GEOCHEMISTRY OF NEARSHORE SEDIMENTS FROM
THE NORTH AEGEAN SEA, GREECE
by
FANI SAKELLARIADOU
A thesis submitted for the degree of Doctor of Philosophy
of the University of London
Applied Geochemistry Research Group
Royal School of Mines
Imperial College of Science and Technology July 1987 2
To my parents Leonidas and Alexandra,
To Leonidas,
To Mary 3 ABSTRACT
The surface sediments of North Aegean Sea comprise modern terrigenous detrital material (introduced by river runoff and coastal erosion), reworked older sediments and biogenic matter.
Fine-grained sediments near major river mouths reflect fluvial input. Sands in the mid-shelf region of Samothraki plateau represent deposition during earlier, lower sea levels. Aluminium and iron rich, fine-grained sediments in Kavala gulf consist of reworked material.
Supply and distribution of sediments on Samothraki and Strymonikos plateaux are affected by a westwards flowing nearshore current and a west-east trending nearshore canyon, respectively. The former transports the Evros runoff westwards and the latter directs the Strymon runoff downslope.
Within the study area, lerissos gulf surface sediments are significantly enriched in Mn, Fe, Ni, Zn and Cu. The enrichments express the mineralization onland (mixed sulphide deposits, Cu-bearing porphyry stocks and Mn-oxide deposits). In buried sediments off Stratoni, Mn, Fe and Zn concentrations increase dramatically close to the surface. This trend is probably related to contamination deriving from coastal mining activities.
The buried sediments of North Aegean Sea mainly comprise clay and carbonate matter. In some of them, an upward migration of Mn due to diagenetic remobilization is observed.
In all surface and subsurface sediments, more than 50% of Ti, Al, Fe, K, Ba, V, Cr, Li, Co, Ni, La, Mg, Be, P, Cu and Zn is hosted in the lattice structure of aluminosilicate minerals. Significant proportions of Mn, Fe, Zn, Pb, Ni and Co are associated with the Fe,Mn-oxides. The carbonates contain mainly Ca and Sr. In the Samothraki plateau surface sediments the Cu affinity for organic matter is pronounced. The highest proportions of Pb, Zn, Cu, Ni and Mn not bound to aluminosilicates occur in the lerissos gulf sediments, confirming that their origin is related to onland mineralization and mining activities.
Although various enrichments were determined in some sediments, economic deposits of any of the elements studied are unlikely to occur in the upper parts of the North 4 Aegean Sea sediments. However, it is demonstrated that the detection of mineralized areas onland (e.g. eastern Halkidiki peninsula) is feasible through the geochemical study of nearshore sediments. 5
Table of Contents
Eaoa
Abstract 3
Table of contents 5
List of Tables 11
List of Figures 15
Acknowledgements 26
Chapter I Introduction
1.1 General introduction, project outline and aims 27 1.2 Thesis structure 30 1.3 General morphological and geological setting 30 1.3.1 Main rivers 30 1.3.2 Composition of rocks on the mainland 31 1.3.3 Composition of rocks on islands 32 1.4 Bathymetry 33 1.5 Hydrography 37 1.6 Nature of sediments offshore 37 1.6.1 Strymonikos area 37 1.6.2 Samothraki plateau 37 1.6.3 General discussion 38 1.7 Previous geochemical and mineralogical investigations 39
Chapter II Analytical techniques
2.1 Bulk chemical analysis 41 2.1.1 Lithium metaborate technique 41 2.1.1.1 Outline of the technique 41 2.2 Mineral acid attack 42 2.2.1 Outline of the technique 42 2.3 Comparison between method A and method B 42 2.4 Selective sequential extraction analytical procedure 50 2.4.1 Introduction 50 2.4.2 Methods 50 2.5 Calcium carbonate and organic carbon determinations 57 2.5.1 Calcium carbonate determination : outline of thetechnique 57 2.5.2 Organic carbon determination : outline of the technique 57 2.6 X-Ray Diffraction 58
Chapter III The Samothraki plateau
3.1 Surface sediments 59 3.1.1 Sediment composition 59 3.1.1.1 Distribution of rock fragments, quartz, heavy minerals, 59 biogenic matter and carbonates 6
3.1.1.1.1 Rock fragments 59 3.1.1.1.2 Quartz 60 3.1.1.1.3 Heavy minerals 60 3.1.1.1.4 Biogenic carbonate particles 61 3.1.1.1.5 Carbonates 61 3.1.1.2 Distribution of organic carbon and results from XRD and 62 SEM studies 3.1.1.2.1 Organic carbon 62 3.1.1.2.2 XRD results 63 3.1.1.2.3 SEM results 64 3.1.2 Bulk geochemistry of offshore sediments and of those on 72 the adjacent coast 3.1.2.1 Silicon 72 3.1.2.2 Calcium 72 3.1.2.3 Aluminium 73 3.1.2.4 Iron 73 3.1.2.5 Potassium 74 3.1.2.6 Magnesium 75 3.1.2.7 Titanium 75 3.1.2.8 Strontium 75 3.1.2.9 Phosphorus 76 3.1.2.10 Barium 76 3.1.2.11 Manganese 77 3.1.2.12 Zirconium 77 3.1.2.13 Chromium 78 3.1.2.14 Vanadium 78 3.1.2.15 Zinc 78 3.1.2.16 Lanthanum 79 3.1.2.17 Nickel 79 3.1.2.18 Copper 80 3.1.2.19 Cobalt 80 3.1.2.20 Beryllium 80 3.1.3 Multivariate Statistical Analysis 91 3.1.3.1 Introduction 91 3.1.3.2 Correlation Matrices 92 3.1.3.3 Cluster analysis 95 3.1.3.3.1 Box-Cox transformed data 95 3.1.3.3.2 Raw, untransformed data 95 3.1.3.4 Factor analysis 99 3.2 Buried sediments 102 3.2.1 Description of the cores 102 3.2.1.1 Core THR 7 102 3.2.1.2 Core THR 22 102 3.2.2 Bulk geochemistry of subsurface offshore sediments 103 3.2.2.1 Core THR 7 103 3.2.2.2 Core THR 22 103 3.3 Summary 106
Chapter IV The Strymonikos gulf, plateau and triangle, and the Kavala gulf
4.1 Surface sediments 108 4.1.1 Sediment composition 108 4.1.1.1 Distribution of rock fragments, quartz, heavy minerals, 108 biogenic matter and carbonates 7
4.1.1.1.1 Rock fragments 108 4.1.1.1.2 Quartz 109 4.1.1.1.3 Heavy minerals 109 4.1.1.1.4 Biogenic carbonate particles 111 4.1.1.1.5 Carbonates 111 4.1.1.2 Distribution of organic carbon and results from XRD and 111 SEM studies 4.1.1.2.1 Organic carbon 111 4.1.1.2.2 XRD results 111 4.1.1.2.3 SEM results 113 4.1.2 Bulk geochemistry of offshore sediments and of those on 120 the adjacent coast 4.1.2.1 Silicon 120 4.1.2.2 Aluminium 120 4.1.2.3 Calcium 121 4.1.2.4 Iron 121 4.1.2.5 Potassium 121 4.1.2.6 Magnesium 122 4.1.2.7 Titanium 122 4.1.2.8 Phosphorus 123 4.1.2.9 Manganese 123 4.1.2.10 Barium 123 4.1.2.11 Strontium 124 4.1.2.12 Zirconium 124 4.1.2.13 Vanadium 125 4.1.2.14 Chromium 125 4.1.2.15 Zinc 126 4.1.2.16 Nickel 127 4.1.2.17 Lanthanum 127 4.1.2.18 Copper 127 4.1.2.19 Cobalt 128 4.1.2.20 Beryllium 128 4.1.3 Multivariate Statistical Analysis 139 4.1.3.1 Strymonikos group 139 4.1.3.1.1 Correlation Matrices 139 4.1.3.1.2 Cluster analysis 142 4.1.3.1.2.1 Box-Cox transformed data 142 4.1.3.1.2.2 Raw, untransformed data 142 4.1.3.1.3 Factor analysis 145 4.1.3.2 Kavala group 148 4.1.3.2.1 Correlation Matrices 148 4.1.3.2.2 Cluster analysis 151 4.1.3.2.2.1 Box-Cox transformed data 151 4.1.3.2.2.2 Raw, untransformed data 151 4.1.3.2.3 Factor analysis 154 4.1.3.3 Comparison between the Strymonikos and Kavala groups 156 4.2 Buried sediments 157 4.2.1 Description of the cores 157 4.2.1.1 Core STR 1 157 4.2.1.2 Core STR 2 157 4.2.1.3 Core STR 4 157 4.2.1.4 Core STR 11 158 4.2.1.5 Core STR 12 158 4.2.1.6 Core STR 27 159 4.2.1.7 Core STR 28 159 4.2.1.8 Core KB 3 159 8
4.2.1.9 Core KB 5 159 4.2.1.10 Core KB 7 160 4.2.1.11 Core KB 11 160 4.2.1.12 Core KB 12 160 4.2.2 Bulk geochemistry of subsurface offshore sediments 161 4.2.2.1 Core STR 1 161 4.2.2.2 Core STR 2 161 4.2.2.3 Core STR 4 161 4.2.2.4 Core STR 11 162 4.2.2.5 Core STR 12 162 4.2.2.6 Core STR 27 162 4.2.2.7 Core STR 28 163 4.2.2.8 Core KB 3 163 4.2.2.9 Core KB 5 163 4.2.2.10 Core KB 7 164 4.2.2.11 Core KB 11 164 4.2.2.12 Core KB 12 164 4.3 Summary 177
Chapter V The lerissos gulf
5.1 Surface sediments 178 5.1.1 Sediment composition 178 5.1.1.1 Distribution of rock fragments, quartz and heavy minerals 178 5.1.1.1.1 Rock fragments 178 5.1.1.1.2 Quartz 179 5.1.1.1.3 Heavy minerals 179 5.1.1.2 Carbonates 179 5.1.1.3 Organic carbon 179 5.1.1.4 XRD results 180 5.1.2 Bulk geochemistry of offshore sediments and of those on 186 the adjacent coast 5.1.2.1 Silicon 186 5.1.2.2 Aluminium 186 5.1.2.3 Calcium 186 5.1.2.4 Iron 187 5.1.2.5 Potassium 187 5.1.2.6 Magnesium 187 5.1.2.7 Titanium 188 5.1.2.8 Manganese 188 5.1.2.9 Phosphorus 188 5.1.2.10 Barium 189 5.1.2.11 Strontium 189 5.1.2.12 Zinc 190 5.1.2.13 Chromium 190 5.1.2.14 Vanadium 190 5.1.2.15 Zirconium 191 5.1.2.16 Nickel 191 5.1.2.17 Copper 191 5.1.2.18 Lanthanum 192 5.1.2.19 Cobalt 192 5.1.2.20 Beryllium 192 5.1.3. Multivariate Statistical Analysis 204 5.1.3.1 Correlation Matrices 204 5.1.3.2 Cluster analysis 207 9
5.1.3.2.1 Box-Cox transformed data 207 5.1.3.2.2 Raw, untransformed data 207 5.1.3.3 Factor analysis 210 5.1.3.3.1 Box-Cox transformed data 210 5.1.3.3.2 Raw, untransformed data 213 5.2 Burled sediments 216 5.2.1 Description of the cores 216 5.2.1.1 Corel 5 216 5.2.1.2 Core I 8 216 5.2.1.3 Corel 10 216 5.2.1.4 Corel 13 217 5.2.1.5 Corel 14 217 5.2.2. Bulk geochemistry of subsurface offshore sediments 217 5.2.2.1 Corel 5 217 5.2.2.2 Corel 8 218 5.2.2.3 Corel 10 218 5.2.2.4 Corel 13 218 5.2.2.5 Corel 14. 218 5.3 Summary 224
Chapter VI Partition Geochemistry
6.1 Introduction - Rationale for partition work 225 6.2 Surface sediments 226 6.2.1 Results 234 6.2.1.1 Exchangeable fraction 234 6.2.1.2 Carbonate hosted fraction 236 6.2.1.3 Reducible fraction 240 6.2.1.4 Organic and sulphide bound fraction 242 6.2.1.5 Residual fraction 246 6.3 Buried sediments 248 6.3.1 Results 255 6.3.1.1 Corel 14 255 6.3.1.2 Core THR 22 259 6.4 Summary 263
Chapter VII Humic Substances
7.1 Introduction and brief review 265 7.2 Studies on humic substances 267 7.2.1 Utility and objectives of the work 267 7.2.2 Methods and analytical techniques used 270 7.2.2.1 Extraction and purification of humic acids 270 7.2.2.2 Copper complexation 270 7.2.2.3 Ash content determination 271 7.2.2.4 Analytical techniques used 271 7.2.3 Results and Discussion 271 7.2.3.1 Elemental composition 271 7.2.3.2 Infra-Red Spectroscopy Studies 273 7.2.3.3 Metal Humic Acids 278 7.2.3.4 ESPR Studies 279 7.2.3.5 Artificially formed Cu(ll)-HA complexes 285 10
Chapter VIII D iscussion
8.1 Nontransition Elements 288 8.1.1 Alkali Metals 288 8.1.1.1 Potassium 288 8.1.2 Alkaline Earth Metals 289 8.1.2.1 Beryllium 290 8.1.2.2 Magnesium 290 8.1.2.3 Calcium 293 8.1.2.4 Strontium 295 8.1.2.5 Barium 296 8.1.3 The Group lllb Elements 297 8.1.3.1 Aluminium 297 8.1.4 The Group IVb Elements 299 8.1.4.1 Silicon 299 8.1.5 The Group Vb Elements 300 8.1.5.1 Phosphorus 300 8.1.6 The Group lib Elements 301 8.1.6.1 Zinc 301 8.2 Transition Elements 303 8.2.1 The Elements of the First Transition Series 303 8.2.1.1 Titanium 303 8.2.1.2 Vanadium 305 8.2.1.3 Chromium 306 8.2.1.4 Manganese 307 8.2.1.5 Iron 309 8.2.1.6 Cobalt 312 8.2.1.7 Nickel 313 8.2.1.8 Copper 314 8.2.2 The Elements of the Second Transition Series 315 8.2.2.1 Zirconium 315 8.2.3 The Lanthanides 317 8.2.3.1 Lanthanum 317 8.3 Concluding statement 321
References 322
Appendix 1 Simplified geological map of the mainland, and Thassos and 332 Samothraki islands
Appendix II Sample collection and preparation 335
Appendix III Geochemical data for 17 elements determined in 43 336 offshore sediments by both methods A and B
Appendix IV Calcium carbonate and organic carbon contents 344
Appendix V Multivariate Statistical Analysis 349
Appendix VI Element partitioning in selected surface and subsurface 352 sediments
Appendix VII Geochemical data 376 11
LIST OF TABLES Page
1.1 The drainage area, average yield per month, average yield per year 31 and total load per year of the Evros, Nestos and Strymon rivers.
2.1 Proportions of Ca, Mg, Sr, Pb, Cr, Mn, Fe, Co, Ni, Cu and Zn leached 53 by : HOAc 25% overnight (A); HOAc 20% for 15 min. (B) ; HOAc 15% for 15 min. (after heating the samples at 350° C for 2 hours) (C).
2.2 Organic carbon and humic acids (H.A.) content of sample STR 8; 55 organic carbon content of the H .A.; and organic carbon content of the sample, present as H.A.
3.1 Mineralogical composition of selected samples from Samothraki 63 plateau, as determined through XRD analysis.
3.2 Robust correlation matrices of the raw, untransformed (a) and the 93 Box-Cox transformed (b) data of the Samothraki plateau.
3.3 Factor analysis of raw data of the Samothraki plateau. Five factor 100 model: factor loadings and variance accounted for by components, computed on the basis of the robust C.M. of the raw data.
3.4 Samothraki plateau raw data. List of eigen values of all the 101 components of the data.
4.1 Mineralogical composition of selected samples from the western 112 part of the study area, as determined through XRD analysis.
4.2 Robust correlation matrices of the raw, untransformed (a) and the 140 Box-Cox transformed (b) data of the Strymonikos group.
4.3 Factor analysis of raw data of the Strymonikos group. Four factor 146 model: factor loadings and variance accounted for by components, computed on the basis of the robust C.M. of the raw data.
4.4 Strymonikos group raw data. List of eigen values of all the 147 12
components of the data.
4.5 Robust correlation matrices of the raw, untransformed (a) and the 149 Box-Cox transformed (b) data of the Kavala group.
4.6 Factor analysis of raw data of the Kavala group. Four factor 155 model: factor loadings and variance accounted for by components, computed on the basis of the robust C.M. of the raw data.
4.7 Kavala group raw data. List of eigen values of all the 156 components of the data.
5.1 Mineralogical composition of selected samples from lerissos 180 gulf, as determined through XRD analysis.
5.2 The robust correlation matrices of the raw, untransformed (a) 205 and the Box-Cox transformed (b) data of the lerissos gulf.
5.3 Factor analysis of Box-Cox transformed data of the lerissos gulf. 211 Four factor model: factor loadings and variance accounted for by components, computed on the basis of the robust C.M. of the Box-Cox transformed data.
5.4 lerissos gulf Box-Cox transformed data. List of eigenvalues 212 of all the components of the data.
5.5 Factor analysis of raw data of the lerissos gulf. Four factor 214 model: factor loadings and variance accounted for by components, computed on the basis of the robust C.M. of the raw data.
5.6 lerissos gulf raw data. List of eigen values of all the components 215 of the data.
6.1 Average proportions of each of 20 elements in the exchangeable (E), 228 carbonate hosted (C), reducible (OX), organic and sulphide bound (OR+S), and residual (R) fractions of selected samples from lerissos gulf (A), Strymonikos group (B), Kavala group (C), and Samothraki plateau (D), and from offshore areas (E); also, means of the 13
previous estimates (F).
6.2 Calcium content determined by gas chromatography and total calcium 238 content associated with the second and third leachate.
6.3 Total calcium carbonate content of selected samples and proportions 239 of the total Ca content of the same samples associated with the second and the third leachate.
6.4 Organic carbon content of selected samples and proportions of total Cu 244 in the fourth leachate of the same samples.
7.1 Sediments from which humic acids have been extracted. 270
7.2 Major elemental composition (on a moisture- and ash- free basis), 272 atomic ratios and ash contents of sedimentary and soil humic acids.
7.3 Humic acids contents of selected sediments; Fe, Cu, V and Mn 278 concentrations in these sediments (A), and in the respective H.A. (B); percentages of total Fe, Cu, V and Mn concentrations : associated with H.A. (C), and bound to the organic/sulphide hosted fraction (D).
7.4 ESR parameters of organic free radicals in sediment, soil and aquatic 280 humic acids.
7.5 ESR data for Fe3+ complexes with sediment humic acids. 282
7.6 Fe, Cu, V and Mn total concentrations (determined by I.C.P.E.S.) 287 in sedimentary humic acids saturated with 0.1 M Cu2+ solution.
8.1 Means (|ig/g) and robust means (pg/g) of raw data, and robust 319 means (|ig/g) of Box-Cox transformed data, of 20 elements determined in the four groups of samples.
II 1.1 K, Be, Mg, Ca, Sr, Ba, La, Ti and P contents (pg/g) of 43 offshore 336 sediments determined by the method A. 14
111.2 V, Cr, Mn, Fe, Co, Ni, Cu and Al contents (pg/g) of 43 offshore 338 sediments determined by the method A.
111.3 K, Be, Mg, Ca, Sr, Ba, La, Ti and P contents (|j.g/g) of 43 offshore 340 sediments determined by the method B.
111.4 V, Cr, Mn, Fe, Co, Ni, Cu and Al contents (|xg/g) of 43 offshore 342 sediments determined by the method B.
IV.1 Calcium carbonate and organic carbon contents of 206 surface 344 offshore samples. 15
LIST OF FIGURES
Page 1.1 Map of Greece. The study area is shown between lines. 29
1.2 Bathymetric map of the Samothraki plateau. 35
1.3 Bathymetric map of the western part of the study area. 35
1.4 Simplified geological map of the mainland, and the Thassos 40 and Samothraki islands.
1.5 Distribution of sediment types in the study area. 36
2.1 Scattergrams of the elements concentrations determined by the 44 LiB02-TR0K method (A) versus the corresponding ones determined
by the GENHF-GEN4 method(B).
3.1 Samothraki plateau : map representing the offshore surface and 65 subsurface sample locations.
3.2 Samothraki plateau : map representing the onshore sample locations. 66
3.3 Samothraki plateau surface sediments : distribution of rock 67 fragments in the coarse fraction.
3.4 Samothraki plateau surface sediments : distribution of quartz in 67 the coarse fraction.
3.5 a Samothraki plateau surface sediments : distributtion of heavy 68 mineral content in the coarse fraction. 16
Samothraki plateau surface sediments : distribution of total carbonates 71
Samothraki plateau surface sediments : distribution of organic carbon. 71
Samothraki plateau and adjacent coast: regional variation in the Si 81 content of the surface samples.
Samothraki plateau and adjacent coast: regional variation in the Ca 81 content of the surface samples.
Samothraki plateau and adjacent coast: regional variation in the Al 82 content of the surface samples.
Samothraki plateau and adjacent coast: regional variation in the Fe 82 content of the surface samples.
Samothraki plateau and adjacent coast: regional variation in the K 83 content of the surface samples.
Samothraki plateau and adjacent coast: regional variation in the Mg 83 content of the surface samples.
Samothraki plateau and adjacent coast: regional variation in the Ti 84 content of the surface samples.
Samothraki plateau and adjacent coast : regional variation in the Sr 84 content of the surface samples.
Samothraki plateau and adjacent coast: regional variation in the P 85 content of the surface samples.
Samothraki plateau and adjacent coast: regional variation in the Ba 85 content of the surface samples.
Samothraki plateau and adjacent coast: regional variation in the Mn 86 content of the surface samples. 17
Samothraki plateau and adjacent coast : regional variation in the Zr 86 content of the surface samples.
Samothraki plateau and adjacent coast : regional variation in the Cr 87 content of the surface samples.
Samothraki plateau and adjacent coast: regional variation in the V 87 content of the surface samples.
Samothraki plateau and adjacent coast : regional variation in the Zn 88 content of the surface samples.
Samothraki plateau and adjacent coast: regional variation in the La 88 content of the surface samples.
Samothraki plateau and adjacent coast: regional variation in the Ni 89 content of the surface samples.
Samothraki plateau and adjacent coast: regional variation in the Cu 89 content of the surface samples.
Samothraki plateau and adjacent coast: regional variation in the Co 90 content of the surface samples.
Samothraki plateau and adjacent coast: regional variation in the Be 90 content of the surface samples.
Cluster analysis of Samothraki plateau data; robust dendograms : 97 (a) Box-Cox transformed data, (b) raw data.
Core THR 7 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 104 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
Core THR 22 : (a) description; (b) downcore plots of K, Be, Mg, Ca, 105 Sr, Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
Strymonikos and Kavala areas : map representing the offshore surface 114 and subsurface sample locations. 18
4.2 Strymonikos and Kavala areas : map representing the onshore sample 115 locations.
4.3 Strymonikos and Kavala surface sediments : distribution of rock 11 6 fragments in the coarse fraction.
4.4 Strymonikos and Kavala surface sediments : distribution of quartz 116 in the coarse fraction.
4.5 Strymonikos and Kavala surface sediments : distribution of heavy 117 minerals content in the coarse fraction.
4.6 Strymonikos and Kavala surface sediments : distribution of selected 117 heavy minerals in the 2 0 to 3 0 fraction.
4.7 Strymonikos and Kavala surface sediments : distribution of biogenic 118 components in the coarse fraction.
4.8 Strymonikos and Kavala surface sediments : distribution of total 119 carbonates*
4.9 Strymonikos and Kavala surface sediments : distribution of organic 119 carbon.
4.10 Strymonikos and Kavala areas and adjacent coast: regional variation 129 in the Si content of the surface samples.
4.11 Strymonikos and Kavala areas and adjacent coast: regional variation 129 in the Al content of the surface samples.
4.12 Strymonikos and Kavala areas and adjacent coast: regional variation 130 in the Ca content of the surface samples.
4.13 Strymonikos and Kavala areas and adjacent coast: regional variation 130 in the Fe content of the surface samples.
4.14 Strymonikos and Kavala areas and adjacent coast: regional variation 131 in the K content of the surface samples. 19
4.15 Strymonikos and Kavala areas and adjacent coast: regional variation 131 in the Mg content of the surface samples.
4.16 Strymonikos and Kavala areas and adjacent coast : regional variation 132 in the Ti content of the surface samples.
4.17 Strymonikos and Kavala areas and adjacent coast: regional variation 132 in the P content of the surface samples.
4.18 Strymonikos and Kavala areas and adjacent coast: regional variation 133 in the Mn content of the surface samples.
4.19 Strymonikos and Kavala areas and adjacent coast: regional variation 133 in the Ba content of the surface samples.
4.20 Strymonikos and Kavala areas and adjacent coast : regional variation 134 in the Sr content of the surface samples.
4.21 Strymonikos and Kavala areas and adjacent coast: regional variation 134 in the Zr content of the surface samples.
4.22 Strymonikos and Kavala areas and adjacent coast: regional variation 135 in the V content of the surface samples.
4.23 Strymonikos and Kavala areas and adjacent coast: regional variation 135 in the Cr content of the surface samples.
4.24 Strymonikos and Kavala areas and adjacent coast: regional variation 136 in the Zn content of the surface samples.
4.25 Strymonikos and Kavala areas and adjacent coast: regional variation 136 in the Ni content of the surface samples.
4.26 Strymonikos and Kavala areas and adjacent coast: regional variation 137 in the La content of the surface samples.
4.27 Strymonikos and Kavala areas and adjacent coast: regional variation 137 in the Cu content of the surface samples. 20
4.28 Strymonikos and Kavala areas and adjacent coast: regional variation 138 in the Co content of the surface samples.
4.29 Strymonikos and Kavala areas and adjacent coast : regional variation 138 in the Be content of the surface samples.
4.30 Cluster analysis of Strymonikos group data; robust dendograms : 143 (a) Box-Cox transformed data, (b) raw data.
4.31 Cluster analysis of Kavala group data; robust dendograms : 152 (a) Box-Cox transformed data, (b) raw data.
4.32 Core STR 1 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 165 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
4.33 Core STR 2 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 166 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
4.34 Core STR 4 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 167 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
4.35 Core STR 11 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 168 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
4.36 Core STR 12 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr,169 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
4.37 Core STR 27 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 170 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
4.38 Core STR 28 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr,171 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
4.39 Core KB 3 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 172 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
4.40 Core KB 5 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 173 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P. 21
4.41 Core KB 7 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 174 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
4.42 Core KB 11 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 175 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
4.43 Core KB 12 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 176 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
5.1 lerissos gulf : map representing the offshore surface and 181 subsurface sample locations.
5.2 lerissos gulf : map representing the onshore sample locations. 182
5.3 lerissos gulf surface sediments : distribution of rock fragments in 1 83 the coarse fraction.
5.4 lerissos gulf surface sediments : distribution of quartz in the coarse 183 fraction.
5.5 lerissos gulf surface sediments : distribution of the heavy minerals 184 content in the coarse fraction.
5.6 lerissos gulf surface sediments : distribution of selected heavy 184 minerals in the 2 o to 3 o fraction.
5.7 lerissos gulf surface sediments : distribution ofj total carconates 185
5.8 lerissos gulf surface sediments : distribution of organic carbon. 185
5.9 lerissos gulf and adjacent coast : regional variation in the Si content 194 of the surface samples.
5.10 lerissos gulf and adjacent coast : regional variation in the Al content 194 of the surface samples.
5.11 lerissos gulf and adjacent coast: regional variation in the Ca content 195 of the surface samples. 22
5.12 lerissos gulf and adjacent coast : regional variation in the Fe content 195 of the surface samples.
5.13 lerissos gulf and adjacent coast : regional variation in the K content 1 96 of the surface samples.
5.14 lerissos gulf and adjacent coast : regional variation in the Mg content 196 of the surface samples.
5.15 lerissos gulf and adjacent coast : regional variation in the Ti content 197 of the surface samples.
5.16 lerissos gulf and adjacent coast : regional variation in the Mn content 197 of the surface samples.
5.17 lerissos gulf and adjacent coast : regional variation in the P content 198 of the surface samples.
5.18 lerissos gulf and adjacent coast: regional variation in the Ba content 198 of the surface samples.
5.19 lerissos gulf and adjacent coast : regional variation in the Sr content 199 of the surface samples.
5.20 lerissos gulf and adjacent coast : regional variation in the Zn content 199 of the surface samples.
5.21 lerissos gulf and adjacent coast : regional variation in the Cr content 200 of the surface samples.
5.22 lerissos gulf and adjacent coast : regional variation in the V content 200 of the surface samples.
5.23 lerissos gulf and adjacent coast : regional variation in the Zr content 201 of the surface samples.
5.24 lerissos gulf and'adjacent coast : regional variation in the Ni content 201 of the surface samples. 23
5.25 lerissos gulf and adjacent coast : regional variation in the Cu content 202 of the surface samples.
5.26 lerissos gulf and adjacent coast : regional variation in the La content 202 of the surface samples.
5.27 lerissos gulf and adjacent co a st: regional variation in the Co content 203 of the surface samples.
5.28 lerissos gulf and adjacent coast: regional variation in the Be content 203 of the surface samples.
5.29 Cluster analysis of lerissos gulf data; robust dendograms: (a) Box-Cox 208 transformed data, (b) raw data.
5.30 Core I 5 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 219 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
5.31 Core I 8 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 220 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
5.32 Core I 10 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 221 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
5.33 Core I 13 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 222 • Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
5.34 Core I 14 : (a) description; (b) downcore plots of K, Be, Mg, Ca, Sr, 223 Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P.
6.1 Map representing the sites of the samples subjected to partition 227 analysis.
6.2 Average proportions of Li, K, Mg, Ca, Sr, Ba, Al, La, Cr, Mn, Fe, Co, 232 Be, V, Ni, Cu, Zn, Pb and P in the five fractions of selected surface samples from the lerissos gulf.
6.3 Average proportions of Li, K, Mg, Ca, Sr, Ba, Al, La, Cr, Mn, Fe, Co, 232 24
Be, V, Ni, Cu, Zn, Pb and P in the five fractions of selected surface samples from the Strymonikos area.
6.4 Average proportions of Li, K, Mg, Ca, Sr, Ba, Al, La, Cr, Mn, Fe, Co, 232 Be, V, Ni, Cu, Zn, Pb and P in the five fractions of selected surface samples from the Kavala area.
6.5 Average proportions of Li, K, Mg, Ca, Sr, Ba, Al, La, Cr, Mn, Fe, Co, 233 Be, V, Ni, Cu, Zn, Pb and P in the five fractions of selected surface samples from the Samothraki plateau.
6.6 Average proportions of Li, K, Mg, Ca, Sr, Ba, Al, La, Cr, Mn, Fe, Co, 233 Be, V, Ni, Cu, Zn, Pb and P in the five fractions of selected onshore samples.
6.7 Average proportions of Li, K, Mg, Ca, Sr, Ba, Al, La, Cr, Mn, Fe, Co, 233 Be, V, Ni, Cu, Zn, Pb and P in the five fractions, based on the element means of their estimates in the five groups of samples.
6.8 Plot of the total calcium carbonate content of selected samples versus 239 the corresponding proportions of the total Ca content leached by NaOAc/HOAc.
6.9 Plot of the proportions of the total Cu leached by H20 2 in selected 245
samples versus the organic carbon content of the same samples.
6.10 Core 114: downcore plots of the proportions of Li, K, Mg, Ca, Sr, Ba, 249 Al, La, Cr, Mn, Fe, Co, Be, V, Ni, Cu, Zn, Pb and P associated with each of the five fractions.
6.11 Core THR 22 : downcore plots of the proportions of Li, K, Mg, Ca, Sr, 252 Ba, Al, La, Cr, Mn, Fe, Co, Be, V, Ni, Cu, Zn, Pb and P associated with each of the five fractions.
7.1 Map showing the sites of the samples from which humic acids have 269 been extracted.
7.2 I.R. spectra of sediment humic acids. 275 25
7.3 Scatter plot of H content versus organic free radical content, in 281 sediment humic acids.
7.4 Scatter plot of C/H ratio versus organic free radical content, in 281 sediment humic acids.
7.5 E.S.P.R. spectra of sediment humic acids. 283
7.6 I.R. spectra of artificially formed Cu(ll) complexes with sediment 286 humic acids, islolated from sample 1.
VI.1 Proportions of Li, K, Mg, Ca, Sr, Ba, Al, La, Cr, Mn, Fe, Co, Ni, Cu, 353 Be, V, Zn, Pb and P associated with each of the five fractions of the surface samples subjected to partition analysis : I 5, I 22, I 38, I 46, I 50, STR 4, STR 50, STR 74, KB 2, KB 31, KB 55, KB 100, KB 113, KB 126, THR 1, THR 4, THR 11, THR 31, THR 60, THR 84, THR 122, B12, B22, B 27, B 83, B 89, BSTR 4.
VI.2 Core 114: Proportions of Li, K, Mg, Ca, Sr, Ba, Al, La, Cr, Mn, Fe, 368 Be, V, Co, Ni, Cu, Zn, Pb and P associated with each of the five fractions of samples collected from 3, 16, 41, 71, 101, 131 and 161 cm depth.
VI.3 Core THR 22: Proportions of Li, K, Mg, Ca, Sr, Ba, Al, La, Cr, Mn, Fe, 372 Be, V, Co, Ni, Cu, Zn, Pb and P associated with each of the five fractions of samples collected from 2, 24, 64, 94, 122, 149 and 176 cm depth. 26
ACKNOWLEDGEMENTS
This research has been accomplished under the supervision of Dr. D.S. Cronan, whom I would like to thank for his guidance, numerous discussions and critical comments.
I would like to thank Dr. S.A. Moorby for his help and suggestions over the first two years of this study.
Dr. P. Bush is thanked for his substantial help during the final stages of this work.
My thanks are extended to the Greek Institute of Geology and Mineral Exploration (I.G.M.E.) and, particularly, to Dr. C. Perissoratis, for providing the offshore surface and subsurface samples and various geological information relevant to the study area.
Further, I express my gratitude to Prof. N. Senesi, of the University of Bari, for inviting me to his laboratory and giving me the opportunity to study humic acids under his guidance.
I would also like to thank EEC for providing funds for the fees of the first two years, and NATO for covering the fees of the third year and supporting me, financially, at the final 11 months of this study.
I would particularly like to thank :
my parents for supporting me both financially and emotionally, and
Leonidas Mantzos for many lengthy discussions and significant suggestions during the course of this work and also for his understanding and valuable encouragement. 27
CHAPTER I
INTRODUCTION
1.1 General introduction.project outline and aims
The area under investigation is the northern part of the Aegean Sea. It stretches eastwards of the Halkidiki peninsula, north Greece, to the Hellenic - Turkish border (Fig. 1.1). It covers an offshore area of 8,116 km2.
The eastern part of the area is physiographically described as the Samothraki plateau and extends from the shoreline of Thraki southwards to the shelf break, and from Thassos to Samothraki island. It covers an area of 5,800 km2 with 140 km of mainland coastline (Fig. 1.2).
The western part of the study area lies between the Halkidiki peninsula and the Thassos island, and from the shoreline up to the shelf break. It covers an area of 2,316 km2 , with 240 km of mainland coastline (Fig. 1.3). Various separate morphological units occur in this part. These are: a. the lerissos, Strymonikos and Kavala gulfs; b. the Strymonikos plateau; and c. the Strymonikos triangle, a triangular feature lying between the Kavala gulf and the Strymonikos plateau.
The aims of this thesis are to study the regional variations in the major and minor element composition of sediments in the North Aegean Sea and to determine the various factors controlling their geochemistry. The influence of such factors as bottom topography, water currents, source area, lithology and mineralogy on the sediment composition has been studied. A subsidiary aspect of the work has been to examine the placer deposits potential in the offshore areas of the northern part of the Aegean Sea. Five hundred and eighty-seven offshore surface samples have been chemically analyzed in bulk. The data obtained were subjected to multivariate statistical treatment (cluster and factor analysis ). The analysis of a hundred and twenty-four beach-, river-, stream-, and lake-sediments collected from the adjacent mainland 28 and the Thassos island aims at examining the importance of river runoff and coastal erosion in determining offshore compositional variations in the area. The sediments in nineteen cores , collected using a 3m gravity corer, were investigated to determine the geochemical evolution of the offshore sediments undergoing diagenesis and the changing elemental supply to them through time. All the samples were analyzed chemically for twenty one elements using Inductively Coupled Plasma Emission Spectroscopy (I.C.P.E.S.), following a fusion attack with lithium metaborate, (Thompson and Walsh,1983; p.85). The data (concerning offshore surface and subsurface, as well as onshore samples) are provided in Appendix VII.
The possible sources of the elements, their mode of incorporation in the sediments, their relative availability to the biosphere and their behaviour in the secondary environment have been investigated by examining the partitioning of elements among the sediment components. This investigation was based on the selective chemical leaching of representative samples. The well known sequential extraction analytical procedure of Tessier et al (1979), with a few modifications and alterations , was followed because it deals with all the phases that need to be identified. Thus, the partitioning of twenty five elements within the exchangeable, carbonate hosted, reducible, organic matter-sulphide minerals, and residual phases was obtained. The data are provided in Appendix VII.
Total carbonate content was determined in many samples. Representative samples were subjected to X-Ray Diffraction analysis. Selected samples were studied using Scanning Electron Microscopy (S.E.M.). This supplementary work was designed to support and clarify suggestions concerning elemental distribution patterns based on geochemical analysis.
Finally, the important role that humic substances play in metal cycling in the marine environment was studied by examining humic acids isolated from some marine and beach sediments. The humic acids were characterized by determining their elemental composition and chemical functionalities. Both Infra-Red (I.R.) and Electron Spin Paramagnetic Resonance (E.S.P.R.) Spectroscopy were applied. The binding ability of humic acids (H.A.) towards paramagnetic metal ions was investigated by complexing Cu+2 in sites of the H.A. that involve functional groups. The resistance of the above complexes to exhaustive water-washing and to proton-exchange by strong acids was investigated in order to examine their association with transport and immobilization phenomena, involving metal ions, in the marine environment. 29
0
F ig . 1.1 Map of Greece. The study area is shown between lines. 30
1.2 Thesis structure
The first chapter of the thesis deals with the geological setting and with some general information about the study area. The second chapter concerns the analytical techniques used. The following chapters (3 to 5) deal with the lithology, mineralogy and bulk geochemistry of both surface and buried sediments collected from the North Aegean Sea. Chapter 6 covers the partition geochemistry of selected surface and buried sediments. Chapter 7 deals with humic acids (H.A.); it includes a brief review on humic material and it covers the study of H.A. present in selected samples. Finally, chapter 8 comprises the interpretation of the geochemical data on the North Aegean Sea sediments.
1.3 General morphological and geological setting
The land to the north of the study area mainly comprises formations consisting of metamorphic and igneous rocks and Neogene-Quaternary sediments (Fig. 1.4). It is drained by large and small rivers and by seasonal streams.
1.3.1 Main rivers
The largest rivers from east to west are: a. The Evros river on the Hellenic - Turkish border. It is the main river of the area, forming an extensive delta before flowing into the sea. The Evros river has a length of 928 km and a drainage area of 27,465 km2. Its average annual water flow is 615 m3/sec and its average yield per year approximately 103 m3/sec (Table 1.1). b. The Filiouris river forming its delta close to Xylagani. c. The Nestos river in the western-most part of the Samothraki plateau. It forms an extensive delta before flowing into the sea. The Nestos river has a drainage area of 4,874 km2, its average yield per year being 58 m3/sec (Table 1.1). d. The Strymon river in the northwest part of the Strymonikos area. This river drains the metamorphic and igneous formations of the Serbo-Macedonian and Rhodope Massifs in Bulgaria; in Greece, its drainage basin consists of Neogene sediments. The basin of Strymon river narrows where the river reaches the sea , 31
and therefore little area is left for the formation of a delta plain. The Strymon river has a total drainage area of 10,937 km2 . Its average yield per year is approximately 110 m3/sec (Table 1.1).
e. The Asprolakkos river, flowing out on the west coast of lerissos gulf.
Table 1.1: The drainage area , average yield per month, average yield per year and total load per year of the Evros, Nestos and Strymon rivers (after Therianos, 1974).
River Drainage Average yield per month area (m3sec'1) (km2)
J F M A MJ
Evros 27,465 212 200 239 136 92 64 Nestos 4,874 106 95 87 86 89 57 Strymon 10,937 104 118 135 181 228 156
River Average yield per month Average Total (m3sec_i) yield load per year per year J A S ON D (m3sec‘ 1') (m3)
Evros 32 22 16 19 72 35 103 130,000 Nestos 24 11 11 13 39 77 58 65,000 Strymon 65 31 27 56 83 135 110 84,000
1.3.2 Composition of rocks on the mainland
The mainland to the north, lying between the Strymon and Evros rivers, comprises crystalline metamorphic rocks of the Rhodope Massif. Marbles and phyllites cover a small area lying westwards of the Holocene alluvial deposits of the Alexandroupolis district. Some granites and granodiorites occur between Aghios Haralambos and the previously mentioned area which is mainly composed of marbles (Fig. 1.4). Mixed sulphide mineralization is hosted in the metamorphic and igneous rocks of eastern Thraki, (Billett and Nesbitt,1986). These deposits occur in a mineralized zone extending throughout the province of Alexandroupolis and as far as the Bulgarian border. The biggest ore body of this group is close to the village of Kirki in a faulted zone several hundred meters long, trending N-S. These deposits are considered to be 32 hydrothermal in origin and to be related to late volcanicity in Thrace. The ores consist of pyrite, galena, sphalerite, chalcopyrite and wurtzite, the Pb-Zn content reaching the 8%. Total ore reserves of all deposits at Kirki are estimated to be several hundred thousand tonnes, (Marinos, 1982). Close to Abdira, amphibolites, gneisses and molassic formations with lignite occur, surrounded by Holocene alluvial deposits (Fig. 1.4). The coastal area lying between Loutra Eleftheron and Kavala comprises granites, granodiorites and monzonites (Fig. 1.4). Near Kavala, Fe and Mn hydrothermal ore deposits occur, with limonite and pyrolusite as the predominant minerals. This ore grade is 27-43% Fe oxide and 18-43% Mn oxide, containing small amounts of Zn, Pb and As. Reserves are estimated to be several hundred thousand tonnes, (Marinos,1982). About 15 km northwest of Nea Peramos mixed sulphide occurrences are encountered.
The Serbo-Macedonian Massif lies on the west of the Rhodope Massif, extending from the river Strymon westwards and constituting the largest part of the Halkidiki peninsula. The deepest horizons of the Serbo-Macedonian Massif, that are exposed, comprise biotitic gneisses, whereas the upper parts of the Massif consist of gneisses, mica schists and mafic volcanics. Amphibolites and gneisses are found in the area lying between the Strymon river and the lerissos gulf. A large active mine of mixed sulphides exists in the Olympias district. Close to Stratoni, extensive masses of amphibolites occur. In the same area, significant mixed sulphide ores (e.g. Madem Lakos), manganese oxide deposits and small Cu-bearing porphyry stocks are found. A mineral processing plant operates on the coast near Stratoni. Granites and granodiorites are encountered in the region close to the town of lerissos and the eastern coastal area of lerissos gulf (Fig. 1.4). Abandoned chromite mines occur further inland west of lerissos. In the southern part of the lerissos g u lf, east of Nea Roda, serpentines, amphibolites, peridotites and dunites with chromite mineralization occur. Also, on the east of Nea Roda, a small ultrabasic igneous complex of ophiolitic character crops out.
1.3.3 Composition of rocks on islands
Thassos and Samothraki islands are largely composed of metamorphic and igneous rocks (Fig. 1.4).
The central part of Thassos and part of its coastal area comprise amphibolites, gneisses and schists with marble intercalations. Most of the rest of the island consists of marbles, while small unmetamorphosed upper Miocene lacustrine deposits lie at 33 the southwestern coast and Holocene alluvial deposits at the northwestern coast (Fig. 1.4). A number of Pb-Zn deposits occur in this island, being hosted by carbonate rocks and schists. According to Vavelidis and Amstutz (1981 ), the carbonate rocks are mainly dolomitic marbles , while the schists are either mica-carbonate-albite-quartzite schists or carbonate-mica-quartzite schists. Ore minerals in the ore-bearing horizons are cerussite, smithsonite, galena, sphalerite, pyrite, marcasite, baryte and Fe.Mn-oxides. (Vavelidis and Amstutz, 1983). "Soft" iron ores of hydrothermal origin, mainly limonite and goethite, form lenses within the marbles; the grade of the ore is about 48% Fe and 2% Mn, with ore reserves in excess of 20 million tonnes (Marinos,1982). These ores often contain baryte and some Cu, As and other impurities.
The northeast, east and west coastal areas of Samothraki island consist of Holocene alluvial deposits (Fig. 1.4). The northwest region is covered by Mio-Pliocene marine deposits. There are small occurrences of Eocene-Oligocene acid to intermediate volcanics on the eastern part of the island. A small occurrence of Upper Eocene conglomerates , marls, sandstones and limestones with lignites lies on the northern coastal area of Samothraki (Fig. 1.4). The remaining central part of the island consists of diabases and Tertiary acid igneous rocks.
1.4 Bathymetry
The area studied is divided geographically by Thassos island into two parts, the Samothraki plateau to the east and the Strymonikos area to the west.
The bathymetry of the area to the east of Thassos island (Fig. 1.2) is different from that to the west. The shelf break is at a water depth of about 120 m. Two embayments occur, one at Porto Lago and one at Xylagani. The major feature of the Samothraki plateau is a terrace occurring in its central part, at a water depth of about 60 to 70 m (Fig. 1.2). This terrace, although it is not well defined, represents an ancient coastline and a nearshore environment which was formed in a period of lower sea-level stands, during the last Ice-Age (Moorby et al, in press ). The sea floor of the Samothraki plateau has a smooth morphology with a slope gradient of about 0.1%. In its eastern part, to the west and north of Samothraki island minor sea floor irregularities occur. They consist of small channels and valleys which trend in a NE-SW direction. A minor swell of the sea floor occurs between Samothraki island and Alexandroupolis district. South-east of the river Filiouris, a canyon cuts through the terrace that is located in the central part of the plateau. As this canyon is not fed 34
by a modern river, it was most probably formed by a pre- or inter- glacial river, at a lower sea level. To the south of the Samothraki plateau the sea floor gradient increases abruptly and the north Aegean trough begins. This trough is an elongate feature with an ENE-WSW orientation and a depth exceeding 1,000 m.
The Strymonikos area contains three gulfs - the Strymonikos, Kavala and lerissos gulfs - the Strymonikos triangle and the Strymonikos plateau. The latter comprises a smooth terrace at a water depth of about 100 to 110 m (Fig. 1.3). The shelf break is at a water depth of 120 to 130 m. The most striking feature of the Strymonikos area is an east-west canyon (Fig. 1.3) lying just offshore between the mouth of the Strymon river and the Loutra Eleftheron district. In the southwest part of the Strymonikos gulf, south of the mouth of the Strymon river , a north-east trending ridge occurs, the Strymonikos ridge.
The lerissos gulf appears to be an embayment of simple structure. The sea floor slopes evenly and steeply down to a flat bottom at about 75 m (Fig. 1.3). A sill, occurring at the outer part of the gulf, separates it from the Strymonikos plateau.
The Kavala gulf is much shallower than the lerissos and its bottom slopes gently. The floor of the gulf is almost separated from the Strymonikos plateau by a sill. This sill is breached by a small submarine valley, the Kavala valley. Off the mouth of the Kavala gulf an upper terrace, smaller than, and north-east of, the one mentioned earlier, occurs at a water depth of about 50 m. The southwestern edge of the upper terrace slopes steeply down to the lower terrace. The area lying between the Kavala gulf and the Strymonikos plateau is triangular, it will therefore be referred to as the Strymonikos triangle (Fig. 1.3).
The isolation of the lerissos gulf and the connection of the gulfs of Kavala and Strymonikos with the Strymonikos plateau indicate that during low sea level stands the basin of the lerissos gulf was not connected with the open sea, whereas the basins of the inner parts of both Kavala and Strymonikos gulfs were.
It was shown by Perissoratis et al (in press) that during upper Pleistocene-lower Holocene low sea level stands, the Nestos river was flowing out on the eastern coast of the Kavala gulf and its route continued counterclockwise. At that time, the river mouth was lying at the channel formed in the middle of the ridge that occurs between Nea Peramos and Thassos. These suggestions were based on seismic profiles and data on land. 35
Fig. 1.2 Bathymetric map of the Samothraki plateau.
Fig. 1.3 Bathymetric map of the western part of the study area.
A: Strymonikos plateau; B: Strymonikos triangle; C: Strymonikos ridge; D: Strymonikos gulf; E: lerissos gulf; F: Kavala gulf. clay
SAMOTHRAKI iSL.
« Fig. 1.5 Distribution of sediment types in the study area o> 37
1.5 Hydrography
Available data from the Hydrographic Service in Greece (pers. commun.) indicate that the currents form a generally counter clockwise gyre in the gulfs of lerissos, Strymonikos and Kavala. On the Strymonikos plateau, the general direction of the water movement is from east to west close to the shore, and from west to east further offshore. On the Samothraki plateau, the prevailing winds are from the south and south-east; the surface currents flow from east to west close to the shore and from west to east further offshore. On the eastern part of the Samothraki plateau, a strong current flows close to the coast ant towards the west.
1.6 Nature of sediments offshore
1.6.1 Strvmonikos area
On the basis of seismic data (3.5KHz and Uniboon), Perissoratis et al (in press) showed that a reasonable thickeness of sediment exists in the following parts of the Strymonikos area: a. in the central parts of the gulfs; b. in the inner shelf; and c. beyond the shelf break.
In these areas an erosional phase was recognized in the profiles, which was attributed to the regressive phase prior to the late Wurmian transgression. Two seismic units, a lower and an upper one, were discerned above the unconformity. The lower unit has a variable thickness since it is filling the channels and valleys of the erosional surface. It consists of coarse sediments, that would have been deposited during the fluvial phase before the onset of the transgression (Perissoratis et al, in press). In the area concerned, the upper unit has a constant thickness. It consists of sills, fine sands and clays that were, probably, deposited later when open marine conditions predominated.
1.6.2 Samothraki. plateau
As a result of the study of the seismic profiles of the Samothraki plateau, Perissoratis et al (1987) suggested that significant thickness of sediment occurs in the following 38 three areas: a. near the Evros river mouth; b. near the Nestos river mouth; and c. seawards of the shelf break.
1.6.3 General discussion
Surface sediments throughout the study area show considerable variation (Fig.1.5). There are marked differences between the shelf sediments to the west and east of Thassos island.
Sands and silty sands occur in both the nearshore and the outershelf parts of the Strymonikos area (Fig.1.5). By contrast, the inner shelf of the Strymonikos area as well as the central parts of the lerissos, Strymonikos and Kavala gulfs are covered by finer-grained sediments (silty clays and clayey silts).
Conversely, on the Samothraki plateau sands and silty sands occur on the greater part of the shelf with scattered occurrences of sands and sand/silt/clay units (Fig. 1.5). Sandy silts and clayey silts occur locally in the following areas: a. close to the Nestos river mouth; b. close to the Evros river mouth; c. at the inner part of the embayment situated close to Porto Lago; d. at the inner part of the embayment located close to Xylagani; and e. southwards of Thassos island.
Part of the area lying between Porto Lago and the Nestos river mouth, as well as a rather more extensive area situated between Alexandroupolis and Samothraki are covered by sands. The sands of the Samothraki plateau are mostly fine grained. Medium to coarse sands occur around the two islands. 39
1.7 Previous geochemical and mineraloaical investigations
There are very few published data on the geochemistry and mineralogy of the North Aegean Sea sediments. Marinos (1955) reported the presence of magnetite with a 6 to
8 % Ti02 content in beach samples collected from the Alexandroupolis district. In a later work, Marinos et al (1976) reported that the Ti-magnetite concentration in these beach sands is between 0.2 and 2.7%, the Ti-magnetite containing 2.7 to 4.8% Ti. The same authors analyzed samples from the sea floor off Alexandroupolis. They determined a 1 to 2.8% Ti-magnetite content in the sand fraction and a total Ti-concentration of 2 to 5%. Later, Karamatzanis et al (1977) reported the presence of hematite, magnetite and titanomagnetite in surface and buried sediments collected off the Alexandroupolis district. Pe and Panagos (1971) studied the mineralogy of beach sands collected from the area lying between the Nestos river mouth and Alexandroupolis. They found that the heavy mineral content of the 2.0 0 to 3.5 0 fraction ranges from 2.22 to 7.41% and that it contains variable proportions of heavy minerals. In 1975, Papadakis reported the presence of black sands in a coastal zone, some three kilometers long, near Loutra Eleftheron. He identified ilmenite, pyrolusite, hornblende, garnet, epidote, zircon and monazite in them, the concentrations of these minerals are not cited in his work. Later, an extensive study was carried out for recent placer deposits op the eastern Macedonia and Thraki beach sands, by the Greek Institute of Geology and Mineral Exploration (I.G.M.E.), (Markoulis et al, 1978; Me Donald, 1979). The aim of this study was to determine the concentrations of the minerals magnetite, ilmenite, rutile, zircon and garnet. Nevertheless, as noted by Me Donald (1979), the preliminary results were not encouraging for further research. Konispoliatis (1984) found up to 25% heavy minerals in the sand fraction of sediments collected around the Strymonikos gulf coast. Perissoratis et al (1985) reported that high heavy mineral concentrations occur in the 2 0 to 3 0 sand fraction of sediments collected from the following areas : (a) off the northern and southern coast of the lerissos gulf; (b) between Loutra Eleftheron and Kavala; (c) close to the mouth of Nestos river; (d) the eastern coast of Thassos island; (e) between Alexandroupolis and the mouth of Evros river; (f) the outer shelf to the west of Thassos island; and (g) the outer and the central parts of the Samothraki plateau. The minerals recognized were amphiboles, pyroxenes, epidote, garnet, zircon, routile, tourmaline, pyrite, Ti-magnetite and magnetite. □ 1 1 0
+ -t- B 2 11 poo Too < poo 3 1 2
...... r m 4 1 3
1 4 1 1 5 ; * *
1 5 e ! 6 ■
7 s 1 6
51 J t * S 8 55 5
e e : 9 40
Key:
1,2,3,4 : Post-tectonic and late-tectonic sediments. 5,6 : Axios zone and circum Rhodope zone. 7,8,9,10 : Metamorphic rocks. Rhodope and Serbo-Macedonian Massifs. 11,12,13,14 : Igneous rocks. For more details see Appendix I, p.332.
Fig.1.4 Simplified geological map of the mainland, and the Thassos and Samothraki islands (modified from Bornovas and Rondogianni-Tsiambaou, 1983).
A: Asprovalta; AB: Abdira; AC: Arapis cape; H: Aghios Haralambos; I: lerissos; K: Kirki; L: Limenaria; NK: Nea Karvali; P: Petrovouni; R: Nea Roda; S: Stavros; XY: Xylagani dissolution factor : a pure number, the inverse of sample weight (g) present in 1ml of final solution.
*
Na is among the elements detected, however as sea derived NaCI was not removed prior to analysis, the distribution of Na is not studied. 41
CHAPTER ll
ANALYTICAL TECHNIQUES
2.1 Bulk chemical analysis
2.1.1 Lithium metaborate technique
The offshore surface and buried sediments and the onshore coastal sediments were analyzed in bulk using Inductively Coupled Plasma Emission Spectrometry (I.C.P.A.E.S.) after fusion with lithium, metaborate. The lithium metaborate technique was introduced first by Suhr and Ingamells (1966) for the fusion of silicates to be analyzed by spectrographic methods. Its main advantage is its efficiency in breaking down almost all minerals so that they can be taken into solution. Acid attacks are frequently unable to effect a complete dissolution of resistant minerals, such as chromite, rutile, zircon,, magnetite and tourmaline. Fusion with lithium metaborate, L iB 0 2, overcomes this problem. However, disadvantages of the method are (i) a slight possibility of cross-contamination between samples, and (ii) a relatively high dissolution factor*(D.F. 1,000) which reduces sensitivity. In addition, in some cases, part of the hot bead produced from the fusion sticks to the graphite crucible causing loss of material. This problem can be overcome by (i) preheating the graphite crucible, and (ii) scratching the crucible with the aid of a spatula in order to loosen any part of the bead adhering.
2.1.1.1 Outline of the technique
100 mg ±. 0.1 mg of finely ground (-200 mesh) sample is weighed into a dry porcelain crucible and mixed thoroughly with 300 mg of LiB02 . The mixture is transferred to a graphite crucible and fused at 1,000° C in an electric muffle furnace for 20 min. The red hot molten bead produced is at once poured into a plastic tube containing 50 ml of 0.4N HNO3. The mixture is immediately shaken till the dissolution of the bead is complete. Then, 5 ml of the solution are diluted to 10 ml with deionized H20. The final solution is centrifuged for 3 min. at 1,000 rpm, in
* order to remove any graphite particles present, and then analyzed on an ARL 3,400 Inductively Coupled Plasma Emission Spectrometer, using the TROK calibration method (Thompson and Walsh, 1983, p.85). 42
This analytical procedure will be referred to as method A. The precision and accuracy were checked by including duplicate samples, blanks and reference materials in each set of samples.
2.2 Mineral acid attack
The samples that were to be subjected to partition analysis, were additionally analyzed using a mineral acid attack (GENHF)*, which gives a final solution matrix of 1M HCI. This was necessary as the solutions produced by the selective leaching also have a 1M HCI matrix. These acid solutions were run on the I.C.P. using a calibration designed for this matrix and called GEN4 (Thompson and Walsh,1983, p.85).
2.2.1 Outline of the technique
100 mg ± 0.1 mg of finely ground (-200 mesh) sample is weighed into a PTFE tube and then 2 ml of HNOg, 1 ml of HCI04 and 5 ml of HF are added. The mixture is heated in a hot block at 90°C for 3 hours and then at 140°C overnight. The temperature is raised to 160°C and the solution is evaporated until dryness. The resulting pellet is allowed to cool; it is leached, at 70°C for 1 hour, with 2 ml of 6M HCI and made up to 10 ml final volume with deionized water. The final solution is analyzed on an ARL 3,400 Inductively Coupled Argon Plasma Emission Spectrometer, using the GEN4 calibration method .
This analytical procedure will be referred to as method B. The precision and accuracy were checked by including duplicate samples, blanks and reference materials in each set of samples.
2.3 Comparison between method A and method B
Forty three samples were analyzed for K, Be, Mg, Ca, Sr, Ba, La, Ti, P, V, Cr, Mn, Fe, Co, Ni, Cu, Zn and Al by I.C.P. using both A and B methods. The two sets of data produced (Appendix III) were compared in order to assess their compatibility. Comparisons were drawn with the aid of a series of diagrams wherein the data obtained for each element through method A is plotted against the corresponding data obtained through method B (Fig. 2.1). The following relationships were observed:
1. The K, Ti, Mg, Ca, Ba, Zn, Ni, Fe, and Sr data obtained through method A correlate very well with the corresponding data of method B.
*HNO3, HCIO4 . HF 43
2. Method A yields consistently higher values of P, Be, V and Cr than those yielded by method B.
3. Method A Al results are in general higher than method B Al results at levels higher than 8% Al content.
4. Method A yields consistently lower values of Mn than those yielded by method B.
5. The Co and La values determined by method A are not compatible with those determined by method B. This could be related to the general occurrence of Co and La at levels close to their detection limits by method A (approximately 20 pg/g for both elements).
6. In a few samples the Cu values determined by method B are higher than the corresponding values yielded by method A. This could be attributed to Cu-adsorption on the surface of the graphite crucibles used for the fusion of the samples.
In conclusion, the two sets of data are overall quite compatible; it is suggested that the differences found are not likely to significantly affect the interpretation of the data. J
•0 *9 '*!
{
V ore n 8‘i n 9*o oo O'O
ro
n a
8*1
vz
CUE 45
0. * 0 . 360- 5^0 . 720- SOO. 1
{ l \
'00 £
'OZ
9P 47
0. 180. 360- 540. . 720. 900.
t r >
0. 16- 32. 48. 64. 80. 48
0. UOi 880. 1320. 1760 . 2 200.
f 49
Fig. 2.1 Scattergrams of the elements concentrations determined by the LiBO?-TROK method f * y (A) versus the corresponding ones determined by the GENHF-GEN4 method(B).
full line: 1:1 correspondance line dotted line: detection limit 50
2.4 Selective sequential extraction analytical procedure
2.4.1 Introduction
Sequential chemical extraction aims to examine the partitioning of the elements among the various geochemical phases of the sediments. It is a time consuming technique but it has the advantage of furnishing detailed information concerning the source, mode of occurrence, biological and physicochemical availability, mobilization and uptake of trace metals in sediments. Metals supplied by depositional, post-depositional, or bioturbation processes can be redistributed according to their adsorption affinities and kinetic responses to the phases present in the sediment (Balistrieri and Murray, 1986). Since sediments can be an important reservoir of trace contaminants such as * heavy metals in a secondary environment, sediment chemistry can help in the evaluation of water quality and potential pollutants. From the environmental point of view the pollution potential of a sediment depends on the availability and non-availability of trace contaminants to biota. In general terms, land derived pollutants and marine derived trace metals that are adsorbed on the surface of sediments are regarded as potentially available for release to the water column. By contrast, elements contained in the crystalline structure of minerals are considered to be unavailable for biological uptake under most conditions.
2.4.2 Methods
Many different procedures for the sediment partition analysis are available; each uses a different combination of chemical agents (Chester and Hughes, 1967; Nissenbaum, 1972; Gupta and Chen, 1975; Tessier et al, 1979; Valin and Morse, 1982). The sequential extraction analytical procedure of Tessier et al (1979) with a few modifications was followed here. It was selected because it deals with all the phases needed to be studied in the present work.
The sequential chemical extraction which was applied examines the partitioning of trace metals within the following five fractions: a. exchangeable; b. carbonate hosted; c. reducible;
i i
* 1 like the sea i 51 d. organic matter-sulphide hosted; and e. residual.
In the following, the main stages of the procedure are described. a. Exchangeable fraction
The exchangeable fraction represents the material adsorbed on the surface of the particles that make up the sediment. Materials contained in this fraction are considered to be the most easily released .
Ammonium acetate, NH4OAc, was used for the leaching of the elements present in the exchangeable fraction, while the Tessier et al (1979) procedure used magnesium chloride, MgCI2 , for this purpose. Problems such as Mg-interference with trace element determination by I.C.P. and the impossibility of the total removal of the Mg introduced as MgCI2 (prior to the analysis by I.C.P.) are among the drawbacks of the use of MgCI2. An alternative reagent is NH4OAc. However,various workers (Jackson,
1958; Chapman, 1965; Wagemann et al, 1977) suggested that NH4OAc not only takes into solution the exchangeable fraction but it also leaches metals associated with the carbonates. Prior to its use, comparative experimental work using both NH4OAc and
MgCI2 was done, in order to evaluate the efficiency of the NH4OAc leach. Two samples with a relatively high carbonate content were used for this purpose. The results obtained show that in samples with 26% and 22% Ca content, NH4OAc leaches 1.15% and 1.14% of the total Ca, while MgCI2 leaches 1.12% and 0.95% of the total Ca, respectively. Thus, the results suggest that NH4OAc does not attack the carbonates appreciably more than MgCI2. Therefore, leaching with NH4 OAc seems to be satisfactory for the present purpose. b. Carbonate hosted fraction
This fraction represents the proportion of elements associated with sedimentary carbonates, mainly calcium carbonate. 52
Experimental work was carried out in order to establish the easiest and most efficient way to attack the carbonate fraction of selected samples using acetic acid, HOAc, under various conditions. Thus,the following experiments were carried out: a. Dissolution using HOAc 25% overnight; b. Dissolution using HOAc 20% for 15min.; and c. Dissolution using HOAc 15% for 15 min., after heating the samples at 350°C for 2 hours. The pretreatment was carried out in order to prevent an attack on oxides present In the samples.
The results obtained (Table 2.1) show that by increasing the strength of HOAc from 20% to 25% and the leaching time from 15 min. to overnight, the amounts of Ca and Sr leached are not significantly increased. By contrast, the amounts of Fe, Ni, Zn, Co, Pb, and Cr leached are increased. When the samples were heated at 350°C and the strength of the acid used was decreased from 20% to 15% , but the leaching time was kept the same (15min.), it was found that the amounts of Ca, Sr, Mn, Co, Pb, Ni and Cu leached were reduced. However, the decrease in Mn, Co, Pb, Ni and Cu is not in proportion to the Ca and Sr values, it is thus suggested that: a. by increasing the strength of the acid and the leaching time, apart from the dissolution of the carbonate fraction , an attack on a different geochemical phase of the samples also occurs; b. by using HOAc 15% all the carbonate is not taken into solution; and c. significant amounts of Mn, Co, Pb, Ni and Cu are not likely to be associated with the carbonate fraction since the reduction in their proportions leached by HOAc 15% are much greater than and disproportionate to reduction in leaching of Ca and Sr. 53
Table 2.1 : Proportions of Ca, Mg, Sr, Pb, Cr, Mn, Fe, Co, Ni, Cu and Zn leached by : HOAc 25% overnight (A) ; HOAc 20% for 15 min. (B); HOAc 15% for 15 min. (after heating the samples at 350° C for 2 hours) (C).
.Sample S3 Mg Si Eh St Mn l.P. (%) (%) (%) (%) (%) (%)
I 5 A 80 53 87 48 8. 23 B 88 57 91 43 3 25 C 66 38 74 34 15 18 • I 22 A 54 16 63 31 1 42 B 59 11 75 31 0 53 C 39 4 64 42 0 25
STR4 A 78 24 81 13 2 33 B • 87 25 84 16 1 22 C 37 9 38 16 1 10
KB 55 A 81 43 88 28 3 46 B 88 45 90 22 3 46 C 35 15 46 22 3 20
KB 113 A 86 79 94 68 20 81 B 86 84 98 47 7 86 C 31 24 41 24 19 40
THR 4 A 85 62 91 53 23 76 B 92 66 92 51 16 78 C 61 36 75 40 20 47
BSTR4 A 67 19 10 14 7 59 B 66 15 10 7 2 42 C 62 6 7 0 4 21
B 12 A 71 32 69 15 2 42 B 77 33 69 15 2 44 C 32 11 29 0 2 18 54
Table 2.1 (continued)
Sample Ee Qq N i Oil Zn
I.D. (%)(%)(%)(%)(%)
I 5 A 6 16 6 42 54 B 4 16 6 41 35 C 6 3 0 20 40
I 22 A 2 17 7 8 10 B 3 17 7 26 1 C 4 9 4 22 14
STR 4 A 2 4 4 9 7 B 1 3 4 13 1 C 2 3 2 9 2
KB 55 A 5 17 3 13 10 B 2 14 5 18 2 C 3 7 2 9 18
KB 113 A 13 24 52 46 35 B 10 16 11 48 16 C 8 10 2 26 20
THR 4 A 8 26 16 32 27 B 6 26 12 42 8 C 6 10 0 19 14
BSTR4 A 10 20 23 17 23 B 6 8 9 40 16 C 3 3 0 15 5
B 12 A 2 7 3 9 13 B 2 7 3 14 2 C 1 4 0 9 19
In the present work, the carbonate phase was selectively dissolved by using 1M sodium acetate, NaOAc, and adjusting the pH to 5 with acetic acid. The leaching time was 5 hours; according to the Tessier et al procedure (1979) this is a sufficient time for finely divided samples without a very high carbonate content. 55 c. Reducible fraction
The reducible fraction consists of these components that will go into solution in a reducing environment; mainly ferromanganese oxides, some amorphous iron oxides and other reducible material (Chester and Hughes, 1967; Cronan, 1976; Moorby and Cronan, 1981). It is well established (Jenne, 1968) that Fe and Mn oxides occur as nodules, concretions, cement between particles, and simple coatings on particles. These oxides are excellent scavengers for trace metals. However, under reducing conditions, Fe and Mn oxides are soluble and in dissolving release the trace metals bound to them. Hydroxylamine hydrochloride, NH2OH.HCI, and acetic acid, HOAc, were
used for the dissolution of the reducible fraction. d. Organic and sulphide associated elements fraction
Hot hydrogen peroxide, H20 2 , in a nitric acid, HN03 , medium was used for the
oxidation of the organic matter present in sediments. This leaching step is conducted under extreme conditions of prolonged heating, as a result both organic matter and sulphides are oxidized. Because of this, it is impossible to allocate the trace metals found in the leachate to either the organic matter or the sulphide minerals. The use of oxidizing solutions stronger than H20 2 , for a complete oxidation of the organic
matter, was avoided because they rely upon the use of strong acids which would also attack the silicate component of the sediments. Experimental results (Table 2.2.) showed that more than 99% of the organic carbon present in sediment occurs within the humics, the latter are oxidized by H20 2. According to Jackson (1958) the organic matter which is not leached by H20 2 most probably consists of paraffin-like
material and resistant structural (non-humified) organic matter residues.
Table 2.2 : Organic carbon and humic acids (H.A.) content of sample STR 8; organic carbon content of the H.A. ; and organic carbon content of the sample, present as H.A.
Sample org C content H.A. content ora. C content orq, 0 cant, Qf sample I.P. of sample of sample of H.A. present as_H A (%) (%) (%) (%)
STR 8 0.15 0.35 42.91 0.15 56 e. Residual fraction
The residual fraction contains these proportions of elements that are hosted in the lattice structure of aluminosilicate minerals (Tessier et al, 1979). Both primary and secondary minerals can hold trace metals within their crystal structure. The residual solid that remains after the four preceding extractions (a, b, c, and d) essentially consists of detrital silicate minerals, clays, resistant sulphides and a small quantity of refractory organic material. The amounts of elements associated with the residual phase are calculated by substracting the amounts associated with the first four fractions from the total sediment analysis obtained after complete dissolution of the sample by a mixture of HF, HCIO4 and HNO3.
The sequential steps of the selective extraction analytical procedure used in this work are the following: a. Exchangeable fraction
1g of sample is extracted with 8 ml 1M NH4OAc, at pH 8.2, for 1 hour, at room temperature, with continuous agitation.
b. Carbonate hosted fraction
The residue from (a) is leached with 8 ml 1M NaOAc, adjusted to pH 5 with HOAc, for 5 hours, at room temperature, with continuous agitation. c. Reducible fraction
The residue from step (b) is extracted with 20 ml 0.04M NH2OH.HCI in 25% (v/v)
HOAc at pH 2. The extraction is performed at 96+3°C, for 6 hours, with occasional agitation. d. Organic matter-sulphide hosted fraction
To the residue from step (c), 3 ml of 0.02M HNO3 and 5 ml of 30% H20 2, adjusted to pH 2 with HNO3, are added. The mixture is heated to 85±2°C for 2 hours with 57 occasional agitation. Then, a second 3 ml aliquot of 30% H20 2 adjusted to pH 2 with
H N 03 , is added. The mixture is heated at 85±2°C, for 3 hours, with intermittent agitation. After cooling, 10 ml of 3.2M NH4OAC in 20% (v/v) HNO3 are added, followed by dilution to a final volume of 40 ml and continuous agitation for 30 min.
The selective extractions were conducted in I.C.P. disposable plastic tubes. After each successive extraction, separation of the supernatant from the residual solid was done by centrifuging the mixture for 30 min., at 2,000 rev/min. The supernatant liquids are then decanted. The supernatants are subjected to a further treatment involving : a. addition of a few drops of concentrated HN03;
b. heating to dryness; and c. dilution with 1M HCI; prior to their analysis by I.C.P. This treatment yielded a final solution matrix of 1M HCI, which was necessary for the I.C.P. analysis.
2.5 Calcium carbonate and organic carbon determinations
The calcium carbonate and the organic carbon contents of 206 surface marine samples were determined by gas chromatography (the results are presented in Appendix IV).
2.5.1 Calcium carbonate determination : outline of the technique
100 mg of finely ground sample are placed in a small bottle. A small suitable tube containing 2 ml H3P 0 4 25% is put in the bottle which is then sealed. The sample is ready for the determination of its CaC03 content, by gas chromatography.
2.5.2 Organic carbon determination: outline of the technique
Once the CaC03 content is determined, a solution of H2S04 and KMn04 is added in the bottle , which is then sealed and heated for 1 hour at 125°C. The sample is ready for the determination of its organic carbon content, by gas chromatography. 58
2.6 X-Rav Diffraction
X-Ray Diffraction (XRD) has been performed on total sample. Selected samples were cavity mounted and a scanning speed of 2° 20 per minute was used, in the range 4-70°. 59
CHAPTER 111
THE SAMOTHRAKI PLATEAU
3.1 Surface sediments
A total of 177 surface sediment: samples were collected from the 5,800 km^ area of the Samothraki plateau. The sample sites (Fig. 3.1) are reasonably evenly distributed throughout the area, however, some clustering occurs south of Alexandroupolis and around the coast of Thassos. In order to shed some more lightpn the geochemistry of the offshore area, 39 samples from the coast of the adjacent mainland and 6 beach samples from the coast of Thassos island were also collected and analyzed (Fig. 3.2).
3.1.1 Sediment composition
3.1.1.1 Distribution of rock fragments, quartz, heavy minerals, biogenic matter and carbonates
The maps (Fig. 3.3 - 3.7.) that present the distribution of rock fragments, quartz, heavy minerals, biogenic matter and carbonates in the surface sediments of the Samothraki plateau are provided by Perissoratis et al (1987).
3.1.1.1.1 Rock fragments
The distribution of rock fragments in the coarse fraction (Fig. 3.3) shows that high concentrations occur in various nearshore areas, in the mid-shelf region, and in a rather narrow belt along the north coast of Samothraki island. This last area of enrichment, having up to 70% rock fragments content, is attributed to the supply of rock fragments from the steep and scree-covered northern coastline of Samothraki island. The absence of any substantial river draining°\his part of the island is present in responsible for the 'small amounts of material, product of chemical weathering;; me adjacent sea fioor. Apart from the previously mentioned areas of enrichment, the 60 plateau has a low to very low rock fragment content.
3.1.1.1.2 Quartz
The distribution of quartz in the coarse fraction (Fig. 3.4) indicates that the quartz content increases from the nearshore areas towards the mid-shelf, where it locally exceeds 40%, and diminishes again on the outer shelf. The main areas of quartz enrichment are located as follows:
a. off Nestos river mouth;
b. in the northeast part of Samothraki plateau; c. off Alexandroupolis district; and d. northwest of Samothraki island, at about 13km distance.
3.1.1.1.3 Heavy minerals
The distribution of heavy minerals in the coarse fraction (Fig. 3.5 a and b) shows that various areas of enrichment occur. They are the following:
a. off Nestos river mouth: amphiboles, pyroxenes and garnet;
b. off Evros river mouth and westwards of the mouth itself as far as Alexandroupolis: amphiboles, magnetite and ilmenite;
c. off Filiouris river mouth, in Xylagani bay: amphiboles;
d. in the mid-shelf terrace: amphiboles and pyroxenes; and
e. in coastal locations of a rather limited extent, lying off the east coast of Thassos island: epidote and amphiboles.
The high amphibole and pyroxene contents that occur close to the mouths of the main rivers are most likely derived from the metamorphic formations of the Rhodope Massif. Alternatively, part of these concentrations could derive from eroded Neogene and Quaternary sediments which are enriched in amphiboles and pyroxenes, having 61
received them from the aforementioned metamorphic formations (second cycle concentrations),(Perissoratis et al, 1985). The weathering products of the metamorphic formations and of the Neogene and Quaternary sediments, that occur in the mainland, are brought to the sea by the rivers. The high magnetite and ilmenite contents that occur off the Evros river mouth, are likely to be associated with the erosion of the metamorphic and igneous rocks of eastern Thraki. Magnetite occurrence in the area off Alexandroupolis district was also reported by Marinos et al (1976).
Epidote enrichment in the nearshore areas of Thassos island is most probably derived from the metamorphic rocks (schists and gneisses) that outcrop near the eastern coast of the island, which are enriched in epidote.
The heavy minerals enrichment that occurs in the mid-shelf terrace, confined to about 60 to 80 m water depth, is associated with a submarine canyon. The presence of this canyon implies that a river was flowing there during a lower sea level stand. As the canyon lies to the east of the heavy minerals enrichment, an ancient westwards transport of the river runoff along the coastline, is suggested. This would have been similar to the present day transport, along the present day coastline to the north.
3.1.1.1.4 Biogenic carbonate particles
The distribution of biogenic constituents (bioclast and foraminiferal tests) in the coarse fraction (Fig. 3.6) shows the following areas of enrichment:
a. an extensive belt stretching off the southeast coast of Thassos island to the north of Samothraki island;
b. off the northeast coast of Thassos island;
c. seawards of the Porto Lago embayment; and
d. eastwards of the Xylagani bay.
3.1.1.1.5 Carbonates
The distribution pattern of carbonate particles (Fig. 3.7) shows that areas of 62 enrichment occur in the eastern part of Samothraki plateau, off the northeast coast of Thassos island, off the Filiouris river mouth and in its vicinity, and in a narrow coastal belt westwards of Alexandroupolis. A relative carbonate enrichment occurs seawards of the shelf break. In most cases, the distribution of total carbonates corresponds relatively well to that of biogenic constituents in the coarse fraction. Hence, it is likely that the largest part of the total carbonate content in the surficial sediments is related to the biogenic material of the sediments.
However, the carbonate enrichment which is located seawards of the shelf break does not correspond to a high content of biogenic constituents in the coarse fraction. This lack of agreement can be related to the fine grained sediment (clayey silt) that is present in the aforementioned area. Although a high content of biogenic constituents in the coarse fraction of a relatively coarse sediment can reasonably be related to a high carbonate content in total sediment, a corresponding relation cannot be applied when fine grained sediment is concerned. Comparison between the distributions of biogenic carbonate and total carbonates would have been more fruitfull if their variations were based on total sample analysis; unfortunately, such information was not available. The fine grained carbonate could still be biogenic in origin.
3 .1 .1 .2 Distribution of organic carbon and results from XRP and SEM studies
In order to obtain more information on the sediments, the organic carbon content was determined, and also selected samples were subjected to XRD and SEM studies.
3.1.1.2.1 Organic carbon
The distribution of organic carbon (Fig. 3.8) in the Samothraki plateau sediments shows that most of them contain less than 1.3% organic carbon. However, areas of relative enrichment in organic carbon occur in the eastern and northern part of the plateau. The highest organic carbon contents (3.5-7%) were determined in the following areas: a. a coastal belt stretching westwards of Alexandroupolis; b. a location lying seawards of the Porto Lago bay; and c. off the northeast coast of Thassos island. 63
The areas lying close to the mouths of the main rivers are relatively poor in organic carbon.
3.1.1.2.2___XRD Results
Eleven samples, collected from the Samothraki plateau area, were subjected to X-Ray Diffraction studies. The minerals recognized are listed in Table 3.1.
Table 3.1: Mineralogical composition of selected samples from Samothraki plateau, as determined through XRD analysis.
Sample I.D. Major minerals Minor minerals Traces
THR 2 quartz plagioclase illite kaoiinite/chlorite K-feldspar
THR 4 quartz plagioclase calcite high Mg-calcite K-feldspar aragonite
THR 9 quartz plagioclase illite K-feldspar kaoiinite/chlorite high Mg-calcite aragonite calcite
THR 24 quartz calcite illite plagioclase K-feldspar high Mg-calcite aragonite
THR 32 quartz K-feldspar calcite high Mg-calcite aragonite plagioclase
THR 41 quartz plagioclase K-feldspar high Mg-calcite calcite aragonite 64
Table 3.1 (continued)
THR 64 quartz K-feldspar plagioclase aragonite high Mg-calcite calcite
THR 72 K-feldspar plagioclase high Mg-calcite hornblende quartz calcite aragonite
THR 75 K-feldspar high Mg-calcite quartz calcite aragonite piagioclase
THR 82 K-feldspar aragonite high Mg-calcite plagioclase quartz ill ite calcite
THR 98 quartz calcite hornblende plagioclase high Mg-calcite aragonite K-feldspar illite kaolinite/chlorite
THR 129 quartz it I ite talc plagioclase K-feldspar kaolinite/chlorite hornblende
THR 150 quartz illite plagioclase K-feldspar calcite high Mg-calcite aragonite
3.1.1.2.3 SEM Results
Selected samples from the Samothraki plateau (the samples THR 3, THR 56, THR 60 and THR 128) were examined by Scanning Electron Microscopy. Feldspar (represented by K, Al and Si) and mica (represented by K, Al, Si, Fe and Mg) were identified. samples
Fig. 3.1 Samothraki plateau : map representing the offshore surface and subsurface sample locations. O) oi Fig. 3.2 Samothraki plateau : map representing the onshore sample locations. 6 6 67
Fig. 3.3 Samothraki plateau surface sediments : distribution of rocK fragments in the coarse fraction.
Fig. 3.4 Samothraki plateau surface sediments : distribution of quartz in the coarse fraction. Fig. 3.5a Samothraki plateau surface sediments : distribution of heavy mineral content in the coarse fraction.
O) oo 69
0 to «■ —■ Kff)
amphibole pyroxene garnet
^ 5 -1 0 B 2 - 4 □ <2 0
epldote metal li c 0 < 4 A - present A abundant
Fig. 3.5 b Samothraki plateau surface sediments : distribution of selected heavy minerals in the 2 0 to 3 0 fraction.
I • r! •
l Fig. 3.6 Samothraki plateau surface sediments: distribution of biogenic components in the coarse fraction.
O 71
% m 'i Km Fig. 3.7 Samothraki plateau surface sediments: distribution of total carbonates CZJ Fig. 3.8 Samothraki plateau surface sediments : distribution of organic carbon. 72 3.1.2__ Bulk geochemistry of offshore sediments and of those on the adjacent coasl Elements are dealt with in order of decreasing abundance (Table 8.1), as this is established on the basis of robust means of Box-Cox transformed geochemical data (for a definition on robust statistics and Box-Cox transformation see Appendix V). 3.1.2.1.1 Silicon The regional variation of the Si content of the sediments (Fig.3.9) is similar to the distribution of quartz in the coarse fraction (Fig. 3.4.). High Si concentrations (up to 25%, with maximum values of over 29%) occur in a broad belt stretching from Thassos island to Alexandroupolis. Seawards and landwards of this belt the Si content decreases and tends to increase again slightly very close to the coast. Values as low as 10% occur in sediments lying northeast of Samothraki, and in the sediments collected from the channel that separates Thassos island from the mainland. The distribution of Si in the onshore samples shows that high concentrations occur in those collected along the coast, and in the sediments collected from the rivers Nestos and Filiouris. 3.1.2.1.2 Calcium The regional variation of Ca (Fig. 3.10) is almost exactly the reverse of that of Si. Calcium distribution matches quite closely the distribution of total carbonates, indicating therefore that in Ca rich samples, the Ca is mainly present as CaCOg. High Ca values occur in sediments lying: a. northeast of Samothraki island; b. off the northeast coast of Thassos; c. seawards of the Filiouris river mouth; and d. seawards of the Porto Lago embayment. By contrast, low Ca values occur in the Si rich belt stretching from Thassos island to Alexandroupolis, off the Nestos river mouth and to the east of it. 73 The distribution of Ca in the onshore samples shows that most of them are poor in Ca, except for the beach samples collected from the eastern coast of the mainland. The Nestos and Filiouris river sediments are Ca poor. 3.1.2.1,3_ Aluminium Aluminium exhibits a regional variation (Fig. 3.11) rather different to those of Si and Ca. The main areas of Al enriched sediments are associated with the Evros and Nestos rivers. The most extensive area of Al enrichment, with maximum values of just over 9%, occurs in an east-west belt parallel to the coastline and about 5km offshore. It has the shape of a tongue and stretches from southeast of Alexandroupolis, towards the west, almost as far as Xylagani. An enrichment in Al also occurs off the mouth of the Nestos river. The highest values (in excess of 8.5%) in this area do not occur immediately off the river mouth, but instead in two separate lobes , lying one to the west and one to the northeast of the river mouth itself. Regions enriched in Al were also found off the mouth of the Filiouris river, and in the fine grained deep water sediments to the east-southeast of Thassos island. Like Si, Al content is low in the carbonate rich sediments lying to the north of Samothraki island. The distribution of Al in the onshore samples (Fig. 3.11) shows that an area of enrichment occurs between the rivers Nestos and Filiouris. By contrast, the coastal samples collected eastwards of the river Filiouris as far as Alexandroupolis are Al poor. Filiouris river sediments are Al rich, with maximum values over 8.5%. Nestos river sediments are generally Al poor, although the samples collected from its eastern distributary are enriched in Al. Finally, the beach samples collected from the northeast, east and southeast coast of Thassos island are poor in Al. 3.1.2.1.4 Iron The regional variation of Fe (Fig. 3.12) is quite similar to that of Al. It shows that Fe enrichments occur as follows: a. in the rock fragment rich sediments along the northern coastline of Samothraki island, with Fe values as high as 9.5%; b. in the Al enriched sediments collected from the belt that lies in the northeastern 74 part of the Samothraki plateau, with Fe content around 5%; c. in two separate lobes off the Nestos river mouth, a distribution similar to that displayed by Al; d. off the Filiouris river; and e. in the deep water sediments southeast of Thassos island. Sediments collected from Filiouris river are Fe rich (about 4.7% Fe content), while the samples from Nestos river are Fe poor. The sediments collected from the eastern distributary of the latter are relatively enriched in Fe. The beach samples are mainly poor in Fe. 3.1.2.1.5 Potassium The distribution of K (Fig. 3.13), like Al and Fe, shows areas of enrichment closely associated with the Nestos and Evros rivers, and less extensively with the Filiouris river. The area of high Fe enrichment lying off the northwest coast of Samothraki island seems to be poor in K. By contrast, K rich sediments were collected further east along the coast of Samothraki, and from a separate area further offshore to the east. The latter area of enrichment is associated with Fe rich sediments (Fig. 3.12). However, the sediments of the offshore area as a whole, do not show great variability in K content, and the K concentration of most of the samples fall within the range 1.4-2.2% . The coastal samples collected from the beach lying eastwards of the Filiouris river as far as Alexandroupolis are poor in K. In contrast, samples collected from (a) the coastal area lying between the rivers Nestos and Filiouris and (b) the area located to the west of the Nestos river mouth are enriched in K, being characterized by K contents higher than 3%. The beach samples that were collected from the northeast coast of Thassos island are also rich in K. Finally, the river sediments collected from Filiouris, Nestos and the eastern distributary of the latter are enriched in K. 75 3.1.2.1.6 Magnesium The pattern of the variation of the Mg concentration in the sediments (Fig. 3.14) does not match closely any of the other major element distribution patterns. The deep water sediments collected southeastwards of Thassos island are enriched in Mg; this enrichment is associated with high contents in Fe and Al. Also, sediments collected off the Nestos river are rich in Mg. The pattern of Mg enrichment off the Evros river is not as marked or as extensive as those corresponding to AI.Fe and K. Two more areas with sediments rich in Mg were detected; they occur (a) off the northwest coast of Samothraki island, and (b) off the river Filiouris. The Si rich mid-shelf belt is very poor in Mg. The distribution of Mg in the onshore samples (Fig. 3.14) demonstrates that: (a) the beach samples are poor in Mg; (b) the Nestos river sediments are poor in Mg; (c) the sediments collected from the eastern distributary of Nestos river are relatively Mg enriched; and (d) the Filiouris river sediments are rich in Mg, containing up to 2.1% Mg. 3.1.2.1.7 Titanium The distribution pattern of Ti (Fig. 3.15) is very similar to that of Fe, displaying the same areas of enrichment off the rivers Nestos, Filiouris and Evros. The highest Ti value (over 9,000 pg/g) occurs in one sample off the northwest coast of Samothraki island. Elsewhere, Ti values do not exceed 5,000 pg/g. Although Ti and Fe have similar distributions, the Ti rich (0.43%) sample from the mid-shelf area south of Xylagani, is nor enriched in Fe. The beach samples are poor in Ti. Nestos river sediments are quite poor in Ti; however, those collected from its eastern distributary are Ti enriched. Filiouris river sediments are rich in Ti. 3.1.2.1.8 Strontium The regional variation of Sr (Fig. 3. 16) follows closely that of Ca(Fig. 3.10). Thus, Sr is enriched in sediments collected from an extensive area northeast of Samothraki island; values higher than 1,000 pg/g were determined there. A high Sr content was found in sediments lying north and northeast of Thassos island, as well as in sediments collected seawards of the coastal area lying between Porto Lago and Aghios Haralambos. 76 The onshore samples are mainly poor in Sr, but several collected from the coastline between Filiouris river mouth and Alexandroupolis are relatively enriched. Finally, both Nestos and Filiouris river sediments are poor in Sr. 3.1.2.1.9 Phosphorus The regional variation of P (Fig. 3.17) shows that a high P content occurs in the Fe and Al rich belt that stretches westwards of the Evros river mouth. High P concentrations were determined off the northwest and northeast coast of Samothraki island. The sediments collected off the rivers Nestos and Filiouris are rich in P. The highest P content, up to 1,000 pg/g, was found in nearshore samples collected northeastwards of the river Nestos. Among the river sediments, those collected from the river Filiouris and the eastern distributary of the Nestos are P rich; in contrast, samples collected from the Nestos river itself are poor in P. A low P content occurs in the beach samples collected along the coastline. 3.1.2.1.10 Barium The regional variation of Ba (Fig. 3.18) shows that high Ba concentrations were found in : (a) the Fe and Al enriched sediments west of the Evros river mouth, however values above the average do not persist as far west as in the distribution patterns of Fe and Al; (b) the sediments lying to the west of Samothraki island; (c) the sediments collected along the northeast coastline of Samothraki island, with values up to 1,000 pg/g or more; (d) the area off the northeast coast of Thassos island, associated with a K and Si enrichment; and (e) the vicinity of Nestos river, in an east-west belt parallel to the coast lying east of the river mouth. The distribution of Ba in the onshore samples (Fig. 3 18) shows an area of enrichment lying between the rivers Nestos and Filiouris, with values exceeding 700 pg/g. By contrast, the beach samples collected from the area eastwards of Filiouris river as far as Alexandroupolis are Ba poor. Finally, the onshore samples from the northeast coast of Thassos island are Ba rich whereas those from the east and southeast coast of the island are poor in Ba. 77 3.1.2.1.11 Manganese The distribution pattern of Mn (Fig. 3.19) shows that high Mn concentrations occur: (a) in the Fe and Ai rich sediments off the river Evros, where they reach levels as high as 1,320 pg/g; like Mg, this area of Mn enrichment is not as extensive as those of Fe and AI; (b) in the Fe and Mg rich sediments lying off the northwest coast of Samothraki island, where they reach levels as high as 1,120 p.g/g; (c) in the deep water sediments confined east and southeast of Thassos island; (d) off the Filiouris river mouth; (e) off the Nestos river mouth; and (f) in several samples collected southeast of the Porto Lago embayment. The behaviour of Mn in the onshore samples shows that the river sediments collected from Filiouris, Nestos and the eastern dstributary of the latter are Mn rich. A Mn content as high as 1,180 jig/g was determined in the Filiouris river sediments. The beach samples collected from the coastal area lying east of the Filiouris river mouth are Mn enriched. 3.1.2.1.12 Zirconium The regional variation of Zr (Fig. 3.20) is quite similar to that of Ti. However, Zr is not enriched in sediments collected off the coast of Samothraki island. Furthermore, the pattern of the Zr enrichment near the river Nestos is rather different to that seen for Ti and several major elements, like AI and Fe. Thus, in the area off the Nestos river, The Zr content is generally higher in the sediments closest to the coast, to the west and to the east of the river mouth, and it gradually decreases with increasing distance offshore. The difference between this pattern and the corresponding regional distributions of Fe, Ti and AI is that the highest enrichment in the Fe, Ti and AI contents was found in two separate small areas centred a little distance off the coast itself. Sediments relatively rich in Zr were also collected from an east-west belt, parallel to the coast of Alexandroupolis district. This Zr enriched belt corresponds closely, but not exactly, to an area of increased heavy mineral abundance. Although the mean Zr content of the sediments is only about 100 jig/g, Zr concentrations exceed 400 jig/g in two samples; one of them was collected from the coastline to the northeast of the Nestos river mouth, and the other from the Zr rich area off Alexandroupolis. 78 The samples collected from both the river Filiouris and the beach lying to the east of this river mouth (Fig. 3.20) are enriched in Zr - their Zr content exceeds 150 pg/g. The river samples collected from (a) the eastern distributary of Nestos river and (b) the river that flows out west of the Porto Lago embayment are also enriched in Zr. Finally, the beach samples collected from the northeast, east and southeast coasts of Thassos island (Fig. 3.20) are Zr poor. 3.1.2.13 Chromium The distribution pattern of Cr (Fig. 3.21) is similar to that of Ti (Fig. 3.15). Chromium enrichments occur in the following areas: (a) off the northwest coast of Samothraki island, in association with Fe and Mg enrichments (like Ti, the highest values of Cr were determined in samples collected from this area); (b) off the river Nestos, in two separate lobes; (c) off the Filiouris river mouth; and (d) to the west of the Evros river mouth. The Ti rich mid-shelf sample has an above average Cr content. The samples collected from (a) the river Filiouris (with 200 pg/g Cr concentrations); (b) the eastern distributary of the river Nestos; and (c) the river that flows out in the west coast of the Porto Lago embayment are enriched in Cr, as well as in Ti and Zr. 3.1.2.14__ Vanadium The pattern of the V distribution (Fig. 3.22) has some similarities with those of Fe and Al. High V content was determined in the sediments lying (a) off the Filiouris river mouth; (b) seawards as well as northeastwards of the Nestos river mouth; (c) off the southeast coast of Thassos island; (d) off the northwest coast of Samothraki island; and (e) in a rather extensive area occurring westwards of the Evros river mouth. The onshore samples are poor in V, apart from those collected from (a) the river Filiouris, with higher than 130 pg/g V content; and (b) the beach close to Aghios Haralambos. 3.1.2.15__ ZiBSL The regional variation of Zn (Fig. 3.23) shows that sediments with high 79 concentrations in Zn were collected from the following areas: (a) off and to the west of the Evros river mouth, in association with Fe and Al enrichments; (b) off the Nestos river mouth; (c) off the southeast coast of Thassos island; (d) northeast of Samothraki island; and (e) seawards of the Aghios Haralambos district. In contrast, the sediments collected off the Filiouris river mouth were found to be depleted in Zn. The samples collected along the beaches of the adjacent land are Zn poor, apart from those collected from Makriammos (an area lying on the northeast coast of Thassos island). The Nestos river sediments are poor in Zn while the ones collected from its eastern distributary are enriched in Zn, with Zn content exceeding 130 pg/g. 3.1.2.16__ U nthanum The distribution pattern of La (Fig. 3.24) has some similarities with those of Fe and Al. High La values were found in the sediments collected from the following areas: (a) off and to the west of the Evros river mouth; (b) eastwards of the Nestos river; (c) northwards of Samothraki island; (d) off the northeast coast of Thassos island; and (e) off the east and southeast coast of Thassos island. The onshore samples are rather poor in La (Fig. 3.24), apart from these collected from (a) the eastern distributary of the river Nestos, and (b) the beach close to Aghios Haralambos , which are relatively enriched in La. 3.1.2.17 __NIcK&i The pattern of Ni distribution in the offshore sediments (Fig. 3.25) demonstrates the following enriched areas: (a) off the Filiouris river mouth; (b) off the Nestos river mouth; (c) westwards of the Evros river mouth; (d) off the northwest coast of Samothraki island, in association with Fe and Mg enrichments; (e) eastwards and southeastwards of Thassos island, in association with Al enrichment; and (f) off the Xylagani district. Among the river sediments, those collected from the river Filiouris are enriched in Ni, with concentrations exceeding 100 p.g/g; in contrast, the Nestos river sediments are poor in Ni. Finally, the sediments collected from both the river flowing out on the west coast of the Porto Lago bay and the corresponding beach are enriched in Ni. 80 3.1.2.18 Copper The pattern of the Cu distribution (Fig. 3.26) has several similarities with that of Ni. High Cu values were determined off the mouths of the rivers Nestos and Filiouris, off and westwards of the Evros river mouth, off the north coast of Samothraki island, off Porto Lago bay, and finally, in the deep water sediments lying southeast of Thassos island. The variation of Cu in the onshore samples (Fig. 3.26) shows that the beach samples are rather poor in Cu. The river sediments collected from both the Filiouris river and the eastern distributary of Nestos river are relatively enriched in Cu. 3.1.2.19 Cobalt The regional variation of Co (Fig. 3.27) shows that the Samothraki plateau sediments are, in general, poor in Co. Samples relatively enriched were collected from the northeastern part of the plateau. The sediments of the remaining offshore area contain less than 30 p.g/g Co, apart from few isolated samples. The onshore samples are Co deficient, except for the Filiouris river sediments which are relatively Co enriched (Fig. 3.27). 3.1.2.20 __ Ber.yLl.Lum The regional variation of Be (Fig. 3.28) has some similarities with that of Fe. Thus, high Be concentrations were determined in the sediments collected from (a) the large Fe and Al rich belt that stretches westwards of the Evros river mouth; (b) the two Fe and Al rich lobes confined one to the west and one to the northeast of the Nestos river mouth; and (c) the area lying off the northeast part of Samothraki island. In addition, samples collected off the Filiouris river mouth and also from the deep water to the southeast of Thassos island are relatively enriched in Be. The onshore samples are rather poor in Be (Fig. 3.28), apart from those collected from (a) the west coast of the Porto Lago bay, and (b) the eastern distributary of the Nestos river. 81 Si Fig. 3.9 Samothraki plateau and adjacent coast: regional variation in the Si content of the surface samples. ///// 12.0-17.5 % 7.5-120 omna 2.5-75 t z z i < 2.5 Fig. 3.10 Samothraki plateau and adjacent coast: regional variation in the Ca content of the surface samples. t ■ .* ^ y 82 Fig. 3.11 Samothraki plateau and adjacent coast: regional variation in the Al content of the surface samples. 15 Km Fig. 3.12 Samothraki plateau and adjacent coast: regional variation in the Fe content of the surface samples. 8 3 d m 3 ijo~~ia < 1.0 Fig. 3.13 Samothraki plateau and adjacent coast: regional variation in the K content of the surface samples. ///// 0,8 cm Fig. 3.14 Samothraki plateau and adjacent coast: regional variation in the Mg content of the surface samples. 84 Fig. 3.15 Samothraki plateau and adjacent coast: regional variation in the Ti content of the surface samples. Ena 250-450 w i/g CUD <250 : Fig. 3.16 Samothraki plateau and adjacent coast: regional variation in the Sr content of the surface samples. 85 400-500 E n a 300-400 ng/g CZ3 < 300 Fig. 3.17 Samothraki plateau and adjacent coast: regional variation in the P content of the surface samples. 450 - 550 S 350-250 E lia 250 - 350 » <250 Fig. 3.18 Samothraki plateau and adjacent coast: regional variation in the Ba content of the surface samples. t 86 o io >600 ///// 400-600 300-400 ng/a in n s 200 - 300 <200 Fig. 3.19 Samothraki plateau and adjacent coast: regional variation in the Mn content of the surface samples. » v Fig. 3.20 Samothraki plateau and adjacent coast: regional variation in the Zr content of the surface samples. l 87 c m <5o Fig. 3.21 Samothraki plateau and adjacent coast: regional variation in the Cr content of the surface samples. cnna 40-70 ng'g t= k 4 0 Fig. 3.2,2 Samothraki plateau and adjacent coast: regional variation in the V content of the - j surface samples. ►! t 88 Fig. 3.23 Samothraki plateau and adjacent coast: regional variation in the Zn content of the surface samples. Fig. 3.24 Samothraki plateau and adjacent coast: regional variation in the La content of the surface samples. F 89 Fig. 3.25 Samothraki plateau and adjacent coast: regional variation in the Ni content of the surface samples. Fig. 3.26 Samothraki plateau and adjacent coast: regional variation in the Cu content of the surface samples. r 90 MO/g Fig. 3.27 Samothraki plateau and adjacent coast: regional variation in the Co content of the surface samples. E=ZI < 15 Fig. 3.28 Samothraki plateau and adjacent coast: regional variation in the Be content of the surface samples. 91 3.1.3 Multivariate Statistical Analysis 3.1.3.1 Introduction Multivariate statistical treatment was applied to the geochemical data obtained from the offshore surface samples. Multivariate analysis of the multi-element data provides useful information on the identification of broad groups of element associations and the estimation of the influence of one element on another. Furthermore, the analysis of multi-element data helps to focus attention on element associations related to particular geochemical or physiographical units (Howarth and Sinding-Larsen, 1983); it also helps answer such questions as which samples are most similar on the basis of their chemical composition. The calculation of the correlation matrix (C.M.) is the starting point for all the multivariate statistical techniques used in this study. To produce a C.M. from regional geochemical data, which represents the most reliable expression of background relationships, one must correct for outliers and transform the data (Coward, 1986). Calculation of robust coefficients has been the method used for the outlier correction (see Appendix V). On the other hand, the application of transformations allows for the well known statistics of normal distributions to be applied to data; it also reduces the effect of outliers. The Box-Cox (power) transformation (Appendix V) was performed for the multivariate distributional normality of the data. In the following, the robust C.M. of the raw, untransformed data values and the robust C.M. of the Box-Cox transformed data values will be compared. The correlation matrices are an expression of relationships between variables; these elemental associations will be presented through the dendograms provided by cluster analysis (Appendix V). However, as cluster analysis generates elemental groupings on the basis of highest correlations, often ignoring in effect intermediate but still significant correlations, the inter-elemental associations determined by cluster analysis will be discussed very briefly. Relationships between elements and the origin of the various elemental associations will be discussed in more detail after factor analysis. This was deemed appropriate as factor analysis, being a more flexible technique, provides more realistic results, since it allows the participation of individual elements in more than one elemental association (factor), which is normally the case in nature. Therefore, the distribution of each element may be described in terms of all the associations the element takes part in, and thus conclusions concerning its various sources may be drawn. The stength of the affinity that each element exhibits towards the various 92 associations in which it may participate is reflected by the magnitude of the respective factor loading. Applying multivariate statistics on the Box-Cox transformed data in conjunction with its robust C.M., background relationships can be studied more efficiently. The application of multivariate statistics on the raw, untransformed data in conjunction with its robust C.M. results in underlining a potentially existing anomalous component in the multivariate data; as anomalous values are moderately allowed to influence the structure of multivariate functions. Although, in first approximation, it may seem inappropriate to examine first the Box-Cox transformed data and then the raw one, as the former is produced from the latter, this will be done due to the importance of establishing background relationships prior to identify a potentially existing anomalous component. 3.1.3.2 Correlation Matrices The robust correlation matrices of the raw, untransformed and the Box-Cox transformed data are presented in Table 3.2 a and b respectively. The lowest value of correlation coefficient (r) which is significantly different from zero at the 0.001 probability level is 0.135 (for the 177 samples of the Samothraki plateau data set), (Fisher,1963, cit. Howarth, 1983, p.400). A comparison of the two correlation matrices shows that there are 27 cases, corresponding to 14.21% of the total number of pairs, where the correlation coefficients have a difference ^ 0.1-0.2. In all the other instances, the robust coefficients calculated using raw data differ less than 0.1 from those calculated using the Box-Cox transformed data. This comparison shows the similarity of the two correlation matrices and suggests that the elemental associations described by them should be almost identical. T a b le 3.2 : Robust correlation matrices of the raw, untransformed (a) and the Box-Cox transformed (b) data of the Samothraki plateau. Wh Hit CA OR BA LA 1 I ZR K 1.00000 BET .75840 1.00000 MU .75280 .53040 1.00000 CA -.8 5 5 0 0 -.6 0 9 9 0 ~ . 10090 1.00000 SR “ .n.5400 -.6 3 6 9 0 -.1 4 7 0 0 .95370 1.00000 BA ,69430 .51160 -.1 2 3 4 0 -.6 0 5 0 0 -.5 6 1 3 0 1 .00000 LA -.1 1 7 0 0 .20500 .31090 .27860 .29060 .00560 l .00000 TI .72110 .00190 .64820 -.7 2 3 9 0 -.7 4 6 9 0 .40220 .11870 l .00000 ZR .63080 .64790 . 10310 -.6 7 1 1 0 -.6 2 5 3 0 .36160 ,06880 .68590 1 . OOOOO V ♦ .57 460 .49490 .53610 -.3 9 3 4 0 -.4 0 8 5 0 .13410 ,00410 .38310 .18970 1.00000 CR .54320 .73900 .60730 -.6 2 9 6 0 -.6 7 7 2 0 .24140 .09380 .90130 .55020 .01610 MN .32210 .57350 .64510 -.29430 -.3 2 0 9 0 .13650 .18030 .70 t 50 .39300 . 48 I 30 FE .66110 .05970 .72840 -.6 2 7 7 0 -.6 6 6 9 0 .33290 .136 70 .93260 .32180 .66000 CU .46000 .6 3 150 .46920 -.4 6 1 2 0 -.4 5 3 6 0 .41540 .28480 .63480 . 39810 . 45/VO NI .4 VI HO .67970 .67450 -.4 7 5 0 0 -.5 1 4 0 0 .19330 . 18110 .77380 .33070 .70000 CU .33090 .59740 .36020 -.4 0 2 4 0 -.3 9 3 2 0 .35960 .28830 .62970 .37240 .4:5700 ZN »50360 .77350 .49300 -.4 6 5 f i 0 -.4 4 5 8 0 .50790 .32980 .77110 .43090 .46 l VO Al. .860H0 .93130 .55570 -.0 0 6 6 0 -.8 0 2 9 0 .33680 .08930 .9-1960 .69420 . 00 I • .0 HI .59640 .19550 -.4 7 7 1 0 -.7 3 7 6 0 -.6 8 0 7 0 . 6 6 3 3 0 - . 40630 .17830 .48250 -.0 7 6 2 0 P . 6 2850 .03490 .57050 -.3 9 9 7 0 -.5 7 5 7 0 .46930 .227VO .87840 .70680 .01.540 I'P MN PL CO NI t.U ZN AL S 1 CR 1 .000O0 MN .66270 1.00000 FE .88530 ♦ 74,5.50 1 .00000 Cil .63400 .50440 .67300 1.00000 NI .81820 .67350 .05000 .64530 l .00000 CU .6741(0 .54490 .60750 .65930 .61390 1.00000 ZN .64340 .60050 ♦77900 .70440 .66760 .80290 1 . OOOOO Al. .00570 .59260 .90530 .67460 .72680 .35200 .70020 1 .OOUOO SI .10730 -.1 0 6 1 0 .02820 .06390 -.0 8 6 4 0 -.0 1 1 9 0 ,01910 .30600 1 . OOOOO P ♦ 75030 . 7619Q.'- .05970 .67790 .68730 .67520 .75820 .84250 .10350 1 . OOOOO 94 © o © c o © o c C COCO c o o o o o c c c c coco >00 c in in rrs m re c U' C o o o o c o c © COO o O' «r «r rv re O If- S3 o S3 in x *r o x re S3 m © ^ © oc cooocoocccco c o c o o x p- c •3 o o fs £ 3 o © © © o © © © o © o <3 •>4 O' «r n ■p o X o © © © r*. © © © o r. ^ o © © c © c © © © © o iv S3 O' X © NA T-* n .*> X k; « S3 © G r. S3 W Ci ^ o © Av © © O © © © © © © © o o o c © o c c o o © © rv © © © © © © © © © © © © Q © o o o o © © o c o © c © © re O' re t iu o £ 95 3.1.3.3 Cluster Analysis 3.1.3.3.1 Box-Cox transformed data Cluster analysis was applied to the Box-Cox transformed data in conjunction with its robust correlation matrix. The dendogram produced is presented in Fig. 3.29 a. It displays graphically the following elemental associations, for which tentative causes are presented : Ca - Sr : It is considered to represent the biogenic carbonate phase of the sediments. Strontium is associated with Ca because it tends to isomorphically substitute for it. The isolation of this class suggests that the rest of the elements determined are not likely to be significantly associated with the carbonates. Cu -Zn -Co-La : It could represent an association related to Zn-Cu sulphide deposits that occur in the mafic rocks of the mainland. Although La is found in the same megabranch with Cu, Zn and Co its association with these elements is, most probably, more fortuitous than real, according to the low correlation coefficients of La with each of Zn, Cu and Co. Mn -P : It could possibly reflect authigenic sedimentation of hydrated manganese dioxide which, having a great adsorptive power (Mason, 1966, p.183), could have adsorbed some P from the water column. Be - Al - Ti - Fe : It probably reflects terrigenous detrital material, mainly clay minerals and possibly small amounts of Fe.Ti-oxides. Mg - V - Cr - Ni : It could represent weathering products of mafic/ultramafic rocks in which minerals containing these elements occur. Zr - Si - K - Ba : It probably reflects detrital material of felsic rocks. The K - Ba association could be considered as representing K-feldspars; the Zr - Si one, probably, indicates the presence of traces of zircon. 3.1.3.3.2 Raw, untransformed data Cluster analysis was also applied to the raw, untransformed data in conjunction with 96 its robust C.M. The dendogram produced is shown in Fig. 3.29 b. It shows that La has moved from the Cu - Zn - Co branch, with which it was associated in the dendogram that corresponds to the Box-Cox transformed data , to the Ca - Sr branch. However, the correlation coefficients of the pairs Ca-La and Sr-La are low. Consequently, the association of La with Ca and Sr is considered to be a fortuitous one and the erratic distribution of La was confirmed. Otherwise, the dendogram is virtually identical to that discussed in section 3.1.3.3.1. This was expected due to the similarity of the correlation matrices (see section 3.1.3.2) used in the two cases. 97 a COPHENETIC CORRELATION COEFFICIENT r 0-6238 d o r d z □ c r z u j j - i u o cr — or •— c e u(nuNUJZQ.m(ih-u.z:>uZN(/)5cco IU _ l — LU cd a t - u. I V l 98 COPHENETIC CORRELATION COEFFICIENT = 0-5384 c eo tceq: — o z u j j ^ l j o d z q :— cr UWJU2E>EQ.mG[hli.UUMNWii:(D LU _J — UJ co cr t— u. Fig. 3.29 Cluster analysis of Samothraki plateau data; robust dendograms: (a) Box-Cox transformed data, (b) raw data. 99 3.1.3,4 Factor .Analysis As the robust C.M. of the Samothraki plateau raw data is similar to the one calculated on the basis of the Box-Cox transformed data, as far as elemental associations are concerned, factor analysis was applied to the raw data in conjunction with its robust C.M. A 5-factor model (Table 3.3) was chosen on the basis of the following: 1. The CattelTs Scree test (Appendix V) shows that only factors characterized by an eigen value (Appendix V) higher than 0.63 are significant. It suggests, therefore, that a 5-factor model would be a satisfactory solution (Table 3.4). 2. The five factors account for a very high proportion, 88.4%, of the total data variability. Six- and seven- factor models were tested for comparison purposes; however, they did not improve the factor solution. In this study, factor loadings higher than 0.5 in absolute value (Table 3.3) are considered to be important. As in the case of the cluster analysis, some possible causes of the elemental associations will be presented, but detailed discussion will follow in Chapter 8 , where the partition analysis results will also be considered. Factor 1: The elements V, Ni, Mg, Fe, Cr, Al and Ti have significant loadings on this factor, which is likely .to reflect terrigenous detrital material brought in by rivers, as these elements are variably enriched in river derived sediments in the area. The large proportion, 24%, accounted for by this factor would suggest a substantial influence of derital river runoff to the offshore compositional variation. The terrigenous derital material mainly consists of: (a) clay minerals represented by Al and Fe, and by absorbed and/or adsorbed V, Ni, Mg, Cr and Ti; (b) traces of primary minerals deriving from mafic/ultramafic rocks, being represented by V, Ni, Mg, Fe, Cr and Ti; and (c) traces of heavy minerals represented by Mg, Fe, Cr and Ti. Factor 2 : This factor accounts for 26.6 % of the data variability. It exhibits high Si, K, Ba, Zr, Al and Be loadings, and high - in absolute value - negative Ca and Sr 100 Table 3.3 : Factor analysis of raw data of the Samothraki plateau. Five factor model: factor loadings and variance accounted for by components, computed on the basis of the robust C.M. of the raw data. Element Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 K -.38 .84 .07 .06 .22 Be -.48 .53 .34 .26 .46 Mg -.72 -.21 .23 .05 .48 Ca .30 -.86 .21 -.18 -.2 3 Sr .37 -.79 .27 -.1 5 -.2 6 Ba .08 .81 .15 .39 -.01 La -.02 -.21 .88 .24 .09 Ti -.57 .44 .08 .27 .61 Zr .02 .63 .11 .13 .63 V -.85 .09 -.1 0 .25 .03 Cr -.62 .28 -.06 .33 .52 Mn -.40 -.06 -.01 .38 .71 Fe -.68 .30 .09 .36 .52 Co -.43 .27 .23 .63 .15 Ni -.80 .13 .04 .38 .28 Cu -.26 .12 .07 .84 .29 Zn -.36 .26 .25 .71 .31 Al -.56 .61 .15 .20 .48 Si .23 .89 -.29 .01 -.0 9 P .37 .36 .18 .41 .66 Varianceaccounted for by components Component Variance percent Cumulative .variance t 23.55 23.55 2 26.59 50.14 3 7.13 57.27 4 14.19 71.46 5 6.95 78.41 101 loadings. The former association probably reflects the presence of detrital material derived from felsic rocks, and the latter one the occurrence of biogenic carbonates. Factor 3 : This is essentially a single element factor, loaded with La. As the corresponding robust C.M. (Table 3.2 a) shows, La displays low correlation coefficients with each of the other elements determined. Lanthanum affinities with other elements are not depicted and La forms, therefore, a class of its own. Factor 4 : This comprises high Cu, Zn and Co loadings and it accounts for 14% of the data variability. This factor may reflect the weathering products of the Zn-Cu sulphide deposits that occur in the mainland. The incorporation of Co in this factor probably shows that Co is an accessory element of these sulphide deposits. Factor 5 : This comprises high Mn, Zr, Ti, Cr and Fe loadings. The Zr, Ti, Cr and Fe association is likely to reflect the presence of traces of heavy minerals. The occurrence of small amounts of Fe,Ti-oxides is probable. Manganese is loaded in this factor, possibly due to the association of Mn-oxides with Fe-oxides. Table 3.4 : Samothraki plateau raw data. List of eigen values of all the components of the data. Component Eioen Value .Qompoaent .Eigen .Value 1 11.40 11 .13 2 3.41 12 .12 3 1.39 13 .11 4 .86 14 .07 5 .63 15 .07 6 .45 16 .06 7 .41 17 .04 8 .36 18 .03 9 .26 19 .02 10 .18 20 .01 102 3,2_ Buried sediments Two cores from the Samothraki plateau were studied. They were collected from the northeast part of the plateau, close to the Evros river mouth (Fig. 3.1). 3.2.1 Description of the_cores The texture of the cores is generally silt to silty mud. Both cores were extensively mottled. The colours varied from dark yellowish brown to olive grey; in addition, greyish black sediments were found towards the base of one of the two cores (Moorby, pers. comm., 1985). 3.2.1.1 Core THR 7 (Fig. 3.30 a) The core liner was full and the surface sediment was most probably lost. This core consists predominantly of silt. Occasional large shells were found at around 4,134,143 and 170 cm depth. A fragment of wood was seen at about 60 cm depth. The colour varies with depth. There is a very thin layer (<2 mm) of moderate yellowish brown sediment (10YR 5/4) at the top of the core. Below it, down to 50 cm depth, the predominant colour is dark yellowish brown (10YR 4/2). Extensive streaked out mottles of olive grey (5Y 4/1) to dark grey (N3) sediment occur in the uppermost 50 cm. Below 50 cm depth towards the base of the core, the colour of the sediment gradually changes from dark yellowish brown (10YR 4/2) to olive grey (5Y 4/1). In the lower part of the core, the mottling decreases. 3.2.1.2 Core THR 22 fFio. 3.31 a) The core liner was full. A lack of browner coloured sediment at the top of the core indicates that the surface sediment has probably been lost. This core consists of silt to silty mud. A strong smell of H 2S was noticed when the core was first opened, implying that the sediments were reducing. The colour varies with increasing depth throughout the core. Thus, the colour changes from dark yellowish brown (10YR 4/2) to olive grey (5Y 4/1), to olive black (5Y 2/1) and greyish black (N2). Mottling is visible throughout the core. The mottles 103 are particularly abundant at 0-25 cm, 38-46 cm, 62-80 cm, 102-118 cm and 167-178 cm. 3JL2__Bulk geochemistry of subsurface offshore sediments In this section an attempt is made to study the variation in the chemical composition of the subsurface sediments with depth. To facilitate this, the chemical data have been plotted for each of the cores (Fig. 3.30 b and 3.31 b). 3.2.2.1 Core THR 7 (Fig. 3.30 b) The Ca content of this core decreases sharply in the uppermost 20 cm, and then it generally increases gradually. The elements K, Be, Mg, Ti, V, Mn, Fe, Cr, Al and P exhibit an inverse relation to Ca, suggesting their close association with the non-carbonate fraction of the sediment. The Cu content of the sediment decreases gradually with increasing depth. Manganese is relatively enriched in the top 50 cm, while its content decreases below this depth. A relative depletion in Mg, Ti, Cr, Mn, Fe and Cu occurs at the base of the core. 3.2.2.2 Core THR 22 fFin. 3.31 b^ The Ca content of this core decreases overall with depth, in parallel with a change of the sediment colour from dark yellowish brown to greyish black. The trend of the downcore distribution of Sr is similar to that of Ca. The elements K, Al and Be follow quite similar, overall, patterns in their downcore variation. The Fe and Ti distributions display similarities. A relative enrichment in K, Be, Ti, Zr, V, Al and Si occurs around 95 cm corresponding to a low Ca content. The Mg content of the sediment appears to increase overall with depth. Copper follows Mn in its downcore distribution; however, the Mn pattern is more pronounced. Manganese decreases in concentration with depth in the uppermost 20 cm; the opposite occurs from 20 to 40 cm; below 40 cm the Mn content decreases gradually. 104 THRRKI 7 Fig. 3.30 CORETHR 7: (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr,Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P (depth in cm, concentrations in p.g/g). s f : shell fragments, m : mottles, w : wood. 105 THRRKI 22 Fig. 3.31 CORE THR 22 : (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr.Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, At, Si and P (depth in cm, concentrations in pg/g). m : mottles. 106 3.3 Summary The Samothraki plateau surface sediments are largely sands and silty sands. However, finer grained sediments occur in the vicinity of the main river mouths and seawards of the shelf break. The buried sediment studied is silt to silty mud, with extensive mottles, and occasional shells and shell fragments present. The regional distribution of Si in the surface sediments is similar to the variation of quartz in the coarse fraction and almost the reverse of the Ca distribution. The distribution patterns of Al and Fe show an offshore enrichment in an extensive belt, stretching off the Evros river mouth, westwards as far as Xylagani, and centred about 5km offshore. This pattern of Al and Fe enrichment corresponds closely with the distribution of the finest sand component (30-4o) of the sediments. In most cases, areas enriched in Ti, Cr and Zr are associated with a relatively high heavy mineral content of the sediment coarse fraction. However, the Ti, Cr and Zr enrichments in the northeastern part of the Samothraki plateau extend much further westwards than that of the heavy minerals in the coarse fraction. A comparison of the robust correlation matrices of the raw and the Box-Cox transformed data shows that the two matrices are similar. Therefore, the elemental associations illustrated by the corresponding dendograms, obtained from the cluster analysis, are almost identical. The following elemental groupings are formed, for which a possible interpretation is presented: (a) Ca-Sr : biogenic carbonate phase; (b) Cu-Zn-Co : association related to Zn-Cu sulphide deposits of the mainland; (c) Mn-P : authigenic sedimentation of hydrated Mn-oxides and P adsorption; (d) Be-AI-Ti-Fe : terrigenous detrital material, mainly clay minerals and probably small amounts of Fe,Ti-oxides; (e) Mg-V-Cr-Ni : weathering products of mafic/ultramafic rocks; and, (f) Zr-Si-K-Ba : detrital material of felsic rocks. Factor analysis has been used on the basis of a 5-factor model. The interpretation of the obtained factors shows that : (a) the river runoff (reflecting terrigenous material) has an important influence to the offshore compositional variation; and (b) biogenic carbonates are present in the sediments. Additionally, both cluster and factor analysis results suggest that the carbonate phase of the Samothraki plateau surface sediments is - apart from Ca and Sr - not likely to host significant amount of the element content of the sediments. 107 The variation in the chemical composition of the subsurface sediments shows that the elements K, Be and Al exhibit similar downcore distributional patterns. These elements are, in general, varying inversely to Ca, which is considered to represent the carbonate fraction. . 108 CHAPTER IV THE STRYMONIKOS GULF. PLATEAU AND TRIANGLE. AND THE KAVALA GULF 4.1 Surface sediments Chapter 4 deals with the sediments that were collected from the offshore area that lies to the west of the Thassos island. A total of 327 surface samples were collected from an area of 2,200 km2; the area has been much more intensely sampled than the Samothraki plateau. The sampling sites (Fig. 4.1) are relatively evenly distributed, with a clustering tendency towards the coastline. In order to achieve a better understanding of the offshore regional elemental distribution, 66 samples were collected from the coastal part of the adjacent mainland and from the coast of Thassos island (Fig. 4.2). 4.1.1 Sediment composition 4.1.1.1 Distribution of rock fragments, quartz, heavy minerals, biogenic matter and carbonates The maps that present the distribution of rock fragments, quartz, heavy minerals, biogenic matter and carbonates in the surface sediments are provided by Perissoratis (Perissoratis et al, in press). 4.1.1.1.1 Rock fragments The distribution of rock fragments in the coarse fraction (Fig. 4.3) shows that the greatest abundances occur in nearshore areas, off the Nea Peramos bay, in the Strymonikos ridge, and also locally in areas lying on the outer shelf. A local rock fragment enrichment occurs in the channel that separates the Thassos island from the mainland. By contrast, the rock fragments content of the sediments collected from the inner shelf and the central parts of the gulfs is low. 109 4.1.1.1.2 Quartz The distribution of quartz in the coarse fraction (Fig. 4.4) is similar to that of rock fragments. Thus, quartz is abundant in the nearshore areas and on the outer shelf, whereas the quartz content diminishes on the inner shelf and in the central parts of the gulfs. In general, sediments rich in quartz occur in areas with sand enrichment. Perissoratis et al (in press) remarked that the grains of quartz are of an angular shape in the sediments collected from the outer shelf whereas they are more rounded in those of the nearshore areas. 4.1.1.1.3 Heavy minerals The distribution of heavy minerals in the 2 0 to 3 0 fraction ( 0.25 to 0.125 mm, coarse fraction; Udden, 1914; Wentworth, 1922; Krumbein, 1936; Friedman and Sanders, 1978), (Fig. 4.5) indicates that high concentrations occur in the following areas: (a) a narrow coastal strip, off the west coast of the Strymonikos gulf; (b) the Strymonikos ridge; (c) just off the coastline between Loutra Eleftheron and Nea Peramos; (d) off the east coast of Kavala gulf; (e) the outer part of lerissos gulf; (f) the outer part of Kavala gulf; and, (g) near the shelf break, at the western part of the outer shelf. The distribution of selected heavy minerals, like amphiboles, pyroxenes, garnets, epidotes and metallics (Fig. 4. 6) shows that: 1. Amphiboles and pyroxenes occur in all the heavy minerals enriched a»3as. However, they are in particular abundant just off the coastline that lies between Loutra Eleftheron and Nea Peramos, and off the west coast of the Strymonikos gulf. 110 2. Garnet is abundant off the west coast of Strymonikos gulf, off the northern part of the mouth of lerissos gulf, at the outer part of Nea Peramos bay, and off the east coast of Kavala gulf. 3. Epidote is abundant in sediments collected between Nea Peramos bay and Thassos island. 4. Magnetite and ilmenite are present off the southwestern coast of the Strymonikos area. The high amphibole and pyroxene content determined in the nearshore sediments is attributed to the erosion of mafic and metamorphic rocks on the mainland. The weathering products of these rocks are brought to the offshore area either by rivers (i.e. Strymon in the Strymonikos gulf) or by the action of coastal erosion (as in the case of the amphibole and pyroxene enrichment in the sector lying between Loutra Eleftheron village and Nea Peramos bay). The heavy minerals enrichments determined in the outer parts of the lerissos and Kavala gulfs and in locations close to the shelf break are associated with areas of relict sediments. Thus, these enrichments are likely to be ascribed to deposition during earlier periods of low sea level. In an attempt to explain the heavy minerals enrichment off the east coast of the Kavala gulf, the following remarks are pertinent: 1. Big rivers are absent from the nearby mainland. 2. Mafic and metamorphic formations do not outcrop on the adjacent mainland. 3. Nestos river was flowing out on the east coast of Kavala gulf, during the upper Pleistocene-lower Holocene lower sea level stands (see section 1.4.). At that time, the river mouth was in the channel formed at the middle of Nea Peramos - Thassos ridge. The heavy minerals enrichment off the east coast of the Kavala gulf is, therefore,likely to be attributed to transport and offshore deposition of heavy minerals rich sediments by the Nestos river. These sediments might have been further subjected to sorting by wave action. 111 4.1.1.1.4 Biogenic carbonate particles The biogenic components of the sediments are represented, primarily, by benthonic foraminiferal tests and, secondarily, by shell fragments, ostracods and planktonic foraminifera (Perissoratis et al, in press). The distribution of biogenic components in the coarse fraction (Fig. 4.7) shows that high concentrations occur in the inner parts of the gulfs and in the inner shelf, corresponding to areas where fine grained sediments predominate. In sandy areas, the proportion of foraminiferal tests diminishes while that of worn and broken shell fragments increases. 4.1.1.1.5 Carbonates The distribution of carbonates (Fig. 4.8) indicates that high concentrations occur in a rather extensive sector that lies in the outer part of the Kavala gulf and to the west of Thassos island, on Strymonikos ridge, in the central part of Strymonikos gulf, and in the central area of the channel which separates Thassos from the mainland. 4.1.1.2 Distribution of organic carbon and results from XRD and SEM studies In order to obtain more information on the sediments, the organic carbon content was determined, and also selected samples were subjected to XRD and SEM studies. 4.1.1.2.1 Organic carbon The distribution of organic carbon (Fig. 4.9) shows an area of enrichment lying off the northwest coast of Thassos island, with values of up to 4.10%. The sediments collected from the Strymonikos triangle and the east part of the Strymonikos plateau are poor in organic carbon; its content ranges from 0.30 to 0.60%. The western part of the Strymonikos plateau is moderately richer in organic carbon than the eastern part of it. 4.1.1.2.2 XRD results The samples KB 31, KB 38, KB 44, KB 54, KB 104, KB 112, KB 135, STR 69, STR 94, STR 99, STR 160 and STR 167 were subjected to X Ray Diffraction analysis. The 112 minerals recognized are listed in Table 4.1. Table 4.1: Mineralogical composition of selected samples from the western part of the study area as determined through XRD analysis. Sample I.D. Major minerals Minor minerals Traces KB 31 quartz calcite kaolinite/chlorite plagioclase high Mg-calcite aragonite K-feldspar illite KB 38 quartz calcite illite plagioclase aragonite K-feldspar high Mg-calcite KB 44 quartz calcite hornblende K-feldspar illite muscovite plagioclase aragonite kaolinite/chlorite KB 54 quartz calcite plagioclase high Mg-calcite illite K-feldspar aragonite kaolinite/chlorite KB 104 quartz K-feldspar calcite aragonite plagioclase illite KB 112 plagioclase aragonite illite quartz hornblende K-feldspar calcite KB 135 quartz calcite K-feldspar aragonite high Mg-calcite plagioclase STR 69 quartz high Mg-calcite aragonite calcite kaolinite/chlorite plagioclase K-feldspar illite 113 Table 4.1 (continued) STR 94 quartz high Mg-calcite calcite K-feldspar plagioclase kaolinite/chlorite illite hornblende aragonite STR 99 quartz K-feldspar kaolinite/chlorite plagioclase illite calcite high Mg-calcite STR 160 quartz K-feldspar kaolinite/chlorite calcite plagioclase illite low christobalite high Mg-calcite STR 167 quartz calcite hornblende plagioclase aragonite illite high Mg-calcite K-feldspar 4.1.1.2.3 SEM results The samples STR 8, STR 83 and KB 148 were examined by Scanning Electron Microscopy. Feldspar (represented by K, Al and Si), mica (represented by K, Al.Si, Fe and Mg), and quartz (represented by Si) were identified in all of them. In addition, sample STR 83 contains also zircon (represented by Zr and Si), sample KB 148 calcium carbonate (represented by Ca), and sample STR 8 hornblende (represented by Al, Si, Fe, Mg, Ca and K). 8' 16 7 5 ^ 1 «ei 8^2 •85 7? j o i 114 ,7 .5 76 *77 -66^r * 88 • _ 6s 6 $8 07 69 * * •87 . 67 6 ^ 17 . ’9 29 64 . 6 _ * 3 6 5 V* 11 *25 *26 30 63 Q 90 ' 6+ :18 10 km ^ • ' . 2 ° .23 31 91 95 58 *57 19 9 a 94 9 6 55 *56 121 151 97 *22 3? 54 ^ 3 3.5 92 93 102 100 ^58^.52 (25 53 ?4 152 *153 . 4 3 10* 47 ^ 36 103 * 45 18 !6, 2y 124 18 . 116 127‘ 123 37 39 4 24 4 jfo 150-128 114 38 411 148i49 ^ 1 2 9 fl? 113 \3 5 r & i.7 Q •172 *81 11Q 58 72 - -81 ,X ^ 86 >?, * t 6 3 121 10^ 6073 74 f9 60 *84 10 3 ’*138 • 85 87- ,164\ 112 #% V 91 ^0 * 0% _ .101 .ioi“^j« ^T tL33g*137 *»• 62 *^9 2 89 88 99 100 s ’20 *105 v»0 165 HI 134] , " 17fc $5 \ 119 .96 98 128 § 7 .141 162 • 136 106 T h a e a o o 53 V .70 JM \ .107 122 106 97 J08 • • *s 26*4 *178 68 ' J66 \ 105 I 9 * .44 126 150 142 161 >132 124 V 110 4 j J C i ^ <45 .125 109 ^ • 133 149 142 160 ,170 \ A®\ .42^ *39 143 107 137 V 2 .3 8 oa 123 151 14*1 A . 24 37 .<► * J59 \ u 3 ( \ 138 i5o a * 3 4 * 35 .1.21 148 169 • r-* GO I 44 <0 S h 3 0 33 . \ 167 29^Vr 23 * ** \ 147 *116 145 158 \ Ka v 112 \ \ .1 1 4 113 146 • ,176 154 .177 Btpymonlkoa group 156 155 • surface sample® a c o r e s Fig. 4.1 Strymonikos and Kavala areas : map representing the offshore surface and subsurface sample locations. NEA KARVALI Fig. 4.2 Strymonikos and Kavala areas : map representing the onshore sample locations. 116 Fig. 4.3 Strymonikos and Kavala surface sediments : distribution of rock fragments in the coarse fraction. - 4 0 40 Fig. 4.4 Strymonikos and Kavala surface sediments : distribution of quartz in the coarse fraction. 117 lOHm STAVROS 6 — 8 IERISSOS Fig. 4.5 Strymonikos and Kavala surface sediments : distribution of heavy minerals content in the coarse fraction. Fig. 4.6 Strymonikos and Kavala surface sediments : distribution of selected heavy minerals in the 2 0 to 3 0 fraction. 10 118 Fig. 4.7 Strymonikos and Kavala surface sediments : distribution of biogenic components in the coarse fraction. ► i i 1 1 9 Fig. 4.8 Strymonikos and Kavala surface sediments : distribution of total carbonates Fig. 4.9 Strymonikos and Kavala surface sediments : distribution of organic carbon. 120 4-1,2 Bulk geochemistry of offshore sediments and of those on_ the adjacent coast Elements are dealt with in order of decreasing abundance, as this is established on the basis of robust means (Table 8.1) of Box-Cox transformed geochemical data of the Strymonikos group of samples (for a definition on robust statistics and Box-Cox transformation see Appendix V). 4.1.2.1__ Silicon The distribution pattern of Si (Fig. 4.10) reflects closely the percentage of sand-sized material in the sediments. An enrichment in Si content occurs in a narrow offshore belt stretching along much of the mainland coastline. However, around the coastline of Thassos island, low concentrations of Si were determined, apart from the coastal area lying in Limenaria bay. In addition, Si enrichments occur in the outer part of lerissos gulf, in the coarse grained sediments of the Strymonikos ridge, and southwards of Nea Peramos bay. By contrast, the central part of Strymonikos gulf and also Strymonikos plateau are, rather, Si poor. Lower than 22% Si content was determined in the sediments of both areas. Low Si values occur in the central part of Kavala gulf, with a tendency for it to increase around Kavala valley. The samples collected from the beaches and the Strymon river sediments are all enriched in Si (Fig. 4.10), with higher than 29% Si content. 4.1.2.2 Aluminium The regional variation of Al (Fig. 4.11) shows that an antipathetic relationship occurs between Al and Si. Thus, offshore sediments collected around the coastline and from the Strymonikos ridge are poor in Al, while a high Al content was determined : (a) in the sediments of the Strymonikos plateau and gulf; and (b) in the samples collected eastwards and westwards of the Kavala valley. Most of the beach samples as well as the Strymon river sediments are rather Al poor (Fig. 4.11). Samples relatively enriched in Al were collected from the beach between Loutra Eleftheron and Nea Peramos. 121 4.1.2.3 Calcium The distribution pattern of Ca in the offshore sediments (Fig. 4.12) shows that Ca is generally low throughout much of the Kavala and Strymonikos gulfs and also on the Strymonikos plateau. The most extensive area of Ca enrichment, with higher than 12% Ca content , occurs in the outer part of Kavala gulf, and off the west coast of Thassos island. In addition, Ca rich sediments were collected from the following areas: (a) the channel which separates Thassos island from the mainland; (b) off the south coast of Thassos; and (c) seawards of that part of the coast lying between Loutra Eleftheron and Nea Peramos. A low Ca content was determined in the onshore samples (Fig. 4.12), apart from the samples collected from the southwest coast of Thassos island. 4.1.2.4 Iron The distribution pattern of Fe (Fig. 4.13) is quite similar to that of Al. Sediments collected from the whole of the Strymonikos gulf, apart from a narrow coastal zone and the Strymonikos ridge, are rich in Fe with higher, than 4.5% Fe content. This area of enrichment extends eastwards over much of the Strymonikos plateau. The sediments of Kavala gulf are Fe enriched, however, to a lesser extent than those of Strymonikos gulf. Kavala valley sediments are Fe poor. The highest Fe content was determined in the sediments of Limenaria bay; Fe values up to 9% occur there. Low Fe concentrations were determined in the onshore coastal samples (Fig. 4.13). 4.1.2.5 Potassium The distribution of K (Fig. 4.14) is different to that of any of the other major elements. The following areas were characterized as being enriched in K : a. a narrow east-west belt, parallel and just off the coastline, stretching from the Strymon river mouth eastwards as far as Nea Peramos; b. eastwards of the Strymonikos ridge; c. off the west coast of the Kavala gulf; 122 d. off the east and southeast coast of the Kavala gulf; and e. southwards of the Limenaria bay. The distribution pattern of K in the onshore samples (Fig. 4.14) shows that a K enrichment occurs along the coastline lying between the Strymon river mouth and the Kavala town, with a K content higher than 2.8%. In addition, K rich beach samples were collected from Olympias district, Asprovalta village, the east coast of the Kavala gulf, and the northwest coast of Thassos island. Strymon river sediments are also K rich. 4.1.2.6 Magnesium The regional variation of Mg (Fig. 4.15) is similar to that of Fe (Fig. 4.13). Strymonikos gulf sediments, apart from those collected from the ridge, are Mg rich. This area of enrichment extends eastwards over much of the Strymonikos plateau. Kavala gulf sediments are enriched in Mg, although not to the same extent as the Strymonikos gulf sediments. Samples collected from the coastal area of the adjacent mainland are Mg poor (Fig. 4.15). 4.1.2.7 Titanium The regional variation of Ti (Fig. 4.16) is quite similar to those of Al (Fig. 4.11) and Fe (Fig. 4.13). The sediments of the Strymonikos plateau and gulf, apart from the ridge and a narrow coastal belt off the north coast, are rich in Ti. High Ti values were determined in the Kavala gulf, except for the valley area. In all the above Ti enriched areas, the Ti content of the sediments is less than 0.5%, but a small area of Ti enrichment occurs off the Stratoni peninsula, in the southern part of the Strymonikos plateau, with Ti values of up to 0.55%. The highest Ti concentration (0.9% Ti) was determined in a Fe rich sample collected from Limenaria bay. By contrast, the channel between Thassos island and the mainland is Ti poor. Onshore samples collected from the coasts are rather poor in Ti. However, the Fe and Mg rich sample collected from the southwestern coast of Strymonikos gulf, and the Strymon river sediments are, relatively, Ti enriched (Fig. 4.16). 123 4.1.2.8 Phosphorus The regional variation of P (Fig. 4.17) does not closely follow that of any major element; however, it does show some similarities with the distribution patterns of Al (Fig. 4.11) and Fe (Fig. 4.13). High P concentrations were determined : (a) in the Strymonikos gulf (apart from the ridge); (b) off the Strymon river mouth and to the east of it, following the submarine canyon; (c) off the north coast of Kavala gulf; and (d) in the Fe rich Limenaria bay samples (with 1,100 pg/g P content). The maximum P content observed (1,200 pg/g ) occurs in one isolated coastal sample collected off the northern part of the Stratoni peninsula. The sediments of the Kavala valley are P rich. By contrast, the Strymonikos triangle samples and the sediments collected from the channel to the north of Thassos are P poor. The onshore samples are, generally, P poor (Fig. 4.17). However, samples enriched in P were collected from : (a) the sediments of the Strymon river and the coast close to its mouth; (b) the river that flows out in the southwest part of the Strymonikos gulf; and (c) the river that flows out close to Loutra Eleftheron. 4.1.2.9 Manganese The regional variation of Mn(pig. 4.18) shows that a high Mn content occurs over much of the Strymonikos gulf and plateau, corresponding to an Fe enrichment. A Mn enrichment was also found in deep water sediments below the shelf break, on the outer part of the Strymonikos plateau. The Fe rich Limenaria bay sediments are also Mn rich. By contrast, sediments collected from the Strymonikos ridge are Mn poor. Samples collected from the Kavala gulf and valley, as well as from the Strymonikos triangle, are poor in Mn. An onshore coastal Mn enrichment (1,500 pg/g) occurs in the Olympias district (Fig. 4.18). River sediments collected from the Strymon , and the river that flows out in the southwest part of the Strymonikos gulf contain up to 1,000 pg/g Mn. 4.1.2.10 Barium The distribution of Ba (Fig. 4.19) does not closely follow that of any major element. Barium is enriched in many of the nearshore sediments, apart from those collected off 124 the southwest and south coast of the Strymonikos gulf and the samples from the coast to the east of Kavala town. Two extensive coastal strips with 1,000 -1,400 pg/g Ba content occur, as follows: (a) from Nea Peramos to Loutra Eleftheron; and (b) along the west coast of Thassos island. The Fe rich Limenaria bay samples contain 3,000 pg/gBa. Most of the onshore samples are Ba rich (Fig. 4.19). Thus, samples collected from the coast lying between the Strymon river mouth and the Nea Peramos bay contain up to 2,200 pg/g Ba. By- contrast, coastal samples from the southwest coast of the Strymonikos gulf, the Olympias district, and Nea Karvali are Ba poor (Fig. 4.19). 4.1.2.11 Strontium The distribution of Sr (Fig. 4.20) closely follows that of Ca. High Sr concentrations were determined in sediments collected from the Strymonikos ridge and triangle, the Kavala valley, and the channel between Thassos island and the mainland. Enrichment in Sr also occurs in a narrow coastal belt stretching between Strymon river mouth and Kavala town. Although the highest Sr content (over 2,000 pg/g) occurs in the most Ca rich samples, Sr enrichments (over 1,000 pg/g) were found in Ca poor coastal samples collected from Loutra Eleftheron to Nea Peramos. The onshore samples from the area between the Strymon river mouth and the central part of the west coast of Kavala gulf are enriched in Sr (Fig. 4.20), while the sediments in the remaining part of the coastal area are rather poor in Sr. 4.1.2.12 Zirconium The regional distribution of Zr (Fig. 4.21) shows that high Zr content occurs in nearshore sediments collected along much of the coastline, and particularly between the Strymon river mouth and the Loutra Eleftheron district, where Zr values reach 350 pg/g corresponding to a Si enrichment. The major part of Kavala gulf is enriched in Zr; the highest Zr values observed occur off the north coast (210 pg/g) and in Nea Peramos bay (250 pg/g). By contrast, samples collected from the Kavala valley are not Zr enriched. Zirconium is enriched in several deep water sediments, to the south of the Strymonikos plateau (up to 340 pg/g) and in the outer part of the lerissos gulf. 125 The highest Zr content observed occurs in the Fe rich Limenaria bay sediments, where concentrations reach 360 pg/g. The onshore coastal Zr variation (Fig. 4.21) shows various areas of enrichment, as follows: a. close to Nea Karvali; b. close to Kavala town; c. in the Loutra Eleftheron district; d. in the Strymon river, and to the west of the river mouth area; e. on the southwestern coast of Strymonikos gulf, as well as in the river flowing there; and f. on the central part of the west coast of Thassos. 4.1.2.13 Vanadium The regional variation of V (Fig. 4.22) is similar to that of Fe (Fig. 4.13). The most extensive area of V enrichment occurs in the western part of the Strymonikos gulf. Sediments collected over much of the Strymonikos plateau, from the Kavala gulf, and from the shelf break below the Strymonikos plateau are enriched in V. The Fe rich Limenaria bay samples contain 320 ^ig/g V. By contrast, the Strymonikos ridge sediments, those from Kavala valley, and the samples collected from the Strymonikos triangle are V poor. The onshore samples are generally V poor (Fig. 4.22). However, a high V content was determined in samples collected from the southwestern coast of the Strymonikos gulf (corresponding to a Mg; Fe, Ti and Zr enrichment) and from the Kavala district. 4.1.2.14___ Chromium The distribution of Cr (Fig. 4.23) is quite similar to that of Ti (Fig. 4.16). Thus, high Cr content occurs: (a) in the central northern part of the Kavala gulf; and , (b) 126 in the major part of the Strymonikos plateau and gulf, except for nearshore areas and the Strymonikos ridge. One offshore sample from the area off the northwest coast of Thassos island contains 750 pg/g Cr. However, this enrichment is only local, and there is no evidence of any extensive Cr enriched area lying there. The maximum Cr concentrations in the rest of the area do not exceed 220 pg/g, which occurs in the Ti and Zr rich sample (STR 112) off the Stratoni peninsula. Chromium concentrations are low in the Kavala valley, the Strymonikos triangle, and Limenaria bay sediments. The differences between the distribution patterns of Cr (Fig. 4.23) and Ti (Fig. 4.16) are as follows: a. the Ti rich sediments collected from the Limenaria bay are Cr poor; and b. the zone of Cr enrichment off the west coast of the Strymonikos gulf is displaced further seawards than the Ti one. The onshore samples are Cr poor (Fig. 4.23). However, the Zr rich beach sample (Fig. 4.21) collected from the Nea Karvali district, and the Mg, Fe, Ti and Zr rich sediment collected from the river that flows out in the southwest coast of the Strymonikos gulf are enriched in Cr. 4.1.2.15 Zinc The distribution pattern of Zn in the offshore sediments (Fig. 4.24) shows that a high Zn content occurs off the Strymon river mouth (170 pg/g Zn), and in the western part of the Strymonikos gulf. Elsewhere, patchy Zn enrichment occurs, without any apparent pattern to it. The Limenaria bay sediments are Zn rich (370 pg/g), and a high Zn content (400 pg/g) occurs in one sample (KB 48) collected off the northwest coast of Thassos. The maximum Zn concentration observed (480 pg/g) occurs in sample KB 121 (Fig. 4.1) collected from the upper terrace. However, this sample does not show unusual concentration of any other element determined. The onshore samples are generally Zn poor (Fig. 4.24). However, local enrichments were determined in the Olympias district, in the vicinity of the Limenaria village, and 127 at a couple of locations on the west coast of Kavala gulf. Finally, some of the Strymon river sediments are enriched in Zn. 4.1.2.16 Nickel The regional distribution of Ni in the offshore sediments (Fig. 4.25) shows that a high Ni content occurs: (a) in the sediments of the western part of the Strymonikos gulf; (b) over much of the Strymonikos plateau; (c) in the west and north parts of the Kavala gulf; (d) in the northern part of the channel between Thassos island and the mainland; (e) in Limenaria bay; and (f) below the shelf break, at the outer part of the Strymonikos plateau. The maximum Ni value observed (580 jxg/g) occurs in the Cr rich nearshore sample (KB 48), off the northwest coast of Thassos. The onshore samples are all rather poor in Ni (Fig. 4.25). 4.1.2.17 Lanthanum The regional variation of La (Fig. 4.26) does not follow the distribution of any other major or minor element examined. Sediments collected from a nearshore, east-west belt, stretching between Loutra Eleftheron village and Nea Peramos bay are enriched in La; some of these samples contain more than 100 pg/g La, reaching a maximum La content of over 200 pg/g. In the sediments of the outer part of lerissos gulf , a La enrichment occurs ; here, the Ti, Cr and Zr rich sample (STR 112) contains 100 pg/g La. A small area of La enrichment also occurs along the coastline just to the east of Kavala town. The Limenaria bay samples are enriched in La (150 pg/g). Onshore La enrichments were determined close to the town of Kavala, in the vicinity of Nea Peramos, just west of the Strymon river mouth, and close to Loutra Eleftheron (Fig. 4.26). 4.1.2.18 Copper The regional variation of Cu in the offshore sediments (Fig. 4.27) shows that high Cu contents occur in sediments collected over much of the Strymonikos gulf, apart from the area close to the Strymon river mouth. The most extensive area of Cu enrichment occurs in the central part of the Strymonikos plateau; it also extends southwards of the plateau, to the shelf break. Sediments collected from various parts of the Kavala gulf are enriched in Cu. Additionally, local Cu enrichments occur off the north, 128 northwest and west coast of the Thassos island. The onshore samples are, in general, poor in Cu (Fig. 4.27). 4.1.2.19_ Cobalt The regional distribution of Co in the offshore sediments (Fig. 4.28) shows that they are , in general, poor in Co, in particular those collected from the Strymonikos triangle and the eastern part of the Strymonikos plateau. However, local small enrichments occur in the western part of the study area. The onshore samples are poor in Co (Fig. 4.28). 4.1.2.20 Beryllium The regional variation of Be (Fig. 4.29) closely follows the spatial distribution patterns of Al (Fig. 4.11) and Fe (Fig. 4.13). A large area of Be enrichment occurs in the Kavala gulf, with a Be content higher than 3.2 pg/g. Sediments collected over much of the Strymonikos gulf (apart from the ridge) and the Strymonikos plateau are rich in Be. The Strymonikos triangle and the Limenaria bay samples are Be poor. The onshore samples are, in general, poor in Be (Fig. 4.29). However, relative Be enrichments were determined in the vicinity of Loutra Eleftheron and in the central part of the west coast of Kavala gulf. 129 Fig. 4.10 Strymonikos and Kavafa areas and adjacent coast: regional variation in the Si content of the surface samples. Fig. 4.11 Strymonikos and Kavala areas and adjacent coast: regional variation in the Al content of the surface samples. 130 R.Str ymon >12.0 E2a 7.5-12.0 5.0- 7.5 Gnna 2 5 - 5.0 t= 3 Fig. 4.12 Strymonikos arid Kavala areas and adjacent coast: regional variation in the Ca content of the surface samples. Fig. 4.13 Strymonikos and Kavala areas and adjacent coast: regional variation in the Fe content of the surface samples. 131 Fig. 4.14 Strymonikos and Kavala areas and adjacent coast: regional variation in the K content of the surface samples. Fig. 4.15 Strymonikos and Kavala areas and adjacent coast: regional variation in the Mg content of the surface samples. 132 Fig. 4.16 Strymonikos and Kavala areas and adjacent coast: regional variation in the Ti content of the surface samples. Fig. 4.17 Strymonikos and Kavala areas and adjacent coast: regional variation in the P content of the surface samples. 133 S 350-450 ESa 250-350 [ ~ ] <250 Fig. 4.19 Strymonikos and Kavala areas and adjacent coast: regional variation in the Ba content of the surface samples. 134 Fig. 4.20 Strymonikos and Kavala areas and adjacent coast: regional variation in the Sr content of the surface samples. Fig. 4.21 Strymonikos and Kavala areas and adjacent coast: regional variation in the Zr content of the surface samples. 135 Fig. 4.22 Strymonikos and Kavala areas and adjacent coast: regional variation in the V content of the surface samples. R .S tr ym o n >130 100-130 S 70-100 E77H3 AO ~ 70 RS/9 0 3 < A0 Fig. 4.23 Strymonikos and Kavala areas and adjacent coast: regional variation in the Cr content of the surface samples. 136 100-130 70-100 40-7 0 <40 Fig. 4.24 Strymonikos and Kavala areas and adjacent coast: regional variation in the Zn content of the surface samples. Fig. 4.25 Strymonikos and Kavala areas and adjacent coast: regional variation in the Ni content of the surface samples. 137 R .S lry m o n >55 45-55 35-45 Ema 25 - 35 n g /g <25 Fig. 4.26 Strymonikos and Kavala areas and adjacent coast: regional variation in the La content of the surface samples. Fig. 4.27 Strymonikos and Kavala areas and adjacent coast: regional variation in the Cu content of the surface samples. 138 >3.2 2.6-3.2 L = a l 2.1 “ 2.6 i E S 3 15-2.1 EZ3 < 1.5 Fig. 4.29 Strymonikos and Kavala areas and adjacent coast: regional variation in the Be content of the surface samples. 139 4.1.3 Multivariate statistical analysis Multivariate statistical analysis was applied to the data concerning the elemental composition of the offshore surface sediments, for reasons similar to those explained in the section 3.3.1. Before being treated statistically, the set of samples studied in this chapter was divided into two groups, the Strymonikos and the Kavala groups. This aims to examine whether the sediments to the west of Thassos island are geochemically homogeneous. The Strymonikos group consists of the samples (STR) collected from the Strymonikos gulf and plateau (Fig. 4.1). The Kavala group comprises the samples (KB) that correspond to the Kavala gulf, the Strymonikos triangle and the offshore area close to the island of Thassos. 4.1.3.1 Strvmonikos group 4.1.3.1.1 Correlation Matrices The robust correlation matrices of the raw (untransformed) and the Box-Cox transformed data are presented in Table 4.2 a and b, correspondingly. The lowest value of correlation coefficient (r) which is significantly different from zero, at the 0.001 probability level, is ^ 0.132 (for the 179 samples of the Strymonikos group data set), (Fisher, 1963; cit. Howarth, 1983). Comparing the two correlation matrices, it was found that there are 43 cases, corresponding to 22.6% of the total number of pairs, where the correlation coefficients have a difference ± 0.1 - 0.2, and there is only one instance where the difference just exceeds ± 0.2 (this corresponds to the pair Ti-Si, which has a 0.202 correlation coefficient). In all the other cases, the differences of the corresponding robust coefficients calculated using either the raw or the Box-Cox transformed data are less than ± 0.1. Although large discrepancies between the two sets of coefficients are not shown, it was felt desirable to use both matrices in cluster analysis in order to be completely sure that there are no significant differences between these two sets. Table 4.2 : Robust correlation matrices of the raw, untransformed (a) and the Box-Cox transformed (b) data of the Strymonikos group. K BE MO CA HR BA LA 1 1 ZR 0 R 1 .OuOOO BE .71230 1.00000 mc .26530 .66200 i.00000 CA -.6 9 8 8 0 -.61390 -.30670 1.00000 SN -.6 1 4 2 0 -.63490 -.49760 .89820 1.00000 BA .47070 .07220 -.4 4 3 4 0 -.2 3 1 2 0 .02290 l .00000 LA .26360 .50400 . 205H0 -.0 4 8 7 0 .03610 .29220 1.00000 T1 .46020 .75240 .79590 -.6 3 9 1 0 -.7 3 5 0 0 -.1 7 1 5 0 .29070 1 . OOOiA) ZR .25230 .20540 -.1 133 0 -.3 0 5 1 0 -.1 6 6 2 0 .3 6 /1 0 .23600 . 305 70 1 . O000O V .24710 ♦46320 .51200 -.4 4 2 0 0 -.5 1 0 8 0 13340 .17740 .59560 .03580 1.00000 CR .11410 .49470 .79590 -.3 2 6 2 0 -.5 1 6 2 0 -.4 2 1 2 0 .21110 .77580 -.0 0 1 0 0 .61310 MN .10410 .42540 .68520 -.1 7 5 4 0 -.3 2 1 8 0 -.3 3 4 7 0 .18320 .65540 .01710 .35900 FE •398/0 .76000 .89860 -.4 9 3 8 0 -•662 50 -.3 4 2 4 0 .23520 .85870 -.0 8 1 0 0 »63580 CO .17270 .35390 .44440 -.1 1 5 3 0 -.2 3 6 9 0 - . 1.4090 ►24980 .39000 -.0 3 3 0 0 .34390 NI .19260 .47770 .75100 -.2 3 2 1 0 -.3 7 1 4 0 -.3 3 6 3 0 .18220 .63810 -.0 3 3 8 0 .62300 Cl) .18740 .45890 .55090 -.2 4 5 1 0 -.32560 -.12770 .27780 .50010 -.0 5 3 4 0 .35520 ZN .37030 .53350 .53230 -.3 5 8 8 0 -.4 1 3 4 0 .00380 .29100 . 56 720 .07500 .34200 AL .72640 .91940 .75730 -.7 2 2 4 0 -.7 8 5 3 0 -.00700 .32920 .86320 .15080 .56630 SI .12750 -.3 3 6 6 0 -.7 2 0 1 0 -.15440 .03720 .62290 10240 - .3 8 5 vo ,40510 -.3 4 8 3 0 P .50890 .72160 .58100 -.5 4 7 7 0 -.5 8 4 0 0 -.0 0 6 1 0 .27130 .78460 .37780 .46880 i:r MR FE eu NJ CU ZN AL HI p i;k 1.00000 MR .59900 i .00000 FF. .HI 070 .66670 1.00000 i:u .46740 .37580 .48900 1.00000 Rl .74800 .59300 .73410 .49870 1.00000 I'll .40770 .44530 .59170 .42830 .51610 1.00000 ZN .50460 .39210 .60010 .42230 .41650 .36920 1 .*>0000 AL .61990 .53530 .87050 .38230 .58450 .49910 .58900 1.00000 SI -.5 2 5 1 0 -.46660 -.66810 -.2 5 9 7 0 -.5 5 1 7 0 -.3 5 9 7 0 -..96 160 -.3 809 0 l . uOOOO P .50920 .55810 .69560 .31000 .46300 .39180 . 49010 .76050 -.26860 1.00000 TO»NDZr.T*r5 t to nznn »ncN- I ► * I I ...... u s u O'A l i © x ® tr e x U to MU N tJt0X©t0N.'XL*‘->L7!0'0' MMO Nl X O' © tC O' 0 to x c UMrf\IUCw>-CtO^HHMI>'OS>UO NJ U O' ©• © ► * O' © (* C f t 0- ©X © ► - O WOO>CM^ODM>MOC © ffitOOXtruXX© h a MUUO I ► » 1 I 1 ** VAUUUIIU^O X NJ ►* < l f 01 t« X v X IP U Nl U t0 O' O' O' © IFCOViii^UO Ut X fJX o o» o £ A X £•«&*<& O 01 © cr o V ^ < C h <0 I I ** M ^ t t U t t v U C rr. O' O NT L» i f N? A a t O' Si X © © *■ * U! X t 0 c D w x x © © © © © WU *- *-* M O ^K U V U ^U G t0 X © C O' > 0 S' >- Ji O V c O' © • * in W ^ C OLf -cr:0'CD'i>t>-M&(>-UNj\)t>':;c ccocooco occccooccoooocccoo M ! » 1 1 1 i l l i i l I 1 1 ! n u . to X Ln c O O' tO Nl V W W H (,« fj * x X O' © W X © 3> »•* cr 0 r- © t - O' tC Nl U O U 1C W fO © OI © Oi Nl X © u © O'Nj *.* c H* NJ © X M ^ >o H © t t *■* C>«1> its © © X u X X VI Oi ++ N2 o its x tc © o> t« © nj to w © © © © © © c © c © © O© © © © © © © o © © o o © c l f-k. 1 I 1 5 ‘ 1 1 1 1 1 1 1 1 i *■* 2 • • tfi t*- X > i t *r *-* O O X X A N) v. U’ O' ZX: N5 NJ ** © X X O b. i t N. £■ vj rz re w (A C it! < c M M N © u © •■* © X •b. «>* O tO w*- 0- f t O' c O I f x UM J> X c fti ■ */ \ ;r © NJ Nl It! ?v O; t o <1 9m* H — O' O' to © c c c o C c c © c c c c © © c c c C O © c c 1 ! 1 ! t 1 1 1 1—1 n X i t O A © c ♦— IK *-* M O ro © *- n: ic © © *- © 3> it. © w c © !>• 17! © tr w it; ni CTi Ai O' © M If. C l? > i O' ‘tt! ^ \ V X JO — \ J f O 0* b © N.' *lfc- mm. fw © < * *w O' V O' O' o X © c © c © © c © © o © © © o © o © © c © © © 1 1 H » « • • \ r* i t fw Zfi c 2 •-* *-* fc* *0 *+ *■ * w o >— ■ © < X c- SC v X b : tr. nit! — j » fC tr. c < c bl c o t r IS *c M tv * } • • • 3> •J n ! r - 03 — t r t r •J! O' J» X O ' c r - Oi © X © TO X n: d < to D !0 N . ♦ - © b l © 2? r . © i . X i t O ' t o V NJ V © < c . C' U ‘ J X © to © t n « w © t r I f . © C r T © © © © © © c © © © c c © * » ♦ XT. N! C i t i t to — © ► * © ► - © » - to © X tr © N! U C tr: t r * \ ’ C t r © ► " © x © © - c © u C X 0> *1 N l O ' © O' © X -bX O' W i t ! i t X 'COM o © © © c o © © C O © © C O c 1 h* • T * ♦ c © i t r- x to to t r to i x to tr o c i t © © O' O' © © Nl x t r © c X t *1 © m 142 4.1.3.1.2 Cluster Analysis 4.1.3.1.2.1 Box-Cox transformed data Cluster analysis was applied to the Box-Cox transformed data In conjunction with its robust correlation matrix (C.M.) (Table 4.2 b). The dendogram produced is presented in Fig. 4.30 a. It displays graphically the following elemental associations, for which tentative causes are presented : Ca - Sr : probably representing the biogenic carbonate phase of the sediments; no affinities with other elements are depicted. K - Be - Al : probably reflecting clay minerals. Zr - Ba - Si : probably representing detrital material which derives from felsic igneous rocks. Cr - Ni - V - P - Ti - Fe - Mg - Mn : probably representing material deriving from the erosion of the mafic/ultramafic formations of the mainland. Co - Cu - Zn : according to the low correlation coefficients (Table 4.2 b) of the Co - Cu, Co - Zn, and Cu - Zn pairs , it is suggested that the formation of this cluster is more fortuitous than real. La : lanthanum forms a class of its own, due to its low correlation coefficients with each one of the other elements determined. 4.1.3.1.2.2 Raw, untransformed data Cluster analysis was applied to the raw, untransformed data in conjunction with is robust C.M. (Table 4.2 a). The dendogram obtained is presented in Fig. 4.30 b. This dendogram is similar, overall, to that (Fig. 4.30 a) corresponding to the Box-Cox transformed data, suggesting the lack of a significant multivariate anomalous component in the Strymonikos data set. 143 COPMENETIC CORRELATION COEFFICIENT a 0.6445 ctiroDZCtK« ~ uj O z lu _j cc cr —• UtOUUNJUZ>(LI-lLi:ESC£DCENCO(/) 144 b COPHENETIC CORRELATION COEFFICIENT s 0-6609 c r o r iu —j —« z z O tu o z> q c cr o: cc « U(n£D(I!£:h-Q.MEEU.UU>U 2 JNmW O LU oc — Fig. 4.30 Cluster analysis of Strymonlkos group data; robust dendograms; (a) Box-Cox transformed data, (b) raw data. 145 4.1.3.1.3 Factor Analysis As the robust C.M. of the Strymonikos raw data is similar to the robust C.M. calculated on the basis of the Box-Cox transformed data (the similarity being shown also through the corresponding dendograms) - as far as elemental associations are concerned - factor analysis was applied to the raw data in conjunction with its robust C.M. The model with 4 factors (Table 4.3) was deemed appropriate to describe elemental associations on the basis of the following: 1. The Cattelf Scree test (Appendix V) shows that only factors characterized by an eigen value higher than 0.98 are significant; consequently, a 4 factor model would be a satisfactory solution (Table 4.4). 2. The four factors account for a high proportion, 76.3%, of the total data variability. Nevertheless, 5-, 6- and 7- factor models were tested for comparison purposes; however, they did not improve the factor solution. In this study, elements with factor loadings higher than 0.5, in absolute value, (Table 4.3) are considered to describe the composition of the relative factors. In the following, an interpretation of the elemental associations obtained from the 4-factor model is attempted. Factor 1: This factor accounts for 34% of the data variability. It comprises high positive Mg, Cr, Fe, Ni, Mn, Ti, V, Cu, Al and P loadings. Factor 1 possibly represents terrigenous detrital material brought in by rivers, mainly the Strymon river, and coastal erosion. Factor 2: This factor accounts for 24.7% of the data variability. It comprises high positive Ca and Sr loadings, possibly reflecting biogenic carbonates. Factor 2 also exhibits high negative K, Al, Be, Ti, P and Fe loadings, which represent terrigenous detrital material, and demonstrates the antipathetic behaviour of this material towards carbonates. Factor 3: This is a single element factor, having a high La loading. The absence of a La affinity for other elements was also demonstrated by cluster analysis. Factor 4: This is a single element factor, having a high negative Zr loading. Silicon 146 just fails to qualify as a member of this factor; it is characterized by a -0.43 loading. Table 4.3 : Factor analysis of raw data of the Strymonikos group. Four factor model: factor loadings and variance accounted for by components, computed on the basis of the robust C.M. of the raw data. Element Factor 1 Factor 2 Factor 3 Factor 4 K -.0 9 -.85 .36 -.03 Be .42 -.68 .47 -.04 Mg .86 -.31 .14 .14 Ca -.0 9 .93 .06 .20 Sr -.3 2 .88 .16 .09 Ba -.6 6 -.30 .46 -.25 La .12 -.01 .86 -.19 Ti .71 -.5 7 .13 -.29 Zr -.0 9 -.1 7 .14 -.92 V .57 -.40 .02 -.06 Cr .85 -.24 .04 -.06 Mn .77 -.09 .11 -.16 Fe .82 -.50 .16 .13 Co .49 -.09 .40 .07 Ni .80 -.18 .14 .01 Cu .53 -.20 .36 .12 Zn .44 -.38 .40 .05 Al .53 -.7 7 .28 -.02 Si -.7 6 -.11 -.09 -.43 to CO P .52 1 .21 -.37 Variance accounted for bv components Component Variance percent Cumulative variance 1 34.00 34.00 2 24.70 58.70 3 10.15 68.85 4 7.45 76.30 147 Table 4.4 : Strymonikos group raw data. List of eigen values of all the components of the data. Component Eigen Value Component Eiaen Value 1 9.54 11 .35 2 3.24 12 .20 3 1.45 13 .18 4 1.03 14 .13 5 .86 15 .09 6 .70 16 .09 7 .63 17 .06 8 .51 18 .05 9 .45 19 .03 10 .40 20 .02 148 4.1.3.2 Kavala group 4.1.3.2.1 Correlation Matrices The robust correlation matrices of the raw, untransformed and the Box-Cox transformed data are presented in Table 4.5 a and b, correspondingly. The lowest value of correlation coefficient (r) which is significantly different from zero, at the 0.001 probability level, is 0.162 (for the 148 samples of the Kavala group data set), (Fisher, 1963; cit. Howarth, 1983). A comparison of the two correlation matrices shows that there are 42 cases, which correspond to 22.1% of the total number of pairs, where the correlation coefficients have a difference ^ 0.1-0.2; in all the other cases, the difference does not exceed — 0.1. The two sets of coefficients do not exhibit large discrepancies; however, it was felt desirable to use both matrices in cluster analysis, in order to be completely sure that there are no significant differences between these two sets. T ab I e 4.5 Robust correlation matrices of the raw, untransformed (a) and the Box-Cox transformed (b) data of the Kavala group. K HE HQ CA SK bA LA ri ZR 0 K 1.00000 bt . 40750 1.00000 HQ -.31730 .50480 1.00000 CA -.8 3 1 0 0 -.6 1 2 3 0 -.027VO 1 .00000 SR 70510 -.64170 -.22170 .71710 1 .OOOOU 8A .54230 -.21760 -.67540 -.2 8 5 4 0 -.0 1 3 2 0 1 .OOGuO l.A -..17300 .03800 .07600 .37070 .42080 -.07730 1.OOOOO 71 . 14840 .85030 .79610 -.4 7 2 7 0 -.6 0 5 7 0 -.46480 .03660 l . OOwuO I k .34 ISO .66740 .34720 -.56240 -,5718o -.0 8 5 4 0 -.0 8 1 3 0 .c>8240 1 « UOl/OU U •123V0 .75430 .73100 -.40410 -.51150 -. 3770'./ . O 7.160 . 886' »0 .56160 1 .oocoo Cft .09640 .76250 .81540 -.40370 - .5 4 U 0 g -.4 7 4 10 , 0 l 450 .73400 .58740 .H4010 HN -.1 2 2 3 0 .53760 .76710 -.12350 '.31 87 v> -.4/570 . t 5780 ,75770 .40660 .67400 FE .06140 .80620 .82830 3764w -.5 3 4 8 0 .53770 .0555u .7 /3 6 0 .57030 .87550 CU. 12420 .28450 •26610 - . 1 76 70 -.7 0 3 /0 -.0 3 7 5 0 .05700 .26680 .31200 .34080 NI • G4UOO .62780 .70430 -.2 8 6 5 0 .42730 - .42650 ,13530 . 75770 .35280 .74620 Cll -. 72450 .17670 .33460 .07840 . 0383i/ -. 260 70 .25030 .25700 .05250 .23400 IH -.03080 .48180 .63670 -.17100 —«28250 -.3 3 1 1 0 .0/VVO .6 6 /7 0 . 42130 .62310 AL .5 0 /3 0 .75040 .47560 -. 72800 .73480 - . 1 1600 -.0 473 0 .86 170 .68320 .76810 SI ,7 V 100 -.0 0 3 8 0 -.6 3 0 3 0 -. 65300 .47130 .65550 -. 37540 -. 2V4H0 .13530 -.2 7 4 6 0 P .08560 .71220 •63600 -.37680 -.4 7 2 6 0 -. 36800 .00710 .77170 .71850 .69850 CK HN Hr. «;u HI CU 78 hL SI 8 CR 1.00000 HN .6VVU0 1 .oo oo o Ft .73610 .//HUO 1.00000 CO .3V >70 .37000 .35770 1.00000 Ml . 83060 .61410 .78480 ,77H20 i . 6GO0' t:u . 25040 .30340 .27130 . O30o0 .35660 1 .OOU00 ZH .60240 .51720 .67820 , 17250 .64730 .21 5.. 0 1 . GOooo M. .76200 .51700 . 80400 .30280 .61170 . 1 37.5/ . 4 V'/2w 1 . v/OOuo SI-. 30060 -.42320 -.3 6 7 3 0 -.0 1 7 4 0 - .2 /4 2 0 - .61880 -.38620 .07000 1.OOOOO P .75030 .60740 .74220 .25430 .53450 . 233 Ml .53450 .68410 -. 17740 1.00000 149 K RE MG CA GU BA LA 1 1 ZK 0 K 1.00000 BE .58000 1.00000 MG -.2 4 2 1 0 .43280 1.00000 CA -.7 5 8 7 0 -.7 1 6 1 0 -.0 5 4 9 0 1.00000 SR -.68400 -.76880 -.3 2 5 3 0 .09370 1.00000 bA .56200 -.1 0 9 7 0 -.74050 -.2 7 7 2 0 -.0 0 4 3 0 1 . OGoOO LA -.2 5 1 9 0 -.0 6 9 6 0 -.0 8 5 8 0 .28540 .3 3 /9 0 .03400 1.00000 TI .29290 .82430 .74570 -.5 7 2 1 0 -.73940 -.3/800 -.04330 l.ooooo • ZR .49240 .62700 .21480 -.67140 - . 6 8 / 2 0 .10770 -.13900 .61950 1.00000 V .26320 .73420 .67600 -.4 8 1 6 0 -.6 3 4 1 0 -.3 3 1 6 0 -.03420 .8/870 .51560 1.OOOOu CR .15430 .71000 .83020 -.42530 -.66980 -.5 1 9 5 0 -.08870 .93290 .4/610 .82140 MN .06750 .54620 .72840 -.2 2 2 8 0 44810 -.4 1 8 0 0 ,05600 ,75690 .31030 . 662*10 EE .1 9 /4 0 .77610 .80890 -.44010 -.6 6 0 0 0 -.4 9 3 2 0 -.06630 .96330 .51000 .85430 CO .17190 .32900 .25290 -.21610 -.29520 - . 0 8 2 / 0 -.01980 . 8/020 ,31730 .35040 Nl .14870 .60970 .69800 -.33020 -.54/90 .44240 ,01160 .76610 .28330 . /4530 nu -.2 1 7 6 0 .15550 .38060 ..08050 -.00370 -.37760 .12380 .26140 .01360 .23700 ZN .07700 .47470 .66740 -.19560 -.3 3 8 1 0 -.3 2 8 6 0 -.03020 .65430 .75880 .59460 Al .61270 .95120 .41480 -.76990 .80 1/0 ■■ .0 1530 -.08190 .86010 ,6f. 170 .76680 SI .77130 .11250 -.66880 -.5 7 9 4 0 -.3 8 2 4 0 . / 4 A 2 0 -.21100 -.1 9 /4 0 .292/0 -.20070 P .20660 .71350 .60330 -.5 5 1 0 0 -.6 5 6 4 0 •.31200 •.12220 .79830 .62890 .70930 CR MN PE CfJ Nl CU ZN Al SI P CR 1.00000 MN .72450 1.00000 • FE .94170 .77830 % 1.00000 CO . 35530 .42660 .35820 1.00000 NI .83240 .62380 V; .79250 .27560 1.00000 CIJ .29320 .30690 v ; .28320 .06840 .42230 1.00000 ZN .69050 .49740 .68130 .10180 .67950 .28510 1,00000 AL .71430 .55120 .79270 .32760 ♦ 60660 .12080 .49850 1.00000 SI -.3 2 1 4 0 -.33040 -.31890 .00270 -.2 4 9 1 0 -.3 *4 6 0 - . 332.70 .16220 1.00000 P .75330 ..37500 .73010 .30010 .56060 .'26 180 . 31750 .69030 -.11030 1.00000 b 0 5 1 151 4.1.3.2.2 Cluster Analysis 4.1.3.2.2.1 Box-cox transformed data Cluster analysis was applied to the Box-Cox transformed data in conjunction with its robust C.M. The dendogram obtained is illustrated in Fig. 4.31 a. It displays graphically the following elemental associations, for which tentative causes are presented: K - Si - Ba : probably representing detrital material deriving from felsic rocks. Ca - Sr : probably reflecting biogenic carbonates; no affinities with other elements are depicted. Cr - Ni - Zn - Mg - Mn : probably representing weathering products of mafic/ultramafic rocks. Be - Al - Zr - V - P - Ti - Fe : this is likely to represent mainly clay minerals. The Fe - Ti subbranch may reflect traces of Fe,Ti-oxides; Zr may represent traces of the heavy mineral zircon. C o : cobalt displays low correlation coefficients with each of the other elements determined and it, therefore, forms a class of its own. La - Cu : The low correlation coefficient of the La - Cu pair suggests that this cluster is more fortuitous than real. 4.1.3.2.2.2 Raw, untransformed data Cluster analysis was also applied to the raw, untransformed data in conjunction with its robust C.M. The dendogram produced is presented in Fig. 4.31 b. It exhibits significant similarities with that corresponding to the Box-Cox transformed data. The mafic/ultramafic component is reflected through the Ni, Zn, Mg, Mn, V, Cr, Ti and Fe association. Comparing the dendograms illustrated in Fig. 4.31 a and b, it is suggested that the Kavala data set is not likely to include a multivariate anomalous component. 152 a COPHENETIC CORRELATION COEFFICIENT = 0.6323 —-era: — z o z uj _j or «-• in o a d cr tt: x:(n(D uzN Ercoa:N >cLi-u.u_iuo(n UJ _» > Q- *” 153 b COPHENETIC CORRELATION COEFFICIENT = 0.77B3 - d lu _i o; o —• 2: O z: ct —• iu cr id cr ce bd(OCQCQ(INa.U2Ni:E>UHlj.DUU(n —• z o z a: UJ Z M Z Z <-J U. Fig. 4.31 Cluster analysis of Kavala group data; robust dendograms (a) Box-Cox transformed data, (b) raw data. 154 4.1.3.2.3 Factor Analysis As the robust correlation matrices of the Kavala raw and Box-Cox transformed data sets are significantly similar, as far as elemental associations are concerned, factor analysis was applied to the raw data in conjunction with its robust C.M. The 4-factor model (Table 4.6) was deemed appropriate to describe elemental associations on the basis of the following: 1. The Cattell Scree test (Appendix V) suggests that only factors characterized by an eigen value higher than 0.9 are significant; consequently, a 4-factor model is likely to be a satisfactory solution (Table 4.7). 2. The four factors account for a very high proportion, 81%, of the total data variability. In spite of that, 5-, 6- and 7- factor models were tested for comparison purposes; however, they did not improve the factor solution. In this study, elements with factor loadings higher than 0.5, in absolute value, (Table 4.6) are considered to describe the composition of the relative factors. In the following, an interpretation of the elemental associations obtained from the 4-factor model is attempted. Factor 1: This factor accounts for 47% of the data variability. It comprises high positive Ti, Fe, Cr, Mg, V, Ni, P, Be, Al, Mn, Zn and Zr loadings. Factor 1 represents terrigenous detrital material. Factor 1 has significant negative Ba and Sr loadings; they reflect the Ba and Sr antipathy for the elements with high positive factor 1 loadings. Factor 2: This factor accounts for 21% of the data variability. It is characterized by high positive Ca and Sr loadings, reflecting carbonates. The elements K, Si, Ba and Al exhibit high negative factor 2 loadings, probably, representing material derived from felsic rocks and demonstrating the antipathy of carbonates for felsic detrital material. Factor 3: This factor accounts for 6.8% of the data variability, which is much less than that accounted for by factors 1 and 2. Factor 3 exhibits only La and Cu high loadings. The low correlation coefficient of the La-Cu pair (Table 4.5 a) suggests that 155 a La-Cu association is not likely to be real. Factor 4: This is a single element factor characterized by a high Co loading. Cobalt forms also a class of its own in the cluster analysis. Table 4.6 : Factor analysis of raw data of the Kavala group. Four factor model : factor loadings and variance accounted for by components, computed on the basis of the robust C.M. of the raw data. Element Factor 1 Factor 2 Factor 3 Factor 4 K .04 -.9 4 -.14 -.05 Be .80 -.4 2 .13 -.08 Mg .87 .39 -.04 -.04 Ca -.3 8 .87 .21 .00 Sr -.5 5 .70 .32 -.02 Ba -.5 8 -.6 3 .22 -.12 La .01 .27 .85 -.18 Ti .97 -.1 2 .02 -.12 Zr .61 -.4 2 -.0 4 . -.25 V .89 -.0 9 .07 -.11 Cr .95 -.04 -.02 -.12 Mn .78 .18 .09 -.23 Fe .97 -.01 .02 -.11 Co .29 -.0 6 .04 -.86 Ni .82 -.01 .17 .02 Cu .33 .15 .56 .33 Zn .73 .08 .08 .15 Al .79 -.5 2 .07 -.08 Si -.3 7 -.8 5 -.21 -.0 2 P .81 -.1 3 .01 -.0 8 Variance accounted for bv components Component Variance percent Cumulative variance 1 47.66 47.66 2 20.99 68.65 3 6.83 75.48 4 5.56 81.04 156 Table 4.7 : Kavala group raw data. List of eigen values of all the components of the data. .Com ponent Eigen .Value. C om ponent Eigen Value 1 9.78 11 .21 2 4.32 12 .18 3 1.15 13 .15 4 .96 14 .10 5 .77 15 .08 6 .73 16 .06 7 .49 17 .04 8 .37 18 .03 9 .32 19 .02 10 .22 20 .01 4.1.3.3 Comparison between the Strvmonikos and Kavala groups Comparing the results .obtained from the Strymonikos and Kavala groups statistical treatment, it is suggested that the sediments encountered in both areas are very similar from a geochemical point of view. 157 4.2 Buried sediments Twelve cores were studied from the area lying west of Thassos island (Fig. 4.1). 4,2J__ Description of the cores The buried sediments consist, in general, of silt and silty sand, with occasional local occurrences of fine gravel, coarse sand and mud grade material. They have an abundance of shells and shell fragments. In some of the cores studied, plant debris occurs. The Strymonikos triangle subsurface sediments contain some coral fragments. In a number of cores, mottles occur. The colours of the buried sediments - described using the Munsell Soil Colour Charts (Munsell Colour Company, 1971) - vary broadly from moderate yellowish brown to greyish black (Moorby, pers.comm., 1985). 4.2.1.1 Core STR 1 fFia. 4.32 a^ This core consists of sediment with a fine silt to mud grain size. There is a thin void at the top. Mottling occurs in the upper 40 cm. The colour varies with depth. The sediment from 2 to 55 cm depth is moderate yellowish brown (10YR 5/4) in colour, and that lying between 55 and 85 cm is light olive grey (5Y 5/2). In the lower part of this core (from 85 to 121 cm depth), olive grey to dark greenish grey (5GY 4/1 to 5Y 4/1) sediment occurs. 4.2.1.2 Core STR 2 (Fid. 4.33 a) This core consists of mud grade material. There is a clear change of the sediment colour downcore. At the top, the sediment is moderate yellowish brown (10YR 5/4) in colour. Around 20 cm depth, the colour gradually changes to light olive grey (5Y 5/2). Below 70 cm, it gradually becomes olive grey (5Y 4/1) and dark greenish grey (5GY 4/1). Mottling, moderate yellowish brown (10YR 5/4), occurs at around 50 cm depth. Dark grey (N3) tiny mottles are frequent in the lower 30 cm of the core. 4.2.1.3 Core STR 4_(Fia,_ A34 .al This core consists of silty mud grade material, with thin voids at both its top and base. Between 2.5 and 20 cm depth, there is a dark yellowish brown (10YR 4/2) 158 sediment, with a thin layer of light brown (5YR 5/6) sediment at the surface. From 20 to 48 cm, the sediment colour varies from dark yellowish brown (10YR 4/2) to olive grey (5Y 4/1), dark grey (N3) and greyish black (N2); olive grey is, however, the predominant colour. From 48 to 120 cm depth, the sediment is dark yellowish brown (10YR 4/2) in colour. From 74 to 76 cm , three silty bands - each only 2 to 3 mm thick - occur, at around 74.5, 75.0 and 75.5 cm respectively, they are black (N1) in colour. The sediment from 20 to 48 cm depth is extensively mottled. Also, two areas of mottling occur between 69 and 85 cm depth; these mottlings have an olive grey (5Y 4/1) to dark grey (N3) colour. Finally, mottling occurs in the lower part of the core, from 105 to 118 cm depth. A smell of H2S was noticeable when the core was opened. 4.2.1.4 Core STR 11 (Fig. 4.35 a) This core consists of sediment with a fine silt to silty mud grain size. There is a thin void at the top. The colour of the sediment changes gradually from moderate yellowish brown (10YR 5/4) in the uppermost part, to light olive grey (5Y 5/2) at about 30 cm. Then, it becomes gradually darker and by 1m depth, the colour is olive grey (5Y 3/2). Some mixing of sediment occurs along the upper part of one side of the core, probably caused by friction on core penetration. Similar mixing, with a moderate yellowish brown (10YR 5/4) sediment, is visible along the lower part of the opposite than the aforementioned side of the core. Throughout the core, shell fragments are common; some of them are up to 1 cm in size. 4.2.1.5 Core STR 12 /Flo. 4.36 a) The upper part (9 cm) of this short core consists of silty sand grade material, with small shell fragments present; the sediment colour is moderate to dark yellowish brown (10YR 5/4 to 10YR 4/2). Between 10 and 32 cm depth , a dark greenish grey (5GY 4/1) to greenish black (5GY 2/1) fine gravel grade material occurs, with an abundance of shell fragments, up to 1 cm in size. From 9 to 10 cm depth, a fairly sharp transition occurs. Between 32 and 33 cm, light olive grey (5Y 5/2), very coarse sand grade material with abundant shell fragments is present. A thin void occurs at the base of the core. Along the one side of the core, some mixing of sediment has occurred. 159 4.2.1 Core STR 27 (Fig. 4.37 al This is a relatively long core consisting of silty mud grade material. Shell fragments occur throughout it. At the top, a void occurs. A thin (2 mm thick) surface oxidized layer, moderate yellowish brown (10YR 5/4) in colour, is present. In the remaining part of the core, the sediment colour is light olive grey (5Y 5/2), changing gradually at about 60 cm to olive grey (5Y 4/1) and dark greenish grey (5GY 4/1). Some mixing of sediment has occurred along the one side of the core. 4.2.1.7 Core STR 28 (Fig. .4.38 al This core consists of silt grade material with abundant shell fragments. From 46 to 50 cm depth, a slightly coarser silt grade material with many shell fragments occurs. Mica flakes are common throughout the core. Near the top, plant debris occurs. Mottling is visible at about 40, 60 and 70 cm depth. The colour of the sediment is olive grey (5Y 5/2); however, it grades to olive brown (5Y 4/4) and olive grey (5Y 3/2) in the uppermost 20 cm, lowermost 10 cm, as well as in the mottles. A thin layer of moderate yellowish brown (10YR 5/4) sediment occurs at the surface of the core. 4.2.1.8 Core KB 3 (Fig. 4.39 a) The upper 120 cm of this core consist of silty mud grade material. From 120 to 133 cm depth, the sediment becomes increasingly coarse, and at the base of the core silt grade material occurs. A void is noticed at the top. The colour of the sediment is moderate yellowish brown (10YR 5/4) in the upper 40 cm. Below it, the colour changes to light olive grey (5Y 5/2) till 110 cm depth, while from 110 cm to the base of the core, the sediment is olive grey (5Y 4/1) in colour. Mottling is visible , in particular between 95 and 115 cm depth. Large conical shells occur at about 50 and 100 cm depth; echinoderm fragments are noticed close to 10 cm depth. Finally, near the base of the core, shell fragments are common. 4.2.1.9 Core KB 5 fFia. 4. 40 a) This core consists of sediment with a fine silt to silty mud grain size. At the surface and at the bottom of the core voids occur. There is a marked colour change from moderate yellowish brown (10YR 5/4) in the upper part, to light olive grey (5Y 5/2) at about 20 and 30 cm depth. Below 90 cm down to the base of the core, some 160 streaks of moderate yellowish brown (10YR 5/4) silt are noticed. At about 85 cm depth, a large conical shell fragment occurs. 4,2.1.10. Core KB 7 (Fig. 4.41 a) There is a 2 cm void at the top and a thinner one at the base of the core. The texture and the colour of the sediment vary with depth. Thus, from 2 to 117 cm depth, there is very fine silt grade material; the colour varies from moderate to dark yellowish brown (10YR 5/4 to 10YR 4/2), with areas of greyish olive (10Y 4/2). Between 117 and 130 cm depth, dark greenish grey (5GY 4/1) silt grade material occurs; shell fragments of gravel size are abundant, and a few large fragments, up to 3 cm long, also occur. These shells are white to yellowish grey (5Y 8/1) in colour. In the lower part of the core, from 130 to 147 cm depth, stiff calcareous mud occurs. Also, some coarse gravel sized shell fragments, as well as a 3 cm sized fragment of hard calcareous rock, were found. In the lower part of the core, the colour of the sediment varies from very light grey (N8) to light olive grey (5Y 6/1). 4.2.1.11 Core KB 11 _(Fia,_4. 42 a) The texture and colour of the sediment vary with depth. In the upper 22 cm of this core, the sediment is very coarse sand to fine gravel, with many shells (some up to 1 cm in size) and shell fragments. The colour in this top layer varies from light olive grey (5Y 5/2) to dark greenish grey (5GY 4/1). Between 22 and 57 cm depth, olive grey (5Y 4/1), medium dark grey (N4) to dark grey (N3) silt grade material, with small shell fragments, occurs. The shell fragments decrease in abundance towards the base of this core. From 57.0 to 62.5 cm depth, mottles of moderate yellowish brown (10YR 5/4) colour occur. In the lower 0.5 cm of the core, the sediment has a moderate yellowish brown (10YR 5/4) colour. 4.2.1.12 Core KB 12__(Fiq^_4,4_3,..ai This core consists of silt grade material. At the top, a 3 cm void occurs. The colour varies with depth. Thus, the sediment from 3 to 30 cm depth is dark yellowish brown (10YR 4/2); in the depth range 30 - 64 cm, the colour changes from yellowish brown (10 YR 4/2) to olive grey (5Y 4/1); in the lower part of the core, olive grey (5Y 4/1) to dark grey (N3) sediment occurs. Coral fragments are abundant at depths between 3 and 64 cm, becoming increasingly common downwards. Those occurring from 30 to 64 cm depth are quite large, being up to 4 cm long and 161 0.5cm in diameter. Below 64 cm, coral fragments are present occasionally, with a large clump of coral found at a depth of 78 - 82 cm. In addition, large shells along with coral fragments occur at depths of 98 to 101 cm. 4J2.2__ Bulk, geochemistry of .subsurface offshore sediments In this section, the variations in the chemical composition of the subsurface sediments are studied. To facilitate this, the chemical data have been plotted for each of the cores. 4.2.2.1 Core STR 1__ (EfcL_.4«32..-b). The Ca and Sr contents increase gradually with depth in an inverse relation to those of Al, Fe and P, showing the close association of the latter group of elements with the non-carbonate fraction of the sediments. Beryllium follows Al in its downcore distribution . The contents of Cu and Zn decrease with depth in the uppermost 50 cm ; below 50 cm, Zn concentrations are low while Cu exhibits an uneven distribution. In the upper 20 cm, Mn decreases sharply with depth; below it, the Mn content is low. Chromium and Mg follow similar uneven downcore distribution patterns. 4.2.2.2 Core STR 2 (Fia. 4. 33 b\ The downcore distribution patterns of .Mn and Fe are similar; both elements decrease in concentration with depth. Additionally, the contents of Ti, V, Al and P decrease overall with depth. Beryllium follows K in its downcore distribution. The concentrations of K and Be decrease in the upper 20 cm and in the lowermost part of the core, while they increase with increasing depth from 20 down to 70 cm. The downcore variations of Mg and Ca are similar; Ca and Mg increase gradually in the uppermost 70 cm. Strontium increases in concentration with increasing depth from 20 to 70 cm. Zirconium increases, overall, with depth. Zinc reaches its highest value at the base of the core. The upper 20 cm of the core are enriched in Mn, Fe, Al, P, V and Ti and deficient in Ca. Conversely, the lower part of the core is enriched in Ca while it is poor in Fe, Mn, Al, Ti, P and V. 4.2.2.3 Core STR 4^_(Fio. 4,34 b l This core, collected off the Strymon river mouth (Fig. 4.1), appears to be, in general, richer than the other cores of the Strymonikos gulf in most of the metals determined. 162 The elements Al, Fe, Mg, Cr and V have almost Identical downcore distribution patterns. Their concentrations are low at 75 cm, in the rest of the core they do not fluctuate significantly. At 75 cm a depletion in Ni and Cu occurs, too. Zirconium follows Si in its downcore distribution. The distribution patterns of Al and Si are quite similar throughout the core, apart from the 75 cm depth; there, high Si and Zr contents occur, corresponding to a deficiency in Al. Although the Zn content is relatively low throughout the core, an enrichment occurs at 66 cm. Strontium distribution pattern has similarities with that of Ba; the trend of Ba is more pronounced. The downcore variations of Ca and Sr display differences, despite of the similar distribution patterns of these two elements in the surface sediments and most of the other cores. Calcium content increases with depth in the upper 66 cm, then it decreases gradually downwards. Strontium decreases in the upper 66 cm of the core, it has its highest value at 75 cm and then it decreases downwards. Manganese increases with depth in the uppermost 36 cm of the core; at 36 cm it reaches its highest content (1,410 pg/g) while below it , it decreases dramatically downwards. 4.2.2.4 _ Core STR 11 (Fid. 4.35 bl The downcore distribution patterns of the elements Fe, Mn, Cr, Mg, Ti, Al, Co and Ni are similar. Their trend of variation is the reverse of the one exhibited by Ca and Sr. The concentrations of the former group of elements tend to increase overall in the upper 70 cm of the core, and then they decrease, gradually, downwards. Thus, at 70 cm depth, a deficiency in the carbonate fraction of the sediments occurs while an enrichment in the former group of elements is noticed. Zirconium follows Si in its downcore variation. The Zn content is low in the upper 100 cm of the core, while in the lower 30 cm it increases dramatically. 4.2.2.5 Core STR 12 (Fio. 4.36 b) The elements Be, Mg, La, Ti, Zr, V, Cr, Mn, Fe, Ni, Cu, Zn, Al and P exhibit an inverse relation to Ca and Sr, in this very short core. The concentrations of the former group of elements decrease with increasing depth while the contents of the latter increase, in parallel with an increase in shell fragments abundance. 4.2.2.6 Core STR 27 (Fig,^,37_bl The downcore distributions of the elements K, Fe, AI,Ti and Be are overall similar. Their patterns are, generally, inversely related to those of Ca and Sr. Between 30 and 163 125 cm depth, the Al, Fe, K, Ti and Be contents decrease gradually with depth while those of Ca and Sr increase. The concentrations of Cr, V, Mg and P tend to decrease, overall, towards the base of the core. The concentration of Zn is low throughout the core, apart from the sediment collected from 30 cm depth. A relative Mn enrichment occurs at 125 cm depth. The downcore Cu and Co distributions are rather variable; however, they tend to decrease overall with increasing depth. 4.2.2.7 Core STR 28 (Fin. 4.38 hi In the upper 20 cm of the core, the concentrations of the elements K, Al, Si, Mg, Fe, Ti, Zr, Mn, Cu and Be increase with depth while the Ca content decreases. Below it, Al, Mn, Be and Ca do not fluctuate significantly in concentration; the values of the former three elements are relatively high while those of Ca are low. The concentrations of Cr, Fe, Ti, V, Ni and Zn increase downwards between 20 and 65 cm depth. The variation of zirconium follows closely that of Si. The downcore distribution trends of Sr and Ba are almost identical; although in most of the surface and subsurface sediments studied Sr follows the variation of Ca, this does not occur in this core. 4.2.2.8 Core KB 3 (Fla. 4.39 b) The elements Al, Si, K, Mg, Fe, Ti, Cr, Mn, P, V and Be display similar patterns of distribution; they are enriched, without fluctuating significantly in concentration, in the upper 110 cm. Below 110 cm, their values decrease dramatically downwards. Calcium and strontium doncore variations show a deficiency of both elements in the upper 80 cm while from 80 to 110 cm their contents increase significantly. In the lowermost part of the core, a Ca and Sr enrichment occurs associated with presence of shell fragments. The Zn content is low throughout the core, apart from an enrichment at 21 cm. 4.2.2.9 Core KB 5 (Fig. 4.40 b) The elements Fe, Al, Ti and Be display similar downcore distribution patterns; their contents increase with depth in the upper 50 cm, they decrease from 50 to 80 cm and then, they increase downwards. Potassium and Mn tend to increase overall downwards. Calcium content increases with depth in the uppermost 20 cm; below it, the downcore Ca variation is inversely related to that exhibited by Fe, Al, Ti and Be. 164 4.2.2.10 Core KB 7 (Fig. 4.41 b) The elements Al, Si, K, Ti, V and Fe have similar downcore distribution patterns; they decrease in the upper 20 cm, increase in the following 30 cm, decrease gradually downwards from 50 to 80 cm, and exhibit their lowest concentrations at the base of the core. Their general trend of variation is the inverse of that displayed by Ca and Sr. Manganese and Mg follow, overall, the general trend displayed by Al, Si, K, Ti,V and Fe, apart from a significant enrichment in the lower part of the core. Between 111 and 123 cm depth, Ca and Sr contents increase by 93% and 81%, respectively, in association with an abundance of shell fragments. Close to the base of the core, a high Ca content occurs; its value is almost six times higher than that corresponding to the top of the core. This is related to the presence of calcareous mud in the lowermost part of the core. 4.2.2.11 Core KB 11 fFio. 4.42 b) The downcore distribution patterns of K, Al, Si, Ba, Zr and Ni are similar. Their concentrations increase with depth in the upper 40 cm while below it, they do not fluctuate significantly. The contents of Fe, Ti, V, Cr and Be increase gradually with depth. Calcium increases in the upper 16 cm; below it, Ca decreases gradually with depth, in parallel with a decrease in the shell fragments abundance. Zinc follows Mg in its downcore distribution. The Mn content increases with depth in the uppermost 40 cm while it decreases in the lower part of the core. 4.2.2.12 Core KB 12 (Fia. 4.43 b) The elements K, Al, Si, Ti, Cr, Mn, Fe, Mg, Ba, Be and V have similar overall downcore distribution patterns; their concentrations decrease in the upper 20 cm while below this depth, they increase gradually downwards. Calcium displays the inverse distribution pattern to that corresponding to the aforementioned group of elements, as far as the upper 70 cm of the core are concerned. Below 70 cm depth, Ca increases in concentration. Zirconium increases in concentration downwards the core. 165 STRYMGN 1 a v b Fig. 4.32 CORE STR 1 : (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr,Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P (depth in cm, concentrations in iig/g). v : void, m : mottles. 166 STRYMGN 2 Fig. 4.33 CORE STR 2: (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr,Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P (depth in cm, concentrations in pg/g). v : void, m : mottles. 167 STRYMON 4 0 j z z z ^ - 5YR 5/6 ,10YR « - 20 ' T i KOYR 4/2, > 5Y4/1, - 4 0 ' N 3, N2 - 60' ^5YR5/6 TI ZR V . m 10YR 4/2 x 'N l ' - -80 ’5Y4/1 f0N3 -100 ™ f o n r h cm a QR MN FE CO NI CU ZN RL SI P b Fig. 4.34 CORE STR 4: (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr,Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P (depth in cm, concentrations in pg/g). v : void, m : mottles. 1 6 8 STRYMGN ll O vW z JSz Z Z Z Z T 1 ^ - 0YR5/4\0YR5/4*. c r - 2 0 ^* ° ! o • 5Y5/2 jl . _ Oo 0 , - 40 * / „sfSf o' 0\ - 6 0 o * o -80 o a o _100 ‘5Y 5Y3/2 3/2 0 -120 U cm ___o_ c K)YR 5/4 a b Fig. 4.35 CORE STR 11 : (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr.Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P (depth in cm, concentrations in pg/g). v : void, c : contamination, sf : shell fragments. 169 STRYMGN 12 — ° l «t o m . i ! - . 5GYVI to - 2 0 ; - vs f cm ,I o o 5GY2/I — 5 Y 5/2 a .40. H------, i t i------1 i------1 ------30 . 70 <00 350 <1000 30000 tO ZVW. 80 14 • 4CT10' 80 50000 70000 100000 300000 100 400 . CR MN FE CO NI cu ZN RL SI b Fig. 4.36 CORESTR 12 : (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr.Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P (depth in cm concentrations in p.g/g). sf : shell fragments, v : void, c : contamination. 170 STRYMON 27 a Fig. 4.37 CORE STR 27: (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr,Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P (depth in cm, concentrations in |ig/g). v : void, o : oxidizing layer, s f : shell fragments, c : contamination. 171 STRYMGN 28 '\ TI ZR V 78000 87000 230000 260000 680. flL S I P b Fig. 4.38 CORE STR 28 : (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr,Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P (depth in cm, concentrations in pg/g). pd : plant debris, m : mottles, sf : shell fragments. 172 KRVRLA 3 0 UZZZZ& JL -20 10Yfi 5 A - 4 0 - - 60 - 5 Y 5/2 - 80 - -100 ■m 5Y4/1 - 120-1 o a o cm °0 | ° O P» Sf b Fig. 4.39 CORE KB 3 : (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr,Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P (depth in cm, concentrations in p.g/g). v : void, ef : echinoderm fragments, s : shells, sf : shell fragments, m : mottles. 173 KRVRLA 5 0 /77771- v 10YR Sjlf -20 -4 0 5Y 5/2 -60 -80 A-st -100 10YR 5/; •120 cm 77777# a Fig. 4.40 CORE KB 5: (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr,Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P (depth in cm, concentrations in pg/g). v : void, sf : shell fragments. 174 KRVRLR 7 0 \10 - 20 - KJY4/2'f* -M- 10YR -6 0 5/4 \ jlOYR ! 4/2 -80- •• -100 . 10Y4/2 8 f — —120 *5GY4/1 .f ^ N8 - u o cm »5Ytf1° Fig. 4.41 CORE KB 7: (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr.Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, At, Si and P (depth in cm, concentrations in p g /g ). v : void, sf : shell fragments. 175 KRVFILR 11 bv Fig. 4.42 CORE KB 11 : (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr.Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P (depth in cm, concentrations in pg/g). s: shells, s f : shell fragments, m : mottles. 176 KRVRLR 12 0. n 0 -2 Cl ef ^ -4 (> -60- -8 0 -100- cm Fig. 4.43 CORE KB 12: (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr.Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Su and P (depth in cm, concentrations in pg/g). c f : coral fragments, v : void, s ; shells. 177 4.3 Summary The part of the North Aegean Sea that lies to the west of Thassos island is mainly covered by fine grained sediments, such as silty clays and clayey silts. However, sands are present along the shoreline, at the outer parts of the gulfs and just landwards of the shelf break. The regional variation of Si in the surface sediments implies that a large part of the sand-sized material consists of silicates. The distribution of Ca is different to that of Si; to a large extent, Ca enrichments correspond to a high carbonate content. Particularly high Ca values were determined around Thassos, in the channel between Thassos and the mainland, in the Strymonikos triangle, and in the Strymonikos ridge. The regional distributions of Al and Fe are similar; an enrichment was determined in the Strymonikos gulf surface sediments (apart from a narrow coastal zone and the north-east trending ridge), extending over much of the Strymonikos plateau. Iron and Al rich fine grained sediments are present in the gulf of Kavala. Although a low heavy mineral content (determined in the coarse fraction) occurs in the northern part of Kavala gulf, enrichments in Ti, Cr and Fe were determined in surface sediments with a high silt content. A K and Sr enrichment occurs in the nearshore sediments lying between the Strymon river mouth and Nea Peramos bay. A Ba enrichment occurs off the west coast of Thassos island. Surface sediments collected from Limenaria bay are rich in Fe, Mn and Ni. A Mn and Zn enrichment occurs in the Olympias bay sediments. A high La content was determined in coastal surface sediments collected eastwards of Loutra Eleftheron. Statistical treatment of the data, concerning the elemental composition of the offshore surface sediments, suggests that the Strymonikos and Kavala groups of samples consist of sediments which are very similar from a geochemical point of view. Cluster and factor analysis confirm that the offshore sediments are composed of terrigenous detrital material and biogenic carbonates. Both cluster and factor analysis applied to raw, untransformed data show that anomalous metal enrichments are not likely to be present. The subsurface sediments mainly consist of silt and silty sand grade material, with abundant shell fragments. The downcore chemical variations illustrate the antipathetic behaviour of elements deriving from clay minerals towards those from carbonate particles. An upward migration of Mn was noticed in some of the cores studied. 178 CHAPTER V THE lERISSOS GULF 5.1 Surface sediments A total of 83 samples of offshore surface sediments were collected from the 116 km2 area of lerissos gulf; Figure 5.1 shows the samples sites which are, fairly, evenly distributed throughout the gulf. In addition to the offshore samples, 19 onshore samples from the coast of the adjacent mainland (Fig. 5.2) have been collected and chemically analyzed. The determination of their elemental content was significant for the interpretation of the offshore elemental distribution. The lerissos gulf is considered separately from the rest of the study area lying to the west of Thassos island because of its bathymetric isolation from the Strymonikos plateau by a sill at the mouth of the gulf. 5,.1-J__ Sediment composition ___Pistribution_of_ rock_-fraqments. quartz and heavy minerals The maps that present the distribution of rock fragments, quartz and heavy minerals in the surface sediments of the lerissos gulf are provided by Perissoratis (Perissoratis et al, in press). 5.1.1.1.1 Rock fragments The distribution of rock fragments in the coarse fraction (Fig. 5.3) shows that high abundances occur in the nearshore sediments, especially in those lying between Stratoni and lerissos, with maximum rock fragments content in the sediments off lerissos district. By contrast, in the central part of the gulf low values of rock fragments occur. 179 5.1.1.1.2 Quartz The distribution of quartz in the coarse fraction (Fig. 5.4) has close similarities with that of rock fragments. Thus, quartz is abundant in the nearshore sediments and diminishes in the central part of the gulf. 5.1.1.1.3 Heavy minerals The distribution of heavy minerals in the coarse fraction (Fig. 5.5) indicates that high concentrations occur, primarily, off the northwest and, secondarily, off the southeast part of the lerissos gulf coast. In more detail, the distribution of selected heavy minerals (Fig. 5.6) shows that: 1. Amphiboles, pyroxenes and garnets are abundant off the southern coast of lerissos gulf. The presence of amphibolites and amphibolitic gneisses on the adjacent mainland is likely to be responsible for the amphibole and pyroxene abundances. The enrichment in garnet could be attributed to the corresponding local occurrences of ultramafic masses (dunites and peridotites) on the southern coast of the gulf (Fig. 1.4). 2. Two local enrichments of pyrite occur: (a) just off Stratoni district,in the northwestern part of lerissos gulf; and, (b) off the southern coast of lerissos gulf. Both these offshore pyrite occurrences are likely to be related to the mixed sulphide mineralization which occurs on the mainland adjacent to these enrichments. 5.1.1.2 Carbonates The distribution of calcium carbonate (Fig. 5.7) shows that, in general, low calcium carbonate content occurs in the lerissos gulf sediments. Two local enrichments are present, containing more than 50% CaCOg. these are located (a) off the southern entrance of the gulf, in the vicinity of cape Arapis, and (b) off the central area of the south coast of the gulf. 5.1.1.3 Organic carbon The variation of organic carbon (Fig. 5.8) shows that in general its content in the 180 lerissos gulf sediments is low. The central part of the gulf, where fine grained sediments predominate, is relatively enriched in organic carbon, with values ranging from 0.6 to 0.9%. By contrast, the coastal sediments are rather depleted in organic carbon, containing less than 0.3%. The highest organic carbon content was determined in the vicinity of cape Arapis. 5.1.1.4 XRD Results The samples I 5, I 22 and I 68 were subjected to X-Ray Diffraction studies. The minerals which have been recognized are listed in Table 5.1. Table 5.1: Mineralogical composition of selected samples from lerissos gulf, as determined through XRD analysis. Sample I.D. Major minerals Minor minerals I 5 quartz aragonite K-feldspar calcite plagioclase I 22 talc high Mg-calcite quartz hornblende dolomite aragonite K-feldspar kaolinite/chlorite calcite I 68 quartz K-feldspar calcite high Mg-calcite plagioclase kaolinite/chlorite illite aragonite 181 0 2 km A T Ofij, NEA RODA • surface samples 0 cores Fig. 5.1 lerissos g u lf: map representing the offshore surface and subsurface sample locations. r V .a s p r o l a k o s i. . lrso gul: a rpeetn te nhr sml locations. sample onshore the representing map lf: u g lerissos 5.2 Fig. * 2 km 2 0 -- * 182 r ► § 183 0 % BUSH 20-30 10-20 5-10 c z ra < 5 Fig. 5.3 lerissos gulf surface sediments : distribution of rock fragments in the coarse fraction. 0 km % m m 20-30 ^ 10-20 m m < 10 * Fig. 5.4 lerissos gulf surface sediments : distribution of quartz in the coarse fraction. ^ 184 0 5 .km % ^■i 11li 6^0 x z m 4 - 6 UTrm < 2 Fig. 5.5 lerissos gulf surface sediments : distribution of the heavy minerals content in the coarse fraction. 0 5 km ♦>10 ^ 5-10 ^<5 l>4 B 2-4 D <2 r epidote •>8 6 4-8 0<4 m e ta llic t- A abundant A present Fig. 5.6 'enssosjjulfsurface sediments: distribution of selected heavy mi minerals In the 2 o to 185 0 5 km % ^ 30-50 ^ 10-30 { n m < 10 Fig. 5.7 lerissos gulf surface sediments : distribution of -total carbonates 0 5 km % m n > 3 IOJTTI .3-6 C U R 3 r K Fig. 5.8 lerissos gulf surface sediments : distribution of organic carbon. f 186 5.1.2 Bulk geochemistry of offshore sediments and of those on the adjacent coast Elements are dealt with in order of decreasing abundance, as this is established on the basis of the robust means (Table 8.1) of Box-Cox transformed geochemical data (for a definition on robust statistics and Box-Cox transformation see Appendix V). 5.1.2.1 Silicon The regional variation of Si in the offshore sediments (Fig. 5.9) shows that high Si values occur in a belt stretching along the coast, apart from the coastal area lying in the northwestern part of the gulf, off Stratoni district. The highest Si concentrations were determined in the southwest region. Low Si content occurs in the sediments lying in the central deeper part of the gulf. Most of the samples collected from the onshore coastal area are rich in Si (Fig. 5.9). The sediments collected from the river Asprolakkos and the one flowing out in the south part of the gulf are Si enriched. 5.1.2.2 Aluminium The regional pattern of Al (Fig. 5.10) indicates that Al is in general antipathetic to Si, in its distribution. Hence, an Al enrichment occurs throughout the central deeper part of the gulf, where Al content exceeds 8%. The onshore coastal samples are, generally, rather Al poor (Fig. 5.10). However, in the south and southeast part of lerissos gulf, Al enriched river sediments do occur. 5.1.2.3 Calcium The distribution pattern of Ca in the sediments (Fig. 5.11) is almost the reverse of that of Si. In general terms, the Ca content of the sediments of lerissos gulf is rather low. However, relatively high Ca values (over 7%) occur in the following two areas (a) a coastal belt which stretches from lerissos town to Nea Roda village, and (b) a coastal location which lies off the northwest coast of Petrovouni peninsula, in the eastern part of lerissos gulf. In addition, two relatively Ca enriched, but local, areas occur in the northern part of the gulf; the one off the northwest coast, in the vicinity of 187 Stratoni district, and the other off the northeast coast. The onshore coastal samples, with the exception of various samples collected from the south and southwest coast of the gulf, are low in Ca (Fig.5.11). 5A £A__ Iran The distribution pattern of Fe (Fig. 5.12) shows that the variation of Fe follows closely that of Al in the offshore area which lies southwards of lerissos town. Nevertheless, the maximum Fe content occurs in the northwest part of the gulf, where an exceptional Fe enrichment (10-12% Fe) was determined in the coastal area which is located just off Stratoni district, corresponding to an Al depletion. The Fe enriched area of the northwest part of lerissos gulf extends southwards and southeastwards to merge with the central region of the gulf itself. The variation of Fe in the onshore coastal samples (Fig. 5.12) indicates its enrichment in the area which lies from Stratoni town to Asproiakkos river. The Asprolakkos river sediments are also rich in Fe. 5.1.2.5 Potassium The distribution pattern of K (Fig. 5.13) is fairly similar to that of Al. Hence, high K concentrations occur in the sediments of the central deeper region of the gulf. The sediments of the coastal areas are in general rather low in K, apart from the following two locations: (a) an extensive area lying off the southeast coast; and (b) a small area lying off the coast in the vicinity of lerissos town. The most extensive sector of K enrichment in the sediments (Fig. 5.13), with K content exceeding 2.2%, occurs in the northern central part of lerissos gulf. The onshore coastal samples are generally low in K (Fig. 5.13). However, relatively enriched samples were collected from : (a) lerissos district and its nearby coastal area; (b) Nea Roda; and (c) the river which flows into the southeast part of the gulf. 5.1.2.6 Magnesium The regional variation of Mg in the offshore sediments (Fig. 5.14) shows that a high Mg content occurs in the sediments collected throughout most of lerissos gulf. Thus, Mg concentrations higher than 2.3% occur in the sediments of (a) the offshore area 188 which lies to the north of the Asprolakkos river mouth, and (b) a small coastal sector located off Nea Roda. In contrast - apart from the samples collected from the lerissos district, the central part of the west coast and the central part of the south coast of the gulf - the onshore coastal samples (Fig. 5.14) seem to be rather depleted in Mg. 5.1.2.7 Titanium The distribution pattern of Ti (Fig. 5.15) is similar to that of Mg. Sediments collected throughout most of the gulf have high Ti concentrations. The highest Ti content, just over 0.5%, occurs seawards of the Asprolakkos river mouth. High Ti concentrations occur in the sediments collected from : (a) an extensive area which lies off the west coast; (b) a less extensive offshore sector located in the southeast part of the gulf; and (c) a small area which lies off the central part of the north coast. In contrast, the narrow coastal strips which stretch along the southwest, south, east and northeast coasts are low in Ti. In addition, the narrow belt which lies off Stratoni district is also Ti poor. The onshore coastal samples, apart from two collected from the central part of the west coast and the central part of the south coast respectively, seem to be low in Ti. 5.1.2.8 Manganese The distribution pattern of Mn (Fig. 5.16) is closely related to that of Fe. Thus, the sediments of the central deeper part of the gulf are enriched in Mn. A high Mn content was also determined in the sediments collected from the offshore region that lies from Stratoni town to the mouth of Asprolakkos river. By contrast, the sediments of the nearshore areas off the S-SW and off the SE coasts of lerissos gulf are poor in Mn. The highest Mn content (9,400 jig/g) occurs in a coastal sample collected from a beach lying in the vicinity of a fertilizer company, in the northwestern region of the gulf. The onshore coastal samples collected from the southwestern and southern parts of the coast are poor in Mn. 5.1.2.9 _PhOSPtLP.m£ The regional variation of P (Fig. 5.17) shows that a high P content occurs in the 189 sediments throughout the central part of the gulf, corresponding to an Al, Fe, K and Mg enrichment and a Ca depletion. Phosphorus concentrations exceeding 500 jig/g occur in the following areas : (a) an extensive region that lies off the central part of the west coast, and extends towards the north central part of the gulf; and (b) a narrow coastal belt which stretches from lerissos town to Nea Roda village. By contrast, a low P content occurs in the sediments lying in the SW region of the gulf, and in those along the south and east part of the coast. Apart from a sample collected from the central part of the west coastline which is enriched in P, as well as in Cr, Mn, Ti and Zr, the onshore coastal samples are poor in P (Fig. 5.17). 5.1.2.10 Barium The distribution pattern of Ba (Fig. 5.18) shows the following variation: 1. The sediments of the deeper central part of the gulf have a rather low Ba content. 2. Four areas of Ba enrichment occur, these are: (a) a coastal area which lies at the northwest part of lerissos gulf, just off Stratoni district; (b) a coastal area which is located off the S-SW coast; (c) a belt that stretches along the southeast and east coast; and (d) a sector that lies off cape Arapis. The onshore coastal samples collected from the northern half of the west coast are poor in Ba. By contrast, high Ba concentrations were determined in the beach samples collected from the southern part of the west coast, the Stratoni district, and also the south coast of lerissos gulf. In addition, more than 700 jxg/g Ba occurs in the river sediments collected from the river that flows into the southeast part of lerissos gulf. 5.1.2.11 Strontium In general, the regional variation of Sr (Fig. 5.19) shows some similarities with that of Ca (Fig. 5.11), although it does not follow closely the Ca distribution pattern. Thus, the N-NW part of lerissos gulf is poor in Sr, with less than 250 |ig/g Sr content. In addition, low Sr concentrations occur throughout the central deeper part of the gulf. High Sr values occur in sediments collected from : (a) a nearshore sector that lies off the N-NE coast; (b) a coastal belt that stretches from Nea Roda to the east, along the south and east coastline of the gulf; and (c) the entrance of lerissos gulf. 190 The onshore coastal samples are rather Sr poor (Fig. 5.19). However, various samples showing Sr enrichment were collected from the S-SW and from the south coast of lerissos gulf. 5.1.2.12 Zinc The regional variation of Zn (Fig. 5.20) follows closely those of Fe and Mn. High Zn values occur throughout the whole gulf, except for a rather narrow belt that stretches along the south and east part of the coast, the entrance of the gulf, and finally the northeast part of the coastline. Zinc concentrations higher than 700 pg/g were determined in a quite extensive area lying off the northwest and west part of the coast. Apart from the sediments collected from the coastal area lying between Stratoni and the central part of the west coast, as well as the Asprolakkos river sediments, the onshore coastal samples are Zn poor (Fig. 5.20). 5.1.2.13 Chromium The distribution pattern of Cr (Fig. 5.21) shows that most of the gulf sediments are enriched in Cr. Nearshore samples collected just east of Nea Roda are highly Cr enriched; one of these samples has a Cr content higher than 0.6%. A Cr enrichment, containing over than 250 pg/g Cr, occurs off the Asprolakkos river, while northwards of the river mouth itself sediments containing more than 350 pg/g were collected. The onshore coastal samples are, generally, rather Cr poor (Fig. 5.21). However, high Cr values were determined in : (a) a beach sample collected from the central part of the west coast; (b) one of the Asprolakkos river sediments; (c) the river sediments collected from the river that flows into the central part of the south coast of lerissos gulf; and (d) the beach samples collected from the vicinity of the outlet of the river mentioned above in (c), up to 0.4% Cr occurs in the beach sample of the small bay that lies to the east of this river. 5.1.2.14 Vanadium The regional variation of V (Fig. 5.22) is fairly closely related to that of Ti (Fig. 5.15). The central deeper part of the gulf is enriched in V. The highest V content occurs in the following two areas : (a) an extensive area that lies off the west coast; 191 and (b) a rather small area which lies just off Nea Roda village. The onshore coastal samples are V poor (Fig. 5.22), apart from that collected from the central part of the west coast (containing up to 220 pg/g V), which is also enriched in Ti, Mn, P and Cr. 5.1.2.15 Zirconium The regional variation of Zr (Fig. 5.23) is similar to that of Ti (Fig. 5.15). Most of the gulf sediments are Zr enriched. The Zr content is higher than 100 pg/g in the samples collected off the west coast, and in the sediments of the small sectors which lie off the north coast and in the southeast part of the gulf. By contrast, low Zr values occur in the sediments of the following areas : (a) off the Stratoni district; (b) off the east coast; and (c) in a narrow coastal belt which stretches along the SW-S-SE coast. Apart from the beach sample collected from the central part of the west coast and a few river sediments collected from the rivers that flow into the south and southeast part of the gulf, the onshore coastal samples are poor in Zr (Fig. 5.23). 5.1.2.16 Nickel The distribution pattern of Ni (Fig. 5.24) follows closely the spatial variation of Cr. Thus, the central deeper part of lerissos gulf is Ni enriched. Sediments with higher than 100 pg/g Ni content were collected from the following areas : (a) off the west coast; (b) off the central part of the south coast; and (c) southweastwards of the centre of the gulf. The onshore coastal samples are rather Ni poor (Fig. 5.24). However, Ni enrichments occur in : (a) some of the Asprolakkos river sediments; (b) the Ti, Mn, P, Cr and Zr enriched sample collected from the central part of the west coast; and (c) the beach samples collected from the central part of the south coast, and the river sediments collected from the river that flows into this area (up to 930 pg/g Ni content was determined in this last situation). 5J..2ZLZ__ Copper. The distribution pattern of Cu (Fig. 5.25) shows that it is similar to the regional distributions of Fe, Mn and Zn. Thus, the major part of the gulf is Cu enriched. The 192 highest Cu content occurs in a coastal region situated between Stratoni and the central part of the west coast. By contrast, most of the remaining coastal areas are Cu poor. In addition, low Cu content was determined in the sediments of the gulf entrance. The onshore coastal samples are poor in Cu (Fig. 5.25), with the exception of those collected from the Asprolakkos river and the beach adjacent to it. 5.1.2.18 Lanthanum The regional variation of La (Fig. 5.26) shows that the major part of lerissos gulf is enriched in La. Lanthanum contents exceeding 40 jig/g were determined in the following locations : (a) an extensive central region lying seawards of the west coast; (b) an area of limited extent confined off the north coast; (c) a small area lying off the north entrance of the gulf; and (d) a small area in the southeast part of the gulf. By contrast, nearshore sediments collected off the northwest, southwest and east parts of the coastline are poor in La. The onshore coastal samples are La poor (Fig. 5.26), except for that collected from the central area of the west part of the coast. 5.1.2.19 Cobalt The distribution pattern of Co (Fig. 5.27) shows that the lerissos gulf sediments are, in general, poor in Co. However, a local coastal enrichment (80 p.g/g Co) occurs just east of Nea Roda, associated with enrichments in Mg, Cr and Ni. The onshore samples are poor in Co (Fig. 5.27), apart from two local enrichments, one in the central part of the west coast and another in the central part of the south coast. Both these onshore Co enriched locations correspond to Mg, Cr and Ni enrichments. 5.1.2.20 Beryllium The distribution pattern of Be (Fig. 5.28) follows closely that of Al (Fig. 5.11). High Be concentrations were determined in sediments throughout much of the gulf, particularly in the deeper central part of it. By contrast, low Be contents occur in sediments along the north, northwest and east parts of the coast. 193 The onshore samples are poor in Be (Fig. 5.28), apart from some of those collected from the south part of the coast. 194 0 u . Si % Fig. 5.9 lerissos gulf and adjacent coast: regional variation in the Si content of the surface samples. o 5 j Km Al % ^ 2 65-8.0 fciiiiif < 5,0 f Fig. 5.10 lerissos gulf and adjacent coast: regional variation in the Al content of the surface samples. r 195 Fig. 5.11 lerissos gulf and adjacent coast: regional variation in theCa content of the surface samples. Fe % mm 55-80 mm 4D-S5 15-4.0 Fig. 5.12 lerissos gulf and adjacent coast: regional variation in the Fe content of the surface samples. 196 Fig. 5.13 lerissos gulf and adjacent coast: regional variation in the K content of the surface samples. o Mg % ■ n i ■ i 'i m t v / s m m 12 -2 0 E-US < 0.7 Fig. 5.14 lerissos gulf and adjacent coast: regional variation in the Mg content of the surface samples. 197 o Ti H9'9 >3750 2500 - 3750 1000-2500 <1000 Fig. 5.15 lerissos gulf and adjacent coast: regional variation in the Ti content of the surface samples. 0 Mn H9^9 V77ft 1000- 4000 S 500-1000 EHS3 < 500 Fig. 5.16 lerissos gulf and adjacent coast: regional variation in the Mn content of the surface samples. 198 5 l J Km p n g /g >550 450-550 250-450 < 250 Fig. 5.17 lerissos gulf and adjacent coast: regional variation in the P content of the surface samples. 5 Krrj Ba (ig/g >450 320-450 280-320 < 280 Fig. 5.18 lerissos gulf and adjacent coast: regional variation in the Ba content of the surface samples. Fig. 5.19 lerissos gulf and adjacent coast: regional variation in the Sr content of the surface samples. 0 5 j km Zn ng/g Fig. 5.20 lerissos gulf and adjacent coast: regional variation in the Zn content of the surface samples. 200 0 5 km Cr ng/g >350 250 -350 mm 150-2 50 ^ 75-150 E03 <75 Fig. 5.21 lerissos gulf and adjacent coast: regional variation in the Cr content of the surface samples. -* km v |ig/g E B >150 m m 10 0 -150 5 0 -10 0 E n a < so Fig. 5.22 lerissos gulf and adjacent coast: regional variation in the V content of the surface samples. 201 0 , Km Zr >100 75-100 50-75 n g /g Fig. 5.23 lerissos gulf and adjacent coast: regional variation in the Zr content of the surface samples. 0 km Ni Z2ZA 70-100 25-70 i^g/g iiimiil <£ 25 Fig. 5.24 lerissos gulf and adjacent coast: regional variation In the Ni content of the surface samples. 202 0 Cu ng/g >75 50-75 20-50 <20 Fig. 5.25 lerissos gulf and adjacent coast: regional variation in the Cu content of the surface samples. 5 Km La n g /g ZZ721 30-40 n n r^ < 2 0 Fig. 5.26 lerissos gulf and adjacent coast: regional variation in the La content of the surface samples. 203 0 5 i. mJ km Co ng/g ezzza> 50 ^ 35-50 tuna 2 0 -3 5 l 3 <20 Fig. 5.27 lerissos gulf and adjacent coast: regional variation in the Co content of the surface- samples. o Be ng/g n > 2 6 wm 1.7-26 Fig. 5.28 lerissos gulf and adjacent coast: regional variation in the samples. Be content of the surface 204 5.1.3 Multivariate Statistical Analysis Multivariate statistical analysis was applied to the data concerning the elemental composition of the lerissos gulf offshore surface sediments, for reasons similar to those explained in section 3.1.3.1. 5.1.3.1 Correlation Matrices The robust correlation matrices of the raw, untransformed and the Box-Cox transformed data are presented in Table 5.2 a and b, respectively. The lowest value of correlation coefficient (r) which is significantly different from zero at the 0.001 probability level is ± 0.360 (for the 83 samples of the lerissos data set), (Fisher, 1963, cit. Howarth, 1983). A comparison of the two correlation matrices shows that there are 48 cases where the correlation coefficients have a difference 0.1-0.2, 30 instances representing a ± 0.2-0.3 difference, 10 pairs reflecting a 0.3-0.4 discrepancy, and one case where the difference exceeds 0.4. They correspond to 25.3, 15.7, 5.3 and 0.5 %, respectively, of the total number of pairs. Therefore, due to the differences found between the two correlation matrices, it was deemed appropriate to transform the data, in order to study the background elemental associations. Cluster and factor analysis have been carried out on both transformed and untransformed data for the study of background relationships and the investigation of a potentially existing anomalous component, respectively. Table 5.2 : The robust correlation matrices of the raw, untransformed (a) and the Box-Cox transformed (b) data of the lerissos gulf. K BE HG CA SR BA LA I l ZR 0 K 1.00000 BE .31070 1.00000 MQ - . 15870 .48410 1.00000 CA -.5 8 9 5 0 -.2 9 5 3 0 .14100 1.00000 SR -.36070 -.41910 -.2 7 4 3 0 .82650 1.00000 BA .63190 -.34100 -.7 2 0 1 0 -.3 6 0 7 0 .07330 1.00000 LA -.20670 .15630 .56610 .41750 .17580 -.5 2 9 7 0 1.00000 TI -.1 4 7 4 0 .51610 .82730 -.1 1 8 0 0 -.4 2 4 7 0 -.6 9 6 5 0 .52850 1.00000 ZR -.0 9 6 9 0 . 7.8750 .59930 -.10880 -.33090 -.4 6 3 6 0 .33250 .73980 l .00000 0 .04020 .52870 .75910 -.1 6 7 2 0 -.4 0 4 2 0 -.4 5 4 7 0 .37480 .78630 .61670 1.00000 CR -.1 2 1 2 0 .30500 .66920 -.0 6 2 5 0 -.2 7 0 3 0 -.4 2 3 8 0 .21150 .54780 .46520 ,55000 HN -.0 3 7 9 0 -.0 4 5 2 0 .44950 -.0 8 8 2 0 -.3 7 9 4 0 -.2 5 7 4 0 .09110 .26120 .16170 .37390 FE ♦ 01200 .47200 .85990 -.1 0 8 7 0 -.5 0 0 9 0 -.56050 .38750 .72320 ,50250 .75600 CO .05030 ,49490 .72630 -.1 2 0 6 0 -.3 7 9 7 0 -.4 7 2 1 0 .42970 .65660 .47680 .68810 NI .01400 .51670 .04900 -.0 9 4 8 0 ~ .33880 -.4 7 6 6 0 .37900 .70560 .57250 .75780 CU .10070 .11300 ,53070 -.1 4 1 7 0 -.42260 -.2 4 7 8 0 .10770 .30080 .25140 .45890 ZN .01720 .02230 .50700 -.10360 -.4 0 3 0 0 -.2 6 0 7 0 . 10810 .30370 .23750 .43680 AL .413150 .08120 .56640 -.4 3 2 5 0 -.5 2 7 9 0 -.2 4 8 6 0 .25130 .67550 .42170 . f>4640 ST .30910 -.3 4 5 7 0 -.7 7 0 1 0 -.5 8 3 4 0 -.1 8 7 7 0 .70930 -.5 9 6 6 0 -.40590 .25590 -.44720 P - . 20140 .34190 .67660 .09420 19560 -.6 4 3 0 0 .66630 .74700 .49930 .63420 CR HN FE r:o NT CU ZN Al ST P CR 1.00000 HN .25600 1.00000 FE .50210 ,74760 1.00000 CO .50630 • 46430 .74040 1.00000 NI .77200 ,29980 .68710 .74170 1.00000 CU .31010 .80500 .80450 .52170 .40320 1.00000 ZN .29220 .96890 .79770 .52370 .35940 .92160 1,00000 AL .33150 .10020 .59240 .57040 .50900 .28750 .17740 1.00000 SI -.3 3 6 5 0 -.3 5 1 4 0 . -.66480 -.4 5 8 5 0 -.4 8 7 7 0 42360 -.3 0 6 6 0 -.3 1 7 6 0 1.00000 P .33240 ,27060 .60840 .49960 .46400 .30390 .28690 .40400 -.52660 1.00000 ro 0 01 1 K BE MG CA SR UA L A 1 I 7R V K 1.00000 BE .33970 1.00000 hQ -.33100 .3 4 5 /0 1.00000 CA -.4 3 5 0 0 -.1 6 4 0 0 .52390 1.00000 SR -.2 3 5 9 0 -.1 7 3 0 0 -.37590 .38740 1.00000 BA .65720 -.2 2 4 3 0 -.7 7 4 5 0 -.6 0 0 2 0 .04700 1.00000 LA -.2 1 0 0 0 .12810 .56180 .54110 .19830 -.5 3 0 0 0 1*00000 TI -.1 2 7 6 0 .50130 . 83830 .13730 -.3 9 3 9 0 -.6 3 8 4 0 , 50/50 1 .00000 ZR - . 0 / 6 2 0 .33430 .64190 .14090 -.2 4 5 6 0 -.4 2 1 4 0 .39660 .77890 1,00000 V .05030 .48450 .63240 .08760 - . 3 / 1 5 0 -.3 8 2 5 0 .34680 .71410 .62370 1.00000 HR -.2 0 0 1 0 ♦40620 .82700 .20910 -.3 9 4 5 0 -.6 2 0 5 0 .3 /8 9 0 ,83910 .67630 .65050 MN -.1 5 2 6 0 . t0170 .76730 .23600 -.7 1 5 1 0 - . 4 9 1 00 .28000 .'59340 .42280 ,55710 FE -.0 0 0 5 0 .35650 .85580 ♦ 20080 ~.7297o - .51.380 ,32540 .71840 .52320 .65940 CO .11930 .53750 .53010 -.0 6 1 8 0 -.3 6 0 9 0 -.4 1 5 4 0 .39630 ,63090 .49450 .60040 NI .05020 .59480 ,73730 .07820 -.35030 - .53860 .43640 ,80730 .67130 .75500 CIJ .23440 .37700 .53030 .00230 -.6 5 3 7 0 -.2 4 4 5 0 .08060 .41500 ,30760 .50690 ZN .00720 .23520 .79400 .23140 -.6 8 8 5 0 -.4 3 8 0 0 .2 /0 4 0 .61530 .46810 .63230 AL .52570 ♦07540 .40260 -.2 8 2 1 0 34070 - . 12950 .22110 .61220 .45170 .58490 SI .30000 -.1 5 9 6 0 -.7 2 3 2 0 -.7 0 5 6 0 .10630 .71490 -.5 1 8 3 0 -.3 7 7 2 0 -.1 8 9 7 0 -.3 3 6 9 0 P -.1 5 0 4 0 .30390 .75110 .34690 -.2 2 2 9 0 -.5 4 6 5 0 .62620 .75540 .57400 .60940 r.R MN FE CO NI CIJ ZN AL SI P CR 1.00000 MN .63140 1.00000 FE .71120 .90460 1.00000 CO .5 9 /7 0 .47230 .56160 1.00000 NI .02100 .52250 .68820 .79050 1,00000 CU .4 /4 1 0 .69340 .7/420 .51180 .58010 1,00000 ZN .65070 .89960 .94450 .52720 .63640 .82900 1.00000 AL .4 /5 5 0 .24300 .49670 .60000 .65170 .45300 . 39040 l .00000 SI -.3 9 7 7 0 -.5 570 0 -.6 4 7 5 0 —.27580 -.4 1 5 0 0 -.4 8 8 5 0 -. 63620 - .16710 1 .00000 P .57060 .55^00 .64900 .46690 .52570 .35320 .55270 .46560 -.4 6 1 9 0 1.00000 b 206 2 0 7 5.1.3.2_____Cluster Analysis 5.1.3.2.1 Box-Cox transformed data Cluster analysis was applied to the Box-Cox transformed data in conjunction with its robust correlation matrix (C.M.). The dendogram produced is presented in Fig. 5.29 a. It displays graphically the following elemental associations, for which a possible interpretation is given below: Ca -Sr : It, probably, represents the biogenic carbonate phase of the sediments. The isolation of this group from the remaining elements studied suggests that none of the other elements is associated with the carbonate phase. Be - Al : It, possibly, represents clay minerals. La -P : The significant association of La with P could be the result of substitution of La in biogenic or detrital apatite. Zr -V -Ti - Cr -Co -Ni : It, most probably, represents detrital material deriving from the erosion of the mafic/ultramafic rocks that occur in the mainland. Mg - Mn - Cu - Fe - Zn : This may represent Fe,Mg-minerals and weathering products of the mixed sulphide deposits, the Cu-bearing porphyry stocks, and the Mn-oxides that occur in the mainland. K - Ba - Si : It probably reflects detrital material of felsic igneous rocks. 5.1.3.2.2 Raw, untransformed data Cluster analysis was also applied to the raw, untransformed data in conjunction with its robust C.M. (Table 5.2 a). The dendogram provided is presented in Fig. 5.29 b. This dendogram indicates more clearly the mineralization component, which is represented by the Mn - Zn - Cu association. 208 a COPHENETIC CORRELATION COEFFICIENT = 0-586* CEorujjcr cr « k o « o z z d u j z cr •— UCOCDCE_JQ.M>h-UUZZEUL-Mi:(r3W oe CC NJ o 209 b COPHENETIC CORRELATION COEFFICIENT = 0.5604 cr o: z z 3 cr a: « o uj o ce •— lu _i cc — (Jl/)EM UJ[LL)Z EL.UNt->CDCLiC0(/) O Uj c u. Fig. 5.29 Cluster analysis of lerissos gulf data; robust dendograms : (a) Box-Cox transformed data, (b) raw data. 2 1 0 5.1.3.3 Factor Analysis 5.1.3.3.1 Box-Cox transformed data The model with 4 factors (Table 5.3) was chosen on the basis of the following: 1. The Cattell Scree test (Appendix V) suggests that only factors characterized by an eigen value higher than 1 are significant; a 4-factor model would, therefore, be a satisfactory solution (Table 5.4). 2. The 4 factors account for a very high proportion, 83.3%, of the total data variability. Nevertheless, 5, 6 and 7-factor models were tested for comparison purposes; however, they did not improve the factor solution. In this study, elements with factor loadings higher than 0.5, in absolute value, (Table 5.3) are considered to describe the composition of the relative factors. In the following, a possible interpretation of the elemental associations obtained from the 4-factor model Is given: Factor 1: This factor accounts for 27.6% of the data variability. It exhibits high positive Ti, Zr, Cr, Ni, V, Mg, P and Co loadings. This elemental association, probably, reflects detrital products of mafic/ultramafic rocks. Two groups of nearshore samples are characterized by high positive scores on this factor. These are: (a) samples I 57, I 59, I 77, I 78 and I 79, off the west part of the coast, the composition of this group probably reflects the presence of amphibolites (ultra-basalts and meta-gabbros) in the corresponding part of the mainland; and (b) samples I 26 and I 29 , off the central area of the south part of the coastline, the composition of this group can possibly be related to a small ultrabasic igneous complex outcropping east of Nea Roda (Fig. 1.4). Factor 1 has a significant negative Ba loading, due to the negative correlation of the elements on it with Ba, as shown in the corresponding C.M. (Table 5.2 b). Factor 2: This factor accounts for 18.1% of the data variability. It comprises high positive Si and Ba loadings, probably associated with detrital material of felsic 21 1 Table 5.3 : Factor analysis of Box-Cox transformed data of the lerissos gulf. Four factor m odel: factor loadings and variance accounted for by components, computed on the basis of the robust C.M. of the Box-Cox transformed data. Element Factor 1 Factor 2 Factor 3 Factor 4 K -.2 6 .47 .11 .75 Be • .41 -.04 .03 .78 Mg .64 -.54 .50 .03 Ca .04 -.90 .00 -.31 Sr -.2 2 -.38 -.85 -.05 Ba -.5 4 .67 -.18 .17 La .39 -.7 0 -.07 .16 Ti .87 -.2 0 .29 .18 Zr .84 -.06 .14 .07 V .65 -.1 3 .37 .32 Cr .79 -.2 0 .37 .09 Mn .37 -.21 .85 -.10 Fe .44 -.2 6 .82 .17 CO .58 -.1 0 .30 .45 Ni .72 -.20 .34 .41 Cu .13 -.10 .83 .34 Zn .34 -.2 5 .87 .12 Al .44 .04 .19 .83 Si -.0 5 .84 -.47 -.08 P .61 -.4 3 .26 .15 Variance accounted for bv components Component Variance percent Cumulative-variance 1 27.65 27.65 2 18.13 45.78 3 23.88 69.66 4 13.62 83.28 2 1 2 igneous rocks, and hiah neaative Ca, La and Mg loadings, probably reflecting carbonate j material. (La substitutes for Ca in hornblende and apatite (Mason, 1966), however, the reasons for its participation in this factor are unclear.) Factor 3: This factor accounts for 23.9% of the data variability. It exhibits very high positive Zn, Mn, Cu and Fe loadings, and a relatively lower Mg loading. The Zn, Mn, Cu and Fe association could reflect the mainland mineralization, which includes mixed sulphide deposits, Cu-bearing porphyry stocks and Mn-oxides. The presence of Mg in this factor, too (apart from its presence in factor 1) could represent the relation of the mainland mineralization to mafic rocks. Samples I 44, I 46, I 48, I 49, I 59, I 51, I 52, I 72, I 73, I 75 and I 78 - collected from the northwest coastal part of the gulf (from Stratoni to the north of the Asprolakkos river mouth, Fig. 5.1) - are characterized by high positive scores on factor 2. This implies the influence of the mainland mineralization on the composition of the corresponding offshore coastal sediments. In addition, it could imply that this coastal area is contaminated by the large mining operation, based close to the sea, just outside the town of Stratoni. Factor 3 has a high negative Sr loading, due to the significant negative correlation of the elements on it with Sr, as shown in the corresponding C.M. (Table 5.2 b). Factor 4: This factor accounts for 13.6% of the data variability. It comprises high Al, Be and K loadings, and is likely to represent clay minerals. Table 5.4 : lerissos gulf Box-Cox transformed data. List of eigen values of all the components of the data. Component Eigen Yalus Component Eigen Value 1 10.12 11 .19 2 3.38 12 .13 3 2.07 13 .08 4 . 1.10 14 .07 5 .80 15 .06 6 .57 16 .05 7 .49 17 .03 8 .36 18 .02 9 .26 19 .01 10 .22 20 .01 213 5.1.3.3.2 Raw, untransformed data Factor analysis was applied to the raw, untransformed data, in conjunction with its robust C.M., in order to investigate whether geochemical outliers occur in the lerissos gulf data. The 4-factor model (Table 5.5) was chosen on the basis of the following: 1. The Cattell Scree test (see Appendix V) suggests that only factors characterized by an eigen value higher than 1 are significant; therefore, a 4-factor model would be a satisfactory solution (Table 5.6). 2. The 4 factors account for a very high proportion, 81.8%, of the total data variability. Factor 1: This is almost identical to the mafic/ultramafic factor of the previous factor analysis (section 5.1.3.3.1.). Factor 2: This is very similar to the previous factor 2. In the present case, however, Sr has a high negative loading, most probably representing its relation to the carbonate phase of the sediments. Factor 3: This factor accounts for 20.7% of the data variability, exhibiting high Zn, Mn, Cu and Fe loadings. Factor 3 is probably the mineralization component. Factor 4: This factor is identical to the clay factor of the previous analysis. Although the robust correlation matrices of the raw and the Box-Cox transformed data exhibit significant differences, the elemental associations formed when cluster and factor analysis were applied to both data do not show big differences. In both cases, the influence of the mainland mineralization on the corresponding offshore sediment composition was picked up. 214 Table 5.5 : Factor analysis of raw data of the lerissos gulf. Four factor model : factor loadings and variance accounted for by components, computed on the basis of the robust C.M. of the raw data. Element Factor 1 Factor 2 Factor 3 Factor 4 K -.31 .42 .08 -.7 5 Be .42 -.06 -.06 -.7 9 Mg .76 -.38 .39 -.2 3 Ca -.12 -.87 -.09 .40 Sr -.41 -.65 -.35 .34 Ba -.7 0 .51 -.14 -.13 La .39 -.68 .03 -.1 5 Ti .89 -.14 .14 -.2 4 Zr .81 .06 .06 -.0 3 V .73 -.07 .31 -.36 Cr .73 .03 .18 .01 Mn .14 .00 .96 .08 Fe .57 -..18 .72 -.31 Co .59 -.14 .43 -.36 Ni .76 -.1 0 .25 -.2 9 Cu .18 -.01 .94 -.11 Zn .19 .00 .97 .01 Al .46 .02 .10 -.8 5 Si -.3 4 .80 -.35 .16 P .63 -.3 9 .17 -.2 0 Variance accounted for bv comoonents Component Variance percent Cumulative variance 1 31.26 31.26 2 15.57 46.83 3 20.68 67.51 4 14.27 81.79 215 Table 5.6 : lerissos gulf raw data. List of eigen values of all the components of the data. ■OgmpQDepl Eiaen Value Component Eigen Yl 1 9.25 11 .14 2 3.40 12 .12 3 2.50 13 .10 4 1.20 14 .08 5 .93 15 .06 6 .74 16 .04 7 . .46 17 .03 8 .38 18 .02 9 .32 19 .01 10 .20 20 .01 216 5.2 Purled sediments Five cores were studied from the lerissos gulf (Fig. 5.1). 5.2.1 D escription o f the cores The buried sediments of the lerissos gulf are in general of silt to silty mud grade. The sediment colours vary from dusky yellow to greyish black, depending on the sampling site and depth. The sediment colours were described using the Munsell Soil Colour Charts (Munsell Colour Company, 1971),(Moorby, pers.comm., 1985). 5.2.1.1 Core I 5 (Fio. 5.30 a) This core consists of silty mud grade material, with a lack of obvious structures. At the top, a thin void occurs. The sediment colour varies with increasing depth, from dusky yellow (5Y 6/4) in the upper part, it gradually changes to dark yellowish brown (10YR 4/2), while in the lowermost 10 cm it becomes greyish olive (10Y 4/2). 5.2.1.2 Core I 8 fFio. 5.31a) This core consists of silty mud grade material, with a lack of visible structures. At the top, a thin void occurs. There is a distinct change of sediment colour downcore. Thus, there is moderate to dark yellowish brown (10YR 5/4 to 10YR 4/2) sediment at the top of this core; below 40 cm, the sediment becomes greyish olive (10Y 4/2) changing gradually to dark greenish grey (5GY 4/1); and finally, between 94 and 97cm, the colour changes sharply to dark yellowish brown (10 YR 4/2). 5.2.1.3 Core I 10__(Fig. 5.32 ..a) This core consists of silt grade material. At the top, a thin void occurs. There is a thin oxidized layer in the upper 2 mm, with a slight brown (5YR 5/6) colour. There is a streak of light brown (5YR 5/6) sediment along one side of the core, which is likely to be surface sediment pushed down the core liner. The remainder of the core is olive grey (N3) to dark grey (N2) in colour. In this core, small amounts of broken shells and some Irregular mottling of dark grey colour (N2) occur. 217 5.2.1.4 _Core I 13 (Fid. 5.33 a) This short core consists of silt grade material. It has a void at the top. An oxidized light brown (5YR 5/6) sediment occurs in the upper 2 mm. The rest of the core is homogeneous, dark grey (N3) to greyish black (N2) in colour. At the base of the core, a thin oxidized light brown (5YR 5/6) layer occurs, similar to that at the surface. 5.2.1.5 Core l 14— (Fia. 5.34 a) This is the longest of the cores collected and studied from the lerissos gulf. A 2 cm void occurs at the top. Between 2 and 2.5 cm depth, a layer of light brown (5YR 5/6) to moderate brown (5YR 4/4) oxidized sediment occurs. There is homogeneous silt grade material between 2.5 cm and 192.8 cm depth, olive grey (5Y 4/1) in colour; some streaks of darker grey silt-clay (N3) occur within it. Finally, from 192.8 to 193 cm depth, a thin layer of light brown (5YR 5/6) oxidized sediment occurs. 5.2.2 Bulk geochemistry of subsurface offshore sediments In this section, the variation in the chemical composition of the subsurface offshore sediments is studied. For this purpose, the chemical data have been plotted for each of the cores. 5.2.2.1 Core I 5 (Fia. 5.30 b) The downcore distribution pattern of Sr is similar to that of Ca. The two elements increase in concentration with depth in the upper 25 cm of the core, decrease between 25 and 50 cm , and gradually increase again below 50 cm down to the base of the core. The elements K, Be, Al and Si exhibit similar downcore variation patterns; they increase in the upper 50 cm and decrease in the lower part of the core. The decrease in the concentrations of Be, Ba, V, Fe, Co, Ni, Cu, Al, Si and P with increasing depth, below 50 cm, is in parallel with a change of the sediment colour, gradually from lighter to darker. Similarities occur among the downcore variations of Cr, Mg and Ti. In the upper 25 cm of this core, the concentrations of Mn, Fe, Cu and Zn decease sharply with depth. Below 25 cm depth, the contents of Mn and Zn are low. 218 £ 2,2, 2 _ Q flis l 9 (Fio. . 5,3.1—bl The elements Kt Be, Al, Fe, Mn and Cu have similar downcore distribution patterns. They decrease in concentration with increasing depth in the upper 20 cm, increase between 20 and 40 cm , and then, decrease downwards to the base of the core. They exhibit an inverse relation to Ca in its downcore variation, suggesting their close association with the non-carbonate fraction of the sediments. The highest Fe, Cu and Zn contents were determined at the top of the core. 5.2.2.3 Core I 10 fFio. 5.32 b) In the uppermost 22 cm of this core a dramatic decrease of the Mn, Fe, Cu and Zn contents occurs with increasing depth (77, 40, 55 and 76 %, respectively). Below 22 cm , these metals do not fluctuate in concentration significantly, being uniformly at low values. In the lower part of this core, K, Ba, Be, Ti and Al decrease In concentration, whereas Ca content increases with increasing depth. 5.2.2.4 Core f 13 (Fig, 5,33 -b l With increasing depth, the elements Mg, Ti, Fe, V, Cr, Zn, Al and Be decrease in concentration whereas Ca, Sr, Mn, Ni, Cu and Zr increase. £2.2,5...ilPlfi I 14 (Fig, 5,34-.h i In the upper 16 cm of this core, a dramatic decrease in the contents of Mn, Fe, Cu and Zn occurs with increasing depth. Below 16 cm, Mn and Cu do not fluctuate in concentration significantly. At 71 cm, an enrichment in Zn, Ni, Co and P was determined. Throughout the core, the elements Ca and Sr exhibit an inverse relation to Al and K. The distribution patterns of Cr, Mg, Ca and Sr are similar. The Cr, Mg, Ca and Sr contents appear to increase, overall, with depth in the uppermost 160 cm, whereas those of K, Ba, V, Fe, Cu, Zn and Al decrease; below 160 cm, the reverse occurs. 219 IERISSOS 5 Fig. 5.30 C0REI5: (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr.Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P (depth in cm, concentrations in pg/g). v : void. 2 2 0 IERISSOS 8 0 -20 -l 10. -40 22000 23000 2.5 2.4 59300 16000 3$000 50000 170 300 W 94 ilo ^H 4000 44001 WO 149 K BE MG CB SR BR Lfl TI ZR -60 -80 -100 cm 660 $60 96000 ^ ° ° ° 43 U 6 143 42. 56 550 78000 #6000 lOOOOO HOCQO 440 5l0‘ MN FE CO NI CU ZN BL SI P Fig. 5.31 CORE I 8 : (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr,Ba, La, Ti, Zr. V, Cr, Mn, Fe, Co, Ni. Cu, Zn, Al, Si and P (depth in cm concentrations in ng/g). H ’ v : void. 2 21 IERISSOS 10 a -80 • i t------1 ----1 210. 270 950 (||00 WOO 70000 16 34 111 I4p44. 270* <730 66000 78000 210000 250000 960 610. CR MN FE CO NI CU ZN RL SI P b Fig. 5.32 CORE 110: (a) description ; (b) downcore plots of K, B.e, Mg, Ca, Sr,Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P (depth in cm, concentrations in pg/g). v : void, sf : shell fragments, o : oxidizing layer, sm : shiny material, m : mottles. 2 2 2 IERISSOS 13 OW W * ™ 5,6 N2 N 3 -20 5YR5/6 cm o a 80. MO 3600 6500 80000 *15000 iO 60 100 lioo 3500 4500 50000 60000 170000 iooooo 4oo 600. CR MN FE CO NI CU ZN AL SI P b Fig. 5.33 CORE 113: (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr.Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P (depth in cm, concentrations in |ig/g). v : void, o : oxidizing layer. 223 IERISSOS 14 __V 0- c>5YR 5/6 -20 -40 OrN* -60 Nsf • -80 -100 5 Y4/1 -120 -140 -160 -200. H------1 -180 164. 230 700 1 b a Fig. 5.34 CORE 114: (a) description ; (b) downcore plots of K, Be, Mg, Ca, Sr.Ba, La, Ti, Zr, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Al, Si and P (depth in cm, concentrations in pg/g). v : void, o : oxidizing layer, sf : shell fragments. 2 2 4 5.3 Summary The distribution of Si in the surface sediments of lerissos gulf strongly reflects that of sand-sized material and quartz; being almost the reverse of that of Ca, which reflects the carbonate abundance. There is, generally, a low Ca content in the lerissos gulf sediments. The antipathetic relation between Si and Ca is, also, reflected by the opposite signs of their loadings on the same factor, determined by factor analysis. Cluster analysis of the geochemical data shows that the biogenic carbonate phase of the sediments does not show an affinity for trace metals. The variation in the chemical composition of the subsurface sediments shows that the downcore distribution patterns of K, Be and Al are similar, exhibiting an Inverse relation to Ca. This illustrates the antipathy of the former group of elements - which are likely to be present, largely, in clay minerals - for carbonates. The lerissos gulf surface sediments are overall significantly enriched in Mn, Zn and Cu, relatively to the samples collected from the remaining part of the study area. In the northwestern nearshore part of the gulf, off Stratoni, a high enrichment in Fe (13%), Mn (9,700 p.g/g) and Zn (3,300 p.g/g) was determined in the surface sediments. Conversely, studying the downcore distribution of these metals, in cores collected from the same coastal area, it was found that their concentrations decrease dramatically with depth in the upper part of the cores. Samples collected from the Asprolakkos river are rich in Fe, Ni, Mn, Zn and Cu; the same applies for the sediments collected off the river mouth. Coastal sediments from the southern part of the gulf, east of Nea Roda, are rich in Mg, Cr, Ti, Ni and Co; they also contain hornblende and talc. The application of cluster and factor analysis to the geochemical data of the lerissos gulf surface sediments indicates the presence of weathering products of mafic/ultramafic rocks, detrital material of felsic igneous rocks, material deriving from the mainland mineralization, and a biogenic carbonate phase. 2 25 CHAPTER VI PARTITION GEOCHEMISTRY 6.1 Introduction - Rationale for partition work The distribution of elements between the various phases making up the sediment can be determined by geochemical partition analysis. Although a time consuming technique, partition analysis allows conclusions to be drawn, concerning the different possible routes by which elements are accumulated into the sediment, with more confidence than on the basis of bulk geochemical analysis alone. The procedure that was used is a modification of the Tessier et al (1979) method, as was described in chapter 2. This method selectively removes, by chemical techniques, certain phases of the sediment for analysis. In that way, studies of various forms and attachments of metals - the exchangeable fraction, carbonate hosted, reducible, organic-sulphide bound, and elements hosted in the lattice structure of aluminosilicate minerals - can be considered individually. Thus, by determining the phase associations of various elements, it becomes possible to investigate both their paths of supply to the sediments and their relative availability for release. For example, in evaluating the pollution potential of a sediment, it is essential to differentiate between the availability or non-availability of trace contaminants to the biota. The non-residual fraction of the sediment represents the reservoir of trace metals available for release if significant changes occur in environmental conditions. By contrast, the trace metals contained in the residual fraction are considered to be unavailable for biological uptake under normal conditions, because they are locked in the crystalline structure of minerals with a very low solubility. It must be borne in mind however, that the distribution of trace metals in any chemically extracted fraction is defined not only by fundamental properties of the system but also, to a great extent, by the method of extraction used. The selection of the samples subjected to partition analysis was based on their metal and organic carbon content, considering the variation of the study area and the need to have a reasonable cover of samples over it. 226 6,2 Surface Sediments A selection of samples collected from the offshore area and the onshore area adjacent to it were subjected to partition analysis. The sample positions are shown in Fig. 6.1. The partitioning of the elements within the five fractions is presented in Fig. VI.1, in Appendix VI, based on their proportions as a percentage of the whole. The average percentage partitioning of each element within each of the five groups of samples individually (I : lerissos gulf, STR : Strymonikos area, KB : Kavala area, THR : Samothraki plateau, B : onshore area), as well as within the whole set of samples are listed in Tables 6.1 and presented in Fig. 6.2 - 6.7. 0 10 k m Fig* 6.1 Map representing the sites of the samples subjected to partition analysis. 227 0 surface samples 0 cores 228 Table 6.1 : Average proportions of each of 20 elements in the exchangeable (E), carbonate hosted (C), reducible (OX), organic and sulphide bound (OR+S), and residual (R) fractions of selected samples from lerissos gulf (A), Strymonikos group (B), Kavala group (C), and Samothraki plateau (D) and from onshore areas (E); also, means of the previous estimates (F). A. lerissos aulf Element E £ £K £B±S B <%) (%) (%) (%) {%) Li 4 4 14 9 70 K 11 2 2 2 82 Be 0 2 27 9 62 Mg 7 4 40 5 45 Ca 12 22 42 3 20 Sr 12 17 33 3 34 Ba 3 5 23 4 65 Al 0 0 1 3 95 La 1 6 21 7 65 V 0 0 12 5 83 Cr 0 1 9 8 81 Mn 1 7 52 5 35 Fe 0 0 8 3 89 Co 1 3 18 8 70 Ni 1 4 17 13 65 Cu 5 4 8 26 58 Zn 5 6 38 13 36 Pb 2 13 58 6 21 P 1 1 12 27 60 B. Strvmonikos a r o u D Element E £ £B±S B (%) (%) (%) (%) (%) Li 2 1 17 11 67 K 7 1 1 2 89 Be 0 1 26 9 63 Mg 8 3 13 6 70 Ca 11 31 34 4 19 Sr 13 9 27 3 49 Ba 2 2 3 2 92 Al 0 0 1 4 95 La 0 3 15 17 64 V 0 0 10 3 87 Cr 0 0 4 5 92 Mn 1 8 40 6 45 Fe 0 0 7 2 90 2 2 9 Table 6.1 (continued) Co 0 3 19 7 71 Ni 0 1 15 10 74 Cu 0 1 9 22 68 Zn 2 3 20 13 63 Pb 1 6 39 14 41 P 1 0 5 27 68 C. Kavala arouo Element E Q £K QB±£ B (%)(%)(%) (%) (%) Li 3 3 22 11 62 K 5 1 1 2 91 Be 1 1 24 7 67 Mg 8 5 26 4 56 Ca 10 18 29 2 41 Sr 7 12 36 2 43 Ba 1 1 4 3 91 Al 0 0 1 4 95 La 0 2 23 10 65 V 0 0 17 5 78 Cr 1 4 16 12 62 Mn 1 4 31 4 60 Fe 0 0 11 6 83 Co 1 7 19 10 63 Ni 1 5 17 12 65 Cu 0 3 9 24 63 Zn 2 5 25 17 51 Pb 1 7 33 7 52 P 1 1 25 19 54 D. Samothraki plateau Element E Q Q i OR-fS B (%) (%) (%) (%) (%) Li 3 2 16 9 70 K 8 1 2 2 87 Be 1 3 36 12 48 Mg 8 3 22 6 61 Ca 8 8 40 3 40 Sr 8 5 41 2 44 Ba 2 4 8 5 82 Al 0 0 1 4 94 2 3 0 Table 6.1 (continued) La 0 2 31 14 53 V 0 0 14 4 82 Cr 0 1 7 9 83 Mn 1 5 43 5 47 Fe 0 0 11 4 86 Co 1 4 23 8 64 Ni 0 3 21 11 65 Cu 5 2 10 28 55 Zn 2 4 28 15 51 Pb 1 8 38 8 43 P 1 1 16 20 63 E. Onshore areas Element E Q QR+S B (%) (%) (%) (%) (%) Li 1 0 10 8 81 K 2 0 1 1 96 Be 0 2 24 9 65 Mg 8 2 20 6 64 Ca 11 21 28 2 37 Sr 7 5 14 2 73 Ba 3 1 3 2 91 Al 0 0 1 3 96 La 0 0 11 9 79 V 0 0 8 3 89 Cr 0 1 6 7 86 Mn 2 6 41 4 46 Fe 0 0 8 4 89 Co 1 3 13 4 79 Ni 0 1 12 10 77 Cu 1 3 8 17 71 Zn 2 5 23 21 50 Pb 2 9 22 5 62 P 0 0 10 29 61 F. Means s? 9 Element £ Q QB±3 B (%) (%) (%) <%) I Li 3 2 16 10 70 K 7 1 1 2 89 Be 0 2 27 9 61 Mg 8 3 24 5 59 Ca 10 20 35 3 31 231 Table6.1 (continued) Sr 9 10 30 2 49 Ba 2 3 8 3 84 Al 0 0 1 4 95 La 0 3 20 11 65 V 0 0 12 4 84 Cr 0 1 8 8 81 Mn 1 6 41 5 47 Fe 0 0 9 4 87 Co 1 4 18 7 69 Ni 0 3 16 11 69 Cu 2 3 9 23 63 Zn 3 5 27 16 53 Pb 1 9 38 8 44 P 1 1 14 24 61 201 /o * 0 1 mM tlla - - - - — E n i u . ,llM, 20 % 0 1 ~ B B i « i l > c 401 !% 20 I llll- Ih L m JI i ox 40i OR Mm — Li K Bt^jGa StBqAI UV Crhnfe GoN Gj* Pb P U K BehgCaSr8aAlLnV CrHnFeCoNiCuZn PbP I STR K B Fig. 6.2 Fig. 6.3 f i9* 6.4 Average proportions of Li, K, Mg, Ca, Sr, Ba, Al, Average proportions of Li, K, Mg, Ca, Sr, Ba, Al, Average proportions of U, K, Mg, Ca, Sr, Ba, Al, V, Be, La. Cr, Mn, Fe, Co, Ni, Cu, Zn, Pb and P V. Be. La. Cr, Mn, Fe, Co, Ni, Cu, Zn, Pb and P V .B e.La, Cr, Mn, Fe, Co, Ni, Cu, Zn, Pb and P in the five fractions of selected surface samples in the five fractions of selected surface samples in the five fractions of selected surface samples from the Kavala area. 232 from the lerissos gulf. from the Strymonikos area. 2°i 20 i % o' — ______E % 0] -■■■■.______m ._____ E % 0 j -■ ■■■- - E %20 I I Q J —------BB«- — —* C Li K ^CaSrBaAlLaCr^FeCoNiOjZnPbP 1) K k H jth S r& iA L iV OMnReCoNlQjZn PbP THR UK Be^jOaSrEfcOl UVCrMnFeCo^CuZhPbP Fig. 6.5 Fig. 6.6 ° Average proportions of Li, K, Mg, Ca, Sr, Ba, Al, Average proportions of Li, K, Mg, Ca, Sr, Ba, Al, Fig. 6.7 V , B e ,L a , Cr, Mn, Fe, Co, Ni, Cu, Zn, Pb and P V, Be, La. Cr, Mn, Fe, Co.Ni, Cu, Zn, Pb and P Average proportions of Li, K, Mg, Ca, Sr, Ba, Al, in the five fractions of selected surface samples in the five fractions of selected onshore samples. V .B e .L a .C r, Mn, Fe, Co, Ni, Cu, Zn, Pb and P from the Samothraki plateau. in the five fractions, based on the element means of their estimates in the five groups of samples. Denominator is the percentage of individual element concentration in the total sample, accounted for by the content of the respective element in the fraction. Each denominator is the average of the respective denominators calculated on the basis of individual samples. 234 MJ__ Resell?. The results are described separately for each of the five fractions. 6.2.1.1 Exchangeable fraction The exchangeable fraction represents the material adsorbed on the sediment. The amounts of metals contained in this fraction are released easier than those in the remaining portion of the non-residual fraction. Ammonium acetate was used as the leaching reagent. The following sequences of solubility of the elements in NH4OAc 1M occur, one for each of the five groups of samples. Ca=Sr>K Mg Cu=Zn>Li>Ba>Pb>La=Mn=Co=Ni=P Be=AI=V=Cr=Fe=Ti 11-12% 7% 1-5% 0% Sr>Ca Mg>K Li=Ba=Zn>Mn=Pb=P Be=AI=V=Cr=Fe=La=Co=Ni=Cu=Ti STR: ------> ------> ------> ------11-13% 7-8% 1-2% 0% Ca Mg>Sr K>li>Zn>P=Mn=Co=Ni=Pb=Cr=Ba=Be AI=V=La=Fe=Cu=Ti KB: -----> ------> ------> ------10% 7-8% 1-5% 0% K=Mg=Ca=Sr Cu>Li>Ba=Zn>Be=Mn=Co=Pb=P AI=V=La=Fe=Ni=Cr=Ti THR: ------> ------> ------8 % 1-5% 0% Ca Mg>Sr Ba=K=Mn=Zn=Pb>Li=Co=Cu Fe=Be=AI=V=Cr=La=Ni=Cr=Ti B : ----- > ------> ------> ------11% 7-8% 1-3% 0% If all the surface samples subjected to partition analysis are considered as forming one set of samples, the solubility sequence is as follows: Ca>Sr>Mg>K Li=Zn>Ba=Cu>Mn=Co=Pb=P Be=AI=La=V=Cr=Ni=Fe=Ti 7-10% 1-3% 0% The above sequences make it clear that among the various elements Ca, Sr and Mg are significantly removed with the exchangeable fraction. This indicates that easily hydrolyzed Ca, Sr and Mg components are likely to be present. In addition, a slight 235 dissolution of the carbonate fraction by NH4OAc 1M at pH 8.2 may have occurred. The association of a proportion of K with the exchangeable fraction is attributed to its contribution from seasalt. A small proportion of Mn (1-2% of the total) was leached by NH4OAC. The exchangeable portions of Zn, Cu, Co, Pb, Cr and Ni represent minor fractions of their total concentrations. Only traces of Al were leached by NH 4OAc, implying that the treatment used does not affect the aluminosilicates. Comparing the five groups of samples according to the composition of their first leachate, a relative enrichment in Cu (5%), Zn (5%), and Pb ( 2%) in the lerissos group of samples (I) is found. It corresponds to a Pb (3%) and Mn (3%) enrichment in the exchangeable fraction of the onshore samples collected from the northwest coast of the lerissos gulf. These enrichments could be attributed to land sources and man-made pollution, as mixed sulphide and Mn-oxide deposits occur in the mainland that surrounds lerissos gulf, and a mining operation takes place in the coastal area of Stratoni. Among the samples subjected to partition analysis, an exceptionally large metal enrichment in the exchangeable fraction composition occurs in some of them. Thus, from the lerissos group, the samples I 46 and I 50 show an enrichment in their exchangeable metal content, the former in Cu (9%) and Zn (7%), and the latter in Cu (6%), Pb (6%) and Zn.(10%). As both samples were collected off Stratoni district, these enrichments probably reflect the pollution of this area from the onshore mining activities. Among the onshore samples, B 83 and B 89, collected from the coastline of lerissos gulf, exhibit metal enrichments in their exchangeable fraction. Sample B 83, collected from a beach area close to a fertilizer factory, shows an enrichment in exchangeable Pb (4%). Sample B 89 has enrichments in exchangeable Zn (5%) and Mn (4%). The river sediment BSTR 4, collected from the eastern distributary of Strymon, is enriched in exchangeable Zn (5%). Only one sample from the Kavala group exhibits a significant metal enrichment in its exchangeable fraction; this is KB 126 which is enriched in exchangeable Zn ( 6%). Finally, the nearshore samples THR 4 and THR 11, collected from the Samothraki plateau, show enrichments, the former in exchangeable Cu (31%) and the latter in exchangeable Zn ( 6%). These enrichments most probably reflect contamination from the mixed sulphide deposits of the Alexandroupolis area. As the amount of exchangeable metals is a good indicator of the pollution potential of a sediment, the samples mentioned above may be considered to be polluted, in a broad 236 sense, containing portions of metals easily available to the biota. 6.2,1,2 Carbonate hosted fraction This fraction represents the elements associated with carbonate minerals in the sediments. Several researchers (Gupta and Chen, 1975; Stover et al, cit. in Tessier et al, 1979; Chester and Hughes, 1967; Perhac, cit. in Tessier et al, 1979) have shown that significant concentrations of trace metals can be associated with carbonates. Ammonium acetate and acetic acid were used as the leaching reagents. The following solubility sequences of elements in this reagent occur, group by group : Ca>Sr Pb Mn>Zn=La Ba>Li=Mg=Ni=Cu>Co>Be=K>Cr=P AI=V=Fe=Ti : ------> --- > ------> ------> ------22-17% 13% 7-6% 1-5% 0% Ca Sr>Mn>Pb Mg=La=Co=Zn>Ba>li=K=Be=Ni=Cu AI=V=Fe=Cr=P=Ti STR : ----- > ------> ------> ------31% 9-6% 1-3% 0% Ca>Sr Pb=Co Mg>Ni=Zn>Cr=Mn>Cu=Li>La>K=Be=Ba=P AI=V=Fe=Ti KB: ------> ------> ------> ------18-12% 7% 1-5% 0% Ca=Pb Sr=Mn>Ba=Co=Zn>Be=Mg=Ni>Cu=La=Li>K=Cr=P AI=V=Fe=Ti THR: ------> ------> ------8 % 1-5% 0% Ca Pb Mn>Zn=Sr>Cu=Co>Be=Mg>Ba=Cr=Ni AI=V=Fe=P=Li=K=La=Ti B: ----- > — > ------> ------21% 9% 1-6% 0% Considering all the surface samples subjected to partition analysis as forming one group, the following solubility sequence occurs: Ca>Sr Pb>Mn>Zn>Co>Mg=Ba=La=Ni=Cu>Li=Be>K=Cr=P AI=V=Fe=Ti ------> ------> ------10- 20% 1-9% 0% The above sequences confirm that biogenic carbonate remains comprise the phase mainly attacked by NaOAc / HOAc at pH 5. 237 Trace metals can be removed from seawater by biogenic processes according to three main ways: i. incorporation of trace metals into the soft parts of the organisms; ii. incorporation of trace metals into the hard skeletal parts of the organisms; and iii. association with body processes, such as digestive and excretory functions (Riley and Chester, 1971). The aforementioned element solubility sequences show that: 1. A significant proportion of Pb was leached by NaOAc / HOAc, implying that a part of the total Pb is supplied biogenically. An association of Pb with the acetic acid leachate has been also noted by Varnavas (1979) and Shearme et al(1983). 2. Still significant proportions - although lower than those of Pb - of Mn, Zn, Co, Ni and Cu were leached by NaOAc / HOAc. This also indicates a biogenic contribution of these metals - however, to a lesser extent than for Pb. 3. Only traces of Fe were found in the second leachate, showing that Fe is not likely to be associated with the biogenic matter, and also that Fe-oxides have not been attacked. 4. Aluminium was not leached by NaOAc / HOAc which implies a lack of attack of the aluminosilicate phase. It was noticed that in some of the samples subjected to partition analysis low proportions of the total Ca and Sr were leached by NaOAc / HOAc corresponding with high Ca and Sr proportions found in the third leachate (leached by NH 2OH-HCI). As low as 1-2% of the total Ca and Sr were removed in the second leachate of samples B 12, I 5, STR 4, THR 31 and THR 122. By contrast, exceptionally high proportions of Ca (48-80%) and Sr (60-74%) were found in the third leachate of the above samples. As the NaOAc / HOAc reagent is used to dissolve the carbonates and a high Ca and Sr content is unlikely to occur in the reducible fraction of the sediments, the anomalous Ca and Sr distribution among the second and third leachate of these samples indicates an incomplete dissolution of the sediment carbonates by NaOAc / HOAc. These samples were collected from areas enriched in carbonates and therefore high 238 proportions of Ca were expected to be found in the NaOAc/Hoac leach. In a number of samples, the concentrations of Ca occurring as carbonates were compared with the total Ca content of the second and third leachate (carbonate content being measured by gas chromatography, see also Chapter II and Appendix IV). This comparison (Table 6.2) shows that the sum of Ca associated with the second and third leachate represents the total amount of Ca present as carbonates. Table 6.2 : Calcium content determined by gas chromatography and total Ca content associated with the second and third leachate. Sample I.D. Ca1 Ca2 (%) (%) 1 5 13 12 1 22 9 7 1 38 3 3 1 46 6 3 1 50 5 5 KB 2 0 0 KB 100 8 6 THR 1 1 1 THR 11 24 15 THR 31 11 9 THR 84 1 0 THR 122 8 9 Ca content determined by gas chromatography; 2: Ca content associatedwith the second and third leachate. According to the data obtained, it is evident that the NaOAc / HOAc leach is not completely dissolving the carbonates; change of pH during the NaOAc / HOAc attack could be responsible for the incomplete attack on the carbonates. Comparison of various samples, regarding the anomalous Ca distribution between the second and third leachate, (Table 6.3, Fig. 6.8 ) shows that the dissolution of carbonates by NaOAc / HOAc is less complete in the samples richest in carbonates. 239 Table 6.3 : Total calcium carbonate content of selected samples, and proportions of the total Ca content of the same samples associated with the second and the third leachate. Sample I.D. CaCOo1 Ca2 Ca3 (%) (%) (%) I 5 61.6 1 48 I 22 22.5 21 60 I 38 8.2 40 24 I 46 14.8 21 36 I 50 11.7 27 44 KB 100 19.5 25 48 THR 11 60.0 10 54 THR 31 28.1 2 74 total CaCOg content determined by gas chromatography; 2 : proportion of the total Ca associated with the second leachate ; 2 proportion of the total Ca associated with the third leachate. Fig. 6.8 Plot of the total carbonate content (%) of selected samples versus the corresponding proportions of the total Ca content (%) leached by NaOAc/HOAc. 240 Finally, it was noticed, that the samples in which this anomalous Ca distribution occurs belong to the sand (coarse, medium or fine) type of sediment. By contrast, this phenomenon was not observed in the samples of a clayey or silty sediment type. Evidently, even though the samples were ground to a fine powder before the partition analysis treatment, the sediment type could have an effect on its dissolution characteristics. The nature of the carbonate is also likely to be important. 6.2.1.3 Reducible fraction Hydroxylamine hydrochloride was used for the removal of the reducible fraction. The major phases attacked are reducible ferromanganese oxides, some amorphous iron oxides and other reducible material (Chester and Hughes, 1967; Cronan, 1976; Moorby and Cronan, 1981). It is well established (Jenne, 1968) that iron and manganese oxides occur as nodules, concretions, cement between particles, or simply as coatings on particles. These oxides are excellent scavengers of trace metals (Cronan, 1980). Under anoxic conditions, Fe,Mn-oxides are unstable and they, therefore, tend to dissolve releasing the trace metals bound to them into solution. The following solubility sequences of the various elements occur, one for each group of samples: Pb>Mn Ca>Mg>Zn>Si>Be>Ba>La Co>Ni>P=V>Li>Cr>Fe=Cu K>AI Ti ------> ------> ------> ---- > — 50-60% 20-45% 8-18% 1-2% 0% Mn>Pb Ca>Sr>Be>Zn Co>Li>La=Ni>Mg V>Cu>Fe>P>Cr>Ba>AI=K Ti STR: ------> ------> ------> ------> — 39-40% 20-35% 13-19% 1-10% 0% Sr=Pb>Mn Ca>Mg>Zn=P>Be>La>U Co>Ni=V>Cr>Fe Cu>Ba>K=AI Ti KB : ------> ------> ------> ------> — 30-36% 22-29% 11-19% 1-9% 0% Mn>Si>Ca>Pb>Be>La Zn>Co>Mg>Ni P=Li>V>Fe>Cu Ba>Cr>K>AI Ti THR: ------> ------> — ------> ------> — 30-45% 20-30% 10-16% 1-8% 0% Mn Ca>Be>Zn>Pb>Mg Sr>Co>Ni>La>Li=P V=Cu=Fe>CR>Ba>AI=K Ti B: --- > ------> ------> ------> — 41% 20-28% 10-14% 1- 8 % 0% 241 Considering ail the surface samples subjected to partition analysis as forming one set of samples, the following solubility sequence occurs: Mn>Pb Ca>Sr>Zn=Be>Mg La>Co>Ni=Li>P>V Fe=Cu>Cr=Ba K=AI Ti ------> ------> ------> ------> ------> — 38-41% ' 24-35% 12-20% 8-9% 1% 0% It has already been mentioned that the NaOAc / HOAc attack did not successfully removed all the sedimentary carbonates. Therefore, the relatively high proportions of total Ca and Sr that occur in the third leachate can be attributed to the remaining carbonates, as these elements are not known to form reducible phases in marine sediments. A portion of the Mg leached by NH 2OH-HCI could also be associated with the carbonates left over from the incomplete dissolution by NaOAc / HOAc. It is also possible that, part of the Pb and Zn proportions found in the third leachate is associated with the carbonates. The large proportions of the total Mn, Fe, Pb, Zn, Co, Ni and Cu removed by NH 2OH*HCI indicate the presence of Fe.Mn-oxides in the sediments and demonstrate the scavenging ability of these oxides for metals. Although this scavenging ability is well established, the suggestion that the trace metals found in the third leachate are associated with the Fe.Mn-oxides rather than with the remaining carbonates is strengthened by the following: 1. The correlation matrices of the data, which represent the bulk geochemistry of the marine sedirrjents, showed that the correlation coefficients of Ca and each of the elements Mn, Fe, Zn, Co, Ni and Cu are very low, in all groups of samples. By contrast, the correlation coefficient of the Ca-Sr pair is high, in all groups of samples. 2. The dendograms, obtained when cluster analysis was applied to the geochemical data produced by the whole sample analysis, showed that in all groups of samples a megabranch containing only Ca and Sr was formed. 3. The results of the experimental w ork-w hich concerns attack on the carbonates by acetic acid under various conditions and is described in Chapter 2 - suggest lack of trace metal association with sediment carbonates. The latter mainly contain Ca and Sr. Comparing the five groups of samples according to the composition of their third leachate, an enrichment in Mn (52%), Pb (58%) and Zn (38%) is found in the 242 lerissos group (I). As high as 44% of the total Zn, and 48 to 62% of the total Pb occur in the reducible fraction of the samples collected off Stratoni; this could be related to the presence of mixed sulphide mineralization in the mainland. An exceptional metal enrichment in the third leachate (76% Pb, 65% Ni, 62% Zn and 65% Co) was also found in the nearshore sample THR 1, collected from the northeast part of the Samothraki plateau; this, too, is likely to be attributed to the sulphide deposits that occur on the corresponding part of the mainland. In general, comparing the proportions of Mn, Fe, Co, Ni, Zn, Pb and V present in each of the non-residual geochemical fractions, the most significant are associated with the reducible phase. 6.2.1.4 Organic and Sulphide bound fraction This fraction represents the elements bound to various forms of organic matter and those present in sulphide minerals. The organic matter in sediments usually comes from the following sources: a. debris remaining after the decomposition of marine organisms; b. metabolic end products from natural biota; c. industrial and domestic wastes; d. agricultural and mining runoffs; and e. accidental spillages of organic substances. Furthermore, numerous forms of intermediates 1 produced by interactions between various decomposed products of living organisms may occur (Yen and Tang, 1977). Although, on average, marine sediments have a rather low organic content, organic molecules are an important part of them, and they are also important in terms of their environment. This importance is illustrated as follows: 1 intermediate is a compound used in an intermediate step in the manufacture of a final product by chemical synthesis. 243 i. Organic molecules in marine sediments possess reactive functional groups. Thus, stable linkages can be formed by the coordination of inorganic cations, such as heavy metals, to these sites. In this manner, marine sediments can accumulate metals. ii. In a following step, the metal complexes and chelates, formed as described in (i) above, could further coordinate inorganic anions, such as sulphate, chloride and phosphate, and could be easily attached or detached under variable redox conditions. In this manner, the transport or the migration of nutrition - important anions could be regulated in sediments. Hydrogen peroxide was used for the leaching of the organic compounds and sulphide minerals, together with elements associated with the latter. The solubility sequences of the various elements in H 202 are presented below, one for each group of samples: P>Cu Ni=Zn Li=Be>Cr=Co>La>Pb>Mn=V=Mg>Ba>Fe=AI=Ca=Sr>K Ti I : ------> ------> ------> ------26-27% 13% 1-9% 0% P>Cu La>Pb>Zn>Li>Ni Be>Co>Mn=Mg>Cr>Ca=AI>Sr=V>Ba=Fe=K Ti STR: ------> ------> ------> — 22-27% 10-17% 1-9% 0% Cu P>Zn>Cr=Ni>Li>Co=La Be=Pb>Fe>V>Mn=Mg=AI>Ba>K>Ca=Sr Ti KB: ----- > ------> ------> — 24% 10-12% 1-7% 0% Cu>P Zn>La>Be>Ni Li=Cr>Pb=Co>Mg>Ba=Mn>Fe=V=AI>Ca>Sr=K Ti THR: ------> ------> ------> — 20-28% 11-15% 1-9% 0% P>Zn>Cu Ni>Be=La>Li>Cr>Mg Mn=Fe=Co>Pb>AI-Cr>Ca=Sr=Ba=K Ti B : ------> ------> ------> — 17-29% 6-10% 1-4% 0% Considering all the surface samples subjected to partition analysis as forming one group, the following solubility sequence occurs: P>Cu Zn>Ni=La>Li Be>Cr=Pb>Co>Mn=Mg>AI=Fe=sV>Ba=Ca>Sr=K Ti ------> ------> ------> — 23-24% 10-16% 2-9% 0% 244 The above sequences show that Cu and P have a strong association with the H 20 2 soluble fraction. This is attributed to the occurrence of these elements largely in organic matter. Nevertheless, Cu in the fourth leachate may not be associated only with organic materials but may also be related to sulphides. This possibility is a more likely situation in the lerissos gulf sediments, on account of their contamination by material derived from the mineralization of the mainland. Conversely, the richest in organic carbon sediments of the study area are those of the Samothraki plateau. Therefore, it is easier to demonstrate the affinity of Cu for organic matter with reference to this area. This affinity is depicted when the organic carbon content of the Samothraki plateau sediments and the proportion of Cu in their fourth fraction are compared (Table 6.4, Fig. 6.9). It is shown thereby, that the amount of Cu associated with the H20 2 soluble fraction increases with an increase in the organic carbon content of the samples (Sakellariadou, 1986). The strong affinity of Cu for organic matter results from the formation of chelate^ compounds between Cu and many naturally occurring organic materials (Fraser, 1961). Table 6.4 : Organic carbon content of selected samples and proportions of total Cu in the fourth leachate of the same samples. Sample I.D. org. C 1 Cu 2 (%) (%) THR 84 0.08 10 THR 1 1.44 16 THR 122 4.07 34 THR 31 6.41 35 THR 11 6.76 50 1: org. C content of total sample; 2: proportion of the total Cu content associated with the H20 2 soluble farction. 1 Chelate compounds are substances that contain a metal bonded directly to nitrogen, sulphur or oxygen atoms (Fraser, 1961). 245 Fig. 6.9 Plot of the proportions of the total Cu (%) leached by H 20 2 in selected samples versus the organic carbon content (%) of the same samples. The concentrations of Al In the fourth leachate show that relatively little of the total Al (up to 4%) was leached by H 20 2 ; therefore, there is a minimal attack of the major aluminosilicate phases by this reagent. A relatively exceptional enrichment in the metal content of the fourth fraction occurs in some of the sediments subjected to partition analysis. The following enrichments were recorded: 1. In samples THR 11, THR 31, THR 122 and KB 100, with organic carbon contents of 6.76, 6.41, 4.07 and 4.05% respectively, it was found that as high as 50, 35, 34 and 41% of the total Cu was leached by H 20 2. The high Cu proportions associated with the fourth leachate of these samples are attributed to their high organic carbon content and the affinity of Cu ions for the organic matter present in these samples. 2. Among the samples collected from lerissos gulf, it was noticed that 36 and 32% of the total Cu of samples I 5 (with 0.5 % organic carbon) and I 50 (with 0.1% organic carbon) respectively, and 13% of the total Co and 18% of the total Ni of sample I 46 (with 0.14% organic carbon) occur in the H 20 2 soluble fraction. It is suggested that these enrichments in the fourth leachate are associated with sulphides, as the organic 246 carbon content of these samples is very low. This suggestion is supported by the presence of Cu-bearing porphyry stocks and mixed sulphide deposits in the mainland adjacent to the gulf. 3. Up to 46% of the total Zn occurs in the fourth fraction of sample B 89. This enrichment can be attributed to traces of sulphides, as mixed sulphide mineralization occurs in the drainage basin of the river from which this sample was collected. 6.2.1.5 Residual fraction This fraction contains most of the elements hosted in the lattice structure of aluminosilicate minerals. These elements are not expected to be released rapidly in the secondary environment under the conditions normally encountered in nature. The various elements occur in the residual fraction in the following sequences; each group of samples is considered separately : Ti>AI>Fe>V>K>Cr Li=Co>La=Ni=Ba>Be>P Cu>Mg>Zn>Mn>Sr Pb>Ca I: ------> ------> ------> ------>80% 60-70% 34-58% 20-21% Ti>AI>Ba=Ci>Fe>K>V Ni>Co>Mg>Cu=P>Li>l_a>Be=Zn Sr>Mn>Pb Ca STR: ------> ------> ------> — >80% 63-74% 41-49% 19% Ti>AI>K=Ba Fe>V>Be>La=Ni>Cu=Co>Cr=Li>Mn Mg>P>Pb>Zn>Sr>Ca KB: ------> ------> ------>90% 60-83% 41-56% Ti>AI>K>Fe>Cr>Ba=V Li>Ni>Co>P>Mg Cu>La>Zn>Be>Mn>Sr>Pb>Ca THR: ------> — :------> ------>80% 61-70% 40-55% Ti>AI=K>Ba>V=Fe>Cr>Li La=Co>Ni>Sr>Cuu>Be>Mg>Pb>P Zn>Mn>Ca B: ------> ------> ------>80% 61-79% 35-70% Considering all the surface samples subjected to partition analysis as forming one set of samples, the following sequence exists: Ti>AI K>Fe>Ba=V>Cr Li>Co=Ni>La>Cu>Be=P Mg>Zn>Sr>Mn>Pb>Ca > ------> ------> >95% 81-89% 61-70% 31-59% 247 The above sequences show that the major part of Al, K, Fe, Ba, V, Cr, Li, Co, Ni, La, Cu, Be, P, Mg and Zn in the sediments remains in their residual fraction. Relatively, lower than usual proportion of Ca is associated with the insoluble solid of sample KB 100 (12% Ca), collected from the channel to the north of Thassos island. This could suggest that most of the Ca is supplied to this area in the form of carbonates. In accord with this suggestion are both the carbonate enrichment at this area (Fig. 4.8) and the presence of marbles on the adjacent to the channel part of Thassos island. A high proportion of La (95%) was found to be associated with the residual fraction of the coastal sample KB 126; this could be explained by the possibility of La-derivation from monazite which according to Perissoratis et al (in press) may occur in the acidic intrusive rocks of the corresponding part of the mainland. 248 6.3 Buried sediments Two cores (I 14 and THR 22) were subjected to partition analysis. The procedure used was the same as the one applied to surface samples. In order to assess variations in chemical partitioning with depth, the percentages of each element associated with each fraction were plotted against depth for each core (Fig. 6.10 and Fig.6.11). In addition, the partitioning of the elements within the five fractions, for each depth separately, is presented in Fig. VI.2 and Fig. VI.3, in Appendix VI, on the basis of the elements proportions as a percentage of the whole. 249 DEPTH (cm) DEPTH (cm) in in o o o oin 0 5 2 DEPTH (cm) DEPTH lea) 100 50 150 50 6.10 70 ie fractions. five the of each with Passociated Pband Cu,Zn, Ni, Co, Fe, Mn, Cr, La, Al, Core 0 10 90 114: downcore plots of the proportions of Li, K, Mg, Ca, Sr, Ba.V.Be, Ba.V.Be, Sr, Ca, Mg, K, Li, of proportions the of plots downcore 65 OE H QR+ s IH CORE OE 14 I CORE 0 15 90 % 50 5 0 0 9 10 5 95 &5 100 90 100 90 85 251 252 CORE THR 22 % CORE THR 22 f > Q100- UJ (cm) 0 200-1 0 10 0 20 OE H 22 THR CORE 20 OE H 22 THR CORE m o 10 o % 20 W) OX R+ S OR + 20 1 0 10 0 10 0 Fe 253 K b f'- CORE THR22 70 100 50 75 100 10 55 100 70 100 50 75 100 % Fig. 6.11 Core THR 22 : downcore plots of the proportions of Li, K, Mg, Ca, Sr,V, Be, Ba, At, La, Cr, Mn, Fe, Co, Ni, Cu, Zn, Pb and P associated with each of the five fractions. 255 6.3.1 Results 6.3.1.1 Core I 14 The solubility sequences of the various elements in the five fractions are listed below, according to depth: i. Exchangeable fraction Sr Mg>Ca>K Cu=Li>Mn=Pb=Zn Be=Ba=AI=La=V=Cr=Fe=Co=Ni=P=Ti 3cm : ----- > ------> ------> ------17% 8-11% 1-2% 0% Sr>Mg>Ca=K Li=Mn=P>Ba=Cu=La Be=AI=V=Cr=Fe=Co=Ni=Zn=Pb=Ti 16cm: ------> ------> ------9-14% 1-2% 0% Sn>Mg>K>Ca Mn>Li=Pb=La Be=Ba=AI=V=Cr=Fe=Co=Ni=Cu=Zn=P=Ti 41cm: ------> ------> ------8 - 12% 1- 2% 0% Sr>K=Mg=Ca Li=Mn=Pb>Ba Be=AI=La=V=Cr=Fe=Co=Ni=Cu=Zn=P=Ti 7 1 c m : ------> ------> ------9-15% 1-2% 0% Sr>Mg>K>Ca Li=Mn>P Be=Ba=AI=La=V=Fe=Co=Ni=Cu=Zn=Pb=Ti 101cm: ------> ------> ------7-12% 1-2% 0% Sr>K>Mg>Ca Mn>Ni=Li=Ba=P Be=AI=La=V=Cr=Fe=Co=Cu=Zn=Pb=Ti 131cm: ------> ------> ------5-9% 1-2% 0% Sr>K=Mg>Ca Li>Mn Be=Ba=AI=La=V=Cr=Fe=Co=Ni=Cu=Zn=Pb=P=Ti 161cm: ------> ------> ------5-10% 1-2% 0% ii. Carbonate hosted fraction Ca>Sr Pb>Mn>La Mg>Li=Ni=Cu=Zn>Ba=K Be=AI=V=Cr=Fe=Co=P=Ti 3cm: ------> ------> ------> ------31-44% 7-13% 1-4% 0% 256 Ca=Sr Mg=La>Li=K=Mn=P Be=Ba=AI=V=Cr=Fe=Co=Ni=Cu=Zn=Pb=Ti 16cm: ------> ------> ------5% 1-2% 0% Ca=Sr La=Mg>Li=K=Mn Be=Ba=AI=V=Cr=Fe=Co=Ni=Cu=Zn=Pb=P=Ti 41cm: ------> ------> ------5% 1-2% 0% Ca=Sr Mg>Li=K=Mn=Pb Be=Ba=AI=La=V=Cr=Fe=Co=Ni=Cu=Zn=P=Ti 71cm: ------> ------> ------5% 1-2% 0% Ca=Sr Mg>Li=K=Mn=P=Zn Be=Ba=AI=La=V=Cr=Fe=Co=Ni=Cu=Pb=Ti 101cm: ------> ------> ------5% 1-2% 0% Ca>Sr Mn Mg=La>Ni>Li=K=Ba=P Be=AI=V=Cr=Fe=Co=Cu=Zn=Pb=Ti 131cm:------> ----- > ------> ------29-34% 15% 1-4% 0% Ca>Sr Mn Mg=La=Pb>P>Li=K=V=Ni=Ba Be=AI=Fe=Co=Cu=Zn=Cr=Ti 161cm:------> ----- > ------> ------26-31% 11% 1-4% 0% iii. Reducible fraction Mn>Pb Zn>Be=Ca>Ni>Co>Li Sr=V=La=Mg>Fe=Cu>Cr=P>K=Ba AI=Ti 3cm: ------> ------— > ------> ------24-29% 9-18% 1-7% 0% Pb>Zn Ca>Mn>Sr>P Be>Mg=La=Ni>Co>Li=V=Cr=Fe=Cu K=Ba=AI=Ti 16cm: ------> ------> ------> ------30-43% 8-15% 1-4% 0% Ca=Pb>Mn>Sr=Zn Be>La=Co>Li=Mg>Ni>V Cu=P=Fe>Cr>K=Ba AI=Ti 41cm: ------> ------> ------> ------20-30% 6-13% 1-3% 0% Ca>Pb>Mn>Sr>Zn Ni=Be>Li=Mg=La=Co>P=V=Fe>Cr=Cu K=Ba=AI=Ti 71cm: ------> ------> 9-14% 1-4% 0% Ca>Pb=Mn>Sr Be>Zn>la>Co>Li=Mg>Ni V>P>Fe>Cr>Ba>Cu>K AI=Ti 101cm: ------> ------> ------> 42-62% 13-29% 1-9% 0% 257 Ca>Mn=Pfc»Sr>Be Co>Zn>Lj=La=Ni>Mg>V P>Cr=Fe>Cu>Ba=K AI=Ti 131cm: ------> ------> ------> ------24-35% 8-17% 1-6% 0% Ca>Mn>Pb>Sn>Be Zn>Mg=La=Co>Ni=P>Li V>Fe=Sr>K=Ba=Cu AI=Ti 161cm: ------> ------> ------> ------28-44% 11-16% 7% 0% iv. Organic and sulphide bound fraction Cu La>P>Li>Ni=Zn>Pb>Be Co>Mn>Mg>Cr>Ca=AI>Sr=Ba=V>Fe=K Ti 3 cm: ----- > ------> ------> — 31% 10-18% 2-9% 0% P>Cu La>Li>Zn>Pb>Co=Ni=Be Mg=Cr=Mn>AI=V=Ca>Sr>Ba>K=Fe Ti 16cm :------> ------> ------> — 28-39% 13-21% 2-7% 0% La>Cu>P>Zn Ni=Co>Pb>Cr=Be=Mg>Li>Mn Ca=Sr=AI=V>K=Ba=Fe Ti 41cm: ------> ------> ------> — 17-22% 7-11% 2-4% 0% Cu>P>La=Zn Li=Ni>Pb>Co>Cr>Be=Mg>Mn>AI Ca=Sr=V>Fe=K=Ba Ti 71cm: > — 19-26% 7-15% 4% 0% Cu>La>Li Pb=Be>Zn>Ni>Ca=Mn=Co>Sr Mg=Cr=P>Ba>K=AI=V=Fe Ti 101cm: ------> ------> ------> — 10-13% 5-9% 1-3% 0% Cu Li=P>La>Zn>Be=Pb Ni>Co>Ca=Mn>Sr>Mg Cr>V=Fe=Ba=K AI=Ti 131cm: -----> ------> ------> ------> ------26% 10-13% 4-8% 1-3% 0% Cu Li>La>Zn Ni>P>Co=Pb>Be>Mg=Ca=Cr=Mn Sr>V=Fe=AI>K=Ba Ti 161cm:----- > ------> ------> ------> — 21% 10-13% 4-8% 1-3% 0% v. Residual fraction Ti>Ai>Ba>Cr=Fe>V>K>P>Co Be>Li>Ni>Mg>La>Zn Cu>Pb=Mn>Sr Ca 3cm: ------> ------> ------> ----- >82% 67-76% 42-59% 28% 258 Ti>Fe>Ba>AI>V>Cr>k>Co>Ni>Be Mg>Mn>Li>La>Cu>Ca=Sr Zn>P>Pb 16cm: ------> ------> ------>83% 67-79% 40-53% Ti>Ba>AI>Fe>V>Cr>K>Ni>Co Be=P>Cu>Mg>La>Mn>Zn Sr=Pb>Ca 41cm: ------> ------> ------>80% 63-78% 53-59% Ti>Ba=Fe>V>AI>Cr>K>Be>Co>Ni Li>La=Mn>Mg>P>Cu=Zn>Pb>Ca>Sr 71cm: ------> ------>81% 67-79% Ti>AI>Ba>Fe>Cr>V>K>Cu>V>Ni Co>Li>Mg>La>Zn>Be Mn>Pb>Sr Ca 101cm: ------> ------> ------> ------>80% 62-78% 37-45% 21% Ti>AI>Ba>Cr=Fe>K>V P>Ni>Mg=Co>Zn>Cu>Li=La>Be Pb>Mn Sr>Ca 131cm: ------> ------> ------> ------>89% 66-79% 45-58% 20-29% Ti>AI>Ba>Fe>Cr>V>K>Co Ni=P>Cu>Zn>Li>Mg=La>Be Pb>Mn Sr>Ca 161cm: ------> ------> ------> ------>81% 67-79% 48-57% 15-29% The data obtained show some interesting variations down the core; specific items are as follows: 1. At 16 and 71 cm depths the proportions of Co and Ni associated with the reducible fraction are lower than those at the other depths. 2. At 71 cm depth the amounts of Pb and Zn bound to the reducible fraction are lower than those at the other depths. 3. At 101 cm depth, as high as 47% of total Pb occurs in the reducible fraction; at this depth, the highest content of reducible Fe (6%) was also found. 4. A decrease in the non-residual Mn, Fe, Ni and Cu of 56, 57, 38 and 26%, respectively, occurs between 3 and 16 cm depth. It corresponds to a large decrease in the total concentrations of these metals. 259 6.3.1.2 Core THR 22 The solubility sequences of the various elements in the five fractions are listed below, according to depth: L__ Exchangeable fraction Sr>Ca>Mg K>Zn Li=Mn>Ba=Cu=P Be=AI=La=V=Cr=Fe=Co=Ni=Pb=Ti 2cm: ------> ------> ------> ------13-16% 7-8% 1-2% 0% Ca>Sr Mg>K Mn>Li=Ba=Cu=Zn=P Be=AI=La=V=Cr=Fe=Co=Ni=Pb=Ti 24cm :------> ------> ------> ------2 1 - 24% 10-15% 1-3% 0% Ca>Sr Mg>K Ba>Zn>Mn>Li=Ni=Cu Be=AI=La=V=Cr=Fe=Co=Pb=P=Ti 64cm :------> ------> ------> ------20-23% 8-14% 1-5% 0% Ca>Sr Mg>K Mn=Ba>Li=Cu=Zn=P Be=AI=La=V=Cr=Fe=Co=Ni=Pb=Ti 94cm : ------> ------> ------> ------22- 27% 8-11% 1-2% 0% Ca>Sr Mg>K Ba=Mn>Li=La=Cu Be=AI=V=Cr=Fe=Co=Ni=Zn=Pb=P=Ti 122cm:------> ------> ------> ------18-28% 10-13% 1-2% 0% Ca>Sr Mg>K Mn>Li=Ba=P Be=AI=La=V=Cr=Fe=Co=Ni=Cu=Zn=Pb=Ti 149cm: ------> ------> ------> ------18-21% 9-11% 1-2% 0% Ca>Sr=Mg>K Mn>Li=Ba=Ni Be=AI=La=V=Cr=Fe=Co=Cu=Zn=Pb=P=Ti 176cm: ------> ------> —------9-13% 1-2% 0% ii. Carbonate hosted fraction Ca Sr=Mn Zn La=Mg=Ba=Pb>Cu=Li=K Be=AI=V=Cr=Fe=Co=Ni=P=Ti 2cm: -----> ------> ------> ------> ------43% 16% 7% 1-3% 0% Ca Sr Mg=Ba=Mn>K Li=Be=AI=La=V=Cr=Fe=Co=Ni=Cu=Zn=Pb=P=Ti 24 cm : -----> — > ------> ------14% 6% 1-2% 0% 260 Ca Sr Mn>Mg=Ba>K Li=Be=AI=La=V=Cr=Fe=Co=Ni=Cu=Zn=Pb=P=Ti 64cm :----- > ------> ------> ------16% 7% 1-3% 0% Ca Mn Sr Pb Zn>Mg>Ni>Li=K=Ba Be=AI=La=V=Cr=Fe=Co=Cu=P=Ti 44% 22% 12% 9% 1-4% 0% Ca Mn Sr Mg>Ni=Zn>Li=K=Ba=Cu Be=AI=La=V=Cr=Fe=Co=Pb=P=Ti 122cm: — > — > —> ------> ------2 2% 12% 7% 1 -4% 0% Ca Mn Sr Mg>Zn>Ni=La>Li=K=Ba=Cu Be=AI=V=Cr=Fe=Co=Pb=P=Ti 149cm: ----- > ----- > — > ------> ------41% 21% 11% 1-4% 0% Ca Sr>Mg>K=Mn Be=Li=Ba=AI=La=V=Cr=Fe=Co=Ni=Cu=Zn=Pb=P=Ti 176cm: — > ------> ------8% 1-3% 0% iii. Reducible fraction Pb>Be>Zn>Mn Ca>Li>Co>Ni=V La>Fe=Sr>P Ba=Cr>Mg Cu>AI=K Ti 2 c m : ------> ------> ------> ------> ------> — 26-36% 15-21% 9-12% 5-6% 1-2% 0% Pb>Ca=Mn>Be>Zn La>Co>Ni=V=Li>Sr=P Fe>Mg>Cr>Ba K=Cu>AI Ti 24cm: ------> ------> ------> ------> — 29-35% 10-15% 4-8% 1-2% 0% Mn Pb>Be=Ca>Zn Ni>Co>V=Fe Li=La>Sr>Ba=Cr=Mg=P Cu>K AI=Ti 64cm: ----- > ------r----- > ------> ------> ------> ------24% 12-17% 7-10% 3-5% 1 - 2 % 0% Pb>Be Mn>Zn>V>Co=Li Ni>Ca=Fe La=P>Cr>Mg=Sr Ba>Cu>K=AI Ti 94 cm: ------> ------> ------> ------> ------> — 28-30% 12-19% 9-10% 4-6% 1-3% 0% Be Ni>Li=V=Mn>Fe=Co=Zn=P=Ca>Mg>Sr>Cr K=Ba=AI=La=Cu=Pb=Ti 122cm: — > ------> ------6% 1-4% 0% Be Mn>Li=Co>V=Ni=Zn=Pb P>Ca=Fe>La=Cr>Mg=Sr=AI>K Ba=Cu=Ti 149cm: -----> ------> ------> ------12% 6- 8 % 1-5%‘ 0% 261 Be>Mn Pb=Co>Ni>V>Li>Fe=Zn=P Ca>Mg=Cr Sr=La>K=Cu>Ba AI=Ti 176cm:------> ------> ------> ------> ------24-33% 11-17% 5-8% 1-3% 0% iv. Organic and sulphide bound fraction Cu P>La Zn>Pfc»Be>Ni Li>Co>Mn=Ba=Ca>Cr Sr>Mg=V>AI>Fe>K Ti 2cm: -----> ------> ------> ------> ------> — 42% 28-34% 12-19% 7-10% 1-6% 0% La>Cu P>Zn>Pb=Be>Ni Li>Mn>Ca>Co Sr>Cr>Mg>Ba=AI=V>Fe>K Ti 24cm :------> ------> ------> ------> — 3 0 - 34% 12-17% 8-10% 1-5% 0% Cu>P>La Zn>Pb>Ni Li>Be>Co=Mn Ca=Ba>Mg>Al=Cr Sr>V>Fe=K Ti 64cm :------> ------> ------> ------> ------> — 3 1 - 39% 17-21% 12-15% 8-10% 2-5% 0% La>Cu=P Zn>Pb>Be Mn=Ni=Li>Ca=Co AI=Mg=V=Cr>Sr>Ba>K Fe=Ti 94cm : ------> ------> ------=------> ------> ------26-30% 13-17% 7-8% 1-4% 0% Pb La>Ni>Li=Cu K=Be=Mg=Ca=Sr=Ba=AI=V=Cr=Mn=Fe=Co=Zn=P=Ti 122cm:----- > ------> ------10% 1-3% 0% La>Cu Zn>Li=Be=Ni>Pb>Co Mn>Ca>Cr=P>Mg=Sr>AI=V=Fe>K=Ba Ti 149cm: ------> ------> ------:------> — 19-26% 7-10% 1-6% 0% P>Cu Be>Li=Ni=Zn>Co=La Mn>Cr=Mg>V>Fe=Ca>K=Sr=AI=Pb Ba=Ti 176cm:------> ------> ------> ------13-18% 6-10% 1-5% 0% v. Residual fraction Ti>AI>K>Fe>Cr>Ba>V Co>Mg>Ni>Li P>La>Be>Sr=Cu>Mn>Zn>Pb Ca 2cm: ------> ------> ------> ------>81% 72-75% 43-57% 12% Ti>AI>Ba>Cr=Fe>K>V Co>Ni=Li>P>Mg>Cu Sr>Zn=Be>Mn>La=Pb Ca 24cm: ------> ------> ------> ------>84% 67-78% 51-58% 18% 262 Ti>AI>Fe>V>K=Cr Co>Li=Ba>Mg>Be=Ni Sr=La=Zn>Pb=P>Mn>Cu Ca 6 4 c m :------> ------> ------> — >89% 72-80% 58-64% 37% Ti>AI>Ba>Cr=Fe>K>V>Co Ni>Li=Mg>Cu P>La>Zn>Be>Sr>Mn>Pb Ca 94cm: ------> ------> —-«------> ------>81% 71-79% 45-67% 11% Ti>AI>P=Cr=Fe=Co=Zn>Cu=Ba=La>V>Li=Be>Ni>Pb>K>Mg=Mn Sr Ca 122cm:------> ------> ------>83% 74% 49% Ti>Ba>AI>Fe>V=Cr>P>K>Co=Pb>Ni>Zn=Li>Mg>Cu Be>La>Sr>Mn Ca 149cm:------> ------> ------>80% 63-78% 29% Ti>AI>Ba>Cr>La>K>Fe>Cu>Pb>V>Sr>Zn Mg>Li>Co>Ni>P>Ca>Mn Be 176cm:------> ------> ------>80% 67-79% 57% The data obtained show some interesting variations down the core; specific items are as follows: 1. The proportions of Cu, P, Pb and Zn leached by H2O2 are lower in the lowermost 82 cm of the core than in the upper 94 cm of it. 2. Comparing the amounts of residual Ca downcore, it is noticed that these are low at 2. 24 and 94 cm depths - where the total Al content is low-and high at 122 and 176 cm depth - where high total Al concentrations occur. 3. At 122 and 149 cm depths, the amounts of reducible Fe, Mn, Co, Ni, Zn, Pb and Cu are lower than at the other depths. 4. At 176 cm depth, the major part of all the elements determined remains in the residual fraction. 263 6.4 Summary The part of the sediment most easily released is the exchangeable fraction, leached by NH4OAC. It demonstrates the presence of easily hydrolyzed Ca, Sr and Mg -components, and of a minor amount of Mn in an exchangeable form; it also shows that a minor K-contribution from the seasalt occurs. Ammonium acetate has slightly attacked the carbonate fraction. Ammonium acetate and acetic acid were used as the second leaching reagent, designed to remove the carbonate material. However, this reagent was not successful in dissolving all of it. A lack of complete dissolution of carbonates was recorded mainly in the carbonates-enriched samples. This seems to have a relationship to sediment type; in coarser grained samples, carbonates were difficult to be dissolved in the NaOAc/HOAc reagent.. Magnesium is partially associated with the carbonates; this association is more pronounced in samples collected off Thassos island. A carbonate related contribution of Pb and Zn occurs in the sediments studied. The reducible fraction contains Fe,Mn-oxides and shows their scavenging ability for trace metals. Comparison of the amounts of trace metals associated with each of the non-residual fractions shows that the majority of them occurs in the reducible fraction; such elements include Zn, Pb, Mn, Fe, Ni and Co. The H20 2 attack has leached both the organic matter and the sulphide minerals. Copper was found to be complexed by organics, especially in the Samothraki plateau sediments. Phosphorus also has an affinity for organic matter. A minimal attack of the major aluminosilicate phases by H20 2 was recorded. It was found that the highest proportions of the non-residual Pb, Zn, Cu, Ni and Mn occur in the lerissos gulf sediments. This could be related to the contamination of the sediments by mining effluents. The contamination of the sediments off Stratoni by the mainland mining activities is also suggested by the decrease of the non-residual Mn, Fe, Ni and Cu proportions with increasing depth, in the uppermost 16 cm of core I 14. The major parts of the elements Al, K, Fe, Ba, V, Cr, Li, Co, Ni, La, Cu, Be, P, Mg and Zn are hosted in the lattice structure of aluminosilicate minerals. In general, the metals determined in the marine sediments subjected to partition 264 analysis were found to occur mainly in the residual and reducible fractions. This is in agreement with the suggestions of Li (1982), and Calvert and Piper (1984); they observed that in marine sediments and ferromanganese nodules, metals tend to primarily associate with aluminosilicates and Mn and Fe oxides. 265 CHAPTER VII HUMIC SUBSTANCES In order to shed more light on the organic fraction of the sediments, humic substances isolated from selected samples are specifically studied. 7.1 Introduction and brief review Humic compounds are acidic, dark coloured, partially aromatic substances, of a complex chemical structure. They are widely distributed in nature; occurring in terrestrial soils, natural waters, marine and lake sediments, peat bogs, shales and brown coals. Based on their solubility properties, humic substances are partitioned into three main fractions : a. fulvic acid, soluble in both alkali and acid, and having relatively the lowest molecular weight; b. humic acid, soluble in alkali and insoluble in acid, having an intermediate molecular weight; and c. humin, insoluble in both alkali and acid, and having relatively the highest molecular weight. In general, humic and fulvic acids are polycarboxylic acids with phenolic, alcoholic and carbonyl groups, with aromatic rings and relatively high content of free radicals (Schnitzer, 1978; Schnitzer and Khan, 1972; Stevenson, 1967); their molecular weight ranges from a few hundred to several thousands. Humification is the result of a complex biochemical and physicochemical transformation. Starting material can be decomposition products deriving from chemical and biological degradation of plant and animal materials, microorganisms and their metabolic products, and microbes through their various synthetic activities. During humification, carboxyl groups are lost and more strongly chelated forms of carbonyl and hydroxyl groups are developped (Stevenson and Goh, 1971). Condensation is a chemical change in which two or more molecules react with the elimination of water or of some other simple substance. 266 The origin of marine humates, as far as terrestrial or marine derivation is concerned, is dubious. Ertel and Hedges (1985) suggested that vascular plant debris -an usually important constituent of coastal sediments - can be a significant source of sedimentary humic substances. A marine origin of marine humates was proposed by Nissenbaum and Kaplan (1972), and Nissenbaum (1974), on the basis of their genetic relation to degraded planktonic material. On the contrary, a continental origin of the marine humic fraction was suggested by Yen and Tang (1977), who proposed that lignins formed on the continents are the most likely source of humic acids found in marine sediments. However, it has also been suggested that both origins may be valid; thus, Rogers and Koons (1968) proposed that much of marine nearshore humic material is likely to be terrestrially derived, whereas offshore humic substances are dominantly deriving from settled marine plankton. Various differences occur between marine and terrestrial humic acids (H.A.); (Rashid and King, 1970 and 1971; Hue et al, 1974; Nissenbaum and Kaplan, 1972), the most important are: a. marine H.A. contain more aliphatic and heterocyclic structures than terrestrial H.A.; I b. marine H.A. are less oxygenated; than terrestrial H.A.; c. marine H.A. contain onlv about half the amount of carboxyl and phenolic groups; than terrestrial H.A.; | d. marine H.A. contain more S and N than terrestrial H.A.; and e. marine H.A. associated with bottom sediments have a low degree of condensation* (Bordovski, 1965). This can, at least partly, be attributed to the occurrence of relatively more anaerobic conditions in sea-bottom deposits than in land-soils - bearing in mind that the hydrogen content is higher when anaerobic conditions predominate. Humates and fulvates are very important constituents of soils and sediments; performing in many functions, like: a. metal - scavenging, - concentration and - transport (Kononova, 1966; Drozdova, 1968; Manskaya and Drozdova, 1968; Swain, 1968). Although various metals 267 can be associated with humic acids, their sorption on them is affected by the metal concentration, the pH of the equilibrating solution, and the amount of H.A. present (Kerndorff and Schnitzer, 1980). According to Nissenbaum and Swaine (1976), most of the metals associated with humic substances are introduced into the sediment during the diagenetic formation of the humic acids - the metals being leached from a mineralogical component or components, either by the humates or their precursors; b. growth regulation - they may either nourish aquatic organisms or be biotoxic; c. toxic material -concentration and -transport. Humic and fulvic acids can operate either as accumulators of trace toxic materials (like pesticides) in sediments or as transporters of them to other environments in which pH or ionic strength would precipitate them (Steelink, 1977); d. sediment formation, by transportation and enrichment of mineral substances in sediments and sedimentary rocks (Manskaya and Drozdova, 1968). Metal ions, oxides and hydroxides, as well as minerals interacting with the functional groups of H.A. form complexes through electrostatic forces and electron pair sharing (Nissenbaum and Swaine, 1975). The H.A. functional groups are oxygen-containing groups, especially -COOand phenolic -OH, in the case of soil humates, and N- and S- containing ligands, in the case of marine humates (Nissenbaum and Swaine, 1976). Diagenesis of the sediment has, in general, little effect on the elemental composition and metal content of the humic acids (Brown et al, 1972). 7.2 Studies on humic substances 7.2.1 Utility and objectives of the work The chemical characterization of humic substances and metal-humic complexes in sediments, and the residual binding ability of H.A. toward metal ions are of environmental importance because they can furnish fundamental information on: a. the origin of sediment humic substances and the humification processes they have 268 undergone; b. the role of humic substances in the transport processes of metal ions from terrestrial sources through aquatic systems, and in the accumulation of metals in aquatic sediments; c. the metal biogeochemistry cycles; and d. the possible function of humic substances as indicators of metal pollution in aquatic systems. Therefore, a number of humic acids isolated from river, brackish-water-lake and marine sediments (Table 7.1, Fig. 7.1) were analyzed for their elemental composition, chemical functionalities and metal-organic complexes, by Inductively Coupled Plasma Emission (I.C.P.E.), Infra-Red (I.R.) and Electron Spin Paramagnetic Resonance (E.S.P.R.) Spectrometry. The principal objectives of this analytical work were the following: a. the chemical and physico-chemical characterization of the humic acids, in order to obtain information on their structural and functional properties; b. the comparison of chemical properties of sediment humic acids with those of soil humic acids; c. the determination of the primary amounts of metals in sediment humic acids and of the nature and stability of metal-humic bindings; d. the evaluation of the residual complexing ability of sediment humic acids toward metal ions, using copper ion as spin probe; and e. the determination of the stability of Cu+2-humic acid complexes toward intense water washing and exhaustive proton exchange. 0 10 km Fig. 7.1 Map showing the sites of the samples from which humic acids have been extracted. Sample 4 combines offshore sediment from two different sites (due to smali amounts 9 6 2 :tcr:a!}; the same-applies for sample 5. 270 Table 7.1 : Sediments from which humic acids have been extracted. Sample N2 Identifier Type 1 Core THR 22 offshore subsurface sediment (0 -1 .7m) 2 Core I 14 offshore subsurface sediment (0 -1 .8m ) 3 STR 8 offshore surface sediment 4 THR (31+122) offshore surface sediment 5 THR (1+11) offshore surface sediment 6 B 72 rive r surface sediment 7 B 22 river surface sediment 8 B 27 brackishwater lake surface sediment 7.2.2 Methods and analytical techniques used 7.2.2.1 Extraction and purification of humic acids The method used for the extraction of the H.A. is the following: The air-dried, fine-ground sediment is treated with 1M HCI in a, volume (V ml) of acid to weight (g) of sample, ratio of 5 to 1. The residue is separated from the supernatant by centrifugation and washed twice with distilled H2O. Then, it is treated with V ml 0.5N NaOH (prepared under N2) for 24 hours at room temperature. The alkaline extract is separated from the residue by centrifugation, and then it is acidified with 1N HCI followed by 6N HCI to final pH1, and allowed to stand at room temperature for 2 hours. The coagulate (H.A.) is separated from the supernatant (fulvic acid , F.A.) by centrifugation (27,000 G, 20 min.). The centrifuged H.A. is purified by repeated redissolution with NaOH and reprecipitation with HCI. Then, it is washed with distilled H2 0 and finally freeze-dried. 7.2.2.2 Copper complexation The method used for the preparation of Cu-complexes with H.A. is as follows: An aliquot of humic acid is placed on a fritted glass funnel and treated to saturation with 0.1 M Cu(CI04)* 6H20. Successively, the copper-saturated aliquot is divided 271 into two portions; the one is washed with distilled water, the other with 0.1 M HCI till free of copper ions. The latter is then washed with distilled water till free of chloride ions. The products of interactions are finally air-dried. 7.2.2.3 Ash content determination The method used for the ash content determination consists of ignition of the humic preparations at 750°C for 4 hours. 7.2.2.4 Analytical.techniques used The following analytical techniques have been used: a. elemental analysis, providing information on the C, H, N and S content of the H.A.; b. Infra-Red spectrometry, in order to shed some light as far as the functional groups, aliphaticity and aromaticity are concerned; c. I.C.P.E.S., for the determination of Fe, Cu, V and Mn in the humic preparations, following a 4N HN03 digestion of humic acid samples; and d. E.S.P.R., as far as organic free radicals and metal ion - humic acid complexes are concerned. 7.2.3 Results and Discussion 7.2.3.1 Elemental composition The major elemental composition (on a moisture- and ash- free basis), the atomic ratios C/H, C/N, O/C, and ash contents of the sedimentary humic acids isolated from freshwater and marine sediments are given in Table 7.2. A comparison of the composition of the humic acids shows that: a. The highest N, H and S contents correspond to the H.A. isolated from the three marine surface sediments (samples 3,4 and 5), the brackish-water sediment (sample 8) and the one river sediment (sample 6), (Fig.7.1); 272 b. The lowest C corresponds to the highest 0 content, determined in the H.A. of sample 2; c. The highest C corresponds to the lowest O content, determined in the H.A. of samples 5 and 1; and d. In general, the sedimentary H.A. exhibit a wide variation in their elemental composition; only those isolated from the three marine surface sediments (samples 3, 4 and 5) have similar compositions. Table 7.2 : Major elemental composition (on a moisture- and ash- free basis), atomic ratios and ash contents of sedimentary and soil humic acids. * HA. Q H N 3 Q. C/H C/N O/C Ash sam. NS % % % % % A.R. A.R. A.R. % 1 59.8 5.3 4.2 <0.1 30.7 0.94 16.6 0.38 11.4 2 40.0 4.5 4.0 <0.1 50.5 0.74 10.0 0.50 10.4 3 52.4 6.7 6.0 3.0 31.9 0.65 10.1 0.46 18.2 4 52.8 6.3 6.2 3.3 31.4 0.70 10.0 0.45 41.8 5 58.6 6.5 ‘ 5.7 2.9 26.3 0.75 11.9 0.34 13.8 6 52.8 6.2 6.3 1.7 33.0 0.71 9.8 0.47 10.5 7 50.4 3.8 3.3 <0.1 42.5 0.90 18.3 0.61 12.3 8 48.7 6.7 6.2 5.8 32.6 0.61 9.2 0.50 12.2 so il** 56.0 4.5 1.6 0.3 37.0 1.00 20.0 0.50 2-4 A.R .: Atomic Ratios * Calculated through the formula : O=100-(C+H+N+S) ** Average values for a range of soils (Schnitzer and Khan, 1972) A further comparison of the sedimentary humic acids studied with soil humic acid - using the average values for a range of soils, according to Schnitzer and Khan (1972) - demonstrates that: a. The sedimentary humic acids in this area have higher N contents and lower C/N ratios, implying their higher content of proteinaceous materials; b. The sedimentary H.A. in this area have lower C/H ratios, indicating their higher Stretching is the vibration n the line of a valence bond, that changes the distance of the two atoms bound by this bond. 273 aliphaticity; c. The sedimentary H.A. in this area often have higher H and S contents; d. C and O contents in the sedimentary H.A. in this area are within the corresponding ranges of the soil H.A.; and e. The sedimentary H.A. in this area correspond to a higher ash, implying an inorganic material association. 7.2.3.2 Infra-Red Spectroscopy Studies The sediment humic acids were examined by Infra-Red (I.R.) spectroscopy; Infra-Red methods being commonly used in H.A. studies because they provide valuable information concerning the nature and arrangement of functional groups. In general, the Infra-Red absorption characteristics of the humic preparations isolated from the 8 samples resemble to each other (Fig.7.2a-h). However,the relative intensities of specific bands vary. The I.R. spectra recorded show that significant bands occur at the following wavelengths; the interpretation being done according to Stevenson and Goh (1971) : 3.400 cm'1, arising from H-bonded -OH groups, including those of -COOH; * 2,900 cm'1, arising from aliphatic >CH stretching; 1,720 cm'1, arising from carbonyls and free carboxyls; 1,660 -1,600 cm'1, arising from aromatic >C=C< vibrations, amide absorption of peptides, H-bonded and conjugated ketones; 1,575 and 1,390 cm '1, characteristics of the -COO' ion; 1,540 cm'1, attributed to the peptide linkage of proteins; 1.400 cm'1, arising from -OH deformation and >C-0- stretching of 274 phenolic -OH groups, or from C-H deformation of >CH2 and -CH3 groups; 1,200 cm'1, assigned to C-0 stretching and -OH deformation of carboxyl groups; 1,050 cm"1, attributed to C-0 stretching of polysaccharide or polysaccharide-like substances. In conclusion, the strong bands dominating the I.R. spectra of the humic preparations are indicative of hydroxyl groups of various types, hydrogen bonded and conjugated carbonyl and amide groups, and polysaccharide components. A predominance of aliphatic over aromatic structures is noticed; among the sediment humic acids examined, the higher aliphatic character is always associated with a larger amide group content. Carboxylate groups prevail over free carboxyls; indicating a high metal binding by humic acids. TRANSMITTANCE (X) A AEUBR (CM‘ WAVENUMBER CD CO CO 00 ro 2 75 276 S4 S5 S 6 i } >' r 277 S 8 Fig. 7.2 I.R. spectra of sediment humic acids. ! f 278 7.2.3.3 Metal Humic Acids A comparison of the contents of the paramagnetic transition metais Fe, Cu, V and Mn determined in the sedimentary humic acids shows (Table 7.3) that: a. high Fe contents (3,700-19,200 jig/g) are associated with metal humic acids; b. Cu contents are generally one to three orders of magnitude lower than the corresponding Fe contents; and c. the contents of Mn (23-128 fig/g) and V (49-892 jig/g) are much smaller than the corresponding Fe contents. Copper exhibits higher portions of total metal content associated with H.A. than the other three metals examined (Table 7.3). This demonstrates its strong affinity towards humics. The association of Cu with organic material has been also indicated by partition analysis as relatively high portions of total Cu were found to be leached by hydrogen peroxide. The data presented in Table 7.3 show that the portions of Cu associated with H.A. do not depend on the total Cu content of the corresponding samples. Table 7.3 : Humic acids contents of selected sediments; Fe, Cu, V and Mn concentrations in these .sediments (A), and in the respective H.A. (B); percentages of total Fe, Cu, V and Mn concentrations: associated with H.A. (C), and bound to the organic/sulphide hosted fraction (D). Sample HA Ee Qu y Mn % (gr) 1 0.148 A 49,500 67.62 125 623 B 3,776 1.05 256 23 C 0.01 0.00 0.30 0.00 D 2.02 28.94 3.20 7.95 2 0.143 A 47,624 59.62 120 1,049 B 13,179 2.08 282 121 C 0.04 0.00 0.34 0.02 D 1.90 25.04 2.92 6.68 3 0.355 A 36,000 40 94 340 B 10,182 426 65 . 92 C 0.10 3.78 0.25 0.10 D 279 Table7.3 (continued) 4 0.141 A 18,450 18.15 61 505 B 19,200 564 145 128 C 0.15 4.37 0.34 0.04 D 5.42 34.71 9.67 6.63 5 0.128 A 33,250 36 92 990 B 5,250 875 168 51 C 0.02 3.11 0.23 0.01 D 3.56 25.56 2.74 4.44 6 0.773 A 37,000 20 107 1,090 B 5,021 236 83 57 C 0.11 9.12 0.60 0.04 D 7 0.109 A 45,000 39 128 1,210 B 10,193 1,354 892 71 C 0.02 3.78 0.76 0.01 D 2.22 18.46 2.58 8.51 8 0.238 A 10,400 7 28 121 B 6,136 273 49 26 C 0.14 9.27 0.42 0.05 D 5.48 26.14 3.89 5.21 A : metal content (|ig/g) of total sediment ; B : metal content (jxg/g) in H.A. ; C : % of total metal content (g) associated with H.A. ; D : % of total metal content (jxg/g) bound to the organic/sulphide hosted fraction. 7.2.3.4 E.S.P.R. Studies The organic free radicals in the sediment humic acids were studied by E.S.P.R.; the data presented in Table 7.4 allow a comparison between these H.A. and the soil and aquatic ones, on the basis of their E.S.P.R. parameters. Thus, it is shown that: a. The organic free radicals in the sediment humic acids examined have the same order of concentrations with those of soil and aquatic humic acids. 280 b. The g1 -values of the organic free radicals in the sediment H.A. are comparable with those in soil and freshwater H.A. This demonstrates that the organic free radicals are of the same nature, and implies therefore that semiquinone radicals are present in the sediment H.A. examined. Table 7.4 : ESR parameters of organic free radicals in sediment, soil and aquatic humic acids. H.A. Free radical cone, g-value Line width sample (spins/g x 10'17) (gauss) 1 3.53 2.0038 6.0 2 3.35 2.0035 7.3 3 1.18 2.0040 6.1 4 1.02 2.0040 7.5 5 3.22 2.0039 6.6 6 0.97 2.0038 5.8 7 7.26 2.0035 6.5 8 0.80 2.0034 5.9 soil1 5 -1 0 2 . 0032- 2.0047 4.8 -5 .2 freshwater1 3-4 2.0038 4.0 marine water1 2.6 2.0030 4.8 1 : Schnitzer and Skinner, 1969; Steelink and Tollin, 1967; Nissenbaum and Kaplan,1972. A comparison of the organic free radical concentrations, H contents and C/H ratios in the sediment H.A. studied shows that a lower free radical content is associated with prevalent aliphaticity (Fig. 7.3 and Fig. 7.4), while higher free radical concentrations are associated with higher aromatic character. 1 The quantity g is a pure number and its value depends on the relative contributions of orbit and spin to the total angular momentum Abragam and Bleaney,1970. 281 H (%) Fig. 7.3 Scatter plot of H content (%) versus organic free radical content (spins/gx10"17) in sediment humic acids. Fig. 7.4 Scatter plot of C/H ratio versus organic free radical content (spins/gxICT17) in sediment humic acids. The E.S.P.R. spectra analysis (Fig. 7.5, Table 7.5) demonstrates the presence of up to 282 four different geometrical forms of ferric iron in complexes with the sediment humic acids. Also, E.S.P.R. evidence is given for vanadyl (II) ions, copper (II) ions and manganese (II) ions complexed by sediment humic acids (Senesi and Sakellariadou, 1987). Table 7.5 : ESR data for Fe3+ complexes with sediment humic acids. H.A. g AH1 g" sample N3 1 4.14 79 5.6 8.7 2 4.14 75 5.6 8.5 3 4.12 75 5.6 8.7 4 4.13 86 5.5 8.5 5 4.14 64 5.6 8.7 6 4.14 64 5.6 8.7 7 4.13 79 — 8.8 8 4.12 75 5.6 8.5 1 The parameter 2 a H is defined as the width of the magnetic resonance line at half the maximum intensity, for a paramagnetic ion or free radical in the process of • i . ' ; paramagnetic resonance | Abragam and Bleaney,197Q. 283 a b 284 c Fig* 7.5 E.S.P.R. spectra of sediment humic acids. a : sa mp I e 1 b; sample 5 c; sample 7 A : organic free radical; E. 3', B ", B '" : Fe3+; C : Cu2+ ; D : V02+ ; E : Mn2+ . r 285 7.2.3.5 Artificially formed Cu(lh-H.A. Complexes The humic acid preparations isolated from the samples 1, 3 and 8 have been saturated with 0.1 M Cu+2 solution and then either exhaustively washed with distilled water or with 0.1 M HCI and successively with distilled water. The total contents of Fe, Cu, V and Mn, determined by I.C.P.E.S., are presented in Table 7.6. It is shown that: 1. Fe contents in the artificially formed Cu(ll)-H.A. complexes, after being washed with either H20 or HCI followed by H20,are almost the 1/3 of the Fe contents in the corresponding original sediment H.A. (before being saturated with Cu+2). This implies a partial replacement of Fe+3 by Cu+2. 2. Cu contents in the exhaustively water washed artificially formed Cu+2-H.A. complexes are 50 to 100 times higher than those in the corresponding original humic acids. This shows the high stability of the Cu+2-H.A. complexes toward water washing. 3. Cu contents in the Cu+2-H.A. complexes, after their HCI treatment, are 5 times higher than those in the corresponding original humic acids. This indicates that the artificially formed complexes have a limited stability to intense proton exchange. 4. V and Mn contents in the two groups of artificially formed Cu(ll)-H.A. complexes, those washed with H20 and those treated with HCI followed by H20, are lower than their contents in the corresponding original H.A. This suggests that V and Mn have been replaced by Cu. The artificially formed Cu(ll) complex with the sediment H.A. isolated from sample 1 was examined by Infra-Red Spectroscopy. The I.R. spectrum of the complex, after being washed first with HCI and then with H20 (Fig. 7.6 b), shows that a strong band occurs at 1,710 cm-1, arising from carbonyls and free carboxyls. On the contrary, the I.R. spectrum of the exhaustively water washed complex (Fig. 7.6 a) shows that at 1,710 cm"1 a weak band occurs; indicating a lower content of the corresponding functional groups. 100. Fig. 7.6 I.R. spectra of artificially formed Cu(ll) complexes with sediment humic acids, islolated from sample 1. 286 287 Table 7.6 : Fe, Cu, V and Mn total concentrations (determined by I.C.P.E.S.) in sedimentary humic acids saturated with 0.1 M Cu2+ solution. HA Ee £u Y Mn sanLEff (M-g/g) (tig/g) Oig/g) (txg/g) 1 a 1,361 45,714 61 12 1 b 1,758 4,701 36 12 3 a 3,142 24,571 49 31 3 b 4,185 2,223 31 45 8 a 2,341 26,667 5 8 8 b 2,550 1 ,485 29 10 a : artificially formed Cu2+- sed. H.A. complexes, exhaustively water washed. b : artificially formed Cu2+- sed. H.A. complexes, washed with 0.1 M HCI and successively with distilled water. In conclusion, the data concerning the complexes of copper ions with sediment H.A. imply that (Senesi and Sakellariadou, 1987): 1. Copper ions are highly adsorbed and/or compiexed by sediment humic acids from this area, by either partial replacement of other metal ions, i.e. ferric iron, or by involvement of originally free humic functional groups. 2. Artificially formed Cu(ll)-sediment H.A. complexes show a high stability toward exhaustive water washing but only a limited portion is stable to intense proton exchange. 288 CHAPTER VIII DISCUSSION The subject of this final chapter is the interpretation of the geochemical data on the North Aegean Sea sediments studied. The variations in the major and minor element distribution, the element associations and the controls on the geochemistry of the sediments studied, will be discussed in the following in an element by element basis. The elements dealt with will be examined in groups according to their arrangement in the periodic table. 8.1 Nontransition elements 8 .1.1 Alkali Metals (Group 1a) These are univalent metals, belonging to Group 1a of the periodic table. 8 .1.1.1 Potassium Potassium is a lithophile element and a major component of K-feldspars, micas and illitic clays; K is an important constituent of granitic igneous rocks (Rose et al., 1979). The highest K robust mean of the Box-Cox transformed data (in the following referred to as robust - Box-Cox - mean) within the study area occurs in the sediments belonging to the Strymonikos group, closely followed by the Kavala group (Table 8.1). The regional variation of K in the surface sediments shows that K enrichments1 occur: a. In the main river samples and the sediments collected off the mouths of these rivers, implying K supply via the fine grained river runoff. b. In the clayey silty sediments of the central part of the lerissos gulf. 1 The terms enriched, depleted, rich and poor - for a particular element - sediment are used in a comparative sense and denote the magnitude of a particular element concentration in relation to the general distribution of the element concerned within the study area. 289 c. In the coastal sediments collected from a narrow belt stretching from the mouth of the Strymon river to Nea Peramos bay. This coastal K enrichment is disproportionate to the respective Al enrichment, and it therefore seems that K in these particular sediments is hosted in K-bearing phases which are mainly other than clay minerals, possibly feldspars. This suggestion is supported by the enrichment in sand sized material in the sediments of this area. The samples collected from the beach adjacent to the area of offshore enrichment, also have high K contents. These results, together with the presence of granites and granodiorites in the respective part of the mainland, suggest that the K enrichment offshore is most likely to reflect coastal erosion and nearshore deposition of the weathering products of the granitic rocks. d. In the sandy coastal sediments - both on- and off- shore - which lie southeast of Olympias . This coastal enrichment is also likely to reflect coastal erosion and nearshore deposition of K bearing weathering products, as an outcrop of piagioclase-microcline-gneiss extends across the corresponding part of the mainland. Statistical treatment of the geochemical data shows that K is mainly associated with Al, Be and Ba. Partition analysis of selected samples gives evidence that in all of them the bulk of K remains in the residual fraction. Consequently, K in the surface sediments studied is hosted in resistant aluminosilicate minerals and their alteration products. In the buried sediments, the downcore distributional pattern of K is similar to that exhibited by Be and Ba, and inverse to that of Ca content (see also section 8.1.2.3). Partition analysis of buried sediments shows that the major part of K remains in the residual fraction. Thus, the downcore behaviour of K is similar to that found in the surface sediments. In conclusion, K in the sediments of the study area is mainly supplied by the fine grained river runoff and locally by coastal erosion. It is mainly hosted in the lattice structure of potash feldspars and clay minerals. 8 .1.2 Alkaline Earth metals (G ro u p 2 a ) These are bivalent metals and they belong to Group 2a of the periodic table. 290 8 .1.2.1 Beryllium Beryllium is a lithophile element which is mainly hosted in the minerals mica and feldspar (Rose et al., 1979). Its most important mineral is beryl, Be3Al2(Si03)g. The charge/radius ratio of Be2+ is ^6.5; Al has nearly as high a charge/radius ratio, namely ^6.0, and therefore some chemical similarities between the two cations exist (Cotton and Wilkinson, 1966). The Be mean values determined in sediments collected from both areas, westwards and eastwards of Thassos island, are similar (Table 8.1). The spatial Be distribution in the surface sediments exhibits similarities with that of Al. Thus, Be is relatively enriched in the fine grained sediments that occur : (a) in the central deeper part of the lerissos gulf; (b) in Kavala gulf; (c) over much of the Strymonikos gulf and plateau; (d) in an extensive belt stretching westwards of the Evros river mouth; and (e) in two lobes lying close to the Nestos river mouth. In the sediments of the above areas, apart from relatively high Be contents, enrichments in Al and Fe were also determined. Statistical treatment of the geochemical data confirmed the association of Be with Al; similar results have been reported by Li (1982), who applying factor analysis on various elements in manganese nodules and associated sediments, concluded that Be is associated with aluminosilicates. Partition analysis of selected samples gave evidence that the major part of Be remains in the residual fraction. All the aforementioned results support the suggestion that Be in the surface sediments studied is mainly hosted in clay minerals. In buried sediments, Be follows K and Al in its downcore distribution, exhibiting an inverse relation with Ca, as the latter mainly reflects biogenic matter (see also section 8.1.2.3). In conclusion, the distribution of Be within the study area is mainly controlled by the variation of the finer-grained material. The highest Be concentrations were determined in sediments made up either of clayey silts and silty clays - in the major part of the area - or of mainly silty sands- as far as the sediments close to the Evros river mouth are concerned. 8.1.2.2 _Magnesium The chemistry of Mg is intermediate between that of Be and that of Ca (Cotton and 291 Wilkinson, 1966). Magnesium enters into the composition of a large number of rock-forming silicates and it is, therefore, a major component of Fe,Mg-minerals, such as biotite, pyroxene, amphibole, olivine, etc. (Read, 1970). Another important source of Mg is dolomite, CaCOg-MgCOg. According to Kukal (1971), Mg in sediments is mostly bounded to clay minerals, like illite and chlorite (Friedman and Sanders, 1978; Mason, 1966). The highest robust - Box-Cox - Mg-mean within the sediments of the study area, occurs in those belonging to the Strymonikos group (Table 8.1). The regional distribution of Mg shows that its highest concentrations occur in the finer-grained sediments. Thus, a comparatively high Mg content (Fig. 4.15) was determined over much of the Strymonikos gulf (apart from the related ridge) and plateau, as well as in the eastern and western part of the Kavala gulf; in these areas, silty clay and clayey silt predominate. In the Samothraki plateau, a high Mg content was determined in the deep water, clayey silt sized sediment, southeast of Thassos. The fine grained sediment westwards of the Evros river mouth and that lying off the Filiouris river mouth, as well as samples collected from the latter river, contain relatively high Mg concentrations (Fig. 3.14); indicating that the terrestrial detrital material supplied by these rivers is Mg enriched. Although the Nestos river sediments are poor in Mg, two offshore lobes of enrichment occur; they lie westwards and northeastwards, respectively, of the river mouth. The sediments of both these lobes are made up of clayey silt; therefore, the enrichments concerned reflect the association of Mg with the finer-grained material. These two lobes of Mg enriched sediments are thought to represent reworking of older sediments (see also section 8.1.3.1). Cluster and factor analysis applied to the geochemical data show that Mg is strongly associated with V, Cr and Ni; this group of elements is regarded as the mafic/ultramafic weathering products factor (Rose et al., 1979). Partition analysis of selected samples gives evidence that more than 45% of the total Mg remains in the residual phase; consequently, the major source of Mg in the sediments studied is resistant detrital minerals (Fe,Mg-minerals and their weathering products). The largest proportions of residual Mg were determined in sediments collected from offshore areas which receive the weathering products of basic and ultrabasic formations. Thus, as much as 91% of the total Mg remains in the insoluble solid of sample THR 84, collected off the northwest coast of Samothraki 292 island, from an area where sandy sediment predominates and rock fragments are abundant. The high residual Mg of the sample concerned is attributed to resistant minerals supplied by the ultrabasic intrusion that occurs on the respective part of the island. On the basis of the partition analysis results and the elemental associations , the high Mg concentrations determined in the coastal sediments collected along the west coast of the lerissos gulf, and close to Stavros and Olympias are likely to reflect weathering material that derives from basalts and meta-gabbros, and amphibolites, respectively (Fig. 1.4). In lerissos gulf, just east of Nea Roda, a coastal Mg enrichment was determined, corresponding to relatively high Cr, Ti, Ni and Co contents in the sediments. In addition, hornblende and talc were recognized in samples of the coastal area concerned. It has already been mentioned (section 1.3.2) that a small ultrabasic igneous complex of ophiolitic character crops out on the mainland adjacent to this coastal area. Hence, the Mg enrichment determined in these coastal sediments most likely reflects material deriving from the erosion of this ultrabasic complex. The presence of hornblende and talc could be attributed to primary crystallization, in the case of the former, and to hydrothermal alteration of the ultrabasic formation, in the case of the latter. Comparing the areas of Mg enriched sediments close to the mouth of the Asprolakkos river (west coast of lerissos gulf) and just east of Nea lRoda, it is obvious that the latter is significantly less extensive than the former. This can be explained, at least partially, by the fact that the latter area receives material deriving from coastal erosion which, therefore, is coarser-grained than that supplied by the river runoff. The latter, being finer, is further transported by current and wave action. Partition analysis also shows that about 8% of the total Mg in the sediments studied is removed with the exchangeable fraction, indicating that easily hydrolyzed Mg components are likely to be present in the sediments. Also, a significant proportion of the total Mg was found to be hosted in the carbonate fraction, probably reflecting the presence of dolomite and high Mg-calcite. Among the samples subjected to partition analysis, the Mg association with the carbonate fraction was found to be more pronounced in the sediments collected off Thassos island. This is probably because the marbles of Thassos are dolomitic (Vavelidis and Amstutz, 1981). In buried sediments, the Mg distribution pattern downcore is, in general, similar to 293 the respective Cr and V trends, demonstrating a common origin for these metals - Mg, Cr and V most probably deriving from the weathering products of mafic/ultramafic rocks. In addition, according to partition analysis results, Mg is mostly hosted in resistant minerals, in agreement with its previously suggested derivation. In conclusion, Mg in the sediments of the study area is associated mainly with the finer-grained material supplied by the rivers; in local coastal areas Mg is supplied by coastal erosion. Magnesium occurs mainly in detrital resistant silicates and their alteration products. Close to Thassos island, a contribution to the Mg content of the sediments from the dolomitic marbles of the island exists. 8 .1.2.3 Calcium Calcium is the third most abundant metal terrestrially (Cotton and Wilkinson, 1966). Its compounds are very abundant, widely distributed and essential to life. Calcium is a major constituent of carbonate minerals, like calcite, dolomite and aragonite. The highest robust - Box-Cox - Ca-mean in the sediments of the study area occurs in those collected from the Samothraki plateau (Table 8.1). This indicates that, as far as background population is concerned, Samothraki plateau sediments are relatively enriched in Ca. Nevertheless, throughout the study area, Ca enrichments are to a large extent associated with a high carbonate content. West of Thassos island, the highest Ca content was determined in sediments collected from shallow water areas, such as the Strymonikos ridge and triangle. The silty sand sediments that lie in the channel between Thassos and the mainland have also a high Ca content; however, the residual Ca in these sediments is relatively lower than in the other samples subjected to partition analysis. Therefore, Ca in the sediments of this channel, seems to be mostly supplied in the form of carbonates; a carbonate enrichment being also found in these sediments (Fig. 4. 8). This is also in accord with the presence of marbles on Thassos. The carbonate enrichment in the channel sediments could possibly be due to current action sweeping away the finer-grained aluminosilicate material. High Ca contents were determined in the coarse grained sediments lying off the west and south coast of Thassos. As shell fragments are abundant in the sediments concerned, the respective Ca content seems to be mainly supplied by shells and shell 294 fragments. However, some of the high Ca concentrations determined in sediments around Thassos can be partially attributed to material deriving from the marbles which occur on the island. The regional Ca variation in surface sediments shows one more area of enrichment; it is an extensive one and it lies in the eastern part of Samothraki plateau. This Ca enrichment is also associated with a high carbonate content. Both partition analysis of selected samples and statistical treatment of the entire geochemical data show that the carbonate hosted fraction of the surface sediments within the study area, contains mainly Ca and Sr, the former representing calcium carbonate and the latter substituting for Ca in the carbonates. The fact that no other element shows preferential affinity for this fraction is hardly surprising since the crystal structure of calcium carbonate is quite unyielding to cation substitution (e.g. Graf, 1960). The small portion of residual Ca - determined by partition analysis - implies that only a small proportion of the total Ca occurs in the lattice of resistant Ca-bearing minerals. In buried sediments, the downcore variation in Ca content shows that Ca is mainly enriched in sediments having abundant shells and shell fragments. The downcore Ca fluctuation is, in general, inversely related to the distributional trends of K, Al, Be and Ba; Ca reflecting principally the carbonate abundance, exhibits an antipathy for the latter group of elements, which are present mainly in clay minerals. Partition analysis of buried sediments also shows that Ca is mainly present in the form of carbonates. The proportion of residual Ca was found to be greater in sediments that contain comparatively higher amounts of total Al. For example, in core THR 22, as much as 49 and 69 % of the Ca is residual in sediments collected from 122 and 176 cm depth. In the samples concerned, Ca is therefore mainly supplied by resistant Ca-bearing minerals. In conclusion, in the sediments of the study area, the distribution of Ca is mainly controlled by the variation in carbonate content. Calcium is principally supplied by shells and shell fragments; some Ca exists also in the form of resistant minerals. Locally, close to Thassos island, some of the Ca content in the nearshore sediments has been contributed by material that derives from the island's marbles. 295 8 .1.2.4 Strontium Strontium is a lithophile element which, due to the size of its ion, can proxy for either Ca or K; therefore, it is either admitted to calcium minerals or captured by potassium minerals (Mason, 1966). Strontium can also substitute for Ca in the skeletal material (Emiliani, 1955). Comparing the four groups of sediments, on the basis of the respective Sr content, it is obvious that the Kavala and Samothraki plateau groups correspond to the highest robust - Box-Cox - Sr-mean (Table 8.1). The regional Sr distribution in the surface sediments shows that, in general, Sr follows Ca in its spatial variation. The strong association between Ca and Sr has been also confirmed when cluster and factor analysis were applied to the geochemical data; Ca and Sr form an elemental grouping of their own, the carbonate factor. Partition analysis reveals that Sr is either hosted in the carbonate fraction or in a residual form. Despite the similarities exhibited by the regional distributions of Ca and Sr in the surface sediments of the study area, an enrichment in Sr was determined in Ca poor sandy sediments collected from a coastal belt stretching from the Strymon river mouth to Nea Peramos bay. As already discussed in section 8.1.1.1, the coastal area concerned receives detrital material deriving from coastal erosion of the granites and granodiorites adjacent to it. Consequently, the Sr enrichment is in accord with the offshore coastal deposition of detrital material from the granitic rocks. The behaviour of Sr in the buried sediments is similar to that in the surface sediments; part of Sr occurs in a residual form and part is hosted in the carbonate fraction. The downcore distribution pattern of Sr is in general similar to that exhibited by Ca. However, as far as the core STR 28 is concerned, Sr does not follow Ca in its downcore variation; Sr and Ba having almost identical downcore trends. Strontium in the buried sediment of core STR 28 probably occurs mainly in detrital felsic material. In conclusion, Sr content in the sediments of the study area is mostly controlled by the Ca distribution, but detrital controls involving granitic derived material are locally important. 296 8 .1.2.5 Barium Barium is a lithophile metai. The order of its abundance in some important rock types, according to Wedepohl (1969-1978) and Turekian (1977), is the following: Gran > Sh > Maf > Ss > Ls (Gran=granite, Sh=normal shale, Maf=basalt and gabbro, Ss=sandstones and quartzites, Ls=carbonate rocks). Comparing the four groups of sediments studied, the highest robust - Box-Cox - Ba-mean was found to occur in the Kavala group of samples (Table 8.1). The regional variation of Ba in the surface sediments shows that the highest Ba concentrations occur: (a) in the fine grained nearshore sediments lying westwards of the Evros river mouth; (b) in the Nestos river sediments and those close to its mouth; (c) in the coastal sediments lying between the Strymon river mouth and Nea Peranos bay; (d) in both onshore and offshore coastal sediments along the southwest and east-southeast coast of the lerissos gulf; (e) just off Stratoni; (f) in the coastal sediments along the west coast of Kavala gulf; (g) in both onshore and offshore coastal sediments along the west coast of Thassos; and (h) off the north coast of Samothraki island. Partition analysis of selected samples, from various locations of the study area, shows that Ba mostly occurs in the residual fraction. It is, therefore, suggested that the major source of Ba in these sediments is resistant detrital Ba-bearing minerals. This suggestion is in accord with the Ba association with the element grouping containing K, Al and Be, in the cluster and factor analysis. This grouping is regarded as representing detrital material that derives from felsic rocks. Considering the previous results, the aforementioned Ba enrichments in the sediments can be attributed as follows : 1. The sediments belonging to the first two groups (a and b) represent terrestrial detrital material, supplied through the runoff of the corresponding rivers. 2. The sediments belonging to groups (c), (d), (e) and (f) reflect deposition of material that is derived by coastal erosion of the granitic rocks that occur in the adjacent parts of the mainland. 3. The high Ba contents in the sediments along the west coast of Thassos (group g) 297 also reflect coastal erosion. In this case, the nearshore material is thought to contain the weathering products of barite mineralization, which according to Vavelidis and Amstutz (1983) is associated with the Pb-Zn ore deposits of the island. 4. The high Ba concentrations off the north coast of Samothraki (group h) correspond to comparatively high Mn values. Both these metal enrichments in the sediments concerned could be associated with the presence of hot springs on the island's coast, which according to Petrascheck (1982) release waters enriched in Ba and Mn. In buried sediments, Ba follows a similar trend to that exhibited by K and Be; most of the Ba present occurs in a resistant form. Thus, the Ba behaviour in the buried sediments studied is similar to that exhibited in surface sediments. In conclusion, the distribution of Ba in the sediments of the study area is mainly controlled by the supply of terrestrial material through river input and coastal erosion; Ba being mostly present in Ba-bearing detrital felsic minerals. 8.1.3 The Group lllb elements 8.1.3.1 Aluminium Aluminium is the commonest metallic element in the earth's crust (Cotton and Wilkinson, 1966), constituting 8% of it (Read,1970). It occurs widely in nature in silicates (like micas and feldspars)and as the oxide (bauxite). Aluminium is concentrated in sediments through the products of chemical breakdown of aluminosilicates (Mason, 1966); these products being largely clay minerals. A comparison of the robust - Box-Cox - Al-means of the four groups of samples shows that the highest one corresponds to the Strymonikos group (Table 8.1). In general, the regional distribution of Al in the surface sediments varies in a similar way to the Fe variation. Aluminium is enriched in sediments of finer than average grained size (mainly silt- and clay-sized sediments), although the Al enrichment west of Evros is associated with an increase in the very fine sand component. The highest Al contents in the surface sediments of the study area occur mainly in 298 those collected from the following regions : a. The major part of the Strymonikos gulf (apart from the ridge) and plateau, where silt- and clay-sized material is common. The Al distribution pattern in this region, shows the influence of the Strymon river in supplying aluminosilicate-rich material to the gulf, and the importance of the west-east trending submarine canyon - just off the coastline - in distributing this material downslope to the deeper areas. Its transport further eastwards is restricted by the shallow plateau, west of Thassos. b. The Kavala gulf (where silt- and clay-sized sediments are common), although no major river currently flows into it. Therefore, this Al enriched fine-grained material, most likely, represents older reworked sediments. The Nestos river flowed into the eastern coast of the gulf , during upper Pleistocene - lower Holocene times (see also section 1.4). c. The Samothraki plateau, where an extensive belt of Al enrichment occurs, stretching off the Evros river mouth westwards as far as Xylagani, and which is centred about 5 km offshore. Here the finest sand component (3 - 4 0 or 0.06 - 0.12 mm ) predominates (Perissoratis et al., 1987). In this belt, Al values reach a maximum of just over 9%. However, onshore coastal sediments collected from the mainland adjacent to the offshore Al rich belt are relatively poor in Al. This large Al enriched belt represents the fine-grained Al rich continental detritus supplied by the Evros river to the offshore coastal area, being transported further westwards by the strong westward-flowing current close to the coast (section 1.5). d. Close to the Filiouris river mouth, where silt- and clay-sized sediments predominate. The respective river sediments are also relatively enriched in Al. This local area of enrichment represents fine-grained Al rich material introduced via the Filiouris river. e. In two separate lobes, one to the west and one to the northeast of the mouth of the Nestos river, where silt- and clay-sized sediments predominate. By contrast, Nestos river sediments are comparatively poor in Al. Thus, the two nearshore lobes of Al rich sediments may represent reworking of older deeper sedimentary horizons. 299 Partition analysis gives evidence that the bulk of A! in the surface sediments studied occurs in the residual fraction; it, therefore, demonstrates that Al mostly occurs in clay and other resistant minerals. In accord with this is the presence of Al in the elemental group containing also K, Be and Ba - regarded as representing detrital terrigenous material (mainly clays), when the geochemical data were treated statistically. In buried sediments, Al exhibits a downcore distribution pattern similar to those of K and Be, with the bulk of Al occurring in the residual fraction on partition analysis. In conclusion, the distribution of Al in the sediments of the study area largely reflects the occurrence and the distribution of finer-grained sediments. Aluminium - occurring in a residual form - is either supplied by modern river runoff and transported by current action or it is reworked from older sediments. 8.1.4 The Group IVb elements 8 .1.4.1 Silicon Silicon - a nonmetallic element - is second only to oxygen in weight percentage of the earth's crust, 28% (Cotton and Wilkinson, 1966). Its compounds, like quartz and various silicates, are the most important rock-forming minerals (Read, 1970). Silicon is the most abundant of the elements determined in the sediments of the study area; its content in the surface samples ranges from 22 to 25%. Throughout the area, the spatial Si distribution in the surface sediments is similar to the regional variation in quartz abundance in the coarse fraction (the two distribution patterns are not identical, however, as Si may also occur in clay matter), and almost the reverse of the Ca distribution. The regional variation of Si in the surface sediments collected from the part of the study area to the west of Thassos, reflects closely the percentage of sand-sized material in the sediment. Thus, the highest Si concentrations were determined in sandy sediments collected from the following areas : (a) a thin coastal strip stretching along much of the mainland coastline; (b) Strymonikos ridge, where comparatively coarse-grained material predominates; (c) the outer part of lerissos gulf, where comparatively coarse-grained sediments also occur; and(d) seawards of the southwest coast of Thassos island. 300 The coastal Si enrichment, mentioned above, determined in both onshore and offshore sediments represents the importance of coastal erosion in Si supply to the nearshore sediments. This is in accord with the low Si content in coastal samples collected from the west coast of Thassos island, which largely consists of marbles. On Samothraki plateau, the variation of Si in the surface sediments is largely dependent on the amount of sand-sized quartz grains in the sediments. The highest Si content was determined in sediments of a broad belt that stretches from Thassos to the west of Alexandroupolis. All over the study area, sediments collected from the main rivers and the coastal areas close to their mouths are relatively enriched in Si; therefore, a corresponding Si supply by fine-grained material introduced by the rivers is implied. Cluster and factor analysis applied to the geochemical data, showed that Si is associated with elements such as K, Al, Be, Ba and Zr; an elemental group considered to represent terrestrial detrital material. The statistical treatment of the data has also confirmed the reverse relation between Si and Ca. In buried sediments, the downcore distribution pattern of Si reflects to a large extent the antipathetic relation between Si and Ca, as the former occurs in terrigenous detrital material and the latter in biogenic matter. In conclusion, the Si distribution in the sediments of the study area is mainly controlled by the quartz abundance in the coarse sediment fraction; Si exhibiting an antipathetic behaviour towards Ca. However, some supply of fine-grained sediment via the rivers also contributes Si. 8.1,5 The Group Vb elements 8.1.5.1 Phosphorus Phosphorus - an element essential to life - occurs in various orthophosphate minerals, in marine sediments notably in fluorapatite, SCagfPO^'CafF.CI^ (Cotton and Wilkinson, 1966). According to Turekian (1977), it exhibits the following order of abundance in some important rock types : Maf > Sh > Gran > Ls > Umaf > Ss (Umaf=ultramafic rocks). 301 A comparison of the robust - Box-Cox - P-means corresponding to the four groups of samples (Table 8.1) shows that the highest average P concentration occurs in the Strymonikos group of sediments. The regional distribution of P in the surface sediments of the study area shows that, generally, it is the finer-grained material, that mainly represents the accumulation of continental detritus derived via the rivers and transported by current action, that is enriched in P. This is in agreement with P belonging to the elemental group reflecting terrigenous detrital material, in the cluster and factor analysis. Partition analysis of selected samples gives evidence that the major part of the P remains in the residual fraction; non-residual P being mainly associated either with organic matter or the reducible fraction. The latter association demonstrates the ability of phosphate anions to coprecipitate onto the surface of Fe and Mn oxides (Bortleson and Lee, 1974); according to Froelich et al. (1982) ferric hydroxides can adsorb P from the water column. Partition analysis has also shown that higher than usual proportions of P occur in the non-residual fraction in samples collected from offshore of relatively higher populated districts. This may reflect presence of agricultural and urban pollutants; P compounds being commonly used as fertilizers and detergents, and occurring in sewage. In buried sediments, the downcore P variation exhibits, in general, a reverse relation to Ca. This is likely to largely reflect P occurrence in clay minerals and, therefore, antipathy for carbonates. The downcore distribution pattern of P also shows a surface enrichment, supporting an anthropogenic supply of P in the sediments. In conclusion, the distribution of P in the sediments of the study area is mainly controlled by the variation of the finer-grained material; P occurring mostly in P-bearing resistant minerals. 8 .1.6 The Group Hb elements 8 .1.6.1 Zinc Zinc is a metal of relatively low abundance in nature, occurring in a number of minerals, such as zinc blende (ZnS), smithsonite or calamine (ZnCOg), willemite (Zn2SiC>4), zincite (ZnO), and several others (Cotton and Wilkinson, 1966). Zinc is 302 associated with Cu and Pb in base metal deposits, and with Mg in some silicates. It can be sorbed by clay minerals, organic matter, as well as on Fe-, Mn- and Al- hydroxides (Rose et al. 1979). The order of Zn abundance in some important rock types is Sh > Maf > Umaf > Gran > Ss > Ls (Wedepohl, 1969-1978). A comparison of the robust - Box-Cox - Zn-means of the four groups of sediments (Table 8.1) shows that the highest mean is found in the lerissos group. It is also in this group that a comparatively large discrepancy between the raw - and the robust - Box-Cox-Zn- mean occurs; implying that Zn-anomalies are likely to exist within the lerissos data set. In lerissos gulf, a coastal Zn enrichment was determined in both onshore and offshore surface sediments collected from the northwest part of the gulf; the Asprolakkos river sediments also having, relatively, high Zn contents. In addition, Fe, Mn and Cu enrichments occur in the sediments of the area concerned. Applying multivariate statistics to the lerissos geochemical data set, a multivariate anomalous component consisting of Zn-Mn-Cu-Fe was noticed. This component is regarded as reflecting mineralization. The sediments of the northwest part of lerissos gulf exhibit, relatively to the other sediments studied, high factor scores on the "mineralization component". On the basis of the above, the Zn enrichment in the sediments of the northwest part of the gulf is thought likely to reflect the influence of the mainland mineralization (consisting of mixed sulphide deposits, Cu-bearing porphyry stocks, and Mn-oxide deposits) on the offshore sediment composition. Coastal erosion and river runoff are likely to supply the mineralization components to the sea. In addition, an offshore discharge of material from the mineral processing plant operating close to the sea, just outside Stratoni town, also supplies Zn rich sediment to the coastal area. By studying buried sediments collected off Stratoni, it was found that a dramatic increase in Zn, Fe and Mn contents occurs close to the surface (Sakellariadou, 1985). This increase indicates that the metal enrichment in the surface sediments is mostly man-made (see also section 8.2.1.4). In the part of the study area lying west of Thassos, two local coastal Zn enrichments were determined : (a) in the sediments lying close to Olympias; and (b) in the sediments along the west and southwest coast of Thassos island. The former represents material deriving from mixed sulphide deposits on the adjacent land, and the latter reflects coastal deposition of the weathering products of the Pb-Zn ore-bearing horizons, which occur on the island. 303 Elsewhere within the study area, Zn is enriched in the finer-grained material that represents continental detritus supplied mainly by rivers, and transported by current action. Partition analysis indicates that the major part of the Zn remains in the residual solid; Zn being, therefore, mostly present in clay minerals. In accord with this is Zn belonging to the elemental group reflecting clay material on statistical treatment of the geochemical data. Considering the non-residual Zn in the sediments, the proportion of Zn bound to the reducible fraction is of primary importance (according to Balistrieri and Murray (1984), Zn can be scavenged from sea water by Mn oxides), and that hosted in the organic/sulphide fraction of secondary. In buried sediments, the major part of the Zn occurs in Zn-bearing resistant minerals; the non-residual Zn is, mainly, associated with the reducible and organic/sulphide hosted fractions. The downcore distribution pattern of Zn varies between the various buried sediments examined. In conclusion, the distribution of Zn in the sediments of the study area shows that it is largely enriched in the finer-grained material that represents terrestrial detritus supplied by rivers. Additionally, Zn contamination is occurring in sediments collected from coastal areas near to which mixed sulphide deposits exist. ..JjaiisitiQn_etemeDte 8.2.1 The elements of the first transition series 8.2.1.1 Titanium Titanium is a metal relatively abundant in the earth's crust, 0.6% (Cotton and Wilkinson, 1966). The main minerals are ilmenite, FeTiOg, and rutile, one of the several crystalline varieties of T i0 2. Titanium occurs in shales and igneous rocks Turekian and (Wedepohl, 1961); it replaces Al in six-coordination and it, therefore, appears in pyroxene, hornblende and biotite (Mason, 1966). As Table 8.1 shows, the highest robust - Box-Cox - Ti-mean corresponds to the Strymonikos group of samples. 304 In the surface sediments of the study area, the regional distribution of Ti is similar to that of Fe. On the Samothraki plateau, high Ti concentrations were determined in sediments collected from the northeast part of the region, close to the Evros river mouth, and also from the area to the northeast of the Nestos river mouth. The Ti rich sediments of both these areas show also an enrichment in the heavy mineral content of their coarse fraction. Consequently, the Ti enrichment can be attributed to presence of Ti in heavy minerals. However, as Fig. 3.15, 3.5 and 3.3 show, the area of Ti enrichment in the northeastern part of Samothraki plateau, extends much further westwards than do the areas of heavy mineral and rock fragment enrichment in the sediment coarse fraction. This implies that, in the area concerned, Ti is likely to occur in phases of a finer particle size in the west than in the east. This is in accord with the suggestion that this material is current-transported from the Evros river area, rather than derived from coastal erosion. Sediments collected off Samothraki island contain high Ti concentrations, and their coarse fraction exhibits an enrichment in rock fragments but not in heavy minerals. Therefore, these Ti high values are attributed to the relative abundance of Ti rich rock fragments. In the part of the study area lying to the west of Thassos island, Ti enrichments occur predominantly in Fe and Al rich, fine-grained sized material, and show a lack of coincidence with areas of high heavy mineral content. Consequently, it seems that west of Thassos, either Ti bearing heavy minerals are absent or they occur in the finer sediment fraction. High Ti concentrations were determined in the silt- and clay-sized sediments over much of the Strymonikos gulf and plateau areas, where continental detritus has been supplied by Strymon river (the river samples are relatively Ti enriched) and distributed through the west-east trending submarine canyon. The silty sediments collected from the northern part of Kavala gulf contain relatively high amounts of Ti, being poor in heavy minerals in their coarse fraction. As no major river currently flows into the gulf, the Ti enrichment in the surface sediments concerned is regarded as the result of the reworking of earlier sedimentary horizons. These horizons are likely to be enriched in heavy minerals, deposited by the Nestos river during an earlier period. Partition analysis gave evidence that the bulk of Ti occurs in the residual fraction, suggesting its presence in terrigenous detrital material. Cluster and factor analysis applied to the geochemical data show that Ti is strongly correlated with Cr, Mg, V, Fe, Ni and Mn; Ti being present in the elemental group regarded as representing the 305 mafic/ultramafic detrital component of the sediments. Sediments collected from the south coastal area of the lerissos gulf, east of Nea Roda, have high scores on the mafic/ultramafic factor; they contain up to 4,500 p.g/g Ti, 6,200 jxg/g Cr and 160 jig/g Ni. This enrichment (as already discussed in section 8.1.2.2) reflects the offshore deposition of material derived from the ultramafic intrusion on the mainland. In buried sediments, the elements Ti, Cr, Mg, V and Ni follow similar downcore distribution patterns; reflecting their common origin - weathering products of mafic/ultramafic rocks. The bulk of Ti in buried sediments occurs in the residual fraction. In conclusion, the distribution of Ti in the sediments of the Samothraki plateau generally shows a correspondence with the heavy mineral content of the sediment coarse fraction; a local association with the abundance of rock fragments also occurs. In the part of the area west of Thassos, Ti exists predominantly in fine-grain-sized (silt and clay grade) material. All over the study area, Ti is supplied through continental detrital material. 8.2.1.2 Vanadium Vanadium has an abundance of about 0.02% in the earth's crust (Cotton and Wilkinson, 1966). Its more important minerals are vanadinite (Pb5(VC>4)3CI), and carnotite (K (U 02 )V 0 4 *3/2 H2 0). Vanadium is a major constituent of mafic minerals; it occurs in pyroxene, amphibole and biotite, substituting mainly for Fe+3 (Mason, 1966), and it is associated with Fe-oxides and organic matter (Rose et al., 1979). A comparison of the four groups of samples shows that the highest robust - Box-Cox - V-mean occurs in the Strymonikos and lerissos groups (Table 8.1). The regional distribution of V in the surface sediments is, in general, similar to that of Fe; V being mainly enriched in the finer-grained material. Applying cluster and factor analysis to the geochemical data, V has been found to be strongly associated with Fe, Ni, Mg, Cr and Ti; vanadium being present in the factor considered to represent residual products of mafic/ultramafic rocks. Partition analysis of selected samples gave evidence that more than 80% of total V occurs in the residual fraction; therefore, V is mainly hosted in the lattice structure of resistant minerals. 306 In buried sediments, the behaviour of V is similar to that exhibited in surface sediments. Vanadium follows Fe in its downcore distribution; the major part of V (>80%) remains in the residual fraction. In conclusion, the distribution of V in the sediments of the study area is to a large extent influenced by the variation of Fe. Vanadium is mainly supplied through the weathering products of mafic/ultramafic rocks; it occurs in resistant V-bearing minerals. 8.2.1.3 Chromium Chromium is a lithophile metal, strongly associated with Ni and Mg in ultramafic rocks (Mason, 1966). Its chief ore is chromite, F e C ^ O ^ In sediments, Cr is largely concentrated in micas and clay minerals, particularly illite (Frblich,1960, cit. in Chester and Hughes, 1969) substituting for aluminium (Mason, 1966). Table 8.1 shows that the highest robust - Box-Cox -Cr-mean, among the four groups of samples, occurs in the lerissos and Strymonikos data sets. The regional distribution of Cr in the surface sediments follows, in general, the variation of Fe, being similar to that of Ti. Thus, on the Samothraki plateau, Cr is enriched in the sediments collected from its northeastern part, close to the Evros river mouth, and also from the area to the northeast of the Nestos river mouth; the coarse fraction of sediments from both these areas being enriched in heavy minerals. This suggests, therefore, that in the sediments concerned, Cr mainly occurs in heavy minerals. However, as far as the sediments of the northeastern part of Samothraki plateau are concerned, Cr enrichment occurs in phases of a finer particle size in the west than in the east, for reasons already discussed in section 8.2.1.1, in regard to Ti. An enrichment in Cr was determined off Samothraki island; it can be attributed to the high, relatively, abundance of rock fragments in the area (see also section 8.2.1.1). In the part of the study area lying to the west of Thassos island, Cr is enriched in Fe- and Al- rich fine-grain-sized material; these sediments are poor in heavy minerals. Therefore, Cr-bearing heavy minerals are either absent from the area concerned, or they occur in the finer sediment fraction. The Cr enrichment in the silty sediments of the northern part of Kavala gulf represent older reworked sediments, deposited by the Nestos river during an earlier period (see also section 8.2.1.1). A high Cr content, associated with Mg, Ti and Ni enrichments, was found in the coastal sediments east of 307 Nea Roda, and in those collected from the corresponding river. These enrichments represent coastal deposition of the weathering products from the Nea Roda ultramafic rocks. The chromite mineralization present in these rocks is also reflected by the enrichments detected. Partition analysis shows that the major part of Cr occurs in the residual solid. Therefore, the primary source of Cr in the sediments studied is Cr-bearing resistant minerals. The remaining part of Cr has been found to be hosted either in the reducible or in the organic/sulphide fraction. Applying cluster and factor analysis to the geochemical data, Cr occurs in the factor regarded as representing the mafic/ultramafic detrital component of the sediments; Cr being strongly correlated with Ti, Mg, V, Fe, Ni and Mn. In buried sediments, Cr exhibits a downcore distribution pattern similar to those presented by Ti, Mg, V and Ni; Cr occurring mostly in Cr-bearing resistant minerals. Consequently, Cr has similar behaviour in the surface and buried sediments studied. In conclusion, Cr in the sediments studied occurs mainly in minerals that derive from the weathering of mafic/ultramafic rocks. The distribution of Cr is similar to those of Fe and Ti. In the Samothraki plateau, Cr enrichments correspond either to a relatively high heavy mineral content or to an abundance of rock fragments in the sediment coarse fraction. In the part of the study area lying to the west of Thassos island, Cr is mainly enriched in silt- and clay- sized sediments. 8.2.1.4 Manganese Manganese is a relatively abundant metal, constituting about 0.085% of the earth's crust (Cotton and Wilkinson, 1966). It is a lithophile element occurring in most mafic minerals; its most important deposit is pyrolusite, Mn02. Table 8.1 shows that within the four data sets, the highest robust - Box-Cox - Mn-mean corresponds to the lerissos gulf group of sediments. It is also in this group that a large discrepancy exists between the raw- and the robust - Box-Cox - Mn-mean. The latter implies that Mn anomalies are likely to occur in the lerissos data set. The distribution of Mn in the surface sediments of the lerissos gulf shows that an exceptionally high content occurs in those collected from the northwest part of it, off 308 Stratoni. These sediments are also enriched in Fe and Zn. Applying factor analysis to the lerissos gulf geochemical set, it was found that the sediments off Stratoni exhibit high scores on the factor loaded with Mn, Zn, Cu and Fe, which is regarded as a mineralization factor (see section 5.1.3.3). Thus, the Mn enrichment in the sediments of the northwest part of lerissos gulf is considered to represent the influence of the mineralization (mixed sulphide deposits, Cu-bearing porphyry stocks and Mn-oxide deposits) on the offshore sediment composition, and the nearshore sediment contamination by the mining operation based close to the sea, just outside Stratoni. Examining buried sediment off Stratoni, it was found that the Mn content - as well as the Fe and Zn contents - increases dramatically close to the surface. In addition, partition analysis gave evidence that the non-residual proportions of Fe, Mn, Ni and Cu decrease with depth in the uppermost 16 cm of core I 14, collected off Stratoni. The above results confirm the contamination problem attributed to the coastal mining activities already mentioned. The Mn enrichment in the upper part of the buried sediments might also be attributed to diagenetic Mn remobilization. In the rest of the study area, Mn enrichments were found in fine-grained sediments over much of the Strymonikos gulf and plateau, on the outer part of the Strymonikos plateau, and close to the Evros river mouth. Two local coastal Mn enrichments were also found in Olympias bay sediments, and in Limenaria bay sediments. The former, associated with a high.Zn content, is related to the abundant Mn-oxides which are present along the ore host horizon of the Olympias Zn-Pb orebody. The latter, associated with a high Fe content,is attributed to weathering products which derive from the Fe, Mn-oxide mineralization, that according to Vavelidis and Amstutz (1981) is associated with the Pb-Zn ore deposits on Thassos island. Partition analysis shows that most of the Mn occurs either in the reducible phase or in the residual solid. Therefore, the largest Mn reservoirs in the sediments studied are reducible ferromanganese oxides and detrital Mn-bearing minerals. A small proportion of Mn (1-2% of the total) was found to be leached by NH4OAc. This exchangeable Mn, probably representing a small amount of Mn present in a reduced form, could be adsorbed on clay minerals. In addition, partition analysis has indicated a small biogenic contribution of Mn, such contribution has also been noted by Shearme et al. (1983) in sediments from the Tag hydrothermal field. Finally, some Mn was found to be hosted in the organic/sulphide fraction. Applying cluster and factor analysis to the geochemical data, Mn belongs to the elemental group regarded as representing the weathering products of mafic/ultramafic rocks. Manganese exhibits a high correlation coefficient with Fe, demonstrating the similar behaviour of these 309 metals. In buried sediments, the major part of Mn occurs in residual minerals. The rest of it is mainly hosted in the reducible fraction, while smaller proportions are associated with the other phases present. In some buried sediments, a Mn enrichment was determined in the top layer of the cores. It most probably indicates an upward migration of Mn due to post-depositional remobilization (Bonatti et al., 1971); according to Brannon et al. (1977) this is the case when reducing conditions underlie a more oxidized surface sediment zone. The Mn enrichment in the uppermost part of the sediment column is a cyclical process. In areas where reducing conditions prevail below the sediment-water interface, Mn4+ in manganese dioxide minerals is reduced on burial to divalent manganese which passes into solution in the interstitial waters of the sediments. It then migrates upwards to become reoxidized in the oxidizing upper sediments and reprecipitates as manganese dioxide (Cronan, 1980). Changes of the Mn content of sediments with depth, due to diagenetic recycling of Mn have also been reported by Lynn and Bonatti (1965), Froelich et al. (1979), and Sawlan and Murray (1983). In conclusion, Mn in the sediments of the study area is mainly supplied in terrigenous detrital material that derives from the weathering of mafic and ultramafic rocks, and occurs in the finer-grained sediments. Locally, in coastal sediments, Mn derives from Mn-oxide deposits of the mainland. Manganese also exhibits some diagenetic remobilization from buried sediments towards the surface in a few cores. 8.2.1.5 Iron Iron is the second most abundant metal, after Al, and the fourth most abundant element in the earth's crust (Cotton and Wilkinson, 1966). The major iron minerals are hematite, Fe2 O3, magnetite, FegO^ limonite, FeO(OH), and siderite, FeCOg. Iron is a siderophile element; in the earth's crust it is chalcophile and lithophile. Iron is a major component of mafic minerals. It exhibits the following order of abundance in some important rock types : Umaf > Mat > Sh > Gran > Ss > Ls (Turekian, 1977). Iron, when in a trivalent state, exhibits a very low mobility due to its precipitation as hydrous iron oxide (Rose et al., 1979). In a reduced sediment zone, Fe can also be immobilized as sulphide, if the Eh is low enough for sulphate reduction (Bonatti et al., 1971). In weathering products, Fe occurs mostly in the form of iron oxides (Rose et al., 1979). 310 Table 8.1 shows that among the four groups of sediments, the highest robust - Box-Cox - Fe- mean occurs in the Strymonikos group. In the lerissos gulf data set, the robust - Box-Cox -Fe-mean differs significantly from the mean of the corresponding raw data; implying, therefore, that Fe anomalies occur within the lerissos gulf data. Throughout most of the study area, the regional distribution of Fe in the surface sediments is similar to the corresponding variation in Al content. Both elements are mainly enriched in sediments of finer grain size. Thus, as already discussed for Al in section 8.1.3.1, the Fe distribution pattern reflects : a. the influence of the Strymon river in supplying aluminosilicate rich material to the Strymonikos gulf, and the importance of the west-east trending submarine canyon in distributing it downslope; b. the reworking of older sediment deposited by the Nestos river during an earlier period, as far as the silt- and clay-sized Al and Fe rich sediments of Kavala gulf are concerned; c. the influence of Evros river in supplying fine-grained Al and Fe rich continental detritus to the offshore area, and the importance of the strong coastal current in transporting it westwards; d. the Filiouris river runoff; and e. the reworking of older deeper sedimentary horizons, close to the Nestos river mouth. A coastal Fe enrichment has been determined in the sediments of Limenaria bay. However, these sediments are, relatively, poor in Al; this suggests that the Fe is not associated with clay minerals in the coastal region concerned. On the contrary, these Al poor sediments are, apart from Fe, rich in Mn and Ni. Iron enrichment in Limenaria bay sediments is considered to be related to the mixed sulphide mineralization present on the island; this mineralization occurs in an area drained by a small river that flows into the bay. Furthermore, mineral processing plants operated by the Germans during the second world war on the coast of Limenaria bay, may have resulted in a corresponding offshore sediment contamination. 311 In the lerissos gulf group of samples, a discrepancy between the distribution of Fe and Al occurs, Fe presenting exceptionally high contents in the area off Stratoni. This enrichment reflects (on the basis of the results obtained from factor analysis, downcore bulk Fe distribution, and partition analysis) the influence of both the mineralization of the adjacent mainland rocks and the mining operation on the offshore sediment composition (see section 8.2.1.4). Partition analysis gave evidence that the major part of Fe is hosted in the residual solid. Therefore, most of the Fe in the sediments studied is present in resistant Fe-bearing minerals, such as heavy minerals, silicates and clays. Applying factor analysis, Fe belongs to the factor regarded as representing clay minerals. Iron was found to exhibit high correlation coefficients with Mg, Cr, V and Ni; this indicates an iron supply through the residual products of mafic/ultramafic rocks. In fact, the coastal samples collected off the west coast of the lerissos gulf and those just east of Nea Roda have high factor scores on the supposed mafic/ultramafic factor; they correspond with presence of amphibolites and ultrabasic igneous rocks, respectively, on the mainland. Partition analysis has also shown that a small amount of the total Fe occurs in a reducible form, and a small proportion is hosted in the organic/sulphide fraction. Furthermore, as far as the organic hosted fraction is concerned, ESR spectra analysis of sediment humic acid preparations has given evidence that from one to four different geometrical forms of ferric iron are associated with and/or bound to sediment humic acids. In buried sediments, Fe has a similar behaviour to that exhibited in surface sediments. Iron varies with depth sympathetically to Mg, Cr and V, demonstrating that these elements have a common origin from mafic/ultramafic rocks. Furthermore, the downcore distribution pattern of Fe is similar to that of Al. This indicates that Fe is accommodated in clay matter, part of which, apparently, comprises weathering products of the mafic/ultramafic rocks. Similar results were obtained by Chester and Hughes (1966) in a North Pacific deep sea clay core; they have shown that Fe is mainly present in the clay and non authigenic phases of the sediment. In conclusion, the distribution of Fe in the sediments of the study area - being similar to that of Al - largely reflects the distribution of finer-grained sediments deriving from terrigenous detrital material. Iron, in the form of resistant Fe-bearing 312 minerals (like Fe,Mg-minerals and their alteration products) is supplied mainly by modern river runoff, to a lesser extent by reworking of older sediment, and locally by coastal erosion. The high Fe contents in the sediments off Stratoni indicate man-made contamination by mining operations there. 8.2,1,6 Cobalt Cobalt is a chalcophile element; in the earth's crust, it is chalcophile and lithophile. It occurs primarily in mafic minerals. It always occurs in nature in association with Ni, and usually also with As (Cotton and Wilkinson, 1966). Its most important minerals are smaltite, CoAs2 , and cobaltite, CoAsS. Cobalt can be sorbed by clay minerals, and hosted in Mn and Fe phases. It also exhibits a close connection with organic matter (Rose et al., 1979). The sediments of the study area contain small amounts of Co. The distribution of Co in the surface sediments shows that a local coastal enrichment occurs east of Nea Roda, associated with high Mg, Cr and Ni contents. This enrichment is attributed to detrital material deriving from the ultrabasic intrusion that outcrops on the mainland. Cluster and factor analysis have shown that : (a) in the lerissos gulf sediments, Co is associated with metals deriving from mafic/ultramafic rocks; (b) in the Samothraki plateau sediments, it shows an affinity for base metal* and (c) in the Strymonikos and Kavala groups of samples, it forms a class of its own. Partition analysis shows that the major part of Co occurs in the residual fraction. Cobalt is, therefore, likely to be mainly present in the lattice of clay minerals; according to Chester and Hughes (1967, 1969), Co in the monovalent state, as CoCI+ or Co(OH)+ , may substitute for hydrogen in the (OH) groups of the clay mineral lattices. The remaining portion occurs mainly in the organic/sulphide fraction, and to a less extent in the reducible fraction. In buried sediments, the downcore distribution of Co, in general, does not follow that of any other element and its concentration levels are low. Cobalt is mostly present in the residual fraction. In conclusion, in the sediments of the study area, Co mainly occurs in Co-bearing resistant minerals, and is present in low concentrations. Z PI a nd C U 313 8.2.1.7 Nickel Nickel is a siderophile element; in the earth's crust, it is chalcophile and lithophile. It is relatively immobile. Nickel occurs primarily in mafic minerals. It is associated with Mg and Co in ultramafic and mafic rocks. In sulphide deposits, Ni is associated with Cu (Rose et al., 1979). Table 8.1 shows that among the four groups of sediments, the Strymonikos and lerissos groups exhibit the highest robust - Box-Cox - Ni-means. In the lerissos gulf group of samples, the raw - Ni-mean differs significantly from the robust - Box-Cox - Ni-mean, implying that in that data set, Ni values higher than those corresponding to the background population occur. The regional distribution of Ni in the surface sediments of the study area shows that it is in general enriched in the finer-grained sediments. Partition analysis gave evidence that the major part of Ni occurs in the residual solid. Consequently, its principal route of supply to the sediments in this region is Ni-bearing detrital minerals. This is compatible with Ni occurring in the elemental group considered to represent detrital material of mafic/ultramafic rocks in the statistical analysis, Ni being highly correlated with V, Mg, Fe, Cr and Mn. In particular, the high Ni content in coastal sediments east of Nea Roda is attributed to material that derives from the ultrabasic intrusion outcropping on land; these sediments also contain high amounts of Mg, Cr, Ti and Co. The non-residual Ni of the surface sediments is mainly distributed between the reducible and the organic/sulphide hosted fraction. In buried sediments, Ni mostly occurs in the residual solid, implying that its principal reservoir is Ni-bearing detrital minerals. In some of them, the downcore variation of Ni exhibits similarities with the patterns of Mn and Fe. However, in several cores Ni follows Fe in preference to Mn; a possible explanation for this could be that Ni, being partly captured by Fe hydroxides and sulphides, is prevented from following Mn in its upwards migration, where the latter occurs (Bonatti et al., 1971). In general, Ni is supplied to the sediments of the study area by terrigenous detrital material and it is accommodated in weathering products of mafic/ultramafic rocks. It occurs in the finer-grained sediments that contain material supplied by river runoff and coastal erosion. 314 8.2,1.9__Copper Copper is a siderophile element; in the earth's crust, it is chalcophile and lithophile. It is widely distributed in nature in the free state, in sulphides, arsenides, chlorides, mafic minerals and carbonates (Cotton and Wilkinson, 1966). It is present in chalcopyrite, CuFeS2 , bornite, Cu5FeS4 , chalcocite, Cu2S, cuprite, Cu20 , and complex Cu-As-Sb-S minerals (Rose et al., 1979). In sulphide deposits, Cu is associated with Pb, Zn and Ni. In weathering products, Cu may occur in sulphides, oxides, basic carbonates, sulphates, Mn-oxides, limonite, and organic matter (Rose et al., 1979). Copper exhibits, in general, a strong affinity for organic matter (Fraser, 1961). Table 8.1 shows that among the four groups of samples, the lerissos gulf data set exhibits the highest robust - Box-Cox -Cu-mean. In this data set too, the largest difference between the robust - Box-Cox - and the raw - Cu-mean occurs. The latter implies that, within the lerissos gulf sediments, Cu anomalies are likely to occur. The regional variation of Cu in the surface sediments of the study area indicates that Cu is mainly enriched in the finer-grained sediments. Hence, on the Samothraki plateau, a high Cu content was determined : (a) in the fine-grained terrigenous detrital material supplied by the Evros river and transported westwards by current action; (b) in the fine-grained Al rich deep water sediments southeast of Thassos island; (c) in the silty material close to Nestos river mouth; and, (d) in both Filiouris river sediments and those off its mouth. In the west part of the study area, high Cu values were determined in the silt- and clay-sized material over much of the Strymonikos gulf and plateau. As far as the lerissos gulf sediments are concerned, it was found that most of them are, relatively, rich in Cu. The highest Cu content occurs in sediments collected from the northwest part of the gulf; most of them have , in addition, high scores on the supposed mineralization factor. This Cu enrichment is attributed to the nearshore deposition of weathering products deriving from the Cu-bearing porphyry stocks and the mixed sulphide deposits occurring in the amphibolitic rocks of the adjacent mainland. Partition analysis shows that the major part of Cu occurs in the residual fraction. This indicates that, in the sediments studied, Cu is mainly hosted in the lattice structure of resistant minerals. According to Graybeal and Heath (1984), Cu can be removed from sediment pore water and be absorbed into the lattices of clay minerals. 315 The non-residual Cu is primarily associated with the organic/sulphide hosted fraction implying (a) copper affinity for organic matter, particularly in the Samothraki plateau sediments, and (b) a possible occurrence of traces of copper sulphide, especially in the lerissos gulf sediments which are poor in organic matter. The strong affinity of copper ions for organic matter was confirmed by the significant proportions of total Cu forming complexes with sediment humic acids. Copper ions exhibit a high adsorption on and/or complexation with sediment humic acids, by either partial replacement of other metal ions or binding with organically free humic functional groups. This property of copper ions needs to be considered when copper transport and immobilization phenomena in the marine environment are considered. Artificially formed complexes of Cu(ll) with sediment humic acids have a high stability toward exhaustive water washing (Senesi and Sakellariadou, 1987). Partition analysis has also shown that a significant portion of the non-residual Cu is hosted in the reducible phase, indicating a Cu presence in reducible ferromanganese oxides. In buried sediments, the major part of Cu is* hosted in the lattice structure of Cu-bearing resistant minerals. The non-residual Cu mostly occurs in the organic/sulphide hosted fraction. In several cores, a Cu enrichment in their uppermost parts was detected, which suggests that pollution occurs in the respective areas. In conclusion, Cu in the sediments of the study area mainly occurs in resistant minerals. It is supplied in the form of terrigenous detrital material, mostly transported by river runoff. In addition, in some areas, a Cu supply from pollution was found. The high affinity of Cu for organic matter has also been confirmed. 8.2.2 The elements of the second transition series 8.2.2.1 Zirconium Zirconium is a lithophile element. It occurs widely over the earth's crust, but not in very concentrated deposits. Its major minerals are zircon, ZrSiC^, and baddeleyite, a form of Zr02 (Cotton and Wilkinson, 1966). Zircon is a common accessory mineral in many sediments, often surviving more than one cycle of weathering and 316 sedimentation (Deer et al.f 1966). The order of its abundance in some important rock types is : Ss > Gran > Sh > Maf (Turekian, 1977). Table 8.1 shows that within the four groups of sediments, the highest robust - Box-Cox -Zr-mean occurs in the Strymonikos group. The distribution of Zr in the surface sediments of the study area shows that, on the Samothraki plateau, most Zr enrichments correspond to a relatively high heavy mineral content in the sediment coarse fraction. In the western part of the northeastern Samothraki plateau, Zr is likely to occur in phases of a finer particle size than in the eastern part. The latter is suggested as the area of Zr enrichment in the sediments concerned extends much further westwards than do the areas of heavy mineral and rock fragment enrichment in the sediment coarse fraction (see also section 8.2.1.1). In the part of the study area lying west of Thassos island, the sediments found to be enriched in Zr do not exhibit heavy mineral enrichment in their coarse fraction. It is, therefore, suggested, that in this region Zr-bearing phases do not occur in the coarse sediment fraction, but are more concentrated in the fine fraction. High Zr concentrations were determined in sandy coastal sediments collected from the northern part of the Strymonikos plateau. As Zr, in the form of zircon, is known to be a primary constituent of igneous rocks, especially the more acid rocks (Read, 1970), these coastal enrichments of Zr may represent offshore deposition of material deriving from the erosion of the granites occurring on the mainland. Furthermore, coastal Zr enrichments in sandy sediments of the Strymonikos gulf could represent an offshore reworking of beach sands, the latter being previously enriched in Zr by coastal erosion. By statistically treating the geochemical data, Zr, being strongly associated with Si and Ba, belongs to the factor considered to represent detrital material of felsic rocks. In some buried sediments, Zr was found to follow Si in its downcore distribution; suggesting, therefore, that zircon is likely to occur in these sediments. Zircon was also recognized in some surface sediments, studied by S.E.M. In conclusion, the distribution of Zr indicates that Zr-bearing phases are mostly present in the sediment coarse fraction, on the Samothraki plateau. This is not the 317 case, as far as the major part of the remaining area is concerned, where Zr occurs in the sediment fine fraction. SJLsL__ The Lanthanides 8.2.3.1 __ Lanthanum Lanthanum is a lithophile element that occurs principally in accessory minerals of igneous rocks (Rose et al.,1979). Its most important mineral in silicic igneous rocks is monazite (RE,Th)P04, which, on erosion, often concentrates as a heavy dark sand. The order of La abundance in some important rock types is : Gran > Sh > Maf > Ss > Ls = Umaf. Lanthanum can partly replace Ca2+ in apatite and hornblende, as well as in allanite which occurs in some granites and pegmatites (Mason, 1966). Table 8.1 shows that the concentrations of La in the sediments of the study area are in general similar throughout it, apart from the lerissos gulf sediments which contain relatively lower amounts of La. The regional distribution of La in the surface sediments shows that it is enriched in the coastal sediments collected eastwards . of Loutra Eleftheron (Fig. 4.26). These sediments have also a heavy mineral enrichment in their coarse fraction (Fig. 4.5). The area of the La rich coastal sediments concerned, is largely adjacent to a mainland occurrence of Tertiary acidic intrusive rocks, which according to Perissoratis et al. (in press) may contain monazite. By subjecting samples from these La rich coastal sediments to partition analysis, it was found that the bulk of La remains in the residual fraction, confirming that in these sediments, detrital material is the principal La source. Elsewhere, partition analysis has shown that up to half of the total La is hosted in the lattice structure of resistant minerals. The remaining portion has been distributed, primarily, in the reducible and, secondarily, in the organic/sulphide hosted fractions. In the surface sediments of the study area, La does not exhibit significant sympathetic behaviour towards the other elements examined, either in its regional distribution or in the elemental associations formed when cluster and factor analysis were applied to the geochemical data. However, in the lerissos gulf sediments, La was found to be associated with P. According to Bonatti et al. (1971), La in marine sediments can be hosted in phosphates. The formation of the La-P elemental group in the present 318 sediments might be regarded as representing a La,P-bearing phase. Both "fresh” apatite, deriving from skeletal remnants (Marchig et al., 1982), and "reworked" apatite, a weathering product of the leucocratic metamorphic rocks of the mainland (Demetriades, 1974), may contribute to this La,P-bearing phase; the presence of La in biogenic apatite is also supported by Kukal (1971). In buried sediments, La does not show an association with any of the other elements studied. The major part of it remains in the residual fraction, the rest being mostly associated with the organic/sulphide hosted fraction and a significant proportion being associated with the reducible phase. 319 Table 8.1 : Means (p.g/g) and robust means (jig/g) of raw data, and robust means (jig/g) of Box-Cox transformed data, of 20 elements determined in the four groups of samples. Element .Samothraki Kavala .Strymonikos leiissea plateau group flip,up oull K R 17,790 21,690 21,010 18,160 R--R 18,470 21,480 21,320 19,370 R-B-C 18,783 20,559 21,359 19,251 Be R 2 2 . 2 2 R--R 2 2 3 2 R--B-C 2 2 3 2 Mg R 13,080 11,510 13,310 17,440 R--R 12,560 11,600 13,710 15,000 R--B-C 12,427 10,889 13,062 10,864 Ca R 85,310 78,710 63,670 55,410 R--R 81,420 74,970 52,070 48,210 R--B-C 63,578 59,880 48,077 40,209 Sr R 654 656 464 440 R--R 628 609 369 407 R--B-C 461 492 338 313 Ba R 374 521 441 388 R--R 368 463 396 405 R--B-C 354 417 370 366 Al R 57,830 61,390 68,730 58,740 R--R 58,850 64,250 72,870 62,350 R--B-C 56,885 65,605 73,139 61,888 Si R 224,500 230,200 228,800 232,500 R--R 225,200 228,900 230,400 243,200 R--B-C 228,202 226,578 227,406 236,100 P R 439 413 513 410 R--R 427 411 519 420 R--B-C 415 408 498 419 Ti R 2,503 2,146 2,947 2,358 R--R 2,492 2,231 3,112 2,517 R--B-C 2,478 2,209 3,106 2,352 V R 76 72 99 93 R--R 72 73 100 94 R--B-C 65 71 89 88 320 Table 8.1 (continued) Cr R 81 66 89 277 R--R 78 65 91 143 R--B-C 76 58 87 89 Mn R 381 273 418 1,560 R--R 360 274 424 1,398 R--B-C 354 282 413 710 Fe R 26,880 23,500 31,770 35,790 R--R 26,200 24,230 33,620 33,160 R--B-C 25,345 23,881 32,482 29,226 Co R 23 22 24 23 R--R 23 21 24 23 R--B-C 22 21 23 21 Ni R 45 42 52 97 R--R 42 39 50 67 R --B-C 36 36 46 45 Cu R 31 25 33 60 R--R 31 24 35 59 R--B-C 26 22 29 37 Zn R 90 75 81 586 R--R 69 68 85 493 R--B-C 52 61 77 195 Zr R 102 93 109 80 R--R 101 91 111 84 R--B-C 9 1 - 88 101 75 La R 40 44 42 30 R--R 40 41 39 30 R--B-C 40 41 38 31 R : means of raw data; R--R : robust means of raw data; R--B-C : robust means of Box-Cox transformed data. 321 8.3 Concluding statement The processes responsible for the supply of elements to the North Aegean Sea sediments are modern river runoff (e.g. Evros river to the Samothraki plateau, Strymon river to the Strymonikos gulf), coastal erosion (e.g. off Nea Roda, off Stratoni, off the coastine from Loutra Eleftheron to Nea Peramos), man-made pollution (i.e. mining and mineral processing activities at Stratoni), and reworking of older sediments deriving from deeper sedimentary horizons (e.g. at the mid-shelf region of Samothraki plateau, and in the gulf of Kavala). The element distribution and transport are affected by currents (e.g. the westwards flowing nearshore current in the northeastern part of Samothraki plateau, transporting the Evros river runoff) and bottom topography, like occurrence of canyons (e.g. the west-east trending canyon in the northern part of Strymonikos plateau, directing the Strymon river runoff downslope), ridges (e.g. Strymonikos ridge), and sills (e.g. the sill at the mouth of lerissos gulf, preventing the dispersion of the contaminant material beyond the gulf). Despite the various metal enrichments defined, it is very improbable that economic concentrations of the metals studied occur in the upper parts of the North Aegean Sea sediments. Therefore, future research in the same area should be focussed on deeper sedimentary horizons. An additional finding of the present study is that mineralized areas onland may be detected on the basis of the spatial distribution patterns of mineralization related elements, also considering parameters like the nature of the sediments, bathymetry and hydrography. 322 REFERENCES Abragam, A. and Bleaney, B., 1970. Electron paramagnetic resonance of transition ions. The International Series of Monographs on Physics, pp. 911. Balistieri, L.S. and Murray, J.W., 1986. The surface chemistry of sediments from the Panama Basin : The influence of Mn oxides on metal adsorption. Geochim. Cosmochim. Acta, vol. 50, pp. 2235-2243. Billett, M.F. and Nesbitt R.W., 1986. Base-metal mineralization associated with mafic and ultramafic rocks, Eastern Rhodope Massif, Greece. Trans. I.M.M., London, vol. 95, pp. B37-B43. Bonatti, E., Fisher, D.E., Joensuu, O. and Rydell, H.S., 1971. Post-depositional mobility of some transition elements, phosphorus, uranium and thorium in deep sea sediments. Geochim. Cosmochim. Acta, vol. 35, pp. 189-201. Bordovskii, O.K., 1965. Accumulation and transformation of organic substance in marine sediments - 3 : Accumulation of organic matter in bottom sediments. Mar. Geol., vol. 3, pp. 33-82. Bornovas, J. and Rondogianni-Tsiambaou, Th. ,1983. Geological Map of Greece. Bortleson, G.C. and Lee, G.F., 1974. Phosphorus, iron, and manganese distribution in sediment cores of six Wisconsin lakes. Limnol. Oceanogr., vol. 19 (5), pp. 794-801. Brannon, J.M., Rose, J.R., Engler, R.M. and Smith, I., 1977. The distribution of heavy metals in sediment fractions from Mobile Bay, Alabama. In: Chemistry of Marine Sediments. Yen, T.F. (ed.), Ann Arbor Science, pp. 125-150. Brown, F.S., Baedecker, M.J., Nissenbaum, A. and Kapplan, I.R., 1972. Early diagenesis in a reducing fjord, Saanich Inlet, British Columbia-Ill. Changes in organic constituents of sediment. Geochim. Cosmochim. Acta, vol. 36, pp. 1185-1203. Calvert, S.E. and Piper, D.Z., 1984. Geochemistry of ferromanganese nodules from DOMES site A, Northern Equatorial Pacific : Multiple diagenetic metal sources in the 323 deep sea. Geochim. Cosmochim. Acta, vol. 48, pp. 1913-1928. Campbell, N.A., 1980. Canonical variate analysis; some practical aspects. Unpub. Ph.D. Thesis, Univ. of London. Chapman, H.D., 1965. Methods of soil analysis. In : American Society of Agronomy. Black, C.A. (ed.), Madison, Wis., pp. 891-904. Chapman, R.P., 1976. Limitations of correlation and regression analysis in geochemical exploration. Trans. I.M.M., (London), vol. 85, pp. B279-B283. Chester, R. and Hughes, M.J., 1966. The distribution of manganese, iron and nickel in a North Pacific deep-sea clay core. Deep Sea Res., vol. 13, pp. 627-634. Chester, R. and Hughes, M.J., 1967. A chemical technique for the separation of ferro-manganese minerals and adsorbed trace elements from pelagic sediments. Chem. Geol., vol. 2, pp. 249-262. Chester, R. and Hughes, M.J., 1969. The trace element geochemistry of a North Pacific pelagic clay core. Deep Sea Res., vol. 16, pp. 639-654. Colley, S., Thomson, J., Wilson, T.R.S. and Higgs, N.C., 1984. Post-depositional migration of elements during diagenesis in brown clay and turbidite sequence in the North East Atlantic. Geochim. Cosmochim. acta, vol. 48, pp. 1223-1235. Cotton, F.A. and Wilkinson, G., 1966. Advanced inorganic chemistry. A comprehensive text, Second Edition. Interscience (publ.), pp. 1136. Coward, R.N., 1986. A statistical appraisal of regional geochemical data from the Southwest Pacific for mineral exploration. Unpub. Ph.D. Thesis, Imperial College, Univ. of London, pp. 279. Cronan, D.S., 1976. Basal metalliferous sediments from the Eastern Pacific. Geol. Soc. Am. Bull., vol. 87, pp. 928-934. 324 Cronan, D.S., 1980. Underwater minerals. Academic Press inc., London/New York/Toronto/Sydney/San Francisco, pp. 362. Davis, J.C.,1973. Statistics and data analysis in geology. Second edition. John Wiley and Sons, New York/Chichester/Brisbane/Toronto/Singapore, p.p.550. Demetriades, S., 1974. Petrological study of the migmatic gneisses and amphibolites of the areas of Rentina - Asprovalta - Stavros - Olympias, (in greek). Unpublished Ph.D. Thesis, University of Thessalonika, Greece, pp. 231. Drozdova, T.V., 1968. . The role of humic acids in concentrating trace elements in soils. Soviet Soil Sci., vol. 10, pp. 1393-1396. Emiliani, C., 1955. Mineralogical and chemical composition of the tests of certain pelagic foraminifera. Micropalaeont., vol. 1, pp. 377-380. Ertel, J.R. and Hedges, J.I., 1985. Sources of sedimentary humic substances : vascular plant debris. Geochim. Cosmochim. Acta, vol. 49, pp.2097-2107. Fraser, D.C., 1961. Organic sequestration of copper. Economic Geology, vol. 56, pp. 1063-1078. Friedman,G.M. and Sanders, J.E., 1978. Principles of Sedimentology. John Wiley and Sons, New York/Santa Barbara/Chichester/Brisbane/Toronto, pp. 792. Froelich, P.N., Klinkhammer, G.P., Bender, M.L., Luedtke, N.A., Heath, G.R., Cullen, D., Dauphin, P., Hammond, D., Hartman, B. and Maynard, V., 1979. Early oxidation of organic matter in pelagic sediments of the eastern equatorial Atlantic : Suboxic diagenesis. Geochim. Cosmochim. Acta, vol. 43, pp. 1075-1090. Froelich, P.N., Bender, M.L., Lendtke, N.A., Heath, G.R. and De Vries, T., 1982. The marine phosphorus cycle. Am. Jl. Sci., vol. 282, pp. 474-511. Graf, D.L., 1960. Geochemistry of carbonate sediments and sedimentary carbonate rocks. Part III - Minor element distribution. Illinois State Geological Survey, Circular 301, pp. 71. Graybeal, A.L. and Heath, G.R., 1984. Remobilization of transition metals in surfacial 325 pelagic sediments from the eastern Pacific. Geochim. Cosmochim. Acta, vol. 48, pp. 9 6 5-975. Gupta, S.K. and Chen, K.Y., 1975. Partitioning of trace metals in selective chemical fractions of nearshore sediments. Environ. Lett., vol. 10 (2), pp. 129-158. Howarth, R.J. and Sinding-Larsen, R., 1983. Multivariate analysis. In : Handbook of Exploration Geochemistry, vol. 2. Statistics and data analysis in geochemical prospecting. Howarth, R.J. (ed.), pp. 207-289. Hue, A.Y., Durand, N. and Jacquin, F., 1974. Characterisation des acides humiques de sediments marins tecents et comparaison avec leurs homologues terrestres, (in french). Bull. E.N.S.A.I.A., Nancy, vol. 16 (1-2), pp. 59-75. Jackson, M.L., 1958. Soil chemical analysis, Prentice-Hall, Englewood, N.J. Jenne, E.A., 1968. Controls on Mn, Fe, Co, Ni, Cu and Zn concentrations in soils and water : The significant role of hydrous Mn and Fe oxides. In : Trace inorganics in water, Advances in chemistry series, Gould, R.F. (ed.), pp. 337-387. Karamantzanis, M.I., Kaklamanis, N., and Traganos, G., 1977. Internal report No 1: Preliminary results from the examination of sand samples from Alexadroupolis, I.G.M.E., (in greek), pp. 11. Kerndorff, H. and Schnitzer, M.,1980. Sorption of metals on humic acid. Geochim. Cosmochim. Acta, vol. 44, pp. 1701-1708. Konispoliatis, N., 1984. Study of the sediments at Strymonikos gulf, (in greek). Unpub. Ph.D. Thesis, National Technical University of Athens, pp. 109. Kononova, M.M., 1966. Soil organic matter. Its nature, its role in soil formation and in soil fertility. Pergamon Press : Oxford/London/Edinburgh/New York/Toronto/Sydney/Paris/Braunschweig, pp. 544. Krumbein, W.C., 1936. Application of logarithmic moments to size-frequency distributions of sediments. Jour. Sedimentary Petrology, vol. 6, pp. 35-47. Kukal, Z., 1971. Geology of recent sediments. Petr&nek, J. (ed.), Academic Press 326 Inc., London, pp. 490. Li, Y., 1982. Interelement relationship in abyssal Pacific ferromanganese nodules and associated pelagic sediments. Geochim. Cosmochim. acta, vol. 46, pp. 1053-1060. Lynn, D.C. and Bonatti, E., 1965. Mobility of manganese in diagenesis of deep-sea sediments. Mar. Geol., vol. 3, pp. 457-474. Manskaya, S.M. and Drozdova, T.V., 1968. Geochemistry of organic substances. Shapiro, L. and Breger, I.A. (eds.), Pergamon, International Series of Monographs in Earth Sciences, vol. 28. Marinos, G., 1955. Deposits of Ti magnetite sand near Alexandroupolis, (in greek). I.G.M.E. Bull., Athens, vol. 1, pp. 13. Marinos, G., 1982 . Greece. In : Mineral deposits of Europe, vol. 2 : Southern Europe. Dunning, F.W., Mykura, W. and Slater, D. (eds.), I.M.M. and Mineralogical Society, pp. 233-253. Marinos, G. Mariolakos, H. and Sabo, B., 1976. Titan-magnetite sand deposits of the beach of Alexandroupolis bay (Thrace, Greece), (in greek). Thalassographica, vol. 1, pp. 1-20. Markoulis, M., Orfanos, B., Kaklamanis, N. and Takousis, D.,1978. Industrial minerals and rocks. Research for heavy minerals at sands of the coastal areas of Eastern Macedonia and Thraki, (in greek). Unpubl. report, I.G.M.E., Athens, pp. 75. Marchig, V., Gundlach, H., Mbller, P. and Schley, F., 1982. Some geochemical indicators for discrimination between diagenetic and hydrothermal metalliferous sediments. Marine Geology, vol. 50, pp. 241-256. Mason, B.1966. Principles of geochemistry. Third edition, John Wiley and Sons, Inc., New York/London/Sydney, pp. 329. Me Donald, E.H., 1979. The mineral sand deposits of Thrace and Macedonia. Unpubl. report on the project UN/TCD/GRE/77/007, pp. 28. 327 Moorby, S.A., Cronan, D.S., Perissoratis, C. and Sakellariadou, F., in press. A statistical analysis of geochemical data in regard to placer mineral exploration in the northern Aegean Sea. Mineralium Deposita, special issue. Munsell, A.H., 1971. Soil, color chart. Munsell Color Comp. Inc., Baltimore Md. Nissenbaum, A., 1972. Distribution of several metals in chemical fractions of a sediment core from the sea of Okhotsk, Israel. J.Earth Sci.,vol. 21, pp. 143-154. Nissenbaum, A., 1974. The organic geochemistry of marine and terrestrial humic substances : Implication of carbon and hydrogen isotope studies. In : Advances in Organic Geochemistry. Nissenbaum, A. and Kaplan, I.R., 1972. Chemical and isotopic evidence for the in situ origin of marine humic substances. Limn, and Ocean., vol. 17 (4), pp. 570-582. Nissenbaum, A. and Swaine, D.J., 1976. Organic matter-metal interactions in recent sediments : the role of humic substances. Geochim. Cosmochim. Acta, vol. 40, pp. 809-816. Pe, G.G. and Panagos, A.G., 1979. Heavy mineralogy of river and beach sands, continental Greece. Neues Jahrb. Mineral. Abh., vol. 136 (3), pp. 254-261. Perissoratis, C., Angelopoulos, I. and Mitropoulos D., 1985. Exploring the offshore area of N.E.Greece for placer deposits : Geologic framework and preliminary results. Proceedings of the NATO ARI "Marine Minerals : Resource Assesment Strategies". Wales. June 1985. Perissoratis, C., Moorby, S.A., Papavasiliou, C., Cronan, D.S., Angelopoulos, I., Sakellariadou, F. and Mitropoulos, D., 1987. The geology and geochemistry of the surficial sediments off Thraki, Northern Greece. Marine Geology, vol. 74, pp. 209-224. Perissoratis, C., Moorby, S.A., Angelopoulos, I., Cronan, D.S., Papavasiliou, C., Konispoliatis, N., Sakellariadou, F. and Mitropoulos, D., in press. Mineral concentrations in the recent sediments off Eastern Macedonia, Northern Greece : Geological and geochemical considerations. Marine Geology. 328 Petrascheck, W.E., 1982. Introduction, In : Mineral deposits of Europe, vol. 2 : Southern Europe. Dunning, F.W., Mykura, W. and Slater, D. (eds.), I.M.M. and Mineralogical Society, pp. 1-12. Rashid, M.A. and King, L.H., 1970. Major-oxygen containing functional groups present in humic and fulvic acid fractions isolated from contrasting marine environments. Geochim. Cosmochim. Acta, vol. 34, pp. 193-201. Rashid, M.A. and King, L.H., 1971. Geochemical characteristics of fractionated humic acids associated with marine sediments. Chem. Geol., vol. 7, pp. 37-43. Read, H. H., 1970. Rutley's elements of mineralogy. Thomas Murby and Co, London, pp. 560. Riley, J.P. and Chester, R., 1971. Introduction to marine chemistry. Acad. Press, pp. 465. Rogers, M.A. and Koons, C.B., 1968. Glacial-eustatic control of organic carbon delta C13 values from post-Pliocene through recent-marine sequences in the Gulf of Mexico. Program 1968. Geol. Soc. Amer. Ann. Mtg., Mexico City. Rose, A.W., Wawkes, H.E. and Webb, J.S., 1979. Geochemistry in Mineral Exploration. Second Edition. Acadenic Press, London/New York/Toronto/Sydney/San Francisco, pp. 657. Sakellariadou, F., 1985. Geochemical investigations in sediment cores from lerissos gulf, N. Aegean Sea. Proceedings of the International Conference "Heavy Metals in the Environment", Athens, Greece, Sept. 1985. Lekkas, T.D. (ed.), vol. 2, pp. 435-437. Sakellariadou, F., 1986. The geochemistry and the partitioning of various metals in selective chemical fractions of sediments collected from the Samothraki plateau, (N. Greece). Proceedings of the International Conference "Chemicals in the Environment", Lisbon, Portugal, July 1986. Lester, J.N., Perry, R. and Sterritt, R.M. (eds), Selper Ltd, London, pp. 578-586. Sawlan, J. and Murray, J.W., 1983. Trace metal remobilization in the interstitial waters of red clay and hemipelagic marine sediments. Earth Planet. Sci. Lett., vol. 64, pp. 213-230. 329 Schnitzer, M., 1978. Humic substances chemistry and reactions. In : Soil Organic matter. Schnitzer, M. and Khan, S.U. (eds.), Elsevier, New York. Schnitzer, M. and Khan, S.U., 1972. Humic substances in the environment. Marcel Dekker, New York. Schnitzer, M. and Skinner, S.I.M., 1969. Free radicals in soil humic compounds. Soil Sci., vol. 108, pp. 383-390. Senesi, N. and Sakellariadou, F., 1987. Characterization of humic acids and their paramagnetic metal ion complexes isolated from marine and beach sediments. Proceedings of the 4 ^ International Symposium on the Interaction between sediments and water. Melbourne, Australia, February 1987. Shearme, S., Cronan, D.S. and Rona, P.A., 1983. Geochemistry of sediments from the Tag hydrothermal field, M.A.R. at latitude 26° N. Marine Geology, vol. 51, pp. 269-291. Steelink, C., 1977. Humates chemistry : Genesis, composition, reactions. Wiley-lnterscience, New York. Steelink, C. and Tollin, G., 1967. Free radicals in soils. In: Soil Biochemistry. McLaren, A.D. and Patterson, G.H. (eds.), Dekker, pp. 147-169. Stevenson, F.J. and Goh, K.M., 1971. Infra-red spectra of humic acids and related substances. Geochim. Cosmochim. Acta, vol. 35, pp. 471-483. Suhr, W.H. and Ingamells, C.O., 1966. Solution technique for analysis of silicates. Anal. Chem., vol. 38, pp. 730-734. Swain, F.M., 1968. Organic geochemistry I. Pergamon Press, New York. Tessier, A., Campbell, P.G.C. and Bisson, M., 1979. Sequential extraction procedure for the speciation of particulate trace metals. Anal. Chem., vol. 51 (7), pp. 844-850. Therianos, A.D., 1974. Water regime and geographical distribution of the river yield at Greece, (in greek). Bull. Geol. Soc. Greece, vol. 10 (1), pp. 28-58. 330 Thompson, M. and Walsh, J.N., 1983. A Handbook of Inductively Coupled Plasma Spectroscopy. Blackie and Son Ltd, pp. 273. Thomson, J., WILson, T.R.S., Culkin, F. and Hydes, DJ., 1984. Non-steady state diagenetic record in equatorial Atlantic sediments. Earth Planet. Sci. Lett., vol. 71, pp. 23-30. Turekian, K.K., 1977. Geochemical distribution of elements. In : Encyclopedia of Science and Technology, Fourth Edition. McGraw-Hill, New York. Turekian, K.K. and Wedepohl, K.H., 1961. Distribution of the elements in some major units of the Earth's crust. Geol. Soc. of Am. Bull., vol. 72, pp. 175-192. Turner, M., 1986. Statistical analysis of geochemical data illustrated by reference to the Dalradian of N.E. Scotland. Unpubl. Ph.D. thesis, Imperial College, Univ. of London. Udden, J.A., 1914. Mechanical composition of clastic sediments. Geol. Soc. Am. Bull., vol. 25, pp. 655-744. van Valin, R. and Morse, J.W., 1982. An investigation of methods commonly used for the selective removal and characterisation of trace metals in sediments. Marine Chemistry, vol. 11 (6), pp. 535-564. Varnavas, S.P., 1979. Geochemical investigations on sediments from the Eastern Pacific. Unpubl. Ph.D. Thesis, Imerial College, Univ. of London, pp. 429. Vavelidis, M. and Amstutz, G.C., 1981. New genetic investigations on the Pb-Zn deposits of Thassos (Greece). In : Mineral deposits of the Alps and of the Alpine Epoch in Europe. Proceedings of the IV ISMIDA 1981, Hans- J.Schneider (ed.), Special Publication No3 of the Society for Geology applied to mineral deposits. Springer-Verlag, Berlin/Heidelberg/New York/Tokyo. Wagemann, R., Brunskill, G.J. and Graham, B.W., 1977. Composition and reactivity of some river sediments from the Mackenzie Valley, N.W.T., Canada. Environ. Geol., vol. 1, pp. 349-358. Wedepohl, K.H., ed., 1969-1978. Handbook of Geochemistry, vol. 1-4. 331 Springer-Verlag, Berlin. Wentworth, C.K., 1922. A scale of grade and class terms for clastic sediments. Jour. Geology, vol. 30, p. 377-392. Wickremeratne, W.S., 1986. Preliminary studies of the offshore occurrences of monazite-bearing heavy-mineral placers, southwestern SriLanka. Mar. Geol., vol. 72, pp. 1-9. Wilson, T.R.S., Thomson, J., Colley, S., Hydes, D.J., Higgs, N.C. and Sorensen, J., 1985. Early organic diagenesis; the significance of progressive subsurface oxidation fronts in pelagic sediments. Geochim. Cosmochim. Acta, vol. 49, pp. 811-822. Yen, T.F. and Tang, J.I.S., 1977. Chemical aspects of marine sediments. In : Chemistry of Marine Sediments. Yen, T.F. (ed.), Ann Arbor Science, pp. 1-38. Deer, W.A., Howie, R.A. and Zussman, J., 1965. Rock-forming minerals. Vol. 1. Ortho- and Ring Silicates. Longmans, Green and Co Ltd, London, pp. 333. Moorby, S.A. and Cronan, D.S., 1981. The distribution of elements between co-existing phases in some marine ferromanganese-oxide deposits. Geochim. Cosmochim. Acta. vol. 45, pp. 1855-1877. Papadakis, 1975. The black sands from Loutra Eleftheron near Kavala, Greece. Sci. Ann. Fac., Phys. and Mathematics, Univ. of Thessaloniki, vol. 15. pp. 331-390. Stevenson, F.J.. 1967. Organic acids in soil. In: Soil Biochemistry. McLaren, A.D. and Patterson, G.H. (eds.), Dekker, pp. 119-146. 332 A P P E N D IX 1 SIMPLIFIED GEOLOGICAL MAP OF THE MAINLAND AND THE THASSOS AND SAMOTHRAKI ISLANDS Simplified geplogical map of the mainland, and the Thassos and Samothraki islands (modified from Bornovas and Rondogianni-Tsiambaou, 1983). Key_i A Post-tectonic and late-tectonic sediments 1 Holocene Alluvial deposits in valleys, plains and coastal deposits. Alluvial fans and older talus cones in the east part of Samothraki island. 2 Pleistocene Lacustrine and continental deposits : clays, loam, sand, conglomerates, red clayey material, etc. Plio-Pleistocene Lacustrine deposits : sand, conglomerates, clays, peat or lignite beds, sometimes red clayey material (north of Porto Lago bay). Marine deposits : sand, marls, clays, conglomerates (vicinity of Loutra Eleftheron). 3 PliQceae Marine deposits : conglomerates, sand, clays, marls, marly limestones and sometimes gypsum beds. Brackish facies deposits are also included. Mio-Pliocene (Upper Miocene-Pliocene) 333 Lacustrine and terrestrial deposits : conglomerates, sand, marls, clayey material, marly limestones, and clays. Sometimes lignite beds. Marine deposits : the above detrital sediments and occasionally gypsum beds. Usually, no red clayey material. Brackish facies deposits are also included. Upper Miocene (and sometimes. Middle Miocene) Lacustrine deposits : marls, clays, sandstones, conglomerates, clays, locally with lignite beds. 4 Oligocene (Eocene-Oligocene-Upper Eocene) Molassic formations : clays, conglomerates, sandstones, marls. In Thraki area, additionally, limestones with lignites. & Axios zone and circum Rhodope zone 5 Triassic-Jurassic Schists, sandstones, marbles or quartzites, phyllites, diabases, limestones. 6 Mesozoic Marbles and phyllites. Q Metamorphic rocks. Rhodope and Serbo-Macedonian Massifs 7 Gneisses, in general, (Paleozoic or older). 8 Paleozoic amphibolites. 9 Amphibolites, gneisses, schists with marble intercalations. 1 0 Marbles or crystalline limestones. D Igneous rocks 334 I Intrusives 11 Tertiary granites, granodiorites, monzonites; Mesozoic and Tertiary in Halkidiki. 1 2 Diabases. 1 3 Paleozoic serpentinites. II Volcanic 1 4 Acid to intermediate. Eocene rhyolites, rhyodacites, dacites, andesites, trachyandesites, trachytes. In Samothraki island, the same formations as above and, additionally, Oligocene pyroclastics : tuffs, ignimbrites. 1 5 Lake 1 6 : active mine $ : idle mine $< : ore occurrence 335 APPENDIX II SAMPLE COLLECTION AND PREPARATION The offshore samples, both surface and buried, were collected by the Greek Institute of Geology and Mineral Exploration (I.G.M.E.); grab samplers (Van Veen or Dietz la Fond) and a 3 m, 250 kg gravity corer were used, respectively. The onshore coastal samples were collected by the author either by hand or by grab samplers. The latter was applied to a few of the river and stream sediments. All the samples were carried to London in a wet state in plastic containers (bags or tubes). They were allowed to dry at room temperature and then big shells present in many samples were removed. In the following stage, the sediments were ground to a fine powder, using a Tema mill with an agate pot. 336 APPENDIX 111 GEOCHEMICAL DATA FOR 17 ELEMENTS DETERMINED IN 43 OFFSHORE SEDIMENTS BY BOTH METHODS A AND B Table 111.1: K, Be, Mg, Ca, Sr, Ba, La,Ti and P contents (jig/g) of 43 offshore sediments, determined by method A. Sample I.D. K Be Mg Ca Sr Ba La Ti P B 12 6,200 0.52 6,700 178,000 650 210 35 1,150 260 B 22 18,300 2.40 21,000 21,000 178 460 32 4,800 530 B 27 22,000 1.82 3,900 18,000 260 600 28 1,790 260 B 83 13,100 0.63 10,300 34,000 86 340 22 920 290 B 89 9,400 3.90 9,000 9,900 60 116 44 1,550 550 BSTR 4 25,000 1.96 4,800 34,000 230 620 34 2,100 440 I 5 7,900 0.60 7,600 240,000 1,440 189 45 690 290 I 22 1,800 0.20 102,000 87,000 270 36 23 183 188 I 38 22,000 3.20 20,000 39,000 270 310 35 3,600 560 I 46 19,000 1.03 16,800 55,000 108 440 33 1,280 480 I 50 10,100 0.50 17,200 64,000 109 430 26 1,840 490 STR 4 12,700 1.61 13,800 141,000 840 250 53 2,300 580 STR 50 19,700 2.40 20,000 34,000 199 290 29 5,300 710 STR 74 26,000 3.20 173 380 31 4,200 730 KB 2 24,000 3.50 9,300 19,800 460 580 43 3,000 440 KB 31 25,000 3.60 17,600 32,000 260 390 34 4,000 490 KB 55 9,700 1.26 9,600 250,000 1,880 195 55 960 320 KB 100 18,000 2.10 14,800 73,000 620 270 41 2,800 440 KB 113 4,600 0.85 23,000 250,000 2,100 104 40 60 350 KB 126 23,000 2.50 3,700 27,000 8901,110 50 1,300 390 337 Table 111.1 (continued) THR 1 19,600 2.20 26,000 19,200 165 260 28 4,100 620 THR 4 7,400 0.90 8,900 220,000 690 55 38 520 300 THR 11 6,100 0.77 14,800 230,000 1,830 109 53 1,080 710 THR 31 10,600 1.07 12,800 111,000 820 122 31 1,790 460 THR 60 26,000 3.20 15,000 15,600 185 470 46 4,500 600 THR 84 6,200 0.91 31,000 82,000 360 210 21 9,900 620 THR 122 15,600 1.40 8,800 130,000 1,000 260 37 1,290 370 I 14/3 3,000 2.70 23,000 38,000 250 320 40 3,800 580 I 14/16 23,000 2.90 23,000 40,000 280 300 42 3,900 570 I 14/41 23,000 3.00 24,000 43,000 310 310 42 4,000 560 I 14/71 23,000 3.00 23,000 41,000 300 310 42 4,000 .590 I 14/101 22,000 2.80 24,000 45,000 310 300 42 4,100 510 I 14/131 22,000 2.70 25,000 54,000 360 290 38 4,100 510 I 14/161 22,000 2.70 27,000 61,000 370 290 43 4,100 520 ! 14/185 23,000 2.90 23,000 43,000 300 310 44 4,000 570 THR 22/2 23,000 3.30 15,500 25,000 200 550 37 4,200 720 THR 22/24 23,000 3.30 16,200 12,600 1621,350 43 4,300 680 THR 22/39 24,000 3.50 16,100 16,800 1671,470 32 4,300 670 THR 22/64 23,000 3.20 16,300 12,800 162 540 32 4,200 630 THR 22/94 26,000 3.60 16,600 10,500 157 440 39 4,500 710 THR22/121 25,000 3.60 16,400 6,800 125 390 32 4,000 680 THR22/149 24,000 3.40 16,900 9,900 143 410 35 4,300 670 THR22/176 26,000 3.60 17,200 9,200 148 440 32 4,400 700 L.D.L. 100 0.40 100 50 4 4 20 60 40 CM X CO CM O U.D. L. 131X103 41X102 51X103 25X104 4X1034X1034X10315X103 338 Table 111.2 : V, Cr, Mn, Fe, Co, Ni, Cu, Zn and Al contents (pg/g) of 43 offshore sediments determined by method A. Sample I.D. V Cr Mn Fe Co Ni Cu Zn Al B 12 93.0 39.0 540 24,200 15.0 27.0 24.0 15.0 29,000 B 22 143.0 220.0 1,180 47,000 36.0 140.0 44.0 69.0 82,000 B 27 43.0 55.0 87 10,500 39.0 22.0 10.0 22.0 53,000 B 83 47.0 48.0 9,400 103,000 15.0 15.0 28.0 15.0 27,000 B 89 86.0 193.0 6,300 200,000 44.0 160.0 730.0 15.0 57,000 BSTR 4 51.0 50.0 660 16,600 20.0 21.0 20.0 350.0 56,000 1 5 16.0 19.0 91 6,100 2.0 10.0 41.0 37.0 22,000 1 22 81.0 3200.0 480 42,000 80.01,620.0 15.0 82.0 7,000 1 38 136.0 141.0 1,500 51,000 20.0 84.0 56.0 280.0 85,000 1 46 62.0 53.0 9,800 83,000 15.0 15.0 68.0 3,300 44,000 1 50 45.0 49.0 4,800 123,000 15.0 15.0 14.0 2,700 25,000 STR 4 105.0 770 31,000 31.0 48.0 20.0 59.0 47,000 STR 50 139.0 147.0 530 49,000 37.0 74.0 41.0 82.0 79,000 STR 74 148.0 111.0 840 52,000 27.0 81.0 24.0 167.0 94,000 KB 2 62.0 39.0 280 29,000 18.0 18.0 24.0 68.0 93,000 KB 31 127.0 117.0 460 43,000 20.0 61.0 30.0 15.0 92,000 KB 55 5.0 39.0 210 14,600 15.0 15.0 27.0 31.0 32,000 KB 100 95.0 88.0 280 32,000 11.0 54.0 23.0 104.0 66,000 KB 113 14.0 12.0 340 4,500 15.0 15.0 6.0 35.0 9,800 KB 126 51.0 20.0 145 12,400 15.0 15.0 20.0 15.0 7,900 THR 1 145.0 260.0 540 51,000 15.0 191.0 56.0 75.0 82,000 THR 4 37.0 33.0 690 13,900 15.0 24.0 16.8 15.0 18,400 THR 11 51.0 83.0 135 18,300 41.0 370.0 30.0 400.0 20,000 THR 31 87.0 88.0 490 22,000 15.0 49.0 14.0 15.0 41,000 339 Table 111.2 (continued) THR 60 124.0 106.0 630 42,000 36.0 56.0 38.0 183.0 90,000 THR 84 290.0 240.0 1,210 96,000 28.0 85.0 44.0 44.0 78,000 THR 122 51.0 49.0 300 14,900 15.0 23.0 24.0 22.0 43,000 I 14/3 153.0 184.0 1,900 59,000 30.0 130.0 112.0 790.0 86,000 I 14/16 145.0 188.0 990 53,000 31.0 125.0 63.0 660.0 86,000 I 14/41 152.0 200.0 1,050 52,000 31.0 130.0 58.0 410.0 87,000 I 14/71 139.0 199.0 890 51,000 35.0 138.0 52.0 890.0 87,000 I 14/101 133.0 190.0 880 48,000 24.0 120.0 47.0 160.0 84,000 I 14/131 126.0 200.0 830 51,000 34.0 126.0 43.0 157.0 81,000 I 14/161 117.0 230.0 780 47,000 44.0 131.0 45.0 159.0 78,000 I 14/185 139.0 197.0 880 52,000 43.0 139.0 65.0 540.0 86,000 THR 22/2 122.0 123.0 530 50,000 23.0 77.0 100.0 210.0 87,000 THR 22/24 134.0 138.0 470 51,000 28.0 111.0 60.0 157.0 89,000 THR22/39 133.0 121.0 800 52,000 31.0 89.0 77.0 230.0 92,000 THR22/64 123.0 123.0 750 49,000 27.0 104.0 61.0 152.0 87,000 THR22/94 144.0 112.0 500 51,000 30.0 86.0 56.0 169.0 96,000 THR22/121 135.0 107.0 370 49,000 23.0 84.0 43.0 103.0 95,000 THR 22/149132.0 123.0 450 50,000 34.0 90.0 41.0 92.0 94,000 THR22/176 131.0 121.0 460 51,000 25.0 87.0 44.0 94.0 96,000 L.D.L. 12.0 20.0 20 100 20.0 20.0 12.0 20.0 50 U.D.L. 4,000 4,000 22,000 264,000 4,000 4,000 4,000 4,000 208,000 3 4 0 Table 111.3: K, Be, Mg, Ca, Sr, Ba, La,Ti and P contents (pg/g) of 43 offshore sediments, determined by method B. Sampie I.D. K Be Mg Ca Sr Ba La Ti P B 12 5,800 .20 6,600 181,000 550 170 12.0 1,220 250 B 22 18,100 2.00 20,000 23,000 156 430 23.0 4,700 530 B 27 21,000 1.20 3,900 19,700 220 550 21.0 1,760 200 B 83 12,600 .20 10,100 35,000 69 28 11.7 950 280 B 89 21,000 1.30 4,500 36,000 210 520 24.0 1,700 310 BSTR 4 25,000 1.35 4,600 36,000 196 570 21.0 2,200 470 1 5 7,500 0.40 7,700 260,000 1,270 182 9.3 810 300 1 22 1,670 0.05 97,000 90,000 230 24 4.5 191 141 1 38 22,000 2.40 19,700 40,000 210 280 28.0 3,800 620 1 46 19,400 0.50 17,000 58,000 94 89 16.8 1,320 490 1 50 9,700 0.10 16,400 63,000 85 34 16.0 1,610 560 STR 4 12,800 1.28 13,900 147,000 780 250 31.0 2,400 450 STR 50 19,600 2.10 20,000 35,000 174 290 27.0 5,700 830 STR 74 23,000 2.90 16,600 29,000 126 360 32.0 4,300 790 KB 2 23,000 3.00 9,500 22,000 430 590 32.0 3,200 510 KB 31 23,000 2.90 16,700 31,000 210 340 34.0 4,100 500 KB 55 9,400 0.78 8,700 240,000 1,580 185 24.0 1,130 320 KB 100 18,100 2.10 15,600 81,000 580 280 35.0 3,000 420 KB 113 4,500 0.05 22,000 290,000 1,930 115 11.6 250 330 KB 126 23,000 2.40 3,700 30,000 850 1,230 44.0 1,590 360 THR 1 18,900 2.00 26,000 21,000 146 300 18.9 4,300 640 THR 4 7,200 0.20 8,900 23,000 690 137 13.9 650 320 THR 11 5,900 0.30 13,800 230,000 1,620 116 21.0 1,050 720 THR 31 11,700 0.79 13,000 119,000 760 199 23.0 1,980 480 341 Table 111.3 (continued) THR 60 25,000 2.90 15,500 17,300 178 490 24.0 4,600 680 THR 84 5,800 0.40 30,000 85,000 330 270 15.8 9,900 630 THR 122 15,700 0.94 9,100 137,000 940 350 29.0 1,420 390 I 14/3 21,000 2.10 21,000 38,000 190 280 28.0 3,700 590 I 14/16 21,000 2.30 22,000 41,000 230 280 28.0 4,100 570 I 14/41 21,000 2.30 23,000 44,000 260 280 30.0 3,900 570 I 14/71 21,000 2.30 22,000 40,000 220 270 28.0 4,200 510 I 14/101 21,000 2.20 23,000 47,000 260 270 27.0 4,300 520 I 14/131 20,000 2.10 24,000 55,000 270 260 29.0 4,100 490 I 14/161 19,800 1.89 25,000 61,000 300 270 30.0 4,000 470 I 14/181 21,000 2.30 23,000 46,000 250 290 30.0 4,100 580 THR 22/2 22,000 2.80 15,500 26,000 180 520 33.0 4,300 770 THR 22/24 22,000 2.80 16,200 13,200 136 1 ,250 27.0 4,300 660 THR 22/39 23,000 3.10 15,900 15,600 123 880 31.0 4,700 710 THR 22/64 23,000 2.90 16,600 12,300 127 500 31.0 4,800 710 THR 22/94 25,000 3.20 16,300 9,100 110 390 30.0 5,000 740 THR 22/121 25,000 3.50 17,500 9,000 127 390 48.0 4,000 750 THR 22/149 25,000 3.20 17,900 10,900 131 420 30.0 4,600 750 THR 22/176 24,000 3.10 18,700 20,000 194 490 102.0 3,900 740 L.D.L 10 0.05 10 6 0.3 0.5 1.0 6 4 U.D.L. 2x103 103 5x103 5x103 104 104 104 5x1044x104 3 42 Table 111.4: V, Cr, Mn, Fe, Co, Ni, Cu, Zn and Al contents (jxg/g) of 43 offshore sediments determined through the method A. Sample I.D. V Cr Mn Fe Co Ni Cu Zn Al B 12 70.0 25.0 550 23,000 10.6 19.2 16.8 26.0 29,000 B 22 128.0 1 60.0 1,210 45,000 31.0 128.0 39.0 89.0 79,000 B 27 28.0 29.0 121 10,400 6.3 15.8 7.0 19.7 53,000 B 83 34.0 25.0 8,700 194,000 23.0 27.0 820.0 6,800.0 26,000 B 89 26.0 13.1 490 10,100 7.2 11.3 4.4 21.0 50,000 BSTR 4 41.0 31.0 690 16,300 9.1 14.8 8.6 37.0 52,000 1 5 21.0 17.1 176 6,400 3.5 8.2 23.0 36.0 23,000 1 22 47.02300.0 530 40,000 62.0 1460.0 7.2 65.0 7,200 1 38 117.0105.0 1,410 47,000 24.0 75.0 61.0 330.0 71,000 1 46 49.0 30.0 9,400 94,000 12.1 18.4 400.0 3,100.0 46,000 1 50 37.0 27.0 4,500 172,000 13.4 35.0 300.0 2,300.0 24,000 STR 4 68.0 57.0 880 30,000 16.0 42.0 14.0 50.0 49,000 STR 50 133.0 123.0 640 47,000 28.0 83.0 42.0 100.0 70,000 STR 74 126.0 86.0 920 49,000 24.0 56.0 45.0 133.0 92,000 KB 2 66.0 30.0 380 28,000 13.4 22.0 18.4 79.0 92,000 KB 31 97.0 86.0 460 39,000 20.0 58.0 28.0 100.0 84,000 KB 55 35.0 19.6 290 13,000 7.0 16.5 10.1 31.0 31,000 KB 100 88.0 81.0 370 32,000 16.8 51.0 31.0 98.0 71,000 KB 113 11.2 6.7 430 4,800 4.8 4.6 3.9 18.5 10,200 KB 126 36.0 4.7 250 12,400 7.6 5.4 2.9 17.5 75,000 THR 1 136.0 210.0 660 50,000 34.0 192.0 50.0 105.0 77,000 THR 4 32.0 24.0 790 13,600 7.6 18.8 5.4 21.0 18,800 THR 11 48.0 42.0 1,320 16,500 34.0 340.0 21.0 340.0 20,000 THR 31 72.0 62.0 610 22,000 13.7 56.0 23.0 56.0 42,000 THR 60 110.0 90.0 750 42,000 25.0 64.0 50.0 177.0 90,000 3 43 Table 111.4 (continued) THR 84 270.0 193.0 1,300 91,000 43.0 84.0 33.0 70.0 81,000 THR 122 49.0 34.0 400 14,900 9.1 29.0 13.3 47.0 45,000 1 14/3 134.0 137.0 1,850 53,000 24.0 130.0 108.0 890.0 73,000 1 14/16 132.0 145.0 1,030 50,000 26.0 109.0 65.0 580.0 74,000 I 14/41 121.0 148.0 1,080 48,000 26.0 108.0 58.0 320.0 85,000 I 14/71 120.0 141.0 890 45,000 26.0 114.0 55.0 270.0 67,000 I 14/101 112.0 130.0 940 45,000 27.0 101.0 49.0 144.0 78,000 I 14/131 110.0 152.0 840 47,000 26.0 110.0 44.0 131.0 72,000 I 14/161 103.0 170.0 820 43,000 25.0 111.0 38.0 127.0 76,000 I 14/185 127.0 161.0 940 50,000 27.0 107.0 60.0 470.0 81,000 THR 22/2 113.0 101.0 580 47,000 24.0 70.0 97.0 199.0 85,000 THR 22/24 120.0 106.0 540 49,000 23.0 85.0 77.0 160.0 90,000 THR 22/39 126.0 108.0 810 48,000 25.0 88.0 123.0 210.0 70,000 THR 22/64 122.0 104.0 780 46,000 26.0 89.0 69.0 152.0 65,000 THR 22/94 131.0 84.0 510 48,000 25.0 69.0 51.0 166.0 72,000 THR 22/121 129.0 95.0 510 52,000 24.0 78.0 42,0 107.0 127000 THR 22/149 132.0 103.0 550 51,000 25.0 76.0 42.0 105.0 100000 THR 22/176 123.0 102.0 700 55,000 25.0 76.0 40.0 105.0 182,000 L.D.L. 0.5 1.0 1.5 4 1.0 1.0 0.5 1.0 5.0 U.D.L. 104 104 5x1 05 50x104 104 104 104 104 5x105 L.D.L.= lower detection limit, U.D.L.= upper detection limit Numbers initalics denote best estimates of below detection limit concentrations. 3 4 4 APPENDIX IV CALCIUM CARBONATE AND ORGANIC CARBON CONTENTS Table IV.1: Calcium carbonate and organic carbon contents of 206 surface offshore samples Sample I.D. CaCOo org. C Sample I.D. CaCOu org. C (%) (%) (%) (%) I 1 1.0 <0.10 I 36 12.4 0.10 I 2 55.2 0.91 I 37 6.1 <0.10 I 3 35.8 0.46 I 38 8.2 0.73 I 4 58.7 1.14 I 39 14.8 0.18 I 5 61.6 0.50 I 40 1.0 0.10 I 6 17.5 0.50 I 41 10.8 0.71 I 7 52.6 0.57 I 42 9.2 0.57 I 8 25.2 0.21 I 43 3.9 0.18 I 9 1.0 0.10 I 44 2.3 0.21 I 10 7.6 0.71 I 45 7.2 0.50 I 11 9.6 0.91 46 14.8 0.14 l 12 10.4 0.84 I 48 20.2 0.14 I 13 8.8 0.82 I 49 13.2 <0.10 I 14 4.3 0.46 I 50 11.7 0.10 I 15 0.2 0.10 I 51 2.3 0.50 I 16 0.2 <0.10 I 52 3.9 0.77 I 17 3.1 <0.10 I 53 5.9 0.68 I 18 1.0 <0.10 I 54 8.4 0.64 I 19 0.4 <0.10 I 55 10.4 0.64 I 20 6.8 0.46 I 56 6.3 0.89 345 Table IV.1 (continued) I 21 8.8 0.48 I 57 1.4 0.18 I 22 22.5 0.64 I 58 1.0 <0.10 I 23 55.6 0.36 I 59 5.9 0.91 I 24 66.0 0.18 I 60 8.0 <0.10 I 25 10.4 0.93 I 61 23.7 0.10 I 26 8.0 0.64 I 62 1.0 0.87 I 27 1.0 0.10 I 63 15.2 0.36 I 28 4.6 <0.10 I 65 8.4 0.87 I 29 7.2 0.10 I 66 10.4 0.71 I 30 13.3 0.18 I 67 9.2 0.59 I 31 10.8 0.50 I 68 1.8 0.73 I 32 10.8 <0.10 I 72 9.4 0.84 I 33 28.3 0.23 I 73 2.3 0.41 I 35 40.9 0.41 I 74 8.4 0.77 SIR 8 11.5 0.15 STR145 21.1 0.22 STR100 13.1 0.73 STR147 30.6 0.22 STR117 11.5 0.80 STR149 30.4 0.54 STR118 13.4 0.74 STR150 15.6 0.54 STR119 13.1 0.69 STR151 20.0 0.69 STR120 12.3 0.58 STR152 18.1 0.78 STR121 10.7 0.88 STR154 21.9 0.30 STR123 12.0 0.73 STR155 20.5 0.28 STR124 12.6 0.65 STR157 24.9 0.60 STR128 12.3 0.91 STR158 24.6 1.49 STR129 9.9 0.99 STR159 22.4 0.34 STR131 10.4 0.99 STR160 28.7 0.30 346 Table IV.1 (continued) STR133 18.9 0.72 STR161 23.3 0.34 STR134 15.3 0.24 STR162 22.2 0.52 STR140 17.2 0.73 STR163 12.3 0.45 STR141 14.5 0.71 STR168 15.3 0.17 STR142 20.8 0.60 STR170 21.9 0.54 STR143 23.8 0.62 STR174 7.7 0.39 STR144 23.0 0.41 STR178 10.4 0.89 KB 1 2.5 0.78 KB 59 2.2 0.71 KB 2 1.2 1.08 KB 61 3.4 0.75 KB 3 12.0 0.91 KB 96 27.4 0.50 KB 21 26.5 0.78 KB 97 38.2 0.50 KB 35 40.6 0.24 KB 100 19.5 4.05 KB 40 54.3 1.51 KB 101 25.1 1.12 KB 43 30.0 3.73 KB 103 28.4 2.07 KB 45 51.4 3.06 KB 104 47.5 2.16 KB 49 22.5 4.06 KB 105 22.9 0.30 KB 50 22.9 3.60 KB 107 18.1 0.30 KB 52 23.3 2.24 KB 120 18.5 0.60 KB 53 37.6 1.47 KB 123 19.1 0.39 KB 54 30.2 1.77 THR 1 2.0 1.44 THR 60 1.9 0.84 THR 2 1.1 0.68 THR 61 1.5 1.60 THR 3 36.3 2.59 THR 64 34.4 2.09 THR 5 22.1 3.65 THR 68 7.5 1.07 THR 6 6.0 1.07 THR 69 8.4 1.07 THR 8 7.8 0.87 THR 71 17.0 1.15 347 Table IV.1 (continued) THR 9 63.2 3.90 THR 72 22.0 3.98 THR 10 2.4 0.94 THR 76 56.2 2.96 THR 11 60.0 • 6.76 THR 78 37.4 1.70 THR 12 6.9 0.61 THR 79 19.0 0.63 THR 12 2.0 0.96 THR 80 16.6 1.07 THR 15 1.3 0.50 THR 81 15.5 1.26 THR 16 1.7 1.06 THR 84 1.6 0.08 THR 20 8.0 0.24 THR 85 18.6 1.22 THR 21 5.1 0.85 THR 87 19.5 1.13 THR 22 60.0 3.57 THR 90 15.5 1.07 THR 25 15.3 2.99 THR 91 12.2 2.73 THR 27 8.6 0.48 THR 93 26.4 0.31 THR 28 47.2 2.31 THR 95 26.8 0.99 THR 29 11.7 2.88 THR 96 17.3 0.90 THR 31 28.1 6.41 THR 99 11.3 0.55 THR 32 35.2 2.33 THR101 22.1 1.89 THR 33 29.2 2.09 THR102 19.7 1.09 THR 34 2.7 0.89 THR105 13.9 1.03 THR 35 21.7 1.05 THR110 13.5 1.07 THR 36 22.8 1.20 THR112 50.5 5.16 THR 37 17.5 1.29 THR118 16.6 1.11 THR 38 19.7 1.11 THR119 55.2 3.37 THR 41 4.10 THR120 26.1 2.68 THR 44 4.0 1.46 THR122 21.1 4.07 THR 45 10.2 0.35 THR125 9.7 0.96 THR 46 24.1 1.09 THR126 1.6 0.94 THR 50 15.0 0.76 THR129 1.6 1.15 348 Table IV.1 (continued) THR 51 23.0 1.53 THR130 1.3 1.28 THR 53 40.6 2.70 THR132 2.4 1.01 THR 55 2.0 0.40 THR133 6.7 1.66 THR 57 1.6 1.13 THR135 16.2 1.34 THR 58 3.1 1.05 349 APPENDIX V MULTIVARIATE STATISTICAL ANALYSIS Definitions and brief explanations of some of the terms used in the multivariate statistical sections will follow. Cluster analysis is an assortment of techniques designed to perform classification by assigning observations to groups in a way that each group is more-or-less homogeneous and distinct from other groups (Davis, 1973). P e n d o a ra m is a diagram of a hierarchical kind that presents graphically the elemental associations provided by cluster analysis (Davis, 1973). In a dendogram, the endpoints of the branches represent the elements; the height at which the branches join corresponds to the element similarity level for admission to a pre-existing group (Howarth and Sinding-Larsen, 1983, p. 221). For the cluster analysis an interactive graphics program (TREES) obtained from Bell Laboratories and extensively modified by Turner (A.G.R.G.) was used. Factor analysis is a number of related computational procedures that share the common objective of attempting to represent the variance of a large number of elements in the original data by a small number of factors. Each of them is a linear function (transformation) of the element concentrations. In this way, a greater efficiency in terms of information compression over the original data, and hopefully also some help concerning the interpretability, are gained (Howarth and Sinding-Larsen, 1983, p. 233). By using factor analysis the element associations inherent in the structure of the correlation matrix are separated into a number of groups of elements that together account for the greater part of the observed variability of the original data. Eiaen value is the total variability accounted for by each factor over all the elements (Howarth and Sinding-Larsen, 1983, p. 236). The loadings of the original variables on the factors constructed are a measure of the correlation between the variables and the factors. The sign of any loading on any one individual component is not important, as far as the magnitude of the loading is 350 concerned. A high positive loading is as significant as a high, in absolute value, negative one; when however the two occur within the same factor they reflect the antipathetic relationship between the corresponding variables. For the generation of the rotated loading matrix an interactive program (BPCA2) was used. CattelFs Scree test is a criterion which is used to help choosing the correct number of factors to be retained (Howarth and Sinding-Larsen, 1983, p. 247). This test is a plot of eigen values versus factor number. It shows a "significant" break in slope from a linear decrease associated with the last few eigen values and attributed to random error terms. Thus, the number of factors that would be a satisfactory solution is defined by the one corresponding to the aforementioned break. In a set of n observations it is possible for a limited number to be separated in value from the remainder; such values are called o u tlie rs. The multivariate outliers may represent anomalous values in terms of unusually high metal contents (Howarth and Sinding-Larsen, 1983, p. 229). The Box-Cox transformation is a power transformation which optimizes the normality of element frequency distributions (Coward, 1986, p. 25). The transformation is of the form: x'= (x^ -1)/ X x'=transformed element value x =untransformed element value X =power The transformation of the data has the effect of redistributing the data values so as to increase the distance between samples in the main body of data (background values) and force the long tail of high raw data values back towards Y axis. That is, data transformation enhances the background relationships between variables and supresses high values (Chapman, 1976). The Box-Cox transformation is preferred to the more usual logarithmic one; in some situations where a transformation is necessary, taking logarithms was found to result in "under normalization" or alternatively produced negatively skewed distributions (Coward, 1986, p. 24). For the Box-Cox (power) transformation an interactive program (BLAM2) was used. The correlation matrix obtained is likely to characterize background population structure (Turner, 1986). 351 Robust statistics trim the data by recognizing and downweighting outlying values (Turner, 1986). Outlying samples adversely affect the calculation of correlation coefficients between elements. This is because background variable means are pulled away from their true values by atypical observations. To overcome the outlier problem (Coward, 1986, p. 27) robust calculation of coefficients has been used. To perform the robust calculation of coefficients a proprietary program (BWTHGRP) written by Campbell (1980) was used. All the statistical programs used in this study were run on the CDC 6,500/174 system at Imperial College. 352 APPENDIX VI ELEMENT PARTITIONING IN SELECTED SURFACE AND SUBSURFACE SEDIMENTS 353 f I B8 I 46 4 5 3 355 3 8 8 S s 8 8 O CNJ STR 74 STR 4 356 20 40 60 80 100 357 8 8 9 S 3 3 o 8 CNJ O in a : »— oo t 8 S o -J- 8 o in ! 358 r KB 55 KB 100 359 ' >~T,' I KB 113 KB 126 360 THR 1 THR re / •/ \ / ■ O 3 _A ; / ; _A 3 0 * / \ \ / 0* ; s ; a w y _ .—. .-r -H— ■■— £7 cx o cr x \ ;/ \ o =*, 7 1 re a ■ “ / 3“ ■ 8 . / V ,/... / V / a a QC ° * ? g- O ! / / 1 3 / \ s 8 3 8 > A ^ . 1 . r a i o 1/1 X o 361 1 8 8 8 3 8 8 8 3 THR 11 THR m 362 8 THR 60 THR 84 363 f CM CM CC THR 122 z n s 3 5 O O'* o 364 t 365 8 8 8 S R CM CM CO 366 CO cn rS rS*— § t 08 80 s 8 TT" j , o cff ,o / s j ^ c/ 1 9 rsj / 3 I 3 / 3 / S J « / - J C — . j / tu u x ^ a UJ U x ^ rv UJ <-t x y cr o cc O cc ♦ ° a o o o f O Q C 03 BSTR O 366 A 366 l 3 6 7 Fig. VI.1 Proportions of Li, K, Mg, Ca, Sr, Ba, Al, Ca, Cr, Mn, Fe, Co, Ni, Cu, V, Be, Zn, Pb and P associated with each of the five fractions of the surface samples subjected to partition analysis : I 5, I 22, I 38, I 46, I 50, STR 4, STR 50, STR 74, KB 2, KB 31, KB 55, KB 100, KB 113, KB 126, THR 1, THR 4, THR 11, THR 31, THR 60, THR 84, THR 122, B12, B22, B 27, B 83, B 89, BSTR 4. I 14/3 I 14/16 8 6 3 1 14/41 I 14/71 369 I 14/101 I 14/131 370 I 14/161 I 14/161 Fig. VI.2 Core I 14 : Proportions of Li, K, Mg, Ca, Sr, Ba, Al, Ca, Cr, Mn, Fe.V, Be, Co, Ni, Cu, Zn, Pb and P associated with each of the five fractions of samples collected from 3, 16, 41, 71, 101, 131 and 161 cm depth. 1 7 3 T H R 22/2 THR 22/24 372 CVJ Os CC -j- THR 22/64 lu *_ j t/i X OC * Q o QC 373 THR 22/122 THR 22/14-9 4 7 3 ’Ti THR 22176 THR 22/176 Fig. VI.3 Core THR 22: Proportions of Li, K, Mg, Ca, Sr, Ba, Al, Ca, Cr, Mn, Fe, V, B e, Co, Ni, Cu, Zn, Pb and P associated with each of the five fractions of samples collected from 2, 24, 64, 94, 122, 149 and 176 cm depth. 375 376 APPENDIX VII GEOCHEMICAL DATA The analyses (bulk and partition) of offshore surface and buried samples, and of onshore samples are stored on microfiche inside the envelope on back cover.