Life on a Young Planet (36 page)

In 1992, Joe Kirschvink, a gifted maverick in Caltech’s geology department, painted an extraordinary picture of late Proterozoic ice ages. According to Joe, glaciation may have begun conventionally, at high latitudes or elevations, but as ice sheets expanded toward the equator, Earth’s climate system approached and then crossed a critical threshold. Ice reflects sunlight back into space (in the language of climatology, it has a high albedo), cooling the planet as it expands. Cold, in turn, facilitates further glacial growth. Thus, there is a positive feedback associated with waxing ice sheets. In Kirschvink’s view, once glaciers expanded to within about 30º of the equator, runaway refrigeration ensued, covering the Earth with ice in as little as a few thousand years. Ice sheets stretched from pole to pole, and sea ice blanketed the oceans, forming, in Kirschvink’s evocative phrase, a Snowball Earth.

At first, few people took Kirschvink’s proposal seriously. For one thing, it required that we set aside accepted models based on Pleistocene climate in favor of a radical alternative, a move rarely favored by innately conservative scientists. Moreover, the Snowball Earth presents a fundamental problem—once the planet gets into this state, it’s hard to get back out.

In 1998, however, stock in the Snowball Earth appreciated dramatically. During a series of late-night conversations in the Harvard Earth Science building, geologist Paul Hoffman and geochemist Dan Schrag conceived of a way to reconcile geological observations of late Proterozoic glaciation with Kirschvink’s vision of a frozen planet. In particular, they interpreted the geology and chemistry of cap carbonates as evidence for Joe’s proposed escape route from the Snowball’s icy grip.

In Hoffman and Schrag’s reworking of the Snowball Earth hypothesis, late Proterozoic ice ages began, much as Kirschvink suggested, with conventional glaciers that crossed a critical latitudinal threshold and then rapidly covered the planet. The global lid of ice nearly shut down primary production, accounting for a putatively extended interval of low C-isotopic values—which Hoffman and colleagues found begins just
below
the tillites at least for one ice age. The icy veneer further prevented oxygen from diffusing into seawater from the atmosphere, resulting in anoxic deep oceans. In the absence of primary production, rates of sulfate reduction also slowed, limiting the production of H
₂
S and, therefore—for the first time since 1.8 billion years ago—allowing iron to build up in oceanic deep waters.

Cold and ice also shut down continental weathering. Thus, glacial ice nearly stopped the two principal processes that remove carbon dioxide from air. But ice couldn’t stop the main engine by which CO
₂
is
added
to the atmosphere—volcanism. As a result, CO
₂
levels gradually rose in the air above the icy wastes. Carbon dioxide is an important greenhouse gas, but Hoffman and Schrag estimate that to vanquish the global ice sheets, atmospheric CO
₂
must have risen to levels as much as 300 to 400 times higher than today’s. It takes time to accumulate such massive stores of carbon dioxide—Hoffman and Schrag suggest several million years, in line with their estimates of glacial duration—but once a critical threshold (again) was passed, deglaciation would have been nearly instantaneous, melting the vast ice sheets (causing a sharp rise in sea level) and catapulting ocean temperature to 40ºC or higher, much hotter than the warmest seas today. Chemical weathering on the hot postglacial Earth would have proceeded at record high pace, feeding large amounts of calcium into the oceans, where cap carbonates precipitated. Weathering also would have drawn CO
₂
from the atmosphere, hastening our planet’s return to normalcy.

What about Snowball biology? If the apocalyptic scenario of Hoffman and Schrag is to be believed, previously productive habitats would have disappeared as ice expanded, eventually restricting most marine organisms to small refuges around emergent hydrothermal vents like present-day Iceland. Then, things would have gotten really bad. Having narrowly escaped an end by ice, the biological world would have been scorched by “fire” as the oceans heated up to temperatures that few eukaryotes can tolerate for long intervals. Despite this, Hoffman and Schrag suggest that the Snowball Earth and its aftermath provided the crucible that forged animal life. This belief rests largely on the stratigraphic appearance of Ediacaran animals above late Proterozoic tillites, but it is bolstered by the suggestion, much debated in genetic circles, that extreme environmental stress can induce mutations that fuel biological innovation.

The Snowball Earth qualifies as a Big Idea, taking in a remarkable sweep of climatic history, geochemistry, and biology. Not surprisingly, then, it has spawned vigorous argument, as Big Ideas usually do. Not all aspects of the debate are germane to this discussion, but as paleontologists
we do need to ask two questions: did some form of Snowball really happen, and if it did, what might we reasonably conclude about its evolutionary consequences?

Two issues illustrate (but do not completely encompass) the nature of current disputes. First, there are questions about water during late Proterozoic ice ages. In the original Snowball scenario, sea ice as much as half a mile thick severely restricts the transfer of water vapor from ocean to atmosphere, slowing Earth’s hydrological cycle almost to a standstill. Yet, those Australian tillites that formed near the late Proterozoic equator reach a thickness of more than three thousand feet, telling us that ice sheets must have continued to grow for a long time after glaciation reached the tropics. Given that low-latitude sea ice both begins and ends rapidly in the Snowball scenario, it isn’t easy to reconcile these tillites with hydrological shutdown. Now, we can relent on sea ice thickness and envision oceans with no more than a thin (maybe a few feet) veneer of ice that would crack and break, allowing the hydrological cycle to continue. But, if we accept this, we must allow carbon dioxide exchange between air and sea, potentially limiting the hypothesized buildup of atmospheric CO
₂
.

A second issue concerns the carbon isotopic record outlined above. C-isotopic values may be low both before and after ice ages, but in the evolving Snowball scenario, they are low for different reasons. Along with graduate student Pippa Halverson and Yale University geochemist Robert Berner, Schrag and Hoffman hypothesize that methane leakage from organic-rich sediments controlled immediately preglacial climates. Biogenic methane is strongly enriched in
¹²
C (recall
chapter 6
), and molecule for molecule, it is far more effective as a greenhouse gas than carbon dioxide. Schrag, Hoffman, and colleagues suggest that, in essence, the Earth became addicted to methane for keeping warm, and when for some reason the supply was shut off, temperatures plunged and ice expanded. (Note that this proposal can work only if oxygen levels remained low on the late Proterozoic Earth; today, methane released slowly from seafloor sediments would react with O
₂
as it ascended through the water column, delivering carbon dioxide to the atmosphere. We’ll return to the question of oxygen below.) In contrast, Schrag and Hoffman ascribe postglacial C-isotopic values to superrapid carbonate precipitation.

Because these hypotheses explain preglacial and postglacial C-isotopic values in different ways, they make no prediction about carbon chemistry
during
the ice ages. Many workers have assumed that C-isotopic values were continuously low through the glacial epoch—hence the hypothesis that primary production fell to extremely low levels—but recent analyses of relatively rare carbonate beds
within
tillite sections call this assumption into question. Three leading Snowball skeptics, Martin Kennedy of the University of California at Riverside, Tony Prave of Aberdeen University in Scotland, and Columbia University’s Nick Christie-Blick (who once gave a lecture at MIT titled “A Neoproterozoic Snowjob,” leaving no ambiguity about his views), have systematically sampled intraglacial carbonate beds from several continents. These rocks turn out to have C-isotopic values that are rather normal for limestones and dolomites of any age. It might be argued that intraglacial carbonates formed by the redeposition of older, preglacial rocks, but the beds include oolites that must have precipitated from ice age seawater. Remembering that the C-isotopic composition of carbonates reflects
proportional
rates of organic- and carbonate-carbon burial, we can conclude either that carbonate deposition plummeted along with primary production during global glaciation or that primary production didn’t drop off so much after all.
²

In fact, most aspects of the late Proterozoic glacial record can be interpreted in more than one way, including the proposed extent of ice cover. Tom Crowley, now at Duke University, has investigated late Proterozoic glaciation using climate models of the type developed to explore twenty-first-century global warming. In his models, ice sheets spread rapidly once they reach a paleolatitude of 30–40º, and sea ice expands
to cover much of the world’s oceans—score one for the Snowball Earth. But unlike the full-tilt Snowball scenario, some iterations of Crowley’s model leave extensive areas of equatorial ocean ice-free. Another difference: Crowley’s ice begins to retreat when atmospheric CO
₂
reaches a modest four to five times preglacial levels.

Who is right? Nobody knows, in part because we still lack observations or measurements that might help us to choose among competing plausibilities. I like many features of the Snowball hypothesis. Nonetheless, I confess a preference for milder, “slushball” variants of late Proterozoic climatic history, versions that begin and end less catastrophically and leave a bit of open water in the oceans, because I find them easier to reconcile with what I see in the field and because they don’t require ad hoc assumptions to explain the survival of diverse eukaryotic organisms. (Bacteria and Archaea are less informative because, as outlined in
chapter 7
, prokaryotic taxa are hard to eliminate under almost any circumstances.)

The final chapters in Snowball research have yet to be written. But we know enough to be clear on what is probably the key point in all of this: even in the mildest permissible scenarios, ice covers much of our planet’s oceans and most of its continental shelves. No resolution of the current debate will restore the comfortable idea that late Proterozoic ice ages were like the Pleistocene, only older. Late Proterozoic glaciation was extraordinary, and it must have left its mark on contemporary biology.

In
chapter 9
, we discussed the early fossil record of eukaryotic organisms. Red algae appeared well before the onset of late Proterozoic ice ages, and so must have survived the vicissitudes of both glaciation and deglaciation. Green algae did, as well, along with relatives of brown algae, dinoflagellates, ciliates, and testate amoebas. If molecular clocks have any merit at all, even microsocopic animals must have weathered some or all of these climatic storms.

The list of survivors expands further when we add inferences from the Tree of Life. For example, fossils of testate amoebae in 750-million-year-old rocks require, at a minimum, that the common ancestor of fungi and animals was also present, because the evolution of identifiable characters in testate amoebae must logically have come after the phylogenetic split between this group and the animals + fungi. In fact,
most
major groups of present-day eukaryotes must have been present before the late Proterozoic ice ages began—a lot of lineages survived climatic upheaval.

This perspective on ice age biology indicates that refuges must have been numerous, widespread, and persistent during the worst of the glaciers—explaining my preference for relatively mild paleoclimatic scenarios. On the other hand, tabulating ice age survivors doesn’t fully address questions about ice age extinctions. After all, while the phylum Brachiopoda survived the great Permo-Triassic mass extinction 251 million years ago, more than 90 percent of all brachiopod
species
disappeared. Many microscopic eukaryotes leave no identifiable fossils, so our ability to evaluate the magnitude of ice-related extinctions is limited. Nonetheless, by tracking the records of eukaryotes that do fossilize, Gonzalo Vidal and I observed years ago that many protists failed to survive late Proterozoic glaciation—climatic shifts
did
prune the eukaryotic tree. That said, the most conspicuous plankton extinction in the Proterozoic record occurred just after the Doushantuo Formation was deposited (chapter 9), perhaps associated with one last cold snap, but
not
in conjunction with global glaciation. Snowball ice was not the only influence on late Proterozoic life.

If patterns of extinction and survival allow a range of ice age scenarios, what should we make of attempts to link late Proterozoic glaciers with biological innovation?

Research on stress-induced mutations remains in intriguing infancy, but as yet we have little evidence that stress facilitates mutations that lie beyond the realm of more mundane genetic processes or that these mutations adapt animals for conditions beyond those that induced the stress. Of course, in “slushball” scenarios for late Proterozoic climate change, surviving populations need not have been subjected to extremely harsh glacial or postglacial conditions. Moreover, the numbers of survivors needn’t have been small. When organisms are tiny, many individuals can fit into a small area—a square meter of beach sand, for example, can harbor millions of nematodes.

The principal influence of global glaciation was probably ecological, regardless of which ice age scenario we choose to favor. Recall the argument, introduced last chapter, that permissive ecology foments biological
revolution. The growth of globe-swaddling ice sheets would have removed biology from much of the Earth’s surface, and when, in due course, the ice receded, huge areas of real estate would have become available for recolonization. Successful colonists likely experienced little competition, permitting novel and, perhaps, poorly functioning variants to survive. Genetic variation is necessary for evolutionary radiation, but is not sufficient. Populations need Lebensraum where nascent novelties can survive and reproduce. That is just what the decay of late Proterozoic ice ages provided.