Present-Day Rhythmic Deposition in the Grand Rhone Prodelta (NW Mediterranean) According to High-Resolution Pollen Analyses

Journal of Coastal Research 21 2 292–306 West Palm Beach, Florida March 2005

Present-Day Rhythmic Deposition in the Grand RhoneProdelta (NW Mediterranean) According toHigh-Resolution Pollen AnalysesCelia Beaudouin†, Jean-Pierre Suc†, Genevieve Cambon‡, Abdelali Touzani§, Pierre Giresse§, Didier Pont††,Jean-Claude Aloısi‡‡, Tania Marsset§§, Pierre Cochonat§§, Daniele Duzer‡, and Jacqueline Ferrier‡

†LaboratoirePaleoEnvironnements etPaleobioSphere (UMRCNRS5125)

Batiment Geode rue Dubois69622 Villeurbanne Cedex,

France

‡Paleoenvironnements etPalynologie

Institut des Sciences del’Evolution (UMRCNRS5554), case 061

Universite Montpellier 234095 Montpellier Cedex 5,

France

§Laboratoire deSedimentologie marine

Universite de Perpignan52 avenue de Villeneuve66860 Perpignan Cedex,

France

††Ecologie des HydrosystemesFluviaux (UMR 5023)

Universite Claude Bernard—Lyon Batiment Forel

6 rue Dubois69622 Villeurbanne Cedex,

France

‡‡Centre d’Etudes et deFormation enEnvironnement Marin

Universite de Perpignan52 avenue de Villeneuve66860 Perpignan Cedex,

France

§§Laboratoire EnvironnementsSedimentaires

Geosciences MarinesIFREMERCentre de BrestBP 7029280 Plouzane, France

ABSTRACT

BEAUDOUIN, C.; SUC, J-P.; CAMBON, G.; TOUZANI, A.; GIRESSE, P.; PONT, D.; ALOISI, C.; MARSSET, T.;COCHONAT, P.; DUZER, D., and FERRIER, J., 2005. Present-day rhythmic deposition in the grand Rhone prodelta(NW Mediterranean) according to high-resolution pollen analyses. Journal of Coastal Research, 21(2), 292–306. WestPalm Beach (Florida), ISSN 0749-0208.

A high-resolution pollen study (136 samples) has been performed on the KTR05 core (738 cm long) located in themodern Grand Rhone prodelta. The objectives were (1) to evaluate the palynological and sedimentological recordquality of a prodelta in comparison with fluvial inputs (2) and to quantify lost sediment (resuspensions) in this area.The core covers more than the last thirty years. By comparison with the modern pollen rain both in the Rhone deltaand in the mid-drainage basin (Lyon), a seasonal forcing in pollen deposition has been demonstrated. Monthly pollensuccessions can be evidenced in a well preserved sequence, providing an estimate of the true sedimentation rate (60cm.yr21). Importance of disturbed or incomplete sequences suggests that most of the sediment first deposited in theprodelta (around ⅔) has been resuspended. Sedimentological and palynological data record not only seasonal butfluvial impact. The progressive shift of the prodelta has been evidenced firstly with sedimentology and secondly withpalynology. Despite high fluvial impact, pollen grains in the KTR05 core are representatives of the vegetation of theRhone drainage basin. The apparent homogeneity of the pollen spectra evidences the high reliability of palynologyapplied on prodeltaic area, at least in a microtidal environment.

ADDITIONAL INDEX WORDS: Palynology, fluvial impact, sedimentation, seasonality, prodelta, Mediterranean sea.

INTRODUCTION

Prodeltas characterise the interface between continentaland marine environments. When marine and fresh watermeet, physico-chemical processes (floculation) quickly bringabout the sedimentation of continental and marine particules(PAUC, 1980). Thus, 75% of the suspended matter of theRhone river is deposited before a 3-mile limit off the rivermouth (ALOISI et al., 1979). The Rhone river is the main sup-ply of sediment to the shelf of the gulf of Lion (ALOISI et al.,1979). Sedimentological processes on this shelf have been

DOI: 10.2112/02-878.1 received and accepted in revision 21 Septem-ber 2002.

studied for twenty years (ALOISI, 1986; ALOISI and MONACO,1975; GOT and ALOISI, 1990; GOT et al., 1985; TORRES et al.,1995; MARSSET and BELLEC, 2002) in order to evaluate theriver impact on marine sedimentation. They usually use geo-chemistry (MONACO et al., 1999; DURRIEU DE MADRON et al.,2000; CHARMASSON et al., 1998; RADAKOVITCH et al., 1999)and sedimentology. But, destroying processes (as resuspen-sions) affecting prodelta sediments are not well-known andnot quantified. Indeed, the prodelta of the Grand Rhone ischaracterised by a fast sedimentation rate which has beenestimated by several authors varying from 10 to about 50cm.yr21. ADDED et al. (1984) were the first to estimate a pro-deltaic sedimentation rate at about 8 to 10 cm.yr21 using ra-

293Pollen Analysis of Prodelta Deposits

Journal of Coastal Research, Vol. 21, No. 2, 2005

Figure 1. Location map of KTR05 core in the Gulf of Lions and of thetwo airborne pollen traps at La Tour de Valat and La Palissade.

Figure 2. Sediment succession of KTR05 core. Sediment is a more orless silty mud. Grey layers express color of the sediment.

dionuclides. ALOISI (1986) estimated a rate of 26 to 34cm.yr21 using seismic profiles by mesuring the height of theprodelta which began to be build in 1869, date of the stabi-lisation of the Rhone course (DUBOUL-RAZAVET, 1958). A rateof 37 to 48 cm.yr21 has been found by CHARMASSON et al.(1998) using radionuclides and 210pb. RADAKOVITCH et al.(1999) have estimated a rate of more than 20 cm.yr21 usingthe same method. These large variations of estimations mayoriginate from the quality of prodeltaic sedimentological rec-ord. The main objectives of this study are also to evaluate thequality of record into prodeltas, and to identify the processesdistorting this record by using geochemical, sedimentologicaland palynological approaches.

KTR05 core (738 cm) was drilled on January 1992, at theend of the TRANSRHO cruise (realised on the vesselL’Atalante; TORRES et al., 1995). It was situated just in frontof the Grand Rhone mouth (distant of about 3 km) within thepelitic prodeltaic sedimentation area (location area: Roustanbuoy; 43818.4709 N latitude; 4851.0209 E longitude; waterdepth: 40 m) (Figure 1).

METHOD

Palynology

Chemical Treatments and Analyses

For palynological study, samples have been regularly takeneach 5 cm. A classical chemical treatment has been per-formed on 10 grams of dry sediment. The method consists indestroying carbonates with HCl (33%), then dissolving silicawith HF (70%). A dense liquor has been applied and sampleshave been sieved at 200 and 10 mm. Finally, a Luber blendinghas been used (150 cm3 of HNO3 plus three drops of pure HClfor 5 minutes). Concentration of pollen grains has been cal-culated with the Cour’s method (COUR, 1974). The errorrange has been calculated with constant errors applied on

294 Beaudouin et al.


Figure 3. Sedimentologic parameters measured along the KTR05 core.

Figure 4. 137Cs content of KTR05 core.

volumes (6 20 ml), on weight of sediment (6 0.1 gram) andon calculation of the slide width to be analysed (6 200 mm).Analyses consisted in two main approaches: (1) a classicalpollen analysis has been carried out on the basis of more than250 pollen grains (without Pinus which is still more or lessover-represented) counted in each sample; (2) dinocysts havebeen considered on some selected parts of the core accordingto the other results. The results have been compared to themodern pollen rain collected during one year from two local-ities in Camargue (Figure 1). For the latter study, two hori-zontal pollen traps for collecting atmospheric pollen rain havebeen set up (COUR, 1974) at La Tour du Valat (near the Vac-cares lagoon) and at La Palissade (near the Grand Rhonemouth). A classical Chemical treatment has been done oneach sample (CAMBON, 1981). For the two localities, the con-tinuous pollen rain record has been analysed every fortnightduring one year (from early May 1992 to late April 1993 atLa Tour du Valat, from early December 1992 to late Novem-ber 1993 at La Palissade). 26 pollen records were obtainedfor each site.

Synthetic Pollen Diagrams Elaboration

Pollen results are presented on synthetic pollen diagrams.It results from the detailed pollen diagram after grouping thetaxa according to the various vegetation sources (altitudinalbelts, latitudinal zones, edaphic associations). Pinus, a gen-erally over-represented element because its pollen grains aregreatly advantaged both in production, transport and pres-ervation, has been excluded of the pollen sum for drawing thesynthetic pollen diagram and represented apart. This elimi-nation has no consequence on the ecological understanding ofthe diagram. Indeed, vegetation belts of the area are char-acterised by several Pinus species. Pollen analyst is unable

to distinguish the pollen of these species. Taxa have beengrouped into nine categories, from the left to the right:

(1) Alnus, Fraxinus, Populus, Salix will be considered asindicators of the riparian woodland which is especially welldeveloped along the lowland Rhone margins. Ulmus campes-tris, another element of the delta riparian woodland, is notso significant in this pollen group because of its scarcity inthe pollen records. The cause is probably the elm disease(FERAULT, 1983), the effect of which has been obvious in theatmospheric pollen record in Montpellier (CAMBON, 1981). Asa consequence, the rare pollen grains of Ulmus campestrishave been placed in the third group (mesophilous elements)with the other species of Ulmus.

(2) Water plants (Potamogeton, Ruppia, Typha, Spargan-ium, Myriophyllum, Cyperaceae etc.) reflect the freshwaterassociations which are greatly expanded in Camargue.

(3) Mesophilous elements (deciduous Quercus, Carpinus,Corylus, Tilia, Ulmus, Hedera, Buxus sempervirens, Ericaceae,Acer, Sambucus, Viburnum, Castanea, etc) are considered asmostly characteristic of the supramediterranean belt (alsocalled ‘‘hill vegetation’’ belt out of the Mediterranean realm).

(4) Altitudinal trees (Betula, Fagus, Abies and Picea) willrepresent the mountainous and subalpine combined belts.

(5) Cupressaceae pollen grains may originate from the Ca-margue (Juniperus phoenicea), from the mesomediterraneanassociations s.l. (Cupressus sempervirens, J. oxycedrus), fromthe successive altitudinal belts (J. communis, J. thurifera, J.nana and J. sabina), or even from gardens (introduced exoticspecies).

(6) Mediterranean xerophytes (Olea, Pistacia, Quercus ilextype, including the evergreen Q. ilex and Q. coccifera species,Cistus, Rhamnus) represent the mesomediterranean belt.



Figure 5. Simplified draft of pollen results (synthetic pollen diagram, Pinus percentages, pollen grains concentration). Pollen grains concentration isexpressed as pollen grain/gram of dry sediment.

(7) Herbs and shrubs (Asteraceae, Poaceae, Brassicaceae,Papaveraceae, Urticaceae, Helianthemum, Hippophae rham-noides, Plantago, Rumex, Polygonum, Sanguisorba, Filipen-dula, Galium, Mercurialis, etc.). For most of these taxa, iden-tification at the species (and often genus) level is question-able. In addition, their geographical distribution area mayconsequently be very large. This means that many of the cor-responding plants may represent several vegetation beltsfrom the Mediterranean shoreline to the alpine belt.

(8) Halophytic plants (Amaranthaceae/Chenopodiaceae,Caryophyllaceae, Plumbaginaceae, Ephedra, Tamarix). Thepollen grains of these taxa are mostly indicative of the Ca-margue halophytic vegetation (brackish and salted lagoons,coastline).

(9) Most of the so-called ‘‘other elements’’ are consideredin this work as not significant, since they include taxa whicheither have an anthropogenic significance (Platanus, Juglans,Cedrus, Aesculus, Liquidambar, Vitis, Cerealia, Rosaceae:gardens, road edges, cultivations) or may grow in a varietyof environments (Ranunculaceae).

Pollen Rain Data from Lyon

In addition, results have been compared to the pollen con-tent of the Lyon atmosphere recorded during the last decade(AFEDA, 1984, 1987, 1989, 1990, and 1992–2000). The entirecore has been compared to the record of pollen rain in Lyon.This place has been chosen as it is situated in the latitudinalmiddle of the Rhone basin. It gives an average for floweringdates with respect to latitudinal and altitudinal thermic gra-dient. In order to point the month of pollination of each taxa,an average has been made covering the last decade. Then,groups of taxa which flower during the same time have beenmade and applied to KTR05 by calculating the percentage ofeach groups along the core. AFEDA’s calendar begins in Feb-ruary and finishes in October. A lot of weeks are missingduring which we know that Cupressaceae are dominant. Theshortest period during which pollen are produced is twomonths. It represents the best resolution that we can expectfor applying to the core. The groups of taxa are:

(1) February–March: Alnus, Cupressaceae, Corylus, Taxus,



Figure 6. Detailed pollen diagram of KTR05 core spotlighting some selected taxa.



Figure 7. Simplified vegetation map of the Rhone drainage basin.

Ulmus. Winter (characterised by Cupressaceae) is includeinto this group.

(2) March–April: Populus, Betula, Fraxinus, Salix, Platan-us, Carpinus, Acer, Buxus

(3) April–May: deciduous Quercus, Juglans, Brassicaceae,Aesculus, Fagus, Moraceae, Phillyrea, Picea, Rosaceae

(4) May–June: Poaceae, Olea, Fabaceae, Ranunculaceae,Rubiaceae, Tilia, Vitis, Papaveraceae, Rumex, Sambucus

(5) June–July: Plantago, Castanea, Cereralia, Ligustrum(6) July–August: Urticaceae, Apiaceae, Asteraceae(7) August–September: Ambrosia, Amaranthaceae-Cheno-

podiaceae, Artemisia(8) September–October: Cedrus, Mercurialis

Sedimentology

Samples for sedimentological analyses (TOUZANI, 1998)have been taken following a less regular range (every 1 to 6cm on the main part of the core) in order to take into accountthe minor lithological variations. Using a Leco induction fur-nace, organic Carbon and inorganic Carbon contents weremeasured. Carbonate content was deduced from inorganicCarbon content (carbonate 5 inorg. Carbon 3 8.33). The sandfraction was determined by washing at 50 mm and was ex-amined using a stereomicroscope.

Geochemistry

Some samples have been used for radionuclide analyses.These samples underwent a direct gamma spectrometry todetermine radionuclide contents. However, in order to getenough material, layers grouping was sometimes necessaryleading to a lower resolution in radionuclide profiles com-pared to other parameters. 62 samples distributed over theentire section and representing sediment layers with thick-ness varying from 2 to 11 cm, were used for direct gamma-spectrometry analysis. These samples were first dried at 608Cto constant weight, then homogenized and packaged into 200cm3 or 20 cm3 calibrated geometries depending on the quan-tity of sediment available. Samples were measured with N-type hyper pure Germanium detector with 0.5 mm thicknessBerrylium window for 24 to 48 hours. Activities, expressedas Bq kg21 dry weight of sediment, and were corrected to thedate of the sample collection.

Spectral Analyses(Software AnalySeries 1.1: PAILLARD, 1996)

The objective was to point at cyclicities observed in the pal-ynological record, sand, CaCO3 and organic Carbon records.They have been compared with periods found on solid andliquid discharges of the Rhone taken every month from 1967to 1996 at Arles (PONT et al., 2002).

First, the data are prepared for processing, taking into ac-count the conditions related to the used methods. It consistsof a (1) re-sampling using a spline interpolation with a con-stant sampling rate corresponding to the mode of the values,(2) a trend removal when necessary, and, (3) a removal of themean.

Second, the spectral analysis is applied using 1) either the

FFT (Fast Fourier Transform) Blackmann-Tukey method ap-plied to the autocorrelation function and which completes astatistical evaluation of the frequencies and 2) the ME (Max-imum Entropy) which provides a good frequential resolution.The spectra display the frequencies from 0 to Nyquist (1/2dt,with dt 5 sampling rate). It is worth noting that a period issignificative, from a statistical point of view, only if it is pre-sent several times depending on the method used (at least 2to 3 times for the Maximum Entropy method and at least 6to 7 times for the Fast Fourier Transform). Both methods arecomplementary because Blackman Tuckey is more reliablethan Maximum Entropy and the latter has a better resolutionthan the former.

SEDIMENTOLOGICAL SETTING OF THE COREKTR05 (TOUZANI and GIRESSE, 2002)

The section shows alternations of centimetric to multicen-timetric light and dark grey laminae already pointed out inthis area by Aloisi and Monaco (1980). This general mode ofsedimentation is interrupted by some non-structured inter-vals of multicentimetric thickness (Figure 2). Dark layers arecoarser and contain between 30 and 40% of sand. They con-tain more organic Carbon (.1.5%) than the lighter layers.Light layers are poor in organic Carbon and rich in CaCO3.



Figure 8. Synthetic presentation of one year of pollen rain record from two localities in Camargue (La Tour du Valat, La Palissade).

Translation between the two laminae can be sharp or gra-dational. Figure 3 shows the evolution of sand, CaCO3 andorganic Carbon contents along the core. Sand content increas-es upward (from 10% to 20% in average) and is characterisedby high variations (from less than 10% to 50%) between 230cm and 50 cm core depth. Organic Carbon has the same trendas sand (from less than 1% to less than 2%). On the contrary,CaCO3 decreases from the bottom to the top of the core (fromless than 50% to 20%). Organic Carbon is made by coarseplant debris. It indicates fluvial input. But Rhone floods aregenerally characterised by a decrease of organic carbon dueto the dilution of organic matter by water. Then, an increaseof organic Carbon associated with an increase of sand indi-cates an increase in terrigeneous flows which does not sig-nificate intensifying floods. Carbonates of the prodelta sedi-ments mainly come from the river. Suspended matter con-tains 30 to 40% of CaCO3 (PONT, 1992) in the northern partof the delta. So, the high content in CaCO3 at the bottom ofthe core, relieved by the increase of sand and organic Carbonmay correspond to the space mobility of alluvion areas froma point relatively far from the depocentre to a point close tothe depocentre (coarser sediments). In detail, dark layerscharacterise peaks in terrigeneous transport. On the con-

trary, light layers characterise moderate terrigenous flowsand a low marine energy. Some ‘‘atypical’’ laminae have alsobeen recorded (relatively sandy light grey layers and rela-tively clayey dark layers). In these cases, they can be relatedto the action of floods.

CHRONOLOGICAL RESULTS OF THECORE KTR05

According to the results obtained by CHARMASSON et al.(1998) on a core sampled nearby KTR05 core in 1990, it wasexpected to obtain radionuclide profiles of both 137Cs and134Cs and to use their ratio to determine apparent sedimen-tation rates. However, due to the delay between samplingand analyses (up to 3.5 years) 134Cs was no more detected inthe samples of KTR05 core owing both to its relatively shorthalf-life (2.1 years) and to its lower discharges by nuclearinstallations. In addition, 137Cs profile is not very informativesince there are a lot of layers lacking gamma spectrometryanalyses (Figure 4), preventing us from finding clear alter-nation of low and high contents that should have been linkedto the Rhone River flow (CHARMASSON et al., 1998).

Nevertheless, 137Cs provided a chronological information:



Figure 9. Curve of Winter (February–April)/Summer (July–October)blooming plants as recorded through their pollen in KTR05 core. Thiscurve reveals some cyclicities which can be considered as belonging toseasonal succession especially in the uppermost part of the core (.257.5cm depth).

Figure 10. Percentages of taxa according to their flowering period (bimonthly grouping) within the reference window 257.5–177.5 cm depth (wellpreserved part of the core). February–March: Alnus, Cupressaceae, Corylus, Taxus, Ulmus. March–April: Populus, Betula, Fraxinus, Salix, Platanus,Carpinus, Acer, Buxus. April–May: deciduous Quercus, Juglans, Brassicaceae, Aesculus, Fagus, Moraceae, Phillyrea, Picea, Rosaceae. May–June: Poaceae,Olea, Fabaceae, Ranunculaceae, Rubiaceae, Tilia, Vitis, Papaveraceae, Rumex, Sambucus. June–July: Plantago, Castanea, Cereralia, Ligustrum. July–August: Urticaceae, Apiaceae, Asteraceae. August–September: Ambrosia, Amaranthaceae-Chenopodiaceae, Artemisia. September–October: Cedrus, Mer-curialis.

the detection of this nuclide in the layers above 595 cm cor-responds to a time window younger than 1958 (date of thefirst discharges by the factory of Marcoule and atmosphericnuclear weapons testings). 4 samples out of 62 analysed havelower 137Cs content than the threshold detection (628–638;650–655; 680–685; 690–695 cm) and 2 are very close to thislimit (568–573: 0.7 Bq.kg21; 720–730: 0.6 Bq.kg21).

POLLEN RECORD

One hundred thirty-six regularly distributed samplesprovided a diversified pollen flora (184 taxa). Figure 6shows the evolution of the proportion of the main taxaalong the core. On average, pollen concentration is higherthan 10,000 pollen grains/gram of dry sediment, and rang-es from 1 279 to 42 673 pollen grains/gram of dry sediment(Figure 5). After the synthetic diagram (Figure 5), all thevegetation belts of the entire Rhone drainage basin arerepresented in the pollen record (Figure 7). In a previouswork (CAMBON et al., 1997), it has been concluded thatsuch a pollen assemblage is well representative of the veg-etation of the region, the transport to the sedimentationarea being forced by fluvial action. This gives a high suit-ability to the pollen diagrams obtained on similar prodel-taic deposits for older periods (Late Pleistocene: ACHERKI,1997; BEAUDOUIN et al., in press; Pliocene: SUC, 1984 and1989) in the same area. Percentages of the different pollengroups are almost constant along the studied section. Butthere are some minor differences which can be explainedby pollen grains productivity (seasonality) and by distanceof transport. Indeed, the evolution of pollen grain produc-tivity is obvious in pollen rain at La Tour du Valat and LaPalissade (Figure 8). Records of pollen rain illustrate theseasonal pollen distribution along the year in relation withplant blooming which can be summarized as follows.



Figure 11. Detail of some sedimentological and palynological parameters. The shaded zone corresponds to the well-preserved sequence. Concentrationis expressed as a number of grain/gram of dry sediment.

The riparian tree pollen grains are recorded in winterand spring, and the water plant pollen grains are recordedin summer. Mesophilous elements pollen is abundant frommid spring to mid summer. The altitudinal tree pollen isscarce but concentrates in spring. Cupressaceae pollengenerally shows low percentages within the sediments be-cause of its poor preservation. Here, it greatly prevails (be-cause of few other plants flowering at the same period)from mid autumn to mid spring. Mediterranean xero-phytes characterise spring and early summer. Bloom ofherbs occurs from late spring to mid autumn, including oneto several peaks of halophytes.

Some differences exists between La Tour du Valat andLa Palissade which can be explained by some aspects ofthe local vegetations (for example, water plants are moreabundant at La Tour du Valat and riparian trees are morefrequent at La Palissade). In the same way, when compar-ing the KTR05 core pollen record (Figure 5) with the at-mospheric pollen record, some aspects appear directlylinked to the mode of transport: for example, the regionalvegetation (riparian, mesophilous and altitudinal groups)is better represented in the KTR05 record thanks to thefluvial input, and the local vegetation (water plants, Med-iterranean xerophytes, herbs and halophytes) is more per-formed in the atmospheric record.

Despite the signal is homogeneous along the core, someseemingly minor variations are mainly perceptible in theupper part of the KTR05 pollen diagram (since 257 cmdepth). From 257.5 cm to 177.5 cm depth (i.e. from samples46 to 61 enclosed), the following succession (respectingmaximum frequencies) has been noted: Cupressaceae andriparian trees—altitudinal trees—herbs (including halo-phytes)—Cupressaceae and riparian trees. This sequenceevokes a seasonal succession ranging from winter–springto winter–spring of the following year. This hypothesis is

also supported by the well-contrasted curve (Figure 9)which represents the ratio between plants flowering inwinter–spring (February to April included) and plantsflowering in summer–automn (August to October includ-ed). Some similar sequences can be recognised in the up-permost part of the core (177–120 cm, 120–50 cm), but thesequence 257.5–177.5 cm is the most obvious and is se-lected as reference.

Compared with the AFEDA calendar, most of the coreshows too short successions of groups of months (less thanthree) to be taken into account. One section has revealed along enough temporal signal (from 257.5 cm to 177.5 cmdepth) (Figure 10). There is a succession from taxa floweringin February–March to March–April and then April–July,July–September, September–March and back again toMarch–May of the following year. Such successions are in-dicated by arrows on Figure 10.

COMPARISON BETWEEN POLLEN RECORD ANDSEDIMENTOLOGICAL-GEOCHEMICAL ANALYSES

Trends in Palynological and Sedimentological Record

Figure 5 shows an upcore decrease of Pinus parallel to thedecrease in CaCO3 (Figure 3). And pollen grain concentrationincreases parallel to sand and organic Carbon (Figures 3 and5). This illustrates the evolution of the shift of the prodeltadepocentre. Indeed, present-day palynological sea-water sam-ples contain few pollen of Pinus when directly under the in-fluence of the river, i.e. located within the fluvial plume. Oncontrary, present-day sea-water samples taken out of the flu-vial plume show low pollen concentration vs. high percentagein Pinus.



Figure 12. Results of spectral analysis realised on KTR05 pollen con-tent. Pollen have been grouped according to blooming months. Thickcurve corresponds to frequences found with the Maximum Entropy andthin curve corresponds to frequences found with the B-Tukey. X-axis rep-resents frequences (cm21) and Y-axis represents the power of spectrum(P(f)).

Preservation of Palynological and SedimentologicalSignal

Two kinds of sequences have been distinguished aftercomparison between the seasonal interpretation of the corebased on pollen record and the lithologic succession (Fig-ure 11): (1) preserved sequence, (2) disturbed or incompletesequence.

Preserved Sequence (177.5–257.5 cm depth)

Pollen data suggest a continuous record along more thanone year. So, the successive months identified according toblooming of plants are used as chronological reference for thecorresponding sediments. Sand percentages vary from 3 to23% except for one level which reaches 50% at 216 cm depth.Organic Carbon reaches 0.9 to 1.5%. Sand percentage in-creases and reaches 10 to 22% in February–April. Duringthese months, organic Carbon decreases. Then sand decreas-es in April–June inversely to organic Carbon. From June toOctober, sand decreases. It increases in October–Februaryand then decreases in March. Pollen concentration increases

relatively regularly from February to October–February andthen decreases in March. Pollen concentration increases rel-atively regularly from February to October–February (fromless than 5000 to less than 20 000 pollen grains/gram of drysediment). It increases since February to October. Then, itslightly decreases during February–March. And finally, it in-creases during March–April. Reworked material (spores, pol-len grains and dinocysts) and modern dinocysts are not fre-quent compared with pollen grain concentration. In somesamples, reworked material are slightly more frequent. Re-worked dinocysts originate from Mesozoic deep-marine sedi-ments which are exposed on large surfaces in the southernAlpine area. Some are dated from Cenozoic marine sedimentssituated in the Rhone valley. Over this well-preserved sec-tion, reworked material increases from April–May to Marchof the next year.

From February to March, the combination of low contentin pollen grains, low content in dinocysts, low content in or-ganic carbon could be due to a greater dilution of transportedmatter. This could correspond to successions of spring floods.Then, the increase of pollen concentration, reworked dino-cysts may correspond to weaker dilution of organic matterduring summer. The weaker variations in sand, organic Car-bon content, pollen grain concentration and reworked dino-cysts could correspond to little short floods.

Modern dinocysts are rare. Their concentration never ex-ceeds 50 dinocysts/gram of dry sediment. They are mainlyrepresented by Lingulodinium machaerophorum (a coastaldinocyst) and some Spiniferites (an oceanic dinocyst). Someof them are corroded and could indicate a poor conservationof organic matter.

Example of Disturbed or Incomplete Sequence(257.5–299 cm depth)

From the palynological record, this interval does not pro-vide evidence for seasonal succession. This corresponds to 3unconformities (Figure 2). Reworked material and moderndinocysts are more frequent and their concentration variesmore than before (from nearly 0 to respectively 321 and 134per gram of dry sediment) (Figure 11). There are some cor-roded specimens. Pollen grain concentration reveals big var-iations from one sample to another (from 5000 pollen grainsto more than 20 000 pollen grains/gram of dry sediment).This disturbed sequence could be the consequence of strongfloods and/or storms. Possible origin of this disturbed se-quence includes oscillations wave packing or undirectionalcurrent clustering.

SPECTRAL ANALYSIS

Pollen Record (Figure 12)

From each sample taken in the core every 3 to 5 cm,different taxa have been identified. Each data set consistsin a percentage of pollen grains or groups of pollen grainsper sample from the top to 595 cm depth (before the non-structured intervals). The re-sampling was carried out us-ing a sampling rate of 5 cm. In the spectra, the frequentialpeaks correspond to centimetric periodicities which can be



Figure 13. Results of spectral analysis realised on the main pollen grains of each group representing months. Thick curve corresponds to frequencesfound with the Maximum Entropy and thin curve corresponds to frequences found with the B-Tukey. X-axis represents frequences (cm21) and Y-axisrepresents the power of spectrum (P(f)).

converted into temporal periodicities by using sedimenta-tion rate considered as a constant value. Two hypothesesare tested: (1) sedimentation rate calculated on the well-preserved sequence of the core (60 cm 5 1 year) and (2)sedimentation rate corresponding to an average calculated

on the entire core (595 cm of sediment deposited during 34years 5 17.5 cm.yr21). Groups of pollen grains correspond-ing to months have been tested (Figure 12). Most groupsreveal periodicities of about 15 cm. Periods of 30 cm, 60cm, 90 cm and bigger than 178 cm are present in several



Figure 14. Results of spectral analysis realised on pollen grains concen-tration. Thick curve corresponds to frequences found with the MaximumEntropy and thin curve corresponds to frequences found with the B-Tu-key. X-axis represents frequences (cm21) and Y-axis represents the powerof spectrum (P(f)).

Figure 15. Results of spectral analysis realised on sedimentological parameters (CaCO3, organic Carbon and sand content). Thick curve corresponds tofrequences found with the Maximum Entropy and thin curve corresponds to frequences found with the B-Tukey. X-axis represents frequences (cm21) andY-axis represents the power of spectrum (P(f)).

groups. The shortest period could correspond to significantperiod but its existence has to be confirmed using a moreclosely spacing. In the same way and as explained previ-ously, a period is significative when it return 3 times atleast along the signal. Here, the ratio of 595/3 (5 about200 cm) gives the longest period to be significative. As aconsequence, all periods higher than 200 cm are not sig-nificant.

According to the sediment thickness considered to berepresentative of one year for what concerns the preservedsequence, the period of 60 cm could illustrate the seasonalcyclicity. Then, the other period of 30 cm corresponds to 6months. If we take into account the average sedimentationrate (17.5 cm.yr21), the period of 15 cm could correspondto one year. But, as explained before, this hypothesis hasto be confirmed using a higher resolution in sampling. Theperiodicities are not obviously present in all the data setswith respect to specific groups of taxa. Then, the mostabundant taxa of each group have been independently test-ed in order to estimate impact of the origin of pollen oncyclicities (Figure 13). Pinus has been analysed too. Sev-eral cycles are recorded and some are not well-marked. Theperiod of about 15 cm is given by most of the taxa. Theperiod of about 30 cm is given by Platanus, Fagus, Plan-tago, Asteraceae and Mercurialis. The period of about 60

cm is given by Picea, Fagus and deciduous Quercus. Theperiod of 90 cm is given by Alnus and Salix (limit). Theperiod of 178 cm is given by Poaceae. A new period (50 cm)is recorded by Salix, Betula, Mercurialis and Pinus. Pollengrain concentration has been tested too (Figure 14). It re-veals a period of nearly 25 and 50 cm.

Granulometry, CaCO3 and Organic Carbon (Figure 15)

The re-sampling was applied using a sampling rate of 4cm. Sand gives several periods (10, 17, 22, 32, and 77 cm).CaCO3 is characterised by periods of 23, 55 and 222 cm. Or-ganic Carbon reveals periods of 18, 30, 55 to 70 and 250 cm.The last period (222 to 250) is an artefact due to a gap insampling.

Periodicities of the Rhone Floods

The spectral analyses of monthly liquid and solid discharg-es in Arles from 1967 to 1996 (PONT et al., 2002) have beendone. Results are shown on (Figure 16). The liquid dischargegives periods of 6 months, 1 and 6 years. The solid dischargegives periods of 4 months, 1 and 5 years.

Signification of the Periods

The periods found by pollen grains may have two origins:they may be due to seasonnal flowering (1 year) or to fluvialinput. Indeed, floods can resuspend sediment along the rivercourse containing pollen grains. The periods given by sand,CaCO3 and organic Carbon can be controlled by the Rhoneriver and/or marine swell.

Pollen grains which give a period of 30 cm are composed ofherbs and trees living in all the drainage basin (except forFagus which is situated in mountains). The herbs genera in-clude several species which cannot be determined by pollenanalysts. These herbs generally have a long flowering period.But Platanus and Fagus have a small period of flowering. Theperiod of 60 cm is given by trees which live in mountains(Picea and Fagus) and in all the drainage basin (deciduousQuercus). Both periods (30 cm and 60 cm) could correspondto the seasonnal flowering. Pollen grains concentration is de-pendent on floods which induce a dilution of water pollengrains content. Floods mainly occur every year. Then, the



Figure 16. Results of spectral analysis realised on monthly liquid and solid discharges of the Rhone river in Arles from 1967 to 1986. Thick curvecorresponds to frequences found with the Maximum Entropy and thin curve corresponds to frequences found with the B-Tukey. X-axis representsfrequences (cm21) and Y-axis represents the power of spectrum (P(f)).

smaller period given by spectral analysis on pollen grainsconcentration could correspond to 1 year (25 cm). As a con-sequence, the most closely period given by pollen grain (30cm) could correspond to 1 year. A period of nearly 30 cm isgiven by Sand and organic Carbon which seems to confirmthis hypothesis.

If the period of 30 cm corresponds to a year, the period of60 cm correspond to 2 years, 90 cm to 3 years and 178 cm to6 years. This hypothesis is consistant with the cyclicities re-corded from the Rhone floods.

DISCUSSION AND CONCLUSION

General shift of the depocentre of the prodelta is clearlyidentified by both pollen concentration, Pinus percentages,sand, CaCO3 and organic Carbon. Variations of percentagesof Pinus is generally attributed to variations of distance fromcoastal shoreline (ACHERKI, 1997; HEUSSER, 1988). From1958 to 1992, the coast has not deeply changed (SUANEZ andPROVANSAL, 1998) whereas percentage of Pinus highly de-creases. It seems that abundance of Pinus is highly influ-enced by fluvial input.

According to SUC (1984) pollen concentration is controlledby river flow and transport of terrigenous material which in-creases pollen dilution. Similar conclusions concern 137Cs con-centration in the Rhone prodelta (CHARMASSON et al., 1998).Low pollen and 137Cs concentrations should correspond tohigh Rhone input and inversely. However, the fact that 137Csprofile is both discontinuous and has been obtained on thick-er sediment portions compared to pollen data does not allowus to draw precise parallel between the two profiles. Calcu-lation of rate of sedimentation on the entire core gives anaverage value of 17.5 cm.yr21. This is low compared with val-ue of 60 cm.yr1 obtained on the preserved sequence (between187.5 to 247.5 cm depth). The latter may nearly equals to thepresent sedimentation rate. Average rates in this area wouldcorrespond to residual average sedimentation rates: such lowvalues are mainly due to resuspensions of particules whichare transfered on different area of the prodelta, the shelf and/or into canyons (MONACO et al., 1999). If we compare theaverage rate of sedimentation for the entire core (17.5cm.yr21) to the sedimentation rate found on the preserve se-

quence (60 cm.yr21), most of the sediment first deposited inthe prodelta has been resuspended (about ⅔). Then, we mustconsider that evidence of this rate of 60 cm.yr21 is exceptionalin such a sediment.

According to spectral analysis, periods revealed by bothpollen grains and sedimentological data seem to be stronglycorrelated with fluvial input. Marine effects, which are main-ly caused by E-SE swells (LACOMBE, 1988), are secondarywith respect to the prevalent Rhone terrigenous input. Thisis consistent with scarcity of sedimentary unconformities andmarine dinocysts within KTR05 core.

Pollen flora is consistent with vegetation of the region(from coastline to mountains). Recorded within prodeltaicsediments, it is able to produce a precise seasonal time-scale because of the almost immediate succession of thephases: blooming of plant, pollen diffusion by air, pollenrain. Transport of pollen masses, whatever the place ofblending into waters along the Rhone, appears to be a rel-atively quick phenomenon. Nevertheless, a part of the pol-len content (probably constituted by specimens which fellin a fluvial agent upstream within the drainage basin) doesnot reach fastly the prodeltaic deposition area. Because oftheir delay in transport, a lot of these pollen grains is con-tinuously arriving into the prodelta. They are probably atthe origin of the apparent homogeneity of the compositepollen signal.

This study is the first to have been so detailed in pro-deltaic deposits. Thanks to the potentiality of the pollensignal, it is possible to evidence a seasonal signature with-in a very complex and composite record. But, there is apredominance of the Rhone terrigenous input on the sea-sonal forcing. Nevertheless, marine effects which provokeirregular disturbances cannot be neglected. In such a ma-rine palynological analysis, dinocysts (reworked or not)which provide informations about coastal or oceanic influ-ences, are to be systematically counted and identified. Sed-imentological data should be taken into account for pa-leoenvironmental interpretations as well in order to un-derstand marine influence on pollen grain assemblages.This study demonstrates the high potentiality of pollen re-cord in fast sedimentary systems in good agreement with



sedimentological parameters. It is demonstrated that pro-delta areas are very propitious for high resolution pollenanalyses whatever the age of the sediment.

ACKNOWLEDGMENTS

The authors are indebted to P. Cour (Montpellier) for pro-viding pollen traps, J.-C. Gleyse (La Tour du Valat) and S.Amico (La Palissade) for sampling pollen filters, S. Charmas-son and M. Arnaud (IFREMER, La Seyne-sur-mer) for pro-viding Caesium data and for the manuscript inprovment, G.Clauzon (Aix en Provence) for asking for realising KTR05core and F. Giraud (Lyon) for improving spectral analyses.We thank crew members of the R.V. L’Atalante for their ef-ficiency in collecting this core.

This work was supported by the National Programme onCoastal Oceanography (PNOC-ECOCOT). This is a contri-bution to the Programme ‘‘Environnement, Vie et Socie-tes’’.

LITERATURE CITED

ACHERKI, N., 1997. Analyse palynologique de quatre carottes du gol-fe du Lion: application a la restitution de la vegetation et du climatdu midi de la France pendant le dernier cycle climatique et a astratigraphie marine. Montpellier, France: University of Montpel-lier 2, Unpublished Ph.D thesis, 149p.

ADDED, A.; FERNEX, F.; RINGOT, J.-L., and SPAN, D., 1984. Carac-teres sedimentologiques et geochimiques du plateau continentaldevant l’embouchure du Grand Rhone. In: BIZON, J.-J., and BUR-OLLET, P.-F. (Eds.), Ecologie des microorganismes en Mediterraneeoccidentale ‘‘ECOMED’’. Paris, France: Association Francaise desTechniciens du Petrole, pp. 33–42.

ASSOCIATION FRANCAISE D’ETUDE DES AMBROISIES (1984, 1987,1989, 1990, and 1992–2000). Calendrier pollinique. Saint-Priest,France.

ALOISI, J.-C., 1986. Sur un modele de sedimentation deltaıque. Con-tribution a la connaissance des marges passives. Perpignan,France: University of Perpignan, Thesis, 162p.

ALOISI, J.-C. and MONACO, A., 1975. La sedimentation infralittorale.Le prodeltas nord-mediterraneens. Comptes Rendus de l’Academiedes Sciences de Paris, ser. D, 280, 2833–2836.

ALOISI, J.-C. and MONACO, A., 1980. Etude des structures sedi-mentaires dans les milieux deltaıques (Rhone). Apport a la con-naissance des conditions de sedimentation et de diagenese.Comptes Rendus de l’Academie des Sciences, Paris, ser. D, 290,159–162.

ALOISI, J.-C.; MILLOT, C.; MONACO, A., and PAUC, H., 1979. Dyna-mique des suspensions et mecanismes sedimentogenetiques sur leplateau contiental du Golfe du Lion. Comptes Rendus de l’Academiedes Sciences, Paris, ser. D, t. 289 (13), 879–882.

BEAUDOUIN, C.; SUC, J.-P.; ACHERKI, N.; COURTOIS, L.; ROBINEAU,M.; ALOISI, J.-C.; SIERRO, F.J., and OBERLIN C. Palynology of theNorthwestern Mediterranean Shelf (Gulf of Lions): First vegeta-tional record for the last climatic cycle. Marine and Petroleum Ge-ology. In press.

CAMBON, G., 1981. Relations entre le contenu pollinique del’atmosphere et le couvert vegetal en Mediterranee occidentale aMontpellier (France), Valencia (Espagne) et Oran (Algerie). Mont-pellier, France: Univ. Montpellier 2, Ph.D. Thesis.

CAMBON, G.; SUC, J.-P.; ALOISI, J.-C.; GIRESSE, P.; MONACO, A.;TOUZANI, A.; DUZER, D., and FERRIER, J., 1997. Modern pollendeposition in the Rhone delta area (lagoonal and marine sedi-ments), France. Grana, 36, 105–113.

CHARMASSON, S.; RADAKOVITCH, O.; ARNAUD, M.; PRUCHON, A.S.,and BOUISSET, P., 1998. Long-core Profiles of 137Cs, 134Cs, 60Co and210Pb in Sediment Near the Rhone River (Northwestern Mediter-ranean Sea). Estuaries, 21(3), 367–378.

COUR, P., 1974. Nouvelles techniques de detection des flux et desretombees polliniques: etude de la sedimentation des pollens etdes spores a la surface du sol. Pollen et Spores, 16(1), 103–141.

DUBOUL-RAZAVET, C., 1956. Contribution a l’etude geologique et se-dimentologique du delta du Rhone. Memoires de la Societe geolo-gique de France, 76, 1–234.

DURRIEU DE MADRON, X.; ABASSI, A.; HEUSSNER, S.; MONACO, A.;ALOISI, J.-C.; RADAKOVITCH, O.; GIRESSE, P.; BUSCAIL, R., andKERHERVE, P., 2000. Particulate matter and organic carbon bud-gets for the Gulf of Lions (NW Mediterranean). Oceanologica Acta,23(6), 717–730.

FERAULT, C., 1983. La graphiose de l’orme. Horticulteurs et Mar-aıchers, 237, 25–29.

GOT, H.; ALOıSI, J.-C., and MONACO, A., 1985. Sedimentary pro-cesses on deltas and shelves of the Mediterranean sea. In: WEZEL.,F.C. AND STANLEY, D.J. (Eds.), Geological evolution of the Medi-terranean Basins, Pergamon Press, Oxford, pp. 355–376.

GOT, H. and ALOıSI, J.-C., 1990. The Holocene sedimentation on theGulf of Lions margin: a quantitative approach. Continental ShelfResearch, 10(9-1), 841–855.

HEUSSER, L.E., 1988. Pollen distribution in marine sediments on thecontinental margin off northern California. Marine Geology, 80,131–147.

LACOMBE, L., 1988. Considerations generales sur l’oceanographiephysique mediterraneenne. In: MINAS, H.J. and NIVAL, P. (Eds.),Oceanographie Pelagique Mediterraneenne, Oceanologica Acta Spe-cial Issue No. 9, pp. 7–12.

MARSSET, T. and BELLEC, V., 2002. Late Pleistocene-Holocene de-posits of the Rhone inner continental shelf (France): detailed map-ping and correlation with previous continental and marine stud-ies. Sedimentology, 49, 255–276.

MONACO, A.; DURRIEU DE MADRON, X.; RADAKOVITCH, O.; HEUSS-NER, S., and CARBONNE, J., 1999. Origin and variability of down-ward biogeochemical fluxes on the Rhone continental margin (NWMediterranean). Deep-Sea Research, 146, 1483–1512.

PAILLARD, D.; LABEYRIE, L., and YIOU, P., 1996. Macintosh ProgramPerforms Time-Series Analysis. Eos. Transactions, American Geo-physicla Union, 77, 379.

PAUC, H.E., 1980. Floculation et potentiel de surface des materiauxen suspension en environnement d’embouchure. Comptes Rendusde l’Academie des Sciences, Paris, ser. B, t. 90, 175–178.

PONT, D., 1992. Caracterisation de la charge solide en suspensiondu Rhone au niveau du palier d’Arles lors d’une crue d’importancemoyenne. Rapport final, Agence de l’eau Rhone-Mediterranee-Corse, University Lyon 2, 25p.

PONT, D.; SIMONNET, J.-P., and WALTER, A.V. (2002). Medium-termChanges in Suspended Sediment Delivery to the Ocean: Conse-quences of Catchment Heterogeneity and River Management(Rhone River, France). Estuarine, Coastal and Shelf Science, 54(1),1–18.

RADAKOVITCH, O.; CHARMASSON S.; ARNAUD, M., and BOUISSET, P.,1999. 210Pb and Caesium Accumulation in the Rhone Delta Sedi-ments. Estuarine, Coastal and Shelf Science, 48, 77–92.

SUANEZ, S. and PROVANSAL, M., 1998. Large Scale Evolution of theLittoral of the Rhone. Journal of Coastal Research, 14(2), 493–501.

SUC, J.-P., 1984. La signification paleoecologique des restes vege-taux. Bulletin des Centres de Recherche Exploration-Production ElfAquitaine, 8(1), 97–104.

SUC, J.-P., 1989. Distribution latitudinale et etagement des associ-ations vegetales au Cenozoıque superieur dans l’aire ouest-medi-terraneenne. Bulletin de la Societe geologique de France, ser. 8,5(3), 541–550.

TORRES, J.; SAVOYE, B., and COCHONAT, P., 1995. The effects of lateQuaternary sea-level changes on the Rhone slope sedimentation(Northwestern Mediterranean), as indicated by seismic stratigra-phy. Journal of Sedimentary Research, B65(3), 368–387.

TOUZANI, A., 1998. L’edification prodeltaıque du Rhone: marqueursde l’evolution seculaire de la sedimentation. Perpignan, France:Unpublished Ph.D. Thesis, Univ. Perpignan. 190p.

TOUZANI, A. and GIRESSE, P., 2002. The Rhone River Prodelta:Short-Term (108–103 Year) Sedimentation Patterns and HumanImpact. Journal of Coastal Research, 18(1), 102–117.



M RESUME M

Une analyse a haute resolution (136 echantillons) a ete realisee sur la carotte KTR05 situee dans le prodelta actuel du Grand Rhone. Il s’agissait (1) d’evaluer lesqualites de l’enregistrement palynologique et sedimentologique dans le prodelta par comparaison avec les apports fluviaux et (2) de quantifier les pertes de sedimentsdues aux remises en suspension. La carotte couvre un peu plus des trente derniere annees. Le contenu pollinique enregistre le forcage saisonnier et l’influence dufleuve. La vitesse de sedimentation reelle a pu etre calculee sur une section preservee (60 cm. an21). Le reste de la carotte (sequences palynologiques perturbees)suggere que les 2/3 du sediment depose a ensuite ete remis en suspension et redistribue sur les fonds marins. La stabilite du signal pollinique suggere la fiabilitede l’enregistrement palynologique dans ce type de corps sedimentaires.

Present-Day Rhythmic Deposition in the Grand Rhone Prodelta (NW Mediterranean) According to High-Resolution Pollen Analyses

Documents