The other day I was on a flight and got this GRAMET in the updated briefing pack, 30 min before EOBT.
In fact all over land, there were scattered Cu with bases of 3-4000 feet. In the general area around SOVAX-TRS there were isolated CB with TS. I didn’t encounter any cells, but there were plenty of layered clouds between SOVAX and ROGMI which were clearly remnants of dissolved CBs. I even got trace amounts of ice over the sea at FL100.
The “official” aviation forecasts (SWC, TAF, GAFOR…) all mentioned the CBs.
I agree that wind forecasts are amazingly accurate. Cloud less so.
BTW: Why doesn’t the magenta line include a descent to my destination? At other times I’ve seen the magenta line descend to ground level halfway to the destination! Is there some logic in this?
This really is a discussion about GFS, GRAMET is just the messenger here 
Those fair weather CU are very hard to predict and GFS is not really that good about it. However, they are usually not a factor for a go/no-go decision. Very frequently the local radio meteorologists are wrong about fair weather Cu in my experience.
Regarding the remnants of dissolving CBs, that’s probably a timing error. According to GFS the dissolution would have happened at that point already. The GRAMET is a point in time, a snapshot. To judge the certainty of the forecast, you have to look at the general weather situation and at a larger time frame. If you see much more cloud 2h earlier, then you get the hint that something is moving through and that timing can be a factor. I personally like to use the MSLP to quickly understand the nature of the weather.
Recently a user suggested that given we know the position of those CBs/TCUs on the GRAMET, we could generate the routing to go around them. Well, aehm, we could
Airborne_Again wrote:
BTW: Why doesn’t the magenta line include a descent to my destination? At other times I’ve seen the magenta line descend to ground level halfway to the destination! Is there some logic in this?
Looks like something is wrong. The GRAMET should terminate at ESSV really.
achimha wrote:
Regarding the remnants of dissolving CBs, that’s probably a timing error. According to GFS the dissolution would have happened at that point already. The GRAMET is a point in time, a snapshot. To judge the certainty of the forecast, you have to look at the general weather situation and at a larger time frame. If you see much more cloud 2h earlier, then you get the hint that something is moving through and that timing can be a factor. I personally like to use the MSLP to quickly understand the nature of the weather.Recently a user suggested that given we know the position of those CBs/TCUs on the GRAMET, we could generate the routing to go around them. Well, aehm, we could
Given the chaotic nature of CB formation, could GFS really forecast the formation and dissolution of individual CBs? ESKN, which was about 25 miles to the SW of my track at the closest point, reported CB with VCSH in the latest METAR in the updated briefing pack. In a METAR shortly before that it reported TS. I checked on the radar pictures before takeoff that this CB was actually dissolving.
Airborne_Again wrote:
In fact all over land, there were scattered Cu with bases of 3-4000 feet.
This may also be a bit of an issue with the way we plot convective clouds. Scattered means probably around 20%, and in order to approximate 20%, with our relatively wide (but already rather distorted looking) Cu icon we can only plot one or two Cu icons over the width of the full GRAMET. I’m open to suggestions on how to more intuitively and more nicely displaying Cu “bands”.
The basic issue IMO is that the GRAMET horizontal axis is very much compressed, compared to the vertical axis.
Airborne_Again wrote:
At other times I’ve seen the magenta line descend to ground level halfway to the destination!
That’s an artefact of the way level changes are handled. Level changes are supposed to happen at the beginning of the leg. That’s apparently ICAO and Eurocontrol expectation. For the last leg, it obviously doesn’t make sense. This will be changed in the next couple of days.
Well, as I said, you’re looking at a snapshot. If a front is moving through and GFS expects at 1400Z the last CBs to be gone while the front actually moves slower then you’re looking at a blue image for 1400Z. If take a GRAMET of 1300Z then there might be CBs.
Always understand what influences the weather to determine how sensitive the GRAMET will be to timing and things developing slightly different from how it was forecast. This is not always easy.
I remember an episode from the last Tannkosh festival. The super duper experts from DWD told all pilots that a cold front was moving in and it would be there by 1200Z after which no flying would be possible for the whole day because it would be strong one with IMC. What followed was a mass exodus where people were standing in line with engines running for 45 minutes and in the overall rush it was a miracle there were no accidents. Well, what actually happened was nothing at all — blue sky the whole day until the night. This was the top notch human forecasters with the best high resolution weather models, decades of experience and a forecast timeframe of only a few hours.
Don’t rely too much on weather forecasts, no matter where they come from.
Given the chaotic nature of CB formation, could GFS really forecast the formation and dissolution of individual CBs?
No; completely impossible.
Part of the problem is that GFS breaks up cloud types into low medium and high. This is based on the fairly distinct physical categories involved so it isn’t completely daft, but if you are generating a 3D model why have this division at all?
GFS totally fails to forecast quite a lot of cloud. It is quite good at forecasting generally convective conditions, which one can then depict as TCU/CB according to the various parameters such as lapse rate, water content, etc. But the only thing which tells you what is actually present is radar and IR. Sferics just tell you when things get really horrible – like a stormscope 
achimha wrote:
Well, as I said, you’re looking at a snapshot. If a front is moving through and GFS expects at 1400Z the last CBs to be gone while the front actually moves slower then you’re looking at a blue image for 1400Z. If take a GRAMET of 1300Z then there might be CBs.
Certainly, if you’re looking at frontal weather. But in my case it was air-mass CBs.
But in my case it was air-mass CBs.
I may think ATPL Met is crap but airmass CBs are formed when there is enough lift and enough water.
Their formation, or lack of, is a purely statistical exercise.
The only way to make it precise would be to have a wx model with say a 1 metre resolution but that will never be possible not least because there is no way to get the data with enough resolution.
The only even remote chance of getting the data is from satellite observations but those don’t give you a “probe” into the 3D atmosphere. They give you various proxies… same with ADS-B and Mode S emissions from airliners; if somebody is actually collecting that stuff, they give you accurate wind data (TAS-GS spread, and headings+tracks) for the various altitudes in question. But there is no way to get hi-res 3D data on temperatures and humidity.
Hence, radar, IR and sferics, used as preflight metars, rule the world
Quick follow up, it was quite a massive TS area that moved in over southern Sweden with frequent TS. Unusual that gramet didn’t show it.
