Running out of Gas: Zircon 18O-Hf-u/pb Evidence for Snowball Earth Preconditioned by Low Degassing
J. Hartmann,G. Li,A. J. West
DOI: https://doi.org/10.7185/geochemlet.1734
2017-01-01
Geochemical Perspectives Letters
Abstract:doi: 10.7185/geochemlet.1734 The general long-term stability of Earth’s climate over geologic time was punctuated by dramatic excursions. Between ca. 2.5 and 0.5 billion years ago (Ga), these events included the globally extensive glaciations known as Snowball Earths, when ice extended to tropical latitudes. Such anomalous periods of time provide unique opportunities for understanding the mechanisms regulating planetary climate and habitability. However, the causes of these events remain enigmatic, in part because there is little information about fluxes in the global carbon cycle in deep time. We propose that the oxygen stable isotope composition in zircons (δOzircon) contains information about past weathering conditions on the continents, imparted during the time between separation of parent material from the mantle (reflected in the Hf model age) and zircon crystallisation (the U/Pb age). A new compilation of coupled 18O-Hf-U/Pb isotopic data shows that the mean δOzircon value varied particularly between 2.5 Ga and 0.5 Ga. The maximum in the δOzircon record, which we interpret as a time of intense weathering, is associated with the Lomagundi Event (~2.22–2.07 Ga), a dramatic carbon isotope excursion thought to reflect enhanced organic carbon burial facilitated by the release of phosphorous during rock weathering. The onset of the Neoproterozoic Snowball Earth events coincides with the minimum in δOzircon, suggesting low silicate weathering rates at the time. This evidence suggests that long-term decreases in the rate of CO2 release to the atmosphere from solid Earth degassing may have preconditioned the global climate system for intense glaciations. Received 13 May 2017 | Accepted 3 August 2017 | Published 15 September 2017 1. Institute for Geology, Center for Earth System Research and Sustainability (CEN), Universität Hamburg, Bundesstrasse 55, D-20146 Hamburg, Germany 2. MOE Key Laboratory of Surficial Geochemistry, Department of Earth Sciences, Nanjing University, 163 Xianlindadao, Nanjing 210023, China 3. Department of Earth Sciences, University of Southern California, 3651 Trousdale Parkway, Los Angeles, CA 90089, USA * Corresponding author (email: geo@hattes.de) Letter Over geologic time, Earth’s climate is regulated by the fluxes of carbon to and from the coupled ocean-atmosphere system, specifically by the balance between solid Earth degassing, the consumption and release of CO2 via rock weathering, and the burial and oxidation of organic carbon (Berner, 2004). For most of Earth’s history, the sources and sinks of atmospheric CO2 have remained in close balance, maintaining the planet’s equable climate (Walker et al., 1983; Berner and Caldeira, 1997). The large carbon cycle anomalies observed in the sedimentary record of the Archean and Proterozoic eras represent puzzling deviations from this general stability (Hoffman et al., 1998; Goddéris et al., 2003; Bekker and Holland, 2012; Lyons et al., 2014). Understanding the causes of these events requires information about the carbon cycle at the time, which remains scarce. Geochemical evidence from marine sediments, such as records of 87Sr/86Sr or δ13C, is subject to multiple interpretations (e.g., Kump, 1989; Goddéris et al., 2017). For example, the evolution of seawater 87Sr/86Sr over geological time ( McArthur et al., 2012) may be influenced by changes in degassing and hydrothermal activity, continental weatherability, and the isotopic composition of rocks undergoing weathering (Kump, 1989). Without independent constraints, marine isotopic records cannot distinguish between these possibilities, each with different implications for the carbon cycle. Zircons may provide information about carbon cycle fluxes that is independent of the multiple factors affecting seawater composition. The Hf isotopic composition of a zircon reflects the time when the grain’s unmixed parent material separated from the depleted mantle reservoir (Griffin et al., 2002; Hawkesworth and Kemp, 2006). In contrast, the U/Pb age of a zircon records the last time the mineral experienced temperatures above the closure threshold for the Pb system (>1000 °C), and thus the crystallisation age (Mezger and Krogstad, 1997). The difference between the Hf and U/Pb ages, here termed ∆THf-U/Pb, represents the time interval during which zircon parent material could have been affected by weathering and magmatic processes (Fig. 1), leaving an imprint on δOzircon (Valley et al., 2005). The contribution from even small amounts of altered crustal material can increase δOzircon values above the primary mantle signature of ~5.3 ± 0.3 ‰, because altered crust is isotopically enriched as a result of weathering by meteoric fluids (Savin and Epstein, 1970; Gregory and Taylor, 1981; Bindeman et al., 2016). The extent of isotopic enrichment for a given zircon grain will depend on the amount Figure 1 Schematic diagram of the age evolution of a zircon grain, showing how combined Hf and U/Pb isotope information can be used to determine the time window (∆THf-U/Pb) during which the δOzircon value is acquired, based on the conceptual model proposed here. The Hf model age constrains the time that parent material separated from the mantle, beginning its journey of reworking via weathering, erosion, and incorporation into melts. This journey ends when the zircon crystallises, as represented by its U/Pb age. Longer ∆THf-U/Pb windows may be affected by mixing and make it difficult to identify the timing when the δOzircon signature was acquired. We screen zircons by ∆THf-U/Pb in order to calculate the time evolution of δOzircon. Geochemical Perspectives Letters Letter 42 of weathering-related contamination of the parent magma, which is related to the alteration state of the crustal material (Valley et al., 2005). More intense chemical weathering and clay formation on the continents will produce more weathered sedimentary material, with heavier δ18O, increasing the potential for elevating δOzircon (Valley et al., 2005; Payne et al., 2015; Bindeman et al., 2016). Thus, a grain’s δOzircon value should preserve some information about the weathering conditions during that grain’s ∆THf-U/Pb time interval. While petrologic processes and magma chamber conditions such as oxygen fugacity may also influence δOzircon by changing fractionation factors (Valley, 2003), we propose that, when averaged over grains from around the world, δOzircon reflects globally averaged weathering conditions during a given ∆THf-U/Pb interval. We have compiled published δ18O-Hf-U/Pb zircon data from predominantly detrital zircons – including in total 4444 grains with coupled information for all three isotope systems (i.e. all three analyses made on the same zircon crystal). These data span 4.4 Ga of Earth’s history. We have calculated ∆THf-U/Pb for each grain and evaluated the δ18O of the zircons that fall within a given ∆THf-U/Pb window. We argue that shorter windows are most likely to provide information about crustal weathering conditions during a given ∆THf-U/Pb window. Long windows may incorporate mixing of material with different Hf-derived mantle extraction ages (Fig. 1) and also reflect the accumulated effects of reworking over long periods of time, potentially subject to long-term tectonic controls on the extent of low temperature contribution to the δ18O signature. Even for short ∆THf-U/Pb, any individual zircon’s δ18O will be influenced by the tectonic and petrologic history of that grain’s parent magma, which is likely specific to its regional origin. Thus, when plotting data from all individual grains, a very large spread is observed in δOzircon values for any given U/Pb crystallisation age (Valley et al., 2005; Payne et al., 2015, 2016). By aggregating many hundreds to thousands of grains from detrital sources, we seek to extract the average conditions that are most likely to represent the state of crustal alteration for defined time slices, rather than the specific history of individual zircons. We have calculated a moving average of δOzircon over time, using a Gaussian filter. Specifically, for each time t, beginning at 4 Ga and proceeding in 50 Ma increments, we begin by calculating a weighting factor for all zircon data points. The weighting factor is determined by each zircon’s age with respect to the normal distribution centred at time t, with standard deviation of 100 Ma, thus giving zircons closer to t a higher weight and effectively giving zero weight to zircons more than a few hundred million years from t. For zircon age, we average the Hf model age and the U-Pb age of each grain, to reflect the midpoint of the assumed processing time during which the grain’s parent material acquired its δ18O composition. We estimate a representative δOzircon value for time t by calculating a mean weighted δOzircon for all data points. The standard deviation of the mean δOzircon for time t is similarly calculated with the Gaussian weighting (±2 standard deviation shown by grey envelopes in the Figures and the Supplementary Videos). We include only zircons that fall within a given ∆THf-U/Pb window, while varying the size of the ∆THf-U/Pb window from 50 Ma to 1500 Ma (see Supplementary Videos, where the number of grains considered and the mean ∆THf-U/Pb for each time window is reported). Figure 2 shows examples, with average zircon δ18O plotted versus time for ∆THf-U/Pb ranging from 50 to 450 Ma. For low ∆THf-U/Pb (<~300 Ma) the number of total data points is low (<1000 grains total), leading to larger uncertainties (Fig. 2). There is thus a tradeoff between focusing on short time windows (over which there are too few data from zircon grains to provide a conclusive picture) versus longer time windows (which average over longer periods of time, reducing the temporal resolution for relating to specific geologic events). For ∆THf-U/Pb of 350–500 Ma, uncertainties are reduced and distinct patterns emerge in the data, with first order minima and maxima