Exploring New Worlds with Webb

Joanna Barstow, joannakbarstow@cantab.net

Exoplanet researchers are counting down the days until the launch of the James Webb Space Telescope. Webb will transform our ability to unveil the atmospheres of planets transiting close to their parent stars. The community is in the process of developing tools, obtaining complementary observations, and planning for the first round of Webb observing proposals.

Transit spectroscopy and Webb

Transiting exoplanets provide an excellent opportunity to study atmospheres in the most extreme conditions. When a planet passes in front of its star, some starlight is filtered through the upper regions of the planet’s atmosphere. Tiny (1:10,000–1:1,000,000) fluctuations in the transit depth correspond to the signatures of absorbers in the planet’s atmosphere.

The measurement of these miniscule signals requires low noise and well-characterized detectors and optics. Webb may not have been designed in the era of transit spectroscopy, but much instrumental testing has taken place with this type of observation in mind (e.g., Ferruit et al. 2014; Greene et al. 2007; Lagage et al. 2010). Starting exoplanet observing programs with the best possible knowledge of instrument behavior will help researchers prepare ahead of time for instrumental challenges. Feasibility studies to examine the likely recoverable information from Webb spectra are also being performed (Barstow et al. 2015; Greene et al. 2016; Figure 1).

Figure 1: Simulated eclipse spectra taken from Greene et al. 2016. The spectra are for a single transit with equal time on the star alone for each of the four instruments. Spectra have been binned to resolution R ≤ 100 (hot Jupiter, warm Neptune, warm sub-Neptune) and R = 35 (cool super-Earth). The simulated spectra include a noise instance and presented as colored curves. The black error bars denote 1σ of noise composed of random and systematic components. Dashed lines show the wavelength range boundaries of the chosen NIRISS, NIRCam, and MIRI instrument modes. This figure was originally presented as Figure 5 of Greene et al. 2016, Characterizing Transiting Exoplanet Atmospheres with JWST, ApJ, 817:17, January 20, 2016. DOI: 10.3847/0004-637X/817/1/17. ©AAS. Reproduced with permission.

Remote sensing and degeneracy

Remote sensing is the process of inferring properties of atmospheres from images and spectra, obtained by telescopes and spacecraft. The information contained in these data is insufficient to fully specify the detailed properties of a planet’s atmosphere. Retrieval techniques based on, for example, optimal estimation algorithms (Irwin et al. 2008; Lee et al. 2012; Rodgers 2000) or Monte Carlo (e.g., Madhusudhan & Seager 2009; Benneke & Seager 2012; Line et al. 2013; Waldmann et al. 2014), are used to infer the most likely state of the atmosphere from the limited information available.

Retrieval is challenging; often, multiple model solutions can be found that produce an equally good match to the data—a problem called degeneracy. For solar system planets that have been visited by in-situ probes, we can include an informed prior in the retrieval that restricts the range of possible solutions. For exoplanets, we don’t have that luxury.

Webb’s broad wavelength coverage (0.6–28 μm) will go some way towards solving the problem. For transit spectra, one of the biggest challenges is presented by the presence of clouds that obscure atmospheric absorption features due to gases and prevent us from seeing deep into the atmosphere (e.g., Kreidberg et al. 2014). But cloud opacity is often a strong function of wavelength, so using a broad wavelength range may help us to avoid this pitfall (Figure 2). When planets are viewed in eclipse, different wavelengths are sensitive to radiation emerging from different levels in the atmosphere, so a broad wavelength range means information about atmospheric structure over the greatest possible pressure range (Figure 3).

Figure 2: Simulated Webb primary transit spectra based on two different atmospheric models for cloudy super Earth GJ 1214 b (15 transits). Both are compatible with current data, but they diverge at wavelengths that will be sampled for the first time with Webb. Existing data are as shown in Barstow et al. (2013b), with the data from Kreidberg et al. (2014) and Fraine et al. (2013) added. The dark/light gray lines show models with the cloud top at 0.1/0.01 mbar. This figure was originally presented as Figure 16 of Barstow et al. 2015, Transit Spectroscopy with JWST: Systematics, Starspots and Stitching, MNRAS, 458, 2657.

Figure 3: Retrieval sensitivity to temperature as a function of pressure and wavelength, for a hot Jupiter orbiting a Sun-like star (top) and a hot Neptune orbiting an M dwarf (bottom), as observed by Webb NIRSpec and MIRI instruments. The contours correspond to a change in observed flux ratio for a 1 K change in the temperature at each level in the atmosphere. The division between the NIRSpec and MIRI instruments is marked as a thin white line. This figure was originally presented as Figure 1 of Barstow et al. 2015, Transit Spectroscopy with JWST: Systematics, Starspots and Stitching, MNRAS, 458, 2657.

Another way of overcoming degeneracy is to exploit the range of different geometries available for transiting exoplanets. In transit, we observe starlight passing through the terminator atmosphere (the region between day and night) on a long path tangential to the planet’s surface (Figure 4). These measurements are sensitive to the presence of high clouds and absorption by specific gases, but relatively unaffected by the atmospheric temperature structure.

Figure 4: The various transit observation geometries. In primary transit, the signal of interest is from the starlight passing through the planet’s atmosphere, whereas for secondary eclipse and phase curve measurements the signal is thermal radiation from the planet itself, emerging from the planet at different phases.

Secondary eclipse spectra probe thermal emission from the planet itself and penetrate deeper into the atmosphere. The temperature structure is encoded in these spectra, as well as absorption by gases. Secondary eclipse spectra at shorter Webb wavelengths will also contain contributions from reflected starlight, providing further information about the presence and scattering properties of clouds. Eclipse spectra probe the starlit (dayside) part of the atmosphere.

For key targets, it’s possible to observe flux from the planet at a range of angles. By observing the system for the duration of the planet’s orbit, monitoring the slight change in flux tells us about the varying contribution from the planet as our view changes from one of the hot dayside to the cooler, fainter nightside. This kind of measurement allows us to make a longitudinal map of the planet’s properties (e.g., Stevenson et al. 2014), which can be compared with circulation models that predict wind speeds and structure (e.g., Kataria et al. 2015).

Although these different measurements probe different regions of the atmosphere, we anticipate that the sensitivity to different parameters can be exploited to break degeneracies. Additionally, measurements made by alternative techniques using ground-based telescopes (e.g., high resolution Doppler spectroscopy, Snellen et al. 2010) can provide necessary prior information.

Preparing for Webb

Currently, efforts are focused on preparing for the Webb Early Release Science (ERS) program. This community-selected program will execute early in Cycle 1, with the goal of producing scientifically compelling data sets in all of the major modes of JWST available to the broad community with no proprietary time. The ERS program will provide proposers from Cycle 2 onwards with information about instrument performance in scientific observations. The exoplanet community is at work identifying suitable targets that can be observed with a range of instrument modes during ERS, to provide maximum information about the telescope’s performance for transit spectroscopy (Stevenson et al. 2016). In preparation, Hubble will likely perform reconnaissance of these targets. More information on the ERS program is available on the Institute's website, https://jwst.stsci.edu/science-planning/early-release-science-program.

In addition, efforts are underway to acquire UV and optical spectra of likely Webb exoplanet candidates while Hubble is still operational, since this wavelength region is critical for studies of clouds in both transit and eclipse. With these legacy spectra and new observations with Webb, we hope to extend studies such as that presented by Sing et al. (2016) to a much wider range of objects.

Towards terrestrial planets?

Webb will be a fantastic tool for comparative spectroscopy of hot Jupiters and warm Neptunes, but what about the more challenging smaller, temperate planets? With Webb, it may be possible to detect evidence of biosignatures such as ozone on an Earth-like planet orbiting in the habitable zone of a nearby cool M dwarf (Barstow et al. 2016). One such system, with three planets orbiting the M8 star TRAPPIST-1, has recently been discovered (Gillon et al. 2016). However, such a measurement would require observation of at least 60 transit events, taking up a significant proportion of the available observing time (Barstow et al. 2016). The decision of whether to pursue such a program would be a difficult one. The science return could be enormous, but the risk is high that biosignatures could be absent or obscured by clouds.

It is likely that Webb will be pushed to the limit of its capabilities for new exoplanet science. Whether or not that is characterization of other habitable worlds, it is sure to be exciting.


Barstow, J. K., Aigrain, S., Irwin, P. G. J., Kendrew, S., & Fletcher, L. N. 2015, MNRAS, 458, 2657

Barstow, J. K., Aigrain, S., Irwin, P. G. J., Kendrew, S., & Fletcher, L. N. 2016, MNRAS, 448, 2546

Barstow, J. K. & Irwin, P. G. J. 2016, MNRAS Letters, in press; arXiv:1605.07352

Benneke, B. & Seager, S. 2012, ApJ, 753, 100

Ferruit, P., et al. 2014, SPIE, 9143, 91430A

Gillon, M., et al. 2016, Nature, 533, 221

Greene, T., et al. 2007, SPIE, 6693, 66930G

Greene, T., et al. 2016, ApJ, 817, 17

Irwin, P., et al. 2008, JQSRT, 109, 1136

Kataria, T., et al. 2015, ApJ, 801, 86

Kreidberg, L., et al. 2014, Nature, 505, 9

Lagage, P. O., et al. 2010, In the Spirit of Lyot: Direct Detection of Exoplanets and Circumstellar Disks. Ed. A. Boccaletti (Paris: University of Paris Diderot)

Line, M., et al. 2013, ApJ, 775, 137

Lee, J.-M., Fletcher, L. N. & Irwin, P. G. J. 2012, MNRAS, 420, 170

Madhusudhan, N. & Seager, S. 2009, ApJ, 707, 24

Rodgers, C. (Ed.) 2000, Inverse Methods for Atmospheric Sounding: Theory and Practice, Series on Atmospheric Oceanic and Planetary Physics, Vol. 2 (London: World Scientific Publishing Co. Pte. Ltd.)

Sing, D. K., et al. 2016, Nature, 529, 59

Snellen, I. A. G., de Kok, R. J., de Mooij, E. J. W. & Albrecht, S. 2010, Nature, 465, 1049

Stevenson, K. B., et al. 2012, ApJ, 754, 136

Stevenson, K. B., et al. 2014, Science, 346, 838

Stevenson, K. B., et al. 2016, PASP, in press; arXiv:1602.08389

Waldmann, I. P., et al. 2014, ApJ, 802, 107