Statistical Transfer Function Research Articles

Freshwater Testate Amoebae Communities from Île de la Possession, Crozet Archipelago, Subantarctica

An ecological study of the freshwater testate amoebae (Protozoa, Rhizopoda) on the subantarctic island Île de la Possession (Crozet Archipelago) revealed 85 taxa, belonging to 21 genera. Twenty-two taxa belonged to the genus Difflugia, typical for freshwater habitats whereas the genus Trinema showed the highest relative abundance. A cluster analysis revealed two communities: a Difflugiella crenulata assemblage and a Trinema lineare–T. enchelys assemblage. These communities, together with the results of a RDA analysis, represented a clear geographical separation on the island. The Difflugiella crenulata assemblage typified locations, with a high specific conductance and neutral to alkaline pH values, on the western part of the island. Aquatic habitats in the larger valleys on the eastern side of Île de la Possession, with a low conductivity and slightly acid pH values, were characterized by the Trinema assemblage. Weighted averaging and calibration were used to develop a statistical transfer function to infer the pH of freshwater bodies from the testate amoebae assemblages. The model is usable over a pH range of 5 to 9.

Soil diatom communities from Ile de la Possession (Crozet, sub-Antarctica)

An ecological study of soil diatoms on Ile de la Possession (Crozet Archipelago, sub-Antarctica) was carried out during the austral summer of 1998/1999. Both diatom and chemistry data were collected. A highly diverse diatom flora of 230 taxa, belonging to 39 genera were identified from 104 samples. Several of them are considered to be new. The most abundant genera were Diadesmis, Achnanthes and Pinnularia. Principal Component Analysis (PCA) was used to classify the samples based on their chemical characteristics. Moisture and nutrients were the main factors separating three groups of samples: dry fellfield soils, soils influenced by marine animals and mesic valley soils. Species assemblages correspond well with this division. Forward Selection with Monte Carlo permutations was used to reduce the number of significant variables. The analysis selected moisture, phosphate and sulphate concentrations. Based on the analysis, the same three groups of the PCA were found. Weighted averaging and calibration were used to develop a statistical transfer function to infer the moisture content of soils from the diatom assemblages.

Moss diatom communities from Ile de la Possession (Crozet, Subantarctica) and their relationship with moisture

More than 190 taxa, belonging to 33 genera, were observed on mosses collected on Ile de la Possession, Crozet Archipelago, Antarctica. Six assemblages could be identified. Habitat type and moisture content of the samples appeared to be the major factors in separating the communities. It was clear that diatom communities growing on mosses found in rivers and brooks are different from assemblages on rocks and in stagnant pools. The Diadesmis contenta assemblage was bound to dry mosses growing on rocks whereas the Eunotia paludosa var. paludosa community preferred acid stagnant pools, situated in the larger valleys. The other four assemblages are typical for circumneutral waterbodies, with the Fragilaria maillardii and the Fragilaria vaucheriae-Fragilaria germainii assemblages found in running water and the Fragilaria pinnata var. pinnata and the Achnanthes confusa-Achnanthes aueri assemblages in more stagnant water. Weighted averaging and calibration were used to develop a statistical transfer function to infer the moisture content of moss samples from the diatom assemblages.

Testing the effects of tropical temperature, productivity, and mixed‐layer depth on foraminiferal transfer functions

Statistical transfer functions relate living planktonic foraminiferal species of the central equatorial Pacific to measured sea surface temperature, integrated primary productivity, and mixed‐layer depth. The faunal estimates successfully reconstruct latitudinal patterns observed in both warm (El Niño, February–March 1992) and cool (La Niña, August–September 1992) seasonal settings. Predictions of mixed‐layer depth appear to be unbiased by temperature or productivity in our data set but tend to underestimate deep mixed layers. Interactions between productivity and temperature, perhaps through their common influence on respiration and growth rates, bias foraminiferal transfer functions for both properties. Paleoceanographic estimates may be improved by accounting for such biological processes that translate the environment into a faunal response preserved in the geologic record.

Precession Control on Sedimentary Burial of Organic Matter and Redox Sensitive Metals in the Tropical Indian Ocean

The total organic carbon (TOC) content of oceanic sediments and the total concentrations of several biomarkers are commonly used as qualitative palaeoproductivity proxies. However, fluctuations of these proxies may also be influenced by the variability of secondary processes taking place in the water column and in surface sediments. Past redox conditions of bottom and sediment pore waters are important parameters that can be studied by measuring the concentrations of redox-sensitive trace metals in the authigenic phases of the sediment. Here, we present a 330 kyr record of organic carbon, total alkenone and authigenic trace metal (U, Cd and Mo) in order to better understand how primary productivity and secondary physico-chemical processes are linked to each other and could explain the TOC variability. Core MD900963 is a 54 m long piston core recovered in the eastern part of the tropical Indian Ocean near the Maldives platform (05~ 73~ 2450 m water depth). At this location the present day bottom water oxygen is about 130 lamol/1 and the primary productivity is about 160 gCm-Zyr -1 (Antoine et al., 1996). Core MD900963 is well dated by orbital tuning of a high resolution 51So record measured on planktonic foraminifera (Bassinot et al., 1994). The studied first 16 meters of the core span the last 330 kyr which translates to an average s e d i m e n t a t i o n ra te o f a b o u t 5 c m / k y r . Palaeoproductivity was estimated from independent biological indices based on statistical transfer functions using coccoliths or planctonic foraminifera distributions (Beaufort e t al., 1997; Cayre et al., 1998). These two records suggest that during a typical 22 kyr cycle the primary productivity varied by roughly a factor of two: between c. 200-300 gCm-2yr -1 during glacial stages (high 5180) and c. 100-150 gCm-2yr 1 warm interstadials (low 8180). The bulk sediment concentrations (in ppm) of U, Mo, Cd and Th were measured by isotope dilution ICP-MS following the methods of Zheng et al. (1998). During the course of our measurements we estimate the reproducibility to about 3 -5% based on replicate analyses of a sediment standard. Authigenic U concentrations were obtained by subtracting a terrigenous component based on the Th concentrations and an aragonitic component based on X-ray analyses of the sediment fine fraction (Haddad, 1994). In general both corrections are small, i.e. less than 10% of the total U concentrations. On the same samples we measured the concentrations of TOC (wt.%) and of C37 alkenones (Tot-C37alk in lag/g), which are synthesized by several species of coccolithophorids. Both proxies were measured by gas chromatographic techniques (see Verardo et al., 1990 for TOC and Sonzogni et al. , 1997 for alkenones). By replicating measurements on sediment standards we estimate that reproducibilities for TOC and Tot-C37alk are on the order of 5 and 15%, respectively. The TOC values are rather low and vary between 0.25 and 0.85% while the Tot-C37alk exhibit larger variations between 0.2 and 7 lag@ Both records are very well correlated and are dominated by a clear cyclicity on the order of 22 kyr, with high concentrations during glacial stages (high 8180) and low values during warm interstadials (low 8180). Authigenic U concentrations range between 0.5 and 5.5 ppm, those of Cd vary between 0.3 and 1.9 ppm and those of Mo between 0.2 and 1.2 ppm which are comparable to the range of values reported by Rosenthal et aL (1995) and Crusius et al. (1996). Authigenic metals concentrations are also characterized by large and systematic fluctuations with a clear 22 kyr cyclicity and are positively correlated with records of TOC and Tot-C37alk. The clearest correlation with biomarkers and the strongest 22 kyr cyclicity are observed for the U record. This is partly due to its rather high values (about 2.5 ppm on