Long-period orbital control on middle Miocene global ... · isotope enrichment events in the late Paleogene and Neogene. The stepwise character of the middle Miocene cooling event

Long-period orbital control on middle Miocene global cooling:

Integrated stratigraphy and astronomical tuning

of the Blue Clay Formation on Malta

H. A. Abels,1 F. J. Hilgen,1 W. Krijgsman,2 R. W. Kruk,1 I. Raffi,3

E. Turco,4 and W. J. Zachariasse1

Received 24 December 2004; revised 7 April 2005; accepted 21 June 2005; published 10 November 2005.

[1] Application of an astronomical age model to a bulk carbonate oxygen isotope record in the Ras il Pellegrinsection on Malta indicates that the major step in the middle Miocene global cooling (13.82 Ma ± 0.03) coincideswith minimum eccentricity values associated with the 400-kyr cycle and minimum obliquity amplitudesassociated with the 1.2-Myr cycle. This orbital configuration is very similar to that found for comparable oxygenisotope enrichment events in the late Paleogene and Neogene. The stepwise character of the middle Miocenecooling event appears to be controlled by the combined influence of the 100-kyr eccentricity cycle and the172-kyr cycle in obliquity amplitude. The integrated stratigraphy further allows extension of the astronomicalpolarity timescale to the top of chron C5ACn. The boundary between the Globigerina Limestone and the BlueClay Formation coincides with the major step in middle Miocene global cooling and provides a level suitablefor placing the physical reference point for the Langhian/Serravallian boundary.

Citation: Abels, H. A., F. J. Hilgen, W. Krijgsman, R. W. Kruk, I. Raffi, E. Turco, and W. J. Zachariasse (2005), Long-period orbital

control on middle Miocene global cooling: Integrated stratigraphy and astronomical tuning of the Blue Clay Formation on Malta,

Paleoceanography, 20, PA4012, doi:10.1029/2004PA001129.

1. Introduction

[2] Following the Miocene climate optimum between 16and 14.5 Ma the middle Miocene global cooling at around14 Ma represents a next important step in the evolutionof Cenozoic climate. The climate change did go alongwith changes in ocean circulation and floral and faunaldistribution, with an increase in the meridional temperaturegradient, and with the permanent installation of a larger EastAntarctic Ice Sheet [see Flower and Kennett, 1994].[3] The cause of the middle Miocene cooling has been

ascribed to increased weathering of silicate rocks due touplift in the Himalayan-Tibetan region [e.g., Raymo andRuddiman, 1992] and to increased burial of organic carbon[e.g., Vincent and Berger, 1985], both leading to thewithdrawal of CO2 from the atmosphere and hence areduction of the greenhouse capacity. However, availablepCO2 reconstructions based on different proxies do notshow convincing evidence for lower atmospheric CO2

values after or during middle Miocene cooling [Pagani etal., 1999; Pearson and Palmer, 2000; Royer et al., 2001].Further, changes in ocean circulation patterns, for example,due to tectonic closure of basins, may have increased

moisture transport or reduced heat transport to the Antarcticregion [Shevenell and Kennett, 2004].[4] Additionally, orbital parameters and especially long-

period variations in obliquity amplitude may have played animportant role in Cenozoic climate change by punctuatinglonger-term trends or by positive feedback mechanisms thatpushed climate into a new state. The influence of long-period orbital cycles as put forward by authors [e.g.,Lourens and Hilgen, 1997] is now held responsible forvarious climate steps in the late Paleogene and Neogeneperiods [Turco et al., 2001; Zachos et al., 2001; Wade andPalike, 2004]. To determine the possible role of long-periodorbital forcing on the middle Miocene cooling it is neces-sary to develop high-resolution astronomical age models forclimate proxy records across the critical time interval.Increased ice volume and lowered ocean temperaturesduring the middle Miocene are particularly evident fromopen ocean benthic oxygen isotope records. The mostprominent enrichment events around 14 Ma have beenlabeled Mi3a and Mi3b [Miller et al., 1991, 1996] andE1, E2, and E3 [Woodruff and Savin, 1991; Flower andKennett, 1993]. The middle Miocene cooling is clearlycharacterized by a first relatively small step at �14.2 Ma(Mi3a and E1 or E2) and a second major step at �13.8 Ma(Mi3b and E3) [see also Shevenell and Kennett, 2004].The second step is associated with a shift to heavier d13Cvalues in benthic foraminiferal carbonate and labeled CM6[Woodruff and Savin, 1991]. The age control of these stableisotope records however is not good enough to examine thepossible role of long-period astronomical forcing in trigger-ing middle Miocene global cooling. Independent dating byastronomical tuning of records across the middle Miocene

PALEOCEANOGRAPHY, VOL. 20, PA4012, doi:10.1029/2004PA001129, 2005

1Department of Earth Sciences, Utrecht University, Utrecht, Nether-lands.

2Paleomagnetic Laboratory, ‘‘Fort Hoofddijk,’’ Utrecht University,Utrecht, Netherlands.

3Dipartimento di Scienze della Terra, Universita ‘‘G. D’Annunzio,’’Chieti Scalo, Italy.

4Dipartimento di Scienze della Terra, Universita di Parma, Parma, Italy.

Copyright 2005 by the American Geophysical Union.0883-8305/05/2004PA001129$12.00

PA4012 1 of 17

climate transition is needed [see also Westerhold et al.,2005].[5] One of the few places where the middle Miocene

climate transition can be studied in continuous marinesuccessions on land is on Malta and Gozo located in thecentral Mediterranean (Figure 1). Earlier studies on theseislands have shown that the major middle Miocene isotopicenrichment event (Mi3b, E3, CM6; �13.8 Ma) coincideswith the transition from the Globigerina Limestone to theBlue Clay Formation [Jacobs et al., 1996; John et al., 2003]and that the cyclically bedded Blue Clay is potentiallysuitable for astronomical dating [Sprovieri et al., 2002;John et al., 2003]. However, the tuned ages of Sprovieriet al. [2002] for calcareous plankton events in the middleand upper part of the Blue Clay Formation are significantlyyounger than tuned ages for the same bioevents in the time-equivalent Monte dei Corvi and Tremiti sections [Hilgen etal., 2003]. We therefore decided to independently establishan integrated stratigraphy and astronomical tuning for theuppermost part of the Globigerina Limestone, and the lowerand middle part of the Blue Clay Formation on Malta, usinghigh-resolution biostratigraphic correlations to the well-tuned Monte dei Corvi and Tremiti sections as a startingpoint. In addition, we established a bulk isotope record tolocate the position of the major middle Miocene globalcooling step and associated events.

2. Geological Setting, Section, and Lithology

[6] The sediments of the Globigerina Limestone and BlueClay Formation on Malta and Gozo were deposited in theMaltese Graben System, a series of Miocene-Quaternaryextensional basins that developed on the African foreland ofthe Sicilian Apennine-Maghrebian fold and thrust belt [DeVisser, 1991; Dart et al., 1993]. The sedimentary successionexposed on Malta and Gozo consists of five differentformations: Lower Coralline Limestone (late Oligocene),Globigerina Limestone (Aquitanian-Langhian), Blue Clay

(Serravallian), Greensand (Serravallian-Tortonian), and Up-per Coralline Limestone (Tortonian) [Felix, 1973]. Allboundaries are conformable except for the contact betweenthe Greensand and the Upper Coralline Limestone Forma-tion. Sediments of the Globigerina Limestone and BlueClay were deposited in a low-energy, open marine environ-ment. Paleobathymetric estimates point to a depth of around500–600 m for the Globigerina Limestone and the BlueClay [Bonaduce and Barra, 2002; Bellanca et al., 2002].Benthic foraminiferal [Bellanca et al., 2002] and claymineral [De Visser, 1991; John et al., 2003] studies of theGlobigerina Limestone and the Blue Clay suggest overalllower bottom water oxygen levels, more humid climateconditions, and more intense continental weathering duringthe deposition of the Blue Clay.[7] The Ras il Pellegrin section (RIP), exposed along the

FommIr-RihBay on thewest coast ofMalta (Figures 1 and 2),covers the middle Globigerina Limestone up to the UpperCoralline Limestone Formation and was selected because ofits excellent exposures and its distinct sedimentary cycles. Atotal number of 540 samples were taken from the top part ofthe Globigerina Limestone up to and including the middlepart of the Blue Clay, measuring in total 48.5 m with anaverage sample distance of 9 cm. A transitional bed separatesthe yellowish marly limestones of the Globigerina Limestonefrom the grey clayey marls of the Blue Clay. No indicationsof erosion or deformation were found at the formationboundary. The measured section ends 20 m below theGreensand Formation and largely overlaps with the nearbyRas il Pellegrin A section studied by an Italian research team[Bellanca et al., 2002; Bonaduce and Barra, 2002; Foresi etal., 2002; Sprovieri et al., 2002]. The Blue Clay at Ras ilPellegrin shows a very distinct and characteristic pattern ofhomogeneous grey and white colored marls (Figure 2). Thepresence of two sapropels and several levels with chondritetrace fossils [Bromey, 1990] point to occasional anoxic ordysoxic bottom water conditions. The sapropels and chon-drite trace fossils occur in the grey marl beds. Finally,volcanic minerals found in residues around 40.45 m pointto an ashfall in the younger part of the Blue Clay.

3. Cyclostratigraphy

[8] In the lower and middle part of the Blue Clay wedistinguished six whitish colored marly intervals, numberedI to VI in Figure 3. These intervals, separated by intervalsdominated by grey marls, correspond to the large-scalecyclicity recognized in the Blue Clay by Sprovieri et al.[2002] and John et al. [2003]. The small-scale cyclicity isless easy to distinguish in the field [see also Sprovieri et al.,2002]. Rather arbitrarily, we define a small-scale cycle toconsist of a greyish marl bed at the base followed by awhitish marl at the top. The Blue Clay part of the studiedsection contains 44 small-scale cycles that can be recog-nized in the field and have been labeled I.1 to VI.16 assubdivision of the larger scale intervals (Figure 3).[9] Discrimination of small-scale cycles is particularly

difficult in interval VI. The calcium and potassium contentin the samples was measured using ICP-OES to obtain aquantitative measure of the lithology and hence the cyclicity(Figure 3). (The 540 samples were taken at an average

Figure 1. Location maps of the Ras il Pellegrin (RIP)section on Malta and the Tremiti section in Italy, bothindicated by ‘X’.

PA4012 ABELS ET AL.: ORBITAL CONTROL ON MID-MIOCENE GLOBAL COOLING

2 of 17

PA4012

distance of 9 cm. A part of each sample was dried, crushed,powdered and homogenized. 125 mg of each sample wassubsequently dissolved in 2.5 mL HF (40%) and 2.5 mLmixing acid (HNO3 (16.25%) and HClO4 (45.5%)) andheated at 90�C in a closed tube for at least 8 hours. Then thesample was dried by evaporating the acids at 160�C anddissolved in 25 ml HNO3 (4.5%). These solutions wereanalyzed by a Perkin Elmer Optima 3000 ICP-OES appa-ratus, at Utrecht University, for the elements Al, Ba, Ca, Ce,Co, Cr, Cu, Fe, K, Li, Mg, Mn, Na, Ni, P, S, Sc, Sr, Ti, V, Yand Zn. The relative errors in duplicate measurements ofinternational standards was for all elements lower than 3%,except for Ce, Co, P and S, Sr and Y. Results were checkedand where needed refined by making use of internationalstandards (ise-921) and Utrecht University in-house stand-ards (SO-1).) Ca mainly reflects the (calcium)carbonatecontent, because multiplying the Ca value by 2.496, assum-ing that all Ca is derived from carbonate, yields an averageCaCO3-content for the Blue Clay in the RIP section of23.9%, which is in agreement with CaCO3 values measured

by Sprovieri et al. [2002] and John et al. [2003], who useddifferent methods than ICP-OES. Ca and K showed a stronginverse relation (95%) and the Ca/K ratio was selected toreduce noise in the individual records (Figure 3). The large-scale cycles, observed in the field, are prominently visible inthe Ca/K record (Figure 3). The 44 small-scale lithologicalcycles distinguished in the field are recognizable in thisrecord as well, although small offsets between lithology andchemical records occur especially in interval V and VI.Blackman-Tukey spectral analysis revealed distinct peaks inthe depth domain (Figure 4a), which correspond to cyclethicknesses of �5.5 and 3.5 m. In addition, several peaksare present in the higher-frequency part of the spectrum andcorrespond to cycle thicknesses between 0.6 and 2.4 m.Band-pass filters were applied to better visualize thesmaller-scale cyclicity in the Ca/K record. The 0.6–2.4 mfilter includes all peaks in the higher-frequency domain. The0.6–1.05 m filter extracts cycles with thicknesses similar tothat of the small-scale cycles observed in the field. Thelatter filter indeed shows a strong relation with these small-

Figure 3. Lithological column of the Ras il Pellegrin section showing field-based small-scale cycle and large-scaleinterval labeling, calcium record, calcium-potassium ratio, 0.6–2.4 and 0.6–1.05 m Ca/K filter, small-scale cyclenumbering defined on the basis of chemical records, bulk carbonate isotope records, and 21-point moving averages of theisotope records. ‘‘C’’ indicates levels with chondrite trace fossils.

Figure 2. Ras il Pellegrin section seen from the southern shore of the Fomm Ir-Rih Bay showingsampling trajectories (solid lines), the boundary between Blue Clay and Upper Globigerina Limestone(dashed line), and large-scale whitish colored marly intervals (I–VI).


3 of 17

PA4012

Figure

3


4 of 17

PA4012

scale cycles and reveals the presence of 8 additional cycles,most of them in interval VI. Small-scale cycles weresubsequently numbered based on successive minima inthe (filtered) Ca/K record.

4. Oxygen and Carbon Isotope Record

[10] Our oxygen and carbon isotope record (Figure 3)based on bulk sediment for the uppermost GlobigerinaLimestone and the Blue Clay Formation in the Ras ilPellegrin section reveals a 0.6 % increase in d18O and a1.25% increase in d13C at the boundary between theGlobigerina Limestone and Blue Clay. (Bulk isotope datawere generated on dried crushed bulk sediment on a SIRA-24 of VG (vacuum generators), at Utrecht University.Results, in per mil (%) relative to the Peedee Belemnitestandard, were checked by making use of international(Naxos 45–125 um, validated with NBS-18 and 19) and

Utrecht University in-house standards (IAEA-CO-1), andby duplicate measurements of the Naxos standard everyeight samples. Relative error in duplicate measurements waslower than 2%.) A 21-point moving average was applied tothe isotope records to better visualize the long-term varia-tions. In the lower part of the Blue Clay, both d18O and d13Creach heavier values in the clay-rich intervals. From intervalIV upward the smoothed records show less prominent long-term variations in amplitude and are also less consistentwith lithology.

5. Magnetostratigraphy

[11] A standard paleomagnetic core was drilled at eachsample site, with an average sample distance of 9 cm. (Oneof each two cores was stepwise thermally demagnetizedwith temperature steps of 20� or 30�C, in a laboratory-builtshielded furnace. The other core was treated by stepwise

Figure 4. Blackman-Tukey power spectra of Ca/K records in the (a) depth domain and in (b) timedomain calculated with the AnalySeries program 1.2 [Paillard et al., 1996]. All spectra are based onrecords from bed I.1 upward to avoid disturbance from the Globigerina Limestone in the statisticalanalysis. The three spectra (r, s, and t) of Ca/K-records in time domain (Figure 4b) result from theapplication of three age models to the Ca/K record in depth domain. These age models are based on(r) astronomical ages of 10 correlated bioevents, (s) the ages for bioevents 2, 4, 7, and 9, and (t) tuning ofthe Ca/K record to precession. The 90% confidence intervals are shown in the upper right corner of eachspectrum. Plots on the right are enlargements of the left plots.


5 of 17

PA4012

alternating field demagnetization with steps of 5 mT at lowapplied magnetic fields to 50 mT at high magnetic fields.The natural remanent magnetization (NRM) was measuredafter each step on a 2G Enterprises DC Squid cryogenicmagnetometer, at paleomagnetic laboratory ‘‘Fort Hoofd-dijk,’’ Utrecht University.) Intervals that were found tocontain a reversal were resampled in more detail to betterconstrain their stratigraphic position.[12] The NRM intensity of the samples from the Globi-

gerina Limestone was very weak (0.04 mA/m) and noreliable polarities could be obtained. The Blue Clay showedmuch higher intensities (up to 125 mA/m). Demagnetizationdiagrams reveal a clear subdivision into two components,with a break at 260�C and 50 mT. In all samples, the low-temperature, low-field component is of normal polarity. Weassume that this component represents viscous magnetiteinduced by the present-day field, which normally has anunblocking temperature of �260�C. The high-temperature/high-field component is of dual polarity and we interpretedthis as the primary signal, the so-called characteristicremanent magnetization (ChRM). Many samples showedan increase in intensity at temperatures higher than 400�C,

indicating a conversion of iron sulphides into iron oxides.Strong samples (>10 mA/m) do not show the intensityincrease at 400�C and their Zijderveld diagram can beinterpreted up to 600�C.[13] Samples classified as reliable have an NRM intensity

higher than 0.1 mA/m at 270�C (or 50 mT) and a clearnormal or reversed component in the interval from 270�C to400�C (or 50 mT to 100 mT). Examples are shown inFigures 5a–5d. Other samples (60%) were consideredunreliable (examples shown in Figures 5e and 5f). Thisprocedure resulted in a magnetostratigraphic polarity patternwith four normal and three reversed intervals (Figure 6).The position of four of the six reversals could be veryaccurately determined. The upper part of the reversedinterval at around 25 m contains an interval of uncertainpolarity, mainly because of white marls that contain lowmagnetic intensity. Further, the precise position of thereversal at around 10 m is uncertain. The uppermost partof the section is also considered ambiguous because of analternation of samples revealing weak reversed polarity withsamples revealing normal ChRM. The open dots in the latterinterval represent samples showing reversed components in

Figure 5. Examples of thermal (th) and alternating field (af) demagnetization (Zijderveld) diagrams ofspecimens in the Ras il Pellegrin section with (a and c) reliable normal and (b and d) reversed polarity andwith (e and f) unresolved polarity. Closed (open) circles denote the projection on the horizontal (vertical)scale. Numbers indicate temperature (th) or magnetic field (af) steps in �C or mT. The stratigraphicposition and the number of the specimen are shown in the upper right corner.


6 of 17

PA4012

Figure 6. Magnetostratigraphic results of the Ras il Pellegrin section. Crosses indicate unresolvedsignal. In the polarity column, black (white) denotes normal (reversed) polarity, and grey indicatesundefined polarity. To the right the correlation to the ATNTS2004 and CK’95 timescales is shown onbasis of the age calibration model discussed in the text.


7 of 17

PA4012

the demagnetization interval between 270� and 400�C.Because of the low intensities (<0.1 mA/m) of thesesamples they are not included as reversed samples.

6. Biostratigraphy

[14] Initially, every other sample was examined for plank-tonic foraminiferal biostratigraphy. Later, all samples werecounted around bioevents. Countings are based on thenumber of a specific taxon in nine fields with a maximumof 20 specimens using a rectangular picking tray of 45fields. The counts are shown as number per field inFigure 7a.[15] Comparison of our biostratigraphic data with those

from DSDP 372 [Turco et al., 2003; E. Turco et al.,unpublished data, 2004] reveals that the formation boundarydoes not contain a major hiatus and is most likely contin-uous. This is consistent with our field observations.[16] The pattern of Paragloborotalia siakensis and Glo-

borotalia partimlabiata in the RIP section reveals that thesection is grossly time-equivalent with the Tremiti sectionand the lower part of the Monte dei Corvi section in Italy[Hilgen et al., 2003], which is confirmed by the LCO ofcalcareous nannofossil Sphenolithus heteromorphus in theTremiti and RIP sections. The importance of the timeequivalency of these three sections is that it allows us toimport astronomical ages from the well-tuned Tremiti andMonte dei Corvi sections to the RIP section using detailedbiostratigraphy. Unfortunately, the poor preservation of theplanktonic foraminifers in Monte dei Corvi precludes adetailed biostratigraphic comparison with the RIP section,while the excellent preservation in Tremiti makes such acomparison possible. For that reason the same biostrati-graphic procedure as used in the RIP section has beencarried out on the Tremiti section with the same highresolution (Figure 7b).[17] Although the semiquantitative patterns of the selected

taxa in both sections are grossly similar, several of theprominent short-term changes are absent in one section orshow a different amplitude. This can among others beexplained by differences in the relative number of plank-tonic foraminifers between both sections (e.g., due to differ-ences in P/B ratio between both sections), by hydrographicdifferences between both locations, or by the fact that thetime-equivalent level was not always exactly sampled. Yetwe need individual prominent short-term frequency changesto correlate both sections on the small-scale cycle level.[18] After carefully scrutinizing the patterns we selected

10 of the prominent, short-term abundancy changes tocalibrate Tremiti to the RIP section (marked 1 to 10 inFigures 7a and 7b). These 10 bioevents are codefined by theposition of other (less) well-marked frequency changes. Forexample, the choice of bioevent 3 is codefined by theposition of the short-term acme of P. siakensis, bioevents4, 5, and 6 are codefined by the position of the paracme top

of P. siakensis, and bioevent 8 is codefined by the promi-nent bioevent 7 below and the first spike of the G. menardiigroup above.

7. Astronomical Tuning

[19] Independent orbital tuning of a section can only beestablished by recognizing characteristic patterns of astro-nomical forcing in lithology and other proxies. Unfortu-nately, the various patterns in the Ras il Pellegrin (RIP)section do not provide sufficient characteristic details thatwould allow an independent tuning. Therefore we usedastronomical ages of 10 bioevents (see section 6) in theTremiti section as a first-order age control on the RIPsection.[20] The Tremiti section contains deep marine sediments

that basically show a regular alternation of indurated,whitish colored, carbonate-rich marls and less-indurated,greyish, carbonate-poor marls. Reddish layers occur regu-larly within the whitish indurated marls. The pattern of thesereddish layers shows a striking resemblance with the sap-ropel pattern at Monte dei Corvi [Hilgen et al., 2003]indicating that the red layers of Tremiti are the equivalentof sapropels. The cyclostratigraphic correlations betweenthe two sections are consistent with planktonic foraminiferalbiostratigraphic data. Hence it is safe to assume that thesame phase relationships between the sapropels and orbitalparameters, i.e., individual sapropels correspond to preces-sion minima and clusters to eccentricity maxima, are alsovalid for the reddish layers at Tremiti. Using these phaserelationships, the Tremiti section can be tuned to precession(Figure 8). The resulting age model provides astronomicalages for the bioevents in the Tremiti section (Figure 7b).The astronomical tuning of the Tremiti section will bediscussed in more detail in a future paper.[21] The astronomical ages of the 10 selected bioevents

(see section 6) were exported from the Tremiti to the RIPsection, which seems justified because both sections arelocated in the central Mediterranean Sea (Figure 1). Thecorrelated bioevents in the RIP section are used as agecalibration points for the astronomical tuning of the section.[22] This first-order age control (Figure 8, solid lines)

indicates that the distinct white part of intervals III, IV, andV correspond to eccentricity minima. However, the whitepart of intervals I and II do not directly correlate witheccentricity minima according to this age model. The agemodel further indicates that the small-scale cycles in the RIPsection are precession controlled, because the number ofsmall-scale cycles based on the combination of lithologyand the Ca/K record agrees well with the number ofprecessional cycles in the astronomical curves between thebiostratigraphically controlled age calibration points. Black-man-Tukey spectral analysis was applied on the resultingCa/K time series. The spectrum reveals power at theastronomical frequencies (Figure 4b), but is less convincing

Figure 7. Semiquantitative abundance patterns for selected planktonic foraminiferal species and of the calcareousnannofossil Sphenolithus heteromorphus in the (a) Ras il Pellegrin section and (b) Tremiti section. No data means that thespecies has not been counted in that particular interval. Numbers (1–10) refer to specific events within the abundancepatterns used for correlation between the Tremiti and the Ras il Pellegrin section.


8 of 17

PA4012

Figure 7


9 of 17

PA4012

Figure 7. (continued)


10 of 17

PA4012

if all 10 calibration points were used to generate the timeseries. This noisy behavior of the spectrum is probably dueto the close spacing of the bioevents in combination withremaining small uncertainties in the position and, hence, ageof the bioevents at Tremiti and RIP. Considerably improvedspectra were generated using only 4, arbitrarily selected,widely spaced bioevents as calibration points. One of thesealmost identical spectra is shown in Figure 4b.[23] For detailed tuning of the Ras il Pellegrin section to

precession we need to know the phase relation betweenthe small-scale cycles and precession. Unfortunately, thedetailed biostratigraphic correlations between RIP andTremiti section do not provide a unique solution.[24] The two sapropels and the presence of chondrite trace

fossils at RIP point to periodically anoxic or suboxicconditions during deposition. These sapropels and chondritebeds are almost invariably associated with Ca/K minimaand thus with the clayey marl beds (see Figure 3). The basicsmall-scale cycle in other middle to late Miocene deepmarine sections in the Mediterranean consists of an indu-rated whitish colored carbonate-rich marl and a softer greycolored carbonate-poor marl. Sapropels, and equivalentreddish layers at Tremiti, develop within the whitish marlwhen minimum precession forcing is strong enough [seeHilgen et al., 2003]. However, occasionally, the part of thewhitish marl bed directly underlying or overlying thesapropel lacks sedimentary expression or is not developed[Hilgen et al., 2003]. As a consequence, the position of thesapropels/chondrite levels in the RIP section does notprovide conclusive arguments for the phase relation of thebasic small-scale cycle to precession. Nevertheless, forthe moment we assume that the sapropels/chondrite levelscorrespond to precession minima and hence that also Ca/Kminima (and thus grey marl beds) correspond to precessionminima and summer insolation maxima. The possibility ofan opposite phase relation implies an age uncertainty for thetuned records of at least half a precession cycle.[25] Furthermore, the relationship between some of the

geochemical proxies on the small-scale cycle scale is notalways consistent throughout the section. Confirmationabout the inferred phase relation may come from futuredetailed stable isotope measurements on benthic and plank-tonic foraminifera and quantitative counts of specific plank-tonic foraminiferal species.[26] From the lowermost age calibration point, the last

common occurrence of S. heteromorphus, toward the baseof the Blue Clay at RIP the tuning is based on correlatingeach successive small-scale cycle to successively olderprecession cycles, apart from the extra thick cycle I.5 thatis assumed to contain an extra cycle, as evident from thed13C record (Figure 8). This downward extension of thetuning may result in an additional age uncertainty of oneprecession cycle for the tuning of the lowermost part of theBlue Clay Formation. When combined with the uncertaintyin the phase relation with precession this would amount to atotal uncertainty of 30 kyr. The entire tuning has not been

adjusted to better fit characteristic patterns in the astronom-ical forcing especially related to precession-obliquity inter-ference, because (1) the phase relation between small-scalecycles and precession needs further confirmation and (2) itis not known which astronomical solution should be usedfor this time interval in terms of tidal dissipation anddynamical ellipticity values [see Lourens et al., 2001;Hilgen et al., 2003]. Nevertheless, applying BlackmanTukey spectral analysis on the tuned Ca/K time series(Figure 4b) resolves the spectral peaks in the precessionand obliquity domain much better than in the BlackmanTukey spectra of the age series based on the astronomicalages for bioevents (Figure 4b). However, it should berealized that this improvement is partly a consequence ofthe tuning itself.[27] Our tuning of the Blue Clay succession on Malta

significantly deviates from the tuning proposed by Sprovieriet al. [2002], because they use incorrect astronomical agesfor selected bioevents as starting point for their tuning[Hilgen et al., 2003]. Moreover, a poor coherency existsbetween their CaCO3- and Globigerinoides spp. records,while the small-scale cycles are not recorded in a consistentway in the lithology and these proxies. Nevertheless,quantitative counts of the warm water planktonic forami-niferal group of Globigerinoides spp. in the Ras il Pellegrinsection show lower amplitude variations in the carbonate-rich parts of intervals III and IV, thereby confirming thatthese intervals correspond to eccentricity minima [see alsoSprovieri et al., 2002].

8. Discussion

[28] The magnetostratigraphy can now be calibrated to theCK95 or ATNTS2004 timescales (Figure 6) confirming theambiguous nature of the magnetostratigraphy in the upper-most part of Ras il Pellegrin. The tuning provides accurateastronomical ages for five magnetic reversals, therebyextending the astronomical polarity timescale (APTS) basedon Mediterranean successions (Table 1). There is an intervalof uncertain polarity in the upper part of chron C5AAr,although the reversal (R > N) above this interval is veryclear. The stratigraphic position, including uncertainty inter-vals, of the reversals are given in Table 1. Discrepanciesexist in the age and duration of all (sub)chrons between thetuned age model of RIP, the ATNTS2004 and CK95 time-scales (Table 1). Confirmation of the newly extended APTSby independently tuned magnetostratigraphies is thereforeneeded.[29] The tuning of the RIP confirms that the formation

boundary between the Globigerina Limestone and the BlueClay Formation and the oxygen and carbon isotope shift atthis boundary most probably correspond to the second andmajor step (Mi3b; CM6) in middle Miocene global cooling.This major isotope enrichment event is now astronomicallydated at 13.82 ± 0.03 Ma and coincides with a period ofminimum amplitudes in obliquity related to the 1.2-Myr

Figure 8. (left) Tuning of the red layers in the Tremiti section to precession (La04 1,1 [Laskar et al., 2004]), (middle)correlation of 10 bioevents between the Tremiti and the Ras il Pellegrin section (solid lines), and (right) tuning of the small-scale cycles in the Ras il Pellegrin section to precession (La04 1,1). See color version of this figure at back of this issue.


11 of 17

PA4012

Figure

8


12 of 17

PA4012

cycle and minimum values of eccentricity as part of both the400- and 100-kyr cycle. This coincidence is very similar tothe orbital configuration found for prominent oxygen isotopeexcursions to heavier values just below the Oligocene-Miocene boundary at 23.13 Ma (Mi-1 [Zachos et al., 2001;Billups et al., 2004]) and for two excursions at 10.4 and11.4Ma (Mi-6 andMi-5 [Turco et al., 2001]).A similar orbitalconfiguration during glacial peak excursions has recentlybeen found in the Oligocene [Wade and Palike, 2004].[30] The connection between late Cenozoic oxygen

isotope excursions and periods of minimum variationsin obliquity and precession amplitudes suggests that the1.2-Myr obliquity cycle and the 100- and 400-kyr eccen-tricity cycle exerted a prominent control on the factorscontrolling the size of the Antarctic ice sheet. Most likely,the reduced amplitudes of obliquity and precession (i.e., attimes of eccentricity minima) would have prevented signif-icant melting of the ice sheet during the warmer summers attimes of obliquity and precession maxima. After the middleMiocene cooling event, the isotope values do not return tovalues from before that time. This suggests that this peculiarorbital configuration was superimposed on a long-termcooling trend, thereby pushing the climatic system into acolder mode. Vincent and Berger [1985] suggest thatatmospheric CO2 drawdown already started before themiddle Miocene, while the cooling step took place at themoment that a threshold was reached. However, severalclimate proxy records reveal a climate deterioration already

from 15 to 14.5 Ma on [e.g., Miller et al., 1991; Shevenell etal., 2004]. A climate modeling study of DeConto andPollard [2003] using a general circulation model, withcoupled components for atmosphere, ocean, ice sheet, andsediment, showed that declining Cenozoic atmospheric CO2

only leads to large ice caps when a threshold is crossed.These ice caps furthermore show large orbital variations inmuch the same way as the late Cenozoic ice sheets on theNorthern Hemisphere.[31] To better depict the long-period influence of eccen-

tricity, as amplitude modulator of precession, and theamplitude modulation of obliquity, the bulk isotope recordsare compared with a target curve (EOA in Figure 9) thatcombines eccentricity (E) with obliquity amplitude (OA)variations. The EOA curve is the sum of the standardizedeccentricity curve and the 100- to 2500-kyr band-pass filterof the positively clipped standardized obliquity curve thatdescribes the envelope of obliquity maxima. We used thisenvelope because low maxima result in cooler summers athigh latitudes that prevent melting and are responsible forsubstantial ice sheet growth.[32] The actual major middle Miocene cooling step lags

the beginning of the period of minimum amplitude varia-tions in obliquity with approximately 40 kyr and theminimum values in the EOA curve with approximately20 kyr. The latter difference results from the additionalinfluence of eccentricity in the EOA curve. The age uncer-tainty of 30 kyr for the tuned ages in the lowermost part of

Table 1. Stratigraphic Position and Uncertainty Interval of Polarity Reversals at Ras il Pellegrin and Their

Corresponding Astronomical Age After Tuning the Section to Precessiona

Chron

Ras il Pellegrin

ATNTS2004,Ma

CK95,Ma D RIP-ATNTS2004 D RIP-CK95

StratigraphicPosition, m Age, Ma

Young 12.820 12.775C5Ar.2n duration 0.058 0.044Old 36.66 ± 0.1 12.916 ± 0.015 12.878 12.819 0.038 0.097C5Ar.2r duration 0.110 0.137 0.172 �0.027 �0.062Young 32.03 ± 0.03 13.026 ± 0.011 13.015 12.991 0.011 0.035C5AAn duration 0.130 0.168 0.148 �0.038 �0.018Old 26.47 ± 0.02 13.156 ± 0.011 13.183 13.139 �0.027 0.017C5AAr duration 0.164 0.186 0.163 �0.022 0.001Young 20.23 ± 0.13 13.320 ± 0.014 13.369 13.302 �0.049 0.018C5ABn duration 0.310 0.236 0.208 0.074 0.102Old 8.55 ± 1.05 13.63 ± 0.034 13.605 13.510 0.025 0.120C5ABr duration 0.058 0.129 0.193 �0.071 �0.135Young 6.39 ± 0.28 13.688 ± 0.038 13.734 13.703 �0.046 �0.015C5ACn duration 0.361 0.373Old 14.095 14.076

aUncertainty in the astronomical age is based on the uncertainty in the stratigraphic position of the reversals and anadditional uncertainty of half a precession cycle due to the possibility of an opposite small-scale phase relation (see text fordiscussion on age uncertainties). Reversal ages and (sub)chron durations are compared with the ATNTS2004 and CK95timescales and differences are shown. ‘‘Young’’ (‘‘old’’) indicates upper (lower) limit of that specific normal chron.

Figure 9. Magnetic polarity timescale of the ATNTS2004, precession and its amplitude modulator eccentricity, obliquity,and its amplitude modulator, the EOA target curve (see running text), the magnetostratigraphy of the section, the tuned bulkcarbonate isotope time series of Ras il Pellegrin, and the correlation of the bulk isotope stratigraphy to benthos isotoperecords of Deep Sea Drilling Program (DSDP) Site 588A [Flower and Kennett, 1993] and DSDP Site 747 [Wright andMiller, 1992]. The latter two isotope series are given in the depth domain. The age scale to the left of each record has beenderived using linear interpolation and extrapolation of sedimentation rates between and beyond selected magnetic reversals,with ATNTS2004 ages of the selected reversals indicated next to the magnetostratigraphy. Isotope events are labeledfollowing Flower and Kennett [1993] and Miller et al. [1996].


13 of 17

PA4012

Figure

9


14 of 17

PA4012

the section is too large to discuss the relationship betweenthe cooling step and the fundamental cycles of precessionand obliquity (i.e., 21 and 41 kyr, respectively) in detail.Therefore conclusions can only be drawn concerning theinfluence of the longer-term modulating cycles of preces-sion and obliquity.[33] Shevenell et al. [2004] investigated the astronomical

configuration at times of the middle Miocene climatetransition in the South Tasman Rise. Their oxygen isotoperecord of the benthic foraminifer Cibicidoides mundulusreveals a clear picture of the middle Miocene climatetransition with the major step toward heavier oxygenisotopes (Mi3b) at around 13.85 Ma, while their Mg/Ca–derived planktonic SST curve shows a gradual stepwisedecrease of temperatures starting at around 14.1 Ma. Thetiming of the major step in middle Miocene global coolingis consistent with our results, although their age model isbased on ages of Berggren et al. [1995] for bio-, magneto-,and isotope-events [Shevenell and Kennett, 2004]. Thealmost identical ages result from the minor discrepanciesbetween the CK95 and astronomical ages for reversalboundaries in this interval (Table 1). Shevenell et al.[2004] were not able to observe a relation between changesin their isotope record and eccentricity due to a lack ofcyclostratigraphic control and the absence of an APTS atthat time. Moreover, they did not observe a link with the1.2-Myr obliquity cycle, because their astronomical targetcurve lacks expression of the distinct 1.2-Myr minimum inobliquity amplitude in this interval. This is rather surprisingbecause they used La2004 to construct their target curve,while a distinct minimum in obliquity amplitudes between13.85 and 13.6 Ma is clearly present in our La2004-derivedtarget curves and the onset of this minimum coincides withthe major step in middle Miocene global cooling [Laskar etal., 2004].[34] Figure 9 shows the bulk carbonate isotope records of

Ras il Pellegrin in a global perspective by comparing ourrecords with those of DSDP site 588A (southwest Pacific[Flower and Kennett, 1993]) and ODP site 747 (KerguelenPlateau [Wright and Miller, 1992]). The chronology of theseopen ocean records is rescaled using ATNTS2004 ages[Lourens et al., 2004] for identifiable polarity reversals.The timing and structure of the geographically distantisotope records reveal marked similarities across the middleMiocene climate transition, despite the fact that the isotopesignal of the Ras il Pellegrin section is measured on bulkcarbonate while the two open ocean isotope records arebased on species of the benthic foraminiferal genus Cibici-doides. This strongly suggests a global control on isotoperecords across the middle Miocene climate transition, al-though the clear stepwise character of the interval before themajor transition seems to be recorded differently at varioussites in different ocean basins [e.g., Shevenell and Kennett,2004]. At the moment, it is therefore premature to connectthe isotope excursion at �14.2 Ma (Mi3a) to the distinctminimum in the EOA curve of the same age. Independentastronomical age control on different records around theworld is needed to substantiate this potential linkage.[35] The next 1.2-Myr minimum in obliquity amplitudes

starts at �12.83 Ma, i.e., from bed VI.12 upward. However,

identical minimal values in the EOA curve as during themiddle Miocene global cooling are only reached around12.6 to 12.5 Ma, so that an isotope enrichment event isexpected to occur just above the RIP section of our study(see Figure 9). However, a recently published high-resolu-tion and astronomical tuned isotope record from the SEAtlantic reveals a marked excursion to heavier valuesaround 13.2 Ma, which is not recorded on Malta or in otherpublished records of this interval [Westerhold et al., 2005].Again, this clearly indicates that additional high-resolutionand well-tuned benthic isotope records are required todepict Antarctic and Arctic ice growth events.[36] Finally, the Ras il Pellegrin section is an excellent

candidate to define the Serravallian GSSP. The criticalinterval is continuous, is well exposed, and yields a goodmagnetic signal, a well-preserved record of microfossils,and bulk sediment stable isotope records framed within anastronomical calibrated age model. The level most suitablefor defining the Serravallian GSSP is the middle Mioceneglobal cooling step (Mi-3b). Indeed, this level correspondswith the lithological transition from the Globigerina Lime-stone to the Blue Clay Formations but field and biostrati-graphical data indicate that the succession is continuousacross this boundary. This proposed level is only slightlyolder than the Sphenolithus heteromorphus LCO, that atpresent serves as a guiding criterion for the Langhian-Serravallian boundary [Lourens et al., 2004], but that isdemonstrably diachronous between the Mediterranean (at13.654 Ma (this study)) and the equatorial Atlantic (at13.523 Ma [Backman and Raffi, 1997]).

9. Conclusions

[37] The major step in middle Miocene climate cooling,characterized by a global shift to heavier d18O and d13Cvalues (Mi3b and CM6, respectively), occurred at 13.82 ±0.03 Ma and coincides with the beginning of a period ofminimum amplitude variations in obliquity related to the1.2-Myr and 172-kyr cycles, and with prominent minima inthe 100- and 400-kyr eccentricity cycles. This peculiarorbital configuration is similar to that found for otherprominent cooling events in the Oligocene and Mioceneand is particularly evident in a new astronomical targetcurve (EOA) that combines the amplitude modulators ofprecession and obliquity. In particular, the coincidence ofminima in the amplitude modulation of precession andobliquity seems imperative to punctuate global climate ona million year timescale or even to invoke irreversibleglobal climate changes. In the middle Miocene this remark-able orbital configuration may thus have triggered theclimate cooling step by crossing a certain threshold and/ordetermined the exact moment of the cooling step thatoccurred superimposed on a longer-term cooling trend.The compilation and comparison of high-resolution, astro-nomically tuned proxy records over the middle Mioceneclimate transition is necessary to further unravel the causesand consequences of this major cooling step.[38] In addition to a better understanding of the middle

Miocene cooling, the Ras il Pellegrin section on Malta hasproven to be of importance for two other aspects as well.


15 of 17

PA4012

The first aspect is the good magnetic signal by which theAstronomical Polarity Time Scale (APTS) is extended downto the top of the C5ACn chron, showing discrepancies ofless than 120 kyr with existing CK95 and ATNTS2004timescales. The second aspect is that the Ras il Pellegrinsection is suitable for defining the Serravallian GSSP, link-ing this stage boundary to a well-defined and globallysynchronous oxygen and carbon isotope event.

[39] Acknowledgments. The authors wish to thank G. Debono of theMaltese Oil Exploration Department and G. Maempel Zammit of the Ghar

Dalom Cave and Museum for the working permissions. G. MaempelZammit is also thanked for bringing the excellent Ras il Pellegrin sectionto our attention. A. Pace of the Superintendence of Cultural Heritage isthanked for understanding the need of research in order to show the value ofa certain area. We hope that our work will contribute to the cultural heritageof the Maltese archipelago by demonstrating the suitability of the Ras ilPellegrin section for formally defining the Serravallian-Tortonian boundaryat the base of the Blue Clay Formation. Silvia Iaccarino, Fabrizio Lirer, andWanda Link are acknowledged for their help in the field, and JiovannaJiannelli, Gerrit van ’t Veld, and Geert Ittman are acknowledged forbiostratigraphical sample preparation. The project of H.A. is supported bythe Dutch Science Foundation (NWO-ALW). We would like to thank L. J.Lourens for helpful comments and two anonymous reviewers for theircritical review and good suggestions, which improved the manuscript.

ReferencesBackman, J., and I. Raffi (1997), Calibration ofMiocene nannofossil events to orbitally tunedcyclostratigraphies from Ceara Rise, Proc.Ocean Drill. Program Sci. Results, 154, 83–99.

Bellanca, A., F. Sgarrella, R. Neri, B. Russo,M. Sprovieri, G. Bonaduce, and D. Rocca(2002), Evolution of the Mediterranean basinduring late Langhian–early Serravallian: Anintegrated paleoceanographic approach, Rev.Ital. Paleontol. Stratigr., 108(2), 223–239.

Berggren, W. A., D. V. Kent, C. C. Swisher, andM. P. Aubry (1995), A revised Cenozoic geo-chronology and chronostratigraphy, in Geo-chronology, Time Scales, and GlobalStratigraphic Correlation, edited by W. A.Berggren, Spec. Publ. Soc. Econ. Paleontol.Mineral., 54, 129–212.

Billups, K., H. Palike, J. E. T. Channell, J. C.Zachos, and N. J. Shackleton (2004), Astro-nomic calibration of the late Oligocenethrough early Miocene geomagnetic polaritytime scale, Earth Planet. Sci. Lett., 224, 33–44.

Bonaduce, G., and D. Barra (2002), The ostra-cods in the paleoenvironmental interpretationof the late Langhian–early Serravallian Ras il-Pellegrin section (Malta), Rev. Ital. Paleontol.Stratigr., 108(2), 211–222.

Bromey, R. G. (1990), Trace Fossils: Biologyand Taphonomy, Spec. Topics Paleontol.,vol. 3, CRC Press, Boca Raton, Fla.

Cande, S. C., and D. V. Kent (1995), Revisedcalibration of the geomagnetic polarity time-scale for the Late Cretaceous and Cenozoic,J. Geophys. Res., 100, 6093–6095.

Dart, C. J., D. W. J. Bosence, and K. R. McClay(1993), Stratigraphy and structure of theMaltese graben system, J. Geol. Soc. London,150, 1153–1166.

DeConto, R. M., and D. Pollard (2003), RapidCenozoic glaciation of Antarctica induced bydeclining atmospheric CO2, Nature, 421,245–249.

De Visser, J. P. (1991), Clay mineral stratigraphyof Miocene to recent marine sediments in thecentral Mediterranean, Geol. Ultraiectina, 75,1–243.

Felix, R. (1973), Oligo-Miocene stratigraphy ofMalta and Gozo, Meded. Landbouwhogesch.Wageningen, 73-20, 1–103.

Flower, B. P., and J. P. Kennett (1993), MiddleMiocene ocean-climate transition: High-resolution oxygen and carbon isotopic recordsfrom deep sea drilling project Site 588A,southwest pacific, Paleoceanography, 8,811–843.

Flower, B. P., and J. P. Kennett (1994), Themiddle Miocene climatic transition: EastAntarctic ice sheet development, deep ocean

circulation and global carbon cycling, Palaeo-geogr. Palaeoclimatol. Palaeoecol., 108,537–555.

Foresi, L. M., S. Bonomo, A. Caruso, E. diStefano, G. Salvatorini, and R. Sprovieri(2002), Calcareous plankton high resolutionbiostratigraphy (foraminifera and nannofos-sils) of the uppermost Langhian– lower Ser-ravallian Ras Il-Pellegrin section (Malta),Rev. Ital. Paleontol. Stratigr., 108(2), 195–210.

Hilgen, F. J., H. Abdul Aziz, W. Krijgsman,I. Raffi, and E. Turco (2003), Integrated stra-tigraphy and astronomical tuning of the Serra-vallian and lower Tortonian at Monte deiCorvi (middle-upper Miocene, northern Italy),Palaeogeogr. Palaeoclimatol. Palaeoecol.,199, 229–264.

Jacobs, E., H. Weissert, G. Shields, and P. Stille(1996), The Monterey event in the Mediterra-nean: A record from shelf sediments of Malta,Paleoceanography, 11, 717–728.

John, C. M., M. Mutti, and T. Adatte (2003),Mixed carbonate-siliciclastic record on theNorth African margin (Malta)—Couplingof weathering processes and mid Mioceneclimate, Geol. Soc. Am. Bull., 115(2), 217–229.

Laskar, J., P. Robutel, F. Joutel, M. Gastineau,A. C. M. Correia, and B. Levrard (2004), Along-term numerical solution for the insolationquantities of the Earth, Astron. Astrophys.,428 , 2 6 1 – 286 , d o i : 1 0 . 1 0 5 1 / 0 0 04 -6361:20041335.

Lourens, L. J., and F. J. Hilgen (1997), Long-periodic variations in the Earth’s obliquityand their relation to third-order eustatic cyclesand late Neogene glaciations, Quat. Int., 40,43–52.

Lourens, L. J., R. Wehausen, and H. J.Brumsack (2001), Geological constraintson tidal dissipation and dynamical ellipticityof the Earth over the past three million years,Nature, 409, 1029–1033.

Lourens, L. J., F. J. Hilgen, J. Laskar, N. J.Shackleton, and D. Wilson (2004), The Neo-gene period, in Geologic Time Scale 2004,edited by F. Gradstein, J. Ogg, and A. Smith,pp. 409–440, Cambridge Univ. Press, NewYork.

Miller, K. G., J. D. Wright, and G. G. Fairbanks(1991), Unlocking the ice house: Oligocene-Mioceneoxygenisotopes,eustasy,andmarginalerosion, J. Geophys. Res., 96, 6829–6848.

Miller, K. G., et al. (1996), Drilling and datingNew Jersey Oligocene-Miocene sequences:Ice volume, global sea level, and Exxon re-cords, Science, 271, 1092–1095.

Pagani, M., M. A. Arthur, and K. H. Freeman(1999), Miocene evolution of atmospheric car-

bon dioxide, Paleoceanography, 14, 273 –292.

Paillard, D., L. Labeyrie, and P. Yiou (1996),Macintosh program performs time-series ana-lysis, Eos Trans. AGU, 77, 379.

Pearson, P. N., and M. R. Palmer (2000), At-mospheric carbon dioxide concentrationsover the past 60 million years, Nature,406, 695–699.

Raymo, M. E., and W. F. Ruddiman (1992),Tectonic forcing of late Cenozoic climate,Nature, 359, 117–122.

Royer, D. L., S. L. Wing, D. J. Beerling, D. W.Jolley, P. L. Koch, L. J. Hickey, and R. A.Berner (2001), Paleobotanical evidence fornear present-day levels of atmospheric CO2

during part of the Tertiary, Science, 292,2310–2313.

Shevenell, A. E., and J. P. Kennett (2004),Paleoceanographic change during the middleMiocene climate revolution: An Antarcticstable isotope perspective, in The CenozoicSouthern Ocean: Tectonics, Sedimentation,and Climate Change Between Australiaand Antarctica, Geophys. Monogr. Ser., vol.151, edited by N. Exon, J. P. Kennett, andM. Malone, pp. 235–252, AGU, Washing-ton, D. C.

Shevenell, A. E., J. P. Kennett, and D. W. Lea(2004), Middle Miocene Southern Oceancooling and Antarctic cryosphere expansion,Science, 305, 1766–1770.

Sprovieri, M., A. Caruso, L. M. Foresi,A. Bellanca, R. Neri, S. Mazzola, andR. Sprovieri (2002), Astronomical calibra-tion of the upper Langhian/lower Serravallianrecord of Ras Il-Pellegrin section (MaltaIsland, central Mediterranean), Rev. Ital.Paleontol. Stratigr., 108(2), 183–193.

Turco, E., F. J. Hilgen, L. J. Lourens, N. J.Shackleton, and W. J. Zachariasse (2001),Punctuated evolution of global climate cool-ing during the late middle to early late Mio-cene: High-resolution planktonic foraminiferaland oxygen isotope records from the Mediter-ranean, Paleoceanography, 16, 405–423.

Turco, E., L. M. Foresi, S. M. Iaccarino, F. Lirer,G. Salvatorini, and M. Sprovieri (2003),Langhian planktonic foraminiferal recordfrom the Mediterranean: Paleoecological andpaleoceanographical implications, paper pre-sented at Geoitalia 2003: IV Forum Italianodi Scienze della Terra, Bellaria, Italy, 16–18Sept.

Vincent, E., and W. H. Berger (1985), Carbondioxide and polar cooling in the Miocene: TheMonterey hypothesis, in The Carbon Cycleand Atmospheric CO2: Natural Variations Ar-chean to Present, Geophys. Monogr. Ser., vol.32, edited by E. T. Sundquist and W. S.


16 of 17

PA4012

Broecker, pp. 455–468, AGU, Washington,D. C.

Wade, B. S., and H. Palike (2004), Ocean cli-mate dynamics, Paleoceanography, 19,PA4019, doi:10.1029/2004PA001042.

Westerhold, T., T. Bickert, and U. Rohl (2005),Middle to late Miocene oxygen isotopestratigraphy of ODP site 1085 (SE Atlantic):New constraints on Miocene climate varia-bility and sea-level fluctuations, Palaeo-geogr. Palaeoclimatol. Palaeoecol., 217,205–222.

Woodruff, F., and S. M. Savin (1991), Mid-Mio-cene isotope stratigraphy in the deep sea:High-resolution correlations, paleoclimatic cy-

cles, and sediment preservation, Paleoceano-graphy, 6, 755–806.

Wright, J. D., and K. G. Miller (1992), Miocenestable isotope stratigraphy, Site 747, Kergue-len Plateau, Proc. Ocean Drill. Program Sci.Results, 120, 855–866.

Zachos, J., N. J. Shackleton, J. S. Revenaugh, H.Palike, and B. P. Flower (2001), Climate re-sponse to orbital forcing across the Oligocene-Miocene boundary, Science, 292, 274–278.

��H. A. Abels, F. J. Hilgen, R. W. Kruk, and

W. J. Zachariasse, Department of Earth

Sciences, Utrecht University, Budapestlaan 4,3584 CD Utrecht, Netherlands. ([email protected]; [email protected]; [email protected]; [email protected])W. Krijgsman, Paleomagnetic Laboratory,

Fort Hoofddijk, Budapestlaan 17, 3584 CDUtrecht, Netherlands. ([email protected])I. Raffi, Dipartimento di Scienze della Terra,

Universita ‘‘G. D’Annunzio,’’ Campus Univer-sitario, via dei Vestini 31, 66013 Chieti Scalo,Italy. ([email protected])E. Turco, Dipartimento di Scienze della

Terra, Universita di Parma, Parco Area delleScienze 157/A, 43100 Parma, Italy. ([email protected])


17 of 17

PA4012

Figure 8. (left) Tuning of the red layers in the Tremiti section to precession (La04 1,1 [Laskar et al., 2004]), (middle)correlation of 10 bioevents between the Tremiti and the Ras il Pellegrin section (solid lines), and (right) tuning of the small-scale cycles in the Ras il Pellegrin section to precession (La04 1,1).

PA4012 ABELS ET AL.: ORBITAL CONTROL ON MID-MIOCENE GLOBAL COOLING PA4012

11 of 17

Figure

8

PA4012 ABELS ET AL.: ORBITAL CONTROL ON MID-MIOCENE GLOBAL COOLING PA4012

12 of 17

Documents

Long-period orbital control on middle Miocene global ... · isotope enrichment events in the late Paleogene and Neogene. The stepwise character of the middle Miocene cooling event