Just measuring antibodies, even if perfectly done, can tell you if you are at herd immunity without knowing fundamental characteristics of the virus. How infectious is someone? How long are they infectious?
This becomes trickier when a portion of the population might already have natural cross-immunity and perhaps antibodies fade but T cells remain trained in such a way that re-infection isn’t a risk.
The fundamental requirement for herd immunity for a isolated population is that the population have a low enough susceptible population (never have been infected and are not naturally immune/vaccinated), that
s * c * t * d < 1
s is the susceptible fraction
c is contacts per day for average person
t is transmission probability per contact
d is the average number of days a person is infectious and contacting others
True herd immunity occurs when the s is reached such that normal behavior (normal c, contact pattern) will not result in an epidemic if the virus is in circulation among the population. But you can appear to reach herd immunity at higher s values if c is still artificially low. Included in this would me quarantined, masks, increased hand washing, temperature screening, etc.
The true R0 of this virus was uncertain because our testing wasn’t in place as it started its natural progression through the population. That’s one reason herd immunity predictions vary. That is further complicated by the fact that some people might have a natural immunity.
I say all that to say you have to know what your target is to even make antibody testing for herd immunity useful. I’m not sure we really know what that is.
As for the question about the antibody lifetime and how you could test for it. First, is someone who loses their antibodies still immune? In most cases I think the answer is yes. So what is providing immunity? T cell response?
It’s possible that antibody tests will overpredict the susceptible population of a lot of people lose their antibodies. But that’s probably the limit of what I can say at this point without venturing into too much conjecture.