http://www.nucleardarkness.org/nuclear/nuclearexplosionsimulator/
Does not allow to detonate nukes above 2 MT, but I did some scaling guesswork:
"Certain Mass Fires" radius:
2 MT: 821 km^2
1 MT: 417 km^2
0,2 MT: 88 km^2
0,02 MT: 10 km^2
0,002 MT: 1,3 km^2
Looks like a factor of ~9 for the area for a factor of 10 in weapon yield. If that does not change, I would expect an area of 25*10^6 km^2 or a radius of 2800 km for an explosion of 10^5 MT. However, this would need a nearly flat explosion, which is a bit unrealistic for such a high yield. Multiple bombs at different places could give that effect, of course.
Another tool, here for the shock wave:
5 psi overpressure: "Complete destruction of ordinary houses, and moderate to severe damage to reinforced concrete structures, will occur within this ring." (does not take heat into account)
100MT: 20.99 km radius
20MT: 12.28 km
2MT: 5.7 km (102km^2)
0,2MT: 2.64 km
0,02MT: 1.23 km
That corresponds to a factor of 2.15 in radius or 4.64 in area for a factor of 10 in yield. As ##10^{1/3}=2.154## and the shock wave is spherical, that looks reasonable.
Scaled to 10^5 MT:
100MT: 208 km radius
A third tool shows effects of the shock wave and heat at the same time, and indicates that indeed the burned area expands quicker with weapon yield than the shockwave. And it shows the difference between an explosion on ground and in the air.