Today I diverted due to fog, ending up with (what will eventually be) at least £200 extra cost and a load of extra hassle.
The quality of forecasting fog (during "nice" high pressure conditions) is very poor.
Peter, did you go missed on the approach or not try? I agree fog forecasts are very hit and miss.
Apart from one case where there was a bunch of amended TAF's were produced and a big bank of fog covered the UK quite rapidly, I've generally thought fog was fairly well predicted. Early this morning I was at London City, and mist was forecast and it was dubious that the BA planes were going to land (though they did, though the Lufthansa's stayed at home) but I don't think the METAR's ever did change to FOG.
Was you flying early in the morning, or were you caught out later in the day?
I can't speak much for the cases to which this doesn't apply, but my home airfield is beside a body of water (a large one, but not the sea or ocean) and I have spoken to the forecasters who have said they are fairly good at getting it right that the fog will form over the water, but it's considerably harder for them to get it right when it comes to deciding if its likely to move inland. I assume this is to do with medium cloud cover, wind speed and direction and temperature, and a bit of luck.
Later in the day. The fog never cleared up, which is slightly unusual.
A fog forecast is basically a wind forecast.
I asked a meteorologist this very question recently. He said its very difficult to forecast when it will clear, but with a smile said you could use the month number (about 09:45 today), with a smile I asked 'Local or UTC' he said 'we don't qualify that because it gives us more room for error'.
In the sort of conditions we've had in the UK for the last few days, would one not just expect it as default each evening?
The variable is of course then how long it takes to clear in the morning (if it clears at all). I guess this comes down to whether it gets hot enough for the sun to burn it off, or if there is any significant wind to clear it.
This is stating the obvious but cloud forms when temp=DP, and it is this convergence of temperatures close to the ground that seems very hard to forecast.
The convergence very rarely happens close to the ground, which is why clouds are nearly always some way up, and even fog is not right down to the surface. I think this is because the ground heats up the air and causes a slight temp-DP separation.
Yesterday's TAFs for EGKA showed the fog forecast to clear, until they finally capitulated when it was totally obvious it was not going to happen. So they clearly had no idea.
There must be a separate weather model for this, because GFS cannot even forecast low cloud usefully.
The problem is that when the effects are somewhat localised, models are hard to create.
Fog clearance surely requires either warming (perhaps if there is higher level cloud the sun doesn't warm the ground sufficiently for it to clear) or wind to disperse it?