Formation Holocene climate history of the Nunatsiavut (northern Labrador, Canada), tutoriel & guide de travaux pratiques en pdf.

Methodology

Sampling

Sampling in Nachvak fjord was carried out in November 2009 during leg4b of the ArcticNet campaign onboard Canadian Coast Guard Ship Amundsen. A long sedimentary sequence (812 cm) was collected at station 602 (59.5265°N, 63.8784°W) using a piston corer triggered by a gravity corer. This piston-core was cut into eight sections of approximatively 100 cm each for easiest storage and transport. The eight sections were described and sampled at the Laboratory of Marine Palynology, ISMER-UQAR (Institut des sciences de la mer, Université du Québec à Rimouski), in Rimouski (Québec). The geophysical analyses (density, magnetic susceptibility) were performed at 0.5 cm intervals for each section. The percentages of organic matter and water content were measured every 1 cm. The grain-size and microfossil analyses (dinocysts, pollen and spores, Halodinium sp., foraminifer linings and pre-Quaternary palynomorphs) were performed at 10 cm intervals for each section in the Aquatic Paleoecology Laboratory (APL – Laval University, Quebec City). The palynological analyses were performed using the standard method described in Rochon et al. (1999) and Richerol et al. (2008a,b; 2012).

Stratigraphic analysis

Grain-size analysis
The particule-size composition of the core was determined at 10 cm intervals using the laser granulometer Horiba® LA950v2. The samples were prepared following the protocol from the Laboratory of Geomorphology and Sedimentology at Laval University (Québec, QC). The free software Gradistat v4 (Blott and Pye, 2001) was used to statistically separate the different grain sizes and their percentages, as well as calculate standard statistical parameters (mean, median and sorting) using the mathematical “method of moments” (Krumbein and Pettijohn, 1938).
Multi-sensor core logger (MSCL)
Low-field volumetric magnetic susceptibility (K), density (d) and fractional porosity (FP) were measured at the Sedimentary Paleomagnetism and Marine Geology Laboratory of ISMER-UQAR (Rimouski, QC) using a GEOTEK™ MSCL at 0.5 cm intervals for each core.
The volumetric magnetic susceptibility (K) is a dimensionless parameter measuring the magnetisation of a volume induced by applying a magnetic field to a material. It is used for marine sediments because variations in density in sediment cores are relatively small (Gunn and Best, 1998). Because of the near zero magnetic susceptibility of the pore-water, the porosity can have an impact on the volumetric magnetic susceptibility by diluting the amount of magnetic carrier mineral in the core and thus underestimating the total magnetic susceptibility. We therefore quantified this effect (Kp) by assuming zero susceptibility for the pore volume (Niessen et al., 1998) :
Kp = K/(1-FP) [1]

Chronological framework

A chronology was established based on 20 AMS-14C ages obtained from shell fragments found at various depths throughout the core. The preparation of the samples was performed at the Radiochronology Laboratory of the Center for Northern Studies (Laval University, Québec City, Canada) and the analyses at the Keck Carbon Cycle AMS Facility (Earth System Science department, University of California at Irvine, USA). The calibration of conventional AMS-14C dates into cal BP dates was done using the software CALIB 7.0 online (Stuiver et al., 2005) and MARINE13 (Reimer et al., 2013). The station 899 in Komaktorvik Fjord (Labrador: 59.28°N, -63.73°W), being the nearest of our coring site, was chosen to estimate a value for marine reservoir corrections (ΔR = 150 ± 40) (Table 4.1).
Table 4.1 : Table detailing the statistical data of the calibration of the AMS-14C dates obtained throughout the core.

Palynomorphs preparation

Fossil palynomorphs preparation (pollen, spores, dinocysts, acritarchs, freshwater and pre-Quaternary palynomorphs) followed standard procedures as outlined in Richerol et al. (2008a,b, 2012). Concentration of the palynomorphs (specimens cm-3) was calculated.
Palynomorphs were counted using a transmitted-light microscope (Leica DM5500B) with a magnification factor of 400× at APL (UL, Quebec City). A minimum of 300 dinocysts were counted in each sample. This method yields the best statistical representation of all the taxa present in the samples. The nomenclature used for the identification of the dinocysts corresponds to Rochon et al. (1999), Head et al. (2001), the index of Lentin and Williams (Fensome and Williams 2004) and Radi et al. (2013). The identification of the pollen and spores was performed using the Laurentian flora of Frère Marie-Victorin (1935) as well as Rousseau (1974), to account for the species encountered in Nunatsiavut, and McAndrews et al. (1973) for the pollen and spores micrographs.
Studies from Zonneveld et al. (1997, 2001) and Zonneveld and Brummer (2000) have shown species sensitivity to oxygen availability. The cysts from the Protoperidinium group, such as Brigantedinium sp., are the most sensitive, followed by the genera Spiniferites, Impagidinium and Operculodinium. The high sedimentation rates measured in the Nunatsiavut fjords helped preserving dinocyst assemblages from oxidation (Richerol et al., 2014), and the most sensitive taxa, such as Brigantedinium spp., showed excellent preservation, with the operculum still attached on many specimens.

Statistical analyses

Biozonations through clustering
In order to statistically distinguish the major changes in the dinocyst and pollen assemblages, we used the R software in combination with the libraries “vegan” and “rioja” to produce a cluster showing the euclidean distance between each sample for each core. Using the “decostand” function, our relative abundance data have first been standardized prior to the transformation of the resultant data matrix to a euclidean distance matrix. The grouping test CONISS was then applied to the distance matrix using the “chclust” function (Borcard et al., 2011). For each core section, a graphical hierarchical clustering has been obtained according to time.
Pollen-based quantitative paleoclimatic reconstructions
The reconstructions of terrestrial environmental parameters were performed with the software “R” and the reconstruction technique of the modern analog (RecMAT). The modern pollen assemblage database includes ‘N = 4,833’ sites from lakes, peats, moss cushions, and other terrestrial deposits of North America and Greenland, distributed among 10 vegetation types (biomes) (Whitmore et al., 2005). The methodology applied has been described by Fréchette et al. (2008). We have chosen to group together the Asteraceae because of an inability to distinguish between different species. This database includes 69 climatic and bioclimatic parameters. The followings were reconstructed and results were compared with the dinocyst-based paleoceanographic reconstructions: mean monthly temperature (°C) and precipitation (mm) for each month, total annual precipitation (mm) and mean temperature of the coldest and warmest month (°C). The climate estimates were calculated from the five best modern analogues. The validation of the database provided strong correlation coefficients and a RMSE (Root Mean Square Error) for each reconstructed parameter (Table 4.2).
Table 4.2 : Table summarizing the information on the parameters reconstructed from assemblages
of fossil pollen and dinocyst (from Richerol et al., 2014) : parameter denomination and codename, unit, coefficient de correlation (R2) and Root Mean Square Error (RMSE) of the validation.
Dinocyst-based quantitative paleoceanographic reconstructions
The reconstructions of sea-surface parameters were performed with the software “R” using fossil dinocyst assemblages, transfer functions and the search of five analogs in the GEOTOP modern dinocyst database (N = 1,441) (Richerol et al., 2008b, 2014). Here we also used the « bioindic » package developed on the R-platform (http://cran.r-project.org/), which is specially designed to offer various types of statistical analyses.

Results

Chronological framework

The 20 AMS-14C dates obtained (Table 4.1) do not show any reworking of sediments and a fitting curve extrapolates a R2 = 0.9884. The time span represented by the core covers the last ~7,000 BP. The calibrated dates yield a mean sedimentation rate of 0.1176 cm year-1 with a bottom date at 6,870 ± 122 cal BP. The last 2,000 years are represented in the upper 175 cm of the core (Figure 4.2).