During the period from Nov. 1984 to Apri. 1985, a marine comprehensive investigation, including biology, physics, chemistry, geology and geophysics, was carried out in the northwestern sea area of Antarctic Peninsula. The survey area is located in 55° to 69°30′W and 60° to 66°55′S. Sampling stations for surface sediments are shown in Fig. 1.

Figure 1 Map showing locations of samples and zones of heavy mineral assemblages.

Figure options

The heavy mineral samples from the surface sediments were studied by the authors under a microscope and using NAA, and much quantitative information was obtained about mineralogy, REE, and trace elements. In the twenty three of heavy mineral samples, approximately thirty one kinds of heavy minerals have been identified, of which the contents of pyroxene, amphibole, epidote, magnetite, ilmenite, pyrite, chlorite, muscovite, biotite, garnet, apatite, zircon, and altered ferromagnesian silicate minerals are relatively high, and the contents of weathered mica, limonite, and hematite relatively low. The observed data for REE, trace (Zr, Hf, Th, Ta, U, Rb, Sr, Zn, Co, Ni, Cr, As, Sc) and major (Fe) elements in the samples are shown in Table 1.

Table 1 Initial variables of heavy minerals, REE and trace elements for statistic analysis.

Table options

The research shows that the REE and trace elements in the surface sediments of the study area, probably, exist mainly in the heavy minerals. At the same station, usually the content of the trace elements listed in Table 1 should exist in the heavy minerals also. It can be seen from this that the heavy minerals are the REE and trace element carrier in the surface sediments of the sea area, and that it is very necessary for us to discuss the relationship between the heavy minerals and REE and trace elements in it. (Xu Butai et al., 1987; Wang Dezi, 1975).

In addition, to obtain more reliable geological information from data analysis and to change the qualitative method used before, the authors analyzed the variables in Table 1 with mathematical statistics (i.e. Q-mode factor and R-mode group analyses).

There are 23 heavy mineral samples and decades of observed data in the study area. Mathematically when number of variables is more than that of samples, the solution of the equation is absent. To find the solution, 31 heavy minerals are divided into 12 kinds and 8 REE are merged into 3 groups, and then a method of group analysis is used. First, Q-mode factor analysis is used for 12 kinds of heavy minerals, and then, group analysis is used for all of elements, and at last, group analysis is used for the heavy minerals. The result is in keeping with expectations.

It is the similarity of their physicochemical properties and their association in the ocean-environment that constitutes the principles of merging into the REE and the heavy minerals. For example, in Table 1, five light REE are merged into two groups, i.e. La+Ce+Na and Sm+Eu; three kinds of heavy REE are merged into one group, i.e. Tb+Yb+Lu. In the heavy minerals, epidotes variable takes epidote as the dominant factor, with less clinozoisite, and zoisite. Schistose minerals refer to chlorite, muscovite, biotite, and weathered mica. Having unstable hydrodynamic nature, they are very easy to be suspended in running water, and being mass-deposited only in a low-energy environment. Accessory minerals refer to garnet, apatite, titanite, zircon, and tourmaline. Having stronger anti weathering and stable hydro-dynamic properties, they are only concentrated in a high-energy environment. Rock fragments or combination minerals refer to altered ferromagnesian silicate minerals. The last heavy mineral variable is other minerals, including andalusite, basaltic hornblende, glaucophane, orthite phlogopite, actinolite, carbonates, rutile, xenotime, and olivine. Due to their very low content (amounting to about 3 percent) and weak influence on the content of the heavy minerals, they are merged into one variable.

Table 4  

Table options

A 3-factor model of the heavy mineral assemblages in the study area has been established by way of Q-mode factor analysis. The variance contributions of three common factors are 66.81%, 19.14% and 5.17%, respectively, and cumulative contribution is 91.12%. This shows that the 3-factor model has over 90 percent of the variation information of the heavy minerals. Three common factors represent three heavy mineral assemblages respectively (Fig. 1, Table 2).

Table 2 Contributions of 3-factor variance and types of heavy mineral assemblages.

Table options

Common factor P1 has 13 samples, representing heavy mineral assemblage-type Ⅰ, and pyroxene, magnetite, and ilmenite occupy a dominant position. Based on location of samples. the variation of accessory mineral content as well as if having pyrite, the type Ⅰ can be further divided into Ⅰ₁ and Ⅰ₂ subtypes. The subtype Ⅰ₁ is distributed in the cast of the study area (Fig. 1), and the heavy mineral assemblage is characteristic of being short of pyrite and of higher accessory mineral content (Table 2). In addition to basic volcanic rocks, on the islands of the eastern study area there are also some metamorphic and intermediate acidity rocks, not only providing the sediments of this region with a lot of pyroxene, magnetite and ilmenite, but also with some accessory minerals (Berkenmajer, 1982). It is clear that the subtype Ⅰ₁ is under the control of sources.

The subtype Ⅰ₂ is distributed over the bay around the South Shetland Is. and the content of accessory minerals is very low, but the pyrite content much high in the heavy mineral assemblage-type (Fig. 1, Table 2). This assemblage feature fully corresponds to that of the minerals of the propylite and basic volcanic rock which outcrop extensively on the South Shetland Is. The subtype Ⅰ₂ is, undoubtedly, affected by the source too, at the same time the cold and close bay is also advantageous to the preservation of pyrite.

The common factor P₂ includes 8 samples, and corresponding heavy mineral assemblage-type Ⅱ, in which rock fragment, green amphibole, and epidote are taken as the dominant component with more schistose minerals, is distributed over the western study area. Such mineral composition should be derived from the weathering products of metamorphic, sedimentary and intermediate acidity rocks on the Antarctic Peninsula (Fig. 1, Table 2).

The common factor P₃ includes only 2 samples, and in the corresponding heavy mineral assemblage-type Ⅲ, main minerals are pyroxene and epidote, and magnetite, ilmenite and green amphibole take the second place. Such an assemblage belongs to a transition type, and its space distribution is also between the assemblage Ⅱ and the assemblage Ⅲ, probably results from mixing of two above-mentioned end member assemblages by currents. It is possible that Bellingshausen current transports the sediments of mineral assemblage zone Ⅰ to the assemblage zone Ⅱ, when it gets from northwest into Bransfield Strait (Shen Yichu et al., 1988) (Fig. 1, Table 2).

12 kinds of heavy mineral variable are divided into two groups (Fig. 2). The first group is made up of green amphibole, schistose minerals, epidote, titanite, and accessory minerals. They originate from metamorphic, sedimentary and intermediate acidity rocks. The second group includes pyroxene, iron oxide, brown amphibole, and pyrite. They have genetic relationship with basic volcanic rock, and pyrite is often the marker of late alteration and mineralization of this rock.

Figure 2 Genetic association of minerals by R-mode group analysis.

Figure options

The result of R-mode group analysis of minerals conforms with that of Q-mode factor analysis, and both show that the detrital sediments in the study area probably come from the source area of two different rocks.

22 elements are divided into 3 groups and 9 subgroups (Fig. 3, Table 3). This fact shows that elements have genetic relationship with rocks. As shown in Fig. 3 and Table 3, the first group consists of 17 elements, and most of them are lithophile ones, having exclusive relationship with intermediate acidity rocks. The second and third groups are composed of two and three elements, respectively. They are basically rock-forming and siderophile elements, having exclusive relationship with basic rocks. In Fig. 3, if correlation factor is taken as 0.6, then three groups can be divided into nine subgroups: (1) LRE and Zr, Hf, Th, and Ta; (2) U; (3) HRE; (4) Ni; (5) Sr, and Rb; (6) Zn; (7) Sc; (8) Cr; and (9) Fe, Co, and As. These subgroups reflect the elementary difference in geochemical property and the closeness of element association in geological process. The nearer to 1.0 correlation factor, the closer elementary association.

Table 3 Grouping of REE and trace elements by group analysis.

Table options

Figure 3 Paragenetic associations of elements by R-mode group analysis.

Figure options

In the study area, the heavy mineral assemblages are divided into 4 types, and 22 elements 9 subgroups by way of methods of Q-mode factor and R-mode group analyses. The relationship between assemblages and the subgroups is shown in Fig. 4.

Figure 4 Paragenetic associations of elements by R-mode group analysis.

Figure options

Mineral assemblage Ⅰ₁ is distributed over the east of the study area, taking pyroxene, magnetite, and ilmenite as dominant minerals, and having more accessory minerals. This area is rich in U, Th, Zr, Hf, Ta, Sc, and Cr.

Mineral assemblage Ⅱ₂ is located in the bay around the South Shetland Is. Here the main minerals are pyroxene, magnetite, and ilmenite, with more pyrite, and being rich in Fe, Co, As, Ni, Sc, Cr, Sr, and Zn. Heavy mineral assemblage Ⅱ is situated in the western study area, taking rock fragment, green amphibole, and epidote as dominant minerals, and RE, U, Yh, Rb, Sr, and Zn are quite abundant.

Heavy mineral assemblage Ⅲ lies in the west of Bransfield Strait. In the assemblage, pyroxene, epidote, magnetite, ilmenite, and green amphibole are much abundant, and the trend of dispersion and enrichment of the subgroup elements is not clear (except Sr). The assemblage is situated between the assemblage Ⅰ and the assemblage Ⅱ, having the dominant minerals from both the assemblages Ⅰ and Ⅱ and with close content. Therefore, it should be born of transporting and mixing of the east and west end member assemblages by currents.

The two heavy mineral groups obtained by way of R-mode group analysis show that the surface sediments of the study area basically come from two kinds of mother rock area. One -of them is intermediate acidity, distributed over Elephant Is. (northeast of the study area) and Antarctic Pen. (southeast of the study area). Characteristic weathering products of this area are green amphibole, epidote, schistose minerals, undissociated altered mafic silicate minerals (rock fragments), and accessory minerals. The other is basic volcanic, namely South Shetland volcanic arc, basalt and basaltic andesite are widely dispersed over it. Characteristic weathering products of it are pyroxene, magnetite, ilmenite, volcanic, and glass. And pyrite is the marker of late alteration and mineralization for this kind of rock.