It's not gasses like poisons that are the culprit for making these things detectable. You have particles that permiate solids, like neutrons. As AJ explains below.
Gallium Arsenide neutron detectors are pretty cheap to make, and easy to deploy. (I'm assuming that's what NYC is using.)
AJ, tell me if I'm remembering my physics correctly here. If the bomb was made with plutonium, it'd be easier to detect than a uranium one, right? (Due to the difference in neutron release rate/ decay energy?)
Plutonium would be easier to make a weapon from. Uranium would be harder to make (even if you got some crudely refined 30% enriched, with an awesome neutron reflector) but also harder to detect.
I THINK (not "know") that Pu239, like I said earlier, if it was well-processed from the breeding nuclear reactor to keep Pu240 at a minimum, is actually less detectable than a U235 sub-critical mass.
This is one of the convenient things about Pu239. It's spontaneous rate of neutron emission is much lower, so it can actually be formed into a supercritical sphere, and not go off on it's own. U235 can't do that. Just bring critical U235 masses near each other by handr and you might get a really nasty "fizzle" that'll kill people and might be a few tons in yield. Aside from the famous death at Los Alamos with U235 pieces, others have died just taking liquid U235 solutions from a skinny flask, like a graduated cylinder, to one with a high volume low surface area shape like an Erlenmeyer.
A nice blue flash of Cherenkov radiation and then a few minutes/hours later, you don't feel so good.
The supercritical Pu239 sphere or pit as it's called can make it easier to transport, and handle the device, but the symmetry and simultaneity needed for detonation, plus all the neutron reflectors and enhancers you might want to get a good yield make it harder to manufacture.
A U235 gun or slamming device is technically easier to produce, but much larger and heavier to transport, and it's much "hotter" just by virtue of the nature of U235 which also makes it a good nuclear fuel.
Of course Pu239 has a much longer logistic trail. First you have to get refined U235, build it into a reactor in such a way it breeds Pu239, then reprocess the fuel to get the Pu out, and in such a manner the Pu240 is minimized, or you start having the same problems with your bomb-pit that you do with U235 in terms of radiation and practicality.
AFAIK, the only reason we made a "little boy" U235 gun-device was that the Trinity device and Fat Man completely exhausted the U.S.'s supply of Plutonium.