SESS 0_retro e indice edi.pub Il Quaternario Italian Journal of Quaternary Sciences 24, (Abstract AIQUA, Roma 02/2011), Congresso AIQUA Il Quaternario Italiano: conoscenze e prospettive Roma 24 e 25 febbraio 2011 STATISTICAL ANALYSIS OF PLEISTOCENE CALCAREOUS MICROPLANKTONIC ASSEMBLAGES: AN ATTEMPT FOR UNDERSTANDING THE VARIATIONS OF ENVIRONMENTAL PARAMETERS Claudia Lupi 1, Miriam Cobianchi 1, Luciani Valeria 2 & Manuela Bordiga1 1 Dipartimento di Scienze della Terra, Università di Pavia, Pavia 2 Dipartimento di Scienze della Terra, Università di Ferrara, Ferrara Corresponding author: C. Lupi. ABSTRACT: Lupi C. et al., Statistical analysis of Pleistocene calcareous microplanktonic assemblages: an attempt for understanding the variations of environmental parameters (IT ISSN 0394-3356, 2011) The application of a multivariate statistical approach to a database concerning the occurrence of the two major calcare- ous planktonic groups (calcareous nannofossils and foraminifera) provides a potential additional instrument to identify the variation over time of the sea surface water parameters RIASSUNTO: Lupi C. et al., Analisi statistica di associazioni microplanctoniche del Pleistocene: un tentativo per la com- prensione delle variazioni dei parametri ambientali (IT ISSN 0394-3356, 2011) L’applicazione di metodi statistici quali l’analisi multivariata su di un database concernente la distribuzione dei due princi- pali gruppi di microplancton calcareo (nannofossili calcarei e foraminiferi) fornisce un ulteriore strumento per identificare le variazioni nel tempo dei principali parametri delle acque superficiali. Key words: calcareous nannofossils, planktonic foraminifera, Principal Component Analysis Parole chiave: nannofossili calcarei, foraminiferi planctonici, Principal Component Analysis 1. INTRODUCTION In the view of the paleoenvironmental reconstruc- tions based on the ecological meaning of fossil species, statistical methods are here used to in- vestigate the ecological preferences of Pleistocene calcareous micro-planktonic species and assem- blages (nannofossils and foraminifera) of the core MD 97-2114 (SW Pacific Ocean) in order to high- light the variability through time of the main sea surface water parameters. Several recent studies, in fact, have demonstrated the effectiveness of multivariate statistical approaches to the Pleisto- cene fossil record in detecting paleoenvironmental changes or paleoecology preferences of extinct species (e.g., MALMGREN & KENNETT, 1976; BOE- CKEL & BAUMANN, 2004; MAIORANO et al., 2009). This approach highlights characteristic assem- blages in order to determine, from a numerical point of view, the relations among the species in- volved in the assemblages. However, the knowl- edge of the ecology for some species presents certain ambiguities, making at times still problem- atical the interpretation of the past environments. Starting from our micropaleontological database we provide statistical and graphical treatments from coccolithophorids and planktonic foraminifera firstly processing the data separately secondly making a composite database. We process to- gether the nannofossils and the planktonic fo- raminifera because these groups characterize the surface water ecosystem. The application of the statistical methods simultaneously to the two major calcareous plankton groups is an innovative step for obtaining combined and crossed information. This approach may represent a potential additional instrument for identifying the puzzling sea surface water parameters whose variations have, over time, limited the planktonic foraminiferal and coc- colithophorid distribution and abundance fluctua- tions. 2. MATERIAL AND METHODS Our case history is the core MD 97-2114 (42° 22.32’S; 171°20.42’W) which was recovered from an IMAGES cruise (International Marine Past Global Change Study) at a water depth of 1,935 m on the north-eastern slope of the Chatham Rise (Southwestern Pacific Ocean). The succession is 28 m thick and exhibits pelagic and hemipelagic sediments. The bio-magnetostratigraphic calibra- tion of the core is accurately constrained (LUPI et al., 2008; VENUTI et al., 2007). A detailed benthic oxygen isotope stratigraphy was recently per- formed by COBIANCHI et al. (in prep.). The age model indicates that the studied succession shows a continuous sedimentary record for the past 1.07 Ma and spans from the Marine Isotope Stage (MIS) 1 to the MIS 29 with an average sediment accumulation rate of ca 2.6 cm/kyr. The micropaleontological content was analysed in 113 samples and the data have been collected using standard methodologies (LUPI et al., 2008) and then processed utilizing the basic and useful statistical methods of mono and multivariate analy- sis of the PAST software (PAleontology STatistic, HAMMER et al., 2001) to perform a Principal Com- 212 - 214 213 ponent Analysis (PCA). The PCA is a procedure for finding hypothetical components which account for as much of the vari- ance in a multidimensional database (DAVIS, 1986; HARPER, 1999). These new variables are liear combinations of the original variables, in our case the nannofossils and planktonic foraminifera gen- era and species. The PCA may be used for simple reduction of the data set to only two/three variables (the two/three most important components), that are correlated with some other underlying vari- ables. For ecological data, the components might be physical parameters (for example water tem- perature or nutrient availability). Moreover, it is possible to verify what degree (loading) the differ- ent original variables (in our case, the taxa) enter into the new variable (the components). These loadings are important when we interpret the meaning of the components. In fact, the loading diagrams reveal the affinities of the taxa with re- spect to the unknown components that influence the planktonic foraminiferal and nannofossil distri- bution in the core. Potentially, using the taxa with known ecology, we can interpret the components as environmental parameters. 3. RESULTS The good to excellent microfossil preservation throughout the core allowed us to identify 48 spe- cies among the nannofossils and 30 species among the planktonic foraminifera. The nannofos- sil assemblages are dominated by the small Gephyrocapsa KAMPTNER, and subordinately by Calcidiscus leptoporus (MURRAY & BLACKMAN) (abundance up to 25% of assemblages), Cocco- lithus pelagicus (WALLICH) (up to 40%), Pseu- doemiliania lacunosa (KAMPTNER), (up to 40%), medium Gephyrocapsa KAMPTNER (up to 40%), Reticulofenestra minuta ROTH (up to 30%), Emil- iania huxleyi (LOHMANN) (up to 20%), Heli- cosphaera carteri (WALLICH) (up to 20%). Since small Gephyrocapsa represents on average almost the 50% of the nannofossil assemblage, this morphogroup tends to obscure the fluctuations within the less common species. Hence, for eco- logical interpretation and multivariate analysis, the small Gephyrocapsa was removed from the quanti- tative analysis. The PCA highlights the two major components driving the composition of assemblages: the first one explains the 51% and the second the 23% of variance. Component 1 is positively loaded by C. pelagicus (0.85), and negatively by P. lacunosa (- 0.44). Component 2 is positively loaded by P. la- cunosa (0.69), C. pelagicus (0.44) and by R. minuta (0.45). The most common planktonic foraminiferal species is Globoconella inflata D’ORBIGNY (34%) followed in abundance by Globigerina bulloides D’ORBIGNY, (19%), Neogloboquadrina incompta (CIFELLI) (11%) and Neogloboquadrina dutertrei D’ORBIGNY (8%). The PCA highlights that the two major com- ponents explain, respectively, the 32% and the 27% of variance. Component 1 is positively loaded by G. inflata (0.45) and negatively loaded by G. bulloides (-0.40). Component 2 is positively loaded by Truncorotalia truncatulinoides (D’ORBIGNY) (0.50) and G. inflata (0.45), and negatively by Truncorotalia crassaformis (GALLOWAY & WISSLER) (-0.55) For the composite database, through the PCA we identify three major components. They account for the 33.5%, 17% and 12% of the variance, respec- tively. Component 1 is positively loaded by Pseu- doemiliana lacunosa (0.49) and negatively loaded by Coccolithus pelagicus (-0.66). Component 2 is positively loaded by C. pelagicus (0.47), and P. lacunosa (0.41) and negatively by Globoconella inflata (-0.55), Finally, none species loads posi- tively component 3, whereas it is negatively loaded by C. pelagicus (-0.46) and G. inflata (-0,69). 4. DISCUSSION AND CONCLUSIONS Using the ecological meaning of the taxa, their af- finities and relations with the unknown principal components (PCs), we can interpret these latter as environmental parameters. However, the ecologi- cal meaning of the some species is still matter of debate (for example Pseudoemiliania lacunosa) due to scarce data available or because they are extinct. Checking the effectiveness of the statistical ap- proach, we observe that the processing of each microfossil group separately reveals two major PCs driving the associations. The PCs found for nannofossils and planktonic foraminifera can be different but they have to be consistent because reflecting the same surface water characters. For each processing we identify a great number of species that makes the discussion very complex and full of ambiguities. Through the use of a com- posite database we reduce the PCs from four to three and the number of species involved de- creases as well. However, these species were al- ready present in the first elaboration confirming that the composite approach could be valid. Among the three parameters, one (PC1) mainly influences only the nannofossils. In the interpreta- tions of the data we have to take in account this constrain searching for a parameter that could in- fluence primarily the nannofossils (phytoplankton) with respect to the foraminifera (zooplankton), as can be for example the light. The final interpreta- tion will be based on the ecological meaning of species with well-known preferences thus helping to clarify the environmental behavior of different Statistical analysis of Pleistocene calcareous ... 214 Ms. received: Testo ricevuto il fossil groups. As an example, in our case history the oligotrophic foraminiferal species G. inflata, negatively loaded with respect to PC2, appears to validate the hy- pothesis about the meaning of a supposed eutro- phic nannofossil species (C. pelagicus), positively loaded with respect to the same component. In conclusion, we proposed here, for the first time, the application of statistical methods simultane- ously to two microfossil groups. The data have to come from the same samples and to be measured in the same units (percentage). This approach shows apparently some vantages. Since it is just an attempt, the idea will be checked in more case- histories adding at the database (matrix) more in- dependent data from isotopic or sedimentologic analyses to verify and corroborate the final inter- pretations. REFERENCES BOECKEL B. & BAUMANN K.-H. (2004) - Distribution of coccoliths in surface-sediments of the south-eastern South Atlantic Ocean: ecology, preservation and car- bonate contribution. Marine Micropaleontology, 51, 301–320. DAVIS J.C. (1986) - Statistics and Data Analysis in Geol- ogy. John Wiley and Sons. HAMMER Ø., HARPER D.A.T. & RYAN P.D. (2001) - PAST: Paleontological Statistics software package for educa- tion and data analysis. Palaeontologia Electronica, 4 (1) http:// palaeo-electronica. org/2001_1/past/issue1_01.htm HARPER D.A.T. (1999). Numerical Palaeobiology. John Wiley and Sons LUPI C., LUCIANI V. & COBIANCHI M. (2008) - Integrated calcareous nannofossil and planktonic foraminiferal bioevents of the last 1.07 Ma: a case study from the East New Zealand Pacific Ocean. Micropaleontology, 54 (5), 463-476. MAIORANO P., MARINO M. & FLORES J.A. (2009) - The warm interglacial Marine Isotope Stage 31: Evidences from the calcareous nannofossil assemblages at the Site 1090 (Southern Ocean). Marine Micropaleontol- ogy, 71, 166-175. MALMGREN B. & KENNETT J.P. (1976) - Principal compo- nent analysis of Quaternary planktic foraminifera in the Gulf of Mexico: Paleoclimatic applications. Marine Micropaleontology, 1, 299-360. VENUTI A., FLORINDO F., MICHELL E. & HALL I. R.(2007) - Magnetic proxy for the deep (Pacific) western bound- ary current variability across the mid-Pleistocene cli- mate transition. Earth and Planetary Science Letters, 259, 107-118. January 13, 2011 13 gennaio 2011 C. Lupi et al.