TAF versus MOS?

AcroBoy

AcroBoy

Line Up and Wait
Joined
Jul 18, 2008
Messages
649
Display Name
Display name:
Jim N
The TAF at my local airport is calling for 300ovc while the MOS is saying 700ovc for the same time period, both around 8 hours from now. This means an approach to 100 feet over ILS mins, or more options like a GPS approach or maybe a chance to circle.

What's the difference in how these forecasts are calculated? Is one more accurate than the other?
 
I don't know your answer but I had to look up MOS. I didn't know they even existed. Thanks for teaching me something. I try to learn something new everyday so, now I'm done, so it's back to sleep until tomorrow! :)
 
In my experience, the TAF is much better, provided you're actually at the airport in question. MOS interpretation in Foreflight seems very much a work in progress.
 
I like to stick with the TAS for now. As a planning tool ,it may be behind the forecast at times.
 
If I remember correctly, when planning for an IFR flight the TAF is the forecast you are supposed to use, not the MOS.

The TAF is good for a certain geographic area around an airport (10 nm, I think), and actually put together by a real live person. An MOS is a statistical model that is created to cover unmanned airports and provide some weather forecasting in more remote areas.


Sent from my iPad using Tapatalk
 
I've seen the MOS in FF. I've always stuck with the TAF's.
 
The TAF in my local area is not as accurate. NOAA site has a better data.
I find the MOS more useful for longer range looks.
 
MOS has problems in mountainous terrain. I've seen it predict low IFR on a severe clear VFR day due to assumptions it makes about dewpoint.
 
Here are some articles from Scott D. at ForeFlight about MOS. For IFR alternate requirements, TAF is the official source. MOS is an "another-tool-in-the-toolbox" feature. It's interesting to compare the two and watch what actually happens. Sometimes MOS is more correct than the TAF, and sometimes it forecasts more positive weather. One of the things I like about it is that updates more often than the TAF and the forecasts go out further than TAFs.

http://blog.foreflight.com/2014/12/14/foreflight-mobile-expands-forecast-weather-with-mos/

http://blog.foreflight.com/2015/08/28/gfs-mos-extended-ceiling-and-vis/

http://blog.foreflight.com/2015/08/31/gfs-mos-forecast-update/
 
Nice to see you posting again, Scott!
 
scottd said:
MOS is tuned for the specific airport. It's designed to know about the airport's location as it relates to terrain, bodies of water, urban vs rural. That's the statistical aspect. It derives the forecast for the specific weather event from a forecast model. I think you'll find MOS to be excellent guidance most of the time especially for airports without a TAF and for guidance beyond the TAF's valid time. Historically, TAFs have outperformed MOS especially in the first 6 hours or so. I really hate to hear pilots complain about MOS or TAFs when they really don't have a representative verification sample.
Click to expand...


MOS has been consistently awful for mountain airports around here Scott. It actually surprises me that you say it's tailored to the airport itself and terrain.

I believe a while back I posted examples where it predicted low IFR for an entire VFR day at Salida, CO.

Granted you have access to a larger sample set than my "check it once in a while and notice it's horrendously wrong about mountain airports" view, but it's worth taking a peek at that data if you have access to the historical stuff, and feel like it.

The underlying model doesn't seem to know anything about adiabatic and katabatic winds nor take them into any account at all, for one. And I really don't expect a wide area model to do so, so that's no skin off of my nose at all. Just reporting it. But without taking those into account, it gets totally lost as it churns through dewpoint data and tries to predict ground fog or similar.

From what little I've looked at it, it seems to handle flatland wetter locations much better. I'd certainly look at its predictions there. But Salida, Alamosa, Buena Vista, Aspen, etc? My general peeks at it from time time time find that it's way off, far more than it's right during changing weather patterns, when it counts.

(All the forecasts and models do great when it's just continuously dumping snow. Haha. Of course. Anything widespread and static, they all tell the same tale. It's the in between where MOS seems to fall apart at high elevation valley airports.)

Like I said, not a complaint. Just an observation.
 
Local conditons trump forecasts.
In my particular case both the AF/TAF and MOS are often way off. KHYX is 13 miles South of Saginaw Bay. I live 21 miles due West of the airport. The TAF and MOS can show VFR or MVFR conditons and I can leave the house in nearly CAVU conditions. As I drive East and enter the air blowing off the bay (wind from NNW most winter days) the ceiling and visibility start coming down. BY the time I reach my hangar it is strictly IMC.
Now these are known local conditions and we expect them (called Lake Effect) but a distant pilot using an AF or TAF or MOS will be in for a surprise.
The saving grace is that we have an AWOS on the field and the alert pilot listening from 40 miles out will know to change plans (or air file)
 
Has anyone done a statistical correlation between predicted and actual weather for mos versus taf at a variety of different locations?

Same question for time of forecast: e.g. mos is more accurate a specific time points, but TAF at others?
 
Remember MOS is an automatic model-generated forecast. Its useful for a longer weather outlook, or for airports that are located far from an airport with a TAF.

TAF is always going to be your best bet, if there is one at the field or very close by. A TAF is issued by a forecaster, who looks at all the information available to him (including the MOS) and then writes the forecast.

Its been my limited, unscientific observation that the MOS is generally more pessimistic than TAF in the Southeast. I can remember a few flights being concerned as the MOS is showing LIFR, while the TAF is reading ceilings a few hundred feet higher. In those instances, I considered the TAF as my best source of information and went flying.
 
scottd said:
I would encourage any pilot to contact their local weather forecast office via e-mail and arrange a tour of their office. Tell them that you are a pilot and you'd like to sit in with the forecaster on the short term desk...they issue the TAFs. Now if there's a busy weather day, they may not be able to spend much time with you explaining how they construct the TAF, but it'll give you a good appreciation for the kinds of issues they face dealing with the local weather.
Click to expand...

How cool. I didn't know this was available. For others interested, on weather.gov, you input your local zip code for local information and at the bottom of that page is contact information for your local office (mine's in Monterey).
 
Last edited:
scottd said:
weather.gov, not weather.com. However, be mindful during any visit that you are essentially an interruption to their day. So, they may have to cancel at the last minute if the weather is causing an "all hands on deck" response.
Click to expand...

Oops. Fixed.
 
Sam D said:
How cool. I didn't know this was available. For others interested, on weather.gov, you input your local zip code for local information and at the bottom of that page is contact information for your local office (mine's in Monterey).
Click to expand...

And if you do want to visit it, the NWS office in Monterey is located real close to KMRY. Monterey Jet Center is the closest FBO (though Del Monte is nicer, and the on-field restaurant is on the other side).

KMRY is just about the sleepiest Class C I've seen. Truly trivial, albeit a bit expensive. Watch the TPA. It's high.
 
Amen on the expensive part. Both of the FBOs at Monterey have raised their ramp fees for small singles to $45, unless you buy 15 gallons of fuel (which is hard to manage in a Skycatcher).
 
Palmpilot said:
Amen on the expensive part. Both of the FBOs at Monterey have raised their ramp fees for small singles to $45, unless you buy 15 gallons of fuel (which is hard to manage in a Skycatcher).
Click to expand...

Yeah, but Del Monte will be satisfied with a "top off."

BTDT. You would have to try to burn 15 gallons getting there from Palo Alto, let alone Reid.

It's expensive fuel. Not Oakland expensive, but expensive nevertheless.
 
MAKG1 said:
Yeah, but Del Monte will be satisfied with a "top off."
Click to expand...

Thanks for the tip.

BTDT. You would have to try to burn 15 gallons getting there from Palo Alto, let alone Reid.
Click to expand...

I wouldn't have to burn off as much, because my club doesn't keep the Skycatchers topped off.
 
N2124v said:
The TAF is good for a certain geographic area around an airport (10 nm, I think)
Click to expand...

5 statute miles from the center of the airport. They're basically point forecasts, especially if the terrain is not uniform in the area.

Personally I like tools like MOS for longer lead times more than as a substitute in between TAFs. I wouldn't base a flight to minimums on it, but then again I wouldn't trust a TAF with my life either.

MOS is most useful to me for getting my backup plans in order. I use the MOS 5 day graphical products, which aren't going to be precise by any means, to judge where the fronts are likely to be and how the temperatures/dew points will be trending. That can help me get a ticket on Southwest before the last minute.
 
I have had really bad luck using MOS like that.

It seems to predict substantially better local weather than really happens. Like, Sunday, it predicted 4700 foot overcast around the Santa Clara Valley. The ceiling was actually broken at 1000-1200 around the region (consistent with the SJC TAF). And once the front came through, the icing AIRMETs soon followed.

That's wrong to the point of uselessness.
 
MOS is just the raw output of a computer model. Treat it as such. Useful for longer term outlook, but not for decision making.

TAF is curated by a human. Still not perfect but published with the intention of being used for decision making.
 
You must log in or register to reply here.
Back
Top