Michael,
I think one of the primary reasons to update the F scale, in the form of the EF-scale, is that it's now thought that much weaker winds can to do the damage that we've previously associated with stronger winds. For example, to produce F3/EF3 damage, you probably don't need as high of winds as was associated with the F3 range; thus, for the same type of damage, the EF3 rating dropped the winds likely associated with that damage. Remember, from a pure damage perspective, there is little difference between the old F and the new EF scales. In fact, the new EF scale has many, many more damage indicators and a more precise "degree of damage" (DoD) subscale that, theoretically, will result in more accurate classifications. It seems probable to me (though I'm no expert on this) that the winds speeds associated with the old F3-F4 levels may well produce damage that is actually high-end F4 and F5 in nature. As such, if we had kept the wind speeds, then adjusting the "damage" levels associated with F3-F5 would essentially homogenize those classes. It seems as though it was determined, however, that it's best to keep the damage associated with the old F and the new EF scales the same across each number, which required the adjustment of the wind speeds.
To me, and this is complete conjecture, it seems more likely that may more tornadoes actually have winds that can do EF3-EF4 damage (again, same damage as F3-F4, it's just the wind speeds have been adjusted downward). This has been talked about a bit in the past, but I know Curtis Alexander is looking at wind speeds in mesocyclonic tornadoes in supercells in the Plains, and his work suggests that the previous notion that "F0 tornadoes are the most common, with an exponential drop-off with increasing F number" is incorrect. This, in turn, may mean that many tornadoes are capable of producing EF2-EF3 damage if they hit something that can be rated as such. For the majority of tornadoes that stay in fields and/or away from substantial structures (damage to which is required for the higher damage ratings), however, it's difficult to rate anything above the lowest F/EF classes. So, perhaps it IS quite common for mesocyclone-produced tornadoes to have 120-150 mph winds at 2 meters AGL (i.e. perhaps it IS common for tornadoes to have the potential to create EF3 damage if they hit something substantial).
"Right. That's bad. Okay. All right. Important safety tip. Thanks, Egon."
