You can't rely on weight alone, because the kind of weight that can't be moved by a tornado, would be very difficult to move at all. You're going to need to be securely anchored to the surface at least some of the time, so I would use a walker with lots of legs and multiple drills on each foot. It would have a turtle like shell with a meter thick, Teflon/Kevlar coated, hard rubber exterior. It would be very heavy and nuclear powered. The vehicle could also be submersible, for traveling through shallow lakes and rivers.
Winds and debris will form laminar flows near the surface that will tend to flip over or at least lift most vehicle designs, so the outer edge of the shell would consist of retractable sections that would conform to the local surface profile and cause a laminar flow over the shell. These would also have a thick brush at the proximal end to the surface, that would help to fill-in the small scale ground features, to reduce wind and debris from getting under the vehicle. At more than a meter thick, they should provide a sufficient barrier, that a small vortex will from near the surface and around the edge of the vehicle, to direct airflow over and around the vehicle.
I would embed thousands of fractal antenna in the outer shell for communications and doppler radar. The later will be needed to detect tornados and large objects heading toward the vehicle. While it probably can't just step out of the way of large projectiles or tornados, it can at least stop, squat and anchor itself.
The vehicle can walk in any direction. It is round for aerodynamic purposes, but also so that it can rotate damaged areas to the leeward side of the vehicle. Shell blocks on the leeward side can be withdrawn into the vehicle and replaced. An internal shop provides repair facilities, particularly rubber extruders, Teflon/Kevlar coaters & antenna printers. When external conditions allow, additional sensors can be deployed as shell blocks when the vehicle reaches any artifacts of interest.
As anyone who has ever lost a boot in the bog can attest, swamps are going to be a problem for a walker. The feet will need to have a vacuum breaking compressed air supply. But some swamps are just going to be too deep even for these. This is where tracks would have to be deployed, but they don't provide anchorage. The vehicle is probably going to be heavy enough that it won't be blown over, but a direct hit from a tornado could be devastating when not anchored to the ground. That's where the doppler radar comes in handy. When you see a tornado coming, you squat down and deploy a number of large augurs, drilling as deep as possible into the muck.
You could conceivably construct a train of these walkers, each with a connecting tunnel between them. If the lead walker got into trouble, the others could help pull it out. It might also be abandoned if necessary. In fact, a train could literally form a walking circle that would eventually surround an artifact. A temporary canopy could be stretched between them, or other structure constructed, providing a shirt-sleeve environment for investigators.
Addressing @mishan's concerns regarding route erosion:
There's three ways to design a vehicle for the high winds specified by the OP:
- Construct a heavy/dense enough vehicle that tornados and hurricane winds cannot lift or blow it off course.
- Construct a vehicle capable of anchoring itself to the terrain.
- Rely on weight for the average conditions and use anchorage for the worst conditions.
Moving a vehicle that satisfies #1 implies a great deal of difficulty propelling the vehicle at all. The weight of the vehicle tends to pulverize the terrain beneath it. Consider that our largest tanks require specially designed transports that spread their load out, in order to move them over even our best constructed roads without damaging them.
#2 implies considerable erosion, particularly where anchor points cannot be reused.
#3 is a good compromise, as the lower weight causes less damage to the terrain in the average case, and more erosive anchorages happen at random locations, due to the nature of hurricane spawned tornados and random debris.
Tunneling would be a partial solution for the specified conditions, but is so expensive that it would not be worthwhile until you intend to have a lot of traffic on that route. The OP's scenario involves researches wanting to study immobile artifacts, within an extremely hostile environment, not heavily visited tourist sites.
The walkers don't require ideal road bed to travel over. They are designed to move through swamps and over highly variable terrain, while protecting their occupants in the face of high velocity projectiles and tornados. They should be able travel through the same areas, many times, but can also route around less favorable terrain, such as that which has been degraded by earlier trips.