Compositional Data Analysis as a Potential Tool to Study the (Paleo)ecology of Calcareous Nannoplankton from the Central Portuguese Submarine Canyons(W off Portugal)

Abstract

Submarine canyons are deep and steep incisions on the continental margins. The physical forcing mechanisms linked with these marine systems, such as the enhancement of upwelling and bottom sediment resuspension, are expected to provide a nutrient source that will increase phytoplankton density (Hickey, 1995, Kampf,2006). Coccolithophores are the predominant phytoplanktonic group within the calcareous nannoplankton and their sensitivity to a variety of surface water environmental parameters makes them important markers of oceanographic processes and proxies of sea surface water masses and temperatures, productivity and past climate changes (e.g. Ziveri et al., 2004; Silva et al., 2008). In the present work we propose to test compositional analysis (Buccianti & Esposito, 2004; Pawlowsky-Glahn & Egozcue, 2006) as a tool to: a) achieve a clearer distinction between opportunistic coastal-neritic species (r-strategists) and typical oceanic species (k-strategists) in the central Portuguese margin, and b) to identify a coccolith assemblage that might reflect favorable environmental conditions found in the vicinity of the canyon that promotes the productivity of calcareous nannoplankton. Our main difficulty will be in distinguishing the ecological signal from the effects of other environmental factors mentioned before (i.e. advection, dissolution, bottom resuspension). One way to infer the species’ ecological inter-relationships is by determining species relative percentages. The main concern is how the closure problem and the inconsistency of percentage determinations will affect our results. Compositional analysis was designed to provide more reliable and thus representative results, since the inference made on the coccolith assemblage features from which the data are drawn is correctly performed from a theoretical point of view (Buccianti & Esposito, 2004). Here, we present the first insights from applying compositional data analysis to coccolith assemblages from 85 surface sediment samples collected from the central Portuguese margin