized every hour at 5-kilometer resolution, with other surface variables at 10 kilometers, and atmospheric variables, like wind speed, at 25 kilometers. This is five times sharper than the previous model, WeatherNext 2, which produced forecasts on a 25-kilometer grid in 6-hour increments.
Popular stories
Latest News
Things that are NOT allowed:
To help keep our community safe and free from spam, we apply temporary limits to newly created accounts: