In the next year, if all goes well, JWST will have begun collecting data on the composition of terrestrial exoplanet atmospheres. This is potentially one of the most exciting developments in the history of science, but it's not going to be easy; here is a very informative preview:

Perhaps the key point is that, with the given signal to noise ratios, it may be possible to derive spectra with remarkable fidelity and spectral resolution, but the weak signal in most or all possible cases means that the number of required observations, to build up the signal, will be prohibitive given the limited lifespan of JWST and the large number of systems that we'll want to observe. Rather than campaigns that produce detailed spectra of many candidate "earthlike" planets, we will see the observation time divided amongst many exoplanets and spectra with moderate detail – but likely enough to determine presence or non presence of key molecules. This still depends, of course, upon the exoplanets themselves, whose atmospheres, surfaces, and clouds may make the signal weaker or stronger in any particular case, and those are variables which we cannot possibly control or predict.

To add some sad detail to this, the paper calculates that for some desired measurements, the number of transits that would have to be observed would be over 100 or even 1000, and this is flatly impossible. If the JWST were devoted to the observation of just one particularly special exoplanet and we wished to ignore all other uses of the telescope, this threshold would still be impossible, and obviously, there is no lack of priority targets for the telescope.

Among some molecules of highest interest, the ease of detection will be, in descending ease, CH4, CO2, H2O, O2, and for the planets in the TRAPPIST-1 system, the number of required observations to provide a useful signal for O2 will be on the order of 40.

It seems likely that what we'll see is campaigns to obtain some spectral data for about 15-25 (that is my sense, not a definitive total) candidate "earthlike" planets over JWST's first three or so years, and then more sustained campaigns to follow up on those planets that look most promising after the initial surveys. Overall, the use of JWST for this type of observation will require a very strategic budgeting of the resource of observation time, giving us a little data about a lot of the candidates, and – hopefully – much better data on the few most promising cases. The end result will depend on details that we can only guess at now.

No matter what turns up from JWST, there will always be the opportunity and need for future instruments to extend the studies outward and examine the candidates a little farther. If JWST's "horizon" for this sort of science is a radius of X parsecs, then a future instrument with 4 times the light gathering would extend it to 2X parsecs, and a volume in space 8 times greater. JWST will be the beginning of a great exploration outwards that will never conclude so long as we can keep building bigger and better instruments, decade by decade.