Development of coccolithophore-based transfer functions in the western Mediterranean sea: A sea surface salinity reconstruction for the last 15.5 kyr

9Citations
Citations of this article
42Readers
Mendeley users who have this article in their library.

Abstract

A new data set of 88 marine surface sediment samples and related oceanic environmental variables (temperature, salinity, chlorophyll a, oxygen, etc.) was studied to quantify the relationship between assemblages of coccolithophore species and modern environmental conditions in the western Mediterranean Sea and the Atlantic Ocean, west of the Strait of Gibraltar. Multivariate statistical analyses revealed that coccolithophore species were primarily related to sea surface salinity (SSS), explaining an independent and significant proportion of variance in the coccolithophore data. A quantitative coccolithophore-based transfer function to estimate SSS was developed using the modern analog technique (MAT) and weighted-averaging partial least square regression (WA-PLS). The bootstrapped regression coefficient ( R 2boot) was 0.85MAT and 0.80WA-PLS, with a root-mean-square error of prediction (RMSEP) of 0.29MAT and 0.30WA-PLS (psu). The resulting transfer function was applied to fossil coccolithophore assemblages in the highly resolved (∼ 65 years) sediment core CEUTA10PC08 from the Alboran Sea (western Mediterranean) in order to reconstruct SSS for the last 25 kyr. The reliability of the reconstruction was evaluated by assessing the degree of similarity between fossil and modern coccolithophore assemblages and by a comparison of reconstructions with fossil ordination scores. Analogs were poor for the stadials associated with Heinrich events 2 and 1 and part of the Last Glacial Maximum. Good analogs indicate a more reliable reconstruction of the SSS for the last 15.5 kyr. During this period, several millennial and centennial SSS changes were observed and associated with sea-level oscillations and variations in the Atlantic Water entering the Alboran.

Figures

  • Figure 1. Maps of the study area. Panel (a): location of the CEUTA10PC08 core (red star). Black arrows trace general surface circulation. Legend: AW – Atlantic water; MAW – Modified Atlantic Water; AC – Algerian Current; NC – Northern Current. Panel (b): location of the 88 core-top samples used for final calibrations. Maps generated with Ocean Data View software (Schlitzer, 2014).
  • Figure 2. Geographical distribution of the main coccolithophore taxa. Panel (a): annual mean salinity at 10 m depth and surface circulation patterns in the study area; permanent trajectories (black arrows) and semipermanent mesoscale features (dashed arrows). Legend: AF – Alboran Front; AOF – Almería-Orán Front; CF – Catalan Front; BF – Balearic Front; AW – Atlantic Water; MAW – Modified Atlantic Water; AC – Algerian Current; NC – Northern Current. Distribution, according to their relative abundance (%), of (b) small Gephyrocapsa (< 3 µm), (c) E. huxleyi (< 4 µm)m, (d) G. muellerae, (e) C. leptoporus, (f) Helicosphaera spp., (g) F. profunda, (h) G. oceanica (< 5 µm).
  • Figure 3. Multivariate analyses. Panel (a): PCA based on the 13 initial environmental variables. Panel (b): CCA ordination plot with the site scores scaled by eigenvalues. The 88 sites are represented with their location in the Atlantic Ocean, Alboran Sea or Balearic Sea. Active and passive environmental vectors are represented by black and gray arrows, respectively. Scaling for the 16 taxa scores (open circles) is shown at the top left corner. mG: medium Gephyrocapsa; Eh (> 4): E. huxleyi (> 4 µm); Eh: E. huxleyi; sGg: small Gephyrocapsa; Gm: G. muellerae; Go: G. oceanica; sGo: small G. oceanica; Rc: R. clavigera; Sspp: Syracosphaera spp.; Of: O. fragilis; Gc: G. cf. caribbeanica; Hspp: Helicosphaera spp.; Cl: C. leptoporus; Uspp: Umbellosphaera spp.; Ug: Umbilicosphaera spp.; Fp: F. profunda.
  • Table 1. Multivariate analyses results. λ1/λ2: individual CCA. Preliminary model coefficients from the MAT and WA-PLS2. R 2 boot : bootstrapped coefficient of determination between the observed and predicted values. RMSEP: root-mean-square error of prediction.
  • Figure 4. Diagnostic graphs of the models: (a) observed salinity values; (b) MAT-predicted salinity values; (c) WA-PLS2-predicted salinity values; (d) observed vs. MAT-predicted salinity values; (e) observed vs. WA-PLS2-predicted salinity values; (f) MAT-predicted salinity values vs. residuals; (g) WA-PLS2-predicted salinity values vs. residuals.
  • Figure 5. Panel (a): SSS reconstructions for the CEUTA10PC08 core derived from the MAT (blue) and WA-PLS2 (green). The thin black lines represent the estimated values. The thick blue and green lines represent these original data fitted to a three-point moving average smoothing spline. Pale blue or green shading represent the error range, and dashed lines indicate current annual mean SSS in the Alboran Sea from the WOA13 (Zweng et al., 2013). Panel (b): dissimilarity between modern and fossil assemblages (black dots) measured by squared chord distance (left axis) plotted vs. age. The red line indicates the 10th percentile. Relative abundance of the species E. huxleyi (> 4 µm) (%; right axis). Panel (c): profiles comparing the PC1fossil (black line) and WA-PLS2-estimated SSS (green line).
  • Table 2. Model coefficients from the final MAT and WA-PLS2 cross-validated by bootstrapping for SSS, after removal of one outlier. R2 boot : bootstrapped coefficient of determination between the observed and predicted values. Max_Biasboot: bootstrapped maximum bias. RMSEP: root-mean-square error of prediction (psu).
  • Figure 6. Paleoenvironmental records in the Alboran Sea: (a) WA-PLS2-SSS reconstruction for CEUTA10PC08 core; age control points marked by triangles. Panel (b): C37 Alkenones from core HER-GC-T1 (Ausín et al., 2015). Panel (c): alkenone-SST from core MD95-2043 (Cacho et al., 2001). Red boxes represent the Alboran cooling events (AC_events). Panel (d): REDFIT periodogram of the SSS reconstruction for the Holocene. The gray bar marks the only significant peak at the 95 % significance level (green line).

References Powered by Scopus

A pervasive millennial-scale cycle in North Atlantic Holocene and glacial climates

2663Citations
N/AReaders
Get full text

A Theory of Gradient Analysis

2015Citations
N/AReaders
Get full text

Holocene climatic instability: A prominent, widespread event 8200 yr ago

1580Citations
N/AReaders
Get full text

Cited by Powered by Scopus

Quantitative reconstruction of primary productivity in low latitudes during the last glacial maximum and the mid-to-late Holocene from a global Florisphaera profunda calibration dataset

37Citations
N/AReaders
Get full text

Integrated stratigraphy and paleoceanographic evolution of the pre-evaporitic phase of the Messinian salinity crisis in the Eastern Mediterranean as recorded in the Tokhni section (Cyprus island)

36Citations
N/AReaders
Get full text

Calcareous plankton and the mid-Brunhes climate variability in the Alboran Sea (ODP Site 977)

21Citations
N/AReaders
Get full text

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Cite

CITATION STYLE

APA

Ausín, B., Hernández-Almeida, I., Flores, J. A., Sierro, F. J., Grosjean, M., Francés, G., & Alonso, B. (2015). Development of coccolithophore-based transfer functions in the western Mediterranean sea: A sea surface salinity reconstruction for the last 15.5 kyr. Climate of the Past, 11(12), 1635–1651. https://doi.org/10.5194/cp-11-1635-2015

Readers over time

‘15‘16‘17‘18‘19‘20‘21‘22‘23‘240481216

Readers' Seniority

Tooltip

PhD / Post grad / Masters / Doc 16

57%

Researcher 8

29%

Professor / Associate Prof. 4

14%

Readers' Discipline

Tooltip

Earth and Planetary Sciences 18

72%

Environmental Science 5

20%

Chemistry 1

4%

Engineering 1

4%

Save time finding and organizing research with Mendeley

Sign up for free
0