The Snowball Earth theory has been gaining momentum since 1992 when Joseph
L. Kirschvink of the California Institute of Technology coined the term.
Kirschvink’s initial model of an ice-covered Earth was just a beginning.
When Paul Hoffman of Harvard University and other colleagues picked up the
hypothesis, packed in more evidence, and tossed it to the science community
again, it really started to stick.

“This session brings together advocates, antagonists, and undecideds who are
experts in geology, atmospheric science, marine geochemistry, and evolutionary
biology,” explained Paul Hoffman, the session chair. “They’ll bring a mix
of new discoveries and contrasting perspectives on a highly controversial
Snowball Earth theory.”

One debated issue concerns the role of methane in the snowball cycle.

Dan Schrag from Harvard University will propose a new, counterintuitive model
for the initiation of a snowball Earth. The model involves a slow leak of
methane from organic-rich sediments into an atmosphere that was less
destructive of methane (i.e. less oxidizing) than the present atmosphere.
Because methane is a very potent greenhouse gas (60x more powerful than CO2,
mole per mole), atmospheric CO2 levels would adjust downward to compensate
for a rise in methane. This is unstable, however, because any interruption
of the methane leak would result in rapid loss of methane through oxidation,
plunging the Earth into a global glaciation. The evidence such a scenario
comes from carbon isotopic records immediately prior to glaciation. This
methane ‘trigger’ for snowball events would be deactivated by a rise of
oxygen (a slow leak could not cause methane to build up because it would be
destroyed too rapidly). The advent of macro-animals is commonly attributed
to a rise in oxygen (which is needed both for their construction and
operation), which might explain why snowball events ceased after animals

Martin Kennedy from the University of California, Riverside, will take a
different stand in his address, “The Snowball Earth: Myth or Methane?”
Kennedy will present physical, geochemical, and theoretical evidence for an
alternative model in which massive destabilization of ice-like methane gas
hydrates coincided with postglacial warming. Gas hydrates are an enormous
and highly unstable reservoir of potential greenhouse gasses (carbon dioxide
and methane) and are increasingly believed by paleoclimatologists to play a
critical role in Earth’s climate system. Their deposition is favored by cold
climatic conditions and would have been at a maximum during the extremely
severe Neoproterozoic ice ages. In contrast to the snowball Earth hypothesis,
the model is based on a “conventional” modern climate system, but scaled to
the colder conditions implied by evidence for low latitude glaciation. The
deposition of cap carbonates and with unusual isotopic values is an expected
outcome of a large-scale methane release.

Another argument concerns a ‘Slushball’ Earth-an idea that predated the
Snowball Earth. One of the questions it now raises is: Did the tropical
seas remain open during Snowball Earth?

Simple climate models imply that ice-albedo feedback makes a partially
ice-covered planet unstable if ice cover exceeds ~50% surface area. The
instability results in total ice cover (i.e. so-called ‘hard’ snowball
Earth). Recently, more complex climate models suggest that a partially ice
covered planet may be stable (although simulation times do not exceed a few
decades) with up to 40% open water in the tropics. This so-called ‘soft’
snowball, or ‘slushball’ Earth appeals to many biologists, who are concerned
for the survivability of eukaryotic algae and early micro-animals if any
existed. Bruce Runnegar from UCLA will make this argument at the session.
In fact, Joe Kirschvink’s original snowball Earth hypothesis allowed for
patches of open water in the tropics, shifting back and forth across the
equator with the seasons.

The counter-arguments are the following: (i) the long-term stability of
‘slushball’ Earth solutions has not been demonstrated; (ii) ice lines will
rapidly recede as CO2 builds up due to complete continental ice cover
(eliminating crustal rock weathering which normally consumes CO2). Thus,
a ‘slushball’ Earth will not be long-lived (as indicated by paleomagnetic
evidence), and would not produce iron-formations or cap carbonates, as
observed in the rock record; (iii) Sea ice in the tropics would be <20 m thick, permitting algae to grow and micro-animals to feed even if no open water were present.

The hydrological cycle on a Snowball Earth is another area of disagreement.

Geological evidence implies that mobile (i.e. thick, wet-based) glaciers
existed at least locally during Neoproterozoic ice ages. If the oceans were
ice covered, where would the moisture come from to allow glaciers to grow?
The presence of thick accumulations of glacial debris, it is argued, negates
a frozen ocean. Jim Walker (University of Michigan) has basically reached
the same conclusion, but not on the basis of geological evidence. His
arguments are based on climate theory and he will examine what the weather
would have been like on Snowball Earth. On theoretical grounds, he has come
to the conclusion that the surface of the ice would have melted during the
summer, providing a moist environment.

The counter argument is that ablation of sea ice in the tropics will provide
sufficient moisture for glaciers to slowly grow along elevated sea coasts
because of the lapse rate (decrease in air temperature, hence moisture-
holding capacity, with elevation). Because a Snowball Earth will be long-
lived (millions of years), even a very slow rate of glacier growth (10 cm
annually) will make a ice 1000 m thick in 10,000 years. Modeling results
were presented at the recent American Geophysical Union meeting in Boston
supporting the buildup of thick glaciers on a Snowball Earth. Moreover, as
atmospheric CO2 builds up from volcanic outgassing, the Snowball Earth will
become warmer and warmer, increasing the sea ice ablation rates and the
glacier accumulation rates.

Another point of Snowball contention concerns the evidence for high orbital

Certain structures like ice or sand wedges are common in permafrost soils
and are normally attributed to large seasonal temperature fluctuations. In
South Australia, these structures occur in an area believed to have been
close to the equator when it was glaciated. Seasonal temperature fluctuations
are normally small near the equator (and diurnal fluctuations are too rapid
to produce the observed structures). The structures have been interpreted
(principally by George Williams at Adelaide University) as indicating that
the Earth had a high orbital obliquity during the pre-Cambrian (i.e. a large
angle between the equatorial and ecliptic planes).

With very high obliquity (over 54 degrees), the tropics would be colder on
a mean annual basis than the polar regions, thus favoring low-latitude
glaciation. Grant Young (University of Western Ontario) will make this
argument in his presentation, “Is the Snowball a “No-ball”?: The Case
against the Snowball Earth Hypothesis.”

“The occurrence of glaciers and highly variable seasonal temperatures in
the tropics is perhaps better explained by an Earth that was radically
different — for example, the Earth’s rotational axis may have been inclined
at a much higher angle than today’s, throwing traditional latitude-related
seasonality into a spin,” Young said. “Another possibility is that the
Earth’s magnetic field was different in the geologic past — instead of the
familiar two magnetic poles, the Earth’s field may have had four or more,
so that the magnetic evidence of tropical glaciations may be spurious.
There is no doubt that the Earth underwent climatic convulsions twice in
its long history but we are far from understanding their causes or the
physical parameters that existed on Earth when they took place.”

The arguments against high obliquity are as follows: (i) it results in hot
summers at all latitudes, which makes it difficult to for glaciers to grow;
(ii) widely observed sedimentary features (e.g. iron-formations, cap
carbonates, large stable isotope shifts) are not explained; (iii) there is
no credible mechanism to lower the obliquity at the end of the pre-Cambrian;
(iv) there may be other means of producing ice/sand wedges without strong
seasonality (e.g. surge-type glaciers). The latter point, at least, will
be brought up by Paul Hoffman for discussion during the Snowball Earth
workshop that immediately follows the session.

This is a sampling of a few of the arguments and different perspectives of
the Snowball Earth session that will take place in Edinburgh. While Paul
Hoffman ‘promises’ to keep silent as he fulfills his role as session chair,
that “hat” comes off when he enters the workshop when the real debates
begin. Weather forecast? Stormy and unpredictable.


During the Earth System Processes meeting, June 25-28, contact the GSA/GSL
Newsroom at the Edinburgh International Conference Centre for assistance
and to arrange for interviews: +44 (0) 131 519 4134

Ted Nield, GSL Science and Communications Officer
Ann Cairns, GSA Director of Communications

The abstract for this presentation is available at:

Post-meeting contact information:

Paul Hoffman

Earth and Planetary Science

Harvard University

20 Oxford St.

Cambridge, MA 02138,

Office Phone: +01 617 496 6380

Ted Nield

Geological Society of London

+44 (0) 20 7434 9944

Ann Cairns

Geological Society of America

+01 303 447 2020 ext. 1156