[update: i think i was wrong in a previous version of this post]. i think you need an extra 1/(1+z)^2 for expansion. see, for example, http://en.wikipedia.org/wiki/Luminosity_distance where there is a (1+z) correction to D (and you want D^2 for area).
that would decrease probabilities by ~1/100.
that seems reasonable given ctdonath's comment - tens of photons a second at a telescope sounds like it should be detectable (IR is harder than optical as it has higher backgrounds).
am also unsure about the ^2 (i think you're worried whether it's intensity or not? i don't know either).