Impaginato Di Donato EEFFFFEECCTTSS OOFF LLAATTEE PPLLEEIISSTTOOCCEENNEE--HHOOLLOOCCEENNEE CCLLIIMMAATTIICC CCHHAANNGGEESS OONN TTHHEE PPLLAANNKKTTOONNIICC FFOORRAAMMIINNIIFFEERRAA IINN TTHHEE GGUULLFF OOFF GGAAEETTAA ((TTYYRRRRHHEENNIIAANN SSEEAA,, IITTAALLYY)) VVaalleennttiinnoo DDii DDoonnaattoo Dipartimento di Scienze della Terra - Università degli Studi di Napoli “Federico II” Tel. 081 5473338 - e-mail: valedido@unina.it ABSTRACT Planktonic foraminiferal assemblages of core G93-C9, recovered in the Gulf of Gaeta (Tyrrhenian Sea) at 212 m of depth, were analy- sed by Principal Components Analysis. The analysis was computed first on raw percentage data and then on logratio transformed data. The most important difference between raw data and logratio data rests in the behaviour of N. pachyderma which in the first case, unlike the second, does not have significant loading on the first component. Sea surface temperatures exerted a strong control on fora- miniferal assemblages. Hydrography and productive system also played an important role in determining composition of assemblages, as shown by the loadings of grazing species on the second component. During the last deglaciation and the Early Holocene eutrophic conditions allowed Globorotalia inflata to expand its distribution toward surface waters of the Gulf of Gaeta. During the Late Holocene, planktonic foraminiferal assemblages evidence a tendency towards oligotrophic conditions. RIASSUNTO Effetti dei cambiamenti climatici tardo pleistocenici-olocenici sui foraminiferi planctonici del Golfo di Gaeta (Mar Tirreno, Italia). Vengono presentati i risultati di una Analisi dei Componenti Principali applicata alle associazioni a foraminiferi planctonici della carota G93-C9, prelevata nel Golfo di Gaeta a 212 m di profondità. L’analisi è stata effettuata prima sui dati originali e successivamente applicando ad essi una trasformazione logratio per correggere l’effetto della chiusura dei dati percentuali. I risultati ottenuti differiscono principalmente nel carico di Neogloboquadrina pachyderma sul primo componente, che risulta quasi nullo con i dati originali e significativo con i dati trasformati. Le temperature delle acque superficiali hanno svolto un controllo primario sulle variazioni delle associazioni, in cui si rico- noscono chiaramente due gruppi di specie con carichi opposti sul primo componente. I punteggi del primo componente lungo la carota definiscono chiaramente i principali eventi climatici avvenuti durante l’ultima deglaciazione. I carichi sul secondo componente di specie prevalentemente erbivore evidenziano l’effetto di fattori trofici ed idrografici. In particolare tra Tardiglaciale e Olocene inferiore si instau- rano condizioni eutrofiche che consentono a Globorotalia inflata, una specie prevalentemente mesopelagica, di proliferare anche in prossimità della costa. Viceversa nel corso dell’Olocene superiore i risultati evidenziano l’istaurarsi di condizioni oligotrofiche. Keywords: planktonic foraminifera, climatic changes, Late Pleistocene-Holocene, Tyrrhenian Sea, Principal Components Analysis. Parole chiave: foraminiferi planctonici, cambiamenti climatici, Pleistocene superiore-Olocene, Mar Tirreno, Analisi dei Componenti Principali. Il Quaternario Italian Journal of Quaternary Sciences 1155(2), 2002, 251-257 11.. IINNTTRROODDUUCCTTIIOONN In the last few decades many studies involving the analysis of foraminiferal assemblages have been carried out on Late-Pleistocene-Holocene sediments in the Mediterranean Sea, allowing detailed reconstruction of the environmental changes occurred during the last cli- matic cycles (Blanc-Vernet et Sgarrella, 1989; Kallel et al.1997, Ariztegui et al. 2000, among others). Relatively few studies, however, focus on neritic or coastal envi- ronments. An opportunity to study the effect of climatic changes in a neritic milieu was offered by an oceano- graphic cruise of the O/V Urania, in October 1993, during which several cores were collected from the Gulf of Gaeta continental shelf. The general results of an integrated palaeontological study on three cores (G93- C5, G93-C8, G93-C9) involving descriptive analysis of planktonic foraminifera, calcareous nannofossils, pollen and ostracoda, were reported by Amore et al. (2000). A detailed analysis of calcareous nannofossil assembla- ges, involving comparisons with other sectors of the southern Tyrrhenian continental shelf, was carried out by Esposito (1999). Due to its continuous stratigraphical record and good quality of fossil assemblages, core G93-C9 offers the opportunity to examine planktonic foraminiferal assemblages in detail. With a view to exa- mining the responses of foraminiferal assemblages to environmental changes, planktonic foraminifera of core G93-C9 are investigated by Principal Component Analysis (PCA). 252 V. Di Donato 22.. MMAATTEERRIIAALLSS AANNDD MMEETTHHOODDSS The study area is located within the Gulf of Gaeta continental shelf, in the southern Tyrrhenian Sea (Fig. 1). The area represents an extension of the Volturno river and Garigliano river coastal plains. Gravity core G93-C9 (Lat.41°02’ 4” N, Long. 13°32’18”E), 470 cm long, was collected at a depth of 212 m, along an ESE- WSW seismic line. The stratigraphy of core G93-C9, whose sediments consist mainly of clays, is summarised in Figure 2. For palaeontological analyses of core G93- C9, 47 samples spaced 10 cm apart were collected. Samples were washed through a 106 µm sieve. From each sample at least 300 specimens of planktonic fora- minifera were collected by using complete splits. The choice of the sieve mesh, which may appear somewhat unusual, requires some explanation. In classical works on oceanic planktonic foraminiferal assemblages large sieve meshes were adopted (i.e. 149 µm) (Imbrie & Kipp, 1971; among others), obtaining very low percenta- ges of undeterminable specimens. This method, howe- ver, also artificially reduces the percentages of small species. At the opposite end, in recent works on Mediterranean planktonic foraminiferal assemblages (Capotondi, 1999; among others) very small sieve meshes (65 µm) were adopted, obtaining higher (and less biased) abundances of small species. Assemblages resulting from treatment with a small sieve mesh, however, contain high percentages of undetermi- nable specimens and, consequently, a lower number of actual specimens. At the same time the number of cen- sored values (i.e. values below the detection limit) for less abundant larger species distinctly increases, with a negative effect on the accuracy of data (Sanford et al., 1993). To avoid such shortcomings it would be neces- sary to increase the number of specimens counted per sample. By adopting an intermediate sieve mesh (i.e. 106 µm) fewer censored values result, together with only slightly reduced percentages of small species. In order to investigate the relationship among planktonic foraminiferal species, in the present paper assemblages are investigated by means of a PCA. The PCA was first computed on the raw data and then by applying a logratio transformation (Aitchison, 1986; Kucera and Malmgren, 1998) to the original variables for closure correction. With regards to the foraminiferal species included in the analysis, Globoturborotalita tenella and Globoturborotalita rubescens were lumped together, while no distinctions were made about Neogloboquadrina pachyderma, since it was almost entirely represented by right coiled specimens. No rotations were applied to the components. Since in the o r i g i n a l d a t a t h e t h r e e m o s t a b u n d a n t s p e c i e s (Turborotalita quinqueloba, Globigerina bulloides and Globigerinoides ruber) constitute more than the 64% of total assemblages, it was preferred to operate on the correlation matrix rather than the variance-covariance matrix, in order to investigate relationships between less abundant species. The results obtained with logratio transformed data are less affected by the choi- ce of the matrix to be factored. The interpretation of assemblages is based on the present distribution of species (Bé & Torlderlund, 1971; Hembleben et al.1983; Pujol and Vergnaud Grazzini, 1995; among others). The analyses were performed with Systat. Figure. 1 - Location map. Ubicazione della carota G93-C9. Depths in Radiocarbon Calendar Core C9 Age source ages ages References (cm) (kyr) (kyr BP) 55 Ecozones 2-1 2.8 2.5 Capotondi boundary et al. (1999) 75 Ecozones 3-2 4.0 3.9 Capotondi boundary et al. (1999) 125 Ecozones 6-5 10.0 10.9 Capotondi boundary et al. (1999) 165 Ecozones 7-6 11.4 12.9 Capotondi boundary et al. (1999) 170 14C AMS dating 12.030±0.100 13.45 (intercept) Tab. 1. Age sources for Core G93-C9. AMS C14 dating were calibrated with CALIB 4.3 (Stuiver and Reimer, 1994). Riferimenti per l’inquadramento cronologico della carota G93- C9. La calibrazione delle età radiocarbonio è stata effettuata con CALIB 4.3 (Stuiver and Reimer, 1994). 33.. CCHHRROONNOOSSTTRRAA--TTIIGGRRAAPPHHIICCAALL FFRRAAMMEEWWOORRKK Only one AMS 14C dating obtained from radiocar- bon measurement of Globigerinoides spp., Globigerina bulloides and Orbulina shells, was available for core G93-C9. In order to improve the age model, some events recorded in planktonic foraminifera assembla- ges were tentatively correlated with the ecozone boun- daries of Capotondi et al. (1999). The age model is shown in Table 1. The climatostratigraphy reconstruc- ted for core G93-C9 by palaeontological proxies (Amore et al., 2000) is reported in Figure 2. The Last Glacial period (LG) is represented from the base of the core up to 205 cm. The deglaciation, in which events corresponding with the Bølling-Allerød and the Younger Dryas were identified, was recorded in the 205-125 cm interval. The upper part of the core represents the Holocene. 44.. RREESSUULLTTSS Percentages of planktonic foraminifera recorded in core G93-C9 are shown in Table 2. Tables 3a and 3b show the loadings for the components corresponding to eigenvalues greater than 1, computed respectively from the correlation matrix of the raw and the log transformed data. The most important difference between raw data and logratio data rests in the behaviour of N. pachyder- ma which in the first case, unlike the second, does not have significant loading on the first component. A minor 253Effects of Late Pleistocene - Holocene ... 2 6,93 19,1 41,9 7,59 0,33 1,32 0 4,29 0 2,97 8,91 0 1,32 5,28 10 4 21 47,3 5,67 0 2,33 0 1,33 0 5 8,33 0 1 4 20 1,97 24 33,2 6,25 1,64 0,33 0 13,8 0 1,97 4,61 0,66 1,97 9,54 30 8,12 27,9 35,4 6,82 0 3,25 0,65 1,95 0 2,6 1,95 1,95 2,92 6,49 40 4,26 12,8 33,4 5,9 0 0,66 0 10,8 0 5,57 5,25 1,31 5,57 14,4 50 27,2 10,6 27,2 4,32 3,65 8,31 0 6,31 0 2,33 1,99 2,33 1 4,65 60 9,24 19,8 34,3 6,6 0,33 2,31 0 1,32 0 4,95 2,97 12,2 1,32 4,62 70 12,6 27,2 27,5 6,62 0,99 3,31 1,66 0,99 0,33 1,99 2,65 6,62 1,32 6,29 80 5,94 13,2 30,4 11,6 12,9 2,31 1,32 3,63 0 4,62 0,66 1,65 2,64 9,24 90 13,4 21,2 22,7 14 14,3 2,49 0,62 1,56 0 2,8 0 0,93 1,25 4,67 100 9,78 19,9 26,8 18 8,2 0,95 0 2,21 0,95 2,52 0 1,58 0,63 8,52 110 17,4 17,1 22,9 11,3 5,12 4,1 3,41 2,39 2,39 0,68 1,02 1,02 0,34 10,9 120 10 20,7 35 12,3 5 1,33 1 1 0 1 3 0,33 0 9,33 130 26,2 43,7 6,47 9,39 6,47 1,29 0,32 0,65 0,32 0,32 0,32 0 0,32 4,21 140 19,3 19,6 23,5 11,8 8,5 3,59 1,96 0 0 0,33 0 0,98 0,33 10,1 150 27,4 18,2 18,5 7,96 3,18 5,73 8,28 1,27 1,59 0,96 0 0,32 0 6,69 160 24,5 21,2 15,4 13,7 5,23 1,96 3,59 2,61 0,33 0,33 0,65 0 0,65 9,8 170 32,7 14,7 18,7 15 6,67 2,67 0 2,67 0 2 0,33 0,33 0,33 4 180 16,8 22,7 25,3 12,8 1,32 2,63 1,32 1,64 0 2,3 0 0,99 0,33 11,8 190 15,4 19,7 33,1 13,1 0,66 0,98 2,95 0,98 1,31 3,28 0,33 0,66 0,66 6,89 200 20,7 17,7 23 11,7 8,33 1,67 1,33 2,33 0 0,67 0 0 0 12,7 210 28,9 28 9,21 6,58 10,5 5,59 1,32 3,62 1,32 0,66 0,66 0 0 3,62 220 30,8 17,5 8,9 2,05 3,42 9,93 4,45 6,85 5,48 0,68 0 0 0 9,93 230 32 27,5 12 2,27 1,29 5,5 5,5 0 2,59 0 0 0,97 0 10,4 240 33,1 20,4 15,3 5,73 10,5 2,55 2,55 0,96 0,64 0 0 0,32 0 7,96 250 46,5 32,9 1,99 2,99 7,31 1,33 2,33 0,33 1,66 0 0,33 0 0 2,33 260 38,4 22,9 4,52 1,94 1,61 7,42 4,84 4,19 3,87 0,32 0,32 0 0 9,68 270 37,3 26,9 3,57 1,3 5,84 5,84 6,82 0,32 2,27 0 0 0 0 9,74 280 48,4 22,5 3,27 3,59 4,9 0,65 1,96 1,31 0 0 0,33 0 0 13,1 290 42,3 28 6 2,33 4,67 3,33 4,33 0 3,33 0,33 0,33 0 0 5 300 46,4 18,8 0 2,96 1,97 7,89 2,96 1,97 4,93 0 0 0 0 12,2 310 36,2 32,1 2,52 2,83 1,57 6,29 7,23 0,94 1,57 0,31 0 0 0 8,49 320 39,4 38,1 1,28 3,21 7,37 1,92 2,24 0 0,96 0 0 0 0 5,45 330 54 27 1,67 1,67 5,67 4,33 0,33 0,33 2,33 0 0 0 0 2,67 340 31 44,8 5,36 1,53 1,15 5,75 0,77 1,92 2,3 0 0 0 0 5,36 350 34,4 34,1 1,3 1,3 5,19 7,79 6,17 0,97 1,3 0 0 0,32 0 7,14 360 42,7 42,7 1,95 0,65 1,95 2,28 0,98 0 1,3 0 0 0 0 5,54 370 31,9 37,2 7,89 2,84 7,57 5,36 1,26 1,26 0,32 0 0,95 0,63 0 2,84 380 51,5 24,1 0,98 0 1,95 3,58 2,61 0,65 2,28 0 0,33 0 0 12,1 390 34,6 35,9 5,08 0,95 4,44 3,81 7,94 1,59 0,95 0 0 0 0 4,76 400 50,8 31,8 0,66 0,98 5,57 1,31 0,98 0 1,64 0 0 0 0 6,23 410 42,5 31,9 2,66 1,66 6,98 2,66 2,33 1 2,99 0 0 0 0 5,32 420 19,4 37 1,42 2,84 7,11 3,79 1,9 1,42 3,32 0 0 0,47 0 21,3 430 36,5 24,3 2,63 1,64 3,29 15,5 6,25 0,66 3,62 0 0 0,33 0 5,26 440 36,9 39,5 2,33 1,99 7,97 2,66 1,33 0 2,99 0 0 0 0 4,32 450 37,3 38 0,95 2,22 6,65 6,01 0,95 0 3,16 0 0 0 0 4,75 460 31,1 26,2 1,97 5,9 7,54 6,89 0,66 1,31 5,25 0 0 0,66 0,66 11,8 470 47 22,8 0,99 0 2,98 10,6 3,97 1,32 3,97 0 0,33 0 0 5,96 Table 2. Percentages of planktonic foraminifera within Core G93-C9. Percentuali dei foraminiferi planctonici della Carota G93-C9. cm T u rb o ro ta lit a q u in q u e lo b a G lo b ig e ri n a b u llo id e s G lo b ig e ri n o id e s ru b e r G lo b o ro ta lia in fla ta N e o g lo b o q u a d ri n a p a ch yd e rm a G lo b ig e ri n ita g lu tin a ta N e o g lo b o q u a d ri n a d u te rt re i G lo b o tu rb o ro ta lit a G lo b o ro ta lia s ci tu la O rb u lin a u n iv e rs a G lo b o ro ta lia t ru n ca tu lin o id e s G lo b ig e ri n o id e s sa cc u lif e r G lo b ig e ri n e lla s ip h o n ife ra in d e te rm in a b le difference regards the third component obtained with the logratio transformations, which corresponds to the fourth component of the original data. In Figure 2 the scores of the first three components computed from the logratio transformed data are plotted along core G93- C9. The following discussion is based on the results achieved with the logratio transformation. 55.. DDIISSCCUUSSSSIIOONN The first three com- ponents, corresponding to eigenvalues greater than 1, take into account respectively 55.48%, 10.64% and 8.26% of total variance. The cumu- lative variance retained is 74.38% of the total variance. As shown in Table 3b, the first compo- nent clearly individualizes two groups of species, the first of which, charac- terised by high positive loadings, is constituted by warm water species. The second group consi- sts mainly of cold water species, with high negati- ve loadings. This structu- re suggests a relation between this component and sea surface tempera- tures (SST). It can be also observed that in core G93-C9 Globigerina bul- loides and Globigerinita glutinata are grouped with cold water species, while Globorotalia inflata is grouped with warm species, as a consequen- ce of its absence during the LG. The comparison of first component scores with algebraic climatic curves for pollen and cal- careous nannofossils (Figure 3), points out a good agreement in the general trends, suppor- ting the interpretation of the first component in terms of changes in SST. The scores of the first component are lower in the interval correspon- ding to the LG. The fol- lowing trend in PC scores points out changes in SST corresponding to the Bølling-Allerød and 254 Younger Dryas events of the last deglaciation. The Holocene is characterised by increasing values of first component scores. The second component is characterised by high loadings of grazing species, negative for N. pachyderma and G. inflata and positive for Globorotalia truncatulinoi- des (Table 3b). The abundance of N. pachyderma, which is usually considered a polar-subpolar species, has been related in the Mediterranean sea to the deve- lopment of eutrophic conditions and a deep chlorophyll maximum (DCM) (Rohling and Gieskies, 1989; Rohling, 1994). The present distribution within the Mediterranean sea of G. inflata appears to be related to a rather cool and deep mixing as well as high primary production levels (Pujol and Vergnaud Grazzini, 1995). Moreover although G. inflata is mainly a mesopelagic species, it reaches high abundances in shallow water when phyto- plankton blooms occur in the mixed layer (Pujol and Vergnaud Grazzini, 1995). In the fossil record this spe- cies reaches high abundances during glacial-interglacial transitions (Blanc-Vernet et al., 1984). As regards G. truncatulinoides, Pujol & Vergnaud Grazzini (1995) sug- gest that vertical mixing and winter convection are the primary factors controlling the distribution of this species within the Mediterranean. In the Gulf of Salerno G. truncatulinoides, although its percen- tages never exceed 13%, it is the only grazing species whose abundance increases during the last 5000 yr, when fossil assemblages evidence a decrease in palaeoproductivity (Buccheri et al., 2002). On the basis of the above mentioned remarks, the second component may be likely related to trophic conditions and convection in the water column. In practice the highest negati- ve scores reached during the last deglaciation- early Holocene interval (Fig. 2) could indicate the establishment of eutrophic conditions in the Gulf of Gaeta. On the contrary positive scores characterise the late Holocene, suggesting a decrease in palaeoproductivity. Besides the “open water” system of the modern Tyrrhenian sea has oligotrophic characteristics (Margalef et al. 1966). The opposite relationship between G. inflata and G. truncatulinoides may appear unexpected, as these species are often associa- ted within the Mediterranean sea (Pujol & Vergnaud Grazzini, 1995). The positive loading of G. truncatulinoides does not imply an oligo- trophic connotation for this species which does not reach high percentages in the late Holocene. Rather it evidences that during the early Holocene environmental conditions were unfavourable for this species. As regards G. inflata, its abundance during the last deglacia- tion has also been recorded in the shallow core C5 recovered in the Gulf of Gaeta at a depth of 111 m bsl (Amore et al. 2000). This means that during the last deglaciation, when the sea level was about 80-90 m lower than at present (Fairbanks, 1989), this species was also abun- dant in shallower water in the Gulf of Gaeta. This behaviour probably derived, in addition to advantageous conditions (such as the presence of transitional water masses), also from enhan- V. Di Donato ced primary productivity levels related to increased con- tinental run-off. This is also supported by the high per- centages of Braarudosphaera bigelowii, a calcareous nannofossil species related to turbidity of surface water and/or low salinity (Müller, 1979), recorded within the same interval (Esposito, 1999; Amore et al., 2000). The composite effect on sea surface temperature changes and trophic regime is highlighted in a plot of core sam- ples on the space defined by first and second compo- nents (Fig. 4). The plot allows two groups of samples to be distinguished, the first of which comprises samples from the LG and the deglaciation, with an upward trend towards negative scores for the second component. The second group is represented by late Holocene samples, with positive scores for both components. The third component is characterised by high posi- tive loadings of Globigerinoides sacculifer and negative loadings of Globoturborotalita and does not have a clear interpretation. The results of the standardized PCA seem to simply highlight the single peak of G. sacculifer within core G93-C9. At present, this tropical species is uncommon within the Mediterranean sea, apart from the Gulf of Lion where this species is very abundant at the PC1 PC2 PC3 PC4 Turborotalita quinqueloba -0,883 0,015 -0,105 0,062 Globorotalia scitula -0,686 0,416 0,192 0,077 Globigerina bulloides -0,563 -0,190 -0,671 0,017 Neogloboquadrina dutertrei -0,544 0,336 0,362 -0,219 Globigerinita glutinata -0,517 0,571 0,404 -0,131 Neogloboquadrina pachyderma -0,180 -0,742 0,351 0,205 Globigerinoides sacculifer 0,456 0,121 -0,127 -0,724 Globoturborotalita spp. 0,543 0,421 0,142 0,554 Globorotalia inflata 0,631 -0,521 0,434 -0,087 Globorotalia truncatulinoides 0,723 0,347 -0,317 0,146 Globigerinella siphonifera 0,779 0,238 -0,053 0,203 Orbulina universa 0,909 0,133 0,029 -0,118 Globigerinoides ruber 0,944 0,025 0,087 -0,111 Eigenvalues 5,909 1,857 1,252 1,044 Percent of Total Variance Explained 45,452 14,288 9,633 8,032 Table 3a - Component loadings, eigenvalues and percent of total variance for PCA computed from raw data. Component loadings, autovalori e percentuale della varianza totale per l’anali- si dei componenti principali relativa ai dati percentuali PC1 PC2 PC3 Turborotalita quinqueloba -0,913 -0,025 0,045 Globorotalia scitula -0,867 0,210 0,014 Neogloboquadrina dutertrei -0,799 -0,041 -0,066 Globigerinita glutinata -0,735 0,279 0,061 Globigerina bulloides -0,692 0,076 0,265 Neogloboquadrina pachyderma -0,620 -0,584 -0,163 Globoturborotalita spp. 0,493 0,351 -0,686 Globigerinoides sacculifer 0,538 -0,069 0,638 Globorotalia inflata 0,596 -0,678 -0,114 Globorotalia truncatulinoides 0,643 0,488 0,025 Globigerinella siphonifera 0,806 0,183 0,271 Globigerinoides ruber 0,829 -0,228 -0,042 Orbulina universa 0,904 0,020 -0,011 Eigenvalues 7,083 1,383 1,074 Percent of Total Variance Explained 54,481 10,638 8,259 Table 3b- Component loadings, eigenvalues and percent of total variance for PCA computed from logratio transformed data. Component loadings, autovalori e percentuale della varianza totale per l’anali- si dei componenti principali dei dati sottoposti a una trasformazione logratio. 255 end of summer (Pujol and Vergnaud-Grazzini, 1995). According to Bé and Tolderlund (1971) this species is more abundant in sea surface waters with intermediate salinities, while the maximum abundance of G. ruber occurs at more extreme salinities, as in the present-day Mediterranean sea. The highest abundances of G. sac- culifer and consequently highest positive scores of the third component in core G93-C9 are reached around 3.5 kyr. This peak, which has been recognised throughout the western Mediterranean (Capotondi et al. 1999) coin- cides, in the Gulf of Salerno, with the re-entry of Styliola subula, a halophile pteropod. Buccheri et al. (2002) relate this event to the increase in sea surface salinity following the end of the sapropel S1 stagnation phase. The pre- sent distribution of G. sacculifer suggests that its abun- dance peak could be related to an enhanced influx of Atlantic waters after the end of the S1 stagnation phase, with sea surface salinities higher compared with the sta- gnation phase, but lower in comparison to the present. 66.. CCOONNCCLLUUSSIIOONNSS The results of a PCA of planktonic foraminiferal assemblages of Gulf of Gaeta, applied first to the raw percentages data and then to the logratio transformed data, differ primarily in the behaviour of N. pachyderma, which has no significant loading on the component rela- ted to SST if the PCA is computed from the raw data. This result is analogous to that obtained by Blanc- Vernet and Sgarrella (1989) in Tyrrhenian and Adriatic Sea cores. However if data are corrected for closure the same species achieves a negative loading on the same component. Apart from the strong influence exerted by chan- ges in SST, hydrography and trophic regime also play an important role in determining composition of assem- blages. In particular second component scores indicate the establishment of eutrophic conditions during the last deglaciation and the early Holocene which allowed G. inflata to expand its distribution toward shallow waters of the Gulf of Gaeta. By contrast the Late Holocene is cha- racterised by a tendency towards the establishment of oligotrophic conditions. AACCKKNNOOWWLLEEDDGGEEMMEENNTTSS I wish to thank Prof. Giuliano Ciampo and dr. Paola Esposito for their reviews of the work and their constant support. Thanks are also due to an anonymous reviewer for his critical and constructive comments. Effects of Late Pleistocene - Holocene ... Figure 2 - Stratigraphy of core G93-C9 and scores of the first two component of foraminiferal assemblages. Carota G93-C9. Stratigrafia, climatostratigrafia e punteggi dei primi due componenti principali delle associazioni a foraminiferi planctonici. 256 V. Di Donato Figure 4 - Plot of core samples on the space defined by first and second components. Numbers refer to depth in core G93-C9. Grafico dei campioni della carota G93-C9 nello spazio definito dai primi due componenti. Figure 3 - First component scores of planktonic foraminiferal assemblages compared with algebraic climatic curves for pollen and cal- careous nannofossils (from Amore at al. 2000). Raffronto tra i punteggi del primo componente delle associazioni a foraminiferi planctonici e le curve climatiche algebriche di pollini e nannofossili calcarei. First component S e co n d c o m p o n e n t RREEFFEERREENNCCEESS Aitchison, J., 1986. The statistical analysis of composi- tional data. Chapman & Hall, New York, 416 pp. Aritzegui, D., Asioli, A., Lowe, J.J., Trincardi, F., Vigliotti, L., Tamburini, F., Chondrogianni, C., Accorsi, C.A., Bandini Mazzanti, M., Mercuri, A.M., Van der Kaars, S., McKenzie, J.A. and Oldfield, F., 2000. Paleoclimate and the formation of sapropel S1: inferences from Late Quaternary lacustrine and marine sequences in the central Mediterranean region. Palaeogeogr. Palaeoclimatol. Palaeocol. 158, 215-240. Amore, F.O., Ciampo, G., Di Donato V., Esposito, P., Russo Ermolli, E. and Staiti, D., 2000. An integra- ted micropalaeontological approach applied to Late Pleistocene-Holocene palaeoclimatic and palaeoenvironmental change (Gaeta Bay, Tyrrhenian Sea). In: Hart, M.B. (Editor), Climates: Past and Present. Geological Society, London, special Publications, 118811: 95-111. Bé A.W.H. and Tolderlund D.S., 1971. Distribution and ecology of living planktonic foraminifera in surface waters of the Atlantic and Indian Oceans. In: B.M. Funnell and W.R.Riedel (Editor), The Micropaleontology of the Oceans. Cambridge University Press, London, 105-149. Blanc-Vernet L. and Sgarrella F., 1989. Assemblages de foraminiferes planctoniques en Mer Tyrrhénienne et en Mer Adriatique pendant les derniers 130000 Ans. Implications Paléoécologiques. Paleogeogr., Paleoclimatol., Palaeoecol., 7766: 107-124. Blanc-Vernet L., Sgarrella F. et Acquaviva M., 1984. Évènements climatiques,hydrologie et Foraminifères en Méditerranée au Quaternarie recent. Bull. Soc. géol. France, XXXXVVII,, 77:1235- 1243. Buccheri G., Capretto G., Di Donato V., Esposito P., Ferruzza G., Pescatore T., Russo Ermolli E., Senatore M.R., Sprovieri M., Bertoldo M., Carella D. and Madonia G., 2002. A high resolution record of the last deglaciation in the southern Tyrrhenian Sea: environmental and climatic evolution. Marine Geology, 118866: 447-470. Capotondi, L., Borsetti, A. and Morigi, C., 1999 - Foraminiferal ecozones, a high resolution proxy for the late Quaternary biochronology in the central Mediterranean Sea. Marine Geology, 115533 ((1144)): 253-274 Esposito, P., 1999. Le associazioni a nannoplancton calcareo del Pleistocene superiore-Olocene dei Golfi di Gaeta, Pozzuoli e Salerno (Mar Tirreno). Biocronologia e ricostruzione paleoclimatica. Tesi di Dottorato, Università Federico II, Napoli, 164 pp. Fairbanks R.G., 1989. A 17,000 year glacio-eustatic sea level record: influence of glacial melting rates on the Younger Dryas event and deep ocean circula- tion. Nature, 334422: 637-642. Hembleben Ch., Spindler M. and Anderson O.R., 1989, Modern Planktonic foraminifera. Springer-Verlag Imbrie, J and Kipp, N.G. 1971, A new micropaleontologi- cal method for quantitative paleoclimatology: Application to a Late Pleistocene Caribbean core. In: K. Tukerian (editor), The Late Cenozoic Glacial Ages. Yale Univ. Press., New Haven, 71-191. Kallel, N., Paterne, M., Labeyrie, L., Duplessy, J.C. and Arnold,M., 1997. Temperature and salinity records of the Tyrrhenian Sea during the last 18 000 years. Paleogeogr. Paleoclimatol. Palaeoecol. 113355:: 97-108. Kucera M. and Malmgren B.A., 1998. Logratio transfor- mation of compositional data – a resolution of the constant sum constraint. Marine Micropaleontology, 3344: 117-120 Margalef R., Herrera J., Steyaert M. et Steyaert J., 1966. Distribution et charactéristiques des com- munautés phytoplanctoniques dans le basin Tyrrhénien de la Méditerranée en function des fac- teurs ambiants et à la fin de la stratification estiva- le de l’année 1963. Bull. Inst. r. Sci. nat. Belg., 4422: 1-56 Müller C., 1979. Les nannofossiles calcaires. La Mer Pelagienne, 210-220. Pujol C. and Vergnaud-Grazzini C.,1995. Distribution patterns of live planktic foraminifers as related to regional hydrography and productive systems of the Mediterranean Sea. Marine Micropaleontology, 2255: 187-217 Rohling, E.J., 1994. Review and new aspects concer- ning the formation of eastern Mediterranean sapropels. Marine Geology. 112222: 1-28. Rohling E.J. and Gieskes W.W.C., 1989. Late Quaternary changes in Mediterranean intermedia- te water density and formation rate . Paleoceanography, 44: 531-545. Sanford R.F., Pierson C.T. and Crovelli R.A., 1993. An objective replacement method for censored geo- chemical data. Mathematical Geology, 2255 ((11)):: 59- 79. Stuiver, M. and Reimer, P. J., 1993. Extended 14C data- base and revised CALIB radiocarbon calibration program. Radiocarbon,, 3355: 215-230. 257Effects of Late Pleistocene - Holocene ... Ms. ricevuto il 30 luglio 2002 Testo definitivo ricevuto il 7 gennaio 2003 Ms. received: July 30, 2002 Final text received: January 7, 2003