Downloaded by guest on September 24, 2021 lzbt .Bell A. Elizabeth Keller Brenhin C. Great Unconformity the of origin glacial Neoproterozoic www.pnas.org/cgi/doi/10.1073/pnas.1804350116 widespread con- this of of record context the the and in unconformity. record, coverage records, sediment impact isotope tinental O bolide and terrestrial Hf zircon the the and including extensive (refs. evidence 5 and parate remains profound and unconformity a 4 pre-Cambrian of observation discontinuous) the Pro- (if bound- basins. yet conformable complete ocean sections; and more ary deep strata a fossiliferous the fossil including revealed to terozoic, any has fauna restricted investigation Cambrian continental have Subsequent the of would Cam- of interval which the precursors “Lipalian” erosion, of and a boundary exposure con- proposed lower of Walcott the dearth brian, a spanning Observing sections subse- (2). formable Walcott basement—was by Precambrian recognized of quently directly atop ubiquity deposited unconformably The sediments (6). and clastic at Sandstone pattern—undeformed dramatically Tapeats this most it Cambrian Vishnu (1), Paleoproterozoic where and al. the Schist States, the between et the nonconformity as Powell United sharp buried, such by the southwestern recognized relief deeply the first of often was of of areas While in lacuna Canyon exposed expo- (1–5). a Grand is Unconformity, correlated mass Unconformity Great temporally and Great the and time as abrupt both known the spatially surface takes a often sure transition of this gradually however, form sediments observed, pres- and accumulating Where heat sure. increasing since under metamorphose abrupt, and recrystallize be not need E explosion Cambrian Unconformity Great continental thermal global subsequent for erosion the this freeboard. with of along consequences average isostatic kilometers, global and vertical a totaling 3–5 heteroge- erosion of spatially glacial by Phanerozoic Neoproterozoic explained of neous be pattern together first-order can the sedimentation and record, impact terrestrial the the bolide excursions, in irregularities These preservational sediment scale. Unconformity, large Great and unprecedented erosion of of crustal event set Neoproterozoic subduction late a a zircon with suggest magmatic in associated that excursions isotope is show hafnium and we Unconformity oxygen Here global Great elusive. remained the has extent that Unconformity global and Great formation the fos- the of for animal explanation complex conclusive earliest or a the in directly sils, of associated occurrence observations long system, the While field with century. Cambrian indirectly a enigmatic the over most for of science the Earth base among the remained 2018) stratigraphic 21, below has Earth’s March review evident in for gap (received often 2018 profound 29, record a November approved Unconformity, and Great Canada, BC, The Victoria, Victoria, of University Hoffman, F. Paul by Edited 53706 WI Madison, Wisconsin–Madison, of University i Kingdom; a at n ca cecs nvriyo itra itra CVW22 Canada; 2Y2, V8W BC Victoria, Victoria, of Australia; University Sciences, Ocean and Earth eateto at,Paeay n pc cecs nvriyo aiona o nee,LsAgls A905 and 90095; CA Angeles, Los Angeles, Los California, of University Sciences, Space and Planetary, Earth, of Department eklyGohoooyCne,Bree,C 94709; CA Berkeley, Center, Geochronology Berkeley rhssdmnaycvrncsaiyrssa et upon contact depth This at basement. crystalline rests metamorphic necessarily or igneous cover sedimentary arth’s e g otws eerhIsiue ole,C 80302; CO Boulder, Institute, Research Southwest eateto h epyia cecs h nvriyo hcg,Ciao L60637; IL Chicago, Chicago, of University The Sciences, Geophysical the of Department .Hr eatmtt nt dis- unite to attempt we Here S1). Dataset | i a,b,1 ihlsL Swanson-Hysell L. Nicholas , nwalEarth snowball o .Husson M. Jon , | lca erosion glacial c osN Mitchell N. Ross , b | eateto at n lntr cec,Uiest fClfri,Bree,C 94720; CA Berkeley, California, of University Science, Planetary and Earth of Department zircon b n hnnE Peters E. Shanan and , f ca n at cec,Uiest fSuhmtn otapo O71J United 1BJ, SO17 Southampton Southampton, of University Science, Earth and Ocean | d ila .Bottke F. William , com/brenhinkeller/GreatUnconformity ac ertebs fteCmra Fg 1A abun- (Fig. Cambrian sediment the continental S1–S3 of Figs. preserved base in the shift near dance global large (within presence a spa- the of corroborate which and records these (7–9) temporal Together constraints. higher-resolution database tial provides Macrostrat refined America) the subsequently North been by has Laurentia pattern thicknesses. in temporal stratigraphic resulting and areas The from Gy basin 1.6 sedimentary past the mapped over flux volume sediment preserved mating 1073/pnas.1804350116/-/DCSupplemental. y at online information supporting contains article Github, This in deposited been has 1 article this for Code deposition: Data under distributed BY).y is (CC article access same open the This Submission.y in Direct PNAS professor a adjunct is article an This is P.F.H., editor, J.M.H.y The as and statement: department context; interest geological of provided Conflict N.L.S.-H. data; paper. isotope y E.A.B. the O and wrote P.B. S.E.P.C.B.K. and record; and Hf impact N.L.S.-H., the zircon contributed E.A.B., T.M.G. contributed P.B., and W.F.B., T.M.G., W.F.B., R.N.M., per- data; R.N.M., J.M.H. analyzed J.M.H., and C.B.K. C.B.K., research; research; designed S.E.P. formed and J.M.H., C.B.K., contributions: Author difficulties presents transition observed con- the However, of either processes. nature or to abrupt strata) the Phanerozoic) Proterozoic principle the of in (erosion in attributed destructive accumulation sediment be (faster might structive 1A) (Fig. to zoic Proterozoic the in rock sedimentary and 4 (ref. al. et Ronov by quantified boundary sed- first Proterozoic–Phanerozoic preserved were the in across variation abundance secular of iment magnitude and extent The Unconformity Global Discontinuous A d owo orsodnesol eadesd mi:[email protected] Email: addressed. be should correspondence whom To eateto ple elg,Cri nvriy et,W 6845, WA Perth, University, Curtin Geology, Applied of Department h Cmra xlso”floigteunconformity. the following in explosion” diversified “Cambrian and the evolved in animals environment multicellular first means con- geochemical the and provide which a sedimentary results the as These on level. glaciation straints base of erosional lowering importance for ero- the glacial the underlining proposed modeling the sion, by by reproduced produced accurately space accommodation be con- shallow can show in seas sedimentation We tinental Phanerozoic of glaciations. extent the Earth” that “snowball dur- Neoproterozoic erosion rapid ing record provide may We ero- unconformity Unconformity. this an Great that sharply the at evidence as often basement known metamorphic years surface or sional half-billion igneous sedimentary past older of the sequence overlies in the that deposited observed rocks been long has It Significance h bevdices rmruhy02km 0.2 roughly from increase observed The j ). e hmsM Gernon M. Thomas , h hcg etrfrCsohmsr,Ciao L60637; IL Chicago, Cosmochemistry, for Center Chicago .y raieCmosAtiuinLcne4.0 License Attribution Commons Creative j eateto Geoscience, of Department f arc Boehnke Patrick , www.pnas.org/lookup/suppl/doi:10. km ∼1 NSLts Articles Latest PNAS 3 yi h Phanero- the in /y aae S2 Dataset and 3 yo preserved of /y IAppendix, SI https://github. c g,h colof School | , ,esti- ), f10 of 1

EARTH, ATMOSPHERIC, AND PLANETARY SCIENCES A proterozoic Eolus granite (Fig. 1), a pluton with an emplacement depth of approximately 10–15 km (3–4.5 kbar) (12), requiring the Ronov,1980 (global)

/yr) 3 erosion of over one-third of the nominal thickness of the conti- 3 Macrostrat (N. Am.) scaled: nental crust over some subset of the ∼0.9 Gy of geologic history fluxN.Am. arealand / areaN.Am. missing from this section.

Proterozoic Posing an additional conundrum in either scenario, the Phanerozoic Phanerozoic–Proterozoic boundary is rather unexceptional from 2 a mantle perspective, with no major variation in mantle poten- tial temperature or tectonic style evident in the continental record (13–16). Consequently, it is difficult to conceive of a model where tectonic sediment supply and basin formation increase profoundly as a result of Neoproterozoic solid-Earth processes alone or one in which dramatically increased tectonic 1 exhumation drives unprecedented erosion. Moreover, while the Rodinian supercontinent cycle features a number of notewor- thy irregularities—including extroverted supercontinent assem- bly (17) and an unusual ore deposit profile (18, 19)—it is unclear how such irregularities could contribute to the formation of the

Preserved sediment volume (km 0 Great Unconformity and associated global preserved sediment 2500 2000 1500 1000 500 0 abundance variations in the absence of significant excursions in Age (Ma) mantle potential temperature. B In either a constructive or a destructive endmember scenario, if global sediment supply from tectonic uplift is held constant near Phanerozoic levels, then the depressed Proterozoic sedi- ment volume in Fig. 1A suggests that on the order of 109 km3 of sediment are absent from the continental crust and deposited instead in the deep ocean basins—either gradually, throughout Cambrian marine sandstone the Proterozoic due to a diminished sediment storage capacity of the continents in a constructive model, or rapidly during an Great Unconformity interval of enhanced erosion near the Proterozoic–Phanerozoic boundary in a destructive model. Indeed, before the plate tec- tonic revolution, the missing sediments from Walcott’s “Lipalian Mesoproterozoic granite interval” were generally expected to reside in the ocean basins (2, 20); their absence, along with the young age of the ocean crust, was considered a significant point of evidence in favor of seafloor spreading and plate tectonics (20). In a plate tectonic model, 50 m much sediment accumulated on the oceanic crust is consumed by subduction—presently at a rate of about 1.65 km3/y (21). Due Fig. 1. The Great Unconformity. (A) Global preserved sedimentary rock vol- to its low density and fusibility, however, subducted sediment in ume increases by more than a factor of 5 across the Phanerozoic–Proterozoic the mantle wedge is often incorporated into new arc magmas (21, boundary in both the estimate of Ronov et al. (4) and a global scaling of 22); consequently, a chemical or isotopic signature of subducted North American units from the Macrostrat database by the area ratio of sediment (if sufficiently voluminous) may be preserved within the global land area to North American land area (a factor of 6.1) according to igneous record. Husson and Peters (8), excluding recent alluvium. (B) The Cambrian Ignacio quartzite overlies the Mesoproterozoic (∼1.35 Ga) Eolus granite at a sharp Zircon Hf and O Isotope Systematics peneplanar nonconformity in the Needle Mountains, CO. One isotopic system amenable to the detection of such a sed- iment subduction signature is the radiogenic hafnium isotope for either endmember model. The estimated volume of pre- system in zircon. In this system, 176Hf is produced by the decay served continental sediment (both in North America and glob- of 176Lu with a 36-Gy half-life. Since lutetium is more compat- ally) does not follow an exponential abundance curve, as would ible in Earth’s mantle than hafnium, the mantle evolves over result from a standard survivorship model (10). Instead, the time toward more radiogenic Hf isotope compositions (e.g., Proterozoic and Phanerozoic preserved sediment abundance higher 176Hf/177Hf) than the crust; this evolution is reported in records are individually roughly constant with age—suggesting terms of εHf or parts per 10,000 relative to the isotopic com- little influence from erosion on epicratonic marine sediment position of average chondrite (CHUR) (24) at any given time. survival at most times in Earth’s history (7–9). Were the step Notably, the common accessory mineral zircon crystallizes with function in preserved sediment abundance observed in Fig. 1A low Lu/Hf and is readily datable by U-Pb geochronology, permit- purely a result of concentrated erosion at or near the base of the ting the accurate calculation of initial Hf isotopic composition at Cambrian, this would involve the erosion of some 80% of the the time of zircon formation. Due to extremely slow diffusion original Proterozoic sedimentary cover (SI Appendix, Fig. S4), in the dense zircon crystal lattice, zircons typically retain their totaling as much as 14 vertical kilometers (11). closed-system isotopic and elemental composition after crystal- Alternatively, a purely constructive interpretation would lization, if not extensively metamict (25). Moreover, zircon is require a roughly fivefold increase in sediment supply and/or produced most voluminously by felsic magmatism (26) partic- continental accommodation space, sustained throughout the ularly in continental arcs (27). Consequently, the erosion of Phanerozoic. However, the observed Great Unconformity is pro- a sufficiently large mass of felsic crust may be expected to foundly erosional in nature, characteristically juxtaposing fluvial increase both the proportion of sediment-filled trenches and sediment with crystalline basement that was formed at great the global rate of sediment subduction, producing a negative depth in the crust. For instance, as shown in Fig. 1B, the Hf isotopic excursion in average global zircon εHfi , considering Cambrian Ignacio quartzite is deposited directly upon the Meso- the strong correlation between trench sediment thickness and

2 of 10 | www.pnas.org/cgi/doi/10.1073/pnas.1804350116 Keller et al. Downloaded by guest on September 24, 2021 Downloaded by guest on September 24, 2021 aeil and (Materials rate subduction analo- Methods sediment crudely of derivative considered O the be and to may Hf gous product standardized This the records. of slope isotope average the calculated sediment the with of of covariance product indicator by the specific by suggested provided more as is a subduction magmas, (21), new balance mass into crustal sediment recycling of anism Eurasia). under continent Indian the of subduction the underplating magma- to by and only anatexis Cenozoic transported to conducive is of temperatures material and proportion depths sedimentary here, small even (32)—but and very tism, a instance global represent of for proportion these leucogranites, small magmas a (Himalayan nonarc are However, magmatism a melting magmas. sediment such new by sed- principle, into produced which In by recycled process records. is geologic iment isotope the Edi- of Hf independent and is and correlation Cryogenian O the zircon between acaran correlation the confirms evident estimate a visually covariance such just moving-window reveals 2 a Fig. correlation; excursion. Hf observed the iso- with oxygen may coincide positive alteration aqueous a low-temperature crust), to reflecting eroded anomaly subducting subjected tope of hydrosphere surface, case the the Earth’s in through (as near processed or has or preex- alteration, crust at the hydrothermal recycled exposed of however, the been that If, near instead reflect crust. pooling continental largely the igneous basalt should isting of melt by composition partial isotope crust by, resulting oxygen deep occur the con- to hot boundary, we were crust–mantle of context recycling this remelting If in signif- scenarios. e.g., setting; a constraints endmember mantle represent thermal two or flux, to sider crustal magmatic is lower global a remelting the favor such of if fraction but icant occur, which may through Hf mechanisms this potential zircon many are There preserved crust. felsic the by in representing excursion Paleozoic, dramatic earliest record. the the isotope most in into the begins which persists far magni- anomaly in and negative eclipsed single Cryogenian are a variations by other spectral tude all significant of However, periods with S10). cycle Fig. timescales, Wilson tectonic at power plate zircon on mean fluctua- global history moderate isotopic this Indeed, in (29–31). the tions cycle to supercontinent in the close Variations to 2A), Lu/Hf. chondritic (Fig. with zircon history reservoir a geological of composition of all out 2A). (Fig. between Ga bins age 4.35 90-My and for Ga mean the 0 of mean uncertainty of SE record 2 continuous and a con is density result sample The resam- 28). temporal 16, via to (13, bias proportional sampling inversely mitigating weights compo- while pling and sample, each this age Hf of in record, sition and uncertainty geologic and propagates the sampling age accurately in While approach U-Pb inescapable (13). are Schoene zircon biases and resam- preservation 29,523 Keller bootstrap weighted of of the method using pling dataset analyses isotopic a O and/or study we Gy, elre al. et Keller Phanerozoic largely a magmas by new followed into Neoproterozoic, recovery. crust late continental the of in across recycling the Unconformity unprecedented in Great an pair- increase the with distinct of boundary, a time Proterozoic–Phanerozoic reveal the the 2D) near (Fig. anomaly results wise The crust. surficially (positive of felsic recycling altered increasing product) (negative consistently decreasing or indicate product) systems isotope both S8 Fig. zircon arc osdrn eietsbuto ob h oiatmech- dominant the be to subduction sediment Considering old, of recycling the requires anomaly isotope Hf this Alone, zircon initial Average average crustal quantify To fa h lblsaehv entaiinlyattributed traditionally been have scale global the at εHf ). and IAppendix, (SI zircons recent more in observed εHf IAppendix SI frmisbodyna eothrough- zero near broadly remains εHf i.S9 Fig. , feouinoe h past the over evolution εHf focrtruhu Earth’s throughout occur εHf ,hglgtn nevl where intervals highlighting ), IAppendix, (SI My ∼500–700 εHf εHf-δ i nzir- in ∼4.4 18 O yGusa enl rmMcota 79,aogwt h corresponding the (23). with Ronov along of (7–9), record Phanerozoic Macrostrat global from by kernel) represented uncertainty Gaussian (age My sediment marine by covered area tal (E reworking. covariance crustal high and decreasing indicate covariance and values high negative Large indicate reworking. values crustal positive increasing Large slope. new standardized average in of dardized recycling product crustal iso- The decreasing hafnium (D) or and magmas. increasing oxygen either indicate zircon both average topes where times indicates covariance (C ( history. Earth’s throughout of volume required the crust Unconformity, of Great subduction and the erosion by with elided consistent potentially is i.2. Fig. h oainebtensadrie zircon standardized between covariance The ) hl h iigo h bevdngtv fiooeanomaly isotope Hf negative observed the of timing the While E D C B A icniooevraiiyadcnietlsdmn coverage sediment continental and variability isotope Zircon vrg zircon Average A) rcino ot mrcncontinen- American North of Fraction ) f- εHf NSLts Articles Latest PNAS vrg zircon Average (B) εHf. δ 18 oainewt stan- with covariance O fand εHf δ 18 .Positive O. | f10 of 3 σ δ 10 = 18 O.

EARTH, ATMOSPHERIC, AND PLANETARY SCIENCES sediment would be large. Using generally conservative estimates with crystalline basement only some tens to hundreds of million for average crust and mantle εHf and continental magmatic years older. In the Mirbat region of Oman, for instance, Stur- flux, we calculate (Materials and Methods) that the observed Hf tian glacial diamictites and syn-glacial sediments unconformably isotope excursion would suggest the recycling of some 2.4 × overlie a juvenile crystalline basement complex with ages rang- 108 km3 of average crust, corresponding to the erosion of ing from ∼810 Ma to as young as 696.7 ± 0.5 Ma (49–51), raising 1.6 km of crust globally if distributed evenly across the conti- the possibility of exhumation of syn-Sturtian phaneritic igneous nents. Accounting for the low recycling efficiency of subducted rocks to the surface during the glacial episode. In sections with Hf into new arc magmas—which is poorly known but likely less less exceptional preservation, juvenile clasts in Neoproterozoic than 50%, considering the immobility of Hf in slab fluids (33)— diamict may provide additional evidence for direct glacial ero- would suggest even larger volumes of subducted crust, ∼3.2 km sion of young crystalline basement: For instance, Sturtian glacial or greater. deposits of the Rapitan Group contain granitic basement clasts as young as 755 ± 18 Ma (52). Since exploitation of a gravita- Neoproterozoic Glaciation and Erosion tional potential energy gradient facilitates rapid glacial erosion Erosional unconformities are common throughout the geologic (35), glacial erosion of young basement may be concentrated record and often have a plausible tectonic cause. The same could in areas of preexisting topography. Critically, Neoproterozoic be said locally for specific exposures of the Great Unconformity glacial erosion need not be spatially uniform to produce the (6). However, it is unclear how any local tectonic explana- observed sediment subduction signature—nor should we expect tion could produce the observed global variations in preserved uniform glacial erosion considering the negligible erosional sediment abundance (Fig. 1) or crustal recycling (Fig. 2). Neo- potential of cold-based ice, the localized erosion of outlet ice proterozoic glacial erosion (34) provides a simple mechanism streams, and the preservation (often in areas of tectonic subsi- which may reconcile rapid global erosion and sediment subduc- dence) of relatively complete sections lacking appreciable glacial tion with the constraints of the sedimentary record. Glaciers erosion (e.g., ref. 53). are unique among erosive agents in their ability to alter ero- Modern glacial erosion rates are highly variable, estimated sive base level: Glaciation promotes continental denudation both to span some four orders of magnitude from ∼0.01 mm/y to indirectly by lowering global sea level (exposing the continents to ∼100 mm/y (54). For comparison, 4 km of erosion over 64 subaerial erosion) and directly through subglacial erosion. While Ma of Neoproterozoic glaciation would require an average ero- rates are variable, in the presence of a large topographic gra- sion rate of only 0.0625 mm/y—nearly two orders of magnitude dient modern subglacial erosion has proved sufficiently erosive slower than recent direct estimates for the modern Greenland to effectively limit global mountain height, evidently outstrip- ice sheet (55); while some such estimates (if reversible pro- ping tectonic uplift rates on the order of kilometers per million cesses are involved) must be corrected for timescale dependence, years (35). the required rate is nonetheless well within the range of physi- Continental glaciation extended to low paleolatitudes in three cal feasibility for glacial erosion. Moreover, while Sturtian and well-established Neoproterozoic intervals: the Sturtian (717– Marinoan glacial deposits evidence accumulation rates 3–10 660 Ma), Marinoan (641–635 Ma), and Gaskiers (∼580 Ma)— times slower than modern equivalents (45, 56), accommodation the first two envisioned as global “snowball” events (36, 37) and space—not depositional process or sediment supply—is likely the Gaskiers as an extensive, but not pan-glacial, event (38). the rate-limiting variable at applicable (>5 My) timescales (56); While ice sheet thickness on a snowball Earth is imperfectly con- in the absence of such accommodation, sediment will not accu- strained and likely heterogeneous (0–6 km) (39–41), glaciation mulate on the continents, but rather in the ocean basins below on all continents analogous to that currently found in Antarc- erosional base level. Consistent with an accommodation-limited tica (∼2 km average thickness) would lower sea level by ∼787 m model, Neoproterozoic diamictites may reach kilometer-scale before isostatic adjustment. After isostatic and local gravitational thicknesses where directly accommodated by local syndeposi- adjustments, modeled freeboard for ice-covered Neoprotero- tional tectonism (46, 47). In the context of global glaciation, zoic continents is variable but positive, with global averages of accommodation must be considered as a competition between 400–650 m for each glacial episode (39). Moreover, if not oth- subsidence and regional upland erosion: Local thermal or tec- erwise constrained by air or water temperature, ice base level tonic subsidence may be thwarted by isostatic rebound from may extend up to 0.89 km below sea level per kilometer of regional erosion (SI Appendix, Fig. S11). ice sheet thickness. Such a configuration would provide a large Delivery of eroded sediment to the deep ocean basins is a gravitational potential energy gradient to drive erosion, while critical requirement for the production of the observed Great isostatically permitting more than 12 km of vertical erosion of Unconformity (where much of the eroded crust is not found else- typical continental crust by a 2-km ice sheet. where on the continents) and is consistent with predictions for The extent of ice-free ocean available to sustain hydrologi- Neoproterozoic glacial erosion. During pan-glacial conditions, cal cycling during such global glaciation is controversial (41, 44). the locus of deposition should shift to deeper waters as a result However, precipitation rates driven by sublimation alone appear of (i) lowered erosional base level; (ii) direct transport of eroded sufficient for the development of localized wet-based ice streams sediment by erosive outlet glaciers [such as those responsible for with high basal sliding velocities and consequent erosive poten- the Chuos paleovalley (47)], which in the present day are often tial (40); evaporation from cryoconite ponds [a notable sink for associated with overdeepened fjords that extend to the edge of solar radiation in a snowball state (45)] might further enhance the continental shelf; and (iii) settling of fine glacial flour in hydrological cycling. Much of the characteristic field evidence deep ocean basins. In more simplistic terms, when all continen- for Neoproterozoic glaciation is unmistakably erosional, includ- tal area is below ice base level during a snowball glaciation, most ing striated pavements, striated and exotic clasts and dropstones, sediment is transported entirely off the continental shelves and and preserved glacial diamictites (36, 46, 47). Although not into the ocean basins, where it is ultimately subducted—just as always well exposed, direct unconformable contact between Neo- suggested by the observed Hf and O isotope records (Fig. 2). proterozoic glacial sediments and Archean to Neoproterozoic Direct and indirect implications are widespread when consid- crystalline basement may be found on most continents (48). ering a geological event as nonuniformitarian as the proposed While the Great Unconformity surface in Fig. 1B allows kilometer-scale Cryogenian erosion, resulting in numerous some ∼0.9 Gy for exhumation of crystalline basement to the testable predictions. For instance, crust exhumed by large-scale surface, other sections may be found where a basement uncon- erosion cools as thermal diffusion adjusts to the new relative formity directly superposes Neoproterozoic glacial diamictites position of the surficial boundary condition. A range of existing

4 of 10 | www.pnas.org/cgi/doi/10.1073/pnas.1804350116 Keller et al. Downloaded by guest on September 24, 2021 Downloaded by guest on September 24, 2021 elre al. et Keller from diameters crater for bins 100-My to in km tallied 2 area area bedrock unit exposed per global km alongside plotted in Ga, area 2.5 past bedrock the over bins 100-My within in tallied to ( known (42). database ages formation with 3. Fig. only My, 700 past the for area com- bedrock of (PASSC) exposed abundance Center follows the closely Science While Space (42). and pilation within Planetary known ages the with from Fig diameter craters. km 10 impact than largest the the but reduce 3A all significantly of should potential Neopro- craters widespread, preservation kilometer-scale if impact diameter, erosion, their Since terozoic to to erosion. relative subject shallow and features, are exhumation surficial by are craters destruction Impact record: impact Neoproterozoic of record a tectonic preserve recent exhumation. to more unlikely of are areas that on activity thermal focused ther- existing is long-term of data proportion the large mochronologic a of since survey continents, the global of required history systematic are a analyses conduct Further to (57–60). exhumation crustal zoic 33 with consistent permissively appear variable, geographically although inversions, thermochronologic B 8 8 2 A n pcfi etbepeito ocrstetretilbolide terrestrial the concerns prediction testable specific One

Craters per 10 yr per cumulative 10 km Impact craters per 100 Myr ◦ 10 15 20 25 2500 10 20 30 40 hw h eodo nw ersra matcaeslarger craters impact terrestrial known of record the shows /mgohra rdet fptnilyrpdNeoprotero- rapid potentially of gradient) geothermal C/km 0 5 1600 0 0 km. >100 h eodo matcaespeevdi at’ otnna crust continental Earth’s in preserved craters impact of record The Diameter Diameter > > > > > > > 100 km 20 km 10 km 100 km 2 km 20 km 10 km bouecae ons(etai)frsvrlsz ranges size several for axis) (left counts crater Absolute A) 2000 1200 2 paetipc rtrn rate cratering impact Apparent (B) (43). axis) (right /y 1500 Sturtian A Marinoan g e (Ma) 800 Gaskiers 1000 ∼100–300 5M (1-σ My ±75

Sturtian onset craters impact >10-km 400 500 5M,updated My, ±75 ◦ – ma a at km (∼3–9 C rmtePASSC the from ) 0 0 0 0.1 0.2 0.3 0.4 0.5

Presently exposed bedrock area (km2 / yr) spriual eakbecnieigta h oe nldsno includes model the that considering remarkable particularly is coverage. sediment marine observed of the record with macrostratigraphic level well this remarkably 4 B, sea corresponds Fig. from in excursion convert illustrated 250-m as to hypso- fraction ) S15 submergence modern continental Fig. a to Appendix, Using isostatic (SI 4A). 250-m (Fig. profile nearly level metric a sea relative case Neoproterozoic free- in each km excursion in Near-modern 3.4–4.5 producing with S13). erosion, reproduced Fig. is glacial Ma Appendix, 750 at (SI freeboard board balance initial influences mass directly via erosion Neoproterozoic of 1A 0.9 Fig. of from Using rate rate assumed My. constant sedimentation 800 net past Phanerozoic the the subsi- over either thermal accumulation isostasy, sediment erosion, and of dence, free- effects continental the of model combining 1D board, a constructed we erosion, terozoic 8. ref. by continental proposed increased as to storage, leads sediment inevitably nonetheless continen- buoyancy erosion geotherm, thermal continental tal the by of moderated advection accommoda- erosional temporarily given new be this may accommodation While space providing tion accumulation. and sediment level for sea space free- relative continental the raising decreases to directly board, Continen- erosion one. contribution through requires thinning it constructive tal contrary, the a such to However, preclude Unconformity; 2E. Great not Fig. does the with by consistent erosion inherited glaciation, be largest recent may (and abundance) most unconformity sediment sub- preserved dramatic by in capture most step to the subject erosion, are sequent surfaces via erosional sediments older Since of erosion. proportion some removing record, crustal (e.g., erosion continuous S4). from survivor- Fig. result exponential Appendix, would SI the that of (10) be discretization curve may ship a record part coverage in sediment considered observed the (67). exhumation Consequently, cratonic of m/My) background (<2.5 a negligible amid if comparatively erosion few glacial extensive, intense a of into focusing This episodes step- discrete denudation 2. by the continental Fig. imposed kilometer-scale of nonuniform, of consequence potential scale conti- expected preservation an the and wise is at excursions record history isotopic discontinuous coverage zircon sediment observed nental the both consistent remarkably with is glaciation Neoproterozoic of timing The Erosion 19). Crustal ref. Rapid in of 2 Consequences (page al. the et of Liu predic- by deposits reported same [mineral] Eon” apparently Phanerozoic toward The (and bias 66). strong “preservational the (65, with stepwise) so consistent likewise uniquely before appears not tion eclogites although glaucophane 64), and assemblages (63, jadeitites (U)HP/LT noted as fragile the such with thermodynamically before consistent abundant of appears less absence prediction This be a Sturtian. should shallow in the geotherm the metamorphism continental low-grade for assemblage prolonged normal wide mineral affinity any survive a that cannot an predict consider which may with to we instance, features record For crust. impact geological bolide of the range from bias tional progressive with erosion. consistent glacial Neoproterozoic Ma—temporally 600 and before Ma craters 700 before area craters unit diameter per <100-km density crater of 3A function (Fig. a strik- as particularly considered is km trend when onset This 4.2–5.8 ing 62). of (61, the depths respectively massive km, predate to 8–11 of eroded and 3 remnants Vredefort, and eroded Fig. Sudbury deeply of craters: both criteria glaciation, Sturtian the of matching craters two h rtodrsceso hs1 reor oe prediction model freeboard 1D this of success first-order The Neopro- rapid of consequences depositional the quantify To the in filter a as acts epoch glacial each discretization, this In preserva- of analysis our extend may we qualitatively, More and ,wt narp rnainof truncation abrupt an with S12), Fig. Appendix, SI km 3 y ayn h oe magnitude model the varying /y, NSLts Articles Latest PNAS k diameter <10-km 0 Ma ∼700 | f10 of 5 ra or

EARTH, ATMOSPHERIC, AND PLANETARY SCIENCES A clastic input. However, there are Ediacaran basins which are 50 not sediment starved. Isostatic freeboard model iii) Chronological bias may result in an underestimation in the Macrostrat sedimentation volume and extent of Ediacaran sediments if ambiguous units 0 are mistakenly assigned to the Cambrian. The residual cur- Marinoan Sturtian Gaskiers 3.4 km glacial erosion 0.9 km3/yr sedimentation rency of the phrase “Precambrian basement” testifies to the 4.5 km glacial erosion historical association of the first sediments above crystalline -50 basement to the Cambrian system. However, we hope that this known problem (72) has been largely corrected over -100 recent decades. While none of the above hypotheses alone is entirely satisfac- -150 tory, all imply a range of testable predictions that may be better understood with future work.

-200 Inferences and Conclusions The first quantification of continental submergence by Egyed (76) indicated dramatic emergence throughout the Phanerozoic Relative continental freeboard (m) -250 (i.e., declining marine coverage), as in Fig. 4B. While the original B Ronov,1994 (global) interpretation of this record has been obviated by plate tecton- 0.4 Macrostrat (N.Am.) ics (77), the paradigm of monotonic continental emergence as a result of global cooling has persisted (78, 79). We suggest that Model this paradigm must be reevaluated. The correspondence between the modal elevation of the continents and global sea level (SI 0.3 Appendix, Fig. S15) is not coincidental, but rather a direct conse- quence of subaerial erosion on a tectonically active Earth (80); given active orogenesis and felsic continental crust, any buoyant continental mass with negative freeboard must thicken by oroge- 0.2 nesis and sedimentation until it reaches zero or slightly positive average freeboard, if not otherwise limited by delamination or gravitational collapse. The negative continental freeboard which 0.1 enabled extensive continental coverage and subsequent recovery Lost to Cretaceous{ (i.e., emergence) throughout the Phanerozoic (9, 76) may thus erosion or lack interior be an anomaly enabled by glacial erosion below ice-free oceanic of sediment supply Continental coverage fraction seaway base level. 0 While nonconformity between sediment and crystalline base- 800 600 400 200 0 ment is ubiquitous on all continents, it is highly diachronous (6). Age (Ma) This diachroneity of amalgamated unconformities has helped to obscure the global significance of Neoproterozoic glacial erosion. Fig. 4. Isostatic global sea level and continental coverage model. (A) Temporal evolution in average continental freeboard driven by erosion, subsequent thermal subsidence, and sediment accumulation. Neoprotero- zoic glacial erosion is distributed in proportion to the duration of each glacial interval. (B) Corresponding modeled continental coverage fraction Laurentide Ice Sheet assuming a constant hypsometric profile, compared with the observed North thickness at 18 ka American record from Macrostrat (7–9) and Ronov’s (23) global record of Phanerozoic marine sediment coverage.

consideration of local tectonics. However, one feature remains problematic: the time delay between the end of Neoprotero- zoic glaciation and the Cambrian increase in preserved sediment abundance. Potential causes for this misfit may fall into three broad categories:

i) Erosional loss of the Ediacaran record, glacial or otherwise, provides the most direct mechanism. Maintaining low pre- served sediment abundance over the 92 My of the Ediacaran (and particularly the 39 My from the Gaskiers to the base of the Cambrian) through erosional means would be trivial

compared with the kilometer-scale erosion we propose for Precambrian 0 1000 2000 3000 m the Cryogenian. While a late Ediacaran glaciation has been bedrock Licciardi et al. suggested (68, 69), precise geochronological constraints are USGS GMNA (2009) (1998) lacking and key observations [e.g., Cloudina in the matrix of an Ediacaran diamictite (70)] have not been replicated. Fig. 5. Capture of the Great Unconformity by Laurentide glacial erosion, illustrated by the correspondence (73) between Precambrian basement ii) Nondeposition resulting from sediment starvation may be exposure as mapped in the Geologic Map of North America (74) and the expected if glacial peneplanation (71) sufficiently reduces the extent of the Laurentide ice sheet at 18 ka as estimated by Licciardi et al. available topography; reduced sediment supply could persist (75). Note the survival of Phanerozoic cover under the ice divide near on tectonic timescales until orogenesis provides a renewed Hudson’s Bay, where basal sliding velocities are low.

6 of 10 | www.pnas.org/cgi/doi/10.1073/pnas.1804350116 Keller et al. Downloaded by guest on September 24, 2021 Downloaded by guest on September 24, 2021 elre al. et Keller relation the following density sample temporal dataset. to the proportional from absent directly weights known largely sample are are Consequently, ages locations While geographic 16). analysis, (13, each O Schoene for and and unique. Keller Hf of are zircon following approach 29,523 resampling of the weighted-bootstrap which record applied of temporal we composition, ), S6 representative isotopic and maximally S5 a Figs. obtain from To Appendix , analyses 35,368 (SI et further of some continents dataset Dhuime by a all (86), augmented in (89), resulting al. data, al. literature et et of (88)/Payne compilation al. Belousova et detrital of Spencer and and datasets igneous (87), al. preexisting for ages the U-Pb from with zircons along compositions isotopic Analysis. O Carlo and Monte and Systematics Isotope Zircon are thermal data and and isostatic code an source at available Computational of freely freeboard. context continental the and of in results model terrestrial interpreted geochemical as and and well integrated stratigraphic as Finally, are geochemistry, datasets. impact isotope bolide O lunar and alongside (85). Hf interpreted is charts zircon Unconformity (COSUNA) compiled Great the America of North record Geologists stratigraphic of Petroleum This Units of Association Stratigraphic American of the Correlation the of digitization by Peters obtained (7) by produced are originally America (macrostrat.org), database North Macrostrat for the from geochemical, data Stratigraphic stratigraphic, datasets. of geological range and a assemble Proterozoic– we the boundary, near record Phanerozoic rock continental the in anomalies investigate To Methods and Materials direct diver- a glaciation. considered the Neoproterozoic be of may to the consequence (5) led such, animals multicellular that As of changes sification 4B). geochemical (Fig. and continental accumulation environmental of sediment history and subsequent Neoproterozoic freeboard the with governing inherently genesis, erosion of constructive glacial Unconformity sedimentary Great and a Phanerozoic erosive by coupled the late explained best preceding that is the record conclude to We relative accumulation Proterozoic. inunda- sediment continental and continen- Paleozoic tion of enhanced cycle with accommo- Phanerozoic sedimentation have a glaciation-derived tal initiating may of for glaciation responsible global lack been Neoproterozoic a Thus, to space. development dation due on regolith accumulation continents and sediment marine the weathering little comparatively of with (84) period experienced prolonged have may a Earth glaciations, (37) the Sturtian evidence ∼0.72-Ga between geologic deposits of lack glacial Paleo- the of for period Given known (83). a glaciation following proterozoic appears isotope Ga O 2.2 and circa corre- Hf but records zircon smaller the billion-year of in set excursions over a Paleoproterozoic that processes lated note we earth context, represent this surface In may of timescales. record, isotopes signal igneous Earth’s a O in preserving and preserved proxy Hf paleoerosion first zircon and the here, crust proposed Archean mity of survival the with (67). lithosphere km/Gy) 50–500 mm/y 0.05–0.5 (i.e., of rates erosion conti- background, modern Phanerozoic unsustainable the high the reconciling to that comparatively relative suggest (82) display rates we erosion may context, nental this epoch In (73, icehouse Quaternary S16). present the Fig. or Appendix, Age glaciation (SI Ice to 81) Paleozoic subject Late been the either have expo- during basement nonorogenic remarkable all Precambrian virtually of A 5); sures bedrock (Fig. erosion. Precambrian glaciation recent and between exposure noted more been by has cap- correspondence deepened collapsed, multiple and unconformities from tured, results Neoproterozoic) likely Unconformity. (e.g., sites Great ancient such ongoing at the 5)—an exhumation at (Fig. However, exposed today commonly even is surface basement Archean or Proterozoic osdrn h lcgncmdlfrteGetUnconfor- Great the for model glacigenic the Considering https://github.com/brenhinkeller/GreatUnconformity. w i = 1  w i X j =1 o ahsample each for n (t .-aRefnen(3 and (83) Rietfontein ∼2.2-Ga i − t 1 j ) 2 + 1 , i ecmie icnHf zircon compiled We r sindinversely assigned are l a etae steuprlmto h compiled the of limit upper the as traced be may as tle in Meanwhile, reservoir basalt. positive primitive most a e.g., the than, crust preexisting zircon- of more only assimilation exhibit samples may and record initial (26) felsic zircon negative predominantly the are which since magmas, conservative bearing is estimate choose we This crust, subducted of volume estimated (13), Schoene and ε from Keller ranging of Lu/Hf values dataset with Neoproterozoic the accordance in obtaining sample in whole-rock ingrowth each Hf for for ratios accounting crustal time average subsequent the calculate any may we history, Earth’s throughout the crust Taking crust. subducted of zircon volume compiled required minimum the estimate may the into sediments accumulated basins). any of ocean subduction complete thus (and with crust consistent modern entirely is in recovery anomalies of the timescale 10Be the cosmogenic (92)—while from magmas by Ma arc (<7–9 suggested magmas eruption) new con- into to is sediment erosion of glaciation recycling Sturtian fast following the excursion with sistent the of onset by scales) gigayear complete Sturtian the is of that onset ery the with coincident Neoprotero- pair [ excursion larger glaciation much initial Paleoproterozoic a an a and with Ma evident: pair nega- 2,200 zoic are circa a beginning events by excursion such an followed Two with immediately peak. peak recovery excursion tive positive pairwise a characteristic as with appear sub- features thus sediment events decreasing subduction or individual increasing duction; either with the S9D). S9E covariance, co-occurring Fig. negative Fig. covariance of Appendix, Appendix, case (SI (SI the signals in product zero isotope covariance–slope to O tends and slope the average Hf with the two covariance Since this the to of of sedi- back product slope (increasing subduction the average examine excursion sediment additionally (decreasing between we recovery baseline), distinguish and to covariance subduction) positive wish ment large we yield anomaly Since recov- and isotope peaks. excursion Neoproterozoic the the Both concert: of in ery decrease or (SI increase either systems signals both S9C for Fig. positive Appendix, is S9 recycling Fig. the increasing with Appendix, that variance, such unit inverted to records nal normalized isotopic both and episodes), detrended recycling crustal are distinct evolu- to crustal opposed secular (as following long-term tion from recycling, covariance sediment extraneous recy- or of dependence in crustal signature illustrated felsic isotope procedure of O a signature zircon the isotope and Hf cling zircon the between covariance implications explored. broader previously the been and not covariance have global thereof systematic their noted 90), previously Hf 31, been and (30, have O zircon atypical Neoproterozoic While of records. isotope characteristics and O isotope magnitude and Hf in between both covariance the notable in are back- excursions tectonic Neoproterozoic normal the this with expected ground, observed Compared the S10 ). roughly are Fig. and with , Appendix variations record, (SI rift, preserved periodicity normal the arc, of such entirety of the (29–31); preservation throughout vary and the throughout magmatism proportion vary the collisional averages as isotope less O cycle undergone and supercontinent has Hf that Zircon mantle. crust convecting felsic Earth’s recy- old, the at reflect of excursions alteration isotope cling silicate Hf aqueous negative of while low-temperature recycling sediment), (i.e., excur- the undergone surface reflect isotope has (∼5.5/mil) O that baseline Positive crust mantle magmas. the new above into for sions crust mean preexisting the of 2). of Fig. recycling (e.g., SE Ma 2 4,350 and and Ma average 0 an between resampled, windows as 90-Ma is presented overlapping dataset through- are results history the Average Earth’s time new uncertainty. out a each analytical that distribution representing such this fully SD and thereby from mean drawn known a is represented with is value variable ratio) random isotope Gaussian Hf a geo- zircon as single each a resampling, (e.g., Throughout measurement analyses. chemical duplicate e.g., of, elimination IAppendix more (SI sampling a distribution produces with temporal weighting replacement, This even weight. with sample to resampled proportional is probability dataset the Subsequently, where t65M.Sneamr eaiecutledebrwl euti lower in result will endmember crustal negative more a Since Ma. 635 at ie h bevdmgiueo h lblH stp xuso,we excursion, isotope Hf global the of magnitude observed the Given the examined we subduction, sediment of importance the assess To the record both timeseries isotope O and Hf zircon average global The 0-ycaatrsi iecl o opeetroe fteoceanic the of turnover complete for timescale characteristic ∼200-My n stenme fsmlsi h aae and dataset the in samples of number the is 1 a(7 1] ai at nadir a 91)], (37, Ma ∼717 fta vrg rs u oagetrcnrbto from contribution greater a to due crust average than εHf fdtse sa siaeo average of estimate an as datastet εHf A hsrwcvrac spstv hr h fadO and Hf the where positive is covariance raw This . and 4 a oal,teesnilyimdae(on immediate essentially the Notably, Ma. ∼340 .Tersligcvrac silsrtdin illustrated is covariance resulting The B). is,t eoeayscale any remove to First, S9. Fig. Appendix, SI fsae h vlto ftedpee man- depleted the of evolution the space, εHf i.S7 Fig. , 6 a n an and Ma, ∼560 NSLts Articles Latest PNAS −33.7 mhszslrepositive large emphasizes ) n bitstemanual the obviates and ) −35ε 176 ffil hog time, through field εHf figot hnthe than ingrowth Hf ε samnmmvalue. minimum a as fo e igneous new of εHf t77M to Ma 717 at t fiooi sig- isotopic εHf 2-yrecov- ∼220-My ssml age. sample is | f10 of 7 fat εHf −34.9 SI

EARTH, ATMOSPHERIC, AND PLANETARY SCIENCES estimating a value of +14ε for the Neoproterozoic (SI Appendix, Fig. S6). if the Sturtian and Marinoan together were to erode 3 km of crust, fol- As seen in Fig. 2, the Neoproterozoic negative εHf excursion ranges from a lowed by 100 m of sedimentation between 635 Ma and 580 Ma, the Gaskiers baseline of +4 ε to a nadir of −8 ε, with a depth of 12 ε, with an average need only erode 100 m of sediment to capture the entire (now) 3.1-km depth of 5.7 ε over the 400-My duration of the excursion. This average depth unconformity. corresponds to 5.7/(14 − (−35)) = 0.12 of the total range between average To quantify the consequences of Neoproterozoic erosion for continental crust and depleted mantle, equivalent to shifting 12% of total continental freeboard and sediment accumulation, we constructed a 1D thermal and magmatism over the duration of the excursion from a mantle source to a isostatic model of the continental crust and lithosphere. On approximately crustal source. gigayear simulation timescales, isostatic adjustment is assumed to be effec- Phanerozoic estimates of rates of volcanic and plutonic magmatism in tively instantaneous, with postglacial viscous mantle rebound (45, 96) likely the continental crust suggest 3–8 km3/y of arc volcanism and plutonism complete within a single-million-year model timestep. However, the ther- along with 0.2–1.5 km3 of intraplate continental magmatism (93). More mal consequences of kilometer-scale erosion may be more protracted. To recent mass balance constraints suggest at least 3.8 km3/y of arc magma- account for thermal subsidence as the advected geotherm decays back into tism is required to avoid long-term crustal destruction. Consequently, we equilibrium, along with the direct isostatic effects of erosion and sedimen- take 5 km3/y as a relatively conservative estimate of total continental mag- tation, our model assumes a coefficient of thermal expansion of 3 × 10−5/K, matism. In this case, shifting 12% of continental magmatism from a mantle a thermal diffusivity of 1 mm2/s, an average crustal thickness of 33 km, an source to a crustal source over 400 My would require the recycling of some average density of 2,818 kg/m3 for the continental crust (97), a mantle den- 2.4 × 108 km3 of average crust. Such a volume corresponds to 1.61 verti- sity of 3,300 kg/m3, and a slightly buoyant mantle lithosphere (3,250 kg/m3) cal kilometers if distributed evenly over the 1.489 × 108 km2 area of the of 100-km thickness, for a total lithospheric thickness (crust + mantle litho- continents. Considering that only a fraction of subducted Hf makes its way sphere) of 133 km, generally intermediate between expected thermal (98) into new magmas (depletion of high-field-strength elements such as Hf is and elastic (96) lithospheric thicknesses. This model was then perturbed a characteristic signature of arc magmatism due to the immobility of these by various scenarios of erosion and sedimentation, with several kilome- elements in aqueous slab fluids) (33), the true value is likely at least twice ters of Neoproterozoic erosion followed by either continuous (0.9 km3/y) 8 3 that, or ∼ 5 × 10 km if this recycling occurs via sediment subduction. or variable (Macrostrat derived, as in Fig. 1A) sedimentation rate. For the purposes of Fig. 4 and SI Appendix, Fig. S13, the total volume of glacial ero- The Terrestrial Bolide Impact Record. To obtain an independent constraint sion was partitioned between the three Neoproterozoic glacial intervals in on the timing and magnitude of Neoproterozoic erosion, we have examined proportion to their duration. However, instead equally distributing erosion the terrestrial impact crater record as compiled in the PASSC Earth Impact between all three glaciations has little impact on the results (SI Appendix, Database (42), with age constraints updated where applicable. Since bolide Fig. S14). impact craters necessarily occur at Earth’s surface, their resistance to erosion To better understand the implications of this model for continental emer- is a function of crater depth. Hypervelocity impact craters are character- gence and sedimentation, the resulting freeboard curve was translated istically shallow features, with an initial depth around 1/10th of the their into expected continental coverage extent, using a present-day hypsometric original diameter or less (94, 95), decreasing above 15 km diameter such that curve (SI Appendix, Fig. S15). The assumption of present-day hypsome- a Lunar impact crater of 100 km diameter may be only 4 km deep (95). Con- try is notably imperfect, but presently unavoidable given an absence of sequently, all but the largest terrestrial impact craters should be susceptible independent constraints on past global hypsometry. Glaciation may signif- to erasure by Neoproterozoic glacial erosion. If the Neoproterozoic glacial icantly alter continental hypsometry—with the potential to either produce erosion hypothesis is correct, we expect a dramatic decrease in impact crater or destroy topographic contrast under different conditions (71). Conse- preservation potential across the Cryogenian for all but the largest class of quently, the global hypsometric gradient is poorly constrained both before terrestrial impacts. While this prediction is broadly confirmed by the raw and in the immediate aftermath of Neoproterozoic glaciation. The assump- impact record alone (Fig. 3A), the signal of preservation is better resolved tion of near-modern hypsometry is more supportable closer to the present by normalizing the impact record to the continental area that was available day (i.e., the past 500 My), which is perhaps unsurprisingly where model for impact cratering at some time in the past and is now again exposed at misfit is lower. the surface. As illustrated in Fig. 4, the model results are remarkably consistent with We explore two such normalizations, (i) to the cumulative area of crust the observed continental coverage extent curve, with continental cover- exposed today that is older than a given impact age and (ii) to the surface age increasing dramatically in the aftermath of Neoproterozoic erosion area of crust exposed today of the same age as a given impact crater (SI and then slowly declining to background as continued sedimentation fills Appendix, Fig. S12 A and B). The first normalization (by cumulative exposed the available accommodation space. While the general agreement between area, as seen in Fig. 3B and SI Appendix, Fig. S12D) is the most conserva- model and observed coverage trends is quite good given the wide range tive in that the presently exposed area bedrock of age X Ma or greater is of uncertainties involved, two particular intervals of misfit are apparent: the maximum exposed surface area that could preserve an impact of age X (i) a period in the middle Cretaceous where observed coverage substan- Ma. This is a maximum extent because, for instance, 1-Ga bedrock may be tially exceeds model expectations and (ii) systematically lower than expected extant at 0.5 Ga, but deeply buried and thus unable to record an impact at coverage before ∼500 Ma. that time. A wide range of factors may introduce such misfit. First, no specific tec- The latter normalization (by the relative area of exposed crust of the tonic or orogenic events are included in our simple 1D model. In this context, same age as a given impact, within some binning resolution) is more the relatively low misfit after ∼500 Ma is arguably surprising and suggests aggressive but may be considered more natural for sedimentary or vol- that the global rates of relevant local processes such as orogenesis and basin canic bedrock, which must have been exposed at the time of deposition formation may not have varied wildly over the past 500 My. Systematic and thus would have been available as a target for bolide impacts at that variation in mantle heat flow may change oceanic spreading rate (99) and time. This normalization results in an even more dramatic discontinuity in midocean ridge height, thus changing average global sea level. Additional preserved cratering rate across the Cryogenian (SI Appendix, Fig. S12C). The misfit may be introduced by erosional or nondepositional unconformity in true preservation signal is likely intermediate between SI Appendix, Fig. S12 the record subsequent to the initial Great Unconformity; continental emer- C and D, but in either scenario strikingly lower preservation potential is gence will be overestimated if we are missing the sediments by which we suggested for impact craters predating Neoproterozoic glaciation. estimate coverage. Any change in the form of the terrestrial hypsomet- ric profile between 800 Ma and today—likely, but difficult to test—would Continental Freeboard and the Sedimentary Record. The Great Unconformity introduce error into the function mapping between continental freeboard is manifest in the macrostratigraphic record of continental sedimentation in and coverage extent. Finally, the accuracy of the observed coverage record the form of a series of approximately stepwise increases in preserved sedi- is entirely dependent on the accuracy of the underlying geochronological ment abundance between approximately ∼720 Ma and ∼500 Ma (Figs. 2E constraints. and 4). In an erosional context, each step may be considered to reflect a One might at first consider this Cretaceous anomaly as a regional bias decreasing probability of any preexisting sediment having survived past a reflecting the well-known Cretaceous Interior Seaway of North America given glacial episode. For instance, sediments older than the Gaskiers may (100, 101) attributable to, e.g., regional tectonics or dynamic topography. have survived only one Neoproterozoic glaciation, while sediments older However, the Cretaceous has long been known as a time of anomalous than the Sturtian must have survived all three. Moreover, since erosive flooding on multiple continents (102), and indeed a positive coverage glaciation tends to capture the evidence of previous erosion, the largest anomaly is observed even in the coarser-timescale global record of Ronov abundance step (and most dramatic unconformity) may be inherited by the (23) as seen in Fig. 4B. Consequently, this excursion may be more con- most recent glaciation, consistent with the results of Fig. 2E. For instance, sistent with a global increase in midocean ridge elevation and spreading

8 of 10 | www.pnas.org/cgi/doi/10.1073/pnas.1804350116 Keller et al. Downloaded by guest on September 24, 2021 Downloaded by guest on September 24, 2021 1 usnJ,Ptr E(08 aueo h eietr okrcr n its transformation: and phase record assisted rock Deformation sedimentary (2012) the CL of Andronicos Nature RA, (2018) Hunter SE 12. Peters JM, continents. the Husson of Denudation 11. (1970) B Gregor 10. 3 elrC,SheeB(02 ttsia eceityrvasdsuto nsecular in disruption reveals geochemistry Statistical (2012) B Schoene CB, Keller 13. 4 odeK,AtrR,vnHnnJ(06 ra hra iegnei h mantle the in divergence thermal great A (2016) J Hunen van RC, Aster KC, Condie 14. 7 upyJ,NneR 20)D uecniet nrvr retoet:Sm-Nd extrovert?: or introvert supercontinents Do (2003) geochemistry RD basaltic Nance continental JB, and tectonics Murphy Plate time. 17. (2018) through B temperatures Schoene mantle CB, and Keller magmas Primary 16. (2017) X Feng J, Ganne 15. 2 aot ,KlmnP 21)Rl facpoessi h omto fcontinental of formation the in crust–mantle processes arc redistribution, of Role (2015) Crustal PB Kelemen (2009) O, Jagoutz JP 22. Morgan P, floor. Vannucchi sea the of PD, spreading that by Clift evolution evidence 21. basin mineralogical ocean and and Continent Geochemical (1961) RS (2017) Dietz RM 20. Hazen AH, Knoll age. C, middle Earth’s Liu (2014) CJ 19. Hawkesworth PA, Cawood 18. 4 ove ,Vror D acetP 20)TeL–fadS–diooi composition isotopic Sm–Nd and Lu–Hf The (2008) PJ Patchett JD, Vervoort A A, Bouvier continents: the 24. on regressions and transgressions Phanerozoic (1994) AB Ronov 23. 8 elrC,SheeB abn ,SmetnK,Hso M(05 Volcanic–plutonic (2015) JM Husson KM, zircon Samperton M, 750 relative Barboni since B, activity Schoene in arc CB, continental Keller of variation nature 28. Episodic Temporal (2017) JS (2017) Lackey CTA, B Lee W, Cao Schoene 27. P, Boehnke CB, zircon. Keller in Diffusion 26. (2003) DJ Cherniak 25. 0 pne J aksot ,Cwo A hieB(03 o l uecniet are supercontinents all Not zircon (2013) in B Dhuime isotopes PA, Hf Cawood C, What Hawkesworth (2017) CJ, K Spencer Suzuki Y, 30. Hibiya K, Itano T, Yamaguchi T, Iizuka 29. 5 gomD,NesnS,Pdre K eeanJ 20)Gailefcslimiting effects Glacial (2009) JE Lesemann VK, Pedersen sheets. SB, ice Nielsen infracambrian DL, by Egholm erosion 35. Deep (1973) of WA characteristics Chemical (1980) White SR sediment- Taylor 34. RJ, pure Arculus AE, of Bence significance DA, Gust and MR, identification Perfit The 33. (2017) al. isotope Hafnium et juvenile TN, The Hopkinson (2016) MJ 32. kranendonk Van CL, Kirkland NJ, Gardiner 31. oeaeaalbeat available are code oeadDt Availability. Data https://macrostrat.org/api and and Code elevation Cretaceous. ridge (103– the of ocean plateau much average for basalt level high sea flood with increased oceanic consistent Ontong–Java potentially and 105), Kerguelen Superchron Normal the has Long Cretaceous Cretaceous and the the with in conjunction flow in proposed heat been mantle increased controversial, While rate. elre al. et Keller .Ptr E usnJ 21)Sdmn yln ncnietladoenccrust. oceanic and continental on cycling Sediment (2017) JM Husson of SE, growth Peters unsteady by 9. driven oxygenation Atmospheric (2017) SE Peters America. JM, North Husson of Macrostratigraphy 8. (2006) SE Peters 7. Unconfor- ‘Great one make unconformities Many (2012) JM the Timmons for KE, trigger Karlstrom a as Unconformity’ 6. ‘Great the of Formation (2012) of RR Gaines analysis SE, Quantitative Peters (1980) KB 5. Seslavinsky AN, Balukhovsky VE, Khain AB, America. Ronov North of 4. interior cratonic the in Sequences (1963) LL Sloss 3. (1914) (1875) CD GB Walcott Goode 2. E, Coues AH, Thompson JW, Powell 1. neapefo h ilmnt-niord ou ahlt,Nel Mountains, Needle batholith, Eolus isograd, sillimanite-in the USA. Colorado, from example An evolution. system Earth for implications Geology reservoir. sedimentary continental the 73–79. pp 489, Paper Special GSA CO), City, Boulder Am, Soc (Geol KE Karlstrom mity’. explosion. Cambrian sedimentation. Phanerozoic 74:93–114. 57. Vol DC), Washington, Press, Baltimore Lord (The Collections Tributaries its and West the of River ihshrceouinaot25Grago. Gyr 2.5 about evolution lithospheric einn . a eceia osrit rmgentn aat n komatiites. and basalts greenstone from Front constraints Geosci Geochemical Ga: 2.5 beginning stp evidence. isotope history. Earth throughout Geosyst Geophys Geochem eyln n hnrzi vlto ftecnietlcrust. crust. continental the of evolution 80–104. Phanerozoic and recycling Nature unique. was assembly Rodinian fCU:Cntansfo nqiirtdcodie n mlctosfrtebulk the for implications and planets. terrestrial chondrites of unequilibrated composition from and Constraints marine CHUR: of of areas and sea the by flooded deposition. areas continental on based approach quantitative aiyadtedfeetaino h otnna crust. continental the of differentiation the and parity compilation. global A Ma: history. Earth throughout abundance rae qa:Gnwn-oii aestudy. case Gondwana-Rodinia equal: created evolution. crust–mantle about us tell onanheight. mountain 84:1841–1844. sources. mantle for Implications basalts: island-arc granites. derived cycles. supercontinent of record a as signal nuRvErhPae Sci Planet Earth Rev Annu rn aynGooy w ilo er fErhsHistory Earth’s of Years Billion Two Geology; Canyon Grand 190:854–857. 45:323–326. 7:543–553. er Nova Terra at lntSiLett Sci Planet Earth Geology Nature abinGooyadPaleontology and Geology Cambrian Nature https://github.com/brenhinkeller/GreatUnconformity. mJSci J Am 460:884–887. at lntSiLett Sci Planet Earth l oplddtst n opttoa source computational and datasets compiled All . at lntSiLett Sci Planet Earth 31:873–876. 25:48–56. 18:872–888. 484:363–366. Geol Sediment arsrtgahcdt r cesbevia accessible are data Macrostratigraphic a Commun Nat 43:363–404. 294:777–801. at lntSiLett Sci Planet Earth Gvrmn rnigOfie ahntn DC). Washington, Office, Printing (Government Lithos ece esetLett Perspect Geochem 467:57–63. at lntSiLett Sci Planet Earth e iea Geochem Mineral Rev mrigTpLf Sci Life Top Emerging Nature c Rep Sci 25:311–325. 274–275:304–327. 8:1950. Geology 481:290–304. Nature 461:85–95. 485:490–493. hmGeol Chem 6:38503. Geol J 228:273–275. Nature 273:48–57. 41:795–798. xlrto fteColorado the of Exploration mtsna Miscellaneous Smithsonian , Geology 460:68–75. 114:391–412. 3:179–189. 53:113–143. 30:227–256. 523:301–307. 2:125–136. 42:503–506. elScA Bull Am Soc Geol d imn JM, Timmons eds , at-c Rev Earth-Sci elScA Bull Am Soc Geol 97: 3 rw 20)Daiyo hra eie stedsiciecaatrsi fplate of characteristic distinctive the is impact regimes an thermal of of Duality signature (2006) Thermal-metamorphic M (1998) Brown G 63. Stevens R, WU RL, Gibson 62. Moln 61. Thermochronol- (2017) AK Ault SN, Thomson S, Marshak WR, Guenthner MS, DeLucia 60. ther- K-feldspar 40Ar/39Ar and Structure (2010) base- MC Quigley York M, Heizler KE, New extreme Karlstrom and rates the erosion 59. long-term Low of (2006) PW Reiners history SA, Bowring RM, thermal Flowers The 58. (1998) TM Harrison MT, Heizler 57. (2011) eds G, Shields-Zhou studies GP, Case Halverson Earth: E, snowball Arnaud on depocenters 48. Sedimentary (2017) slope al. et PF, glaciogenic Hoffman 47. diamictites–Neoproterozoic Vreeland (2000) ME geology- Cryogenian and McMechan dynamics climate 46. Earth Snowball of (2017) coupling al. Asynchronous et Earth: PF, Hoffman Snowball (2017) 45. ME Zugger JF, Kasting and D, world Pollard the 44. of map geological Generalized (1995) Canada of Survey Geological 43. Earth’ ‘snowball Neoproterozoic (2001) (2000) PASSC WR Peltier 42. SK, the Baum between TJ, conflict Crowley a WT, there Hyde Is 41. (2003) C Ritz A G, 1. Ramstein formation: F, Earth Fluteau snowball Y, during Donnadieu variations level 40. Sea (2013) WR Peltier and Y, glaciation Gaskiers Liu the of 39. Geochronology snowballs: Dodging (2016) al. et chronology: JP, Cryogenian Pu A (2015) 38. FA Macdonald AD, Brandon JV, Strauss AD, glaciation. Rooney Neoproterozoic for 37. context palaeogeographic A (2009) ZX Li PF, Hoffman 36. 6 atnC,Sde M(06 lwntsdmn cuuainst nwalEarth snowball sets accumulation sediment net Slow beneath (2016) erosion Rapid PM (2012) Sadler D evacuation CA, Mair sediment Partin A, Sole 56. and I, Bartholomew erosion P, of Nienow T, Rates Cowton (1996) 55. Tre- J the Bogen of L, timing relative Hunter and B, origin Hallet the 54. on Constraints (2012) al. et CV, Rose 53. Neopro- in stones stranger of geochronology U-Pb (1997) ME Villeneuve GM, evo- Ross Neoproterozoic (2014) 52. BF Windley V, Pease MJ, Whitehouse chronostratigraphic NE, the Rantakokko on 51. constraints Geochronologic (2007) al. et glacially SA, Neoproterozoic Bowring A (2006) 50. U Wiechert A, Cozzi JL, Etienne PA, Allen R, Rieu 49. L5103.TMG cnwegsfnigfo h ntdKingdom United the NE/R004978/1. from Grant Grant Council Council funding Research Environment Research and Natural acknowledges Australia EAR-1150082 T.M.G. by Grants Berkeley supported FL150100133. NSF was Compu- (through by R.N.M. Energy supported DE-FG02-97ER25308. Foundation ICER-1440312. of Contract was Getty under Department development sup- fellowship US Macrostrat Gordon Graduate the acknowledges Science and from tational C.B.K. and Ann anonymous Center) manuscript. the Geochronology and the from Cawood, improved port Peter greatly Macdonald, reviewers Francis discussion. valuable ACKNOWLEDGMENTS. etnc ic h Neoarchean. the since tectonics Africa. South dome, Vredefort the in event the for deposits. implications Cu-Ni-PGE and vein-type Canada, of Complex, Igneous genesis Sudbury Range, North the of units cycle supercontinent the to surface Earth. Unconformity snowball Great and the of denudation links Boulder, ogy Am, Soc (Geol MA Lamb LS, model. section Beard 331–352. crustal PJ, pp tilted Umhoefer CO), the eds of 463, Reevaluation Paper block: Special Butte GSA Gold the of history mal dates. (U-Th)/He ancient by documented stability continental studies. K-feldspar 40Ar/39Ar 103:29795–29814. from determined ment Namibia. northern in Formation Chuos Sturtian the from Columbia. British northeast 246–261. Mountains, Rocky deposits, geobiology. model. climate global a with flow sea-glacier Technical Canada), Ontario, Ottawa, 2915d. Canada, Report of Survey (Geological databases linked Canada). Brunswick, New Fredericton, model. climate/ice-sheet coupled a with simulations improved An modeling. interpretation: numerical Earth with snowball understanding the and deposits glacial Neoproterozoic analysis. preliminary biota. Ediacaran the of appearance first the glaciations. Neoproterozoic synchronous long-lasting Two Palaeoecol Palaeoclimatol Palaeogeogr pr rmalyugrgailepisodes. glacial younger all from apart sheet. ice implications. Greenland their the and data field of review 12:213–235. A glaciers: by 319–320:241–250. zona of Report ages Technical Canada), and Ontario, 1997-F. Ottawa, provenance Canada, for of Survey Implications (Geological deposition Cordillera: the Canadian of diamictites, geochronology terozoic refined a on Oman. based of basement Sultanate Arabian region, Marbat eastern the of lution Oman. of Sultanate Supergroup, Huqf 307:1097–1145. Neoproterozoic the with of associated framework carbonate cap Oman. ‘atypical’ southern area, and Mirbat palaeovalleys, succession bedrock margin basin influenced 36. Glaciations proterozoic rF akno H oe C(01 utpehdohra rcse nfootwall in processes hydrothermal Multiple (2001) PC Jones DH, Watkinson F, ar ´ δ 3 nml eo h n-roeinglaciation. end-Cryogenian the below anomaly 13C c Adv Sci at matDatabase Impact Earth 3:e1600983. Geology epy e oi Earth Solid Res Geophys J Go o odn odn,Gooia oit eorNo. Memoir Society Geological London), London, Soc (Geol ..Mlo,MJ elr n .Sheeprovided Schoene B. and Keller, M.J. Maloof, A.C. Geology 46:167–170. Geology 40:343–346. elScSe Publ Spec Soc Geol 277:158–172. Geology TePaeayadSaeSineCentre, Science Space and Planetary (The cnGeol Econ 34:961–964. at lntSiLett Sci Planet Earth Geology Geology epy e Atmos Res Geophys J 118:4410–4424. 44:1019–1022. NSLts Articles Latest PNAS 96:1645–1670. Nature 26:787–790. ai Res Basin h elgclRcr fNeo- of Record Geological The 44:955–958. 392:101–127. epy e oi Earth Solid Res Geophys J Geosphere Geology 405:425–429. ulCnPtGeol Pet Can Bull Geology lblPae Change Planet Global 208:101–112. at lntSiLett Sci Planet Earth 18:471–496. 122:5157–5171. 43:459–462. 13:811–837. 34:925–928. | mJSci J Am f10 of 9 48:

EARTH, ATMOSPHERIC, AND PLANETARY SCIENCES 64. Harlow GE, Tsujimori T, Sorensen SS (2015) Jadeitites and plate tectonics. Annu Rev 86. Belousova E, et al. (2010) The growth of the continental crust: Constraints from zircon Earth Planet Sci 43:105–138. Hf-isotope data. Lithos 119:457–466. 65. Sizova E, Gerya T, Brown M (2014) Contrasting styles of Phanerozoic and Precambrian 87. Dhuime B, Hawkesworth CJ, Cawood PA, Storey CD (2012) A change in the continental collision. Gondwana Res 25:522–545. geodynamics of continental growth 3 billion years ago. Science 335:1334–1336. 66. Palin RM, White RW (2016) Emergence of blueschists on Earth linked to secular 88. Spencer CJ, et al. (2014) Proterozoic onset of crustal reworking and collisional changes in oceanic crust composition. Nat Geosci 9:60–64. tectonics: Reappraisal of the zircon oxygen isotope record. Geology 42:451–454. 67. Blackburn TJ, et al. (2012) An exhumation history of continents over billion-year time 89. Payne JL, et al. (2016) Strengths and limitations of zircon Lu-Hf and O isotopes in scales. Science 335:73–76. modelling crustal growth. Lithos 248–251:175–192. 68. Chumakov NM (2010) Precambrian glaciations and associated biospheric events. 90. Zheng YF, et al. (2004) Zircon U-Pb and oxygen isotope evidence for a large-scale 18O Stratigr Geol Correl 18:467–479. depletion event in igneous rocks during the Neoproterozoic. Geochim Cosmochim 69. Germs GJB, Gaucher C (2012) Nature and extent of a late Ediacaran (ca. 547 Ma) Acta 68:4145–4165. glacigenic erosion surface in southern Africa. S Afr J Geol 115:91–102. 91. MacLennan S, et al. (2018) The arc of the Snowball: U-Pb dates constrain the Islay 70. Chumakov NM (2011) Glacial deposits of the Bokson Group, East Sayan Mountains, anomaly and the initiation of the Sturtian glaciation. Geology 46:539–542. Buryatian Republic, Russian Federation. Geol Soc Lon Memoirs 36:285–288. 92. Morris JD, Leeman WP, Tera F (1990) The subducted component in island arc lavas: 71. Egholm DL, et al. (2017) Formation of plateau landscapes on glaciated continental Constraints from Be isotopes and B–Be systematics. Nature 344:31–36. margins. Nat Geosci 10:592–597. 93. Crisp JA (1984) Rates of magma emplacement and volcanic output. J Volcanol 72. Snyder FG (1947) The problem of the Lipalian interval. J Geol 55:146–157. Geotherm Res 20:177–211. 73. White WA (1972) Deep erosion by continental ice sheets. Geol Soc Am Bull 83:1037– 94. Smith EI (1971) Determination of origin of small lunar and terrestrial craters by depth 1056. diameter ratio. J Geophys Res 76:5683–5689. 74. Garrity CP, Soller DR (2009) Database of the Geologic Map of North America–Adapted 95. Pike RJ (1974) Depth/diameter relations of fresh lunar craters: Revision from from the Map by J.C. Reed, Jr. and others (2005), US Geological Survey Data Series, spacecraft data. Geophys Res Lett 1:291–294. 424. Available at http://pubs.usgs.gov/ds/424/. Accessed August 11, 2017. 96. Lambeck K, Smither C, Johnston P (1998) Sea-level change, glacial rebound and 75. Licciardi JM, Clark PU, Jenson JW, Macayeal DR (1998) Deglaciation of a soft-bedded mantle viscosity for northern Europe. Geophys J Int 134:102–144. Laurentide ice sheet. Quat Sci Rev 17:427–448. 97. Laske G, Masters G, Ma Z, Paysanos M (2013) Update on CRUST1.0–A 1-degree global 76. Egyed L (1956) Determination of changes in the dimensions of the Earth from model of Earth’s crust. Geophys Res Abstr 15:EGU2013-2658. Palægeographical data. Nature 178:534. 98. Artemieva IM (2006) Global 1 x 1 thermal model TC1 for the continental lithosphere: 77. Armstrong RL (1969) Control of sea level relative to the continents. Nature 221:1042– Implications for lithosphere secular evolution. Tectonophysics 416:245–277. 1043. 99. Larson RL (1991) Latest pulse of Earth: Evidence for a mid-Cretaceous superplume. 78. Hargraves RB (1976) Precambrian geologic history. Science 193:363–371. Geology 19:547–550. 79. Lee CTA, et al. (2017) Deep mantle roots and continental emergence: Implications for 100. Williams GD, Stelck CR (1975) Speculations on the Cretaceous paleogeography of whole-Earth elemental cycling, long-term climate, and the Cambrian explosion. Int North America. The Cretaceous System in the Western Interior of North America: Geo- Geol Rev 164:1–18. logical Association of Canada Special Paper, ed Caldwell WGE (Geol Assoc Canada, 80. Whitehead JA (2017) Dimensions of continents and oceans–Water has carved a Waterloo, Ontario, Canada), pp 1–20. perfect cistern. Earth Planet Sci Lett 467:18–29. 101. Eriksen MC, Slingerland R (1990) Numerical simulations of tidal and wind-driven 81. Lopez-Gamund´ ´ı OR, Buatois LA (2010) Introduction: Late Paleozoic glacial circulation in the Cretaceous Interior Seaway of North America. Geol Soc Am Bull events and postglacial transgressions in Gondwana. GSA Special Paper 468, eds 102:1499–1516. Lopez-Gamund´ ´ı OR, Buatois LA (Geol Soc Am, Boulder, CO), pp v–viii. 102. Hancock JM, Kauffman EG (1979) The great transgressions of the Late Cretaceous. J 82. Herman F, et al. (2013) Worldwide acceleration of mountain erosion under a cooling Geol Soc 136:175–186. climate. Nature 504:423–426. 103. Courtillot V, Olson P (2007) Mantle plumes link magnetic superchrons to phanerozoic 83. Gumsley AP, Chamberlain KR (2017) Timing and tempo of the great oxidation event. mass depletion events. Earth Planet Sci Lett260:495–504. Proc Natl Acad Sci USA 114 :1811–1816. 104. Tarduno JA, Cottrell RD, Smirnov AV (2002) The Cretaceous superchron geodynamo: 84. Swanson-Hysell NL, et al. (2010) Cryogenian glaciation and the onset of carbon- Observations near the tangent cylinder. Proc Natl Acad Sci USA 99:14020–14025. isotope decoupling. Science 328:608–611. 105. Zhang N, Zhong S (2011) Heat fluxes at the Earth’s surface and core–mantle bound- 85. Childs OE (1985) Correlation of stratigraphic units of North America–COSUNA. AAPG ary since Pangea formation and their implications for the geomagnetic superchrons. Bull 69:173–180. Earth Planet Sci Lett 306:205–216.

10 of 10 | www.pnas.org/cgi/doi/10.1073/pnas.1804350116 Keller et al. Downloaded by guest on September 24, 2021