EARLY MARS: RECONCILING CLIMATE AND GEOCHEMISTRY
R. Wordsworth1,2,*, A. H. Knoll2, J. Hurowitz3, M. Baum2, B. Ehlmann4,5, J. Head6, K. Steakley7 and F. Ding1.
1
School of Engineering and Applied Sciences, Harvard University. 2Department of Earth and Planetary Sciences,
Harvard University. 3Department of Geosciences, Stony Brook University. 4Division of Geological and Planetary
Sciences, California Institute of Technology. 5Jet Propulsion Laboratory, NASA. 6Department of Geological Sciences, Brown University. 7Space Science and Astrobiology Division, NASA Ames Research Center / Bay Area Environmental Research Institute. *rwordsworth@seas.harvard.edu.
Introduction: The last few years has seen substantial
progress on our understanding of the faint young Sun
problem and the nature of Mars’s early climate. Nonetheless, significant uncertainties remain. Here, we discuss some of our progress on the early Mars climate
problem since the last Mars atmosphere workshop,
with a focus on two issues: the degree to which the
early climate was permanently vs. episodically warm,
and the extent to which scenarios for warming can be
reconciled with the rich geochemical record observed
by both rovers and orbiters.
Observational evidence: Extensive geologic evidence indicates that 3-4 Ga, surface conditions on
Mars were dramatically different, with multiple episodes of fluvial erosion, aqueous alteration and sediment deposition [1]. The most plausible explanation
for this is greenhouse warming from a thicker early
atmosphere, although the details continue to be debated [2,3]. Various analyses suggest that in total, between 104 and 107 years of warm conditions were required to erode observed valley networks, deposit sediment in craters and form weathering sequences [4-7].
In addition, abundant ancient exposures of unaltered
igneous minerals such as olivine and the relative absence of surface carbonates on Mars indicate that large
bodies of surface liquid water were likely not present
over periods much longer than a few million years
[8,9]. Today Mars’ surface is highly oxidized, but its
mantle appears more reduced than Earth’s, and some
Mars meteorites preserve non-zero sulfur massindependent fractionation (MIF) signatures [10], suggesting anoxic atmospheric intervals. Rover observations at both Meridiani Planum and Gale Crater have
also shown strong variability in mineral redox chemistry [11], with oxidizing surface conditions around the
Noachian-Hesperian boundary ca. 3.5 Gya (Figure 1)
suggested by the presence of concentrated hematite
and manganese oxide [11-13]. The manganese deposits
in particular appear to require the simultaneous presence of liquid water and strong oxidants such as O2,
ultraviolet radiation or possibly chlorates [12,14].
Climate and water cycle modeling: Recent work
indicates that a thicker CO2 atmosphere combined with
smaller amounts of reducing gas species (H2, CH4) can

warm early Mars to the melting point of liquid water
[15-18]. In contrast, warm conditions in highly oxidized atmospheres appear much more difficult to
achieve. Clouds (CO2 or H2O) may provide some additional warming, although not enough to warm early
Mars alone given our current understanding of cloud
microphysics (see also Ding et al., this conference).
Adiabatic cooling under a thicker CO2 atmosphere
provides a recharge mechanism for valley network
source regions, either as precipitating snow [2] or rainfall [19] (Figure 1).
Stochastic atmospheric evolution model: To investigate how these diverse strands of geologic evidence
can be reconciled with climate modeling results, we
have developed a stochastically forced atmospheric
evolution model (Fig. 2). We represent the release of
reducing gases to the atmosphere due to volcanism
[15], meteorite impacts [16,17] and crustal alteration
[17,20] in a generalized way by randomly sampling
from a power law distribution. Escape of hydrogen and
oxygen to space via diffusion-limited and non-thermal
escape processes is also included, as is oxidative
weathering of the crust. Water loss to space is constrained via D/H data. Changes in surface temperature
over time due to greenhouse warming by CO2 and reducing gases are also taken into account. We find that
the model produces rapid and repeated fluctuations
between warmer, more reducing and colder, more oxidizing conditions during Mars' early history (Figs. 23), with integrated warm periods sufficient to match
geomorphic observations for mean rates of reducing
gas input in the range of upper limits from volcanism,
impacts and crustal processes.

Figure 1: Schematic of the water-limited ‘icy highlands’ scenario for early Mars (from [2]).

A
equiv. pressure [bar]

10

0

oxidizing (O 2 )
reducing (H 2 )

10

-2

10

-4

4

B

3.5

3

2.5

2

1.5

1

0.5

0

t [Gya]

300
mean T

Tsurf [K]

280

surf

= 222 K

260
240
220
200

C

4

3.5

3

2.5

2

1.5

1

0.5

0

t [Gya]
5

t warm [My]

4

p

p

2
1
0

D

CO2

3.5 Gya = 2 bar

3

p
4

3.5

3

2.5

2

1.5

CO2

CO2

3.5 Gya = 1.5 bar
3.5 Gya = 1 bar

1

0.5

t [Gya]

Figure 2: Observations vs. stochastic evolution model predictions over Mars' history. A) Timeline of major events on
the martian surface from geologic observations. B) Changes
in the net redox state of the atmosphere vs. time due to the
competing effects of episodic release of reducing gases, atmospheric escape and surface weathering. C) Corresponding
surface temperature evolution and D) cumulative integral of
time spent with surface temperature >273 K.

0

Figure 3: Integrated duration of warm intervals as a function
of H fluxes into the martian atmosphere. Colored lines show
the integrated warm period from 4.1 Gya onwards as a
function of the average rate of reducing gas input to the
atmosphere over the entire 4.5 Gy simulation interval. Colors
indicate the assumed variability ratio (specifically, the
maximum possible input rate in a 0.1 My interval relative to
the time-varying mean value).

Sulfate, Hematite and Manganese Chemistry: We
hypothesize that the increase in sulfate deposition from
the Noachian into the Hesperian was also linked to
changes in Mars’s atmospheric composition through
time, for three reasons. First, oxidation of the upper
martian mantle could potentially have moderately increased the fraction of S outgassed by the Hesperian
[21]. Second, preservation of evaporite deposits
against later remobilization would have only been favored toward the waning stages of the Noachian, from
3.5 Ga onward [22]. Third, sulfate minerals can be
converted to reduced minerals like pyrite or pyrrhotite
when dissolved under reducing conditions, particularly
during high temperature post-impact conditions that
would have occurred frequently during the early to
mid-Noachian. Redox fluctuations can also help explain the appearance of sedimentary hematite and
manganese oxide deposits. Our model predicts intervals of elevated atmospheric oxygen throughout the
Noachian and Hesperian. In these intervals, upper layers of the regolith would slowly oxidize via dry weathering reactions, formation of oxychlorine species and
adsorption of volatile species like H2O2. During warming intervals, some of this regolith would have been
transported as sediment to standing bodies of water.
The extremely slow rate of reaction of H2(aq) with oxidized species like Fe3+(aq) [13,23] mean these sediments
would have remained out of chemical equilibrium with
the atmosphere, plausibly leading to the local formation within the sediments of oxidized minerals such
as hematite over time. Warming of early Mars on 104107 year timescales under O2-rich atmospheric conditions is not currently predicted in state-of-the-art climate models, but if it did occur, it could also have led
to the formation of diagenetic hematite and manganese
in Hesperian sediments. In any case, shorter periods of

warming (up to a few years) would still have occurred
episodically when the atmosphere was O2-rich, via the
mobilization of surface ice deposits by redox-neutral
bolide impacts [24]. This would have provided additional routes for the formation, transport and deposition of highly oxidizing minerals, providing a plausible
explanation for the Mn-rich fracture-filling materials in
the Kimberley Formation at Gale Crater [12].
Conclusions and Future Directions: An episodically
fluctuating scenario for Mars' chemical and climate
evolution appears consistent with several key features
of the planet's geological record. It also predicts formation scenarios for various aqueous minerals that can
be tested further by upcoming rover missions and
eventually by investigation of returned samples. In
addition, future modeling work should investigate the
atmospheric and aqueous chemistry of varying redox
states on early Mars in greater detail.
References: [1] Carr., M. H. (2007), vol. 6. Cambridge
University Press. [2] Wordsworth, R. (2016), Ann Rev
Earth and Plan Sci, 44(1). [3] Haberle, R. M. et al.
(2017), Cambridge University Press. [4] Hoke, M. R.
T. et al. (2011), Earth Plan Sci Lett, 312(1):1–12. [5]
Bishop , J. L. et al. (2018), Nat Astro, 2(3):206. [6]
Fukushi, K. et al. (2019), Nat Comm, 10(1):1–11. [7]
Kite, E. S. (2019), Space Sci Rev, 215(1), p. 10. [8]
Niles, P. B. et al. (2013), Space Sci Rev, 174(1-4):301–
328. [9] Ehlmann, B. L. and Edwards, C. S. (2014),
Ann Rev Earth Plan Sci, 42:291–31. [10] Lorand, J.-P.
et al. (2020), Meteorit Plan Sci, 55:2097–2116. [11]
Hurowitz, J. A. et al. (2017) Science, 356, eaah6849.
[12] Lanza, N. et al. (2016) Geophys Res Lett,
43:7398–7407. [13] Hurowitz, J. A. et al. (2010). Nat
Geosci, 3:323. [14] Mitra, K. et al. (2020) AGU Fall
Meeting 2020. [15] Ramirez, R. M. et al. (2014), Nat
Geosci, 7(1):59–63. [16] Haberle, R. M. et al. (2019),
Geophys. Res Lett, 46:13355-13362. [17] Wordsworth,
R. et al. (2017), Geophys Res Lett, 44:665-671.
[18] Turbet, M., Boulet, C. and Karman, T., 2020. Icarus, 346, p.113762. [19] Turbet, M. and Forget, F.,
2021. arXiv preprint arXiv:2103.10301. [20]
Chassefière, E. et al. (2016), Meteorit Plan Sci,
51(11):2234–2245. [21] Gaillard, F., et al. (2013),
Space Sci Rev, 174:251–300. [22] Milliken, R. E., et
al. (2009), Geophys Res Lett, 36:L11202. [23] Tosca,
N. J., et al. (2018), Nat Geosci, 11:635–639. [24] Palumbo, A. M. & Head, J. W. (2018), Meteorit Plan Sci,
54:687–725.