Orbital forcing of ice sheets during snowball Earth

Abstract

The snowball Earth hypothesis-that a runaway ice-albedo feedback can cause global glaciation-seeks to explain low-latitude glacial deposits, as well as geological anomalies including the re-emergence of banded iron formation and "cap" carbonates. One of the most significant challenges to snowball Earth has been sedimentological cyclicity that has been taken to imply more climate dynamics than expected when the ocean is completely covered in ice. However, recent climate models suggest that as atmospheric CO₂ accumulates, the snowball climate system becomes sensitive to orbital forcing. Here we show the presence of nearly all Milankovitch (orbital) cycles preserved in stratified banded iron formation deposited during the Sturtian snowball Earth. These results provide evidence for orbitally forced cyclicity of global ice sheets that resulted in periodic oxidation of ferrous iron. Orbital glacial advance and retreat cycles provide a simple mechanism to reconcile both the sedimentary dynamics and the enigmatic survival of multicellular life during snowball Earth.

Fig. 1. Spatial and temporal distribution of banded iron formation.

a Palaeogeographic reconstruction ca. 700 Myr ago (updated from ref. ) showing the locations of Cryogenian banded iron formation (BIF), and hypothesized cryoconite pans (gray) where cyanobacteria and eukaryotes may have taken refuge during snowball glaciation. b Lithologic map of the two study locations at Oraparinna and Holowilena, South Australia (updated from ref. ). c Simplified lithostratigraphy of the Sturtian glaciation in the Flinders Ranges, South Australia. d Fine-scale rhythmic banding observed in the Holowilena Ironstone at Holowilena.

Fig. 2. Stratigraphy, magnetic susceptibility, and δ⁵⁶Fe isotope geochemistry at Oraparinna.

(Left) Stratigraphic log featuring the key sedimentological facies of BIF. Lower-case letters refer to outcrop photographs (Supplementary Fig. 4). The colors shown, extracted from photographs, are accurate representations of the rock sequence variability. The lower faulted contact is at −12 m (see Supplementary Fig. 5 for additional field data). (Right) Magnetic susceptibility with 2σ uncertainty (black; Supplementary Table 1). δ⁵⁶Fe isotope variability from two sections at Oraparinna: light red is the same section as ours, and dark red is the section of ref. ~6 km along strike (Supplementary Fig. 6); slightly less positive values at our section likely indicate relative proximity to the ice grounding line, consistent with more stratified diamictite than in the section of ref. . Best-fit degree-2 polynomials for each dataset (dashed lines). Note peak values for both trends occur in the middle of the succession. Rock magnetic experiments were performed on samples between 51 and 76 m (Fig. 5 and Supplementary Fig. 10).

Fig. 3. Time series analysis of iron formation.

a Fast-Fourier transform (FFT) of magnetic susceptibility at Holowilena. Red spectra are FFT recomputed after subtracting the strongest signal (~4 m; ~97 kyr). Two distinct signals in the ~100 kyr bandwidth are shown as two overlapping transparent dark blue bands. b Bandpass filters of significant cycles identified with FFT at Holowilena. Raw magnetic susceptibility data are shown in gray. c FFT of magnetic susceptibility at Oraparinna. d Bandpass filters of significant cycles identified with FFT at Oraparinna. Envelopes of bandpass filters exhibiting amplitude modulation characterized using the Hilbert transform (Methods). Raw detrended data are shown as the gray line.

Fig. 4. Hilbert transform of putative precession signal in order to extract putative eccentricity signal.

a Hilbert transform (red) extracted from precession bandpass filter (blue; Fig. 3b). b Fast-Fourier transform (FFT) of the Hilbert transform, indicating power in the eccentricity band (compare to Fig. 3a), as expected.

Fig. 5. Thermal susceptibility experiments.

a Oraparinna. Thermal susceptibility comparison between samples across a short eccentricity cycle (~100 kyr; left) and a long eccentricity cycle (~400 kyr; right). Stratigraphic heights relate to the Oraparinna section (Fig. 2). Unblocking temperatures near ~675 ˚C and ~585 ˚C indicates the presence of hematite and magnetite, respectively. Note hematite is dominant, but magnetite content is variable. Hematite-pure lithologies yield high magnetic susceptibilities and hematite-poor lithologies yield low susceptibilities. b Holowilena. Sample BIF006 from the lower 5 m of the section (Supplementary Fig. 1) exhibits a dominant magnetite peak and subordinate hematite peak, whereas sample HOL-1 from the upper 5 m contains pure hematite. Supplementary Fig. 10 shows the results of other, corroborating rock magnetic experiments. Purple and red vertical bands in the background show typical ranges of (titano-)magnetite and hematite, respectively.

Fig. 6. Petrography of iron formation.

a High susceptibility, hematite-pure hand sample (HOL-1) from Holowilena (Fig. 5b). Note red jasper is laterally discontinuous and interpreted to be chemically precipitated. b Photomicrographs under reflected light show randomly oriented euhedral hematite laths (white) in a quartz, chert, and clay matrix (gray). Note large gray intraclast at the bottom, also containing hematite laths. c SEM images of polished thin section (top) and rock chip (bottom) exhibiting randomly oriented hematite laths in two and three dimensions, respectively. The textures strongly suggest that the hematite is authigenic (i.e., chemically precipitated) and not detrital.