sign in sign up
<<
Home , Atmospheric chemistry

Draft version December 27, 2018 Typeset using LATEX twocolumn style in AASTeX61

BIOSIGNATURE ANISOTROPY MODELED ON TEMPERATE TIDALLY LOCKED M-DWARF PLANETS

Howard Chen,1, 2 Eric T. Wolf,3, 4 Ravi Kopparapu,4, 5 Shawn Domagal-Goldman,4, 5 and Daniel E. Horton1, 2

1Department of and Planetary Sciences, Northwestern University, Evanston, IL 60202, USA 2Center for Interdisciplinary Exploration & Research in Astrophysics (CIERA), Evanston, IL 60202, USA 3Laboratory for Atmospheric and Space , Department of Atmospheric and Oceanic Sciences, University of Colorado Boulder, Boulder, CO 80309, USA 4NASA Astrobiology Institute Virtual Planetary Laboratory, Seattle, WA 98194, USA 5NASA Goddard Spaceflight Center, Greenbelt, MD 20771, USA

ABSTRACT

A planet’s atmospheric constituents (e.g., O2,O3,H2Ov, CO2, CH4, and N2O) can provide clues to its surface habitability, and may offer biosignature targets for remote life detection efforts. The plethora of rocky exoplanets found by recent transit surveys (e.g., the Kepler mission) indicates that potentially habitable systems orbiting K- and M-dwarf stars may have very different orbital and atmospheric characteristics than Earth. To assess the physical distribution and observational prospects of various biosignatures and habitability indicators, it is important to understand how they may change under different astrophysical and geophysical configurations, and to simulate these changes with models that include feedbacks between different subsystems of a planet’s . Here we use a three-dimensional (3D) - Climate model (CCM) to study the effects of changes in stellar spectral energy distribution (SED), stellar activity, and planetary rotation on Earth-analogs and tidally-locked planets. Our simulations show that, apart from shifts in stellar SEDs and UV radiation, changes in illumination geometry and rotation-induced circulation can influence the global distribution of atmospheric biosignatures. We find that the stratospheric day-to-nightside differences on tidally-locked planets remain low (< 20%) across the majority of the canonical biosignatures. Interestingly however, secondary photosynthetic biosignatures (e.g., C2H6S) show much greater (∼67%) day-to-nightside differences, and point to regimes in which tidal-locking could have observationally distinguishable effects on phase curve, transit, and secondary eclipse measurements. Overall, this work highlights the potential and promise for 3D CCMs to study the atmospheric properties and habitability of terrestrial worlds.

Keywords: astrobiology – planets and : – planets and satellites: terrestrial planets arXiv:1810.12904v3 [astro-ph.EP] 26 Dec 2018

Corresponding author: Howard Chen, Northwestern University [email protected] 2 Chen, H., Wolf, E. T., et al.

1. INTRODUCTION large telescopes (E-ELT, GMT, and TMT); these obser- A promising approach in the hunt for life beyond vatories will enable spectroscopic observations of rocky Earth is through the detection of biosignatures – biolog- planets around K- and M-type stars. It is therefore likely that our first opportunity to measure atmospheres ically produced compounds such as O2, CH4,N2O, and of rocky worlds will be tidally-locked terrestrial planets CO2 – in the atmospheres of terrestrial planets orbit- ing the putative habitable zones (HZs) of nearby stars around K- or M-dwarf stars. (Lovelock 1975; Sagan et al. 1993; Seager et al. 2009; Characterization of exoplanets primarily involves Kasting et al. 2015). measuring starlight and terrestrial thermal emissions In recent years, the convergence of our ability to de- absorbed by planetary atmospheres as a function of tect, confirm, and characterize extrasolar planets has wavelength. For transit , which will be the profoundly strengthened the prospects of finding life main tool for obtaining spectra from planets around on other worlds. Consistently improving measurements M-dwarf stars, observations are biased towards atmo- of stellar mass, radius, and distance allows more ac- spheric constituents across the terminators. Therefore, curate constraints on their attending planets (Mann interpreting spectroscopic observations requires infer- et al. 2015). Large-scale observational surveys such ring both the and distribution of de- as the M-Earth project, TRAPPIST survey, Hungar- tectable gases. Such properties can be predicted by 3D ian Automated Telescope Network (HATNet), Kepler global climate and chemistry-climate models (GCMs Space Telescope, and Transiting Exoplanet and CCMs). GCMs and CCMs are numerical mod- Survey (TESS) have detected planets in the habit- els that employ laws of physics, fluid motion, and in able zones around these stars (Thompson et al. 2018) the case of CCMs, chemistry to simulate movements, and will continue to monitor closer and brighter sys- interactions, and climatic implications of a planet’s at- tems for Earth-sized planets (Barclay et al. 2018). Si- mospheric constituents and boundary conditions. multaneously, follow-up characterization efforts of these Previous simulations of atmospheres of tidally-locked confirmed planets were able to resolve atmospheres of planets performed with 3D GCMs have demonstrated much smaller planets than past efforts (e.g., HAT-P- that habitable states of tidally-locked planets are strong 26b; Wakeford et al. 2017). Looking ahead, a va- functions of: (i) Coriolis force (Yang et al. 2016; Way riety of instruments are being designed with life de- et al. 2016; Kopparapu et al. 2016), (ii) stellar energy tection goals in mind. This includes ground-based distribution (SED) and bolometric stellar flux (Koppa- observatories such as the European Extremely Large rapu et al. 2017; Wolf 2017), (iii) atmospheric mass Telescope (E-ELT), Giant Magellen Telescope (GMT), (Wordsworth 2015), and (iv) radiative transfer scheme and Thirty-Meter Telescope (TMT), as well as space- (Yang et al. 2016). Despite the ability of GCMs to simu- based missions such as the James Webb Space Tele- late key climatological factors, as demonstrated by these scope (JWST), Large UV/Optical/IR Surveyor (LU- studies, their foci have primarily been on questions of VOIR), Origins Space Telescope (OST), and Habitable habitability, rather than the and distri- Exoplanet Imaging Observatory (HabEx). HabEx and butions of biologically-produced gases and habitability LUVOIR in particular would enable characterization of indicators. potentially habitable Earth-sized rocky planets in our To study effects of tidal-locking on atmospheric chem- istry and molecular spectroscopic signals, models ca- solar neighborhood (. 100 parsecs; Bolcar et al. 2016; Mennesson et al. 2016; Batalha et al. 2018). pable of resolving chemical speciation, reactions, and The recent discoveries of Proxima Centauri b (Anglada- transport are needed. To date, exoplanet atmospheric Escud´eet al. 2016) and the TRAPPIST-1 system (Gillon photochemical predictions have largely relied on one- et al. 2017) demonstrate that analyses of small rocky dimensional global-mean -climate mod- planets are within reach. However, many of these plan- els (e.g., Kasting et al. 1984; Segura et al. 2005; Mead- ets orbit extremely close to their host M-type stars ows et al. 2018a). These 1D models have been used to (0.02-0.2 AU) and are susceptible to trapping by tidal- simulate synthetic spectra of hypothetical rocky plan- forces (Tarter et al. 2007). Tidally-locked but poten- ets under the influence of different host SEDs (Rauer tially habitable planets are expected to be common in et al. 2011; Rugheimer et al. 2015). However, 1D mod- HZs of low-mass stars (∼15%; Dressing & Charbonneau els employ relatively simple eddy-diffusion parameteri- 2015) – which dominate our solar neighborhood stellar zations for vertical transport and do not account for at- population (∼70%; Henry et al. 2006). Concurrently, mospheric dynamics, climate heterogeneities, or 3D geo- our earliest opportunity for a biosignature search will metric effects critical to observations. These factors are likely come from the JWST and ground-based extremely important as advection and diffusion can affect concen- Tidal-Locking & Biosignatures 3 tration, distribution, and ultimately the composition of et al. 2012). These conditions include atmospheric gases −2 an (Seinfeld & Pandis 2012). In addition to N2 (78% by volume), O2 (21%), and CO2 (2.85×10 %) altering photochemistry, as shown by 1D models, shifts (MacFarling Meure et al. 2006). In addition, the model in stellar SED can influence atmospheric circulation and simulates the free-running evolution of H2Ov and O3, climate (e.g., Shields et al. 2014; Fujii et al. 2017). At- while CH4 and N2O surface fluxes are latitudinally vari- −7 −7 mospheric chemistry and dynamics are thus interactive, able (global mean CH4: 7.23×10 and N2O: 2.73×10 and should ideally be simulated using fully-coupled 3D mol mol−1). Throughout the remainder of the paper, we model components. refer to this Earth-Sun simulation as the baseline. Here, to better understand the observational poten- We also modify CAM-chem to simulate tidally-locked tial of tidally-locked planets, the integrated effects of planets with initially Earth-like atmospheric composi- atmospheric chemistry, photochemistry, and circulation tions forced by M-dwarf SEDs. This SED was obtained are considered over the 3D geometry of a planet’s atmo- from an open-source dataset of an M6V star, Proxima sphere. In this Letter, we simulate Earth-analogs and Centauri, compiled by NASA’s Virtual Planetary Lab- tidally-locked planets around M-dwarf stars using a 3D oratory (VPL) team and is available at http://vpl. CCM, while seeking to (i) elucidate the photochemical astro.washington.edu/spectra/stellar/. We ex- nature of Earth-like worlds, (ii) demonstrate the utility plore two SED-types (active and quiescent) that bracket of 3D CCMs in terrestrial exoplanet studies, (iii) and the endmember ranges of stellar activity. VPL Proxima advance model comparison efforts between 3D and 1D Cen. data is assumed to be moderately-to-highly active. research communities. To construct a quiescent M-dwarf SED, we swap out UV bands (λ < 500 nm) of the original Proxima Cen. data 2. MODEL DESCRIPTION & EXPERIMENTAL with that of a low-activity star (HD114710). For all exo- SETUP Earth simulations, we assume tidal-locking (i.e., trapped in 1:1 spin-orbit resonances), with orbital periods of 50 In this study, we employ the Community Atmosphere Earth days. While we do not use self-consistent stellar- Model with Chemistry (CAM-chem), a subset of the Na- flux orbital period relationships (e.g., Kopparapu et al. tional Center for Atmospheric Research (NCAR) Com- 2016; Haqq-Misra et al. 2018), the idealized case stud- munity Earth System Model (CESM v.1.2), to investi- ied here highlights the value of using CCMs for modeling gate atmospheres of Earth-like planets. CAM-chem is chemical processes on slowly and synchronously rotating a 3D global CCM that simulates interactions of atmo- planets. spheric chemistry, radiation, , and dy- For both Earth and tidally-locked exoplanet simula- namics (for a complete model description, see Lamarque tions, we set orbital parameters (obliquity, eccentricity, et al.(2012)). CAM-chem combines the CAM4 atmo- and precession) to zero, such that top-of-atmosphere in- spheric component with the fully implemented Model cident stellar flux is symmetric about the equator. In- for and Related Chemical Tracers (MOZART) cident bolometric stellar flux for all simulations is set chemical transport model. CAM-chem resolves 97 gas to 1360 W m−2. The substellar point for all simula- phase species and linked by 196 chemical and tions is fixed at () latitude = 0◦ and longitude = photolytic reactions. CAM, the atmosphere component 180◦, in the Pacific Ocean. Note that other studies (e.g., of the model, has seen wide applications in problems of Lewis et al. 2018) have shown that surface type beneath paleoclimate and exoplanets (e.g., Wolf & Toon 2015; a substellar point can modify availability, Kopparapu et al. 2016), whereas CAM-chem has largely influencing water vapor-induced greenhouse and been limited to studies of present Earth. All simulations radiative effects, and possibly atmospheric chemistry. presented were run for 30 Earth years and reported re- In all simulations, we assume present Earths conti- sults are averaged over the last 20. nental configuration, topography, mass, and radius. We We simulate Earth-analogs and tidally-locked planets use prognostic atmospheric and oceanic components of and assess the sensitivity of atmospheric biosignatures CESM, as well as prescribed preindustrial land, surface to three primary variables: (i) stellar spectral energy ice, and sea ice components. Horizontal resolution (lat- distribution, (ii) stellar UV radiation, and (iii) plane- itude × longitude) is set to 1.9◦ × 2.5◦ with 26 vertical tary rotation period. To simulate Earth-analogs, we use atmospheric levels and model top of 1 mb (∼50 km). a preindustrial Earth setup forced by solar spectral ir- The land model is Community Land Model version 4.0 radiance data (Lean et al. 1995), i.e., apart from the or- with non-interactive surface features. The ocean com- bital parameters described below, our Earth-analog sim- ponent is a thermodynamic slab model with prescribed ulation uses identical boundary and initial conditions to Earth in 1850, prior to anthropogenic influences (Taylor 4 Chen, H., Wolf, E. T., et al.

Table 1. Model Comparisons of Approximate Global-mean Mixing Ratios of Various Gases Rauer et al. Rugheimer et al. Study This work (2011) (2015) NCAR’s Model 1D photochemical EXO-P CAM-chem active M6 stellar SED data AD Leonisa Proxima Cen.c modelb Bond Albedo N/Ad 0.07 0.46

Tsurf (K) 298 300 242

−1 −11 −9 −13 O3,surf (mol mol ) 8 × 10 10 9.4 × 10

−1 −4 −3 −4 CH4,surf (mol mol ) 10 1 × 10 3.4 × 10

−1 −6 −6 −6 N2Osurf (mol mol ) 2.0 × 10 1.7 × 10 2.4 × 10

−1 −2 −2 −4 H2Osurf (mol mol ) 7.0 × 10 5.0 3.5 × 10 e T100mb (K) 251 245 200

−1 −7 −6 −7 O3,100mb (mol mol ) 10 10 5.2 × 10

−1 −5 −3 −4 CH4,100mb (mol mol ) 10 1 × 10 3.1 × 10

−1 −5 −7 −6 N2O100mb (mol mol ) 10 5 × 10 2.0 × 10

−1 −6 −4 −6 H2O100mb (mol mol ) 10 7.0 × 10 5.8 × 10

Approximate atmospheric (units of K) and mixing ratios of various gas phase species (units of mol mol−1) simulated on Earth-like planets by various authors. All data reported are taken from simulations forced by active stellar SEDs a b that range from M6V to M8V dwarf stars. Active M3.5eV star, Teff = 3300 K, M = 0.41M . Part of a suite of stellar SED model data generated by Rugheimer et al. (2015) to represent the most active M-dwarf observations (e.g., PHOENIX or MUSCLES database). cSED data from Virtual Planetary Laboratory’s stellar spectrum database (Meadows et al. 2018a). dData not provided. eValues reported at 100 mb. heat flux values sourced from dynamical ocean simula- 3. RESULTS tions (e.g., Danabasoglu & Gent 2009). Three general observations can be made from our sim- Consistent with 1D studies (e.g., Segura et al. 2005, ulated 3D global distributions of O3, CH4,N2O, and Rugheimer et al. 2015) and in alignment with our lack DMS on Earth-like and tidally-locked planets (Figures of terrestrial exoplanet observations, we assume atmo- 1 and 4): (i) Changes in mixing ratios of O3, CH4, and spheric compositions, biological production, and dry de- N2O are primarily due to different levels of stellar UV position rates of gaseous species the same as those of flux amongst the three SED datasets. (ii) Introduction preindustrial Earth. Apart from CH4 and N2O, global of tidal-locking modifies globally homogeneous gas dis- surface gas flux inputs are based on spatially-explicit tributions that characterize Earth-like scenarios. (iii) preindustrial monthly averages (e.g., DMS; Kettle & An- Heterogeneous surface-to-atmosphere flux distributions dreae 2000). Due to SED sensitivities, CH4 and N2O (e.g., DMS) can influence the resultant mixing ratios of surface flux boundary conditions are estimated via an- atmospheric constituents. cillary CCM simulations that allow for the emergence of To facilitate analysis of our results, we define a day- stellar SED-dependent flux magnitudes (i.e., WACCM; to-nightside mixing (mole) ratio contrast as: Neale et al. 2010). Emergent SED-consistent N2O and rday − rnight CH4 flux estimates are temporally and spatially fixed rdiff = (1) rglobe in active and quiescent M-dwarf simulations at CH4: −4 −3 −1 − 3.5×10 and 2.3×10 mol mol and N2O: 2.5 ×10 6 i.e., the relative difference between the two hemi- −5 −1 and 3.2×10 mol mol , respectively. Given uncertain- spheres, where rday is the dayside hemispheric mixing ties inherent in flux estimates, sensitivity experiment ratio mean, rnight the nightside mean, and rglobe the and day-to-nightside mixing ratio comparisons should global mean. The degree of anisotropy is loosely en- focus on relative rather than absolute differences. capsulated in this parameter, which is analogous to the Tidal-Locking & Biosignatures 5

Figure 1. Global distribution of O3, CH4, and N2O mixing ratios and relative humidity for Earth-like non-tidally-locked (P = 24 hrs) Solar-SED simulations (a, d, g, j) and tidally-locked (P = 50 days) active (b, e, h, k) and quiescent (c, f, i, l) M-dwarf SED simulations. Evidence of circulation- and photochemical-induced biosignature anisotropy are apparent. Day-to- nightside mixing ratio contrasts (rdiff ) for tidally-locked simulations are reported, while relative humidity is averaged across the globe (RHglobe). Gas mixing ratios are pressure-weighted vertical averages over the top of the model atmosphere (1-to-100 mb). Relative humidity is reported for the 200 mb pressure surface. Note differences in scaling factors used amongst experiments and constituents. Dashed-lines indicate locations of terminators. definition used by Koll & Abbot(2016) in the context (Figures 1a-c). However, the quiescent SED, produces of temperature contrasts. Values of rdiff for each respec- lower ozone concentrations above the tropopause (Fig- tive experiment are shown in Figures 1 and 4 and will ure 3c). These differences reflect specific stellar activity be discussed throughout the paper. inputs. Quiescent M-dwarfs emit lower UV in the range responsible for ozone production (160 < λ < 240 nm). Moreover, calculated day-to-nightside mixing ratio dif- ferences rdiff are higher (∼19%) in the active M-dwarf 3.1. Ozone Distributions, Water Vapor Mixing Ratios, SED scenario. and Temperature Profiles Modulations to ozone concentration have major in- Ozone production and destruction depend on stellar fluences on distributions of other biosignature gases. UV activity, availability of molecular and atomic oxy- This is due to substellar hemispheric production of ex- gen, and ambient meteorological conditions (P,T ). As cited state atomic O(1D) and constituent fami- 1 our simulated M-dwarf SED is moderately active in the lies of HOx. Both O( D) and HOx constituents are reac- UV bands, our results show similar quantities of ozone tive radicals important for atmospheric biogenic organo- between the baseline Earth-Sun and tidally-locked cases 6 Chen, H., Wolf, E. T., et al.

1 Altitude (km) Pressure (mb)

1 Altitude (km) Pressure (mb)

Figure 2. Stratospheric O3 and OH production rates as functions of longitude for Earth-like non-tidally-locked (P = 24 hrs) Solar-SED simulations (a, d) and tidally-locked (P = 50 days) M-dwarf SED (b, c, e, f) simulations. Photolytic processes drive ozone and hydroxyl production and help to explain many of the observed biosignature gas distributions in Figure 1. Note vertical axis begins at 50 mb (∼20 km). Dashed-lines indicate locations of terminators. compounds and hydrocarbons (e.g., CH4, CH3, HCL, H2S). 1 As ozone is photochemically produced, horizontal ad- O( D) + H2Ov → 2OH (R1) 1 3 vection carries a portion to the nightside as evidenced by O( D) + M(N2, O2) → O( P) + M (R2) its presence in both hemispheres. The lifetime of ozone 3 O( P) + O2 + M → O3 + M (R3) (∼15 days), in conjunction with day-to-nightside trans- port, is sufficient enough to sustain some nightside O3, but not efficient enough to fully mix the atmosphere, The higher ozone mixing ratios on tidally-locked allowing day-to-nightside contrast (∼19%; Figure 1b). nightsides is explained by these reaction pathways (Fig- Conversely, the product O(1D) shows limited transport ure 1b-c). Reaction R1, which creates the hydroxyl effects due to a short lifetime (< 5 secs), reflected in its radical OH, predominantly occurs on the dayside due 1 large rdiff value (∼300%; not shown). O( D) is rapidly to the abundance of H2Ov (Figures 2e-f). In R2, sin- removed by either the R1 or R2 reaction pathway (Jacob glet oxygen returns to the triplet ground state, which 1999): can then recombine with oxygen to form ozone via R3. Considered together, significant dayside UV ozone destruction and enhanced removal of O(1D) by water Tidal-Locking & Biosignatures 7 vapor offset higher ozone production rates (Figure 3b), in which capturing feedbacks between 3D dynamical which helps to explain lower dayside ozone mixing ratios processes, solar forcing, and atmospheric chemistry is (Figure 1b-c). critical. Interaction of stellar UV photons with O2 and O3 can also be seen in vertical temperature profiles. On Earth- 3.2. Types I and II Biosignatures: 3D CH4 and N2O like planets, stratospheric temperature is primarily a Abundances function of incident UV flux (200 < λ < 310 nm) due to CH4 and N2O are important biosignatures produced the role of O3 absorption of shortwave photons. In our by a myriad of bacterial metabolic pathways (Des simulations, this feature is apparent in global and hemi- Marais et al. 2002; Schwieterman et al. 2018). In Figures spherically averaged profiles. Our simulations indicate 1d-f, we show modeled CH4 distributions. High CH4 that upper-stratospheric increase (and in- mixing ratios for planets orbiting quiet M-dwarfs were versions weaken) as UV radiation levels increase; from first noted by Segura et al.(2005) using a 1D model, quiescent SED, to active SED, to the baseline Earth caused by reduced photochemical removal by reaction simulation (Figure 3a). Enhanced UV absorption by O3 with OH: and O2 increase temperatures above the tropopause, re- ducing the vertical gradient and inversion strength. CH4 + OH → CH3 + H2O (R4) We now turn to discussing testability of our 3D model predictions. Based on simulated ozone distributions, the In our 3D CCM, we find similar global mean CH calculated r (∼20%) is notable but unlikely to be 4 diff increases in tidally-locked simulations (Figures 1e-f). discernible with current observational capabilities (Bur- However, active and quiescent simulations have low CH rows 2014). However, this task may prove viable us- 4 r values (13.7% and 2.9%, respectively). Low r ing future instruments (e.g., Greene et al. 2016). Proe- diff diff values are explained by a mixture of competing pro- drou & Hocke(2016) reached a similar conclusion by cesses. First, upwelling in tidally-locked simulations oc- comparing total column ozone of a tidally-locked Solar- curs exclusively below the substellar point, as evidenced SED Earth and found a ∼23% difference between mean by upper-tropospheric moisture patterns (Figures 1j- ozone columns during four arbitrary phases due to vary- l). Compared to the baseline, tidally-locked meridional ing viewing angles. overturning circulation is strengthened, which brings Compared with 1D model studies, our 3D simulations greater moisture aloft. This, in conjunction with the produce similar ozone mixing ratios (Table 1). How- dayside abundance of OH, removes CH via R4. In- ever, we find substantially different Bond albedos, tem- 4 creased dayside OH production (Figures 2e-f) is a con- perature, and water mixing ratio profiles. As a conse- sequence of abundant O(1D) and H O (R1), both of quence of increased dayside water vapor-induced opacity 2 v which are sparse on the nightside. These processes com- and substellar on tidally-locked planets, global- bine to limit dayside CH and produce lower r values mean surface temperatures of both tidally-locked simu- 4 diff than expected. lations are ∼40 K colder than the baseline (Figure 3a), N O is primarily destroyed by UV photons (λ < 220 while Bond albedos are substantially higher (Table 1), 2 nm) and photo-oxidation by reactions with stratospheric in agreement with GCM studies (Yang et al. 2013; Kop- O(1D) (Figure 1g-i). Hence predicted N O concentra- parapu et al. 2016, 2017). Colder global (and nightside) 2 tions around active M-dwarfs are lesser than those with temperatures produce lower global water vapor mixing quiescent SEDs. For both active and quiescent SED ratios than predicted by 1D models with clear-sky as- simulations, higher concentrations within the substellar sumptions (i.e., pure water vapor without clouds). Con- hemisphere are found (Figures 1h-i), similar to CH be- versely, dayside H O mixing ratios are greater due to 4 2 v havior. humid updrafts at the substellar point (Figures 1j-l). Interestingly, simulations forced by the active SED Curiously, the quiescent M-dwarf SED simulation (Fig- have greater stratospheric r (CH : 13.7% and N O: ure 1l) has lower global-mean relative humidity than the diff 4 2 6.9%; Figures 1e-h) than those forced by quiescent SEDs Earth-analog (Figure 1j) and the active M-dwarf case (CH : 2.9% and N O: 1.1%; Figures 1f-i). Higher (Figure 1k). This is due to increased ozone mixing ratios 4 2 r values for active SED cases is somewhat counter- and degree of UV absorption, which limit the photoly- diff intuitive as one might expect that enhanced photolytic sis of H O , in the more active SED simulations. Hence 2 v destruction on planets around active M-dwarfs should counterintuitively, more dayside H O destruction is ex- 2 v suppress day-to-nightside contrasts. However, simula- perienced by the simulation under lower UV radiation. tions forced by active SEDs have more isothermal atmo- Such behaviors exemplify the value of CCM simulations, spheres (i.e., weaker temperature inversions; Figure 3a), 8 Chen, H., Wolf, E. T., et al.

48 (a) (b) (c) QGM

ADS

AGM Altitude (km) 32 Earth

16

0

v 48

32 Altitude (km)

Pressure (mb) Pressure (mb) 16

(d) (e) (f) 0

Figure 3. Vertical profiles of global-mean temperature (a) and mixing ratios of various gas phase species (b, c, d, e, f). QGM (quiescent global-mean) denotes global-mean values from simulations forced by a quiescent, AGM (active global-mean) denotes those forced by an active M-dwarf SED, and ADS (active day-side) denotes dayside-mean values from simulations forced by an active M-dwarf SED. Note axes are log-scaled and begin at planetary surfaces (∼1000 mb). which promote vertical mixing of surface gases above the Similar to the above DMS behavior, a potential con- tropopause, contributing to higher rdiff values. sequence of tidal-locking is the relegation of CH4 and N2O production to a single hemisphere, i.e., processes 3.3. Effects of Surface Fluxes on Atmospheric of methanogenesis and denitrification favor anaerobic Distribution: Case of Dimethyl Sulfide conditions and may be disfavored on photosynthetic On tidally-locked planets, phototrophs are unlikely to oxygen-producing daysides. Spatially-variable surface emit biogenic gases globally (i.e., the assumption for to atmosphere flux distributions of CH4 and N2O there- all biosignature gases considered thus far); rather, ally- fore could exhibit higher rdiff than the values predicted derived emissions are likely to be restricted to the day- here (. 10%; Figures 1e-f and 1h-i). side. To see how a biosignature gas (e.g., DMS; C2H6S) may behave on a tidally-locked planet, we conduct three 4. DISCUSSION experiments, each with a different DMS flux distribution Here we discuss possible areas of future advancement, assumption, i.e., Earth-like, tidally-locked with global as well as the observational relevance of this study. DMS flux, and tidally-locked with dayside DMS flux In this CCM study, we find that factors that deter- (Figure 4). We find that global DMS emissions result mine biosignature concentration and distribution on a in substantially lower rdiff values (∼0% and ∼43%; Fig- habitable tidally-locked planet are species dependent. ures 4a-b) compared to a strictly dayside DMS emission For example, ozone mixing ratios are primarily driven assumption (∼67%; Figure 4c). The larger value of rdiff by photolytic production and destruction, while ozone in the latter simulation is due to the relatively short distribution and nightside sustenance are controlled by lifetime of DMS (Kloster et al. 2005). its transport and lifetime. An additional consideration, Tidal-Locking & Biosignatures 9

Figure 4. Global distribution of photosynthetic biosignature dimethyl sulfide (DMS, (CH3)2S) for Earth-like non-tidally-locked (P = 24 hrs) Solar-SED simulations (a), tidally-locked (P = 50 days) active M-dwarf SED simulations with a global DMS flux assumption (b), and tidally-locked with a dayside DMS flux assumption (c). Day-to-nightside mixing ratio differences (rdiff ) for the tidally-locked simulations are reported. Phototrophs are assumed to be only present on the permanently lit day-side in panel (c), which results in an enhanced stratospheric day-to-nightside mixing ratio contrast. Dashed-lines indicate locations of terminators. here demonstrated in our DMS simulations, is the spa- Earth-like planets, with similar atmospheric formation tial variance of gas fluxes. Given that habitable exoplan- histories, ecospheres, and biological signatures resulting ets are likely to possess heterogeneous ecologies, whose from oxygenic (Meadows et al. 2018b). fluxes will interface with attendant atmospheric struc- This study demonstrates that fully-coupled CCMs are ture and circulation patterns, spatially-heterogeneous particularly promising for studies that seek to assess the surface fluxes could have observationally-distinguishable roles of and feedbacks between different stellar SEDs, bi- effects on atmospheric spectra. For example, dayside up- ological behaviors, and atmospheric compositions. Such welling could facilitate vertical mixing of surface gases efforts are consistent with recent reviews, discussing as- into the upper atmosphere (Figure 1e-f), while night- pirational goals and the future of exoplanet biosigna- side radiation inversions could trap constituents near ture research (Catling et al. 2018; Walker et al. 2018). the surface, limiting vertical mixing, day-nightside inter- Extrasolar astrophysical radiation environments and/or actions, and potentially observability. These scenarios, atmospheric conditions may alter biological activity, as in which atmospheric dynamics, photochemistry, sur- life is both photochemically and climatologically medi- face flux sources, and feedback processes play impor- ated. Living are highly receptive toward UV tant roles highlight the utility of 3D CCM simulations. emissions such that UV-B (290 < λ < 320 nm) pho- However, due to non-linear interactions and internal at- tons hinder metabolism, photosynthesis, and thus bio- mospheric variability, disentangling drivers of emergent logical production rates (Jansen et al. 1998). Moreover, behavior is challenging and will likely require the tools of due to different climate and redox conditions, for ex- modern atmospheric and computational science, includ- ample on anoxic Archean Earth (∼3.0 Ga), hydrocar- ing Lagrangian tracking of constituents (e.g., S¨olch & bons and organosulfur biosignatures (C2H6, CH4, OCS, K¨archer 2010), single- and multi-model ensembles (e.g., DMS etc.) could rise to more prominent abundances Kay et al. 2015), and statistical analyses focused on de- and hence may be conducive to remote detection (Haqq- tection and attribution (e.g., Horton et al. 2015; Diffen- Misra et al. 2008; Domagal-Goldman et al. 2011; Hu baugh et al. 2017). et al. 2012; Arney et al. 2016). Despite these challenges, the introduction of 3D In terms of the potential observational implications CCMs to exoplanet biosignature prediction efforts of- of our simulations, our results agree with those of Se- fers substantial research potential. Future applications gura et al.(2005), Rauer et al.(2011), and Rugheimer are likely to consider a wider variety of Earth-like bio- et al.(2015), i.e., rocky planets orbiting active and quiet spheres, e.g., markedly disparate atmospheres of Earth M-dwarfs should have deeper absorption depths, par- throughout geologic time (Rugheimer & Kaltenegger ticularly for ozone and secondary biosignatures such as 2018; Arney et al. 2016), and should be expanded and (as seen in transit and emis- to include biologically constrained models and mod- sion spectra). This makes habitable zone planets orbit- ules (Catling et al. 2018; Walker et al. 2018). Such ing M-dwarfs favorable targets. 3D predictions from our applications will facilitate egocentricity avoidance – a CCM simulations may be confirmed by remote observa- commonly acknowledged goal of the field (e.g., Seager tions. Phase curve analysis can potentially resolve 3D et al. 2013). Until then, use of default Earth conditions atmospheric structures of super-Earth and Earth-sized may restrict the relevance of 3D CCM findings to truly terrestrial planets (Stevenson et al. 2014; Kreidberg & 10 Chen, H., Wolf, E. T., et al.

Loeb 2016). For example, thick substellar clouds could nature gases on Earth-like and tidally-locked planets appear characteristically for planets with specific spit- as a function of stellar spectral type, stellar activity, orbit resonances (Yang et al. 2013). and planetary rotation period. Qualitatively similar In terms of biosignature measurements on tidally- to 1D models, we find increased mixing ratios of bio- locked planets, different longitudinal gradients of genic compounds (e.g., O3, CH4, and N2O) for both ac- gaseous constituents may affect measurements of vari- tive and inactive M-dwarf SEDs. These increases are ation spectra (peak amplitude of the phase curves) most pronounced for planets around quiet M-dwarfs. extracted from thermal phase curves (Selsis et al. 2011). Even though the effects of tidal-locking are noticeable in As variation spectral signals depend on amplitude-peaks our simulations, they are not yet discernable with cur- in orbital light curves, there may be added anisotropy rent observational techniques, i.e., the primary biosigna- due to time-varying longitudinal gas distributions on tures simulated in this work (O3, CH4,N2O) show low tidally-locked planets in each orbit (assuming null obliq- (. 20%) day-to-nightside mixing ratio contrasts. Con- uity, as seen in Figure 1). Compared to non-tidally versely, simulated day-to-nightside differences of photo- locked fast rotators (with similar stellar UV activity synthetic compounds (e.g., DMS) are found to be nearly and orbital period), we predict that more pronounced 70% and underscore the need for heterogeneous 3D re- absorption signals would be seen in variation spectrum alism in modeling biosignatures and their photochemi- on tidally-locked planets, driven by greater difference cal derivatives. Overall, this study serves as a stepping between maximum and minimum phase amplitudes due stone for future applications using 3D CCMs to study to uneven hemispheric gas distributions. Emission spec- the habitability and spectroscopic observability of ter- trum at maximum phase (direct line-of-sight) should restrial exoplanets. correspondingly see similar behavior, at least for a few IR-windows (e.g., between 3-9 µm; Selsis et al. 2011). H.C. thanks J. Schnell and A. Loeb for helpful discus- For direct imaging, one possibility is that these features sions. E.T.W. thanks NASA Habitable Worlds Grant may be more prominent during certain orbital phases. 80NSSC17K0257 for support. Goddard affiliates are Ozone observability, for example, may decrease during thankful for support from GSFC Sellers Exoplanet En- secondary eclipses as the dayside with reduced ozone vironments Collaboration (SEEC), which is funded by abundance would be Earth-facing. Radiative transfer the NASA Planetary Science Divisions Internal models, using our CCM results as inputs, will be needed Funding Mode. The VPL team at University of Wash- to quantitatively assess observational prospects of the ington is thanked for the stellar spectra data. Computa- above. tional, storage, and staff resources were provided by the QUEST high performance computing facility at North- 5. CONCLUSIONS western University, which is jointly supported by the This Letter reports numerical simulations using a cou- Office of the Provost, Office for Research, and North- pled 3D CCM to explore global distribution of biosig- western University Information Technology.

REFERENCES

Anglada-Escud´e,G., Amado, P. J., Barnes, J., et al. 2016, Catling, D. C., Krissansen-Totton, J., Kiang, N. Y., et al. Nature, 536, 437 2018, Astrobiology, 18, 709 Arney, G., Domagal-Goldman, S. D., Meadows, V. S., et al. Danabasoglu, G., & Gent, P. R. 2009, Journal of Climate, 2016, Astrobiology, 16, 873 22, 2494 Barclay, T., Pepper, J., & Quintana, E. V. 2018, ArXiv Des Marais, D. J., Harwit, M. O., Jucks, K. W., et al. 2002, e-prints, arXiv:1804.05050 Astrobiology, 2, 153 Batalha, N. E., Lewis, N. K., Line, M. R., Valenti, J., & Diffenbaugh, N. S., Singh, D., Mankin, J. S., et al. 2017, Stevenson, K. 2018, ApJL, 856, L34 Proceedings of the National Academy of Sciences, 114, Bolcar, M. R., Feinberg, L., France, K., et al. 2016, in 4881 Proc. SPIE, Vol. 9904, Space Telescopes and Domagal-Goldman, S. D., Meadows, V. S., Claire, M. W., Instrumentation 2016: Optical, Infrared, and Millimeter & Kasting, J. F. 2011, Astrobiology, 11, 419 Wave, 99040J Dressing, C. D., & Charbonneau, D. 2015, ApJ, 807, 45 Burrows, A. S. 2014, Proceedings of the National Academy Fujii, Y., Del Genio, A. D., & Amundsen, D. S. 2017, ApJ, of Science, 111, 12601 848, 100 Tidal-Locking & Biosignatures 11

Gillon, M., Triaud, A. H. M. J., Demory, B.-O., et al. 2017, Mennesson, B., Gaudi, S., Seager, S., et al. 2016, in Nature, 542, 456 Proc. SPIE, Vol. 9904, Space Telescopes and Greene, T. P., Line, M. R., Montero, C., et al. 2016, ApJ, Instrumentation 2016: Optical, Infrared, and Millimeter 817, 17 Wave, 99040L Haqq-Misra, J., Wolf, E. T., Joshi, M., Zhang, X., & Neale, R. B., Chen, C.-C., Gettelman, A., et al. 2010, Kopparapu, R. K. 2018, ApJ, 852, 67 NCAR Tech. Note NCAR/TN-486+ STR Haqq-Misra, J. D., Domagal-Goldman, S. D., Kasting, Proedrou, E., & Hocke, K. 2016, Earth, Planets, and Space, P. J., & Kasting, J. F. 2008, Astrobiology, 8, 1127 68, 96 Henry, T. J., Jao, W.-C., Subasavage, J. P., et al. 2006, AJ, Rauer, H., Gebauer, S., Paris, P. V., et al. 2011, A&A, 529, 132, 2360 A8 Horton, D. E., Johnson, N. C., Singh, D., et al. 2015, Rugheimer, S., & Kaltenegger, L. 2018, ApJ, 854, 19 Nature, 522, 465 Rugheimer, S., Kaltenegger, L., Segura, A., Linsky, J., & Hu, R., Seager, S., & Bains, W. 2012, ApJ, 761, 166 Mohanty, S. 2015, ApJ, 809, 57 Jacob, D. 1999, Introduction to atmospheric chemistry Sagan, C., Thompson, W. R., Carlson, R., Gurnett, D., & (Princeton University Press) Hord, C. 1993, Nature, 365, 715 Jansen, M. A., Gaba, V., & Greenberg, B. M. 1998, Trends Schwieterman, E. W., Kiang, N. Y., Parenteau, M. N., in plant science, 3, 131 et al. 2018, Astrobiology, 18, 663 Kasting, J. F., Chen, H., & Kopparapu, R. K. 2015, ApJL, Seager, S., Bains, W., & Hu, R. 2013, ApJ, 775, 104 813, L3 Seager, S., Deming, D., & Valenti, J. A. 2009, Astrophysics Kasting, J. F., Pollack, J. B., & Crisp, D. 1984, Journal of and Space Science Proceedings, 10, 123 Atmospheric Chemistry, 1, 403 Segura, A., Kasting, J. F., Meadows, V., et al. 2005, Kay, J., Deser, C., Phillips, A., et al. 2015, Bulletin of the Astrobiology, 5, 706 American Meteorological Society, 96, 1333 Seinfeld, J. H., & Pandis, S. N. 2012, Atmospheric Kettle, A., & Andreae, M. 2000, Journal of Geophysical chemistry and physics: from to climate Research: Atmospheres, 105, 26793 change (John Wiley & Sons) Kloster, S., Feichter, J., Maier-Reimer, E., et al. 2005, Selsis, F., Wordsworth, R. D., & Forget, F. 2011, A&A, Biogeosciences Discussions, 2, 1067 532, A1 Koll, D. D. B., & Abbot, D. S. 2016, ApJ, 825, 99 Shields, A. L., Bitz, C. M., Meadows, V. S., Joshi, M. M., Kopparapu, R. k., Wolf, E. T., Arney, G., et al. 2017, ApJ, & Robinson, T. D. 2014, ApJL, 785, L9 845, 5 S¨olch, I., & K¨archer, B. 2010, Quarterly Journal of the Kopparapu, R. k., Wolf, E. T., Haqq-Misra, J., et al. 2016, Royal Meteorological Society, 136, 2074 ApJ, 819, 84 Stevenson, K. B., D´esert,J.-M., Line, M. R., et al. 2014, Kreidberg, L., & Loeb, A. 2016, ApJL, 832, L12 Science, 346, 838 Lamarque, J.-F., Emmons, L. K., Hess, P. G., et al. 2012, Tarter, J. C., Backus, P. R., Mancinelli, R. L., et al. 2007, Geoscientific Model Development, 5, 369 Astrobiology, 7, 30 Lean, J., Beer, J., & Bradley, R. 1995, Geophysical Taylor, K. E., Stouffer, R. J., & Meehl, G. A. 2012, Bulletin Research Letters, 22, 3195 of the American Meteorological Society, 93, 485 Lewis, N. T., Lambert, F. H., Boutle, I. A., et al. 2018, Thompson, S. E., Coughlin, J. L., Hoffman, K., et al. 2018, ApJ, 854, 171 ApJS, 235, 38 Lovelock, J. E. 1975, Proceedings of the Royal Society of Wakeford, H. R., Sing, D. K., Kataria, T., et al. 2017, London Series B, 189, 167 Science, 356, 628 MacFarling Meure, C., Etheridge, D., Trudinger, C., et al. Walker, S. I., Bains, W., Cronin, L., et al. 2018, 2006, Geophysical Research Letters, 33 Astrobiology, 18, 779 Mann, A. W., Feiden, G. A., Gaidos, E., Boyajian, T., & Way, M. J., Del Genio, A. D., Kiang, N. Y., et al. 2016, von Braun, K. 2015, ApJ, 804, 64 Geophys. Res. Lett., 43, 8376 Meadows, V. S., Reinhard, C. T., Arney, G. N., et al. Wolf, E. T. 2017, ApJL, 839, L1 2018a, Astrobiology, 18, 630 Wolf, E. T., & Toon, O. B. 2015, Journal of Geophysical Meadows, V. S., Arney, G. N., Schwieterman, E. W., et al. Research (Atmospheres), 120, 5775 2018b, Astrobiology, 18, 133 Wordsworth, R. 2015, ApJ, 806, 180 12 Chen, H., Wolf, E. T., et al.

Yang, J., Cowan, N. B., & Abbot, D. S. 2013, ApJL, 771, L45 Yang, J., Leconte, J., Wolf, E. T., et al. 2016, ApJ, 826, 222