The Snowball Earth Hypothesis: Where It Came From, Where It’s Going Linda Sohl Center for Climate Systems Research, Columbia University Low-Latitude Glaciation in the Neoproterozoic: The World’s Most Severe Ice Age? Media Blitz! horizon A step back in time timeline rev5 The Neoproterozoic snowball glaciation is of special interest because it occurred shortly before major innovations in biological diversity. Signs of cold climates… On the left - scattered dropstones in laminated siltstone from the Kingston Peak Formation (Sturtian) in Death Valley, indicative of a glacial marinee depositional environment. On the right - a fossil sand wedge (white triangle shape) of the Whyalla Sandstone (Marinoan) in frost-shattered Mesoproterozoic quartzite (darker) in South Australia, indicative of a cold arid environment close to the glacial front (periglacial setting). … followed by rapid warming? Two shots of Marinoan cap carbonate (Nuccaleena Fm) in the Flinders Ranges, South Australia. On the left, one can see a distinctive disrupted layer Martin Kennedy has associated with methane de-gassing. On the right, a characteristically finely laminated horizon that is the more common appearance of the cap carbonate all over the globe. Distribution of Neoproterozoic glacial deposits Evans 2000 (Evans, 2000) The map shows the modern distributions of glacial deposits that could reasonably assigned to the Neoproterozoic glaciations. And so the debates begin… • Harland (1964) proposes the existence of a “great Infracambrian glaciation” • • The community disagrees - vehemently! Distribution of climate-sensitive sediments briden [Converted] Though not shown on this plot, Harland’s colleagues would have been expected glacial deposits to be restricted to mid- to high latitudes - just as in the most recent ice age. The debates continue… • Numerous debates over veracity of the glacial nature of sediments • • Astrophysicists and geophysicists weigh in • • Improvements in paleomagnetic techniques lead to many tests of “low-latitude glaciation” Father of the Snowball Earth JoeK Joe Kirschvink coined the term “snowball Earth” in 1992. The Original Snowball Earth Hypothesis (Kirschvink, 1992) • Concentration of continental land masses at low to mid-latitudes led to global cooling by impacting planetary albedo • • Widespread pack ice led to ocean stagnation, resulting in the return appearance of banded iron formations for the first time in > 1 billion years These points are from Kirschvink’s original 2-page paper tucked into the massive (1348 pages!) book, The Proterozoic Biosphere, edited by Bill Schopf. The hypothesis occurred to Joe as he was preparing a series of continental reconstructions for the book, based upon the 1980’s wave of paleomag studies suggesting that Neoproterozoic glacial deposits were predominantly at low latitudes. meert 1994 This Meert and Van der Voo paper takes issue with the interpretation of nearly every glacial deposit tested for low paleolatitudes - except for the Elatina Formation of South Australia. Paleomagnetic Data – Trezona Bore Section, Flinders Ranges (Sohl et al., 1999) Paleomagnetic samples taken from this section of the Elatina Formation capture reversals in the polarity of the Earth’s magnetic field - a major indication that the low paleolatitude determined for this rocks is original, and not a later magnetic overprint. Summary of Paleomagnetic Results from the Elatina Formation, Central Flinders Ranges, South Australia A) B) A) For 58 sites: In situ Tilt-Corrected Dm = 213.9˚ Dm = 212.1˚ Im = -20.6˚ Im = -16.9˚ k = 7.1 k = 9.9 a95 = 7.6˚ a95 = 6.2˚ B) For 3 sections: In situ Tilt-Corrected Dm = 223.1˚ Dm = 214.9˚ Im = -17.9˚ Im = -14.7˚ k = 11.5 k = 94.9 a95 = 38.1˚ a95 = 12.7˚ In situ Tilt-Corrected (Sohl et al., 1999) The tilt-corrected plots of the Earth’s average magnetic pole position calculated from multiple sites/sections are better grouped, another indication that the magnetization preserved in the Elatina Formation is original. The grouping of poles near the outer edge of the plots indicates a low paleolatitude for South Australia (I.e., being nearly 90 degrees away from a magnetic pole places South Australia in the tropics). Paleolatitude of Australia During the Marinoan Glaciation Location of glacial deposits (Sohl et al., 1999) Founder of the “new” snowball Earth PFH Paul Hoffman of Harvard University, co-creator of the “hard” snowball Earth hypothesis. The New Snowball Earth Hypothesis (Hoffman et al., 1998) •Primarily intended to account for carbon isotopic data (d13C = 0 to -5‰) in cap carbonates •Suggests that carbon isotopic values reflect mantle values in an ocean isolated from the atmosphere Paul Hoffman took the Elatina paleomagnetic results, Joe Kirschvink’s hypothesis, and ran with it. Hoffman especially liked the extensive sea ice cover feature of Krschvink’s hypothesis, but took it further by initially requiring a solid shield of ice to totally isolate atmospheric and oceanic carbon reservoirs. The Snowball Earth (Hoffman and Schrag, 2000) The four stages of the snowball Earth, already modified from Hoffman et al. 1998 to accommodate criticisms that a completely shut down hydrologic cycle would never permit the kind of ice sheet build-up needed to create the volume of glacial sediments that exist. In this version, the “deep freeze” total ice cover only applies to Stage 2; Hoffman and Schrag posit some unknown mechanism, perhaps climate amelioration produced by initial CO2 build-up, to allow adequate ice sheet development (Stage 3). From the Scientific American article. Snowball Earth Hypothesis: Freezing Phase • Primary productivity in surface ocean ceases • Surface ocean entirely frozen over (runaway ice albedo feedback; suggested by energy balance models) • Atmospheric CO2 increases to ~120,000 ppm owing to virtual shut-down of hydrological cycle and silicate weathering 1.Applies to Stages 2 and 3 in SciAm diagram. 2.Hoffman wanted the primary productivity in the surface ocean to be wiped out - the “Strangelove ocean” - so that he could explain cap carbonate C isotope values as being largely the result of mantle C contribution (around -5 per mil del 13C). 3. Surface ocean has to be frozen to kill nearly everything off and seal off ocean from atmosphere. 4. 120,000 ppm comes from Caldeira and Kasting (1992), the amount of CO2 they needed to get the earth to recover from total glaciation WITH modern solar luminosity. Snowball Earth Hypothesis: Melting Phase •Catastrophic melting of ice driven by greenhouse effect •Renewed silicate weathering draws down atmospheric CO2, and delivers needed alkalinity and base cations to ocean. Precipitates Carbonate. Cap carbonate records transfer of excess atmospheric CO2 to the oceans •Trend of increasing carbon isotopic depletion upwards in the cap carbonates is due to Rayleigh distillation 1.Stage 4 of SciAm diagram. 2.Catastrophic melting bit encouraged by Pierrehumbert’s EBM results suggesting warming to 50 deg C. 3. Next points are Dan Schrag’s concept, to help explain the process by which the cap carbonates are deposited. Problems with the new Snowball Earth • Necessary continental configuration not applicable to both glacial intervals • Estimate of duration of glacial interval based upon incorrect basin subsidence calculations • No evidence for mass extinctions • Glacial sediments cannot be created in absence of hydrologic cycle, and are too voluminous to be created solely at the end stage of glaciation • Iron formations are limited in occurrence 1.Refer back to slide showing glacial deposit distribution. 2. Hoffman et al.’s (1998) original estimate of duration based upon the time need to make space on the Namibian passive margin, through thermal subsidence, for the ~450 m thick post-glacial interval that Hoffman interpreted as entirely cap carbonate. (Martin Kennedy interprets the same section as having just a couple tens of meters of cap, if I recall correctly.) 3. Paleobiologists say that the fossil record, such as it is, does not support the idea of a mass extinction that wiped out nearly all forms of life. In fact, an analysis of microfossil species abundance done by Andy Knoll (Harvard) suggests there might even have been an increase in species abundance during and after the Sturtian glaciation. Kevin Peterson (Dartmouth) does think it’s possible that there was an evolutionary bottleneck associated with the Varanger glaciation, based on his molecular clock studies for the time of divergence of various metazoan body plans. A number of paleobiologists think there are enough problems with the whole concept of molecular clocks that Kevin’s view is not the broader view. 4. Criticism already addressed actually by the diagram for Hoffman and Schrag (2000), but which was vocally dismissed by Hoffman in the beginning. 5. The iron formations that occur in Neoproterozoic successions are nothing like the classic banded iron formations of the Paleoproterozoic in that a) they are largely iron-rich sediments rather than distinctly hematite or magnetite layered, b) they are much more limited in size, and c) they are typically confined to rift-basin settings, which would be relatively easy to “seal off” by ice and make anoxic, given their size and geometry. Is a “hard” Snowball Earth really necessary? One alternative explanation for carbon isotope excursions - the methane clathrate hypothesis - negates the need for a totally frozen surface ocean Methane hydrate flame1 (Mahajan, 2007) Methane hydrate (or methane clathrate) is frozen methane surrounded by a cage of water ice crystals. It is stable only under sufficiently high pressures and/or low temperatures. Dissociating methane from destabilized hydrates can enter the atmosphere, where it amplifies the greenhouse effect. types_of_deposits (Courtesy USGS) Neoproterozoic methane hydrate deposits were most likely similar to modern Arctic deposits, in permafrost. Methane Hydrate Hypothesis (Kennedy et al., 2001) •Methane hydrates may have been more abundant during the Proterozoic ice ages than any other time in Earth history • –Coldest intervals in Earth history –Abundant area available for permafrost development –Rapid flooding of continental basins and shelves – • • Modern Cold Seep Features Recovered secondary hydrates from the Cascadia Margin (Bohrmann et al., 1998) Cold Seep Facies in Cap Carbonates •Brecciation •Cement-lined cavities •Internal sediment fill •Deep water depositional setting 10 cm Isotopic Evidence from the Congo Craton •Values indicate a rapid excursion and long-term recovery • •Cap carbonate deposition occurred over a brief interval (likely <10 k.y.) Time Model Stratigraphic Section A rapid injection of methane into sea water would cause a sharp reduction in the del 13C isotope curve, followed by a gradual recovery to normal ocean isotopic values (around 0 per mil). Isotopic Evidence for Clathrate Destabilization •Predictions: •Rapid release of depleted 13C (~-60‰) produces an instantaneous drop in marine 13C •Return to normal values takes several residence times of C (>100,000y) (Kump, 1991) But wait - there’s more! •New paleomagnetic data from cap carbonate in Australia presents a different time scale for the end of glaciation – •New age dates suggest that there may be only one true Neoproterozoic snowball glaciation • •Climate models present a range of possible environmental conditions, depending on the model and starting assumptions • New information requires reassessment of ALL hypotheses, with some of the latest data promising a major re-think of how snowball glaciations could have begun and ended. Using the GISS GCM to simulate Neoproterozoic climates Forcings investigated include decrease in solar luminosity, continental configuration, atmospheric CO2 levels, and ocean heat transports GCMs can be used to explore climate scenarios when inadequate paleoclimate data exist. Firefly:Users:lsohl:Desktop:fig4-SBall.mov This movie shows the progress of one simulation towards an equilibrium state, in which snow and ice cover is extensive but doesn’t encompass the entire surface ocean. GISS GCM Simulation results •Only most extreme combination of forcings •permits the growth of ice sheets on land; •surface ocean does not freeze over Summary •The “hard” Snowball Earth hypothesis (Hoffman et al., 1998; Hoffman and Schrag, 2000) is incorrect on key points • •A “slushball” Earth likely presents a better portrait of the environment circa 640 million years ago