ArticlePDF Available

Application of multivariate statistical analyses to Itrax TM core scanner data for the identification of deep-marine sedimentary facies: A case study in the Galician Continental Margin

Abstract and Figures

The validity and usefulness of multivariate statistical tools for the facies characterization in deep-marine environments have been applied on the geochemical, sedimentological and magnetic data from a piston core extracted from the Transitional Zone in the Galician Continental Margin. The combination of geochemical profiles of Fe, Mn, Ti, Ba and Ca and magnetic susceptibility (MS) obtained using the ItraxTM Core Scanner at the University of Vigo, together with the grain-size, grey level and R (red) G (green) B (blue) colour analyses have allowed characterizing and classifying the sediments of the core PC13-3 using SPSS package v. 23. Cluster Analysis (CA) displays, in the first level of the hierarchy, two major groups that correspond with clay-silt and sand facies. In a second level, it is possible to observe six subfacies that match de visu preliminary classification and allowed us to complete and improve the characterization and the facies limits in the whole core. Discriminant Analysis (DA) confirmed the validity of the cluster analyses and enhanced the results of the classification. The Principal Component Analysis (PCA) shows four principal components: coarse lithogenic fraction (PC1), fine lithogenic fraction (PC2), high density fraction (PC3) and biogenic fraction (PC4). These results are in concordance with the Pearson correlation coefficient and the SEM observations. In general terms, the results confirm the utility of the multivariate statistical methods applied on high resolution geochemical and magnetic data acquired with ItraxTM corer scanner, as a quick and complementary tool in sedimentary facies analysis and description in deep marine environments.
Content may be subject to copyright.
Contents lists available at ScienceDirect
Quaternary International
journal homepage: www.elsevier.com/locate/quaint
Application of multivariate statistical analyses to Itrax
TM
core scanner data
for the identication of deep-marine sedimentary facies: A case study in the
Galician Continental Margin
A.E. López Pérez
a
, D. Rey
a
, V. Martins
b,c
, M. Plaza-Morlote
a
, B. Rubio
a,
a
GEOMA, Dpto. Geociencias Marinas y O.T, Universidade de Vigo, 36310, Vigo, Spain
b
Universidade Do Estado Do Rio de Janeiro, Av. São Francisco Xavier, 524, Maracanã, CEP 20550-013, Rio de Janeiro, RJ, Brazil
c
GeoBioTec, Dpto. Geociências, Universidade de Aveiro, Campus de Santiago, 3810-193, Aveiro, Portugal
ARTICLE INFO
Keywords:
Galicia continental margin
Sedimentology
Facies analysis
Multivariate statistical analysis
Itrax
TM
core scanner
ABSTRACT
The validity and usefulness of multivariate statistical tools for the facies characterization in deep-marine en-
vironments have been applied on the geochemical, sedimentological and magnetic data from a piston core
extracted from the Transitional Zone in the Galician Continental Margin. The combination of geochemical
proles of Fe, Mn, Ti, Ba and Ca and magnetic susceptibility (MS) obtained using the Itrax
TM
Core Scanner at the
University of Vigo, together with the grain-size, grey level and R (red) G (green) B (blue) colour analyses have
allowed characterizing and classifying the sediments of the core PC13-3 using SPSS package v. 23. Cluster
Analysis (CA) displays, in the rst level of the hierarchy, two major groups that correspond with clay-silt and
sand facies. In a second level, it is possible to observe six subfacies that match de visu preliminary classication
and allowed us to complete and improve the characterization and the facies limits in the whole core.
Discriminant Analysis (DA) conrmed the validity of the cluster analyses and enhanced the results of the
classication. The Principal Component Analysis (PCA) shows four principal components: coarse lithogenic
fraction (PC1), ne lithogenic fraction (PC2), high density fraction (PC3) and biogenic fraction (PC4). These
results are in concordance with the Pearson correlation coecient and the SEM observations. In general terms,
the results conrm the utility of the multivariate statistical methods applied on high resolution geochemical and
magnetic data acquired with Itrax
TM
corer scanner, as a quick and complementary tool in sedimentary facies
analysis and description in deep marine environments.
1. Introduction
In general, sedimentological facies classication is based on visual
description/interpretation and qualitative analysis of the sediment
core. Currently, this methodology is still widely used and provides good
results. A large number of works show examples of this, such as
Lamourou et al. (2017) who identied six sedimentary facies based on
microscopic observations in Quaternary deposits in the Gabes Gulf lo-
cated in the southeast of Tunisia coast. However, there is a tendency to
quantify or systematize the classication of facies employing multi-
variate statistical analyses such as Cluster Analysis (CA), Discriminant
Analysis (DA) or/and Principal Component Analysis (PCA). Barbera
et al. (2009) used PCA and discriminant function analyses on miner-
alogical (x-ray diraction) and geochemical (x-ray uorescence) data to
demonstrate provenance and continental sedimentary history in mu-
drocks. Rey et al. (2008) characterized ve magnetochemical facies to
determinate dierent sedimentary marine environments using CA on
geochemical and magnetic data acquired with XRF-CORTEX (core
scanner Texel) and cryogenic magnetometer. Margalef et al. (2013)
performed facies analysis using PCA on Fe, Ti and Ca data measured
with Itrax
TM
corer scanner, along with other discrete analysis (TC, TN
and δ13C) and macrofossil analysis in marine sediments located at the
central South Pacic Ocean. Baumgarten et al. (2014) carried out CA on
XRF data got with an Avaatech XRF core Scanner III to dene lacustrine
sediment characteristics. Flood et al. (2015;2018) used grain size,
mineralogy and geochemistry data (obtained through Itrax
TM
corer
scanner) to dene grain size variability, provenance and depositional
environments from a ne tidal estuary sediments using a multivariate
statistical methodology based on PCA and CA. Recently, Nugroho et al.
(2017) used CA and DA to characterize marine sedimentary facies and
depositional environments using grain size statistical parameters and
compositional data.
https://doi.org/10.1016/j.quaint.2018.06.035
Received 31 January 2018; Received in revised form 4 May 2018; Accepted 21 June 2018
Corresponding author.
E-mail address: brubio@uvigo.es (B. Rubio).
Quaternary International xxx (xxxx) xxx–xxx
1040-6182/ © 2018 Elsevier Ltd and INQUA. All rights reserved.
Please cite this article as: López Pérez, A.E., Quaternary International (2018), https://doi.org/10.1016/j.quaint.2018.06.035
Other studies evidence the utility of high-resolution data obtained
with the Itrax
TM
core scanner to identify facies and microfacies in
varved lakes core sediments, such as demonstrated Dulski et al. (2015).
Despite these interesting works, the use of multivariate statistical
methods in the facies analysis and depositional environment char-
acterization is very scarce when compared with other elds such as
environmental pollution and quality control in sediments (Rubio et al.,
2000;Martins et al., 2016) and groundwaters (Tlili-Zrelli et al., 2013).
The previous studies have demonstrated the advantage of using
multivariate statistical analysis (CA, DA and PCA) combining dierent
geochemical, magnetic and grain-size data against de visu descriptions
to classify facies in deep-marine environments with statistical
condence. CA allows a quick classication of the samples by grouping
samples with similar characteristics, while the DA provides a statistical
assessment and renement of the CA grouping. At the same time, PCA
allows dening new variables or components related to the sedi-
mentological and geochemical properties of the sediment.
This paper will explore how this approach can give greater con-
sistency, reliability, sensitivity and objectivity to facies classication
than the more common de visu procedure, particularly when it is based
on a large and diverse number of variables (i.e. geochemical, sedi-
mentological and magnetic). The main advantage of these statistical
methods lies in the fact that these analyses constitute a fast exploratory
method, supported by statistical parameters that improve the facies
distinction with very subtle changes. De visu classication is very de-
pendent of the observer's experience and could lead to errors in rela-
tively homogeneous sedimentary records, with subtle changes in grain-
size and variations in magnetic and/or geochemical properties. This
study, unlike to the previous referenced works, uses magnetic sus-
ceptibility data (1 cm) and raw high-resolution geochemical, colour and
grey level data (1 mm) (smoothed to each cm to improve results) ob-
tained with the Itrax
Tm
corer scanner. This high-resolution data allow a
better discrimination of the facies classication, even at millimeter
scale. The combination of these high-resolution data with the tradi-
tional lower resolution grain size data supposes an advantage in facies
description because let detect subfacies and subtle limits along the
whole core, very dicult to detect in the de visu classication. Selecting
the appropriate variables dataset allows discriminating between sedi-
mentological and provenance processes in deep marine environments,
mainly between pelagic and hemipelagic processes.
Our approach allows dierentiating a high number of facies in
comparison with the previous de visu procedure based on a high-re-
solution dataset obtaining by XRF-scanner and supported by statistical
analyses. This fact is representing a considerable advantage with the
Fig. 1. Bathymetric map of the Galicia continental margin with the PC13-3 core location and the four main morphostructural provinces in the study area: Deep
Galicia Margin (DGM), Galicia Bank (GB), Transitional Zone (TZ) and the Galicia Interior Basin (GIB).
Table 1
Statistical values from the variables used for the statistical described of the
PC13-3 core.
Variable Mean S.D. Minimum Maximum
Clay (%) 28.42 23.61 1.89 78.94
Silt (%) 26.40 11.68 7.82 56.20
Sand (%) 45.19 32.62 0.00 89.49
MGS (μm) 88.39 69.28 3.62 245.86
Sorting (μm) 66.50 43.28 4.88 152.83
Fe (p.a.) 17,581.06 12,549.91 6265.55 78,318.91
Ti (p.a.) 379.95 347.47 90.82 2161.64
Ba (p.a.) 43.71 23.55 15.09 161.55
Mn (p.a.) 223.86 157.58 66.82 950.64
Ca (p.a.) 162,539.43 39,283.20 53,594.73 227,924.18
GL 33,555.91 110.21 33,368.82 33,867.82
MS (10
5
SI) 6.29 7.59 0.20 48.40
Red 231.44 13.92 161.73 242.36
Green 209.29 23.56 136.45 253.09
Blue 171.20 39.91 90.00 250.00
106 samples used in the analysis. MGS = Mean Grains Size, GL = Grey Level,
MS = Magnetic Susceptibility and S.D. = Standard Desviation.
A.E. López Pérez et al. Quaternary International xxx (xxxx) xxx–xxx
2
classical visual description facies classication commonly used and
constitutes a rened approach of the using of multivariate statistical
methods in the facies classication eld.
2. Materials and methods
This work is based on a 4.28 m long PC13-3 piston core taken at
1688 m depth in the Transitional Zone (TZ) province (Ercilla et al.,
2008,2011;Vázquez et al., 2008) in the Galicia Continental Margin.
The core was collected during the Burato 4240oceanographic cruise
on board the R/V Sarmiento de Gamboa in September 2010 (latitude
42°4304.01N, longitude 11°0919.43W) (Fig. 1). The TZ is char-
acterized by three giant pockmark structures that have been related to
large-scale uid escapes. PC13-3 is extracted at NW of one of these
structures, known as Gran Burato, which has a circular morphology of
4 km in diameter, with maximum depths of 375m, and is characterized
by high slopes. The facies classication of the PC13-3 core will allow
knowing the aection in the local sedimentation of the uid escape
processes.
Optical and radiographical images were obtained with the Itrax
TM
Core Scanner at the University of Vigo, as well as geochemical and
magnetic susceptibility data, using the Mo-tube with a voltage of 30 kV
and an exposure time of 20 s. The high-resolution XRF geochemical raw
data (1 mm step size) were smoothed using a 1 cm running mean to
validate and improve their reliability (Rodríguez-Germade et al., 2013).
Radiographic data were exported to grey-scale data les with the Re-
dicore software of the core scanner. Colour data were obtained in RGB
values from the optical images obtained with the Itrax
TM
core scanner.
Grain-size distributions were determined from discrete samples
collected every 4 cm using a laser diraction particle size analyzer
Coulter LS230 (Beckman) at the Department d´Estratigraa, Paleontologia
I Geociències Marines de la Universitat de Barcelona.
The petrology of the core was studied employing a JEOL JSM-6700f
Scanning Electron Microscope (SEM), operating in back-scattering
mode (BS), located at the C.A.C.T.I. of the University of Vigo.
The statistical analyses (CA, DA and PCA) were carried out using the
SPSS package v.23 for a total of 15 variables analyzed in 106 samples
(1590 data points).
3. Results and discussions
3.1. General sediment properties
Table 1 summarizes descriptive statistical values for the grain-size,
geochemical and magnetic sediment properties obtained by the SPSS
v.23 software. In general terms, PC13-3 core contents an average sand,
silt and clay percentage of 45.19 ± 32.62%, 26.40 ± 11.68%,
and 28.42 ± 23.61%, respectively. The mean grain size is
88.39 ± 69.28 μm(φ= 3.50 ± 3.85). Regarding the sorting, this
parameter shows a value of 66.50 ± 43.28 μm (3.91 ± 4.53 φ), so the
general description of the core corresponds with very ne sand very
poorly sorting. The statistic results of RGB components present mean
values of 231.44 ± 13.92 for the red, 209.29 ± 23.56 for the green
and 171.20 ± 39.91 for the blue, corresponding with sienna tonalities.
Regarding the geochemical results, elements such as Fe, Ti, Ba and
Mn show mean values of 17,581.06 ± 12,549.91 peak areas (p.a.),
379.95 ± 347.47 p. a., 43.71 ± 23.55 p. a. and 223.86 ± 157.58 p.
a. respectively, are being the iron the metal element that presents more
variability between maximum and minimum values. Ca shows a mean
value of 162,539.43 ± 39,283.20 p. a., as well as the highest dier-
ence between the maximum and minimum value. Respect to the MS the
mean value obtained is 6.29 10
5
± 7.59 SI and also presents a high
variability between samples (range varied from 0.20 10
5
- 48.40 10
5
SI). Finally, the Grey Level (GL), a parameter related to the density,
shows mean value of 33,555.91 ± 110.21.
Pearson correlation matrix (Table 2) show most variables are well
correlated (p < 0.01). Clay and silt show a noticeable positive corre-
lation (r = 0.850) and these both variables present negative correlation
with Sand, Mean Grain Size (MGS) and sorting (r = 0.844,
r=0.920 and r = 0.891 for the clay and r = 0.665, r = 0.728
and r = 0.685 for the silt respectively). Sand shows a positive cor-
relation between MGS and Sorting (r = 0.952 and r = 0.948 respec-
tively). These correlations could indicate a large variability of the grain-
size in the sedimentary record, being sand the most abundant size in the
core (high positive correlation between sand and MGS). Moreover high
positive correlation is noticeable between Fe vs. Ti (r = 0.938 and
p < 0.01) and is remarkable the positive correlation between MS vs,
Fe, Ti, Ba, and Mn. This could be related to lithogenic components with
high metallic elements content. Regarding Ca, this element shows po-
sitive correlation with RGB variables and a negative correlation with
others metallic elements (Fe, Ti, Ba, and Mn) and MS, that could be
associated to biogenic components with high Ca and low metallic ele-
ments.
3.2. Statistical analysis
3.2.1. Statistical analysis applications
Cluster Analysis (CA) is a statistical exploration tool that allows to
group samples by its degree of similarity. This analysis is widely used to
study the pattern of distribution and provenance of sediments in
Table 2
Pearson correlation matrix for the variables.
Clay Silt Sand MGS Sorting Fe Ti Ba Mn Ca GL MS Red Green Blue
Clay 1.000 .850** -.844** -.920** -.891** -.071 -.185 -.127 .103 .078 -.449** -.272** .092 .426** .462**
Silt 1.000 -.665** -.728** -.685** .134 .061 .056 .279** -.077 -.404** -.042 -.068 .200* .240*
Sand 1.000 .952** .948** .196* .336** .192* .018 -.088 .329** .350** -.223* -.437** -.477**
MGS 1.000 .995** .160 .296** .167 -.074 -.062 .411** .324** -.155 -.466** -.503**
Sorting 1.000 .172 .308** .182 -.052 -.059 .380** .331** -.175 -.472** -.505**
Fe 1.000 .938** .777** .666** -.577** -.024 .701** -.768** -.812** -.814**
Ti 1.000 .813** .641** -.472** -.009 .688** -.728** -.797** -.798**
Ba 1.000 .727** -.471** -.225* .538** -.692** -.647** -.624**
Mn 1.000 -.527** -.405** .503** -,589** -.512** -.487**
Ca 1.000 .150 -.521** .636** .662** .637**
GL 1.000 .116 .107 -.216* -.227*
MS 1.000 -.543** -.726** -.715**
Red 1.000 .739** .709**
Green 1.000 .977**
Blue 1.000
106 samples were used for the correlation analysis. MGS = Mean Grains Size, GL = Grey Level and MS = Magnetic Susceptibility.
**p < 0.01.
*p < 0.05.
A.E. López Pérez et al. Quaternary International xxx (xxxx) xxx–xxx
3
depositional environments based on grain-size, geochemical and/or
magnetic data (Kim et al., 2013;Kolesnik et al., 2017). Discriminant
Analysis (DA) allows obtaining a discriminant function based on linear
combinations of the variables and allows predicting and dierentiating
the ones, which belong to a particular group of the samples. Principal
Component Analysis (PCA) reduces the dimensionality of the dataset by
linear compilations of correlated variables called Principal Components
(PCs). DA and PCA are typically used in assessment pollution or in the
distribution of metals elements in marine sediments (Rubio et al., 2000;
Farmaki et al., 2014), but very rarely for facies classication.
3.2.2. Cluster analyses
Cluster analysis (CA) was run in Q mode using the unweighted pair
group method with arithmetic mean (UPGMA method) and the
Euclidean squared distance as the similarity coecient. Previously,
with the objective of testing the uniformity and normality of the data,
Kolmogorov-Smirnov and Saphiro-Wilk tests were performed on all
variables, conrming the non-normal distribution of the variables ex-
cept for the variables of silt, MS and green colour. All data were nor-
malized using a logarithmic transformation to obtain better results and
avoid the eect of dierences in magnitude and variance of the data
(Rubio et al., 2000).
Fig. 2 shows the obtained dendrogram, where two main clusters
(CLA and CLB) in the rst level of the hierarchical dendrogram and six
subclusters (S1 to S6) in the second hierarchy level are identied. CLA
comprises all samples with sand content below of 17% and CLB samples
with sand percentage higher than 17%. CLA is divided into two sub-
facies, S1 and S2. Cluster S1 groups samples with sand percentage be-
tween 0 and 2% and S2 clusters samples with sand percentage between
4% and 17%. CLB group includes subclusters CL1 and CL2. CL1 contain
two subfacies depending on the sand content: S3 enclose samples with
sand percentage between 17% and 66% and S4 groups samples with
sand percentage higher than 69%. Both subfacies do not include sam-
ples with high content of metallic elements, and high values of MS. CL2
includes S5 and S6 subfacies, which grouped samples with higher
content in Fe, Ti, and MS in the whole core. S5 shows lower values of
Fe, Ti and MS (19,800.91 p. a., 435.86 p. a. and 12.41 10
5
SI re-
spectively) than S6, which displays the highest values for Fe, Ti and MS
parameters (50,959.20 p. a., 1303.41 p. a. and 17.02 10
5
SI respec-
tively).
3.2.3. Discriminant analysis
DA was performed by the stepwise method to obtain the percentage
of correct prediction to validate statistically the dierent groups ob-
tained by cluster analysis. Prior to the DA, Box's M Test was carried out
to check the validity of the hypothesis of equal covariance. Results
(p < 0.05) reject the null hypothesis of equality of matrices of covar-
iance, so DA was performed obtained a percentage of 98.1% of correct
predictions of samples classied by CA (Table 3 and Fig. 3).
Only two samples (47 and 60) show a dierent classication in DA,
which diers from the CA results. Sample 47 was classied by CA in S4
meanwhile the DA predicted that it pertains to S5 in the highest group
classication and S4 in the second highest group. The probability of
pertaining to the predicted highest group has a value of 0.576 and
shows a Mahalanobis distance of 8.976. On the other hand, the prob-
ability of correct classication in the second highest group has a value
of 0.361 and shows a Mahalanobis distance value (9.910) very similar
in comparison to the distance indicator in the highest group.
Additionally, sample 47 present a grain-size, geochemical and magnetic
values situated in CA in the high limit of S4 close to the low limit of S5.
Regarding the sample 60, it was clustered in S3 using CA meanwhile DA
grouped it in S2 in the highest group and S3 in the second highest group
classication. The value of probability of belong to the assigned highest
group is 0.997, and its Mahalanobis distance from the group centroid of
S2 is 10.712. Otherwise, the probability to pertain to the second pre-
dicted highest group is 0.03, and its Mahalanobis distance (22.500) is
much higher than the rst predicted group distance. Moreover, this
Fig. 2. Dendrogram obtained using UPGMA method and the Euclidean squared
distance as the similarity coecient. A total of 106 samples were used in CA.
A.E. López Pérez et al. Quaternary International xxx (xxxx) xxx–xxx
4
sample has a percentage of sand of 17.18%, being the grain-size lower
limit of S3 suggested by CA. These grain-size values are more similar to
the samples grouped in S2 (mean sand percentage of 10.13%) than
samples in S3 (mean sand percentage of 53.08%).
Taking into account all these considerations from DA (probabilities
and Mahalanobis distances) and the geochemical, magnetic and grain-
size subcluster limits values, we determinate that sample 47 was cor-
rectly classied in S4 using CA meanwhile sample 60 was not correctly
classied in S3. This sample pertains to S2 as demonstrated the statis-
tical results obtained by DA (p = 0.997 and Mahalanobis
distance = 10.712). Moreover, the low value of Wilks's lambda (0.001),
along with the high value of chi-square (1331.03), allowed us to ensure
the validity of the groupings of facies classication (p < 0.0001). Thus
we can determine the useful and complementarity of DA in facies
classication to verify and improve CA results, due to this analysis
allow obtaining statistical parameters that validate the classication.
Table 4 contains the correlations of the variables with the ve rst
discriminant functions and indicates the variables selected and used in
the discriminant analysis (sand, sorting, Fe, Mn, clay, SM and silt). This
suggests that these variables have more weight in the dataset than the
rest of variables.
3.2.4. Principal component analysis
Principal component analysis (PCA) was performed on all data to
obtain principal components that allow describing and characterizing
the sediments and geochemical properties of the core. For this purpose
normalized and standardize data using logarithmic transformations
were again applied without rotation of the matrix. Moreover, the MGS
was removed from the matrix owing to it is related to the grain size and
show high correlation with the sand.
Four components have been extracted, explaining the 87.07% of the
total variance (Table 5). The PC1 groups the variables of sand, sorting,
Fe, Ti, Ba Mn and MS, which shows high correlation. Moreover, it is
remarkable their negative correlation with clay, Ca and RGB colour
variables. This component represents 48.78% of the total variance of
data. PC2 explains 26.44% and groups silt, clay and metal transition
variables and show negative signicant correlation of GL. PC3 re-
presents 6.87% of the variance of all data and only shows signicant
negative correlation of GL parameter. Finally, PC4 represents 4.98% of
the total variance and shows signicant positive Ca correlation. Results
Table 3
Results of correct predictions of samples classied by CA.
Classication results
a
Subfacies Predicted Group Membership Total
S1 S2 S3 S4 S5 S6
Original Count S1 27 0 0 0 0 0 27
S20700007
S3011800019
S4000201021
S5 0 0 0 0 22 0 22
S6 0 0 0 0 0 10 10
% S1 100.0 0.0 0.0 0.0 0.0 0.0 100.0
S2 0.0 100.0 0.0 0.0 0.0 0.0 100.0
S3 0.0 5.3 94.7 0.0 0.0 0.0 100.0
S4 0.0 0.0 0.0 95.2 4.8 0.0 100.0
S5 0.0 0.0 0.0 0.0 100.0 0.0 100.0
S6 0.0 0.0 0.0 0.0 0.0 100.0 100.0
a
98.1% of original grouped cases correctly classied.
Fig. 3. Plot of the canonical discriminant functions obtained by DA.
A.E. López Pérez et al. Quaternary International xxx (xxxx) xxx–xxx
5
of the PCA are in concordance with the Pearson correlation matrix.
PC1 was interpreted as a coarse lithogenic component rich in metal
transition elements and high MS. Meanwhile, variables groups in PC2
allowed us to describe this association as ne lithogenic component.
PC3 was identied as a high-density component related to the ne li-
thogenic component. Both, PC2 and PC3 show a signicant negative
correlation of GL. PC2 also shows a signicant positive correlation with
clay and silt. This means that samples with high content in ne-grain
sediments have low values of GL, an indicative of high density, because
the porosity in clay and silt fraction is lower than in sands. PC4 was
described as biogenic component.
The interpretation of these four components can be observed by
SEM (Fig. 4). Fig. 4a and b shows well-preserved foraminifera sands
with terrigenous components of dierent sizes and Fig. 4c and d shows
magnetite and ilmenomagnetite respectively. These terrigenous com-
ponents constitute the coarse lithogenic rich in transition elements
(PC1) and the ne lithogenic component (PC2) related to the high
density component (PC3). Finally, Fig. 4e shows a small proportion of
well-preserved foraminifera in a coccolithophoridae matrix (Fig. 4f).
This matrix, along with the well-preserved foraminifera along the
whole core, denes the biogenic component (PC4).
3.3. Sediment properties of subfacies classication
Fig. 5 displays the facies classication obtained after CA and DA
application (Fig. 5a). This gure compares the previous visual de-
scription of the core (Fig. 5b) following by the photography (Fig. 5c),
radiography (Fig. 5d) and grain-size data distribution of the core
(Fig. 5e). Note that the limits between the dierent facies and the re-
cognition of the high magnetic susceptibility facies along the core have
signicantly improved. Taking into account the new facies classica-
tion for the PC13-3, Table 6 shows the average values for the dierent
variables for each subfacies.
Subfacies S1 presents average sand, silt and clay percentages of
0.25%, 35.85% and 63.89 respectively and mean values of Fe, Ti, Ca
and MS of 12,891.33 p. a., 213.88 p. a., 166,495.64 p. a. and 2.63 10
5
SI respectively. Volumetrically it represents 8.04% of the core.
Meanwhile, S2 represents volumetrically 26.29% of the core and pre-
sents sand, silt and clay content of 10.13%, 46.72% and 43.15% for
each variable and Fe, Ti, Ca and MS average of 13,367.84 p. a., 279.69
p. a., 178,198.01 p. a. and 4.74 10
5
SI respectively. S3 represents
16.65% of the core and displays an average mean percentage of sand,
silt, and clay of 53.08%, 24.94%, and 21.98% respectively. Moreover
Fe, Ti, Ca and MS show an average of 11,835.36 p. a., 252.76 p. a.,
189,383.77 p. a. and 3.01 10
5
respectively. Regarding S4 represents
volumetrically 19.69% of the whole core and shows mean values of
sand, silt, and clay of 80.97%, 13.27%, and 5.77% respectively. Also, it
shows an average value for Fe, Ti, Ca and MS of 11,920.71 p. a., 242.37
p. a., 185,141.13 p. a. and 2.90 10
5
SI respectively. S5, that represent
volumetrically the 19.21%, contents percentages of sand, silt and clay
of 70.15%, 19.37% and 10.48% respectively and Fe, Ti, Ca and MS
values of 19,800.91 p. a., 435.86 p. a., 129,951.01 p. a. and 12.41 10
5
SI for each variable mentioned. Finally, S6 shows a mean percentage of
sand, silt, and clay of 50.29%, 30.24% and 19.47% respectively and
represents 10.12% of the whole core. Regarding Fe, Ti, Ca and MS
parameters, S6 displays values of 50,959.20 p. a., 1303.51 p. a.,
115,241.88 p. a. and 17.02 10
5
SI respectively.
3.4. Sedimentological signicance of the CA, DA and PCA results
A precise combination of the variables in the matrix dataset, and the
use of limits obtained by the statistical analyses, allow interpret pro-
cesses (pelagic and hemipelagic) and provenance of the sediment re-
cord (detrital and biogenic), considering condence intervals for CA,
DA and PCA, combined with Pearson correlation and SEM observations.
The variables dominant in hemipelagic detrital facies are Fe, Ti, and
MS, in pelagic biogenic facies is mainly Ca.
S1 and S2 clustering in CLA by CA, classied samples corresponding
to a clay-silt grain size that show high Ca content, low Fe and Ti content
and low-susceptibility. These samples dened as Ca-rich low-
Table 4
Structure matrix.
Structure matrix
Variables Function
1234 5
% Sand .878
-.284 .104 -.031 .258
MGS
a
.492 -.462 -.353 .057 -.116
Sorting .474
-.353 -.220 .138 -.096
Fe .082 .674
-.593 .120 -.065
Ti
a
-.025 .557 -.548 .079 -.126
Mn .023 .526
-.212 -.105 .085
Ba
a
.016 .467 -.419 .093 -.005
Red
a
.032 -.431 .169 .013 -.030
Ca
a
.041 -.284 .122 .147 .054
GL
a
.035 -.256 -.010 -.070 -.013
% Clay -.327 .500 .585
.479 .176
Blue
a
-.047 -.316 .541 -.019 -.014
Green
a
.020 -.379 .524 .024 -.029
MS .083 .340 -.244 -,549
.461
% Silt -.165 .511 .460 -.001 -.536
Pooled within-groups correlations between discriminating variables and stan-
dardized canonical discriminant functions. Variables ordered by absolute size of
correlation within function.
*. Signicant correlation between each variable and every discriminant func-
tion.
Bold are the r values which are statistically signicant at the level p < 0.0001.
a
This variable not used in the analysis.
Table 5
PCA results.
PCA results
Components Eigenvalues Explained variance
(%)
Accumulated variance
(%)
1 6.83 48.78 48.78
2 3.70 26.44 75.22
3 0.96 6.87 82.08
4 0.70 4.98 87.07
Components loading
Component 1 Component 2 Component 3 Component 4
% Clay -.430 .852* -.160 .079
% Silt -.176 .853* -.174 .165
% Sand .513 -.744* .260 .075
Sorting .501 -.790* .203 .075
Fe .884* .315 -.130 .184
Ti .895* .199 -.016 .314
Ba .787* .309 .267 .220
Mn .643* .528 .310 .035
Ca -.664 -.288 .079 .642*
GL .082 -.618* -.706* .161
MS (10
5
SI) .789* .044 -.183 .041
Red -.804* -.278 -.002 .114
Green -.942* .055 .198 .104
Blue -.939* .100 .186 .087
106 samples used in the analysis.
* Signicant variable (p < 0.001) for every component.
Bold are the r values which are statistically signicant at the level p < 0.0001.
A.E. López Pérez et al. Quaternary International xxx (xxxx) xxx–xxx
6
susceptibility silt-clay facies (Car-lok silt-clay facies) and described as
silt-clay pelagite. S3 and S4 included in CL1 of CLB, content for-
aminifera-sand samples characterized by high Ca content, low Fe and Ti
content and low-susceptibility. These samples named as the Ca-rich
low-susceptibility sand facies (Car-lok sand facies) and interpreted as
sand pelagite. S5 and S6 grouping in CL2 of CLB, display foraminifera-
sand samples with the highest content of Fe, Ti and most upper MS
values. These samples classied as Fe-Ti high susceptibility sand facies
(Fe-Ti sand facies) and described as hemipelagic magnetic layers inter-
bedded in the pelagic sediment that could be related to IRDs layers
deposited during the Heinrich Events. CA, DA and PCA allow to identify
three dierent clay-silt and sand facies that correspond to pelagic and
hemipelagic sediments.
4. Conclusions
The combination of high-resolution Itrax
TM
core scanning determi-
nation of Fe, Ti, Ca, Mn, Ba and magnetic susceptibility proles with
colour RGB, grey line data, detailed grain-size and other grain-size
parameters (MGS and sorting) in the same data set, have allowed to
characterize and to classify the sediments of the PC13-3 core using
multivariate statistical method through the SPSS package v.23.
Descriptive statistics results, combined with the SEM observations, al-
lowed us to describe the sediment of the study core as a very ne for-
aminifera-sand very poor sorting. CA shows two major facies (CLA and
CLB) and six subfacies that correspond with the hemipelagite and pe-
lagite in a previous visual classication. CA results allowed us to
complete and improve the characterization and the limits of the facies
and subfacies of the core, allowing establishing better limits for subtle
dierences. DA allowed statistically validates the clusters obtained and
improved their results. DA results showed that, overall, more than
98.1% of the samples grouped by the CA are properly classied.
Moreover, the low value of lambda Wilks statistic (0.001), along with
the high value of chi-square (1331.03) allowed validating the facies
classication made by CA (p < 0.0001). Thus, the combination of CA
and DA constitute a complementary multivariate statistical tool in the
Fig. 4. SEM micrographs obtained at dierent depths of the PC13-3 core. a) Fe-Ti sand facies at 35 cm composed by foraminifera and terrigenous components. b) Fe-
Ti sand facies at 54 cm composed by foraminifera and terrigenous components. c) Magnetite located at 105 cm in Fe-Ti sand facies. d) Ilmenomagnetite located in Fe-
Ti sand facies at 105 cm e) Car-lok silt-clay facies at 400 cm f) Car-lok silt-clay facies at 400 cm with optical magnifying where it is possible to recognize the
coccolithophoridae matrix.
A.E. López Pérez et al. Quaternary International xxx (xxxx) xxx–xxx
7
eld of facies classication because establishes a robust statistical
methodology to determinate the facies classication and conrm their
validity. The combination of both analyses allows us to obtain a sta-
tistical value by DA that provides a reliance statistical weight to the CA
classication, conrming the utility and condence of these kinds of
tools in marine facies classication. Moreover, PCA shows four prin-
cipal components, described as coarse lithogenic fraction (PC1), ne
lithogenic fraction (PC2), high density fraction (PC3) and biogenic
fraction (PC4). These results are in concordance with the Pearson cor-
relation coecient and the SEM observations. We can conclude that
multivariate statistical analyses (CA, DA, and PCA) constitute a useful
and fast complementary tool in facies classication applied to Itrax
core scanner data that let improves the visual facies characterization in
deep-marine environments.
Fig. 5. a) Facies description obtained using multivariate statistical methods b) previous visual facies description c) optical and d) radiographical images obtained
with the ItraxCorer scanner, followed by a grain size distribution (e). Note the improvement on the facies classication by using multivariate statistical methods.
A.E. López Pérez et al. Quaternary International xxx (xxxx) xxx–xxx
8
Acknowledgements
This work was funded by the Gran Burato 2010, 2011 convenia
between the University of Vigo, CSIC, and Xunta de Galicia, and the
MINECO Project CGL2008-034774-E. We want to thank the captains
and crew of the R/V Sarmiento de Gamboa, the UTM technical support
and the GB4240 and GB2011 cruise participants. A.E. López-Pérez was
awarded a PhD grant by the Xunta de Galicia (Department of Culture,
Education and University Planning).
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://dx.
doi.org/10.1016/j.quaint.2018.06.035.
References
Barbera, G., Lo Giudice, A., Mazzoleni, P., Pappalardo, A., 2009. Combined statistical and
petrological analysis of provenance and diagenetic history of mudrocks: application
to Alpine Tethydes shales (Sicily, Italy). Sediment. Geol. 213, 2740. http://dx.doi.
org/10.1016/j.sedgeo.2008.11.002.
Baumgarten, H., Wonik, T., Kwiecien, O., 2014. Facies characterization based on physical
properties from downhole logging for the sediment record of Lake Van, Turkey. Quat.
Sci. Rev. 104, 8596. http://dx.doi.org/10.1016/j.quascirev.2014.03.016.
Dulski, P., Brauer, A., Mangili, C., 2015. Combined μ-XRF and microfacies techniques for
lake sediment analyses. In: Micro-XRF Studies of Sediment Cores, Developments in
Paleoenvironmental Research. Springer, Dordrecht, pp. 325349. http://dx.doi.org/
10.1007/978-94-017-9849-5_12.
Ercilla, G., García-Gil, S., Estrada, F., Gràcia, E., Vizcaino, A., Váquez, J.T., Díaz, S., Vilas,
F., Casas, D., Alonso, B., Dañobeitia, J., Farran, M., 2008. High-resolution seismic
stratigraphy of the Galicia Bank Region and neighbouring abyssal plains (NW Iberian
continental margin). Mar. Geol. 249, 108127. http://dx.doi.org/10.1016/j.margeo.
2007.09.009.
Ercilla, G., Casas, D., Vázquez, J.T., Iglesias, J., Somoza, L., Juan, C., Medialdea, T., León,
R., Estrada, F., García-Gil, S., Farran, M. l., Bohoyo, F., García, M., Maestro, A., 2011.
Imaging the recent sediment dynamics of the Galicia Bank region (Atlantic, NW
Iberian Peninsula). Mar. Geophys. Res. 32, 99126. http://dx.doi.org/10.1007/
s11001-011-9129-x.
Farmaki, E.G., Thomaidis, N.S., Pasias, I.N., Baulard, C., Papaharisis, L., Efstathiou, C.E.,
2014. Environmental impact of intensive aquaculture: Investigation on the accumu-
lation of metals and nutrients in marine sediments of Greece. Sci. Total Environ.
485486, 554562. http://dx.doi.org/10.1016/j.scitotenv.2014.03.125.
Flood, R.P., Orford, J.D., McKinley, J.M., Roberson, S., 2015. Eective grain size dis-
tribution analysis for interpretation of tidaldeltaic facies: West Bengal Sundarbans.
Sediment. Geol. 318, 5874. http://dx.doi.org/10.1016/j.sedgeo.2014.12.007.
Flood, R.P., Barr, I.D., Weltje, G.J., Roberson, S., Russell, M.I., Meneely, J., Orford, J.D.,
2018. Provenance and depositional variability of the thin mud facies in the lower
ganges-brahmaputra delta, West Bengal sundarbans, India. Mar. Geol. 395, 198218.
http://dx.doi.org/10.1016/j.margeo.2017.09.001.
Kim, W., Doh, S.-J., Yu, Y., Lee, Y.I., 2013. Magnetic evaluation of sediment provenance
in the northern East China Sea using fuzzy c-means cluster analysis. Mar. Geol. 337,
919. http://dx.doi.org/10.1016/j.margeo.2013.01.001.
Kolesnik, A.N., Astakhov, A.S., Kolesnik, O.N., 2017. Recent deposition environments in
the Chukchi Sea and adjacent areas of the Arctic Ocean: evidence from Q-cluster
analysis of sediment compositions and grain sizes. Russ. Geol. Geophys. 58,
14681477. http://dx.doi.org/10.1016/j.rgg.2017.11.011.
Lamourou, A., Touir, J., Fagel, N., 2017. Reconstructing the Holocene depositional en-
vironments along the northern coast of Sfax (Tunisia): mineralogical and sedi-
mentological approaches. J. Afr. Earth Sci. 129, 713727. http://dx.doi.org/10.
1016/j.jafrearsci.2017.01.031.
Margalef, O., Cañellas-Boltà, N., Pla-Rabes, S., Giralt, S., Pueyo, J.J., Joosten, H., Rull, V.,
Buchaca, T., Hernández, A., Valero-Garcés, B.L., Moreno, A., Sáez, A., 2013. A 70,000
year multiproxy record of climatic and environmental change from Rano Aroi peat-
land (Easter Island). Global Planet. Change 108, 7284. http://dx.doi.org/10.1016/j.
gloplacha.2013.05.016.
Martins, M.V.A., Helali, M.A., Zaaboub, N., Boukef-BenOmrane, I., Frontalini, F., Reis, D.,
Portela, H., Clemente, I.M.M.M., Nogueira, L., Pereira, E., Miranda, P., El Bour, M.,
Aleya, L., 2016. Organic matter quantity and quality, metals availability and for-
aminiferal assemblages as environmental proxy applied to the Bizerte Lagoon
(Tunisia). Mar. Pollut. Bull. 105, 161179. http://dx.doi.org/10.1016/j.marpolbul.
2016.02.032.
Nugroho, S.H., Putra, P.S., Yulianto, E., Noeradi, D., 2017. Multivariate statistical analysis
for characterization of sedimentary facies of Tarakan sub-basin, North Kalimantan.
Mar. Georesour. Geotechnol. 111. http://dx.doi.org/10.1080/1064119X.2017.
1399178.
Rey, D., Rubio, B., Mohamed, K., Vilas, F., Alonso, B., Ercilla, G., Rivas, T., 2008. Detrital
and early diagenetic processes in Late Pleistocene and Holocene sediments from the
SW Galicia Bank inferred from high-resolution enviromagnetic and geochemical re-
cords. Marine Geology, Geological characterization of the Galicia Bank Region
(Atlantic Ocean, NW Iberia) 249, 6492. http://dx.doi.org/10.1016/j.margeo.2007.
09.013.
Rodríguez-Germade, I., Rubio, B., Rey, D., Vilas, F., Martins, V., Rocha, F., 2013.
Evaluation of data processing techniques for marine sediments analyzed with an Itrax
Core Scanner. Geogaceta 53, 8587.
Rubio, B., Nombela, M.A., Vilas, F., 2000. Geochemistry of major and trace elements in
sediments of the Ria De Vigo (NW Spain): an assessment of metal pollution. Mar.
Pollut. Bull. 40, 968980.
Tlili-Zrelli, B., Azaza, F., hamzaoui-, Gueddari, M., Bouhlila, R., 2013. Geochemistry and
quality assessment of groundwater using graphical and multivariate statistical
methods. A case study: grombalia phreatic aquifer (Northeastern Tunisia). Arabian
Journal of Geosciences 6, 35453561. http://dx.doi.org/10.1007/s12517-012-
0617-3.
Vázquez, J.T., Medialdea, T., Ercilla, G., Somoza, L., Estrada, F., Fernández Puga, M.C.,
Gallart, J., Gràcia, E., Maestro, A., Sayago, M., 2008. Cenozoic deformational struc-
tures on the Galicia Bank region (NW Iberian continental margin). Mar. Geol. 249,
128149. http://dx.doi.org/10.1016/j.margeo.2007.09.014.
Table 6
Mean values of variables for each facies.
Facies Car-lok silt-clay facies Car-lok sand facies Fe-Ti sand facies
Subfacies S1 S2 S3 S4 S5 S6
Variables Mean Mean Mean Mean Mean Mean
Clay (%) 63.89 43.15 21.98 5.77 10.48 19.47
Silt (%) 35.85 46.72 24.94 13.27 19.37 30.24
Sand (%) 0.25 10.13 53.08 80.97 70.15 50.29
MGS (μm) 8.60 20.81 82.60 190.65 120.91 82.03
Sorting (μm) 12.12 24.86 71.03 125.27 86.38 71.34
Fe (p.a.) 12,891.33 13,367.84 11,835.36 11,920.71 19,800.91 50,959.20
Ti (p.a.) 213.88 279.69 252.76 242.37 435.86 1303.41
Ba (p.a.) 34.71 39.18 31.83 36.77 48.29 97.52
Mn (p.a.) 188.57 247.65 163.23 123.48 256.76 547.69
Ca (p.a.) 166,495.64 178,198.01 189,383.77 185,141.13 129,951.01 115,241.88
GL 33,502.90 33,550.34 33,506.21 33,654.56 33,579.39 33,534.10
MS (10
5
SI) 2.63 4.74 3.01 2.90 12.41 17.02
Red 237.49 234.95 237.28 235.87 227.09 202.05
Green 227.22 221.85 223.76 206.97 190.71 170.56
Blue 203.78 193.27 196.22 165.04 137.43 107.71
106 samples used in the analysis. MGS = Mean Grain Size, GL = Grey Level and MS = Magnetic Susceptibility.
A.E. López Pérez et al. Quaternary International xxx (xxxx) xxx–xxx
9
... They also defined the major stratigraphic sequences and large geomorphological features of the margin. The margin has also been the subject of numerous research focused on how sedimentary and oceanographic processes driven by glacial/interglacial cycles during the Late Pleistocene and Holocene have modelled the main geomorphological features that conformed the present-day submerged landscape of the Galician Continental Margin Rey et al., 2008;Voelker and de Abreu, 2011;Martins et al., 2013;Salgueiro et al., 2014;Hanebuth et al., 2015;Mena et al., 2015Mena et al., , 2018Zhang et al., 2016;Plaza-Morlote et al., 2017;López Pérez et al., 2019;Petrovic et al., 2019). The sedimentary record is such that permits detailed palaeoclimatic/palaeoceanographic reconstructions of this part of the North Atlantic of global significance. ...
... One remarkable finding is the intensification and deepening (sinking to ~2000 m) of the Mediterranean Outflow Water (MOW) during stadials conditions (Voelker et al., 2006;Voelker and de Abreu, 2011), which according to Ercilla et al., 2011could be related to different erosive and depositional bottomcurrent features at the study area. During cold stadials also took place the deposition of ice-rafted debris (IRD) caused from the melting of icebergs discharges over North Atlantic (Heinrich, 1988), resulting in a detrital imprint on the sediments of the margin Rey et al., 2008;Martins et al., 2013;Plaza-Morlote et al., 2017;Mena et al., 2018;López Pérez et al., 2019). These works demonstrated that the geographic situation of the Galician Continental Margin makes it a place of interest to study variations of oceanographic and sedimentary conditions associated with climatic changes during the Late Pleistocene and Holocene (Plaza-Morlote et al., 2017). ...
... The statistical facies analyses were carried out using the SPSS package v.23 with the methodology proposed by López Pérez et al. (2019). In order to get a reliable facies classification of the cores, different multivariate statistical analyses (Cluster Analysis and Discriminant Analysis) were applied in a large high-resolution geochemical, magnetic and sedimentological dataset. ...
Article
Studies of the most surficial sedimentary record from passive continental margins provide crucial knowledge about sedimentary dynamics and its changes through recent geological times. These studies allow understanding in detail the influence of both tectonic activity and long-term bottom-current circulation over the Late Quaternary sedimentary dynamics. Using a large dataset of multibeam bathymetry, chirp and multi-channel seismic (MCS) records, ROV seabed direct observations and a magneto-chemical facies characterisation, we provide a well-dated record of the tectono-sedimentary processes on an isolated high marginal platform over the Late Quaternary. Our results display several structural and geomorphological features and tectonic pulses that indicate intense faulting, folding, and deformation of the most recent sedimentary cover. Furthermore, we document four main sedimentary systems acting at the study area and controlled by different water masses (MOW, LSW and NADW): the (hemi)pelagic, bottom current-controlled (hemi)pelagic, contourite and downslope sedimentary systems; as well as a new typology of contourite associated to giant depressions, named as pockmarks-related drift. The record also shows erosive features and extremely low sedimentation rates for the last 172 cal ka BP. Results suggest that the topographic configuration of the high marginal platform and reorganizations of the water masses associated to climate changes causes a constriction of the water masses that induce an enhancement of the bottom-current activity, favouring erosion, winnowing and redistribution of sediments. Therefore, this study provides new insight into the tectonic control and bottom current activity effect over the Late Quaternary sedimentary cover from an isolated and elevate offshore morphostructural province, located at the Galician Continental Margin.
... They also defined the major stratigraphic sequences and large geomorphological features of the margin. The margin has also been the subject of numerous research focused on how sedimentary and oceanographic processes driven by glacial/interglacial cycles during the Late Pleistocene and Holocene have modelled the main geomorphological features that conformed the present-day submerged landscape of the Galician Continental Margin Rey et al., 2008;Voelker and de Abreu, 2011;Martins et al., 2013;Salgueiro et al., 2014;Hanebuth et al., 2015;Mena et al., 2015Mena et al., , 2018Zhang et al., 2016;Plaza-Morlote et al., 2017;López Pérez et al., 2019;Petrovic et al., 2019). The sedimentary record is such that permits detailed palaeoclimatic/palaeoceanographic reconstructions of this part of the North Atlantic of global significance. ...
... One remarkable finding is the intensification and deepening (sinking to ~2000 m) of the Mediterranean Outflow Water (MOW) during stadials conditions (Voelker et al., 2006;Voelker and de Abreu, 2011), which according to Ercilla et al., 2011could be related to different erosive and depositional bottomcurrent features at the study area. During cold stadials also took place the deposition of ice-rafted debris (IRD) caused from the melting of icebergs discharges over North Atlantic (Heinrich, 1988), resulting in a detrital imprint on the sediments of the margin Rey et al., 2008;Martins et al., 2013;Plaza-Morlote et al., 2017;Mena et al., 2018;López Pérez et al., 2019). These works demonstrated that the geographic situation of the Galician Continental Margin makes it a place of interest to study variations of oceanographic and sedimentary conditions associated with climatic changes during the Late Pleistocene and Holocene (Plaza-Morlote et al., 2017). ...
... The statistical facies analyses were carried out using the SPSS package v.23 with the methodology proposed by López Pérez et al. (2019). In order to get a reliable facies classification of the cores, different multivariate statistical analyses (Cluster Analysis and Discriminant Analysis) were applied in a large high-resolution geochemical, magnetic and sedimentological dataset. ...
Conference Paper
The analysis of five piston cores from the Transitional Zone of the Galicia continental margin provides information of the sedimentary processes occurred during the last 170 kyr BP as well as the sedimentary characterization of the “Gran Burato” giant pockmark. ROV images together with bathymetric and ultra-high resolution parametric echosounder data helped us to recognize their sedimentary setting. Facies classification was based on grain-size, magnetic susceptibility, high-resolution XRF, RGB colour and grey level data obtained by the ITRAXTM core scanner. The qualities of the results are confirmed by a set of multivariate statistical analyses and SEM observations. The study area is dominated by pelagic, hemipelagic and IRD/detrital sedimentary facies. The pockmark structure shows accumulation of pteropods at the bottom and erosional features on the slopes, as confirmed by the truncated reflectors in the parametric profiles. Sedimentation rates have lower values during glacial periods. Glacial/interglacial cycles together with topography are the main factors controlling the sedimentation in this area.
... The study of the most surficial sedimentary cover of the margin revealed the palaeoclimatic and palaeoenvironmental interest of this area to understand in detail the climatic, environmental and oceanographic changes during the Late Pleistocene and Holocene at these latitudes (e.g. Alonso et al. 2008;Rey et al. 2008;Bender et al. 2012;Martins et al. 2013Martins et al. , 2015Hanebuth et al. 2015;Plaza-Morlote et al. 2017;Mena et al. 2018;Petrovic et al. 2019;López-Pérez et al. 2019, 2021a. ...
... Statistical facies analyses of the core samples were carried out using the SPSS v.23 software using the procedure described in López-Pérez et al. (2019). The facies classification was based on the application of several multivariate statistical analyses (cluster analysis and discriminant analysis) to the large high-resolution geochemical, magnetic and sedimentological dataset from the core. ...
Article
Full-text available
Surficial sediments on the seafloor from passive continental margins can provide insight into recent Late Quaternary sedimentary dynamics acting over offshore sedimentary systems. This work focuses on the study of some particular ferruginous tubular structures resembling bioforms (FTB) located in the distal Galician Continental Margin (NW Iberian Margin) at water depths between ~ 1550 and ~ 2200 m. The characterisation of these structures made it possible to study in depth their formation environment and subsequent sedimentary evolution during the Late Pleistocene and Holocene. The FTB consist of goethite with a framboidal texture. They were interpreted as formed by an initial pyrite precipitation in reducing microenvironments conditioned by the activity of sediment-dwelling organisms during the early diagenesis. This is followed by the oxidation of pyrite by a combination of hydrothermal fluids and erosional processes, which triggers the formation of the framboidal oxyhydroxides. The data allowed obtaining a comprehensive understanding of the environmental context and the significance of these ferruginous tubules, as there are no previous studies in the scientific literature that describe these structures in a source-to-sink sedimentary system.
... In summary, our analyses of turbidite T25 demonstrate that Zr/Rb peaks indicate increased abundances of coarse particles. High-resolution ITRAX data are useful for identifying very thin-bedded turbidites and sedimentary facies, especially in fine-grained trench sediments (e.g., Pérez et al. 2019). ...
Article
Full-text available
The southwestern Ryukyu Trench represents the ultimate sink of sediments shed from Taiwan into the Philippine Sea, which are mainly transported to the trench by turbidity currents via submarine canyons. Here, we present trench turbidites intercalated with hemipelagites in a gravity pilot core and a piston core acquired on the Ryukyu Trench floor at 6147 m water depth. We performed X-ray fluorescence core scans (ITRAX profiles), magnetic measurements, and organic geochemistry analyses to discriminate turbidites from hemipelagites. We identified 36 turbidites (0.9–4.2 cm thick) based on visual core descriptions and Ca/Fe ratios in the ITRAX profiles. Three of these turbidites show magnetic signatures indicating the presence of pyrrhotite and peaks in the magnetic susceptibility profile, suggesting that Taiwan-sourced sediments are transported to the Ryukyu Trench by long-runout turbidity currents. Pyrrhotite is also present in hemipelagites of the upper part of the retrieved cores, indicating a dominant sediment source in Taiwan over the last several thousand years. Ca/Fe and Zr/Rb ratios in the ITRAX profiles mark distal turbidites (about 1–3 cm thick), and Zr/Rb peaks mainly reflect grain size changes. Detailed analyses of a representative turbidite show good correlation between Ca/Fe and Zr/Rb peaks with upward-coarsening and upward-fining trends that delimit the turbidite. Sedimentary organic matter in hemipelagites is characterized by higher total organic carbon and total nitrogen contents and higher δ ¹³ C values than that in turbidites. Our multi-proxy approach employing high-resolution XRF core scans to differentiate turbidites from hemipelagites contributes to establishing a comprehensive view of modern trench sedimentation from Taiwan to the southwestern Ryukyu Trench.
... Evidence for shallow gas in the proximal northwest Iberia continental margin has been described since the early 2000s (Durán et al., 2007;Ferrín et al., 2003;García-García et al., 2003;García-Gil et al., 2015). However, the possibility of hydrate occurrence did not emerge until a decade later based on the presence of several seabed features related to fluid escape imaged in the Transitional Zone (Druet, 2015;Ercilla et al., 2011;López Pérez et al., 2019;Ribeiro, 2011). Some of the fluid escape structures have a seafloor expression (e.g., pockmarks), while others were detected by seismic amplitude anomalies. ...
Article
Large national programs in the United States and several Asian countries have defined and characterised their marine methane hydrate occurrences in some detail, but European hydrate occurrence has received less attention. The European Union-funded project “Marine gas hydrate – an indigenous resource of natural gas for Europe” (MIGRATE) aimed to determine the European potential inventory of exploitable gas hydrate, to assess current technologies for their production, and to evaluate the associated risks. We present a synthesis of results from a MIGRATE working group that focused on the definition and assessment of hydrate in Europe. Our review includes the western and eastern margins of Greenland, the Barents Sea and onshore and offshore Svalbard, the Atlantic margin of Europe, extending south to the northwestern margin of Morocco, the Mediterranean Sea, the Sea of Marmara, and the western and southern margins of the Black Sea. We have not attempted to cover the high Arctic, the Russian, Ukrainian and Georgian sectors of the Black Sea, or overseas territories of European nations. Following a formalised process, we defined a range of indicators of hydrate presence based on geophysical, geochemical and geological data. Our study was framed by the constraint of the hydrate stability field in European seas. Direct hydrate indicators included sampling of hydrate; the presence of bottom simulating reflectors in seismic reflection profiles; gas seepage into the ocean; and chlorinity anomalies in sediment cores. Indirect indicators included geophysical survey evidence for seismic velocity and/or resistivity anomalies, seismic reflectivity anomalies or subsurface gas escape structures; various seabed features associated with gas escape, and the presence of an underlying conventional petroleum system. We used these indicators to develop a database of hydrate occurrence across Europe. We identified a series of regions where there is substantial evidence for hydrate occurrence (some areas offshore Greenland, offshore west Svalbard, the Barents Sea, the mid-Norwegian margin, the Gulf of Cadiz, parts of the eastern Mediterranean, the Sea of Marmara and the Black Sea) and regions where the evidence is more tenuous (other areas offshore Greenland and of the eastern Mediterranean, onshore Svalbard, offshore Ireland and offshore northwest Iberia). We provide an overview of the evidence for hydrate occurrence in each of these regions. We conclude that around Europe, areas with strong evidence for the presence of hydrate commonly coincide with conventional thermogenic hydrocarbon provinces.
Article
The use of core scanners to perform X-ray fluorescence (XRF) spectroscopic analysis can only obtain the intensities of the target elements, which is not conducive for application in marine geology research. In this paper, using a core scanner, in situ measurements were performed on 15 components: Al2O3, SiO2, K2O, CaO, TiO2, MnO, Fe2O3, V, Cr, Zn, Rb, Sr, Y, Zr, and Ba. We explored the feasibility of reducing the effect of interstitial water by normalizing the original intensity using the intensity of Ca and the ratio of coherent to incoherent and attempted to introduce a genetic algorithm-back propagation neural network model and use its nonlinear fitting capability to correct the matrix effect. The prediction precision of this method was 0.6–15.4%. The proposed method is suitable for the rapid analysis of major and minor components in marine sediment core samples, while taking full advantage of the high-resolution of the XRF core scanner.
Article
Full-text available
Deep‐sea sediments record palaeoenvironmental and oceanographic changes over time through variations in their magnetochemical and sedimentological properties. This study presents the Late Quaternary sedimentary reconstruction of three sediment cores extracted from a source‐to‐sink sedimentary system located in a high marginal platform of the north‐west Iberian Margin. The aim of this reconstruction is to explore the palaeoclimatic, palaeoproductivity and diagenetic changes over the last 172 cal ka BP attending to geochemical, magnetic and sedimentological variations. The record is comprised of four sedimentary units consisting of pelagic, hemipelagic and detrital facies, the latter associated with ice‐rafted debris events at Heinrich Stadials. These findings allowed the reconstruction of the lithostratigraphy attending to climatic changes, covering the last eleven Heinrich Stadials. These units and facies successions are controlled by the superposition of the Milankovitch and Dansgaard–Oeschger cycles. Low sedimentation rates have been reported, especially during cold periods. Palaeoproductivity was assessed using total organic carbon and calcium carbonate content, together with biogenic barium (Babio) and organic phosphorous (Porg), both calculated via total digestion analysis using standard calculations. Mass accumulation rates for total organic carbon, Babio and Porg were calculated. Furthermore, a sequential extraction was performed to explore the different phases of the barium within sediments: biogenic barite (Babar), barium associated with iron and manganese oxides (Baox) and Ba associated with aluminosilicates (Badet). Results display a limitation in the use of the biogenic barium and phosphorous and mass accumulation rates, suggesting detrital Ba (Badetr) contribution during Heinrich Stadials, which could mask the calculations. In contrast, only calcium carbonate fluctuates attending to glacial/interglacial cycles. Regarding some diagenetic proxies (i.e. Mn, Mn/Fe and Mn/Ti), they reveal oxidation fronts of different generations, suggesting that the early diagenesis has not reached steady‐state, because the redoxcline fluctuates. This fact suggests an erosional activity that causes the continuous oxygenation of the sediment and non‐steady‐state conditions.
Article
The purpose of this study is to differentiate and characterise contouritic sands in two different locations with variable sediment compositions (siliciclastic and bioclastic) based on a multiproxy approach that includes the analysis of sedimentary texture, semi-quantitative geochemistry, microfacies and ichnological information, as well as a Principal Component Analysis (PCA) applied to geochemical and sedimentary data on sediment cores. The integration of sedimentological analyses and the PCA permits also to the differentiation between fine-grained deep-water deposits such as hemipelagites, muddy contourites and hyperpycnites. A depositional model is proposed here, based on geochemical and sedimentological characteristics of the Holocene-highstand Mozambique Channel upper slope sands, and glacial-lowstand Corsica Contourite Depositional System middle slope sands. The upper continental slope of the Mozambican margin is characterised by siliciclastic sandy contourites, muddy contourites and muddy hyperpycnites. Mozambique siliciclastic sandy contourites constitute large accumulations of well-sorted very fine to coarse sand with evidences of strong winnowing and reworking under high-energy conditions. The sedimentary facies represents highstand contourite sands and shows a reversely-graded trend. The contourite drift on the Pianosa ridge (eastern flank of the Corsica Trough) consists of bioclastic sandy and muddy contourites and hemipelagites. Bioclastic sandy contourites are made up of shallow-marine winnowed bioclasts with a reversely- and normally-graded trend and represent lowstand contourite sands. The PCA in the two environments —showing a distinctive geochemical signal— allows for differentiation of the contourite deposits. In siliciclastic sands, reworking is marked by an accumulation of Si, Zr, and Sr in fine- to medium-grained sands. In bioclastic sands, reworking is less evident but it is characterised by accumulations of Ca and Sr. The reworking and winnowing bottom current effects are also observed at the microfacies scale. Both types of contourite deposits show evidences of intermittent depositional conditions depending on the ichnodiversity, distribution and abundance of trace fossils. This work is useful to discriminate similar fine-grained deposits in different continental margins which would contribute to a better understanding of sedimentary deposits and processes in deep-marine environments.
Article
Full-text available
The introduction and application of high resolution XRF core scanners has made an enormous contribution to palaeoclimate science over the last two decades. These core scanners have become essential analytical tools for the investigation of sedimentary and other natural cumulative records of environmental change. Their widespread adoption by the environmental and geoscience communities has had a significant impact on the ability to systematically record high resolution chemical profiles in marine and lake sediments and soils. The range of applications has broadened beyond sedimentology and palaeoclimatology and now extends into scanning samples of speleothems, corals, mumiyo mounds, rocks, sedimentary nodules, archaeological artefacts, tree sections and melt segregations. Other areas exploiting the capability of XRF core scanners are mineral exploration, geotechnical characterisation, soil remediation studies and carbon capture and storage research. As a consequence of their ability to efficiently determine detailed, non-destructive elemental records in sediments and soils, scientific papers published since the year 2000 now number more than 1000.
Article
Full-text available
Deposition environments in the Chukchi Sea and adjacent areas of the Arctic Ocean are studied by the Q-cluster analysis of compositions, particle sizes, and other properties of surface bottom sediments. Analysis of more than 4700 numeric values allowed mapping fifteen clusters distributed over the seafloor according to deposition environments. Chemical and statistical data confirm the predominance of clastic sedimentation and mainly mechanic sorting of sedimentary material. At the same time, the major-element composition trends correspond to Si decrease and Al increase seaward. Biogenic deposition is of inferior scale and shows up as relative enrichment in total organic carbon, Ca, Mg, Ba, Sr, and some other biogenic and chemogenic components in sediments. Clastic and biogenic deposition, with accumulation of Fe, Mn, V, Ni, Cr, Co, and other elements, as well as precipitation of authigenic phases, occurs within areas of seafloor having particular water chemistry, such as the Herald Canyon, the outer shelf, and the deepwater Arctic Ocean.
Article
Full-text available
A study of sediment dynamics has been conducted on the Tarakan sub-basin, North Kalimantan, Indonesia using multivariate analysis. Multivariate statistical techniques can be used to determine sediment with similar characteristics and be a good proxy to recognize sedimentary facies and depositional environment. These methods have been applied to characterize the sedimentary facies at Tarakan sub-basin. A total of 23 samples were taken from several locations on the different depths of 56–2554 m that represent varieties of environment. The study is interpreted using compositional data analysis associated with a grain size trend analysis (GSTA), principal component analysis (PCA), and cluster analysis (CA). The GSTA value showed the dominance of poorly sorted silt indicating that the sediment is mostly deposited in low-energy depositional. The CA and the PCA determined three distinct sedimentary facies: shelf facies, upper bathyal facies, and lower bathyal facies. The facies were grouped as similar sediment and depositional environment. Sedimentological variables applied in the characterization were described to be an important tool for the interpretation of depositional environments – indirectly showing hydrodynamic energy.
Article
The Sundarbans is one of the largest coastal wetland sites in the world and covers an area of approximately one million hectares of the western delta of the Ganges and Brahmaputra (G-B) rivers (located across Bangladesh and India). Since the late Holocene, the western delta has not been directly fluvially sourced, due to the Ganges shift to the east (present-day Bangladesh). The depositional facies (Thin Mud 2 Facies) of the late-Holocene abandoned western region (The Sundarbans) is derived from dominant estuary-tidal dynamics, however the provenance of the associated TMF sedimentation in this far western zone (Indian Sundarbans per se) is as yet equivocal. In this study, sediment cores from the Indian Sundarbans (Saptamukhi-Thakuran estuary) were closely examined for grain-size distributions (GSDs), mineralogy through X-ray diffraction (XRD), and geochemistry with X-ray fluorescence (XRF). The TMF in the West Bengal Sundarbans has been determined to show intensively weathered, terrestrial sediment, derived principally from the Ganges Alluvial Plain (GAP). There is a predominance of quartz, mica and clay minerals, with quartz interpreted as a product of low-relief tropical weathering sourced via the G-B Rivers draining the Himalayas. Lithofacies interpreted through GSD analysis of the TMF is indicative of a muddy tidal flat environment with aggradation and a general fining-up trend between the adjacent estuaries. The sediment provenance indicates a continuing G-B sediment source, which moves westward along the Bay of Bengal, from the active delta front and is then reworked over the far-western abandoned delta by tidal–estuarine forcing.
Article
A sedimentological and mineralogical study of sedimentary cores allowed reconstructing the evolution of depositional environments along the Northern coast of Sfax (Tunisia). The aim of this research work is to identify the factors controlling the sedimentation from the Holocene to the Present time. Three 30-m sediment cores collected by drilling at 30 m water depth were analyzed for their color, magnetic susceptibility signal, grain size by laser diffraction, organic matter content by loss of ignition, carbonate content by calcimetry and mineralogy by X-ray diffraction on bulk powder and clay <2 μm. They broadly present the same sedimentological and mineralogical features. Microscopical observations of petrographic slides allowed identifying six main sedimentary facies. Bulk mineralogical assemblages comprised clay minerals, quartz, calcite, gypsum and K-feldspars were examined. Considerable change was observed in the carbonate content that mimicked the bioclaste abundance and diluted the detrital minerals (clay minerals, quartz and feldspars). The gypsum mainly occurred in the lower sedimentary columns (SC12 and SC9) and in the upper/middle of core SC6. The clay fraction was made of a mixture of kaolinite, illite, smectite and palygorskite with no clear variation through core depth. Both grain-size parameters and magnetic susceptibility profile showed a sharp transition in the upper 2–5 m of the sedimentological columns. Coarse, sandy to gravely sediments characterized by a low magnetic susceptibility signal were replaced by fine bioclastic-rich clayey sediments. The analysis of vertical succession of depositional facies revealed a fluvial depositional environment (coastal plain) basically marked by fluvial channels and inundation plains at the bottom of all cores. However, core-top sediments recorded a littoral marine environment with sand depositions rich in gastropods, lamellibranches and algæ. Depositional facies, sedimentological and mineralogical parameters were consistent with a transition from a fluviatile depositional environment with some emersion phases marked by the gypsum precipitation, to a marine littoral environment. Such evolution was accompanied with a relative sea-level rise which flooded the fluvial system at the coastal plain during the Holocene, in agreement with sea-level fluctuations in southeast Tunisia during the Holocene.
Chapter
This contribution presents an evaluation of still unexplored potentials, limitations and technical details of µ-XRF element scanning particularly for varved sediment analyses using the vacuum device EAGLE III XL. For this case study a 33 cm long interval of exceptionally well-preserved sub-millimetre scale calcite varves of the interglacial lake deposits from Piànico has been selected. In addition to the demonstration of the very good repeatability of XRF element scans, one focus of this paper is on discussing a suitable scanner resolution in terms of a compromise between short analyses times and a resolution that allows capturing the geochemical signature even of seasonal sub-layers. By combining scanner data with microscopic sediment inspection geochemical signatures of various micro-facies as, for example, varves, detrital layers and matrix-supported and clay layers are explored. The potential of counting varves using the seasonal signature of specific elements is discussed and potential error sources are disclosed by comparison with microscopic varve counts. Detailed grid- µXRF-scanning provides for the first time an insight into the internal structure of varves presented as high-resolution 3D-images. Finally, an approach of calibrating scanner data through comparison with ICP-MS analyses is introduced. Sensitivity factors for each element have been defined as ratios between mean count rates and bulk element concentrations.
Article
Lake Van (Turkey) is the 4th largest terminal lake in the world and is located at a key position for climatic reconstruction. The ICDP project ‘PALEOVAN’ is a deep-drilling campaign initiated in the summer of 2010 to enhance the understanding of paleoclimatic and paleoenvironmental conditions in the Middle East for a period of 550,000 years. Multiple coring of two sites (Northern Basin and Ahlat Ridge) at a water depth of up to 360 m has been performed. The sedimentary record is mainly composed of clayey silts and tephra deposits that were supplied by four volcanic sources: 1) the Süphan volcano, located on the northern shore, 2) the Nemrut volcano, 15 km west of the westernshore, 3) the Incekaya volcano, on the southwestern shore and 4) intralake eruptive centers. The dominant chemical composition of the volcanic sources is known from studies of land deposits.
Article
The Grombalia coastal aquifer, situated in Northeastern Tunisia, is a water source for public, agricultural, and industrial supplies in the region. The overexploitation of this aquifer, since 1959, and the agriculture activities led to the degradation, by places, of the water quality. The present study implemented graphical, modeling, and multivariate statistical tools to investigate natural and anthropogenic processes controlling Grombalia groundwater mineralization and water quality for promoting sustainable development. To attempt this goal, groundwater was collected from 33 observation wells in January 2004, and samples were analyzed for 10 physicochemical parameters (temperature, pH, salinity, Na+, Ca2+, K+, Mg2+, Cl−, HCO3−, and SO 42−). Hydrochemical facies using Piper diagram indicates a predominance of a mixed facies, of the Na-Cl-HCO3 type, or Na-Ca-Cl-SO4 type, and, with less expansion, Na Cl type. The main factors controlling Grombalia groundwater mineralization seem to be mineral dissolution of highly soluble salts especially, the halite dissolution existing in the surface salty deposits and, with less importance, the ion exchange and reverse ion exchange process with clay minerals existing in the aquifer. The comparison of the major ions of the Grombalia groundwater, with the World Health Organization norms of potability (WHO 2004), reveals that these waters cannot be used for human consumption without any treatment. Most waters of the Grombalia aquifer, with a relatively high salinity, are not suitable for irrigation, in ordinary conditions. Nevertheless, they can be used for permeable soils, with an adequate drainage and applying an excess of leaching water.