Mars as an Exoplanet: Lessons from a Planet at the Edge of Habitability

Isotopic chronometers and dynamical models together suggest that Mars formed rapidly and then stalled at sub–Earth mass because its feeding zone was truncated. Hafnium–tungsten measurements indicate that Mars accreted most of its mass within $\lesssim$2–-4 Myr after calcium-aluminum–rich inclusions (CAIs), consistent with a stranded planetary embryo rather than the product of late giant impacts (Dauphas & Pourmand, 2011). Two end-member architectures can reproduce the Earth/Mars mass contrast and aspects of the asteroid belt: (i) the Grand Tack, in which Jupiter’s inward migration to $\sim$1.5 AU and subsequent reversal depleted solids beyond $\sim$1 AU and starved the Mars region (Walsh et al., 2011); and (ii) “annulus” initial conditions that confine most terrestrial material to $\sim$0.7–1.0 AU, naturally producing an over-massive Earth/Venus and an under-massive Mars (Hansen, 2009). Subsequent simulations show that both families of models can match the Earth/Mars mass ratio while reproducing the belt’s excitation and compositional mixing (Raymond et al., 2009; O’Brien et al., 2014).

Today Mars orbits with semi-major axis $a\simeq 1.5237$ AU, eccentricity $e\simeq 0.0934$, and inclination $i\simeq 1.85^{\circ}$ (J2000); these values and their uncertainties are provided by modern, refereed planetary ephemerides (e.g., Jet Propulsion Laboratory (JPL) Planetary and Lunar Ephemerides DE440 and DE441) and are consistent over multi-decade fits to spacecraft and radar tracking (Park et al., 2021). Mars’ comparatively large $e$ and $i$ among the terrestrial planets reflect long-term secular forcing dominated by the giant-planet modes, and its orbit undergoes substantial quasi-periodic and chaotic variations in $e$ and $i$ over $10^{6}$–$10^{8}$ yr (Laskar & Robutel, 1993; Laskar et al., 2004). Mars, in turn, exerts its own influence within Solar System dynamical interactions, playing a role within Earth’s Milankovitch cycles (Kane et al., 2025). Ensemble integrations of the full Solar System show that although Mercury can occasionally be driven to extreme eccentricities, the probability of catastrophic inner-planet instability over the next $\sim$5 Gyr is at the percent level; Mars remains dynamically long-lived aside from its chaotic secular excursions and strongly chaotic obliquity history (Laskar & Gastineau, 2009; Zeebe, 2015; Mogavero & Laskar, 2021).

Orbital spectroscopy and in-situ analyses show extensive Noachian-age alteration by liquid water: phyllosilicates followed by Hesperian sulfates, consistent with a progressively drying planet. This stratigraphic transition from ”phyllosian” to ”theiikian” provides a timeseries of water–rock interaction, pH, and redox conditions (Ehlmann & Edwards, 2014; Bibring et al., 2006). The NASA Curiosity rover established that an ancient, long-lived lacustrine system was once present at Gale Crater with neutral pH, low salinity, and essential nutrients—demonstrably habitable for chemolithoautotrophs. The Perseverance rover confirmed a delta–lake system within Jezero Crater, including flood deposits recording hydrologic variability (Grotzinger et al., 2014; Mangold et al., 2021).

Paleomagnetic and crustal remanence data constrain the history of Mars’ global dynamo, though that history is more complex than a simple early cessation. Analyses of Mars Global Surveyor data placed the end of the dynamo in the Noachian, at $\sim$4.0–4.1 Ga, based on the absence of strong crustal magnetization over the large impact basins Hellas, Argyre, and Isidis (Lillis et al., 2013). However, MAVEN magnetic field measurements over the $\sim$3.7 Ga Lucus Planum lava flows reveal confined near-surface magnetization consistent with a dynamo that was active near the Noachian–Hesperian boundary, substantially later than previously thought (Mittelholz et al., 2020). This finding implies that the dynamo may have postdated much of the geomorphic and mineralogic evidence for surface liquid water described above, complicating simple narratives in which dynamo cessation triggers atmospheric loss and climate decline. Furthermore, it remains unclear whether a global magnetic field provides a net protective effect against atmospheric escape. While an intrinsic dipole deflects some solar wind ion pickup, it can also facilitate loss through polar wind outflow and cusp ion escape, such that the presence of a magnetosphere does not guarantee reduced atmospheric erosion (Gunell et al., 2018; Brain et al., 2013). These nuances are important for interpreting exoplanet scenarios where the presence or absence of an intrinsic magnetic field is often assumed to be a binary indicator of atmospheric retention.

Mars undergoes chaotic obliquity variations that modulate insolation distribution, volatile redistribution, and atmospheric pressure via seasonal and long-term CO₂ condensation-sublimation at the poles and regolith (Laskar et al., 1993, 2004). Without a large stabilizing moon comparable to Earth’s, Mars’ spin-axis tilt has ranged from $\sim$0°to $>$60°on timescales of $\sim$$10^{5}$–$10^{6}$ yr (Ward, 1973; Touma & Wisdom, 1993), driving order-of-magnitude swings in polar insolation and seasonal CO₂ cycling (Jakosky et al., 1995; Forget et al., 2006). At high obliquity, polar volatiles migrate toward the equator and atmospheric mass may temporarily increase, whereas at low obliquity, CO₂ can collapse onto the poles and reduce surface pressure below the triple-point threshold for liquid water. These mechanisms provide a natural laboratory for atmospheric stability thresholds and collapse pathways that offer insights relevant to sub-Earth exoplanets, including tidally locked worlds where qualitatively similar, though dynamically distinct, cold trapping of volatiles can occur (Wordsworth, 2015). Mars also exerts its own secular forcing on the inner Solar System, contributing to Earth’s Milankovitch cycles (Kane & Miles, 2025), further illustrating how planetary architecture shapes long-term climate evolution.

Mars’ thin CO₂ atmosphere (with O, CO, O₂, N-bearing species, and HOx) is shaped by photochemical cycles that regenerate CO₂ from CO/O and regulate odd-hydrogen chemistry (Nair et al., 1994; Krasnopolsky, 2006). The rapid photodissociation of CO₂ by solar UV radiation produces CO and O, but catalytic cycles involving odd-hydrogen species (H, OH, HO₂) efficiently recombine these products and maintain the observed $\sim$95% CO₂ abundance. This balance is sensitive to the water vapor profile, dust loading, and UV flux (parameters that vary with orbital state and atmospheric dust events) making the Martian photochemical system a real-world testbed for photochemical models of thin CO₂ atmospheres on rocky exoplanets (Lefèvre & Forget, 2009). In the exoplanet context, Mars photochemistry constrains the production rates of abiotic O₂ and O₃, which are key false-positive biosignature gases expected on irradiated, low-outgassing worlds. The Martian example also provides empirical constraints on the thermodynamic conditions under which CO₂ condensation occurs, which are relevant to the nightside cold traps of tidally locked exoplanets, where the atmospheric circulation regime and heat-transport efficiency introduce additional controls on collapse thresholds (Wordsworth, 2015).

The present-day Martian atmosphere provides direct evidence that small rocky planets are highly susceptible to long-term volatile loss. Although the present atmosphere of Mars is relatively thin, the early atmosphere was likely far more substantial, since that is required to account for the hosting of surface liquid water (Pollack et al., 1987; Jakosky & Phillips, 2001; Wordsworth, 2016; Kite, 2019; Warren et al., 2019). For example, Joiret et al. (2025) estimated a conservative minimal mass of the primordial Martian atmosphere that implies a surface pressure of at least 2.9 bar. This primordial atmosphere, present during and immediately following planetary formation, was itself shaped by early volatile processing including hydrodynamic escape during magma-ocean stages, isotopic exchange between the atmosphere and solidifying mantle, and impact-driven loss and delivery (Pepin, 1991; Kasting, 1993; Elkins-Tanton, 2008; Pahlevan & Stevenson, 2007). The subsequent atmosphere that hosted the surface liquid water recorded in the geologic record (beginning $\sim$4.0–4.3 Ga) may have differed substantially from this primordial state.

More broadly, the atmospheric evolution of any planet depends critically on the initial volatile inventory incorporated during formation, which in turn reflects the composition, location, and timing of accreting materials (Jakosky & Treiman, 2023). Given that the initial incorporation of volatiles into planets remains poorly understood even within our own Solar System (where we have extensive observations of planets, asteroids, and meteorites) the range of possible atmospheric outcomes for Mars-sized exoplanets with unknown formation histories is correspondingly large. This uncertainty in initial conditions must be borne in mind when extrapolating from Mars to exoplanet populations.

Lyman-$\alpha$ observations and analyses of combined data from MAVEN and the Emirates Mars Mission (EMM) reveal that hydrogen escape—often diffusion-limited—can be seasonally or dust-storm enhanced; large dust events correlate with spikes in upper-atmosphere H and increased escape, demonstrably linking climate and loss (Chaffin et al., 2014; Heavens et al., 2018; Chaffin et al., 2021). Recent observations from MAVEN have provided direct quantitative constraints on the dominant non-thermal escape pathways operating at Mars today, helping to anchor escape models for small exoplanets subjected to elevated XUV and stellar winds. For example, measurements of ion escape rates indicate that atmospheric loss is strongly modulated by solar wind conditions, including dynamic pressure and interplanetary magnetic field orientation, with significant enhancements observed during space weather events such as coronal mass ejections (Jakosky et al., 2015, 2018). MAVEN observations further demonstrate that pickup ion escape, in combination with solar wind-driven sputtering, represents an ongoing mechanism for atmospheric erosion in the absence of a global magnetic dynamo (Brain et al., 2016). Additionally, photochemical escape of oxygen, driven by dissociative recombination of O${}_{2}^{+}$ in the upper atmosphere, currently exceeds ion escape as an O loss channel and can operate regardless of whether a global dynamo is present (Lillis et al., 2017).

Enrichment of heavy isotopes (³⁸Ar/³⁶Ar, D/H, ¹⁵N/¹⁴N) from in-situ measurements indicates substantial atmospheric loss by mass-selective processes Jakosky et al. (2017). Such isotopic forensics calibrate retrieval expectations for exo-atmospheres once precision spectroscopy becomes feasible (Mahaffy et al., 2013; Webster et al., 2013). Early Mars likely experienced hydrodynamic H escape during magma-ocean/outgassing phases (Pepin, 1991; Tian et al., 2009) and episodic atmospheric loss due to large impacts (Melosh & Vickery, 1989); both processes leave distinct isotopic and volatile-inventory signatures relevant to young exoplanets. These results suggest that episodic climate variability may contribute to long-term volatile depletion and provide an empirical benchmark for evaluating atmospheric retention on Mars analog exoplanets, particularly those orbiting active stars with elevated stellar wind fluxes (Brain et al., 2026).

Liquid water is the essential solvent for life as we know it, and the history of its abundance on Mars is central to assessing planetary habitability. The past presence of water, and thus potential habitable environments, is recorded in the Martian geologic record through geomorphological features such as catastrophic outflow channels, fluvial channel systems and valley networks, sedimentary fan deposits, and features potentially consistent with a former ocean. These geomorphological indicators are strongly supported by orbital spectral observations that reveal a wide variety of hydrated minerals and mineraloids, including Fe/Mg-smectite clays, chlorites, kaolinite, mica, hydrated silica, zeolites, sulfates, serpentine, and carbonates. Together, these mineral assemblages indicate widespread and diverse water–rock interactions over Mars’ history (e.g., Bibring et al., 2006; Milliken et al., 2008; Mustard et al., 2008; Ehlmann et al., 2009, 2010, 2011; Wray & Ehlmann, 2011), with Fe/Mg-smectite clays being the most abundant, followed by chlorites.

Longer-lived and widespread liquid water on Mars most likely occurred during the pre-Noachian to Noachian periods. Shorter-lived episodes of liquid water at the surface and in the subsurface, along with associated water–rock interactions, likely persisted into the Hesperian and Amazonian periods and may have formed protected or transient habitable environments (Cockell, 2014; Westall et al., 2015). Specifically, sedimentologic facies and geochemical characteristics at both ongoing rover missions sites show that lakes at Gale crater and deltas at Jezero crater persisted for at least centuries to millennia (likely longer), with redox couples (Fe/S) and nutrients sufficient for microbial metabolisms. The preservation potential of fine-grained mudstones and deltaic foresets informs taphonomic criteria for future biosignature searches on exoplanets (e.g., depositional “traps” inferred from global photometry and spectra) (Grotzinger, 2014; Mangold, 2021).

In addition to lacustrine settings, hydrothermal environments represent prime targets in the search for biosignatures, as they provide sustained heat and liquid water and can concentrate biological materials. Moreover, the highest potential for biosignature preservation is found in geologic materials that experienced limited diagenesis and metamorphism, minimizing overprinting of primary signatures (e.g., Sun & Milliken, 2018; Voigt et al., 2024; Millan et al., 2025, Opal-A), as well as limited exposure to the Martian surface and atmosphere, where oxidizing conditions and radiation would otherwise degrade or destroy the initial signature. The taphonomic criteria established by Mars exploration (particularly the preservation potential of fine-grained mudstones, deltaic foresets, and hydrothermally altered materials) provide a transferable framework for identifying high-priority biosignature targets on rocky exoplanets, where analogous depositional “traps” might be inferred from disk-integrated photometry and spectra.

The transit technique of exoplanet discovery has thus far detected the vast majority of known exoplanets (Christiansen et al., 2025). The photometric depth induced by a planetary transit scales as a simple relationship between the radius of the planet and host star, as: $\delta\approx(R_{p}/R_{\star})^{2}$. For example, a Mars-size ($R_{p}\approx 0.53$ $R_{\oplus}$) planet transiting a late M-dwarf host star ($R_{\star}\approx 0.12$ $R_{\odot}$) will yield a transit depth of $\delta\sim 1.6\times 10^{-3}$, or $\approx$1600 ppm. Although a transit signature of this amplitude is readily accessible from both ground and space-based facilities, such a planet-star configuration represents the best-case scenario. A Mars-size planet transiting K dwarf ($R_{\star}\approx 0.7$ $R_{\odot}$) or G dwarf ($R_{\star}\approx 1.0$ $R_{\odot}$) host star will yield transit depths of 49 ppm and 24 ppm, respectively. Such transit depths are challenging even from space-based photometry, particularly in the face of intrinsic stellar variability (Aigrain et al., 2004; Ciardi et al., 2011; Morris et al., 2020; Fetherolf et al., 2023). However, as described in Section III, several transiting Mars-size planets have indeed been detected, albeit within the high incident flux regime where mass measurements are enabled via radial velocities (RVs) or TTVs within systems where orbital resonance is present (Grimm et al., 2018; Agol et al., 2021).

The detection of true Mars analogs (at relatively low flux) will face the challenge of building up signal-to-noise over multiple transits over an extended period. Although results from the Transiting Exoplanet Survey Satellite (TESS) have significantly added to the known inventory of small planets around M-dwarfs, the 27-day sectors observations decreases sensitivity to longer period planets (Ricker et al., 2015; Kane et al., 2021c; Brady & Bean, 2022). On the other hand, the upcoming PLAnetary Transits and Oscillations of stars (PLATO) mission is designed for long-baseline, high-precision photometry on bright nearby stars (Rauer et al., 2014, 2025). PLATO will continuously monitor selected fields for 2–-3 years, enabling detection of terrestrial planets with much longer orbital periods than those represented in the TESS inventory, including Earth analogs around Sun-like stars and potentially planets in the outer HZs of M-dwarfs. By extending transit searches to periods of $\sim$1 year or longer, PLATO will complement the results of the Kepler and TESS missions, and may yield systems of multiple transiting terrestrial planets in strong resonance configurations (Heller et al., 2022; Matuszewski et al., 2023; Boettner et al., 2024).

Many of the initial exoplanet detections were the result of RV observations through spectroscopic observations of bright stars in the solar neighborhood (Butler et al., 2006; Schneider et al., 2011; Christiansen et al., 2025). The push towards RV precisions that enable terrestrial exoplanet detection has required the evolution of Extreme Precision RVs (EPRVs) and the diagnosis of contamination from stellar activity (Saar & Donahue, 1997; Fischer et al., 2016; Vanderburg et al., 2016; Cale et al., 2021). Numerous EPRV instruments have been developed and deployed, with most aiming for precisions better than $\sim$50 cm/s over long periods of observation (Gibson et al., 2016; Jurgenson et al., 2016; Schwab et al., 2016; Seifahrt et al., 2016; Pepe et al., 2021).

The RV amplitude of the Sun due to the gravitational influence of Mars is $K\sim 7.8$ mm/s, which is substantially beyond the reach of current EPRV instruments. By comparison, the RV amplitude induced on the Sun by Earth is $K\sim 8.9$ cm/s. Figure 3 shows the maximum (edge-on orbit) predicted RV amplitude for a Mars-mass planet in a circular orbit as a function of incident flux for three different host stars: an M-dwarf (solid line), a K-dwarf (dashed line), and a G dwarf (dot-dashed line). The RV amplitude for a Mars-mass planet at Mars equivalent flux around a K dwarf and M-dwarf are $K\sim 1.3$ cm ($\sim$0.815 AU) and $K\sim 4.1$ cm/s ($\sim$0.185 AU), respectively. In order for a Mars-mass planet to produce a similar RV amplitude to an Earth analog requires a flux of 10 $F_{\oplus}$ when orbiting an M-dwarf. Thus, the most promising near-term opportunity for the RV detection of Mars-mass planets requires targeting low-mass stars and short orbital periods.

The astrometric method of exoplanet detection, relying on precise measurements of stellar positions, has advanced considerably in recent years, primarily through the data releases from the Gaia mission (Perryman et al., 2014; Brandt, 2018, 2021; Gaia Collaboration et al., 2021). For example, the Gaia data have been used to refine the measured properties of known exoplanets and their host stars (Stassun et al., 2017; Berger et al., 2018; Fulton & Petigura, 2018; de Laverny et al., 2025) and have demonstrated that some previously detected RV companions do not lie in the planetary-mass regime (Kiefer, 2019).

The semi-amplitude of the astrometric effect scales linearly with the semi-major axis and mass of the planet, and thus favors massive planets at large separations from the host star. For example, a Neptune/solar analog located at a distance of 35 pc would produce an astrometric semi-amplitude of 44 $\mu$arcsec (Kane, 2011), or 154 $\mu$arcsec if located at 10 pc. By comparison, a Mars/solar analog located at 10 pc would induce a maximum astrometric semi-amplitude of only 49 nas. A Mars orbiting at Mars-equivalent flux around a K-dwarf and M-dwarf, also at 10 pc, would induce maximum astrometric semi-amplitudes of $\sim$37 nas and $\sim$20 nas, respectively. Thus, a Mars analog at 10 pc will produce an astrometric signature that lies considerably below the noise floor of Gaia ($\gtrsim$10–20 $\mu$as) or any currently planned astrometric mission.

Gravitational microlensing probes exoplanets through the transient magnification of a background “source” star as a foreground “lens” system closely aligns with the line of sight. In the planetary regime, the observable signature is a short-lived perturbation to an otherwise smooth point-lens light curve (Paczynski, 1986; Mao & Paczynski, 1991; Gould & Loeb, 1992; Paczynski, 1996; Gaudi, 2012). Microlensing is particularly sensitive to cold planets on wide orbits (typically near the Einstein-ring scale) and to planetary systems at kpc distances (Albrow et al., 2001; Cassan et al., 2012; Gaudi, 2012). This regime naturally includes terrestrial and sub-terrestrial planets in Mars-mass bins, provided that the planetary perturbation is sufficiently well sampled and not erased by finite-source effects.

Modern ground-based microlensing planet searches rely on wide-field, high-cadence monitoring of the crowded Galactic bulge from multiple longitudes. Major survey facilities include the Optical Gravitational Lensing Experiment (OGLE-IV) (Udalski et al., 2015), the Microlensing Observations in Astrophysics survey (MOA-II) (Sako et al., 2008), and the Korea Microlensing Telescope Network (KMTNet) (Kim et al., 2016). Despite these advances, detecting Mars analogs from the ground remains difficult for three related reasons: (i) the intrinsic rarity of microlensing alignments (necessitating very large monitored star counts), (ii) the relatively short duration of low-mass planetary perturbations, and (iii) systematic and sampling limitations from atmospheric seeing, blending in dense fields, weather, and diurnal gaps (Gaudi, 2012; Zang et al., 2025). In particular, small-planet anomalies can be missed or poorly constrained when data are interrupted during the perturbation (Zang et al., 2025). Even so, the ground-based low-mass frontier is approaching the Earth-mass regime in favorable (typically high-magnification) events. For example, Zang et al. (2021) reported KMT-2020-BLG-0414Lb with a planet-host mass ratio $q\sim 10^{-5}$. By contrast, Mars analogs around typical bulge lenses (0.3 $M_{\odot}$ at 6 kpc) correspond to $q\sim 10^{-6}$, where the signal timescale shrinks and the detection cross-section drops by an additional factor $\sim\sqrt{10}$. Consequently, present ground-based statistical constraints on cold-planet demographics are best developed at higher mass ratios rather than at Mars-like ratios.

The Nancy Grace Roman Space Telescope will conduct a dedicated Galactic Bulge Time-Domain Survey (GBTDS) with a high-cadence, near-infrared wide-bandpass sequence designed explicitly for microlensing demographics, including sensitivity to low-mass planets (Sajadian, 2021; Fatheddin & Sajadian, 2023). Space-based observations address the principal ground limitations by providing weather-free, high duty cycle during observing seasons, stable point-spread functions and calibrated photometry, and substantially reduced crowding/blending relative to typical ground seeing (Gaudi, 2012; Terry et al., 2025). Roman yield forecasts discuss total event samples of order $\sim 5\times 10^{4}$ microlensing events over the nominal six-season survey, depending on the event definition and impact-parameter cut (Saggese et al., 2025). Forecasts further indicate that Roman is expected to detect bound planets above $\sim$ 0.1 $M_{\oplus}$ (i.e., Mars mass) in statistically useful numbers and maintain sensitivity down to $\sim$ 0.02$M_{\oplus}$ under favorable circumstances (Penny et al., 2019). Overall, gravitational microlensing remains one of the most promising techniques for discovering cold, low-mass exoplanets at wide separations, with the caveat of relatively brief planetary deviations requiring sufficient observational cadence to characterize the planetary system.

Current ground-based direct-imaging surveys remain fundamentally optimized for young, self-luminous giant planets rather than cold Mars analogs. Even with extreme Adaptive Optics (AO), coronagraphy, and modern post-processing, survey-grade near-IR performance is typically in the $\sim$$10^{-6}$–-$10^{-7}$ contrast regime at arcsec separations (Chomez et al., 2025). Demographic constraints from large ground surveys therefore concentrate on companions of a few to $\gtrsim 10$ $M_{J}$ at multi-AU to $\sim$100 AU scales, not on sub-Earth rocky planets (Nielsen et al., 2019). For calculating the planet-star flux ratio of Mars analogs in reflected light, we adopt the formalism of Kane & Gelino (2010, 2011a). Assuming a geometric albedo of $A_{g}=0.15$, a planetary radius of $R_{p}=0.53$ $R_{\oplus}$, a semi-major axis of $a=1.52$ AU, and a Lambertian reflectance, we calculate a planet-star flux ratio of $\epsilon=1.1\times 10^{-11}$, which is several orders of magnitude fainter than present contrast floors. Thermal emission has a less demanding intrinsic contrast (Marley et al., 1999; Sudarsky et al., 2000; Kane & Gelino, 2011b; Guzewich et al., 2020), which, for a temperate Mars analog around a solar twin, is of order $\sim 10^{-8}$ near wavelengths of 10 $\mu$m, but ground-based sensitivity at thermal-IR wavelengths is critically limited by high thermal backgrounds (Rousseau et al., 2024).

Future space-based, high-contrast observatories are expected to shift toward the feasibility of Mars-analog detectability with the Habitable Worlds Observatory (HWO), explicitly conceived as a large UV/optical/IR facility designed to directly image and spectroscopically characterize potentially habitable planets (Harada et al., 2024; Kane et al., 2024; Stark et al., 2024b; Tuchow et al., 2024). Post-processing frameworks and operations concepts for a HWO-class coronagraph are being developed around the need to measure flux ratios of $\sim 10^{-11}$ (McElwain et al., 2025). For a Sun-like star at $d=10$ pc, a Mars-orbit analog has an angular separation of $\theta=0.152\arcsec$, which is outside the anticipated visible-light Inner Working Angle (IWA). However, a robust detection will require a substantially deeper search than for an Earth analog, as the reflected light flux ratio of a Mars analog ($\epsilon\sim 10^{-11}$) is roughly an order of magnitude fainter than for an Earth analog ($\epsilon\sim 10^{-10}$), requiring correspondingly longer integration times. This implies that Mars-analog yields will likely be driven by targeted deep observations of the nearest and most favorable stars rather than uniform completeness across the full exoEarth sample (Stark et al., 2024a). Stellar-type and distance trade-offs may be considered, meaning late-type stars offer higher reflected-light contrast but smaller angular separations. These couplings motivate yield optimization (and wavelength/IWA trade studies) as the operational path to recovering Mars-analog completeness once HWO-class stability and contrast are achieved (Stark et al., 2024a). A dedicated mid-IR space interferometer, such as the Large Interferometer For Exoplanets (LIFE), would greatly expand the search domain, enabling detection/characterization of hundreds of small planets, including potential Mars analogs (Quanz et al., 2022).

Atmospheric characterization of Mars-like exoplanets is scientifically compelling because even a tenuous CO₂-rich atmosphere can profoundly affect surface habitability and attest to a planet’s volatile history (Tinetti et al., 2005; Wolfe & Robinson, 2024). Alternatively, the lack of a substantial atmosphere may be consistent with strong escape mechanisms dominating over the atmospheric evolution, as has been inferred from JWST observations in several cases (Greene et al., 2023; Zieba et al., 2023). For example, a positive detection of H₂O, O₃, or other photochemical tracers may indicate remnant oceans or past high-energy irradiation, providing key benchmarks for comparative planetology (Schwieterman et al., 2018; Fujii et al., 2018; Ostberg et al., 2023). To investigate the observable signatures, we simulated transmission and reflectance spectra of a Mars analog using the Planetary Spectrum Generator (PSG; Villanueva et al., 2018). Each model uses the same Martian planetary configuration. The model utilizes a surface pressure and temperature of $\sim$6 mbar and $\sim$216 K, respectively. The bulk composition of the model atmosphere is dominated by CO₂ (96.934%), with N₂ (1.897%) and O₂ (0.144%) as secondary constituents, and trace amounts of CO, H₂O and O₃. The model atmosphere also includes two aerosol species; dust (2.921 ppm) and water ice (0.947 ppm), both of which have particle radii of 1.222 $\mu$m (Millour et al., 2015, 2024). These aerosols, along with the surface albedo ($A=0.2448$) and emissivity ($\epsilon$ = 0.7555) are wavelength-dependent, and the spectral signature’s detectability is influenced by the spectral energy distribution of the host star (Shields et al., 2016; Wunderlich et al., 2019). For solar-type stars, reflected light dominates at shorter wavelengths, meaning that there are stronger signals occurring at shorter wavelengths. On the other hand, M-type stars dominate in the near-IR, and will therefore have stronger features in that wavelength regime. This also means that surface and atmospheric features in the near-IR will become more prominent compared to that of a solar-type star. We note that these parameters are specific to present-day Mars, and that variations in surface pressure, albedo, or atmospheric composition (as might arise under different stellar spectra or earlier evolutionary states) would affect the depth and detectability of spectral features. For example, increasing the dust opacities could raise the effective albedo and reduce emission feature depths, while increasing the pressure would broaden CO₂ absorption features and enhance transmission signals.

Figure 4 shows the simulated transmission spectrum of a Mars analog around a Sun-like star at a distance of 10 pc, over a wavelength range of 0.2–10 $\mu$m (modeled on JWST/NIRSpec PRISM). The left and right vertical axes shows the depth of the absorption features for a G2 dwarf host and an M5 dwarf host, respectively. Strong CO₂ absorption features appear at 2.7 and 4.3 $\mu$m, in analogy to those seen in Earth’s spectrum, and span several atmospheric scale heights. The corresponding transit depth variation is extremely small: $<10^{-6}$ for a Mars-size planet and Sun-like star, making detection challenging even with JWST. Indeed, Kreidberg & Stevenson (2025) noted that rocky exoplanet spectra require multi-scale-height features for firm detections. Figure 4 demonstrates that, aside from the strongest bands, most absorption features remain near or below the noise floor, reflecting the high mean molecular weight and compact atmosphere of Mars analogs (Lustig-Yaeger et al., 2023; Rigby et al., 2023).

Direct imaging in reflected light offers complementary spectral information. Figure 5 displays model reflected spectra of a Mars analog observed at phase angles of 0°, 90°, and 135°, at a distance of 10 pcs, and assuming a 15 m coronagraph-equipped telescope over 0.2–10 $\mu$m (Kawashima & Rugheimer, 2019), based approximately on the Large Ultraviolet Optical Infrared Surveyor (LUVOIR) A-VIS instrument (The LUVOIR Team, 2019). The reflected spectra in Figure 5 show modest signatures of CO₂ and surface mineral bands, with overall brightness declining at larger phase angles. As noted by Roccetti et al. (2025b), reflected-light spectra as a function of wavelength and phase angle encode atmospheric and surface properties (e.g., albedo variations and molecular absorptions such as O₂, H₂O, CH₄, CO₂). By comparing the three spectra in Figure 5, the full-phase (0°) spectrum is approximately uniformly bright across wavelength, while at 90°–135°the absorption features are weaker and the flux falls sharply at NIR wavelengths due to the reduction in Rayleigh and forward Mie scattering (Roccetti et al., 2025b, a). These variations can be exploited to retrieve composition: for example, deep CO₂ bands at 4.3 $\mu$m and weaker H₂O or O₃ features will modulate the reflected color differently from bare surface (García Muñoz & Cabrera, 2018). Thus, direct imaging at HWO-class contrasts ($\sim$$10^{-10}$) is expected to yield more abundant molecular information for nearby Mars analogs than transit spectroscopy alone.

Studying Mars provides numerous lessons that translate directly into the interpretation of rocky exoplanet data. Regarding escape-driven desiccation and abiotic O₂, long-term H escape from H₂O can leave O₂/O₃-rich atmospheres without the need to invoke biological processes. However, it is important to note that coupled H and O escape may prevent significant O₂ accumulation in the atmosphere (McElroy, 1972; Jakosky et al., 2018), and thus the degree to which abiotic O₂ can build up depends on the balance between production and loss channels. Mars further constrains rates and controlling physics across dust seasons, informing false-positive assessments around M-dwarfs (Heavens et al., 2018). Mars provides empirical tests of general circulation models (GCMs) and analytic criteria for CO₂ freeze-out, and these results inform models of tidally locked exoplanets in/near the HZ, where atmospheric mass and heat transport determine survivability (Wordsworth, 2015), though caution is warranted when extrapolating GCMs validated under present-day Martian conditions to dynamically distinct regimes where boundary conditions and processes may differ substantially. The noble-gas and hydrogen isotope fractionation under known escape regimes calibrate how we read future high-precision spectra of small exoplanet atmospheres (Mahaffy et al., 2013). Mars also informs the potential for habitable windows in the absence of plate tectonics. Sedimentary facies at Gale/Jezero demonstrate that long-lived surface waters can exist transiently, even as a planet trends toward aridity—guiding expectations for temporal habitability on small exoplanets (Grotzinger et al., 2014).

As detailed in Section IV, the detection of Mars analogs is a challenging task as the required sensitivity lies at the edge or beyond for the majority of exoplanet detection techniques. The most promising near-term discovery pathways include transit+TTV observations of compact systems around late M-dwarfs (Gillon et al., 2017; Agol et al., 2021) and the Roman microlensing survey of bulge fields (Penny et al., 2019). The long-term goals are to characterize the atmospheres of low-mass planets, empirically modeling their evolutions from early degassing stages through to later more tenuous states. Validation of retrieval techniques based on disk-integrated spectra of Solar System planets have provided invaluable insight into the requirements of achieving such goals (Kane et al., 2021b; Robinson & Salvador, 2023). Combining thermal-infrared observations of the nearest Mars-analog exoplanet candidates (Wolfe & Robinson, 2024), reflected-light imaging with HWO-class facilities, and mid-infrared interferometry with concepts such as LIFE (Quanz et al., 2022) will ultimately quantify the occurrence rate and atmospheric properties of true Mars analogs, although confirmation of their exact mass, composition, and habitability will remain observationally demanding.

In the most recent planetary science Decadal Survey, Origins, Worlds, and Life¹¹1https://doi.org/10.17226/26522, Mars features heavily both as a worthy Solar System target in its own right and as a basis for better understanding rocky exoplanets. Through a series of Strategic Research (SR) activities within 12 Priority Science Questions, key Solar System science topics can be addressed by studying myriad aspects of the Mars system. For example, in Priority Question 3, ”Origin of Earth and Inner Solar System Bodies”, SR 3.4 includes the objective to ”Determine the formation time of Mars through isotopic analyses of diverse Martian samples,” something that could be achieved through, for example, the return to Earth of samples from the Red Planet. Similarly, in Priority Question 6, ”Solid Body Atmospheres, Exospheres, Magnetospheres, and Climate Evolution” there are two SRs specific to understanding the atmospheric evolution of Mars. One, SR 6.1, states ”Constrain the earliest stages of atmospheric evolution on Venus, Mars, and Titan by measuring noble gas abundances and isotopic fractionation to sufficient precision to quantify their minor isotopes.” For the second, SR 6.2, the objective is to ”Determine how and why Mars’s climate has changed over orbital timescales by performing radar and spectroscopic mapping of the polar layered terrain and by making in situ measurements of their structure and composition (thickness of layers, dust content, and isotope ratios) and their local meteorology (including volatile and dust fluxes).”

These are specific tasks that can be carried out at Mars or on Earth with Mars samples, and are directly in service of better understanding the modern state of the planet and how it got that way. In that same Priority Question, however, SR 6.6 reads ”Investigate the microphysical parameters that influence the formation of clouds in planetary atmospheres (primarily… Mars…) by determining the distribution, nature, and abundance of clouds and the composition and particle size of the droplets comprising them and cloud condensation nuclei around which they form.” This investigation would have direct bearing on efforts to characterize rocky-exoplanet atmospheres in terms of distinguishing potential cloud spectral signatures from that of the bulk atmosphere.

Moreover, Origins, Worlds, and Life also includes an exoplanet-focused chapter, written explicitly with the goal of using that understanding of Mars to further our (remote) exploration of terrestrial exoplanets. For instance, in the Priority Question 12, ”Exoplanets”, SR 12.10 is to ”Determine the properties of the atmospheres of terrestrial planets (i.e., … Mars) that would be observable on exoplanets to build a foundation for atmospheric characterization of analogue exoplanets through coordinating in situ/remote sensing measurements and theoretical studies of wind velocities, radiative balance, cloud dynamics, and atmospheric compositing as function of orbital phase, local time, and solar conditions.” SR 12.10 states ”Improve exoplanet habitability predictions for cold, low-mass planets by determining the key factors that made Mars habitable 3–4 Ga, via a combination of in situ geological and atmospheric analysis and sample return, orbital observations, and climate modeling.” And in SR 12.11, the decadal survey advises ”Study methods to discriminate past and present false positive biosignatures on Solar System bodies (e.g., abiotic O₂ on… Mars) from true biosignatures to inform false positive discrimination methods for exoplanets through in situ, remote sensing, theoretical/modeling studies, analog field research, and laboratory studies that characterize remotely observable properties of these features.”

Together, these activities if implemented would advance not only our understanding of Mars’ past and present states but would more firmly place it as a foundation for how we identify and interpret Mars-analog exoplanet features and properties.