Just spotted this:
The reason I am doing this is that Astrometrica gets this wrong. In fact it reports two different SNRs in two different places
If by two different places you mean two different positions within the same image, it may not be wrong. As I noted in my earlier response the PSF can be position-dependent. The sky brightness will almost certainly be position-dependent (try measuring LBVs in the Andromeda galaxy!). Anyway, a fainter star (lower signal) will have a lower SNR than a brighter one even if the background is the same.
I am not quite understanding what you mean, in other words. I am but a bear of very little brain.