I find that TCs losing their deep convection tend to have the winds "lift" well above the boundary layer, and as a rule, find that assigning a 60% to 80% reduction ratio from h70 (based mostly on a blend of taking the amount and intensity of convection into consideration, along with persistence, and the surrounding surface pressures) actually works quite well. A cyclone with as weak as convection as Bertha now has could easily be running as low as 70% at the surface, adjusted from 700mb, but 80% is probably as good as any right now, accounting for the nearby location of the shortwave ridge, and overall MSLP.