Airborne_Again wrote:
Not the cloudbase, but poor visibility could make it illegal.
It doesn’t say anywhere in the regs. that you must have forward visibility, all it says that visibility should be Xkm. You can have it sideways and still be legal.
I like to see that you appear to have developed a decent idea about what model does what well and which does not. That makes a lot of difference to many people who take models as 100% accurate and bitch if they are off, particularly long range.
The one bit which I really have absorbed in 15 years of weather service is how different the models are and how much work and constant observation it takes to halfways know how to interpret them. Many of the meteo automatic sites don’t do that justice, even less so because most of them only use free models such as GFS, which is a global low res model. GFS has its strenghts and weaknesses like many others, it is pretty good to see large weather phenomena but it is unsuitable for local forecasts, neither was it intended for them. However, their wind forecasts are usually quite ok if you go for mid-long range. I used GFS in its older form (less accurate) for a heavy irons flight planner (mainly MD11, A330 and that size) on long haul and it usually generated very accurate forecasts even then.
High res models in Europe are often not available for free and therefore are not getting used by the products which would improve massively from their use.
The simple thing to know about forecast models is that without a good interpretation guide they are very much less usable than if you know them well, have read about where they are strong and weak and possibly get some first hand experience on them. Quite a lot of batchelor and master thesis’ are written on single parameters within model forecasts, which should show the magnitude of the problem of interpretation. I usually walk away from our instruction days where models are discussed, how they have changed and updated with a serious headake. It also is noteworthy that often enough models which have been dead stable on something all of a sudden change with the update of one or two parameters, in the course of development. That is when the fun really starts as improving e.g. rain pobability may screw up other things.
For local forecasts nothing beats a good forecaster. And that means, most local forecasts of weather apps are to be taken with a barrel of salt, unless they are accompanied by a proper worded forecast. The Meteo Swiss app does this rather well for Switzerland, it combines automated forecasts with the text forecasts issued by the forecasters. For me, the text forecasts are one very important element which many today disregard. To get a proper picture however, they are quite important too.
Cobalt wrote:
Is this still the case? I thought that went long ago.
It’s still the case, and not the case, according to SERA:
The definition of Special-VFR is “a VFR flight cleared by air traffic control to operate within a control zone in meteorological conditions below VMC”
The condition for SVFR is ceiling below 1500 feet or less than 5 km visibility.
The minima for VFR (in F, G) is clear of clouds, 1.5 km visibility and visual ground.
The minima for VFR in controlled airspace is 5 km visibility and 1000 feet vertical distance from clouds.
SERA does not define any minima for SVFR itself, only the condition for where it is needed (what is no longer VMC in a control zone)
One more thing: really heavy rain can block the air filter. I have never seen this on the TB20 but have heard of it. Alternate air may be recommended in the POH for heavy rain. I tend to just watch the MP gauge carefully.
Mooney_Driver wrote:
100% accurate and bitch if they are off, particularly long range.
A 100% accurate model for weather will be unstable as hell over time (it gets accuracy from frequent calibration/parameters estimation not much from underlying models physics or computer power) and will definitely fail for forecasts on large space ranges/time horizons.
I guess what will tend to generally work is a local weather model (e.g. account for local features, say Alps or just hills/lacs behind) that gets in-sync with a global engine (e.g. GFS like) while integrating with its reliable data (e.g large scale wind/pressure gradients) on large scales, a bit like having specific local add-on for GFS, but the problem with this fine-tuning is always the question of which variable you want to work well (while others may blow-up !!)…
For pilots, it will be good to have something that is fine-tuned by weather specialists to give something like an accurate binary signal of VMC/IMC for flying usage, but then it gets crazy when you are close to the boundary unless you relax some of the accuracy (in practice, IMC/VMC is far than just black-white as the rules says, this is the cause why most of us may push into mVMC/deepIMC without even noticing)
Of course, you can’t judge all of it by a simple “good/bad” as we pilots do (the guys who write physics model/simulation code as you mention can write you zillions of pages on comparing models accuracy, stability and forecasting power plus a whole doc package on uncertainty in parameters calibration/estimation…)
In my previous/current engineering jobs (aeronautical/finance), I always get puzzled by the large gap between the low level guy who does model equations/simulation code and the top guy who is interested in something rather binary (will those jet engine blow up between inspection? will we lose money beyond X?) without going trough those 500 pages documentation, at least we should expect a confidence level/probability before defining a realistic set-up for various tweaks/knobs to get a proper answer…so aviation is not an exception neither
Just ask how many pilots have read and looked in details at aircraft POH performance numbers and compare it to actual performance of their aircraft for their specific usage? Most of us will just blame it on ISA or DA, even tough they end-up with a 160hp engine instead of 180hp due to fishy QA 
Ben wrote:
It doesn’t say anywhere in the regs. that you must have forward visibility, all it says that visibility should be Xkm. You can have it sideways and still be legal.
Sorry, but this is completely wrong.
The visibility requirements refer to “flight visibility” (SERA.5001). “Flight visibility” is defined as “the visibility forward from the cockpit of an aircraft in flight” (SERA definitions §80) — so it is not even just the meteorological visibility in the forward direction. A dirty or badly scratched windshield can reduce the flight visibility. This is nothing new either — it has “always” been like that. (“Always” meaning at least since I started flying in 1983.)
VMC minima are really basic stuff. One of the first things you learn in air law TK.
LeSving wrote:
SERA does not define any minima for SVFR itself, only the condition for where it is needed (what is no longer VMC in a control zone)
Sorry to be so direct, but this is wrong. SERA.5010(a) – SVFR minima are clear of cloud, surface in sight, 1,500m (800m if helicopter), also speed limit 140kt.
Also, that is not what I meant – genuine SVFR has lower minima than VFR for the same airspace, by definition.
But in the UK there was some different kind of “SVFR” clearance, for example the London CTR, which was class A, and required either an IMC rating or 10k+ visibility, and there was a licencing restriction on PPLs without IMC rating that imposed higher minima than the airspace minima in certain “notified” CTRs
I thought these specialities were now all gone with FCL removing all restrictions based on the licence, and SERA/the reclassification of the London CTR removing them for the London CTR. Are there any left over?
Cobalt wrote:
Are there any left over?
These has been explicitly removed from the ANO in 2016 edition for those who holds UK PPL and NPPL to be aligned with what SERA permits (including that negative 300m gap between 1.5km and 1.8km for take-offs of IMC holders) , so as far as FCL is concerned the restriction has been removed
The caveat is that legacy SVFR minima still appear in ATC procedures regrading specific zone transits, while the ANO2016 leave door open for not issuing that clearance if it is not aligned with ATC/CTR requirements, so I guess it is just a matter when ATC ops gets updated to SERA?
See,
https://www.caa.co.uk/Blog-Posts/The-revised-Air-Navigation-Order/ local copy
Usually, you will see a lot of conflicts between what rules applies between operator (Part-CAT/NCO?), Aerodrome/ATC (Part-ADR.OPS) and licence (FCL) as these migrations goes around
Airborne_Again wrote:
Sorry, but this is completely wrong.
The visibility requirements refer to “flight visibility” (SERA.5001). “Flight visibility” is defined as “the visibility forward from the cockpit of an aircraft in flight” (SERA definitions §80) — so it is not even just the meteorological visibility in the forward direction. A dirty or badly scratched windshield can reduce the flight visibility. This is nothing new either — it has “always” been like that. (“Always” meaning at least since I started flying in 1983.)
VMC minima are really basic stuff. One of the first things you learn in air law TK.
In this case I have learnt something new. Every day is a school day. Thanks.
Ben
Cobalt wrote:
Sorry to be so direct, but this is wrong. SERA.5010(a) – SVFR minima are clear of cloud, surface in sight, 1,500m (800m if helicopter), also speed limit 140kt.
So it’s the same as before then. Hmm
