Temperature prediction is, in my experience, pretty good.
The 1200Z Obs seem to be missing, but the soundings seem to show the freezing level above FL100 over England:
Herstmonceux
00Z 04 Nov 2010 sub-zero above 670 hPa (1.2 degC at 700 hPa)
00Z 05 Nov 2010 sub-zero above 670 hPa (2.4 degC at 700 hPa)
Nottingham
00Z 04 Nov 2010 sub-zero above 680 hPa (0.8 degC at 700 hPa)
00Z 05 Nov 2010 sub-zero above 690 hPa (1.4 degC at 700 hPa)
Albermarle
00Z 04 Nov 2010 sub-zero above 770 hPa (2.2 degC at 850 hPa)
12Z 04 Nov 2010 sub-zero above 740 hPa (5.0 degC at 810 hPa)
00Z 05 Nov 2010 sub-zero above 780 hPa (4.4 degC at 850 hPa)
There's a difference between trusting freezing level forecasts where the minimum level is dictated by terrain, and trusting them where the minimum level is dictated by airspace or ATC needs. In this case, if the freezing levels had been lower than anticipated, you might have had to go below controlled airspace or abort the flight.
I'm not suggesting that this would have been a pleasant flight, and whether I would have attempted it depends on the circumstances and alternatives. But you did ask "how", and I still believe it would have been better at FL90 than at low level.