Depth criteria seems to be a common problem when detecting vortices in near range storms. The WSR-88D is often able to resolve the fine velocity structure of near range storms, and this can include numerous vortices that are closely spaced. Three dimensional, vertical association becomes a difficult task.
In this case, the MDA has found the tornadic vortex (#8) at the lowest few tilts. At tilt 3, though, another shear segment is identified as well (#26). When looking at tilt 4, there is a broad area of shear. We would like to define it as the vertical continuation of the tornadic mesocyclone (#8). The algorithm, however, has chosen #26 instead, seeing that it is of closer proximity to the tilt 4 shear segment.
Mesocyclone candidate #8 is cut off at a depth of less than one km, and was thusly discarded as insignificant. The resulting mesocyclone (#26) is identified with its base at the third tilt at a location that is several km removed from the tornado.
As always, a careful inspection of the data by the human forecaster can alleviate these problems. Problems such as these are perhaps more important to overcome at times when the radar signatures are not so strongly suggestive of a tornado as those examined here.