The task of identifying young nearby stars is a difficult one, particularly when the stars are members of the nearby probably unbound moving groups/associations/whatchamacallits
Usually, young star research is done on open clusters, OB associations, or other groupings that are far enough away from us that the cluster covers a tiny area of the sky and we can be reasonably sure that all the members of the cluster are within some well-defined region of the sky (although the reverse is usually not true). This is not really helpful when the cluster is composed of perhaps 100 stars, scattered over a quarter of the sky. Instead, we have to rely on different parameters.
Unfortunately, none of these methods are foolproof.
UVW match to cluster: Stars that formed together should still be moving together, even hundreds of millions of years after they formed.
On the minus side: As Lopez-Santiago et al. (2006) determined, and as Eric Mamajek has been saying for years, there are far more random field objects that will just happen to match the motion of your cluster. That’s what happens when you have ~100 star systems in an unbound association (which implies a larger range of acceptable velocities) covering a volume of space with perhaps 100,000 star systems in it.
Also: There is no particular reason to demand that a young star should be a part of an association. We assume stars form in batches, but is that true? And we know that stars can get ejected from binaries at very high velocities, so what’s to stop a single member of a cluster from appearing here, while the rest of its members are 100 parsecs away?
Above the main sequence: Young stars that are still in their pre-main-sequence contraction phase will be physically larger in radius than main-sequence stars of the same color, as follows from L= 4 * pi * r^2 * t^4
On the minus side: The pre-main-sequence lifetimes of stars can be brief, and the portion of that lifetime where a star will be NOTICEABLY above the main sequence is briefer still. In M dwarfs, it does take perhaps 200 Myr for an M0 star to reach the main sequence (Dotter et al. 2008) with longer times for less massive stars, but there are enormous metallicity-related differences in luminosity too. You also have to be certain the star is not elevated above the main sequence because it is an (as-yet) unresolved binary.
Also: This requires a parallax. Or, if it’s a suspected member of an association, you could use an isochrone to get predicted distance, and then make sure the resulting UVW velocity is also appropriate for the association in question.
Low surface gravity: Physically larger stars will have lower surface gravities.
This suffers from all the same problems as identifying pre-main-sequence stars. In mid-M and cooler dwarfs, though, you can use several gravity-sensitive absorption features, like Na I, K I, Ca I, (Allers et al. 2007) and probably their singly-ionized companions (Na II, K II, and Ca II are all strongest in giants and weakest in high-gravity situations, Na I, K I, and Ca I are strongest in high-gravity situations and weakest in giants). TiO and VO are also sensitive to gravity, but they MAKE the spectra of low-mass objects with their massive absorption bands, so unless you’re doing spectrophotometry or really high-resolution stuff it’s probably not so useful.
Lithium: Virtually all lithium in the universe is primordial, the result of a tiny fraction of nucleosynthesis during that phase of the big bang. Normal fusion processes in stars will destroy it but not create it. Therefore, if you see any lithium in the star, it must not have destroyed it yet. (da Silva et al. 2009)
Neglecting that we don’t quite have the mechanism down (Yee & Jensen 2010), my personal problem with lithium is that it’s destroyed very quickly in M3-M4 dwarfs, the most massive fully convective objects (above that mass, matter isn’t cycled directly into the core and isn’t destroyed as quickly; below that mass, fusion is slow and weak and just plain takes longer to destroy the lithium). By requiring lithium we guarantee it’s young (very young?) but I’m still interested in “young” objects that are old enough to have fused all their lithium.
Also note that this test seems dependent on accretion and rotational velocities (can’t find a reference now, but SOMETHING has to explain the massive differences in abundances among well-determined members of the same 12 Myr old cluster). And, it doesn’t work on brown dwarfs of less than 60 solar masses; their cores never reach temperatures sufficient to fuse lithium.
Disks: Young stars will still have their protoplanetary disks around them.
Like Lithium, this only works when REALLY young, like less than 10 Myr old. The other thing to recognize is that debris disks (only dust, not gas+dust) also exist around older stars like Vega and Fomalhaut. Not that Vega and Fomalhaut can be that old, considering that they’re both A stars with lifetimes of ~1 Gyr… but they’re still young enough that what we see is cold dust from an asteroid belt smashing itself to pieces, not a close-in gaseous disk capable of forming planets.
Chemical Composition: Castro et al. (1997) and da Silva et al. (2013) both find that Barium is conspicuously abundant in the young moving groups Ursa Major and Argus (respectively). The general impression I get is that some process (lots of AGB stars?) within Scorpius-Centaurus has been enriching the cloud with S-process elements specifically; as time goes on, the forming stars have higher and higher Barium abundances. There are other leads listed in those papers, but the general picture is that chemical abundances are another good marker for youth.
Activity: This one is a difficult one. Flares, X-ray emission, H-alpha emission, NUV and FUV emission… all are linked to coronal and chromospheric activity, and (ultimately) the rotational velocity of the star. Spin charged gasses, get magnetic fields.
This works great for massive objects, where X-ray emission and fast rotation are unusual (well, except that O,B, and A stars are renowned fast rotators). It doesn’t work so well for M dwarfs, where, due to low stellar winds, they spin fast almost forever: West et al. (2008) lists M7 stars as taking longer than 8 billion years to stop having H-alpha emission. Proxima Centauri (6 Gyr?) is a flare star. Barnard’s Star (8-10 Gyr) is a flare star. Conversely, brown dwarfs are supposed to have very weak magnetic fields, and I’ve heard a talk where someone claimed they don’t spin down at all. And then there are the oddballs: in my next paper (in the hands of a referee now!) I have two very young (~8 Myr old) M0-M3 dwarfs with NO X-ray emission and no flares. Every other one of these properties says they’re TW Hydra members, but they’re not active in that way, at least not right now.
Others: There are probably other ways to tell if a star is young, geared specifically to their spectral types. In particular, something is up with the so-called red L dwarfs that’s making them red.