GFA cloud top map
5 min read

Rumors have swirled for years, but now it’s really happening: the text-based Area Forecast (FA) will officially disappear on October 10, 2017, to be replaced by the Graphical Forecast for Aviation (GFA). On the surface, this seems like an inevitable step in the transition from coded text products to graphical, interactive weather maps. But before we relegate the FA to the dustbin of history, we should consider a few important details. This transition may not be quite so innocuous.

Area Forecast

The Area Forecast isn’t pretty, but those cloud top forecasts will be missed.

The FAA is right about one thing: there’s a lot to dislike about the FA. With six huge forecast areas for the US and only three updates per day, it lacks precision. It’s also difficult to read, with dozens of obscure contractions, all capital letters and significant limitations on the number of characters that can be used. It is a weather product made for the teletype era, not the internet era.

These limitations have been clear for decades, and the decision to replace it was hardly made overnight. Scott Dennstaedt, a former National Weather Service meteorologist, points out that the FA-to-GFA transition has actually been going on for close to 15 years (Canada has had a GFA for many years).

In the US, the modernization push started with AIRMETs. By eliminating forecaster lingo (like the once-popular “ICIP”), the text for AIRMETs was standardized and a new graphical AIRMET map – the G-AIRMET – was born. This product is now operational, and will be coming soon to ADS-B broadcasts and probably SiriusXM data plans too.

It’s a major improvement for pilots in both accuracy and ease of use. The old AIRMETs were “time smeared,” trying to account for potential weather conditions over a 6 to 12-hour outlook period by averaging conditions over that time. This led to wildly pessimistic reports and earned the nickname “ScareMETs.” The modern G-AIRMET is issued for smaller areas and at 3-hour intervals – with conditions valid at that issue time instead of the entire forecast period. This practically eliminates the time smear issue and makes it far easier to interpret. You might find yourself actually looking at AIRMETs again.

The AIRMET upgrade seems like a perfect test case for the text-to-graphical transition: a confusing, dated forecast was replaced by a modern, easy-to-read chart. Copy and paste for the GFA, right?

There’s just one problem: the new GFA is not a reliable replacement for the FA yet.

While there are many nice features on the new website, the cloud tops report is a major step backwards compared to the FA. Compare this layer to a satellite image or PIREPs (or the FA) and you’ll find a huge difference. While the GFA map might show “TOP 080,” the actual cloud tops might be above 25,000 feet. That’s a major problem for a Cirrus pilot trying to escape ice, and there are no other widely available options for finding cloud top forecasts.

Why the discrepancy between the old and new forecasts? The GFA’s cloud top forecast is derived from a computer model (the Rapid Refresh or RAP), and it’s almost 100% automated. This model is fairly accurate with stratiform clouds, but it’s not good at all for deep, moist convection like you find around thunderstorms or fronts. The old FA, in contrast, had lots of human input to improve accuracy.

The FAA’s own Aviation Weather Services publication explains the problem: “NWP [numerical weather prediction] models are much more accurate forecasting winds aloft than they are forecasting cloud tops. Pilots should not expect an equally high degree of accuracy with these forecasts, especially in areas where there is a high degree of variability in the forecast cloud top heights.”

GFA cloud top map

The GFA cloud top map looks great, but it isn’t very accurate right now.

Those sentences calls into question why the FAA is making this transition – if pilots can’t trust the cloud tops report, why is it the new standard? Increased efficiency and reduced costs are the most likely explanations, but those will certainly appeal to policymakers much more than pilots.

We need to make clear that the current GFA is not meeting general aviation pilots’ needs. The National Weather Service has a roadmap of upgrades to the GFA, so with some pressure the accuracy of the cloud top forecast might improve over time. At this point, the best option would be to introduce some human forecasting skills back into the equation.

There is some precedent for this. The Collaborative Convective Forecast Product went from human-only to automated and then to automated with a way for humans to amend it. Now it’s back to being a predominantly human-created product, and it has become an excellent tool for planning during thunderstorm season. Perhaps the forecasters’ time saved by the elimination of the FA could be applied to improving the GFA.

In the meantime, pilots should be very careful with the GFA tool. In particular, consider what products are observations and which are forecasts – back up a forecast with a look at current conditions. A new cloud top product is also coming to ADS-B, but it’s based on the Rapid Refresh as well, so skepticism is in order both during pre-flight planning and in the cockpit.

The National Weather Service proudly says, “Aviation weather users have found that pictures are worth a thousand contractions.” That’s absolutely true, but only if those pretty pictures are accurate. For cloud tops, at least, they aren’t.

John Zimmerman
19 replies
  1. Ed
    Ed says:

    I was unaware of the overall inaccuracies of the GFA cloud top forecasts, although I definitely noticed that it would show tops at 6000 right in the middle of a thunderstorm! Thanks for the info. One item missing from the GFA, which was available (sort of) in the FA was the position and movement of fronts. I still have to pull up the Surface Analysis and Prog Charts to really see the “big picture”.

  2. Scott
    Scott says:

    One other shortcoming of the GFA is that forecasts end at the U.S. border. Understandably, the GFA was designed to replace the FA which didn’t extend into Canada or Mexico. However, other forecasts like icing and turbulence were not in the FA, but the GFA cuts those off at the border. Personally I believe cutting forecasts in this way is a poor decision for pilots flying along the borders. It’s hard to get a sense of trends of weather as it moves into the U.S. The data comes from models that have a North American domain so there’s no real excuse not to include it.

  3. Tom M
    Tom M says:

    The Skew-T diagrams (Op40 / Bak 40) might be the only place to go for cigs / tops as indicated by the Red & Blue lines converging.

    • Scott Dennstaedt
      Scott Dennstaedt says:

      Tom – Unfortunately, it’s not that easy. Much of the cloud cover out there is produced through ascending air…the dewpoint depression is a great way to determine the vertical extent of stratus-type clouds, but not cumuliform clouds. Also, the GFA uses the same model as the Skew-T (Op40). The GFA tends to do somewhat well when there are active areas of convection producing precipitation. But, if there’s convection blossoming, I’ve seen the GFA tops forecast be way off. And if there’s an area of towering Cu that is not producing any precipitation, I’ve found that to be problematic as well.

      • Pat
        Pat says:

        Scott-I am really liking your weather courses, that said it is very useful to know the top of horizontal layer even in the presence of connective activity. Staying out of the horizontal stuff is essential to being able to use view out the window to avoid the vertical stuff & is helpful in avoiding long term exposure to icing conditions.

  4. John Swallow
    John Swallow says:

    “Try it; you’ll like it!” comes to mind.

    Been using them for years up here and find them head and shoulders above that which they replaced. I think the majority of GA in the USA will come to that conclusion.

    Fly on!

  5. Frank Ladonne
    Frank Ladonne says:

    You are (inadvertently I’m sure) making the case for privatization of ATC. The FAA is painfully slow and ponderous and thus has, in all aspects, left us with 1950’s technology.

  6. John Swallow
    John Swallow says:


    If your comment “you are …. making the case for privatization of ATC” was in response to my comment, I hadn’t thought about it that way. I think the GFA would have come in regardless.

    We’ve had a private, not-for-profit ATC up here for years as you know. When flying commercially, I had no quibbles with it; the same holds true now that I’m flying privately. True, it costs me about $50 USD/year, but compared to the cost of AVGAS, that’s peanuts.

    I admit that I don’t understand all the ins and outs of the FAA ATC system, but I confess that I’m a bit mystified about the vehement resistance to the privatization of the ATC in the USA. We had a fairly smooth transition to privatization; maybe because the only thing that really changed was the name. On Sunday night, people were working for Transport Canada; when they went to work on Monday, they were working for NavCanada.


    • Robert Thomas
      Robert Thomas says:

      The fear is not that privatizing the US system would end up like Canada, most I think would be okay with that. The fear is that we end up with a system more like the European model. I’ve yet to meet a European pilot that has good things to say about flying there. The comments run from it sucks, to tolerable. But good never is mentioned….

  7. Frank Ladonne
    Frank Ladonne says:

    Precisely, John.

    Lately there’s been a HUGE “kerfuffle” in the US about ATC privatization. Fear of the unknown. I cannot IMAGINE how a private company could manage ATC WORSE than the FAA has and the fact that it has taken YEARS, MANY years to upgrade weather reporting for pilots from the Stone Age proves the point.

    • David Megginson
      David Megginson says:

      We’re hijacking the comment thread here with a different topic, but I’d still like to add that over the past 15 years, I’ve had almost nothing but positive experiences with both US socialised ATC and Canadian privatised ATC. Flying single pilot (especially IFR), I consider the controllers to be partners in my flight, not obstacles I need to overcome.

      I guess that whatever the management structure up above, the women and men at the radar scopes just shrug, put on their headsets, and focus on providing us with good, professional service.

  8. Frank Ladonne
    Frank Ladonne says:

    Agree. The “folks behind the scopes” are the best. The management bureaucracy, however-Stone age. The embracing of new technology- the force of a shadow and the speed of a glacier.

    You fly a lot of IFR; me too.

    Make sense to you that we put in a BUNCH of time planning a route, file it and then the clearance is altogether different. Okay, then we launch and the route we are told to actually fly is altogether different again. REALLY?

    Need to get Neanderthal FAA out of this business NOW and usher in 21st century technology.

    • David Megginson
      David Megginson says:

      I get frustrated with the routing changes too, but I also understand the challenges of trying to route a slow plane like my Cherokee through busy terminal airspace, especially when they don’t know how bad things will be when we finally arrive in the sector. It must be like trying to route an old person with a walker across the ice in the middle of an NHL game.

      Sometimes I think they’re optimistic and give us an initial clearance for the best case scenario, then have to take it away in flight when things get to crazy.

      It’s getting slightly better now that we can see our initial ATC routings online (or in mobile apps like Garmin Pilot) an hour or two in advance. At least we know what’s coming, even if we can’t do much about it.

  9. Jim
    Jim says:

    I realize this is a little off topic, but on the FAA privatization, there is more behind the scenes than people realize. I work for the Friendly Aunt Agatha, and when the privatization issue came up last time, it never made it out of committee. The Controllers Union was all behind it, but the bill did not have a “No Strike Clause” in the proposal. It almost made it through until a very astute group noticed that fact, and the proposal died. The other issue here in this proposal, is the the major air carriers will have a significant seat on the new “corporation”. Can you imagine how well this new corporation will run if the air carriers can’t even manage their own business to the satisfaction of the flying public? All one has to do is turn on the news and see all the stories of air carrier bad behavior. This proposal right now did take out the user fees for GA, so Avgas will still be high, because of the taxes. Maybe Canada has had success with the privatization, but at the risk of sounding prideful, the US system is much larger, much more complicated and thousands more flights daily. The US system is getting modernized, and that fact is in operation now. ADS-B, GPS, and many other new enhancements. There will be growing pains, and a new “corporation” will have to deal with that as well. The lack of modernization lays part of the blame on the air carriers. With a 2020 mandate the air carriers are not ready. They have a significant amount of the fleet to upgrade. How well would this be accomplished if they were self managed?

  10. Cary Alburn
    Cary Alburn says:

    Getting back on the original topic, the lack of cloud top accuracy in the graphic presentation shouldn’t be of much concern to Mr. & Ms. Average GA Pilot, because it just doesn’t matter to the majority. What percentage of the GA fleet can top anything that is building? What percentage of the fleet, for that matter, is oxygen equipped? Or pressurized? I submit that most of us will fly as we always have, slogging along either VFR under the clouds or IFR in the clouds below altitudes that require us to put on oxygen.

    For the relatively small percentage of the fleet which can fly above most weather, the only real way to know the tops has always been to climb up there and take a look, or ask for PIREPS from those who are already up there. So in that respect, there’s nothing new.

    So in total, it would appear that for us average GA folks, and really everyone, the graphics will be a major improvement.

    • David Megginson
      David Megginson says:

      What Cary Alburn writes is true for convective activity, but in the winter, it is quite possible for our little planes to fly above lake-effect weather SE of the Great Lakes. Those clouds typically top out around 7–8,000 ft, and can be full of ice. Sometimes they appear as solid decks, and sometimes, as long streamers following the wind direction.

      If you have passengers who suffer from motion sickness, it’s also good to know if you’ll likely be able to get them up above normal cumulus into clear, smooth air.

  11. DO
    DO says:

    Your point about cloud tops is well taken. Another concern is legally filing IFR to a a non-TAF destination.

  12. Vicki B.
    Vicki B. says:

    Even the bases are wildly inaccurate. Over the past few months, I have been comparing the two. The area forecast with human intervention is much more accurate that the GFA. Sky clear all day is what the GFA predicts for the MidAtlantic during the summer months, but the usual pattern is for some small cumulus around with 3000 to 5000 ft bases. Someone unfamiliar with the weather patterns in an area could get a surprise.

    As for avoidance of icing, you need to be able to predict where the visible moisture (clouds, rain) will end. Otherwise, yet another nasty surprise.

Comments are closed.