GFS or EURO: Which is More Accurate?

19376 reads link

By Zachary Duhaime on November 10, 2012, 4:00pm Last modified: November 13, 2012, 4:05pm

Use your ← → (arrow) keys to browse more stories.

Next Article
ECMWF model forecast, left. GFS model forecast, right.


        Weather forecasters just don’t stick their heads out of the window to make their weather forecast. Nor do they just see what other stations are forecasting. In reality, it’s a lot more complicated than that. They use many different things in forecasting the weather. There are two main things meteorologist used the most in creating their forecasts. The first main thing is reading up on what the National Weather Service puts out in their daily update. This is important because the national Weather Service has the highest authority throughout the whole weather department. The second and probably most important tools meteorologists use to help them build their forecasts are weather models. These models can forecast just about everything possible. From hurricanes to snowstorms and temperatures to wind gusts, these models are just one of a kind!

       Without these models, it would be nearly impossible to be accurate enough in warning the public when the time comes. Being able to predict major weather events days ahead is key in weather forecasting. The more time people have to prepare for major storms, makes for less chaos and people can figure out place to go if they have to evacuate as well as allowing the needed time to get necessary supplies in case the power is out for days.  These models have helped predict storms days and even over a week from when they actually hit. The only let down is that while in some cases these storms can be forecasted days before they hit, most don’t hit. This is why there is a huge question about the “hype” factor that plays out in every major weather event.

      For example, with Hurricane Irene, both the GFS and the EURO/ECMWF forecasted a massive storm to form somewhere along or just off the East coast. Most meteorologists began watching these models about 5 days before Irene made landfall here in CT. Most people thought because we have not seen a storm like that in decades, so they though that it would be impossible for use to get hit. Along with that, a lot of people also thought that it was just a lot of “hype” and that as 99% of all storms do, that Irene as well would miss us completely. Well obviously that was just the opposite. Hurricane Irene DID in fact crawl up the east coast and made landfall in extreme southwest CT as a tropical storm and knocked out power to hundreds of thousands and turned beachfront communities in to beaches themselves.

       Now not even 2 months after Irene came another historic storm and it happened to fall right on Halloween. This was in fact a major nor’easter that clobbered the Northeast with several inches of snow and left over 2 million in the dark. Once again, lots of citizens though since it was only late October that it was some kind of “trick’ or joke that we would have a massive snowstorm on Halloween. And… it happened!  Now with this storm, the GFS model did not even pick it up until just a day or so before its arrival. And for the EURO… This model not only forecasted the “October nor’easter” days in advance, but also was dead on in where the snow/rain line was going to end up. (

    So after seeing how these models are used and what they do to help out everyone from the meteorologists themselves to regular everyday lives, here is more information about each individual model.

      Let’s start off with the American model which is the GFS. And yes that does stand for something. GFS is short for Global Forecast System. This system uses global computer models as well as variational analysis used by the National Weather Service (NWS) to make weather forecasts. There are several different mathematical equations that go into forecasting weather. This model has the capability to forecast up to 16 days in advanced though anything past 3 or 4 days out would NOT be considered accurate. So the reason why this is called the “American model” is because simply, it was founded in America and it has been known to be one of, if not the best forecasting models out there to predict the weather for America and beyond. (

     Now, after hearing about the American model, here is everything you need to about the European Model (ECMWF/EURO). So while the abbreviation was pretty easy to figure out with the GFS model, it’s just the opposite for the EURO. The ECMWF model stands for European Center For Medium-Range Weather Forecasts. This modeling system was founded and used mostly by the European’s until the National Weather Service got their hands on it because they got to witness first-hand how accurate this model can sometimes be.  Just like the GFS model, several mathematical equations along with several smaller models come together and put into one and only one model run which will be the official run by the ECMWF. (

     These models basically compete against each other to see which one can put out the most accurate forecast with the most time possible before a storm hits. But, they also have a lot in common. Though meteorologists can “run” these models ay just about any given time, they both update about four times a day at around the same time. (8 am/pm and 2am/pm). Both of these “model runs” will show the updated forecasts that usually change every run after new data which includes air pressures, patterns and the Jetstream. The more changes from run to run usually means that there is just too much time to tell exactly what will happen when but if it is accurate, a few days before a storm hits just about every model run will be very close to the one before and after each one.

      Besides those two models, there are a few others that are worth mentioning. One of the models is called the GEM (Global Environmental Multiscale Model) and the other called the NAM (North American Mesoscale Model). These are also two fairly reliable models that the NWS also uses especially when making long-range forecasts. ( /)

     Let’s put these models to the test. A great example of how accurate on inaccurate these two models were when forecasting Hurricane/Superstorm Sandy. Granted the fact that this storm was a combination of a nor’easter and a hurricane, these models did pretty well. At the end, both models just about nailed the landfall of sandy and what the affects this storm would have on the east coast. But… there is one catch. While both models did forecast sandy correctly, the EURO forecasted this storm almost a week in advanced while the GFS took its grand old time and took about 3 days later to finally “agree” with the EURO. ( ) But it does not stop there… Just nine days after “Superstorm Sandy” clobbers the east coast, yet another storm forms and it’s a nor’easter! But this time, instead of rain and wind, we get very heavy snow and along with some wind. Now this is where things become interesting. Both models did depict where and when this nor’easter was going to set up. But what comes next could be one of the worst model forecasts in New England’s history! Not one and I mean one of ANY of the several computer models which include both the GFS and the EURO forecasted a FOOT of snow for the New Haven area! Not only did this area get a lot of snow, this area had the highest snowfall totals of the entire storm! Some readings were as high as 13.5 inches in Clintonville CT! The forecast only called for a light coating if that. So it’s safe to say that weather is and will never be an exact science. ( )

         So after hearing all about how weather forecasts are made and the main two forecasting models used in weather forecasting in the U.S., it could be concluded that there is a lot of conflicting evidence that goes into deciding which model is more accurate. The bottom line, there is NO “declared winner”. There are so many variables that go into forecasting major weather events such as snow storms, hurricanes and nor’easters. These models are among many other models that forecasting these weather events. Besides that, these are the most reliable models out of them all. But it is too hard to decide which model is better due to the fact that weather is just so unpredictable and can change within seconds. Its weather and it’s a science of no other kind!

Use your ← → (arrow) keys to browse more stories.

Next Article


comments powered by Disqus

Zachary Duhaime

Town: Hamden, CT  

Reporting for WXedge since February 2012.

Articles: 78

Zachary Duhaime's Bio

Become a WXedge become a contributor

Let Your Voice Be Heard

Have a question? A comment? A complaint? Our team is here to service your every need. Go ahead, let them have it.