Regular and irregular climatic cycles and the Belemnite Marls ...

Journal of the Geological Society, London, Vol. 147, 1990, pp. 915-918, 4 figs. Printed in Northern Ireland

SHORT PAPER

pliedby the number of samples collected (915cm) by an acceptable 3%. These figures compare favourably with Sellwood’s(1972) value of 8%cmforthesameinterval. Calcium carbonate and total organic carbon(TOC) contents were determined coulometrically using aStrohlein Coulomat 702 (Fig. 1; Jenkyns 1988). The Belemnite Marls arecharacterized by sinusoidal variations in carbonate and organic-carbon contents (Fig. 1). The maxima and minima of %CaCO, and %TOC occur at the centres of beds although there is an overlap in the fields of composition of the different rock types (Fig. 2). The%TOC shows ageneral non-linear inverse correlation with the %CaCO, (Fig. 2A). However, this correlation is partly caused by varying dilution of the small amount of organic matter by the large variations incalcium carbonatecontents, so TOC has been re-expressed on acarbonate-free basis(Fig. 2B). This reduces thedegree of correlation, but there remains nonetheless a nonlinear inverse relationship between the two components. If the measured variations in %TOC had resulted from carbonate dilution of a constant amount of organic carbon in the non-carbonate fraction, Fig.2Bwould have shown a horizontal band of points. The time series possess a relatively simple structure. Decimetre-scale variations in carbonate and TOC contents largely correspond to couplets of light mar1 with dark mar1 or laminated shale. Modulating the couplet compositions are metre-scale variations in average carbonate and organic-carbon levelswhich produce bundles of beds visible in the field (Fig. 1). Finally there is agradual upsection decrease in the thickness of the couplet and bundle variations. To investigate the structure of the time series further, Fourierspectraand cross spectra were generatedforthe %CaCO, and %TOC(carbonate-free) time series. Variance or power spectra can be employed to detect regular cyclicity denoted by spectral peaks (Jenkins & Watts 1968). Coherency spectra record the degree of linear correlation of amplitude variatons intwo time seriesfora series of frequencies; phasespectra indicate the relative timing of the variations. Standard Blackman-Tukey spectral methods (i.e. the truncated lagged-autocovariance method) were used; the time series were initially linearly detrended and 116lags were employed (Jenkins & Watts 1968). The interpretation of spectra derived using a thickness- ratherthana timescale isdiscussed elsewhere (Weedon 1989). Both variance spectra show pair a of peaks corresponding to cyclesof300 and 37.5cm(Fig. 3A).The wavelengths of these peaks correspond to the thickness of the bundles and couplets in the lower two-thirds of the section. Due to the frequency resolution or bandwidth of the spectra, the bundle cycle wavelength may lie anywhere between 191-705 cm, whereas the wavelengths indicated by the couplet peak lie between 35-40cm. Therefore, although the couplet Variations can be described as‘regular’, the bundle cycle wavelength might vary considerably. Atthe frequencies of these cycles(0.33 and 2.67cycles per metre) the carbonate and TOC (carbonate-free) variations are significantly coherent or correlated (Fig. 3B). Thethinner couplets in thetop third of the section explain a slightshelfin the variance spectra and probably account forthe broad region of high coherency between 3.0 and 4.5 cycles permetre.Atthefrequencies of the

Regular and irregular climatic cycles and the Belemnite Marls (Pliensbachian, Lower Jurassic, Wessex Basin) G .P .W E E D O N ’ & H . C.JENKYNS’ ‘Department of Earth Sciences, The University, Cambridge CB2 3EQ, UK ’Department of Earth Sciences, The University, Oxford OX1 3PR, U K

Time series of carbonate and organic-carbonvalues are reported for 10.5 m of the Belemnite ’Ibe light and dark mad beds form decimetre-scale couplets and m e t r e - d e bundles.Light marls are carbon-poor and carbonate-rich whereas dark mnrls and laminated shales are carbon-richandcarbonate-poor.Sediment composition appears to have been indirectly controlled byclimate; the regolar couplets may record orbital-precessioncycles (20 La). However, the bundlescannot be linked to changesineccentricity;ratherthey apparently record irregular,large-amplitude climatic variations with periods of a few hundred thousand years. It shodd not be assumed that 100-400ka cyclicity in climate-related sedimentary sequences can be explained only by orbital forcing.

Mnrls.

Ithas recently been suggested thatthe LowerLias of the Blue Lias (HettangianBritain, and in particular Sinemurian), mightoweitscyclically bedded nature to orbital-climatic control of sedimentation (House 1985, 1986; Weedon 1985). This interpretation reflects the current interest in ancient sedimentary cyclicity and Milankovitch orbital cycles (Berger 1988,1989; papers accompanying et al. 1990). Because they appear Smith1989;Fischer relatively untouched by diagenesis, theBelemnite Marls were selected for detailed stratigraphic sampling and geochemical analysis as part of an extensive study of Lower Lias mudrocks. Herethe time series from theBelemnite Marls are used to reassess the merit of the orbital-climatic hypothesis as applied to the British Lower Lias.

Previous work. Lang et al. (1928)first described the sequence of theBelemnite Marls in detail;their bed numbering scheme is followed here. The sequence consists of varying contents of micrograde carbonateand claywith small amounts of organic matter,pyrite and bioclasts. Sellwood (1970, 1972) used theoccurrence of burrowmottled bed contactsandthelaminationtodemonstrate the primary distinction of rock types,andnotedthatthe proportion of bioclasts varies according to the rock type as does the diversity of trace fossils. Time series. Individual light marl/dark mar1 couplets can be traced continuously for several kilometres along the DorsetCoast. Bed numbers 110a-117 of Lang et al. (1928) were selected for sampling andremeasuredbetween Stonebarrow Hill and St Gabriel’s Mouth near Charmouth. The thinnest bed present is about 3 cm thick so, to avoid aliasing, samples were obtainedatconstant 3cm intervals. 350 sampleswere collected regardless of burrow-mottling and bed contacts, but visible fossils (mainly belemniteguards) were avoided. The measured thickness of beds 110b-116, 885cm, differs from that im-

915

916

SHORT PAPER

A %TOC

.m

-

Fig. 1. Calcium-carbonate and organic-carbon time series from the

U.jamesoni Zone (Pliensbachian) of the Belemnite Mark, Charmouth, Dorset. Bed numbers after Lang et al. (1928), ammonite zonation after Palmer (1972). The depth scale is in centimetres and all 350 samples were collected at 3 cm intervals. Samples were analysed inrandom order and measurements were made in duplicate and averaged. The gradual changes in composition from sample to sample probably reflect bioturbational smoothing and homogenization during crushing. bundle and couplet cycles there is a difference in phase of 180°, confirming the inverse relationship between the two measured components (Figs 2B and 3C).

0

:

:

:

:

:

50

40

:

60

+ :

:

:

l0

l

80

%CaC03

Mar1 Dark Mar1 o Laminated Shale c Light X

Fig.

2. Relationship between carbonate and organic-carbon contents. Mean compositions are: all rock types(350 samples) 51.7% CaCO, and 1.69% TOC; light marls (162 samples) 58.8% CaC0, and 0.94% TOC; dark mark (105 samples) 46.7% CaCO, and 1.84% TOC; laminated shale (83 samples) 44.0% CaCO, and 2.95% TOC.

Diagenesis. Thereareno early diagenetic limestones or stylolites in the Belemnite Mark, but a least some diagenetic dissolution is suggested by the absence of fossils that would have had aragonitic shells. Closed-system redistribution of carbonatecannot explain the inverse correlation of %TOC(carbonatefree) with %CaCO, (cf. Ricken1986; Fig. 2B). Instead,thelargevariations of carbonate in the couplet cycles are best interpreted as primary (Sellwood1970). The bundle cycles equally appear to be primary as their lowest carbonate contents are associated both with carbonate-poor beds that have aboveaverage organic-carbon contents measured (as and carbonate-free) and with the occurrence of laminated shales (Fig. 1). Carbonatecontentsandcementation were not,therefore, simply linked tothe local abundance of organic carbon and the supply of early diageneticcarbonate ions (Figs 1 & 2). So any cementationrelated to degradation of organic matter andlor bioclast dissolution either occurred uniformly, regardless of primary compositions, or emphasized original carbonate variations. Cycliaty and its environmental causes. Spectral analysis has demonstratedthe constancy of couplet thickness in the lower two-thirds of the section studied. This ismost plausibly explained by aconstantperiod of deposition for each couplet cycle. The thinnercoupletsand bundles

+ :

30

A

l

Fig.3.

(A)Variance-, (B)coherency- and (C) phase-spectra for the %CaCO, and %TOC (carbonate free) time series. The bandwidth defines the frequency resolution of the spectra. Both 300 cm spectral peaks are significant at the 80% level. The 37.5 cm peak for %CaCO, is significant at the 98% level and the 37.5 cm peak for %TOC carbonate free is significantat the 90% level.

G . P. W E E D O N & H . C . JENKYNS

in theupperpart of this section andthe even thinner couplets of Lang et al.'s beds 117-120 (not analysed) thus imply a gradually reduced rate of accumulation, as does the relatively high concentration of macrofossils_in these strata. Any sedimentoiogical modeifor the Belemnite Marls must account for varying bottom-water oxygenation as indicated by thetrace fossils, laminatedshaleand TOC values. It must also explain the relationship between %TOC andthe proportion of clay to carbonate (Fig. 2B). The changesin oxygenation levelscould have been caused by variations in: water-column density stratification linked to the supply of development of a thermocline and/or low-salinity water from runoff; and/or surface productivity determined by the supply of nutrientsfrom runoff or upwelled nutrient-rich water (e.g. Weedon 1985; Herbert & Fischer1986; House 1986). The carbonate/clayratio was determined by the relative production of carbonate-walled versus organic-walled plankton and the relative influx of clay and possiblycoastally-derived micrograde carbonate.As global ice volumes were apparently small at this time, any eustatic rise and fall of sea level, over periods of tens- to hundreds-of-thousands of years, was probably too small to effectsignificant changes in thenature of the sediment (Frakes 1979; Weedon 1985).Alocal tectonic/subsidence control of water depth as an explanation for the cyclicity is discounted based on the regularity and estimated period of the couplet cycles and the symmetrical, sinusoidal shape of both the couplet and bundle cycles. It is not yet possible to choose between the plausible models, but those favoured all imply some form of climatic variation. The time series described here resemble those of Herbert & Fischer(1986)from the mid-Cretaceous of the Italian Apennines.Theydemonstrated couplets, bundles andlong-wavelengthcycles attributed tothe 21,100 and 410ka orbital cycles and, as in this study, showed that low-frequency cycles controlled average couplet composition and the occurrence of laminated shales.

Regular andirregularclimaticcycles. To investigate the time series of the Belemnite Marls further,the bundle cyclewasdefinedby repeated smoothing of the carbonate data usinga three-point moving average (Fig.4). The couplet cyclewas then isolated by subtracting, point by point, thesmoothedcarbonate curve fromthe original time series (Fig. 4). Three recent timescales ascribe 7.0-8.0Ma to the five ammonite zones of the Pliensbachian(Hallam et al. 1985; Haq et al. 1988; Harland et al. 1990). The 10.5m section contains 38 couplets and 6recognizable bundles (Fig. 4) and spans 43%by thickness of the U. jamesoni Zone usingSellwood's(1972)log. This suggests that the section represents some 645 ka. Allowing for uncertainties in estimating the temporal proportion of the U. jamesoni Zone which the section represents, and the errors in dating, the section might represent between 323 and 1290 ka. Thus on average the couplet cycleshad an estimated period of between 8 and 34ka and the bundles an average period of between 54 and 215 ka. If the section contains undetected hiatuses, the cycles would have had shorter durations than those estimated. The couplet cyclesshow large, gradual variations in amplitude with10-13 packets of 2-5 cycles (Fig. 4, cf. Weedon1989). Considering the regularity and approxim-

917

%CaC03 Measured

Bundle cycle8 %CaC03 Running mean

Couplet cycles %CaCO, mlnus Mean %CaC03 -20-10

10 2 0

Fig.4. Isolated bundle and couplet cycles. The bundle cycles were isolated by smoothing the carbonate time series until all couplet-scale variations were removed. This involved 90 applications of a three-point moving average with weights: 0.25,0.5,0,25 and end-weights:0.5,0.5. Couplet cycles were isolated by subtracting the smoothed curve from the original data. All curves have the same horizontal scaling. Note that the envelope around the couplet cycles, which definesthe amplitude packets, is out of phase with the bundle cycles. ate period of the couplets, they probably relate to either the precession or obliquitycycleswhich,in the Early Jurassic, had periods calculated as 20 and 37 ka (Berger et al. 1989). The estimated period and particularly the amplitude variability favours theformer. The variable number of couplets per amplitude packet might then result from undetected hiatuses and small changes in the period of the precession cycle (Berger 1989; Weedon 1989). The bundle cycles,definedby the smoothed carbonate curve, are highly variable both in terms of wavelength and amplitude and they contain from 4 to 11 couplets (Fig. 4). Some of the variation in wavelengthcan be attributed to the presumed decrease in sedimentation rate at the top of the section. Based onthe dating alonethe bundle cycles might be ascribed to the 100 ka eccentricitycycle.However, due to the interaction of different orbital elements, variationsineccentricity(i.e. the 410 and 100 ka cycles) automatically determinethe amplitude of the precession index(Fischer et al. 1990). Thus if the couplets record precession andthe bundles eccentricity, the amplitude of the couplet variations should be directly related to the phase of, i.e. the position within, the bundle cycle (Herbert & Fischer1986; Weedon 1989). In this case, however, the amplitude of couplet variations is unrelated to the shape of the bundle cycle(Fig.4). This observation cannot be explainedbyeccentricity-drivenbundling and an incomplete record. Insteadthe bundle cycleis irregular both in terms of thickness and probably, given the variable number of couplets in each bundle, period (Figs 1 & 4). Note that if the couplets related to obliquity rather than precession the bundles wouldhave hadan average

918

PAPER

SHORT

period of234 ka(i.e.(38couplets X 37 ka)/6bundles) which again rules out the eccentricity cycle. Assuming thatthe strata1 bundling was caused by the same environmental phenomena that produced the couplets, the former probably reflects irregular, largeamplitude climatic variations which had periods of tens- to a few-hundred-thousand years. Such climatic variations are not currentlypredicted by the orbital-climatic hypothesis. However, they have recently been described from Pleistocenedeep-searecordswhere,although enigmatic, they have been tentatively linked to the change from predominantly 40 ka to 100 ka climatic variationsatabout 700 ka B.P. (Ruddiman et al. 1989). As a non-linear, dynamicsystem it is possible forshort-term forcing of the climate togenerate long-term climatic variability (Nicolis & Nicolis1984; James & James 1989). Thus another possibilityis that non-linearity (and ‘chaotic’ behaviour) has caused orbitally-forced climatic variations to lead to longperiod irregular climatic oscillations that form part of a ‘red noise’ climatic spectrum (LeTreut & Ghil 1983). Hence it should not be assumed that bundles of strata in climate-related cyclic sequences can be attributed only to the regular 100 and 410 ka orbital eccentricity cycles. We urge caution when ancient deep-sea cyclic sequences are used to refine absolute timescales (cf. House 1985; Gale 1989; Fischer et al. 1990). G.P.W. was variously supported byBP’S Stratigraphy Branch and Research Fellowships from NERC and Downing College, Cambridge. H.C.J. was supported by a grant from BP Research to study the Dynamic Stratigraphy of the Wessex Basin. Our thanks to M. Heslop (Cambridge) and S . Wyatt (Oxford) for their valiant technical help and to D. G . Smith for an incisive review.

References BERGER,A. 1988. Milankovitch theory and climate, Reviews of Geophysics 26, 624-657. -1989. The spectral characteristics of Pre-Quaternary climate records, an example of the relationship between astronomical theory and J. C. (eds) geosciences. In: BERGER,A.,SCHNEIDER,S., DUPLE~SY Climate and Geosciences, Reidel, NATO AS1 Series C, 285,47-76. -, LOUTRE,M. F. & DEHANT, V. 1989. Pre-Quaternary Milankovitch frequencies. Nature, 342, 133. FISCHER, A. G., DE BOER, L.P. & ~ E M O L I SILVA, I. 1990. Cyclostratigraphy. In: GINSBURG, R. N. & BEAUDOIN,B. (eds) CretaceousResources,Eventsand Rhythm, Reidel, NATO AS1 Series C, 304, 139-172. FRAKES, L. A. 1979. Climates Throughout Geologic Time. Elsevier.

GALE,A. S. 1989. A Milankovitch scale for Cenomanian time. Terra Nova, 1,420-425. HALLAM A., HANCWK J. M., BREOUE LA, J. L., L ~ W R IW. E , & CHANNELL, J. E. T. 1985. Jurassic to Palaeogene. In: SNELLING, N.L. (ed.) The Chronology of the Geological Record. Geological Society, London, Memoir, 10, 118-140. HAO,B. L., HARDENBOL, J. & VAIL,P. R. 1988. MesozoicandCenozoic chronostratigraphy and cycles of sea-levelchange. In: WILGUS,C. K., HASTINGS, B. S . , KENDALL, G., POSAMENTIER, H. W., Ross, C . A. & VANWAGNER, J. C. (eds) Sea-level ChangesanIntegrated Approach. SpecialPublication of the Society of EconomicPaleontologistsand Mineralogists. 42,71-108. HARLAND, W. B., ARMSTRONG, R. L., Cox, A. V., CRAIG,L. E., SMITH,A. G. & SMITH,D.G. 1990. A Geologic TimeScale 1989. Cambridge University Press. HERBERT, T. D. & FISCHER, A. G. 1986. Milankovitch climatic origin of mid Cretaceous black shale rhythms in central Italy. Nature, 321, 739-743. HOUSE,M. R. 1985. A new approach to an absolute timescale from measurements of orbital cycles and sedimentary microrhythms. Nature, 315, 712-725. - 1986. Are Jurassic sedimentary microrhythms due to orbital forcing? Proceedings of the Ussher Society, 6, 299-311. JAMES, 1. N. & JAMES, P. M. 1989. Ultra-low-frequency variability in a simple atmospheric circulation model. Nature, 342,53-55. JENKINS, W. M. & WATIS,D. G . 1968. Spectral Analysis and its Applications. Holden-Day. anoxic event: JENKYNS,H. C. 1988. The early Toarcian (Jurassic) stratigraphic, sedimentary, and geochemical evidence. American Journal of Science, 288, 101-151. LANG,W. D., SPATH,L. F., Cox, L. R. & MUIR-WOOD, H. M. 1928. The Belemnite Marls of Charmouth, a series in the Lias of the Dorset Coast. Quarterly Journal of the Geological Society of London, 84, 179-257. LE TREUT,H. & GHIL,M. 1983. Orbital forcing,climatic interactions, and glaciation cycles.Journal of Geophysical Research, 88, 5167-5190. NICOLIS,C. & NICOLIS, G. 1984. Is there a climatic attractor? Nature, 311, 529-32. PALMER, C. P. 1972. A revision of the zonal classification of the Lower Lias of the Dorset Coast in south-west England. Newsletters on Stratigraphy, 2, 45-54. RICKEN, W. 1986. Diagenelic bedding: a model for marl-limestone alternations. Lecture Notes in Earth Sciences No. 6, Springer Verlag. D. G., CLEMENT,B.M. & RUDDIMAN, W. F., RAYMO, M. E., MARTINSON, BACKMAN, J. 1989. Pleistocene evolution: northern hemisphere ice sheets and North Atlantic Ocean. Paleoceanography,4, 353-412. SELLWOOD, B. W. 1970. The relation of trace fossils to small-scale J. C. sedimentary cycles in the British Lias. In: CRIMES,T. P. & HARPER, (eds) Trace Fossils. Geological Journal Special Series No. 3, 489-503. - 1972. Regional environmental changesacross a LowerJurassic stage boundary in Britain. Palaeontology, 15, 125-157. SMITH,D. G . 1989. Milankovitchcyclicity and the stratigraphic record-a review. Terra Nova, 1, 402-404. WEEDON, G. P. 1985. Hemipelagicshelf sedimentation and climaticcycles: the basalJurassic (Blue Lias) of South Britain. EarthandPlanetary Science Letters, 16, 321-335. -1989. The detection and illustration of regular sedimentary cycles using Walshpower spectra andfiltering,withexamplesfrom the Lias of Switzerland. Journal of the Geological Society, London, 146, 133-144.

Received 9 February 1990; revised typescript accepted 26 June 1990.