FFF
Have you asked the UKMO about this apparent discrepancy? It's worth asking. They would need real examples.
If you can demonstrate objectively (with long term data) that you can generate more accurate TAFs than the existing ones, please do so.
You mention F215. That one is like the old joke about British Standards - they are wonderful, so many to choose from
On any given day and for any of the areas listed by number, there is usually a choice of very different conditions - for the same spot.
Which is not to say that TAFs are particularly accurate. They aren't. Chaos will never be abolished - except when the weather comprises of a high pressure that has been in place for last few weeks and doesn't look like it will shift (summer 2003 perhaps) in which case anybody can do it well enough themselves by looking out of the window and looking at the MSLP chart.
Incidentally, I am pretty sure that virtually all of the free weather sites display data which is either directly from GFS, or is lifted from other websites which use GFS. A lot of "lifting" goes on in this business, with websites getting data off one another. The NOAA site I gave before had put in a graphic thingy to stop this. Nobody running a free website is going to put data online which they pay money for to the UKMO. The only exception I can think of is stuff which the UKMO charges for but which it is obliged to distribute under international agreements, and which some foreign (typically American) very kindly put online
Obviously you know this but it's worth stating in case somebody has a go at me. In the end, forecasts will never be accurate enough to rely on, so a Plan B is always needed. Having an instrument rating of some sort provides additional options there, and even more options if the flight is carried out IFR to start with.
p.s. I think an airfield needs to generate just *one* METAR before the TAF comes out, but could be wrong.
Foxmoth
This sort of fast moving weather is not forecastable more than a day ahead, I think.