lable at ScienceDirect

Quaternary International 206 (2009) 52–61

Contents lists avai

Quaternary International

journal homepage: www.elsevier .com/locate/quaint

A high-resolution Holocene record on the Southern Brazilian shelf:Paleoenvironmental implications

Michel Michaelovitch de Mahiques a,*, Ilana Klein Coaracy Wainer a, Leticia Burone a, Renata Nagai a,Silvia Helena de Mello e Sousa a, Rubens Cesar Lopes Figueira a, Ilson Carlos Almeida da Silveira a,Marcia Caruso Bıcego a, Daniel Pavani Vicente Alves a, Øyvind Hammer b

a Institute of Oceanography, University of Sao Paulo, Praça do Oceanografico, 19105508-120 Sao Paulo, Brazilb Natural History Museum, University of Oslo, Oslo, Norway

a r t i c l e i n f o

Article history:Available online 1 November 2008

* Corresponding author. Tel.: þ55 11 30916609; faxE-mail address: mahiques@usp.br (M.M. de Mahiq

1040-6182/$ – see front matter � 2008 Elsevier Ltd adoi:10.1016/j.quaint.2008.09.010

a b s t r a c t

A high-resolution, multi-proxy record has been used to determine the environmental changes during theHolocene on the southern Brazilian shelf. Present oceanographic conditions reveal wind and freshwaterinput as the determinants of short-term productivity changes in the study area. Magnetic susceptibilityand grain-size variations, together with proxies of productivity (organic carbon, carbon accumulationrate, Ba, Sr, and Ca content, Ba/Al, Ba/Ti, and Al/Ti ratios) were analyzed and compared with proxies ofredox condition (V/Ti ratio), terrigenous input (Fe/Ca and Ti/Ca ratios), as well as other Element/Ti ratios,to evaluate the paleoceanographic and paleoclimatic changes over the period.The core covers a time interval of about 7650 years, with sedimentation rates varying from 0.025 to0.250 cm a�1, which represent time intervals of between 8 and 80 a per sample. There is a clear change inthe sedimentation rate at about 2800 B.P.All grain-size and elemental results indicate the occurrence of conspicuous changes between 5200 and5000 cal. B.P., as well as between 3000 and 2800 cal. B.P. A comparison of the results with the palyno-logical information available from the adjacent continental areas suggests that the sedimentary changesin this last interval may be correlated with the onset of modern climatic conditions in South America, andespecially, with the onset of the Plata Plume Water, a water mass that carries cold, less saline waterstowards the north. However, minor changes are observed at ca. 1500 B.P. and are correlated with anincrease in the atmospheric humidity. Furthermore, a time-series analysis undertaken using severalproxies indicated the occurrence of Sub-Milankovitch cycles, which may be compared with thosereported worldwide.

� 2008 Elsevier Ltd and INQUA. All rights reserved.

1. Introduction

The present oceanographic conditions offer the southern Bra-zilian shelf a privileged status for the study of short-term (seasonaland decadal) changes in the wind-driven currents and freshwaterdischarge regimes of the Southwestern Atlantic Ocean. In this area,the wind-dependent northward displacement of cold and lesssaline waters originating in the Rıo de La Plata and the discharge ofthe southern Brazilian lagoons (Piola et al., 2000; Piola and Romero,2004) control the seasonal variation in primary productivity (Ciottiet al., 1995). Sedimentological evidence of the northwarddisplacement of this water flow can be observed both in the organicand inorganic constituents of the bottom sediments and can be

: þ55 11 30916610.ues).

nd INQUA. All rights reserved.

traced as far north as the 24�S parallel (Mahiques et al., 1999, 2004).Also, this northward flow, which was originally attributed to theMalvinas Current, has been used to explain the occurrence of cold-water foraminiferal forms on the Southwestern Atlantic Shelf(Stevenson et al., 1998).

By assuming that the Rıo de La Plata river mouth has undergoneonly very slight changes in its geographical position over the last7500 years (Cavallotto et al., 2004; Violante and Parker, 2004) andthat the relative sea-level crossed its present height approximatelyat this time (Mahiques and Souza, 1999), and the later oscillationsin the Holocene sea-level were unable to expose significantportions of the shelf (Suguio and Martin, 1978; Angulo et al., 1999,2006), the middle and late Holocene paleoceanographic changes onthe Southwestern Atlantic shelf were mainly dependent on thewind regime and freshwater input, and that the millenial variationsin the sedimentary pattern may have been resulted from the vari-ations in this wind-driven water mass.

M.M. de Mahiques et al. / Quaternary International 206 (2009) 52–61 53

The aim of this work is to identify the environmental Holocenechanges on the Southwestern Atlantic shelf based on a high-reso-lution, multi-proxy analysis of a sediment core, sampled froma location in an area of high level of primary productivity, deter-mined by the displacement of the northward flow described earlier.Paleoproductivity, redox conditions, and terrigenous input proxies,together with grain-size and magnetic susceptibility variationswere considered to allow the analysis of the main temporal changesand variability, as well as to correlate with the known climaticchanges that affected the South American continent during theHolocene. This work is the first to utilize a multi-proxy approach inpaleoenvironmental studies on the Southwestern Atlantic shelf.

2. Study area

The northernmost part of the Southern Brazilian margin isknown as the Sao Paulo Bight, which is an arc-shaped featureextending from 23�S to 28�S (Zembruscki, 1979) (Fig. 1). The oceanfloor of the Sao Paulo Bight shows a rather complex morphologyinvolving channels, canyons, and considerable variations in theslope morphology (Furtado et al., 1996). The shelf break is located ata water depth of approximately 140 m with the upper-slopeshowing an average gradient of approximately 1�. On the innershelf, the sedimentation is mainly determined by the Plata PlumeWater flow (Moller et al., 2008), which carries sediments from theRıo de La Plata and, to a lesser extent, from the southern Braziliancoastal lagoons (Mahiques et al., 2004). On the middle and the

Fig. 1. Location of the core analyzed in this paper.

outer shelves as well as on the upper-slope, the sedimentaryprocesses are influenced by the flow of the Brazil Current (BC) alongthe western Atlantic continental margin (Mahiques et al., 2002,2004).

The distribution of surface sediments on the Southeastern Bra-zilian margin was extensively studied during the 1970s, and isdescribed in the studies of Rocha et al. (1975) and Kowsmann andCosta (1979). In general, the present sea-floor is covered by veryfine siliciclastic sands and silts, with variable amounts of clay andCaCO3. Coarser sediments and carbonate gravel and boulder facies,found on the outer shelf, represent <5% of the present bottomsediments, and are generally related to relict sediments, depositedunder lower sea-level conditions. More recently, Mahiques et al.(2002, 2004) re-evaluated the sedimentary characteristics of theSoutheastern Brazilian shelf and proposed hydrodynamic modelsfor the sediment deposition in that area.

The displacement of the La Plata Plume over the continentalshelf is highly dependent on both river discharge and wind regime.Piola and Romero (2004) reported a 1000 km displacement of theplume during the winter season, extending as far north as 27�S, anda retreat to 32�S during summer, by associating this variability withthe influence of the along-shore wind stress. Also, in a multi-yearstudy using satellite images, Gonsalez-Silvera et al. (2006) recog-nized a marked temporal variability of the plume, the most effec-tive northward penetration of the plume, occurring under theconditions of high river discharge and southerly winds. Under theEl Nino conditions, characterized by the blockage of the southerlywinds, the plume extended southward.

3. Methods

A 506 cm piston core was sampled at the coordinates26�59016.800S–048�04033.600W, at a water depth of 60 m, onboardthe R.V. ‘‘Prof. W. Besnard’’. Prior to its opening, the core wasanalyzed for magnetic susceptibility using a Bartington MS2Csensor. At the laboratory, the core was sub-sampled continuously atintervals of 2 cm, and the sub-samples were immediately frozen forfurther freeze-drying.

Owing to the lack of suitable carbonate materials (e.g., fora-minifers, molluscs), approximately 7 g of total sediment at each50 cm was separated for AMS radiocarbon dating at Beta Analytics.Calibrated ages were calculated using the Calib software, version5.0.2html, available at http://calib.qub.ac.uk/calib/, with the stan-dard marine correction of 408 yrs and a regional reservoir effect ofDR¼ 82.0� 46, corresponding to the average value of the threesamples to that area reported by Angulo et al. (2005), and theMarine04 Calibration Dataset (Hughen et al., 2004).

The establishment of a reliable model for depth–age relation-ships is a key question for all the studies involving sedimentaccumulation, and has been recently criticized by Telford et al.(2004). A basic assumption determined by the latter authors is thatthe age–depth models are only meaningful and useful whenestablished using calibrated radiocarbon ages. To evaluate theeffect of the age–depth models on the values of sedimentation ratesand, consequently, sediment accumulation, we tested severalinterpolation procedures (polynomial, linear interpolation betweenradiocarbon values, cubic spline, and mixed-effect). Polynomial,linear interpolation, and cubic spline models were calculated usingthe DepthAge software, developed by Louis Maher Jr., available atthe INQUA file boutique at http://www.geology.wisc.edu/~maher/inqua.html (last accessed December 15, 2006). The mixed-effectmodel was calculated using the Cagedepth and Cagenew functionsas described by Heegard et al. (2005), available at http://www.uib.no/bot/qeprg/Age-depth.htm (last accessed January 3, 2007).

As stated by Grant and Dickens (2002), the linear interpolationleads to artificial sedimentation rate values, and, in our case, there

Table 1Results of AMS radiocarbon datings in the organic matter of selected samples.Calibrated ages were calculated using a SW Atlantic reservoir effect (DR¼ 87,U¼ 46), using data from www.calib.org (accessed June 15th, 2006).

Depth Conventionalradiocarbonage (yr B.P.)

1s d13C(& PDB)

Calibratedradiocarbonage (yr B.P.)

1s

0–2 1410 40 �19.6 860 7050–52 1680 40 �19.8 1160 7098–100 2080 50 �19.8 1560 90148–150 2560 50 �20.3 2120 90198–200 2750 40 �20.9 2370 80248–250 3090 50 �20.6 2780 70298–300 3230 50 �20.2 2930 100352–354 3770 40 �19.6 3600 80398–400 4490 40 �19.9 4550 100448–450 5340 40 �20.3 5630 60504–506 7290 40 �20.4 7660 60

M.M. de Mahiques et al. / Quaternary International 206 (2009) 52–6154

was no significant difference among the fourth-order polynomial,cubic spline, and mixed-effect estimates, with a correlation coeffi-cient >0.999. Nevertheless, owing to the possibility of better errorestimation in this last model, the mixed-effect procedure for thedepth–age modelling was used. Grain-size analysis was performedin a Malvern Mastersizer 2000, on decarbonated samples.

Organic carbon was determined using an LECO CNS2000analyzer. Approximately 200 mg of dry sediment of each samplewas treated with 10% HCl to remove CaCO3, and then freeze-driedand analyzed. The LECO soil standards and blanks were analyzed ascontrols for each set of 30 samples. Organic carbon accumulationrate (orgC AR), in gp cm�2 k yr�1, was calculated in accordance withthe formula described by Grant and Dickens (2002):

ðorgC ARÞ ¼ ð%orgCÞ=100*SR*DBD*1000

where SR is the sedimentation rate (cm a�1) obtained from the agemodel and DBD is the dry bulk density (g cm�3), calculated from:

DBD ¼ ð1� z=100Þ*rs

where z is the wet porosity and rs is the density of the sedimentparticles, previously determined as 2.35 g cm�3.

The wet porosity (z) was calculated using the formula proposedby Clifton et al. (1995):

z ¼ ½ðW=100Þ*rs�=f½ðW=100Þ*rs� þ ½1� ðW=100Þ�*rwg

where W is the water content (in weight %), rs the density of thesediment particles, i.e., 2.35 g cm�3, and rw the density of the porewater, assumed to be 1.0 g cm�3.

Elemental analysis (Al, B, Ba, Ca, Cr, Fe, Mg, Mn, Sr, Ti and V) wasperformed using the ICP-OES technique. Approximately 1 g of drysediment was digested with 10 mL of 1:1 HNO3 at 95 �C for 15 min.After cooling, another 5 mL of concentrated HNO3 was added andthe solution was heated for 30 min. This second procedure wasrepeated until the digestion of the sample was completed. About2 mL of water and 3 mL of 30% H2O2 were added while heating,until the elimination of the organic matter was completed. Thesolution was then filtered through a Whatman 41 filter, and 10 mLof concentrated HCl was added to the digestate. Finally, the solutionwas filtered again in a Whatman 41 filter and the filtrate wascollected in a 100 mL volumetric flask. The volume was completedand the solution was analyzed in a Perkin Elmer Model 2100 DVICP-OES. Measurement precision for all the elements was betterthan 5%. Method accuracy was obtained by analyzing certifiedsediment (HR1) and was conformed by joining proficiencypromoted and accredited by RTC, NELAP (National EnvironmentalLaboratory Accreditation Programme).

Fig. 2. Age–Depth model, based on 11 AMS radiocarbon dating. Interpolation has beenobtained with the mixed-effect model as described in Heegard et al. (2005).

4. Results

4.1. Age mode and sedimentation rates

Table 1 presents the results of radiocarbon dating, and there wasno age inversion in the radiocarbon dates.

The core covered an age range of 7650 a with sedimentationrates varying from 0.025 to 0.250 cm a�1, which represent a timeinterval of between 8 and 80 a for each sample, the main changeshaving occurred at around 3000 B.P. (Figs. 2 and 3). The changes inthe sedimentation rates observed in this study seem dramatic forsuch a short time period. Nevertheless, this has already beenverified in other shelf and upper-slope areas. Knies (2005) observedchanges in the sedimentation rates by a factor of 12 between theHolocene and older layers in the shelf sediments, off northernNorway. The same behavior was observed for an equivalent periodon the shelf off northeastern Iceland (Andresen et al., 2005).

4.2. Grain-size

Almost all of the parameters show a pattern of three timeintervals, with transitions occurring between 5200 and 5000 B.P.,and 3000 and 2800 B.P. This pattern is reflected in the grain-size(Fig. 3). From the base to ca. 5200 B.P., sediments were essentiallycomposed of sandy silts, with clay content varying from 5% to 10%and sands ranging from 20% to 40%. A progressive increase in thesilt content, followed by diminishing sand (10–25%) values wasobserved from this date up to 3000 yrs B.P. Finally, after this age,the grain-size remained approximately constant, as observed bythe mean diameter and silt content, the latter corresponding to>80% of the grain-size distribution.

4.3. Terrigenous input and redox proxies and other element/Tiratios (Fig. 3)

Terrigenous input was evaluated through the use of magneticsusceptibility values and Fe/Ca and Ti/Ca ratios (Arz et al., 1998). Anexpected significant correlation (R¼ 0.82) between the Fe/Ca andTi/Ca ratios was observed, with a much smaller but significant

Fig. 3. Along-core variation in sedimentation rates, grain-size, terrigenous input and redox proxies and other Element/Ti ratios.

M.M. de Mahiques et al. / Quaternary International 206 (2009) 52–61 55

correlation between magnetic susceptibility and those elementalratios. As a rule, terrigenous input, as defined by the Fe/Ca and Ti/Caratios, follows a pattern that is similar and significantly correlated(R> 0.40) with that observed in the clay content, with the occur-rence of three phases in the time interval considered. Higherterrigenous input appears to have occurred during the LateHolocene.

The content of V and, specifically, the V/Ti ratio has been used asa proxy for the redox condition (Moreno et al., 2004). Together withCr/Ti, B/Ti, Mg/Ti, and Mn/Ti ratios, they all exhibit the same three-phase division stated earlier, with a highly significant correlationamong these ratios and the sedimentation rate.

4.4. Productivity proxies (Figs. 4 and 5)

Barium is one of the most widely used proxies for estimatingpaleoproductivity (Dymond et al., 1992; Paytan et al., 1996).Nevertheless, the number of its adherents (Jacot Des Combes et al.,1999, 2005; Calvert and Fontugne, 2001; Pfeifer et al., 2001; Pra-kash Babu et al., 2002; Wei et al., 2003) is as large as those of itsdetractors (Averyt and Paytan, 2004; Anderson and Winckler,2005; Mora and Martinez, 2005) for its use as a reliable indicator ofvariations in productivity over time. The problems reported arerelated to the impossibility of calculating Baxs values from the Ba/Alratios (Kasten et al., 2001; Moreno et al., 2004), owing to thereactivity of Al, its use as a ‘‘normalizing’’ element, the absence ofregional Ba/Al background values, or even inconsistencies in thealgorithms used for the productivity estimates.

Averyt and Paytan (2004) indicated the discrepancies in theresults obtained from the different parameters related to the Bacontent, as well as the Al accumulation rate and Al/Ti ratio. Ina more general analysis, Anderson and Winckler (2005) criticizedthe utilization of Ba/Ti and Al/Ti ratios, which may be influenced bythe spatial and temporal variability in the Ti flux. According tothem, the dissolution of CaCO3 during the Holocene might beresponsible for the increase in the concentrations of barite as wellas excess of Al, which might raise difficulties in the utilization ofthese proxies in paleoproductivity estimates. Pattan et al. (2003),on studying the glacial–interglacial variations in the southeastern

Arabian Sea, indicated that the Al/Ti ratio cannot be used as a reli-able paleoproductivity proxy.

Furthermore, with respect to the data, the following questionsmust be addressed:

1) Can Al be used as a reliable normalizer element? Murray andLeinen (1996) analyzed the samples from the Equatorial Pacific andreported the anomalous values in the Al/Ti ratio, related to the non-terrigenous character of Al. This aspect was evaluated using the Alvs. Ti plot, assuming that this last element is exclusively terrigenousand non-reactive (Figs. 4a and 5). The correlation is considered asstatistically significant (0.797). Nevertheless, when the Ba contentas well as Ba/Al, Ba/Ti, and Ba/Ca ratios is compared with the orgCand carbon accumulation rate (Fig. 4b–i), the Ba/Al plots exhibitedthe smallest values of correlation, despite being statisticallysignificant. Further, it is worth noting that several plots, such as Bavs. orgC, showed a dispersion that was not explicable by a linearequation, but probably only by a polynomial or exponential equa-tion. In fact, there exists a clear temporal dependence in the rela-tionship between Ba and Ba-related parameters with orgC andcarbon accumulation rate.

2) Can Baxs be calculated? None of the Ba/Al values were>0.0075, which may be considered as a terrigenous backgroundvalue (Taylor and McLennan, 1985), or even >0.0028, determinedby Klump et al. (2000) in the Chilean continental margin surfacesediments. Hence, owing to the lack of regional background valuesfor Ba/Al, the determination of Baxs and the calculation of paleo-productivity are somewhat deficient. Nevertheless, it is possible, atleast, to compare the data qualitatively to establish the temporalchanges in productivity.

With respect to this data, there are statistically significantcorrelations (R> 0.50) among organic carbon and Ba/Ca, Ba/Ti, andTi/Al. Thus, these parameters can be used as reliable proxies forproductivity.

As observed in almost all of the other parameters, the paleo-productivity proxies exhibited a temporal variation characterizedby two main changes at the intervals of 5200–5000 and3000–2800 B.P. From the base of the core to 5200–5000 B.P. orgCand carbon accumulation rates, as well as Ba, Al, and Ca exhibitedtheir lowest values. The interval between 5000 and 3000 BP is

100 200 300 400 500

Ti (mg.kg-1

)

5000

10000

15000

20000

25000

30000

Al (m

g.kg

-1)

0.40 0.60 0.80 1.00 1.20 1.40

orgC(%)

8

12

16

20

24

28

Ba

(m

g.k

g-1)

0 0.002 0.004 0.006

Carbon Accumulation Rate

(g.cm2.yr

-1)

8

12

16

20

24

28

Ba

(m

g.k

g-1)

0.4 0.6 0.8 1.0 1.2 1.4

orgC(%)

0.0008

0.0012

0.0016

0.0020

0.0024

Ba/A

l

0 0.002 0.004 0.006Carbon Accumulation Rate

(g.cm2.yr

-1)

0.0008

0.0012

0.0016

0.0020

0.0024

Ba/A

l

A BC

DE

Legend

Top to 3,000 yrs B.P.

3,000 to 5,000 yrs B.P.

5,000 yrs B.P. to bottom

FG

HI

orgC(%)

0.0200

0.0400

0.0600

0.0800

0.1000

0.1200

0.1400

Ba

/T

i

0.4 0.6 0.8 1.0 1.2 1.4 0 0.002 0.004 0.006

Carbon Accumulation Rate

(g.cm2.yr

-1)

0.0200

0.0400

0.0600

0.0800

0.1000

0.1200

0.1400

Ba

/T

i

0.4 0.6 0.8 1.0 1.2 1.4

orgC(%)

0.0004

0.0008

0.0012

0.0016

0.0020

0.0024

Ba

/C

a

0 0.002 0.004 0.006

Carbon Accumulation Rate

(g.cm2.yr

-1)

0.0004

0.0008

0.0012

0.0016

0.0020

0.0024

Ba

/C

a

Fig. 4. Dispersion plots of A) Al vs. Ti; B) Ba vs. orgC; C) Ba vs. carbon accumulation rate; D) Ba/Al vs. orgC; E) Ba/Al vs. carbon accumulation rate; F) Ba/Ti vs. orgC; G) Ba/Ti vs. carbonaccumulation rate; H) Ba/Ca vs. orgC and I) Ba/Ca vs. carbon accumulation rate. Statistically significant values (p< 0.05) of Pearson Correlation Coefficient are pointed.

M.M. de Mahiques et al. / Quaternary International 206 (2009) 52–6156

OrgC

(%)

TotN

(%)

C/N

(wt)

C.A.R.

(g.cm.-2

yr-1

)

Al

(mg.kg-1

)

Ba

(mg.kg-1

)

Ca

(mg.kg-1

)

Sr

(mg.kg-1

)

Ba/Al Ba/Ca Ba/Ti Ti/Al

Cal. A

ge (yr B

.P

.)

0

1000

2000

3000

4000

5000

6000

7000

8000

0.3

1.5

0.04

0.18

6 14 0.00

0

5000

5 30 5000

30 100

0.00

05

0.00

05

0.03

0.15

0.01

0.03

0.01

0

1000

0

2500

0

0.00

25

0.00

25

Fig. 5. Along-core variations in organic Carbon (orgC), total Nitrogen (totN), C/N ratio, carbon accumulation rate (C.A.R.), Al, Ba, Ca and Sr contents, and Ba/Al, Ba/Ca, Ba/Ti and Al/Tiratios.

M.M. de Mahiques et al. / Quaternary International 206 (2009) 52–61 57

characterized by low but increasing values in orgC and carbonaccumulation rate. The Ba/Ti and Ba/Ca ratios had their lowestvalues in this interval, indicating a correspondingly lower produc-tivity period during the Holocene in that area. There was a signifi-cant increase in the productivity after 3000 B.P., as shown by thebehavior of orgC, carbon accumulation rate, Ba, Al, and other Ba/Alratios. This increase coincides with the increase in the terrigenousinput and sedimentation rate in that area.

5. Discussion

5.1. Sea-level changes

The Middle and Late Holocene sea-level change curves of theBrazilian coast are still a matter of debate. Originally, they havebeen defined on the basis of >700 radiocarbon dates (Suguio et al.,1985; Angulo and Suguio, 1995). Angulo and Lessa (1997) andAngulo et al. (1999, 2006) have criticized the former curves andproposed a new Holocene sea-level change model for the Braziliancoast, based exclusively on vermetide tubes, which has since beencorroborated by the studies of Ybert et al. (2003) and Caldas et al.(2006). The main controversy relating to these curves is associatedwith the magnitude of the sea-level oscillations during this period,where the first curve model considered two or three periods of sea-level lower than the present one after 5100 BP, and the secondmodel assuming the present sea-level as the lowest of the last 7000years. Martin et al. (2003) produced the first reservoir correctedand astronomically calibrated sea-level change curve for a sector ofthe Brazilian coast; their results indicate the occurrence of threemain events of submergence of the coast (7800–5600, 3700–3500,and 2300–2100 cal BP), intercalated among the periods of emer-gence. On the other hand, Angulo et al. (2006) summarized theinformation about a large data set in which there were evidences ofa progressive decline of the sea-level, since the mid-Holocenemaximum.

Independent of the model adopted, the mid-Holocenemaximum in the study area would represent an increase of about3 m above the present sea-level. This represents a less than theaverage wave-height in the Southern Brazilian shelf and does not

indicate a significant change in the coastline of the area, apart fromthe development of some lagoonal systems. In this sense, no sedi-mentological or geochemical changes were recognized which couldbe attributed to the mid-Holocene sea-level maximum or any otherpossible sea-level oscillations.

5.2. Climate changes

The Holocene climate history of South America is still not yetfully understood, although the record is improving rapidly. Ingeneral, it would seem that the climate of the Holocene has beenknown to transit between dry and wet conditions. A discussionabout the climatic changes in South America from the Last GlacialMaximum by Adams and Faure (1997), and a good review of SouthAmerica in the early to mid-Holocene by Araujo et al. (2005) exactlypoint out these transitions.

As a rule, important environmental variations along the corewere noted not only in the productivity but also in the terrigenousinput and shelf dynamics. Some events were very well marked –the break in the depositional process prior to and after the transi-tion of 3000–2800 B.P. being the most conspicuous. The causes ofthe oscillations observed along the core can be analyzed from thepoint of view of the climate variability affecting South America, themain factor of which is the wind regime (Ledru et al., 1996, 1998;Behling, 1997, 1998; Gaiero et al., 2004; Gilli et al., 2005).

In addition, the causes of the oscillations observed along thecore can also be analyzed from two points of view of the climaticvariability affecting South America, associated with changes in thewinds: these could change owing to the tropical controls, such asshifts in the intertropical convergence zone (ITCZ) or subtropicalimpacts, such as changes in winds associated with the passage offronts, or even strengthening/weakening of the subtropical West-erlies, possibly associated with different phases of the annularmode. Changes in the wind and their sources account for theassociated environmental changes in moisture.

Considering subtropical Brazil and Argentina, there are contra-dictory signs of the timing of environmental changes. Ledru (1992),based on the terrestrial proxies showed that in southcentral Brazil,conditions appeared to have been somewhat cooler than the

0.3

0.2

0.1

0.0

-0.1

-0.2

-0.3

-0.4

Y

0 1000 2000 3000 4000 5000 6000 7000 8000Yr B.P.

0.000 0.001 0.002 0.003 0.004Frequency (1/yr)

0

10

20

30

Po

wer

0.010.05

6,70

0 2,68

01,

848

1,27

61,

030

A

B

Fig. 6. A) Lomb periodogram of the detrended values of Axis-1 Factor Score (83% of thetotal variance) obtained by a Principal Component Analysis, using elemental values asdata source; B) Sinusoidal fit on the same data, using the 2680, 1848, 1276 and 1030-year cycles.

M.M. de Mahiques et al. / Quaternary International 206 (2009) 52–6158

present. Drier conditions (signifying lesser forests) than the lateHolocene may have existed in the subtropical Brazil from about8000 to 3000 cal B.P., indicated by the charcoal phases followed bysavanna incursion (Vernet et al., 1994). Other evidence of more aridconditions than the present one around 8000 cal B.P. was froma pollen record at Lago do Pires; 17�570S, 42�130W (Behling, 1995).According to Islebe and Hooghiemstra (1997), there has beenevidence of a relatively dry phase presumed to have occurred inPanama, Guatemala, and Costa Rica between about 7000 and5000 cal B.P. Van Geel et al. (1996) summarized the evidence forboth the Colombian Andes (lowering of altitudinal forest limits)and subtropical Chile showing cooler, moister conditions around2600 cal yrs B.P. Araujo et al. (2005) summarized the climate vari-ations in the early to mid-Holocene in the central–southeasternBrazil. According to them, some studies (e.g., Barberi et al., 2000;Ferraz-Vicentini and Salgado-Labouriau, 1996; Salgado-Labouriauet al., 1997, 1998) showed that the Last Glacial Maximum (LGM) thatwas considerably dry, was followed by an increase in moisture,beginning around 6850–5710 cal B.P., which eventually led to themodern climatic conditions. Other studies in the same regionpointed to a different scenario, where the LGM was consideredto be followed by a less dry period, however, around9490–4460 cal B.P., the conditions were presumed to have returnedto dry period (e.g., Behling, 2002; Ledru et al., 1996, among others).

It is also generally accepted by the cited studies, that since ca.5400 cal B.P., the wind changes are associated with a gradual trendtowards drier conditions over northern South America. Haug et al.(2001), through analysis of the Cariaco Basin record, demonstratedthat this trend towards drier conditions is followed by precipitationminima during the time interval 3800–2800 years ago. Theydescribed that these regional precipitation regime shifts are closelytied to the changes in the mean latitudinal position of the InterTropical Convergence Zone (ITCZ).

Ledru et al. (1996), in a description of the climatic variability ofthe last 55,000 years in Southeast Brazil, described the occurrenceof a drier and warmer period, between 7000 and 4000 B.P., fol-lowed by a subsequent increase in the moisture, as a consequenceof oscillations in the position of the ITCZ. In a more detailed paper,Ledru et al. (1998) describe a persistent increase in moisture, from7000 yrs B.P. up to the present, by virtue of changes in the windregime, dominated by the advance of polar masses. The role playedby the wind was also stressed by Behling (1997, 1998), whodescribed the northward advance of the Araucaria forests, from2850 B.P. until present, as the result of intensification of the windregime, increase in moisture, and decrease in atmospherictemperature.

Thus, in view of the modern controls of the productivity on theshelf, these results can be used to infer paleoclimatic changesassociated with the wind regime and continental input. As a rule,the period of lower productivity corresponds to the intervalbetween 5200 and 3000 years, which may be compared with thelower humidity values on the continent as well as weaker windsover the South American continent. The increase in the moistureand the intensification of the southerly winds, related by theauthors mentioned earlier, led to an environmental change in theshelf regime, with a consequent increase in the sedimentationrates, productivity, and terrigenous input. Oscillations occurredafter 3000 years, shown by the decrease in several parameters at ca.1500 B.P., which also agree with the paleoclimatic inferences setout by Behling (1997).

The relationship between the sedimentary data and the mainclimatic cyclicities reported worldwide was analyzed by per-forming a Time-Series Analysis. The occurrence of cycles in theHolocene climate and ocean circulation has been reported byBond et al. (1997, 2001), who identified a 1470� 500-year peri-odicity in the North Atlantic surface winds and hydrography. This

cyclicity, which has been attributed to solar activity and markedthrough the Holocene (Campbell et al., 1998; Bianchi and McCave,1999; Viau et al., 2002) has also been lately identified in otherparts of the world, thus giving it a global character (Russell andJohnson, 2005).

More recently, this single-dominant period cyclicity was ques-tioned by Clemens (2005), who indicated the occurrence of twostronger spectral peaks, at 1667 and 1190 years, and suggested thatmillenial variability can be determined by nonlinear interactions(heterodyning) between centennial periodicities owing to solaractivity.

Thus, to evaluate the occurrence of cycles in the sedimentationof the study area, a Time-Series Analysis (Lomb periodogram) wasundertaken with the aid of the software PAST (Hammer et al.,2001). Prior to analysis, the data were detrended to remove theeffect of very low frequencies (e.g., periods longer than the time

Table 2Significant frequencies obtained in the Lomb Periodogram, for each parameter analyzed. Highest peaks of each parameter are highlighted.

Al Ba B Ca Cr Fe Mg Mn Sr Ti V Sus C S Median Sed Rate Axis-1 PCA scores

6700 6700 6700 6700 6700 6700 6700 66705960

5360 53604180 3880 4520

2550 2680 2820 2550 2680 2550 2680 2550 3150 2550 2820 2360 26801790 1850 1850 1850 1910 1850 1790 1790 1600 18501280 1280 1250 1310 1280 1220 1250 12801050 1030 1030 870 1030 1030 880 1010 1030 1030

M.M. de Mahiques et al. / Quaternary International 206 (2009) 52–61 59

span) on the results. Also, to reduce noise, a Principal ComponentAnalysis (PCA) was performed with the standardized values,including all the elements, and the Axis-1 (83% of the total vari-ance) scores were submitted to spectral analysis. Axis-1 corre-sponds mainly to the input of Fe and Al, which may assume theterrigenous input as the most important factor in the variability ofthe system.

An example of the results is shown in Fig. 6, and the significantfrequencies (p� 0.05) for each parameter are presented in Table 2with the highest peak being highlighted. The spectrograms formost of the elements as well as the PCA scores showed five mainperiodicities:

a) From 3880 to 6700 BP – accounting for the low-frequencyoscillations. Considering the time interval these periods are toolong to give reliable information concerning periodicity;

b) From 2360 to 2820 BP (centered at 2680 BP, according to thePCA analysis) – accounting the highest power in Cr, Fe,magnetic susceptibility, sedimentation rate, and PCA Axis-1score. This cycle may be compared with the 2750 a cycleobserved both in the GISP-2 ice core and Santa Barbara Basinsediment core (Nederbragt and Thurow, 2005);

c) From 1600 to 1910 BP (centered at 1848 BP) – as identified inMg. This cycle is found in most of the elements, but we cannotassume that this cycle corresponds to the 1470� 500 a cycle ofBond et al. (1997, 2001). On the other hand, Clemens (2005)observed a conspicuous 1667 a cycle in the records of HuluCave and GRIP;

d) From 1220 to 1310 BP (centered at 1276 BP) – accounting themost powerful cycle in Al distribution. This cannot be discardedas the second harmonic for the 2360–2820 a cycle. Furthermore,it can also be compared with the 1160–1190 a cycle reported byClemens (2005), for the GRIP and Hulu Cave d18O records;

e) From 880 to 1030 BP (centered at 1030 yrs BP) – as observed inSr, Ti, and V. This cycle can be compared with the 1000 acyclicity reported for thermohaline circulation (Chapman andShackleton, 2000), GISP-2 d18O and atmospheric D14C (Neder-bragt and Thurow, 2005), North American temperatures (Viauet al., 2006), etc. This cycle is still maintained based on severalparameters (Axis-1 of the PCA, B, Sr, Mn, and Ti) when theperiodogram is executed for the most recent interval, i.e.,coretop to 3000 BP.

These results indicate that the parameters related with terrig-enous input (Fe, Ti and magnetic susceptibility) and Eh (Cr, B, Mg, Vand Mn) display cycles in their distribution. Both may be associateddirectly with the rainfall, and thus with the wind regime, whichcontrols the amount of sediment transported to the shelf.

6. Conclusions

A high-resolution, Holocene multi-proxy record obtained for thefirst time on the Southwestern Atlantic shelf allowed us to

recognize the main changes in the wind and La Plata riverdischarge. Two transitions, at 5200–5000 and 3000–2800 BPmarked the boundaries of three time intervals with differentsedimentary characteristics, related to environmental changes.

The first phase corresponds to the interval prior to 5200 BP,which is marked by the lowest sedimentation rate values as well asintermediate values of the terrigenous input and productivity,when compared with the rest of the phases. It corresponds toa phase of increasing sea-level, which lasted until 5600 BP.

The interval between 5200 and 3000 BP marks the period oflowest terrigenous input and productivity, which may be associatedwith the weakening of southerly winds and a decrease in humidityin South America, as confirmed previously in the palynologicalrecord for the continent. It corresponds to a phase of decreasingsea-level, from the mid-Holocene maximum until altitudes about1 m above the present sea-level.

The period after 3000 BP marks the highest sedimentation rates,terrigenous input and productivity, corresponding to the estab-lishment of the modern conditions of wind and river discharge. Atleast one interval of the weakening of the climatic conditions couldbe observed in this period, which corresponds to the period ofstabilization of the sea-level.

The Time-Series Analysis allowed recognition of four significantcycles, which may be compared with the cyclicity described byseveral other researchers, mainly in the northern hemisphere. Thesignificant cycles centered at 2680, 1848, 1276, and 1030 BP foundin the spectrograms of several parameters were comparable withthose of GISP-2, GRIP, Hulu Cave, Santa Barbara basin, D14C, andNorth American temperatures.

Despite their significant correlation with orgC and carbon accu-mulation rates, the utilization of Ba and Ba-related ratios as proxiesfor productivity has only been used qualitatively in this study.Regional studies are required for a better understanding of thebackground terrigenous values of Ba and, consequently, Baxs.

Acknowledgments

The authors wish to thank Mr. Marcelo Rodrigues, SamaraCazzoli Y Goya, Clodoaldo Vieira Tolentino and Edilson Faria fortheir help in the sediment sampling and laboratory analyses.Thanks are due also to the crew of the R.V. ‘‘Prof. W. Besnard’’.

This paper is a contribution to the IGCP526 Project (Risks,Resources, Record of the Past on the Continental Shelves). Financialsupport was provided by Fapesp (process 03/10740-0) and CNPq(300381/2004-2).

References

Adams, J.M., Faure, H., 1997. Preliminary vegetation maps of the world since the lastglacial maximum: an aid to archaeological understanding. Journal of Archae-ological Science 24, 623–647.

Anderson, R.F., Winckler, G., 2005. Problems with paleoproductivity proxies. Pale-oceanography 20, PA3012. doi:10.1029/2004PA001107.

M.M. de Mahiques et al. / Quaternary International 206 (2009) 52–6160

Andresen, C.S., Bond, G., Kuijpers, A., Knutz, P.C., Bjork, S., 2005. Holocene climatevariability at multidecadal time scales detected by sedimentological indicatorsin a shelf core NW off Iceland. Marine Geology 214, 323–338.

Angulo, R.J., Giannini, P.C.F., Suguio, K., Pessenda, L.C.R., 1999. Relative sea-levelchanges in the last 5500 years in southern Brazil (Laguna-Imbituba region,Santa Catarina State) based on vermetid 14C ages. Marine Geology 159,323–339.

Angulo, R.J., Lessa, G.C., 1997. The Brazilian sea-level curves: a critical review withemphasis on the curves from the Paranagua and Cananeia regions. MarineGeology 140, 141–166.

Angulo, R.J., Lessa, G.C., Souza, M.C., 2006. A critical review of mid- to late-Holocenesea-level fluctuations on the eastern Brazilian coastline. Quaternary ScienceReviews 25, 486–506.

Angulo, R.J., Souza, M.C., Reimer, P.J., Sasaoka, S.K., 2005. Reservoir effect of thesouthern and southeastern Brazilian coast. Radiocarbon 47, 67–73.

Angulo, R.J., Suguio, K., 1995. Re-evaluation of the Holocene sea-level maxima forthe State of Parana, Brazil. Palaeogeography, Palaeoclimatology, Palaeoecology113, 385–393.

Araujo, A.G.M., Neves, W.A., Pilo, L.B., Atui, J.P.V., 2005. Holocene dryness andhuman occupation in Brazil during the ‘‘Archaic Gap’’. Quaternary Research 64,298–307.

Arz, H.W., Patzold, J., Wefer, G., 1998. Correlated millennial-scale changes in surfacehydrography and terrigenous sediment yield inferred from last-glacial marinedeposits off northeastern Brazil. Quaternary Research 50, 157–166.

Averyt, K.B., Paytan, A., 2004. A comparison of multiple proxies for exportproduction in the equatorial Pacific. Paleoceanography 19, PA4003. doi:10.1029/2004PA001005.

Barberi, M., Salgado-Labouriau, M.L., Suguio, K., 2000. Paleovegetation and paleo-climate of ‘‘Vereda de Aguas Emendadas’’, central Brazil. Journal of SouthAmerican Earth Sciences 13, 241–254.

Behling, H., 1995. A high-resolution Holocene pollen record from Lago do Pires, SEBrazil – vegetation, climate and fire history. Journal of Paleolimnology 14,253–268.

Behling, H., 1997. Late Quaternary vegetation, climate and fire history of theAraucaria forest and campos region from Serra Campos Gerais, Parana State(South Brazil). Review of Palaeobotany and Palynology 97, 109–121.

Behling, H., 1998. Late Quaternary vegetational and climatic changes in Brazil.Review of Palaeobotany and Palynology 99, 143–156.

Behling, H., 2002. South and southeast Brazilian grasslands during Late Quaternarytimes: a synthesis. Paleogeography, Palaeoclimatology, Palaeoecology 177,19–27.

Bianchi, G.G., McCave, I.N., 1999. Holocene periodicity in North Atlantic climate anddeep-ocean flow south of Iceland. Science 397, 515–517.

Bond, G., Kromer, B., Beer, J., Muscheler, R., Evans, M.N., Showers, W., Hoffmann, S.,Lotti-Bond, R., Hajdas, I., Bonani, G., 2001. Persistent solar influence on NorthAtlantic climate during the Holocene. Science 294, 2130–2136.

Bond, G., Showers, W., Cheseby, M., Lotti, R., Almasi, P., Demenocal, P., Priore, P.,Cullen, H., Hajdas, I., Bonani, G., 1997. A pervasive millennial-scale cycle in NorthAtlantic Holocene and glacial climates. Science 278, 1257–1266.

Caldas, L.H.O., Oliveira, J.G., Medeiros, W.E., Stattegger, K., Vital, H., 2006. Geometryand evolution of Holocene transgressive and regressive barriers on the semi-arid coast of NE Brazil. Geo-Marine Letters 26, 249–263.

Calvert, S.E., Fontugne, M.R., 2001. On the Late Pleistocene–Holocene sapropelrecord of climatic and oceanographic variability in the eastern Mediterranean.Paleoceanography 16, 78–94.

Campbell, I.D., Campbell, C., Apps, M.J., Rutter, N.W., Bush, A.B.G., 1998. Late Holo-cene 1500 a climatic periodicities and their implications. Geology 26, 471–473.

Cavallotto, J.L., Violante, R.A., Parker, G., 2004. Sea-level fluctuations during the last8600 years in the de la Plata river (Argentina). Quaternary International 114,155–165.

Chapman, M.R., Shackleton, N.J., 2000. Evidence of 550-year and 1000-yearcyclicities in North Atlantic circulation patterns during the Holocene. Holocene10, 287–291.

Ciotti, A.M., Odebrecht, C., Fillmann, G., Moller Jr., O.O., 1995. Freshwater outflowand Subtropical convergence influence on phytoplankton biomass on thesouthern Brazilian continental shelf. Continental Shelf Research 15, 1737–1756.

Clemens, S.C., 2005. Millennial-band climate spectrum resolved and linked tocentennial-scale solar cycles. Quaternary Science Reviews 24, 521–531.

Clifton, R.J., Watson, P.G., Davey, J.T., Frickers, P.E., 1995. A study of processesaffecting the uptake of contaminants by intertidal sediments, using the radio-active tracers: 7Be, 137Cs and unsupported 210Pb. Estuarine Coastal and ShelfScience 41, 459–474.

Dymond, J.R., Suess, F., Lyle, M., 1992. Barium in deep-sea sediment: a geochemicalproxy for paleoproductivity. Paleoceanography 7, 163–181.

Ferraz-Vicentini, K.R., Salgado-Labouriau, M.L., 1996. Palynological analysis ofa palm swamp in Central Brazil. Journal of South American Earth Sciences 9,207–219.

Furtado, V.V., Bonetti Filho, J., Conti, L.A., 1996. Paleo river morphology and sea levelchanges at southeastern Brazilian continental shelf. Anais da Academia Brasi-leira de Ciencias 68, 163–169.

Gaiero, D., Depetris, P.J., Probst, J.L., Bidart, S.M., Leleyter, L., 2004. The signature ofriver- and wind-borne materials exported from Patagonia to the southernlatitudes: a view from REEs and implications for paleoclimatic interpretations.Earth and Planetary Science Letters 219, 357–376.

Gilli, A., Ariztegui, D., Anselmetti, F.S., Mckenzie, J.A., Markgraf, V., Hajdas, I.,McCulloch, R.D., 2005. Mid-Holocene strengthening of the Southern Westerlies

in South America d sedimentological evidences from Lago Cardiel, Argentina(49�S). Global and Planetary Change 49, 75–93.

Gonsalez-Silvera, A., Santamaria-Del-Angel, E., Millan-Nunez, R., 2006. Spatial andtemporal variability of the Brazil–Malvinas confluence and the La Plata Plumeas seen by SeaWiFS and AVHRR imagery. Journal of Geophysical Research 111,C06010, doi:10.1029/2004JC002745.

Grant, K.M., Dickens, G.R., 2002. Coupled productivity and carbon isotope records inthe southwest Pacific Ocean during the late Miocene – early Pliocene biogenicbloom. Palaeogeography, Palaeoclimatology, Palaeoecology 187, 61–82.

Hammer, Ø., Harper, D.A.T., Ryan, P.D., 2001. PAST: paleontological statistics soft-ware package for education and data analysis. Palaeontologia Electronica 4 (1),9. <http://palaeo-electronica.org/2001_1/past/issue1_01.htm>.

Haug, G., Hughen, K.A., Sigman, D., Peterson, L., Rohl, U., 2001. Southward migrationof the Intertropical Convergence Zone through the Holocene. Science 293,1304–1308 (2001).

Heegard, E., Birks, H.J.B., Telford, R.J., 2005. Relationships between calibrated agesand depth in stratigraphical sequences: an estimation procedure by mixed-effect regression. Holocene 15, 612–618.

Hughen, K.A., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C.J.H.,Blackwell, P.G., Buck, C.E., Burr, G.S., Cutler, K.B., Damon, P.E., Edwards, R.L.,Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Kromer, B., McCormac, F.G.,Manning, S.W., Bronk Ramsey, C., Reimer, P.J., Reimer, R.W., Remmele, S.,Southon, J.R., Stuiver, M., Talamo, S., Taylor, F.W., Van Der Plicht, J.,Weyhenmeyer, C.E., 2004. Marine04 marine radiocarbon age calibration,26–0 ka BP. Radiocarbon 46, 1059–1086.

Islebe, G.A., Hooghiemstra, H., 1997. Vegetation and climate history of montaneCosta Rica since the last glacial. Quaternary Science Reviews 16, 589–604.

Jacot Des Combes, H., Caulet, J.P., Tribovillard, N.P., 1999. Pelagic productivitychanges in the equatorial area of the northwest Indian Ocean during the last400,000 years. Marine Geology 158, 27–55.

Jacot Des Combes, H., Caulet, J.P., Tribovillard, N.P., 2005. Monitoring the variationsof the Socotra upwelling system during the last 250 kyr: a biogenic andgeochemical approach. Palaeogeography, Palaeoclimatology, Palaeoecology223, 243–259.

Kasten, S., Haese, R.R., Zabel, M., Ruhlemann, C., Schulz, H.D., 2001. Barium peaks atglacial terminations in sediments of the equatorial Atlantic Oceandrelicts ofdeglacial productivity pulses? Chemical Geology 175, 635–651.

Klump, J., Hebbein, D., Wefer, G., 2000. The impact of sediment provenance onbarium-based productivity estimates. Marine Geology 169, 259–271.

Knies, J., 2005. Climate-induced changes in sedimentary regimes for organic mattersupply on the continental shelf off northern Norway. Geochimica et Cosmo-chimica Acta 69, 4631–4647.

Kowsmann, R.O., Costa, M.O.A., 1979. Sedimentaçao quaternaria da margem conti-nental brasileira e das areas oceanicas adjacentes. In: REMAC PROJECT (FinalReport). Petrobras, Rio de Janeiro, pp. 1–55.

Ledru, M.P., 1992. Late Quaternary environmental and climatic changes in CentralBrazil. Quaternary Research 39, 90–98.

Ledru, M.P., Braga, P.I.S., Soubies, F., Fournier, M., Martin, L., Suguio, K., Turcq, B.,1996. The last 50,000 years in the Neotropics (Southern Brazil): evolution ofvegetation and climate. Palaeogeography, Palaeoclimatology, Palaeoecology123, 239–257.

Ledru, M.P., Salgado-Labouriau, M.L., Lorscheitter, M.L., 1998. Vegetation dynamicsin southern and central Brazil during the last 10,000 a B.P. Review of Palae-obotany and Palynology 99, 131–142.

Mahiques, M.M., Mishima, Y., Rodrigues, M., 1999. Characteristics of the sedimen-tary organic matter on the inner and middle continental shelf between Gua-nabara Bay and Sao Francisco do Sul, south-eastern Brazilian margin.Continental Shelf Research 19, 775–798.

Mahiques, M.M., Silveira, I.C.A., Sousa, S.H.M., Rodrigues, M., 2002. Post-LGMsedimentation on the outer shelf – upper slope of the northernmost part of theSao Paulo Bight, southeastern Brazil. Marine Geology 181, 387–400.

Mahiques, M.M., Souza, L.A.P., 1999. Shallow seismic reflectors and upper Quater-nary sea level changes in the Ubatuba region, Sao Paulo State, SoutheasternBrazil. Revista Brasileira de Oceanografia 47, 1–10.

Mahiques, M.M., Tessler, M.G., Ciotti, A.M., Silveira, I.C.A., Sousa, S.H.M.,Figueira, R.C.L., Tassinari, C.C.G., Furtado, V.V., Passos, R.F., 2004. Hydrody-namically driven patterns of recent sedimentation in the shelf and upper slopeoff Southeast Brazil. Continental Shelf Research 24, 1685–1697.

Martin, L., Dominguez, J.M.L., Bittencourt, A.C.S.P., 2003. Fluctuating Holocene sealevels in eastern and southeastern Brazil: evidence from multiple fossil andgeometric indicators. Journal of Coastal Research 19, 101–124.

Moller Jr., O.O., Piola, A.R., Freitas, A.C., Campos, E.J.D., 2008. The effects of riverdischarge and seasonal winds on the shelf off southeastern South America.Continental Shelf Research 28, 1607–1624.

Mora, G., Martinez, J.I., 2005. Sedimentary metal ratios in the Colombia Basin asindicators for water balance change in northern South America during the past400,000 years. Paleoceanography 20, PA4013. doi:10.1029/2005PA001132.

Moreno, A., Cacho, I., Canals, M., Grimalt, J.O., Sanchez-Vidal, A., 2004. Millenial-scalevariability in the productivity signal from the Alboran Sea record, Western Medi-terranean Sea. Palaeogeography, Palaeoclimatology, Palaeoecology 211, 205–219.

Murray, R.W., Leinen, M., 1996. Scavenged excess aluminum and its relationship tobulk titanium in biogenic sediment from the central equatorial Pacific Ocean.Geochimica et Cosmochimica Acta 60, 3869–3878.

Nederbragt, A.J., Thurow, J., 2005. Geographic coherence of millennial-scale climatecycles during the Holocene. Palaeogeography, Palaeoclimatology, Palaeoecology221, 313–324.

M.M. de Mahiques et al. / Quaternary International 206 (2009) 52–61 61

Pattan, J.N., Masuzawa, T., Divakar Naidu, P., Parthiban, G., Yamamoto, M., 2003.Productivity fluctuations in the southeastern Arabian Sea during the last 140 ka.Palaeogeography, Palaeoclimatology, Palaeoecology 193, 575–590.

Paytan, A., Kastner, M., Chavez, F.P., 1996. Glacial to interglacial fluctuations inproductivity in the equatorial Pacific as indicated by marine barite. Science 274,1355–1357.

Pfeifer, K., Kasten, S., Hensen, C., Schulz, H.D., 2001. Reconstruction of primaryproductivity from the barium contents in surface sediments of the SouthAtlantic Ocean. Marine Geology 177, 13–24.

Piola, A.R., Romero, S.I., 2004. Analysis of space–time variability of the Plata Riverplume. Gayana 68, 482–486.

Piola, A.R., Campos, E.J.D., Moller Jr., O., Charo, M., Martinez, C., 2000. Subtropicalshelf front off eastern South America. Journal of Geophysical Research 105 (C3),6565–6578.

Prakash Babu, C., Brumsack, H.J., Schnetger, B., Bottcher, M.E., 2002. Barium asa productivity proxy in continental margin sediments: a study from the easternArabian Sea. Marine Geology 184, 189–206.

Rocha, J., Milliman, J.D., Santana, C.I., Vicalvi, M.A., 1975. Southern Brazil. Uppercontinental margin sedimentation off Brazil. Contributions to Sedimentology 4,117–150.

Russell, J.M., Johnson, T.C., 2005. Late Holocene climate change in the North Atlanticand equatorial Africa: Millennial-scale ITCZ migration. Geophysical ResearchLetters 32, L17705. doi:10.1029/2005GL023295. 2005.

Salgado-Labouriau, M.L., Barberi, M., Ferraz-Vicentini, K.R., Parizzi, M.G., 1998. A dryclimatic event during the late Quaternary of tropical Brazil. Review of Palae-obotany and Palynology 99, 115–129.

Salgado-Labouriau, M.L., Casseti, V., Ferraz-Vicentini, K.R., 1997. Late Quaternaryvegetational and climatic changes in cerrado and palm swamp from CentralBrazil. Palaeogeography, Palaeoclimatology, Palaeoecology 128, 215–226.

Stevenson, M.R., Brito, D.D., Stech, J.L., Kampel, M., 1998. How cold water biotaarrive in tropical bay near Rio de Janeiro, Brazil? Continental Shelf Research 13,1595–1612.

Suguio, K., Martin, L., 1978. Quaternary Marine Formations of the States of Sao Pauloand Southern Rio de Janeiro. International Symposium on Coastal Evolution inthe Quaternary. Special Publication No. 1. Institute of Geosciences of theUniversity of Sao Paulo. 1–55.

Suguio, K., Martin, L., Bittencourt, A.C.S.P., Dominguez, J.M.L., Flexor, J.M.,Azevedo, A.E.G., 1985. Flutuaçoes do nıvel do mar durante o Quaternariosuperior ao longo do litoral brasileiro e suas implicaçoes na sedimentaçaocosteira. Revista Brasileira de Geociencias 15, 273–286.

Taylor, S.R., McLennan, S.M., 1985. The Continental Crust: Its Composition andEvolution. An Examination of the Geochemical Record Preserved in Sedimen-tary Rocks. Blackwell, Oxford. 311p.

Telford, R.J., Heegaard, E., Birks, H.J.B., 2004. All age–depth models are wrong: buthow badly? Quaternary Science Reviews 23, 1–5.

Van Geel, B., Burman, I., Waterbolk, H.T., 1996. Archaeological and palaeoecologicalindications of an abrupt climate change in The Netherlands, and evidence forclimatological teleconnections around 2650 BP. Journal of Quaternary Science11, 451–460.

Vernet, J.L., Wengler, L., Solari, M.E.,1994. Fire, climate and vegetation in central Brazilduring the Holocene – data from a soil-profile with charcoal (Salitre, Minas-Gerais). Comptes Rendus de l’Academie des Sciences, Serie II 319, 1391–1397.

Viau, A.E., Gajewski, K., Fines, P., Atkinson, D.E., Sawada, M.C., 2002. Widespreadevidence of 1500 a climate variability in North America during the past14 000 a. Geology 30, 455–458.

Viau, A.E., Gajewski, K., Sawada, M.C., Fines, P., 2006. Millennial-scale temperaturevariations in North America during the Holocene. Journal of GeophysicalResearch 111, D09102. doi:10.1029/2005JD006031.

Violante, R.A., Parker, G., 2004. The post-last glacial maximum transgression in thede la Plata River and adjacent inner continental shelf, Argentina. QuaternaryInternational 114, 167–181.

Wei, G., Liu, Y., Li, X., Chen, M., Wei, W., 2003. High-resolution elemental recordsfrom the South China Sea and their paleoproductivity implications. Paleo-ceanography 18. doi:10.1029/2002PA000826.

Ybert, J.P., Bissa, W.M., Catharino, E.L.M., Kutner, M., 2003. Environmental and sea-level variations on the southeastern Brazilian coast during the Late Holocenewith comments on prehistoric human occupation. Palaeogeography, Palae-oclimatology, Palaeoecology 189, 11–24.

Zembruscki, S.G., 1979. Geomorfologia da margem continental sul brasileira e dasbacias oceanicas adjacentes. In: Chaves, H.A.F. (Ed.), Geomorfologia da margemcontinental brasileira e areas oceanicas adjacentes. REMAC Project Series, vol. 7.Petrobras, Rio de Janeiro, pp. 129–177.