Based on85C/Full Power
I see your logic, but don't agree this can be inferred from the graph; that at 70degC the body temperature will rise 85DegC at 100% power. There is not a reason presented to provide that inference is true. The same inference cannot be made of transistors, so why here?From Yageo resistors:
At 155C the resistor has a 0 watts rating. So that is to say that this resistor should not get above 155C.
I think at 70C air temp and 100% power the body of the resistor is at 155C.
So (155-70=85) My guess is that 100% power causes a temp rise if 85C.