[ventus45]

Aussie, when you combine raw primary radar returns from multiple different radar stations into "one situation display" there will be some degree of "disjointed plots" for a whole host of reasons.
First, in the 2D map, or "flat earth" if like, display, there will be minor position differences due to the co-ordinates system used, the curvature of the earth between the radar stations, their different site altitudes etc.
That is frustrating enough, but when you add the third dimension, and try to get a 3D plot with altitude, the results are problematic.
Each individual primary radar head is operating in a different bit of atmosphere.
Refraction and ducting effects make altitude determination for any "one head", from only "a few returns", difficult.
You need to "track it for a while" so that the computers can generate a 3D track file that "makes sense". And that is just for "one head".
Now try to combine the plots from "multiple heads".
Now, depending on the "coverage" of each head, there may be some "overlaps", where the target is seen by two or more heads "at the same time".
Combining them to produce one definitive x y z position (lat lon alt) is not a simple matter, and not likely to be very accurate "in the third dimension - altitude".
So I think we can "discount" the "altitude jumps" as a "red herring".
What matters is "the track" the aircraft took, ie, the x y (lat lon) at time "t" plots.