Mate, the rug under your house of cards as to why it couldn't possibly be an anomaly was that (in order of your post right at the top above):
There's not enough sulphide in the core to explain it: Irrelevant, as discussed above.
It doesn't agree with the AMT anomaly: Irrelevant, as discussed above.
It doesn't sit anywhere near the footwall contact: Irrelevant, as discussed above.
Then finally, in your post, after about 2/3 of a page, we get to whether it's a model or real - looking only at the Z component.
Firstly, and on this I'll agree with you about reporting, we don't know which of the three profiles is the Z - there's no labels. 2 out of 3 profiles show what is obviously an unambiguous response at late time. I'm sure even you wouldn't disagree as you chose not to mention it. My suspicion is that it's the top profile which is the Z as it has the most read station on profile - that's where they honed in on the anomaly and then read the other profiles to see if they got confirmation. That's how I read it.
Standard practice is to read Z component first to see what's out there, using a variety of transmitter loops to cover a variety of potential geometries. Saves time and money. If something kicks in Z, then you tighten station space and read X and Y around the Z anomaly to confirm it. If nothing kicks in Z, you pack up and go home. If nothing kicks in either Y or X, you scratch your head, then usually pack up and go home. If something kicks on 2 out of 3 components at least - chances are it's real and needs further work. That's how every crew I've ever worked with does it.