1

WHEN DID MARS BECOME BIPOLAR?: AN ANALYSIS OF THE KEY FACTORS IN
THE LATE NOACHIAN-AMAZONIAN CLIMATE TRANSITION FROM AN
ALTITUDE-DOMINANT TEMPERATURE DISTRIBUTION (ADD) TO A LATITUDEDOMINANT DISTRIBUTION (LDD).
James W. Head1, Robin D. Wordsworth2 and James L. Fastook3, 1Brown University, Providence, RI 02912 USA,
2
Harvard University, Cambridge, MA 02138 USA, 3University of Maine, Orono, ME 04469 USA
(james_head@brown.edu).
Introduction: The history of the atmosphere and climate of Mars (Fig. 1) is one of the major current outstanding question in planetary science [1-2]. Mars today
has an ~6.5 mbar CO2-dominant atmosphere, and a hypothermal (~210K MAT), hyperarid polar desert climate
(10-20 precipitable microns of H2O in the atmosphere)
[3], with surface water sequestered in the polar caps [4].
Yet in earlier history (Noachian, N), there is strong evidence for sustained periods of abundant surface liquid
water, both flowing (valley networks) and standing
(open-closed basin lakes, oceans) (Fig. 1), correspondingly higher atmospheric pressure (Patm ~1-2 bar?) [5],
and surface temperatures (MAT in excess of 273K) [6-7]
(Fig. 2).
Two general classes of models have been proposed to
account for these observations (Fig 2): ‘Warm and Wet’
scenarios [e.g., 6-7] propose N clement conditions (MAT
>273K), with rainfall, runoff, fluvial, lacustrine activity,
transitioning to the type of climate observed today. ‘Cold
and Icy’ scenarios (e.g., 8-9) point to the ‘faint young
Sun’, and predict a resulting MAT of 225K (Fig. 2), with
273K MAT reached nowhere on Mars even with Patm
between 1-7 bars; thus, this ambient ‘Cold and Icy’ climate requires significant transient input of greenhouse
gases to elevate temperatures to cause melting of snow
and ice to produce the observed fluvial and lacustrine
features [e.g., 10-11
Currently debated issues: These include: What was
the nature of the ambient (background) N climate? How
long did it last and was it sustained or were changes episodic? What caused its transition to the climate of today
and when and how did this occur? What was the nature
of the hydrological system (horizontally stratified or vertically integrated) [12] and how did it change with time?
What was the budget of surface/near-surface water [4]
and how was it distributed? What were the conditions
that led to observed fluvial, lacustrine and possibly oceanic environments (duration, periodicity, episodicity)?
What were the warming greenhouse gases, their sources
[e.g., 10-11], and the mechanisms for sustaining them in
the atmosphere? What was the mean annual temperature
(MAT) as a function of time and did global temperature
distribution (GTD) change? What are the atmospheric
loss rates to space [13] and how did they vary with time?

Major obstacles to resolving these questions: These
include: 1) Identification of sufficiently robust greenhouse gas sources to produce and sustain a clement N
ambient climate, or a prolonged transient heating event in
the ‘Cold and Icy’ scenario, 2) Mechanisms to account
for the decrease in Patm from N-H conditions (>1 bar?) to
today (Patm 6.5 mbar)(Fig. 3), and 3) Identification of the
fate of the very large volumes of surface water required
by an ambient N clement climate [4].
Here we propose a scenario that addresses several of
these obstacles and makes a number of testable predictions for future research and exploration. We start with
the currently observed climate as a known benchmark
and then assess the most parsimonious Noachian conditions that could have led forward to this benchmark
(Figs. 2,3), tracking the distribution and fate of water,
and accounting for the geologic observations (Fig. 1).
Current Mars Climate Benchmark: The extremely
low current MAT (~213K) and Patm (~6.5 mbar) (Figs. 23) result in very low atmosphere water content [14], poor
atmosphere-surface thermal coupling, and surface temperature distributions [3] that are dominated by latitudedependent (Fig 4a), not altitude-dependent (Fig 4b), effects. This has three effects: 1) The equator-to-polar temperature gradient is significant; despite 213K MAT,
equatorial peak daytime and seasonal surface temperatures can exceed 273K [15]; 2) Water is metastable, with
ice ablation dominated by sublimation [16]; 3) The North
and South polar regions are cold traps that sequester the
vast majority of surface water in thick ice caps [3-4].
Amazonian Mars Climate: Amazonian climate history is thought to be largely similar to that of today [17],
with spin-axis/orbital variations (primarily obliquity [18])
from time to time mobilizing polar ice and transporting it
to lower latitudes to form local and regional glacial deposits [19-21]; evidence for associated Amazonian ice
melting is minimal and linked to local conditions [22-23].
The global water budget remains substantially the same
(within a factor of two of today) [4]. These observations
suggest that any major transition from 1) N to current
benchmark Patm (Figs. 2-3) and 2) N to current global
water inventory [4], must have occurred before the Amazonian.
The Nature of the Noachian Atmosphere and Climate: A common characteristic of all models for the ear-

2

ly Mars climate is the adiabatic cooling effect (ACE) [89, 24-26] (Fig. 4b). If the atmospheric pressure exceeds a
few tens of millibars, atmosphere-surface thermal coupling becomes effective and altitude-dependent atmospheric temperatures begin to dominate over the latitudedependent temperatures that characterize the Amazonian
benchmark atmosphere (compare Fig. 4a-LDD; b-ADD).
This effect is one of the mainstays of all N climate models. Under ADD conditions, there is little to no North
polar cap, Tharsis is a second pole, analogous to the Tibetan Plateau on Earth today, and the south polar region
is the third locus of snow and ice accumulation (Fig. 4b).
A Parsimonious Scenario for the Noachian Ambient
Climate and its Transition to the Current Benchmark
Climate: On the basis of the roles of XUV-driven loss,
the integrated impact flux and Early Noachian basin formation (Hellas, Argyre, Isidis) in stripping a significant
part of the primary atmosphere [27], and the relatively
low contributions of middle Noachian volcanic outgassing to the secondary atmosphere [1,4,28], it is interesting to explore scenarios where 1) the Middle to Late Noachian Patm was in the several hundred millibar range [5,
29], and 2) where the surface water budget was within a
factor of two of its current value [4]. Under these conditions, the MAT is predicted to be ~225K the ADD would
dominate the surface water budget distribution (Fig. 4b),
and snow and ice would preferentially accumulate in
three cold traps (Fig. 2, right): 1) the south circumpolar
region, 2) the southern uplands, and 3) the Tharsis rise.
The north polar region, situated deep in the relatively
warmer northern lowlands, would not be the site of significant snow and ice accumulation and there would be
no significant north polar ice cap [8-9, 30-31]. This scenario would define the ambient Noachian climate (Fig.
4b), but does not account for the abundant observed fluvial and lacustrine activity.
Nature of the Middle-Late Noachian to HesperianAmazonian Transition Period: We propose that a modest decrease in Patm of the ambient Noachian climate
could plausibly account for the observed Late Noachian
fluvial and lacustrine activity. In this scenario, modest
atmospheric loss during this period could initiate a transition from a global adiabatic cooling dominant atmospheric regime (altitude dominant temperature distribution;
ADD; Fig. 4b) to a global latitude dominant atmospheric
temperature regime effect (LDD; Fig. 4a) similar to today’s benchmark climate. Climate models suggest that
this transition period would have the following elements:
1) Mean Annual Temperature: MAT would not vary
significantly, but global temperature distribution would
transition from ADD (topography) patterns to LDD (latitude) patterns.
2) Reorganization of water ice cold traps: Mars would
become bipolar, similar to the current benchmark climate; the southern uplands and Tharsis ADD cold traps

would be replaced by a robust LDD North polar ice cap,
and the southern ice cap would become smaller (it is currently 2.5x less in area than in the N [32]).
3) Equatorial and mid-latitude surface temperatures
begin to rise: During the transition to dominantly LDD,
warmer temperatures (including any >273 K peak summer daytime/seasonal (PDT/PST) temperatures [33]
would migrate from the northern lowlands toward equatorial regions, causing regional MAT to increase. Under
peak seasonal conditions, transient heating and melting of
snow and ice sequestered in the ADD southern uplands
cold traps could occur (Fig. 2, right), such as is observed
in the Antarctic Dry Valleys [34]. Such a transitional
period could produce a prolonged phase of periodic melting and fluvial/lacustrine activity, easily capable of producing sufficient meltwater to fill the observed lakes
[35]. Meltwater would return to cold traps between peak
T phases; recycling reduces the total amount of water
required from 5000 m GEL [36] to a more plausible 640
m [37]. Over time, water would be preferentially lost
from the uplands to the newly growing North Polar Cap.
4) Migration of the Equilibrium Line Altitude (ELA) to
Higher Altitudes: During the ADE-LDE transition, the
global distribution of warmer MAT isotherms would migrate from the lowest regions (Northern Lowlands, Hellas; Fig. 4b) toward a latitude dominant distribution (Fig.
4a). Accompanying this transition would be a rise in altitude of the ELA position into the snow-and-icedominated mid-latitude-equatorial highlands cold traps.
5) Formation of Observed Fluvial and Lacustrine Features: Rise of the ELA into the icy highlands is predicted
to cause ablation of snow and ice and ultimate demise of
the icy highlands, as water migrated to the newly growing
North Polar Cap. Gradual sublimation of high-altitude
snow and ice and its migration to new cold traps (primarily the North Pole) is accompanied by snow and ice melting during periods when PAT and PST exceed 273K.
Such PAT/PST phases [33] of heating and melting of
snow and ice would be sufficient to provide volumes of
meltwater comparable to those required to form the observed fluvial and lacustrine features [35, 37].
6) Predicted Stratigraphy and Timing of Fluvial and
Lacustrine Deposits: In this transition period, melting
episodes are predicted to be largely seasonal (PDT, PST)
with periodic annual and decadal warming variations
extending the duration, analogous to the types of seasonal
top-down melting episodes seen in the McMurdo Dry
Valleys [34]. Following these peak warm episodes,
meltwater is predicted to return to the cold traps until
ablation has depleted the reservoirs at the end of the transition to a bipolar Mars. Stratigraphy of lacustrine deposits should exhibit multiple layers related to this activity.
Currently unknown is the duration of this transition, but it
is likely to be less than 107-108 years [38], during which

3

time spin axis/orbital changes (obliquity/eccentricity)
may also influence the duration of melting phases.
7) Growth of the North Polar Cap: During this climate
transition, Mars becomes bipolar, with the North Polar
Cap growing towards its current large volume at the expense of the previous ADD cold traps. Due to known
Amazonian obliquity variations, the original deposits
from the pre-Amazonian North Polar Cap have been recycled back and forth to mid-latitude ice-deposits numerous times [19-21].
8) Predicted Temporal Changes During the Transition: The hypothesis presented here predicts that ablation and melting of ‘icy highlands’ cold traps should proceed from lower to higher elevations as the ELA migrates
vertically, governed by the transition toward global latitude-dependent distribution (from Fig. 4b to 4a). These
predictions can be tested with analysis of the stratigraphy
and timing of features related to the rising ELA [39-42].
9) Final Desiccation of the Equatorial Region: The final stage of the transition is marked by the depletion of
the ADD snow and ice reservoirs in the equatorial and
mid latitude cold traps (Figs. 2-right; 4a), and progressive dehydration of the upper part of the cryosphere to
produce the global distribution of surface and nearsurface water seen today [43-44].
We describe this scenario as ‘parsimonious’ because
it: 1) involves a plausible Noachian Patm, 2) utilizes
known global atmospheric effects (Patm-dependent ADDLDD conditions), 3) requires minimal changes in global
MAT (Figs. 2-3), 4) may require no major and persistent
influx of warming greenhouse gases, 5) calls on a plausible global water budget throughout [4], 6) requires modest atmospheric loss to space [13], 7) provides a more
plausible D/H ratio history, and 8) requires no Tharsisinduced true polar wander (TPW) to account for valley
network distribution patterns [45].
Tests of the Model: Here we describe the significant
questions that this scenario raises and how the hypothesis

might be further tested, refined, or rejected: 1. Is a Noachian Patm of several hundred mbar [5, 29] plausible and
what is the geologic evidence for this? 2. What is the
Patm tipping point at which the ADD dominant scenario
begins to decay to the LDD dominant scenario (compare
Fig. 4b,a), how long does this transition take, and does it
change global atmospheric circulation patterns significantly? 3. What are the effects of variations in obliquity
during the transition, including potential very low obliquity-induced atmospheric collapse? 4. When did the
Tharsis Rise form (including possible TPW [45]) and
what effect did it have on the atmosphere and climate? 5.
Are documented geologic events in the Hesperian transitional period (Fig. 1) (e.g., volcanic resurfacing, sulfate
deposits [46-47], outflow channel formation) consistent
with this scenario? 6. Are the major periods of mineralogical alteration (Fig. 1) (phyllosilicates, sulfates, anhydrous oxidation [2]) consistent with this scenario? 7.
Are the major findings of the robotic surface exploration
missions MER, MSL (Gale CBL and Jezero OBL), and
Zhurong (southern Utopia Planitia) consistent with this
scenario? 8. Are the predicted rates of volatile loss to
space envisioned by this scenario consistent with
MAVEN results [13]? 9. Are the observed characteristics, distribution and duration of fluvial and lacustrine
environments (valley networks and lakes) [27,38] and
crater degradation history [48] consistent with this scenario? 10. Are the South polar/circumpolar deposits
(Dorsa Argentea Formation [32]) and their timing consistent with this scenario?
Current Work: We are currently exploring several of
these questions using geologic observations and mapping
(the Hesperian sulfate transition period [46-47]) and climate modeling (the nature of the change from ADD to
LDD; Fig. 2).

4

Fig. 1. Diagrammatic representation of the main themes in the geologic [1] and alteration history [2] of Mars.

Fig. 2. Mean Annual Temperature (MAT)-Time (Noachhian-Early Hesperian) relationships: Left: Baseline examples. Middle:
“Warm and Wet” Ambient Climate Scenario [6-7] with vertically integrated hydrological system (inset) [19]. Right: “Cold and
Icy “ Ambient Climate Scenario [9], with horizontally stratified hydrological system (inset) [19] and the distributi of snow and
ice above an Equilibrium Line Altitude (ELA) of +1 km [34].

Fig. 3. Requirements for changes from a Noachian Ambient Climate to that of
todayin atmospheric pressure (left) and temperature (right) for the two models.

Fig. 4. The modern LDE compared to the
early Mars ADE: a) MAT from 3D GCM at
125 mbar Patm (Amazonian). b) MAT from
3D GCM at 1 bar Patm (Noachian?) [8-9].

References:
1. Carr & Head, 2010, EPSL 294; 2. Bibring et al, 2006, Science 312; 3. Haberle et al (2017) The Atmosphere and Climate of
Mars, Cambridge; 4. Carr & Head, 2015, GRL 42; 5. Kite, 2019, Space Sci. Rev. 215; 6. Craddock & Howard, 2002, JGR 107;
7. Ramirez & Craddock, 2018, Nature Geosci. 11; 8. Forget et al., 2013, Icarus 222; 9. Wordsworth et al., 2013, Icarus 222; 10.
Wordsworth et al., 2021, Nat. Geosci. 14; 11. Wordsworth et al., 2017, GRL 44; 12. Head, 2012, LPSC 43, #2137; 13 Jakosky et
al., 2018, Icarus, 351; 14. Trokhimovskiy et al., 2015, Icarus, 251; 15. Haberle et al (2017) The Atmosphere and Climate of
Mars, Cambridge; Ch. 4; 16. Hecht, 2002, Icarus 156; 17. Haberle et al (2017) The Atmosphere and Climate of Mars, Cambridge; Ch. 16; 18. Laskar et al., 2004, Icarus 170; 19. Head, et al., 2003, Nature 426; 20. Madeleine et al., 2009, Icarus 203; 21.
Forget et al., 2006, Science 311; 22. Weiss et al., 2017, GRL 44; 23. Fassett & Head, 2007, Icarus 189; 24. Palumbo & Head,
2018, GRL 45; 25. Steakley et al., 2019, Icarus 330; 26. Gurzewich et al., 2021, JGR-P 126; 27. Fassett & Head, 2014, Icarus
211; 28. Grott et al., 2011, EPSL 308; 29. Bristow et al., 2017, PNAS 114; 30. Wordsworth et al., 2015, JGR-P 126; 31. Wordsworth, 2016, Ann. Rev. Earth Planet. Sci. 44; 32. Head & Pratt, 2001, JGR-P 106; 33. Palumbo et al., 2018, Icarus 300; 34.
Head & Marchant, 2014, Ant. Sci. 26; 35. Fastook & Head, 2015, PSS 106; 36. Luo et al., 2017, Nature Comm. 8; 37. Rosenberg et al., 2019, Icarus 317; 38. Buhler et al., 2014, Icarus 241; 39. Boatwright & Head, 2021, Planet Sci J, 2; 40. Boatwright &
Head, 2022, Planet Sci J, 3; 41. Bouquety et al., 2020, Geomorphology 350; 42. Fastook & Head, 2021, LPSC 53, 1287; 43.
Feldman et al, 2002, Science 297; 44. Boynton et al. 2002, Science 297; 45. Bouley et al., 2016, Nature 531; 46. Head & Wilson, 2020, LPSC 51, 2048; 47. Kreslavsky & Head, 2020, LPSC 51, 1828; 48. Mangold et al., 2012, JGR 212.