"Brooksig" suggested that he had complete faith in the Met models.!!!
No I didn't!!!
In the absence of any alternative I just asked you for your evidence of inaccuracy, within the limits of the modelling as offered. I have absolutely no doubt that the models ARE less accurate than 'spot' measurements using a probe on an aircraft but erring on the side of safety.
As has been pointed out already, actual measurement of all the FLs and all areas surrounding any particular airport just cannot be done. For a start, you can't easily add extra (recce) aircraft to the traffic already there!
And, finally, what use would 'the equivalent of a METAR' be at the start of a 7 hour flight? Modelling should give a usable forecast of the VA situation on arrival at (eg.) LHR - actual measurements from 7 hours ago would be useless. The key problem is how often does the actual track of an ash cloud vary from the prediction: in other words, how often is the short-term weather forecast incorrect? You can see an example of the model being 're-trimmed' during today for 0600 tomorrow: the southern boundary is now further south than predicted 8 hours ago.
In practice, what improvement of accuracy / sensitivity could be achieved or is needed?