Having grown up in California I am well aware of the ability of a wildfire to be influenced by everything from wind and pressure systems, to terrain and resources. This does not correspond significantly to the requirements needed to effect a tornado before forming. A wildfire may very well exist for many days before containment, whereas a tornado may last minutes. A wild fire is still stationary (though expanding), and can be confronted due to it's location being known.
What you propose is to designate where a system might create a tornado, designate a response team, move resources into place, track said system with technology beyond what we have today, figure out where to fire a missile for optimum effect, issue the appropriate warnings to those underneath the blast, hope the explosion takes place high enough to avoid damage yet low enough to manage the change in atmosphere proposed, and repeat as needed.
Unlikely at best......silly at it's worst.