Meyers And Peters - Department Of Geoscience

S.R. Meyers, S.E. Peters / Earth and Planetary Science Letters xxx (2011) xxx–xxxwhere Δt time interval, Xbt number of gap bound packagesthat cross both the lower and upper time interval boundaries,XbL number of gap bound packages that only cross the lower timeinterval boundary, XFt number of gap bound packages that onlycross the upper time interval boundary.3. ResultsMTM power spectral analysis indicates that most of the varianceobserved in the Ased data series can be explained by a stochasticautoregressive model (at the 99% confidence level), with the exceptionof a region of power centered on a frequency of 1/56 Myr (Fig. 3c).MTM harmonic F-test results confirm the presence of a highlysignificant (N99.9% confidence level) phase-coherent sinusoid inAsed, exhibiting a periodicity of 56 Myr (Fig. 3b). The varianceattributable to this frequency component (gray area in Fig. 3c), oncedeconvolved from the Ased power spectrum (Thomson, 1982), yieldsan estimate of the residual spectrum that is consistent with theautoregressive red noise model. The 56 Myr cycle accounts for 14% ofthe observed variance in the Ased power spectrum (b1/20 Myr 1), anduncertainty in its true periodicity is estimated to be 56 Myr / 3 Myr( / 0.5 fray; fray Rayleigh resolution). EHA of the Ased recordreveals the temporal evolution of this 56 Myr beat (Fig. 3d), which ispersistent throughout most of the Phanerozoic, with the notableexception of a gap at approximately 325 Ma.3SSA extracts two dominant modes of variability in the Ased dataseries: (1) a sinusoidal mode with variance concentrated at afrequency of 1/56 Myr 1 (Fig. 3a and Fig. 4b), and (2) a longerterm Phanerozoic mode (Fig. 4a). Together, these modes of variabilityaccount for 95% of the variance in Ased (Fig. 4c), with the 56 Myroscillation (Fig. 4b) representing 19% of the total variance. Thus, theSSA results for the Ased data series afford strong independentconfirmation of the MTM and EHA results, and also corroborate adiminishment of the 56 Myr cyclic signal at 325 Ma (Fig. 3). The SSAresults further highlight a relatively diminished 56 Myr cyclicvariability in Ased during the Cenozoic.Finally, comparison of the extracted 56 Myr rhythm (Fig. 4b) withthe independently estimated macrostratigraphic rates (Fig. 4d)illustrates that most of the observed oscillations are associated withan increase in macrostratigraphic package initiation and/or truncationrate (asterisks in Fig. 4d). Based on these results, we conclude that the56 Myr Ased cycle is a consequence of changes in both packageinitiation and truncation rate, but the relative importance of these twodistinct phenomena is variable. The highest magnitude packageturnover rates (initiation-truncation; Fig. 4d) are observed at theCretaceous–Paleogene boundary, the Permian–Triassic boundary, andthe middle Ordovician ( 465 Ma), all of which are preceded bypronounced negative turnover rates that reflect a reduction in thearea of sedimentation. In total, these truncation, initiation, andturnover rate responses are diagnostic of geographic expansion oferosion/non-deposition (sea-level fall and “sequence boundary”Fig. 3. Cyclic variation in North American sedimentation. a, Area of preserved sedimentary strata (Ased) in North America during the Phanerozoic (blue), with North Americansequences (Peters, 2008; Sloss, 1963) and a 56 Myr oscillatory component extracted using SSA (red). b, MTM harmonic F-test spectrum for Ased, calculated using three 2π prolatetapers (Thomson, 1982). Dashed lines indicate the 99% and 99.9% confidence levels for F-values. c, MTM power spectrum for Ased, calculated using adaptive weighting and three 2πprolate tapers (Thomson, 1982). Total power has been divided by the 99% confidence level for a robust AR(1) red noise model (Mann and Lees, 1996). Grey area indicates powerattributable to the 56 Myr phase-coherent sinusoid identified in part b, following spectral reshaping (Thomson, 1982). d, EHA of Ased, using three 2π prolate tapers and a 150 Myrmoving window. A linear trend has been removed from each 150 Myr data window prior to harmonic analysis.Please cite this article as: Meyers, S.R., Peters, S.E., A 56 million year rhythm in North American sedimentation during the Phanerozoic, EarthPlanet. Sci. Lett. (2011), doi:10.1016/j.epsl.2010.12.044

4S.R. Meyers, S.E. Peters / Earth and Planetary Science Letters xxx (2011) xxx–xxx(d)(b)MacrostratigraphyPackage RatesPackageTruncationRate(Myr-1)(c)56 Myr "M-curve"oscillatory 56 Myr"M-curve" component cycle2 19%1 2 3 4 5 6 7 8 910Sedimentary Rock Area(x106 km2)2 95%****************(e)ProposedNorth AmericanCratonic Sequences} Tejas/Cenozoic} Zuni/Cretaceous-0.15-0.1-0.0500.050.10.15 76%00.050.10.150.20.25Time uncation(Myr-1)} Jurassic} Triassic} Permain} Kaskaskia 2/Carboniferous} Kaskaskia 1} Tippecanoe} SaukFig. 4. Singular spectrum analysis (SSA) results for the Ased data series, as compared to macrostratigraphy package rates. The Ased data series is shown as a thin black line, and the SSAreconstructions are shown in red. SSA utilizes an embedding dimension of 80 with the Vautard and Ghil covariance estimator (Ghil et al., 2002), and extracts four dominantreconstructed components (RCs). a, RC 1 and RC 2 (combined), which comprise the long term Phanerozoic “M-curve”. b, RC 3 and RC 4 (combined) constitute an oscillatory pair withvariance concentrated at a frequency of 1/56 Myr 1. c, Total SSA reconstruction, which incorporates RC 1, RC 2, RC 3 and RC 4, and accounts for 95% of the variance in Ased.d, Package initiation rate (green), package truncation rate (blue), and initiation rate-truncation rate (black), determined by macrostratigraphic analysis. Initiation and truncationrates are calculated according to the estimated per-package per million-year rate adapted from Foote (2000). e, Designation of major North American cratonic sequences based onmacro-cyclostratigraphic analysis. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)development) followed by sea-level rise and the restoration ofsediment accumulation and preservation. This is in contrast to theTriassic/Jurassic boundary, for example, where truncation andinitiation rate responses are characterized by a degree of covariation,indicating changing loci of deposition on the continent, but also ageneral reduction in the area of sedimentation (Fig. 4).The Permian–Triassic and middle Ordovician turnover maximanoted above approximately coincide with minima in the 56 Myr Asedcycle (Fig. 4). The lack of a strong 56 Myr Ased minimum associatedwith the Cretaceous–Paleogene boundary is potentially due to theincreased influence of terrestrial sediment during the Cenozoic (up to72% of the area), in contrast to pre-Cenozoic sediments, which aredominated by marine strata (Peters, 2006; averaging 92% of the area).This hypothesis is supported by a separate analysis of the area ofmarine strata (Amar), documented below, which indicates a 55 MyrAmar minimum associated with the Cretaceous/Paleogene boundary(Figs. 5 and 6).Preliminary analysis of the coverage area of marine strata (Amar),yields similar results, although generally of lower statistical significance (Fig. 5). The noisier nature of Amar (relative to Ased) maycorrectly characterize the nature of marine sedimentation on NorthAmerica, or alternatively, it could be due to inaccuracies in thepartitioning of marine and terrestrial deposits in the current version ofthe macrostratigraphic database. Given the present status of the Amardata series, we tentatively identify a 55 Myr / 3 Myr ( / 0.5 fray)cycle that accounts for 8% of the variance observed in the Amar powerspectrum (b1/20 Myr 1), a 55 Myr oscillatory signal extracted by SSA,constituting 13% of the Amar variance (Figs. 5a and 6b), and a long-termPhanerozoic mode accounting for 83% of the Amar variance (Fig. 6a).A notable distinction between Amar and Ased is the indication of astrong 55 Myr SSA oscillation in marine sedimentation during theCenozoic, and distortion (elongation) of the SSA oscillation during the“Zuni” cycle (compare Figs. 4b and 6b). Both Amar and Ased reveal adiminishment of the oscillatory signal centered on 325 Ma. Minimain the 55 Myr Amar rhythm approximately coincide with maxima inpackage turnover rates observed at the Cretaceous–Paleogene boundary, the Permian–Triassic boundary, and in the middle Ordovician(Fig. 6).4. DiscussionBuilding upon the previous work of Sloss (1963) and Peters(2008), our macro-cyclostratigraphic analysis of North Americamotivates a redefinition of the North American cratonic sequences.We propose nine individual 56 Myr-long cratonic sequences(Figs. 4e and 6e), including four units previously defined qualitativelyby Sloss (Sauk, Tippecanoe, Zuni/Cretaceous, Tejas/Cenozoic), twounits previously defined by Peters (Triassic and Jurassic, whichtogether comprise part of Sloss' Absaroka), and three new sequences(Kaskaskia 1, Kaskaskia 2/Carboniferous, Permian, which togethercomprise Sloss' Kaskaskia and part of the Absaroka). With theexception of Kaskaskia 1 and 2, each of these cratonic sequences isassociated with an increase in macrostratigraphic package initiationand/or truncation rate (asterisks in Figs. 4d and 6d). Our proposedseparation of Kaskaskia 1 and 2 is based on two criteria: (1) theobservation of a strong 56 Myr oscillation, extracted via singularspectrum analysis, which defines Kaskaskia 1 (Figs. 4b and 6b) and (2)the occurrence of a substantial change in the nature of sedimentation(from siliciclastic-dominated to carbonate-dominated) betweenKaskaskia 1 and 2, which coincides with the Devonian/Carboniferousboundary (Peters, 2008).Cycles in sedimentation of similar duration have also beenidentified in the European stratigraphic record spanning the upperNorian to middle Eocene (Smith and McGowan, 2005), supporting thehypothesis that the 56 Myr oscillation in sedimentation may beglobal, and at least partially influenced by eustatic sea-level change.Additional evidence for a eustatic origin of sedimentation cyclescomes from a complimentary analysis of the coverage area of marinestrata in North America (Amar), which exhibits a 55 Myr oscillation(Fig. 5).Please cite this article as: Meyers, S.R., Peters, S.E., A 56 million year rhythm in North American sedimentation during the Phanerozoic, EarthPlanet. Sci. Lett. (2011), doi:10.1016/j.epsl.2010.12.044

S.R. Meyers, S.E. Peters / Earth and Planetary Science Letters xxx (2011) xxx–xxx5Fig. 5. Cyclic variation in North American marine sedimentation. a, Area of preserved marine strata (Amar) in North America during the Phanerozoic (blue), with North Americansequences (Peters, 2008; Sloss, 1963) and a 55 Myr oscillatory component extracted using SSA (red). b, MTM harmonic F-test spectrum for Amar, calculated using three 2π prolatetapers (Thomson, 1982). Dashed line indicates the 99% the confidence levels for F-values. c, MTM power spectrum for Amar, calculated using adaptive weighting and three 2π prolatetapers (Thomson, 1982). Total power has been divided by the 99% confidence level for a robust AR(1) red noise model (Mann and Lees, 1996). Grey area indicates power attributableto the 55 Myr phase-coherent sinusoid identified in part b, following spectral reshaping (Thomson, 1982). d, EHA of Amar, using three 2π prolate tapers and a 150 Myr movingwindow. A linear trend has been removed from each 150 Myr data window prior to harmonic analysis.The identification of a significant, long-period rhythm in the extentof sedimentation on the North American continent is a novel resultthat demands the presence of a large-scale oscillatory forcingmechanism. DeCelles et al. (2009) recently proposed such amechanism for generating long-term cycles in orogenic activity, andthis hypothesis may help to explain the origin of cratonic sequences(Sloss, 1963) and the quantitative periodic signal documented here.The orogenic oscillator hypothesis (DeCelles et al., 2009) invokesupper plate processes operating at ocean–continent convergentmargins, and a 50 Myr cycle in magmatic activity in the SierraNevada batholith is comparable to the 56 Myr cycle we observe inthe extent of sedimentation. A link between orogenesis andsedimentation is expected because crustal uplift and subsidence areone determinant of the extent to which sediments are delivered toand accumulate on continental crust. Importantly, the history ofocean–continent convergence in North America is dominated by eastcoast orogenesis during the Palaeozoic, and west coast orogenesisduring the Mesozoic and Cenozoic, potentially yielding two geographically distinct oscillators. Diminishment of the 56 Myr Asedcycle during the Carboniferous could be the consequence ofinterference between an eastern (Acadian orogeny) and western(Antler orogeny) oscillatory forcing during this time.Quasi-periodic mantle plume formation (Mjelde and Faleide,2009; Prokoph et al., 2004; Ribe et al., 2007), hypothesized to occurbased on experimental laboratory simulations (Schaeffer and Manga,2001), represents an alternative mechanism for generating theobserved 56 Myr beat in North American sedimentation. Mantleplume formation is expected to influence sedimentation on thecontinents by forcing eustatic sea-level changes and by affecting theaverage freeboard of the continents. The link between mantle plumeformation and continent-scale stratigraphic patterns would be furtheraccentuated by climate interactions (Miller et al., 2005), wherebyglobal warming due to increased rates of mantle degassing results inhigher sea-levels (e.g., via ice sheet melting and thermal expansion)and wider areas of marine sedimentation on the continents.A longer-term Phanerozoic mode in sedimentation (Figs. 4a and6a), linked to the well-known “M-curve” or Wilson cycle ofsupercontinent coalescence and breakup (Heller and Angevine,1985; Peters, 2008), is also identified quantitatively in these data.This important Phanerozoic mode defines two mega-sequences, onethat spans the Palaeozoic and another that spans the Mesozoic/Cenozoic. Our analysis indicates that this signal is indistinguishablefrom a stochastic autoregressive process at the 99% confidence level(Figs. 3c and 5c), although the length of the data series limits ourability to detect oscillatory modes on such timescales. This resultimplies that the physical mechanism responsible for the supercontinent cycle and mega-sequence formation is potentially comprised oftwo components: (1) a Gaussian white-noise source (for example,Please cite this article as: Meyers, S.R., Peters, S.E., A 56 million year rhythm in North American sedimentation during the Phanerozoic, EarthPlanet. Sci. Lett. (2011), doi:10.1016/j.epsl.2010.12.044

6S.R. Meyers, S.E. Peters / Earth and Planetary Science Letters xxx (2011) xxx–xxx(d)(c)(b)2 13%2 2*1 2 3 4 5 6 7 8 910Marine Area(x102 km6)} Jurassic} Triassic} Permain} Kaskaskia 2/Carboniferous} Kaskaskia 1} Tippecanoe} Sauk***} Tejas/Cenozoic} Zuni/Cretaceous**(e)ProposedNorth AmericanCratonic Sequences00.050.10.150.20.25Time (Ma)"M-curve"2 83%00.050.10.150.20.250.30.350.4"M-curve"55 Myr 55 50400450500550MacrostratigraphyPackage Rate(Myr-1)InitiationTruncation(Myr-1)Fig. 6. Singular spectrum analysis (SSA) results for the Amar data series, as compared to macrostratigraphy package rates. The Amar data series is shown as a thin black line, and the SSAreconstructions are shown in red. SSA utilizes an embedding dimension of 80 with the Vautard and Ghil covariance estimator (Ghil et al., 2002), and extracts four dominantreconstructed components (RCs). a, RC 1 and RC 2 (combined), which comprise the long term Phanerozoic “M-curve”. b, RC 3 and RC 4 (combined) constitute an oscillatory pair withvariance concentrated at a frequency of 1/55 Myr 1. c, Total SSA reconstruction, which incorporates RC 1, RC 2, RC 3 and RC 4, and accounts for 96% of the variance in Amar.d, Package initiation rate (green), package truncation rate (blue), and initiation rate-truncation rate (black), determined by macrostratigraphic analysis. Initiation and truncationrates are calculated according to the estimated per-package per million-year rate adapted from Foote (2000). e, Designation of major North American cratonic sequences based onmacro-cyclostratigraphic analysis. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)(a) Area of North(b) Global diversityAmerican strataof short-lived genera050MarineStrata100Total Sed.Strata150Time (Ma)2002503003504004505002345Area2(x106 km )601002003004005006007008009001000550due to changes in heat production within the core), and (2) an Earthsystem component/interaction with the appropriate memory timescale (response time) to integrate the white-noise, producing theobserved “M-curve”.In addition to yielding insight into the characteristics of NorthAmerican stratigraphic sequences and the nature of cyclic controls onsedimentation on the continents, our results have important implications for the Phanerozoic record of biologic diversity. Previous workhas identified a pronounced 62 Myr / 3 Myr cycle in the number ofmarine genera (Rhode and Muller, 2005), which is statistically similarto the 56 Myr / 3 Myr Ased cycle that we observe in North Americansedimentation (Fig. 7). Thus, cyclical variation in global diversity islikely to be either the consequence of a fossil preservation bias (Petersand Foote, 2001; Smith, 2001; Smith, 2007; Smith and McGowan,2007), or, more likely, related to global environmental changes thatare directly (Hallam, 1989; Johnson, 1974; Newell, 1952; Sepkoski,1976; Simberloff, 1974) or indirectly linked to the expansions andcontractions of epicontinental seas (Hallam and Wignal, 1999; Peters,2008). The strong covariation of sedimentologic and faunal recordsduring the Phanerozoic is further underscored by the coincidence ofmany North American cratonic sequences with geologic systemboundaries, most of which were originally defined on the basis ofbiological turnover, but that also occur in phase with a prominent, butas-yet poorly understood, large-scale rhythm in our planet.Number of genera5. ConclusionsFig. 7. Comparison of the 56 Myr oscillatory signal extracted from the North Americanstratigraphic record, and global fossil biodiversity data. (a) The 55 Myr SSA oscillatorycomponent derived from Amar (red line) and the 56 Myr SSA oscillatory componentderived from Ased (blue line). (b) The diversity of short-lived genera during thePhanerozoic (Rhode and Muller, 2005). (For interpretation of the references to color inthis figure legend, the reader is referred to the web version of this article.)In this study, we apply emerging macrostratigraphic andcyclostratigraphic methodologies to quantitatively evaluate temporalmodes (“orders”) of stratigraphic variability across the NorthAmerican continent during the Phanerozoic. MacrostratigraphicPlease cite this article as: Meyers, S.R., Peters, S.E., A 56 million year rhythm in North American sedimentation during the Phanerozoic, EarthPlanet. Sci.

Stephen R. Meyers⁎, Shanan E. Peters Department of Geoscience, 1215 W. Dayton Street, University of Wisconsin-Madison, Madison, WI 53706, USA article info abstract Article history: Received 20 August 2010 Received in revised form 14 Dece