According to the National Weather Service, a five-day temperature projection in 2018 as a two-day outlook back in 1988. Nowthe accuracy and timeliness of weather forecasts are about to get even better, thanks tothe National Oceanic and Atmospheric Administration’s (NOAA) to the country’s flagship weather model, the Global Forecast System (GFS), this month. The improvementwill reportedly grow the range of reliable forecasts to a week or longer.
Here’s how it works:When you check the weather to plan for that camping trip next weekend, the information you’re currently looking at comes from meteorologists using computer-based models to predict conditions. However, they’re not just copying those predictionslike a student cribbing answers from a textbook. They only use those for guidance. Each representation has different strengths and specialties, and it’s up to users drawing ontheir experience to figure out which ones would help them the most tocreate an accurate forecast.
Some models focus on large-scale weather systems around the world, while others zero inon smaller-scale conditions,like thunderstorms and sea breezes. The GFS, as its name suggests, operates on a global scale, running four times per day to provide meteorologists around the world with guidance on what’s coming their way.
These models are powered by a dynamical core—a collection of mathematical equations that explain atmospheric physics using available data and makingpredictions based on that information. That’s where the GFS is getting a boost.
The GFS’s new core—called (FV3)—is designed to partially account for a global model’s general inability to resolve smaller-scale features. The upgraded physical equations in the FV3 give it a more accurate, detailed view of the atmosphere, allowing it to “see” storm systems and different weather features much more clearly than previous iterations.
There are plenty of perks to building a new foundation for a major weather model. The FV3 core allows itto run faster and more efficiently on new computers, which is necessary formeteorologists to keep it fresh and current in the coming years. The upgrade will also remedy some clunkiness of the previous versions(for example, it will offer smoother, more realistic rainfall maps), giving forecasters a truer sense of what the model thinks will happen in a certain area.
Previous dynamical cores couldn’t resolve the updrafts that power storm development. The FV3 core gives the GFSthe ability to resolve that issue, a boon to forecasting since thunderstorms have such a large influence over regional weather.
This is not the first time NOAA has upgraded the GFS. rangefrom minor tinkering to major overhauls. For example, the resolution quadrupled between 2002 and 2015, and that resolution will likely improve in coming years. This is the first time the core has been upgraded in 40 years.
NOAA originally planned to roll out this upgrade at the end of 2018, but the government shutdown—which hit the National Weather Service particularly hard—pushedthe planned release date back to March 2019. That lengthy delay gave meteorologists a chance to in the model, which meant the model routinely ran too cold and produced too many “phantom”snow and ice storms. Experts kept the model on hold for a couple of months until they were satisfied they’d resolved the issues.
The GFS has lagged behind other countries’global-forecacsting models in terms of overall accuracy. This upgrade is part ofthe agency’sEarth Prediction Innovation Centerplan to advance what“will advance U.S. weather modeling and reclaim international leadership in the area of numerical weather prediction.”The code for the FV3 core will also be available to a wider research community for continuous testing and upgrades for the model.
On a more practical level, having more accurate forecasts a week or more in advance will helpeveryone from emergency officials preparing for dangerous conditions to someone planning their next adventure.
The GFS and many of its counterparts are from the U.S. government, which means there are a ton of sites out there that offer weather-model output. and are two of the most widely used modeling websites in the field. However, it’s only a good idea to check the weather models if you know what you’re looking at. For the average user who just wants to check the weather, there are a handful of trustworthy apps available that use the GFS and other sources to make informed forecasts. And they are about to get a lot more accurate.