With the uneven weld prep and the lower part with 10mm of land, I think the 1st question is can it be reliably welded with complete penetration and fusion?
With the appropriate welding process and appropriate back gouging, there will be no problem assuring CJP; however, if not, there will be discontinuities that UT will be able to detect.
The 3,5mm (10/2 - 3/2) of land either side of the root bevel on the lower face (in the diagram) may be difficult to achieve good fusion. As is often the case, if it presents a challenge to the welder, then it will present a challenge to the UT tech as well. Of course if the very best welders and UT techs in the country have been hired then this will mitigate the risk. If not, shear wave UT will not detect LOF in those areas I have highlighted.
I fail to see why this detail should present a serious problem to either the welder or the UT technician. The detail looks like a groove detail that could be used for a horizontal butt joint. It would be welded from one side and back gouged from the second side before the second side is welded. The UT could be performed from both sides of the joint and both from the top and bottom sides (as viewed in the sketch the poster provided).
There should be sufficient beam spread that even if the sound beam didn't hit the incomplete fusion perpendicular to the groove face some of the sound would be reflected back to the transducer. The reflected sound would be detectable in either the first leg or the second leg as the technician "scrubs" the transducer toward and away from the joint.