INTRODUCTION

Mesozoic magmatism widespread in South Africa and East Antarctica forms one of the World’s largest igneous provinces. It spans an area of ca. 3 000 000 km2 and represents a product of the Karoo mantle plume activity (Riley et al., 2005; Luttinen and Furnes, 2000; Duncan et al., 1997; Zhang et al., 2003; Jourdan et al., 2007а, 2007b). The duration of this magmatism is estimated differently by different researchers (Duncan et al., 1997; Dalziel et al., 2000; Marsh et al., 1997; Lawer et al., 1992; Curtis et al., 2008; Svensen et al., 2012; Luttinen et al., 2018), but is at least 20–30 Ma (Riley et al., 2006), which significantly exceeds the activity time interval for typical trap provinces (Cortillot and Renne, 2003). The peak of magmatic activity occurred within a narrow range of 182–183 Ma in the African area and within 184–178 Ma in the Antarctic sector (Duncan et al., 1997; Zhang et al., 2003; Jourdan et al., 2005, 2007b; Luttinen et al., 2015). Within East Antarctica (Queen Maud Land and its continental margin), Mesozoic magmatism spans an area over 700 000 km2 and is represented by dikes and minor intrusions of basic, more rarely alkaline and felsic composition, and mafic volcanic sequences filling riftogenic depressions (Leitchenkov et al., 1996, 2016). In addition to large area and long duration, the Karoo plume magmatism is characterized by the presence of high-Mg incompatible-element depleted rocks, which are scarce in other igneous plume provinces, as well as by magmatic zoning defined by a spatial distribution of high- and low-Ti melts (Heinonen et al., 2010; Luttinen et al., 2018).

The composition of mantle derivatives is mainly determined by the evolution of melts and mantle source, i.e., by primary geochemical characteristics. Plume-derived magmas are sufficiently heterogeneous in composition (Korenaga, 2004). According to the modern concepts, this is likely because the ascending plume additionally contains subducted sediments (Weaver et al., 1986), oceanic crust (Hofmann and White, 1982), continental lithosphere (McKenzie and O’Nions, 1983), or matter representing a mix of all or some of these components (Zindler and Hart, 1986). In addition, the plume may contain primitive mantle component, since hot spots are thought to consist of primitive mantle that experienced only insignificant melting during the early stages of the Earth’s evolution (Honda et al., 1993; Moreira et al., 2001). The study of the largest plume provinces provides insight into a composition of mantle sources for each definite case. Deciphering the evolution of the Karoo plume, one of the Earth’s largest plumes, its interaction with continental lithosphere, and compositional peculiarities of magmatic rocks are an urgent problem of modern geology. To unreveal magma source and its evolution, we studied dolerites from the western Queen Maud Land, the origin of which is related to the main stage of the evolution of the Karoo-Maud plume in Antarctica. Only few bedrock samples were collected during Russian Antarctic Expeditions, but their detailed study can provide the deeper insight into Mesozoic magmatism in Antarctica.

GEOLOGY OF THE WESTERN QUEEN MAUD LAND

The Queen Maud Land (QML) occupies a specious area between 20° W and 45° E in East Antarctica and differs from other Antarctic areas in the presence of a continuous mountainous chain up to 3 km high developed along coast. The western QML is occupied by the Vestfjella, Heimefrontfjella, and Ahlmannryggen and the Borg massif (Fig. 1).

Fig. 1.
figure 1

Modified geological map (Heinonen et al., 2010) of the western Queen Maud Land (wQML). (1) rocks exposed from beneath ice: (a) Archean granitoids (~3 Ga); (b) Mesoproterozoic volcanosedimentary complex (Ritscherflya Supergroup; 1130–1100 Ma) intruded by Jurassic dolerite dikes, (c) Mesoproterozoic gneisses (1200–1000 Ma) intruded by Jurassic dolerite dikes; (d) Jurassic basalts and dolerite dikes; (2) Archean craton (mainly after geophysical data); (3) Proterozoic mobile belt subjected to the Early Paleozoic tectonothermal reworking (based on geological and geophysical data); (4) boundary of tectonic provinces; (5) sampling localities with indication of sample numbers.

Full size image

The western Queen Maud Land (wQML) differs in a heterogeneous tectonic structure (Grikurov and Leitchenkov, 2012). It is made up of the ~3 b.y. old Grunehogna Craton regarded as a part of the Kaapvaal Craton (Elliot and Fleming, 2000; Luttinen and Furnes, 2000), which is surrounded in the east and southeast by the Mesoproterozoic mobile belt representing the Antarctic extension of the Natal Belt of Africa (Fig. 1). The Grunehogna Craton is overlain by the metamorphosed volcanosedimentary rocks of the Mesoproterozoic Ritscherflya Supergroup (Moyes et al., 1995; Groenewald et al., 1995).

Within the wQML, the plume magmatism is represented by mafic intrusions and volcanic complexes peaked at 180 Ma (Riley et al., 2005; Heinonen and Luttinen, 2008; Heinonen et al., 2010; Luttinen et al., 2015). Mesozoic flood basalts were found in the Vestfjella, Kirwanveggen, and Heimefrontfjella (Fig. 1). The largest-volume basaltic eruptions are known in the Vestfjella and southern Kirwanveggen (Harris et al., 1990). The thickness of the lava sequences reaches 900 m in the north and 400 m in the south of the Vestfjella at a general dip 10° west (Furnes et al., 1987). The flood basalts are cut by dolerite dikes and sills and gabbroic intrusions (Spaeth and Schüll, 1987). The base of the lava flows in the Vestfjella is not exposed, but as seen in other regions, these basalts rest on the Paleozoic continental sediments overlying Precambrian basement (Furnes et al., 1982, 1987). The dolerite dikes are also developed in the Kirwanveggen and Ahlmannryggen (Fig. 1).

The wide compositional diversity of magmatic rocks varying from mafic dikes and sills to ultramafic basic and felsic rocks (lamproites, subalkaline meymechites, quartz diorites) and the presence of high-Mg incompatible element-depleted magmas against the diversity of geochemical types (Luttinen et al., 2002; Riley et al., 2005; Heinonen and Luttinen, 2008; Luttinen et al., 2015) emphasize a complicated process of magma generation from initiation to crystallization. The highest Mg and respectively, highest temperature rocks weakly contaminated during crystallization are of most suitable for deciphering the magma generation conditions and composition of primary melts (Howarth and Harris, 2017; Hole, 2018). The oldest rocks occur only in the Ahlmannryggen (Riley et al., 2005). The studied dolerite dike samples from the Queen Maud Land were collected during 33rd Russian Antarctic Expedition, and sampling localities are shown in Fig. 1.

ANALYTICAL TECHNIQUES

Whole-rock compositions of the basalts (Table 1) were analyzed by X-ray fluorescence at the GEOKHI RAS (Moscow) on an PANalytical B.V. AXIOS Instrument equipped with an X-ray tube with a 3 kV Rh anode. The record was conducted using a scanning channel with analyzing crystals (PE-002-С, PX-1, GeIII-C, LIF-200, and LIF-220), and detector. Samples for analysis were manufactured by pulverizing the material (300 mg) to 200 mesh and then pressing into 20-mm pellets with the addition of polystyrene in the proportion 5 vs 1. A separate portion of material was used to determine L.O.I.

Table 1.   Content of major (wt %), trace (ppm) elements and isotope ratios in the studied samples of the Queen Maud Land
Full size table

The contents of lithophile elements was determined at the Central Analytical Laboratory of the GEOKHI RAS (Moscow) using an original technique of open-system digestion of 100 mg samples. A brief description of the technique is given below. A weighed sample was placed in a 60 mL Teflon glass with a spherical floor and mixed with concentrated acids (4 mL HF and 1 mL HNO3). Then, the glass was tightly closed with a massive cap, placed in volume heating block with thermocouple preliminarily heated to 50°С, and held for 72 h. Samples of unknown composition were digested together with standard samples (BCR-2 and BHVO-2). To estimate the element detection limit, the same procedure was carried out also for blanks. After digestion, the samples were evaporated at 140°С to wet salts. The cooled samples were mixed with 2 mL HF and 0.5 mL HClO4 (concentrated), the glasses were closed with caps with built in air refrigerators and kept for 12 hours at 140°С. Then, the samples again evaporated to wet salts, mixed with a mixture of concentrated 1.5 mL HNO3 and 1 mL HCl and heated to complete salt dissolution. Then, the samples were evaporated to a volume of ~0.5 mL.

Obtained solutions were transferred into a plastic tube and filled with 2% HNO3 to a volume of 25 mL. For ICP-MS analysis, obtained solutions were additionally diluted 20 times with 2% HNO3. Indium was added to solution as an internal standard before measurements to correct for instrumental drift. The concentration of the internal standard in samples was 10 μg/L. The measurements were carried out using an XSeries II (Thermo Scientific) inductively coupled plasma quadrupole mass-spectrometer at a plasma power of 1400 W, a plasma-forming gas (Ar) flow rate of 13 L/min, and auxiliary Ar flow rate of 0.95 L/min, Ar flow rate into nebulizer of 0.87 L/min, and analyzed sample flow rate of 0.8 mL/min. Under such conditions, the level of oxide ions CeO+/Ce+ was no more than 2%, while the relative fraction of two-charged ions (Ва++/Ва+) is no more than 3%. The measured spectra were processed using iPlasmaProQuad software (GEOKHI RAS), which provides import of all measured data into MS Access database and their processing with built-in facilities (including approximately 60 queries, with calculations). The processing involves plotting calibration curves, calculation of isotope concentrations, introduction of corrections for the internal standard, control of measurement reliability, estimation of uncertainties (with allowance for error accumulation at different stages of analysis), consolidation of data, verifying for correction of analysis of standard samples, and other functions.

The chemical composition of olivine and orthopyroxene was studied in grains selected from samples on a JEOL JXA 8230 Superprobe at the Institute des Sciences de la Terre (ISTerre), University Josef Fourier, Grenoble, France following method of high-precision trace-element determination (Batanova et al., 2015) (Table 2). According to this method, olivine grains, in addition to major components (Mg, Fe, Si), were analyzed for Na, Al, P, Ca, Ti, Ni, Mn Zn, Cr, and Co at an accelerating voltage of 25 kV, beam current (measured on a Faraday cup) of 900 nA. Trace elements were measured on five spectrometers with wavelength dispersion, while major components, on an energy dispersive spectrometer. Counting time was 12 minutes. The San Carlos olivine standard (USNM 111312/44) (Jarosewich et al., 1981) was analyzed as unknown three times every 30 measurements in order to monitor potential instrumental drift. This made it possible to control and correct the drift of device. The reproducibility of analysis was estimated using the control olivine sample as 2 standard deviations from the mean, accounting for 4–10 ppm for most trace elements. 15 ppm, for sodium, and 300 ppm for major end members.

Table 2.   Compositions of olivines and host orthopyroxenes
Full size table

The Sr, Nd, and Pb isotope compositions in the dolerites were determined at the Center for Isotopic Research of the Karpinsky All-Russian Research Geological Institute (VSEGEI, St. Petersburg) (Table 1). The elements were extracted using chromatographic and ion-exchange separation with previously described technique (Luchitskaya et al., 2017). The blanks during the analytical work were no more than 0.01 and 0.1 ng for Rb and Sr, and 0.02 ng for Sm and Nd, and 0.01 ng for Pb. The element contents were determined by isotope dilution with addition of calibrated isotope tracer. The element isotope compositions were measured on a multicollector solid-state TRITON (CIR VSEGEI) mass spectrometer in a static mode. Normalizing values are 88Sr/86Sr = 8.375209 and 146Nd/144Nd = 0.7219. The quality of measurements was controlled by the intermittent measurement of the JNdi-1 standard isotope composition: 143Nd/144Nd 0.512109 ± 0.000006 (n = 22), NIST-981: 206Pb/204Pb 16.913 ± 0.001, 207Pb/204Pb 15.451 ± 0.001, 208Pb/204Pb 36.594 ± 0.001 (n = 12), NBS-987 87Sr/86Sr 0.710225 ± 12 (n = 12).

COMPOSITION OF wQML DOLERITES

Figure 2 shows variations of TiO2 and K2O contents (Table 1) versus MgO content in the studied mafic rocks from the western QML compared to the general variations of plume magmas distributed in South Africa (Karoo province) and East Antarctica (Queen Maud Land). It is seen that the majority of samples are ascribed to the low-Ti type, which is widespread in the Karoo and QML provinces. Its typical representatives could be the Mesozoic dolerites of the Muren dike dolerite complex (not shown in the figure) located in the western QML (Vuori et al., 2003). Most part of these rocks represents strongly differentiated varieties with MgO content decreasing from 7 to 2 wt %.

Fig. 2.
figure 2

Compositions of studied basalts (Table 1) and comparative characteristics of high- and low-Ti magma types developed within the Karoo province in South Africa and Queen Maud Land according to (Luttinen, 2018). The compositions of Ahlmannryggen Fe-picrites fall in the field of high-Ti basalts. Basalts: (1) QML, (2) Karoo province; (3) samples. 33921-7, 333905-7, Ahlmannryggen, (4) other studied samples of the wQML (Fig. 1).

Full size image

The highest-Mg rocks most suitable for the reconstruction of primary melts are dolerite samples from the Ahlmannryggen massif with MgO content reaching 12 wt %. They plot in the basaltic group with elevated TiO2, but sufficiently low K2O (Fig. 2). According to compositions presented in (Luttinen et al., 1998), the dolerites of the Ahlmannryggen are ascribed to the high-Fe magnesian mafic rocks.

Dolerites represented by samples 33921-7 and 333905-7 of the Ahlmannryggen consist of 40–60 vol % plagioclase, ca. 20 vol % orthopyroxene, near 10 vol % clinopyroxene, and ore minerals. Olivine (1–3 vol %) forms fine rounded scarce grains and sometimes clusters in groundmass. In some cases, the former presence of olivine is inferred from the morphology and secondary products of completely altered crystals (33905-7). The groundmass consists of plagioclase, Ti-magnetite, and olivine microphenocrysts subjected to partial alteration (development of secondary amphiboles after orthopyroxene and olivine, partial saussuritization of plagioclase, and formation of chlorite). Some orthopyroxene crystals contain inclusions of high-Mg and high-Ni olivine (Fig. 3).

Fig. 3.
figure 3

Inclusions of olivines Fo85 and Fo87 in orthopyroxenes from dolerite samples 33921-5 and 33905-7 of the Ahlmannryggen (Table 2).

Full size image

Dolerites from samples 33921-7 and 333905-7 have the following average composition: SiO2—49.4, TiO2—3, MgO—11, Al2O3—10.5, FeO—13.2, CaO—9, Na2O—1.65, K2O—0.35 (wt %). They plot in the composition field of high-Ti and high-Fe basalts of the Karoo province. It should be emphasized that other studied samples are correlated with differentiated varieties of the QML plume magmas and are distributed south of the Ahlmannryggen and rather represent derivatives of low-Ti magmas. The TiO2 content varies within 1.6–2.3 wt % (at MgO 4.6–6.9 wt %). In spite of the sufficiently strong rock alterations, K2O in them varies from 0.3 to 1 wt %.

High-Ti magnesian ferrobasalts with characteristic cumulate trend of Mg enrichment were previously found within the Vestfjella and Ahlmannryggen (Riley et al., 2005; Heinonen and Luttinen, 2008; Heinonen et al., 2010; Luttinen et al., 2015). This group of magmas (type G-3 after Lambart et al., 2013) was distinguished at Ahlmannryggen by the presence of olivine phenocrysts with orthopyroxene inclusions. The studied dolerites of samples 33921-7 and 333905-7, unlike these magmas, are characterized by the crystallization of magnesian orthopyroxene with inclusions of small grains of magnesian olivine (Fig. 3). At the same time, these magmas have similar major- and trace-element composition (Fig. 4).

Fig. 4.
figure 4

Variations of major elements in Fe-picrites related to melting of pyroxenite source. Compositions: (1) ferropicrites according to (Luttinen, 2018); (2) ferropicrites (type G-3, Lambart et al., 2013) developed in the Ahlmannryggen area; (3) calculated model primary melts related to melting pyroxenite mantle (Lambart et al., 2013); (4) studied samples of picrite basalts (Table 1).

Full size image

COMPOSITION OF OLIVINE AND EQUILIBRIUM ORTHOPYROXENE

Orthopyroxenes from Ahlmannryggen dolerites (Table 2) are close in composition to the previously studied orthopyroxenes from high-Fe magnesian rocks developed in the Karoo and QML provinces, which are supposedly related to melting of a pyroxenite source (Kamenetsky et al., 2017). The Mg number of the studied orthopyroxene varies within Mgat# 84–88 and corresponds to conditions of equilibrium crystallization with Fo84–87 olivine. A decrease of Mg number in orthopyroxene is accompanied by an increase of TiO2 content from 0.26 to 0.43, Al2O3 from 0.9 to 1.4, CaO from 1.4 to 2.2, and a decrease of SiO2 from 56 to 54.8 (wt %).

It was found that orthopyroxene grains (samples 33921-7 and 333905-7) contain small olivine crystals (Table 2, Fig. 3), composition of which is shown in (Fig. 5). Major peculiarity of this olivine is the elevated Ni content, which is typical of olivine from a pyroxenite source in plume magmas (Sobolev et al., 2005, 2007). In particular, the evolution trends of Ahlmannryggen olivines are close to those of the plume magmas of the Karoo province (Letabo area) in South Africa and Siberia (Fig. 5), which are unambiguously related to melting of a pyroxenite source (Sobolev et al., 2007; Kamenetsky et al., 2017). In spite of the fact that olivines from Siberian traps (Gudchikha Formation) are more fractionated, their extrapolation to olivines equilibrated with a mantle source indicates their similarity with the studied Ahlmannryggen olivines. Olivines of the Gudchikha Formation mark the early stage of the Siberian plume activity, and, as was shown, were generated through melting of an olivine-free source (Sobolev, 2009). Mafic magmatism of the Letabo area is dated at 184.2–181.2 Ma (Duncan et al., 1997), i.e., belongs to the early stages of the Karoo plume activity in South Africa. Olivines from magnesian olivine porphyry rocks of the Letabo area (Kamenetsky et al., 2017) in terms of Ni, Mn, and Ca contents are comparable with studied olivines (Fig. 5). With decreasing Mg# from 87 to 84 Fo in olivines of the QML, the NiO content decreases from 0.5 to 0.4 wt %, Cr2O3 from 0.78 to 0.68, while MnO increases from 0.15 to 0.19 wt %. It is seen in Fig. 5f that in terms of 100 × Mn/Fe–Ni/(Mg/Fe)/1000, Ahlmannryggen ferropicrites fall in the field of pyroxenite mantle derivatives. Olivines from melts in equilibrium with a peridotite source differ in the highest 100 Mn/Fe and the lowest Ni/(Mg/Fe)/1000) values, while the composition field of olivines in melts generated from a pyroxenite source have the highest Ni/(Mg/Fe)/1000 and the lowest 100 Mn/Fe. The calculation of the amount of pyroxenite component in the source from Ni/Mn ratio in olivine using the proposed relations (Sobolev et al., 2007): XPxNi = 10.54× NiO/(MgO/FeO)—0.4368; XPxMn = 3.483 – 2.071 × (100Mn/Fe) yields almost 100% of melting pyroxenite, i.e., the compositions of the studied magmas could correspond to the initial composition of picritic melts formed through melting of an olvine-free source.

Fig. 5.
figure 5

Variations of trace elements in olivines of the Ahlmannryggen, mafic rocks from the Karoo province, and pyroxenite-related magmas of the Siberian plume. (a–e) variations of Al2O3, NiO, Cr2O3, СаО, and MnO contents versus olivine composition. Compositions of olivines from basalts of the Karoo area (triangles), Siberian plume (spots), Ahlmannryggen (triangles) form separate clusters. The elevated contents of these elements in the studied samples are close to those in the Siberian traps derived from pyroxenite source. Data on the traps of the Karoo and Siberian plume are from (Sobolev et al., 2007, 2009). (f) variations of olivine composition in the diagram 100Mn/ Fe–Ni/(Mg/Fe)/1000 (Sobolev et al., 2007), which allows discrimination between pyroxenite and peridotite sources of primary melts (shown as fields). Fe- and Mg/Fe-normalized contents of trace elements in olivine makes it possible to estimate their concentrations in olivines with similar Fe content and Mg/Fe ratio to avoid the effects of melt differentiation and reflect true variations of Ni and Mn in melt.

Full size image

GEOCHEMICAL FEATURES OF MELTS

Variations of primitive mantle-normalized lithophile element pattern in the studied samples (Fig. 6a) demonstrate large compositional dispersion, which was previously noted for basaltic flows of QML (Luttinen et al., 1997, 1998). Unlike the studied basalts of the Kirwanveggen and Vestfjella (Table 1), the high-Mg Ahlmannryggen dolerites have sufficiently persistent weakly enriched LILE pattern: (La/Sm)n = 0.9–1.1 (Fig. 6a). The trace-element pattern for lavas of the Siberian traps derived from a pyroxenite source is shown for comparison (Fig. 6b). At generally similar distribution patterns, the Siberian traps differ in the elevated contents of the most incompatible elements compared to the andesitic glass formed likely through melting of a fragment of the Gondwana lithosphere in the South Atlantic spreading zone (Kamenetsky et al., 2001). The composition of this glass is frequently used as the reference composition for evidence of melting of typical mantle beneath the southern part of the Gondwana supercontinent, which experienced plume impact in the Mesozoic.

Fig. 6.
figure 6

Lithophile element distribution: (a) in the studied basalts of the Vestfjella and Ahlmannryggen; (b) in magmas of the Ahlmannryggen (samples 33921-5, 33905-7) compared to their contents in glasses of the South Atlantic (S18-60/1—Kamenetsky et al., 2001) and Siberian traps derived through melting of pyroxenite mantle (Sobolev et al., 2009). Lithophile elements are normalized after (Sun and McDonough, 1989).

Full size image

Figure 7 shows variations of isotope compositions of Ahlmannryggen and Vestfjella basalts (Table 3) compared to their variations in the magmas of the Karoo and QML provinces. The studied samples fall in the field of enriched magma varieties of these provinces (Fig. 7а) and differ from QML depleted basalts. It should be noted that the revealed isotope systematics of high- and low-Ti magmas of South Africa differed from that of the Antarctic magmas, which show both depleted and enriched signatures. However, it is of great importance that the isotope compositions of dolerite samples 33921-7 and 333905-7 in the isotope diagrams 206Pb/204Pb vs. 208Pb/204Pb and 206Pb/204Pb vs. 87Sr/86Sr (Figs. 7b and 7c) fall in the field of high-Ti basalts of South Africa and significantly far from other basaltic rocks of the QML (Heinonen et al., 2010; Kamenetsky et al., 2017). They are close to the previously mentioned basalts from the Letabo province (Nuanetsi). The isotope signatures of the studied Ahlmannryggen dolerite samples are characterized by decreased values of 206Pb/204Pb: 17.33–17.37, 207Pb/204Pb: 15.37–15.52, 208Pb/204Pb: 37.40–37.79 and relatively increased 143Nd/144Nd: 0.51249–0.51259 and Sr 87Sr/86Sr: 0.7049–0.7063, which to a first approximation could be regarded as characteristics of a pyroxenite mantle source of enriched magmas (Fig. 7). In terms of isotope-geochemical composition, they are most close to mafic (oceanic) crust or mafic eclogite, which simultaneously bears signatures of a source depleted in lithophile elements (lowered 206Pb/204Pb) and enriched in some of them (for instance, in Rb owing to an increased 87Sr/86Sr introduced by sedimentary component (Hawkesworth et al., 1984).

Fig. 7.
figure 7

Isotope characteristics of melts related to melting of pyroxenite source. (a) Sr-Nd isotope systematics of Ahlmannryggen and Nuanetsi basalts compared to high- and low-Ti magmas of the Karoo and Queen Maud Land provinces. Shown are magma compositions: (1) Karoo (South Africa), (2) Queen Maud Land; (3) Ahlmannryggen picrites (Table 1); (4) basalts of the QML (Table 1). Nd and isotope compositions are recalculated to the eruption time (180 Ma). (b, c) variations of isotope ratios in magmas related to the Karoo plume. (1–2) South Africa basalts: (1) high-Ti, (2) low-Ti; (3) basalts and dolerites of different geochemical types, Queen Maud Land; (4) Ahlmannryggen picrites (Table 1); (5) Vestfjella basalts and dolerites, QML, (6) Nuanetsi high-Ti ferrobasalts; (7) enriched (EM I, EM II, HIMU) and (DM) depleted model sources after (Armienti and Longo 2011). Abbreviation L-Ti means low-Ti basalts, H-Ti—high-Ti basalts. Compiled using data in Table 1 and from (Heinonen et al., 2010; Heinonen et al., 2016; Luttinen, 2018; Luttinen et al., 2015) for the Queen Maud Land and (Ellam and Cox, 1989; Heinonen et al., 2014, Jourdan et al., 2004) for South Africa.

Full size image

DISCUSSION AND GEODYNAMIC INTERPRETATION

The contribution of a pyroxenite source in the formation of plume-related traps is also suggested in some papers (Hirschmann, Stolper, 1996; Elkins et al., 2019; Sobolev et al. 2009), however, unambiguous and direct evidences for its existence are extremely scarce, because it is stable within a narrow Р-Т range (Sobolev et al., 2009). Numerous works show the presence of such melts in different proportions in the parental magmas (Lambart et al., 2012; Matzen et al., 2017; Søager et al., 2015; Yang et al., 2016), but only our studies discovered practically undifferentiated magmas derived through melting of a pyroxenite source, and thus demonstrated their existence. The appearance of such magmas at the initial stages of plume activity could reflect directly the composition of ascending plume magma reaching surface. The experimental studies of pyroxenite melting at pressures of 20–25 kbar showed the existence of melts close in composition to the studied samples (Lambart et al., 2013, 2016). It was shown (Heinonen et al., 2013) that the primary melts of Ahlmannryggen ferropicrites were formed through fractionation of olivine and orthopyroxene. In this work, the compositions of model primary melts calculated by addition of fractionated olivine and orthopyroxene in proportions (2–3% Fo84 + 6–9% Opx) yielded compositions close to the compositions of dolerites 33921-7 and 333905-7 (Fig. 4). Since these compositions are close to model primary melts, they could represent to a first approximation undifferentiated melts formed through melting of a pyroxenite source. Similar melts differ in the low MgO content near 10 wt %, which is typical of the studied rocks. In turn, the melts derived from pyroxenites should be characterized by low SiO2 and high FeO contents, as well as high CaO/Al2O3 ratio (Elkins et al., 2019). However, this is not typical of the Ahlmannryggen melts, which is determined by the wide heterogeneity of eclogites, which give rise to pyroxenite during upwelling of mantle diapir. Therefore, obtained geochemical and isotope characteristics of the studied samples are considered as reflecting the composition of a pyroxenite source of magmas formed by impact of the Mesozoic Karoo plume on the lower parts of lithosphere or in the upwelling plume proper.

The South Africa and western QML magmas related to the Karoo plume activity show a significant geochemical heterogeneity. Such heterogeneity of plume magmas is confirmed by numerous studies of magmatic rocks of igneous provinces of different age (Cox, 1989; Courtillot and Renne, 2003; Encarnación et al., 1996; Elliot and Fleming 2000; Jourdan et al., 2007; Heinonen et al., 2013; Callegaro et al., 2013; Sobolev et al., 2009, and others). Its formation could be explained by both assimilation of diverse crustal rocks and delamination of lithospheric roots during plume impact, the low-degree melting of the lower lithosphere, and subsequence mixing of magmas in variable proportions. It is difficult to distinguish a definite or prevailing mechanism. However, the detailed study of magmatism of the Hawaiian Archipelago (Hawaiian plume has evolved and continues to evolve beneath thinned oceanic lithosphere), as well as Siberian traps (plume upwelling beneath a thickened lithosphere) showed that mantle plumes could bear numerous pyroxenite fragments, whose melting is established from geochemical features of liquidus olivine (Sobolev et al., 2007, 2009; Yang et al., 2016). Formation temperature of Ahlmannryggen melts from mantle pyroxenite determined by different methods varies within 1500–1700°С (Hole, 2015; Heinonen et al., 2015). According to model (Sobolev et al., 2007), mantle jet containing ca. 10–15% recycled eclogite begins to melt at depths near 200 km, thus forming SiO2-rich melts, which in turn, interact with peridotite to form a hybrid pyroxenite (Sobolev et al., 2007). Numerical modeling showed that ascending mantle plume has a potential temperature no less than 1650оС (Sobolev et al., 2009). The formation of a thermal boundary layer between a relatively cold depleted lithosphere and hot core of mantle jet prevent lithosphere heating to jet temperature even during 10 Myr and cannot melt lithospheric mantle. Therefore, lithospheric pyroxenite cannot be a source for Ahlmannryggen melts. It is more probable that upwelling mantle plume has already contained the fragments of dense recycled crust (Stroncik and Devey, 2011; Day et al., 2009). Generation of melts enriched in Ti, Fe, and Mg is related to melting of a pure pyroxenite source, which should occur at the early stage of plume emplacement; otherwise, they would be mixed with melts derived from peridotite mantle. In spite of the fact that the reliable geochronological determinations are few in number, the Ahlmannryggen picritic magmas (dolerite samples 33921-7 and 333905-7 as typical representatives) with an age of 190 Ma (Riley et al., 2005) and the Vestfjella with an age of 183–180 Ma (Luttenen and Furnes, 2000) likely reflect this early stage.

The Ahlmannryggen melts show specific LILE distribution pattern similar to that of the Gudchikha basalts of the Siberian trap province, which correspond to the initial stage of the Siberian Plume evolution with an age of 250 Ma (Fig. 6). The main peculiarity of these magmas is the presence of Ni-rich liquidus olivines, which indicate a direct relation with melting of a mantle pyroxenite component in upwelling plume (Sobolev et al., 2007, 2009). This component differs in a sufficiently smoothed normalized lithophile element pattern, with enrichment in more incompatible elements relative to more compatible elements (Th/Nb)n = 1.2–1.17; (Zr/Y)n = 4.8–6.2, which is determined by the presence of garnet in a source (Hirschmann and Stolper, 1996; Tuff et al., 2005). Magmas of similar composition could be formed by 40% partial melting of pyroxenite, which was derived by the interaction of partial melt (60%) of recycled eclogite (CRC) with primitive mantle peridotite (PM) (Sobolev, 2009).

Figure 8 shows isotope characteristics of end members of mantle sources: depleted peridotite source (DM), which generates pyroxenite-free oceanic type melts and pyroxenite (Px) mantle source, whose melting produces melts similar to those of the Ahlmannryggen magmas. Tholeiitic magmas of the Bouvet triple junction in the South Atlantic shown in this plot revealed an intermediate composition and are characterized by the wide variations in proportions of enriched pyroxenite component in a source region (Migdisovа et al., 2017). A source of picritic magmas (based on the composition of analyzed samples) forming through melting of pyroxenite mantle has the following isotope parameters: 206Pb/204Pb 17.35, 207Pb/204Pb 15.45, 208Pb/204Pb 37.5, 143Nd/144Nd 0.5125, and 87Sr/86Sr 0.706. Unlike pyroxenite source occurred in the Bouvet hot spot area, pyroxenites formed at the initial stage of the Karoo–Maud plume activity were significantly depleted in U and Th. It was shown (Sushchevskaya et al., 2019) that such isotopic characteristics could be produced through two-stage formation of a parental mantle source. The early stage, was responsible for the formation of ancient recycled component, i.e., oceanic crust mixed with sediments. At the next stage, this component was converted into mantle pyroxenite through the interaction with depleted mantle peridotite (Sushchevslaya et al., 2019). Diverse paleotectonic models for the Late Precambrian evolution of this region (Ellam and Cox, 1989; Jacobs et al., 1993; Groenewald et al., 1995; Ferraccioli et al., 2005; Grosch et al., 2007; and others) suggest ancient subduction of oceanic crust at 700–800 Ma during amalgamation of the Gondwana supercontinent.

Fig. 8.
figure 8

Isotope characteristics of pyroxenite source according to data on Ahlmannryggen picrite magmas. Compositions of oceanic magmas peridotite source and of melts from spreading zones of the Bouvet triple junction are from (Sushchevskaya et al., 2003).

Full size image

As mentioned in the Introduction, plume magmas could be mixed with material of subducted slabs to acquire composition differing from typical mantle derivatives (Weaver et al., 1986; McKenzie and O’Nions, 1983; Zindler and Hart, 1986). Subduction and plume tectonics played a significant role in the formation of the lithosphere and the Earth’s crust of Antarctica during Proterozoic and Phanerozoic (Gorczyk et al., 2018; Veevers, 2012). In the Late Proterozoic (800–600 Ma), the western part of the Queen Maud Land represented a mobile belt (magmatic arc) with inferred subduction of the Mozambique ocean and subsequent collision of continental blocks of Antarctica at the final stage of the Gondwana amalgamation (Pauly et al., 2016; Jacobs et al., 2020). The material of the Late Proterozoic subducted slabs through mantle convection could affect the composition of the Early Jurassic plume magmas, but it is more probable that the geochemical specificis of the Karoo plume was determined by later geodynamic processes. The final assembly of Gondwana in the Early Paleozoic was followed by steady subduction with Paleopacific active margins along West Antarctica, New Zealand (with surrounding submarine plateau) and East Antarctica. In the Late Paleozoic–Early Mesozoic, subsiding slab determined the regimes of compression and extension of the Antarctic lithosphere in back-arc settings, as well as, at significant distance from active margin. In the Permian, the ascending plume collided with subsiding slab, which led to flattening of its trajectory, the development of compressional strains in the lithosphere bottom, and the formation of the Gondwana within-plate folding (Dalziel et al., 2000). Further mantle convection led to the break up of subsiding slab, contamination of magma, and manifestation of Jurassic magmatism at the surface of the Gondwana supercontinent (within the Queen Maud Land and South Africa) (Dalziel et al., 2000), while the mantle composition reflects the interaction of ascending mantle plume with upper mantle (Hastie et al., 2014; Heinonen et al., 2013, 2014, 2016).

Numerical modeling showed that a recycled crustal material, if present in upwelling plume, will be melted at a depth of 150–170 km to form andesitic melts, the further interaction of which with peridotite substrate could produce reaction pyroxenite (Sobolev et al., 2005). During further upwelling, at depths of 150–120 km, this reaction pyroxenite melted to form pyroxenite melts. Thick lithosphere on the way of ascending plume leads to melting of the reaction pyroxenite (Natali et al., 2017). Forming melts could reach surface under favorable conditions, when lithosphere is subjected to early destruction (Gorczyk et al., 2018). At a depth ca. 100 km, mantle peridotite is involved in melting.

Forming low-Ti magmas are widely distributed and related to melting of metasomatized lithospheric mantle, which contains enriched ЕМII type component (Fig. 7) typical of all trap magmas (Melankholina and Sushchevskaya, 2019). These low-Ti magmas reach surface near the marginal parts of the thick Zimbabve–Kaapval craton (Peters et al., 1991) and Queen Maud Land mobile belt, which represents the extension of the Natal orogenic belt formed during disintegration of the Proterozoic Rodinia supercontinent (Jacobs et al., 1993).

A spatial geochemical zoning of flood basalts, in particular, the development of low- and high-Ti basalts, plays an important role in determining the position of plume center (Heinonen et al., 2010, 2018; Natali et al., 2017; Luttinen et al., 2010; Luttinen, 2018). The high-Ti basalts, and especially ferropicrites likely mark the central, high-temperature part of the plume that was emplaced in the Africa–Antarctica lithosphere in the Nuanetsi–Ahlmannryggen areas (Fig. 9). Plume zoning in the Antarctic sector with uplifted central part around 500 km in radius (spanning the studied area, Fig. 1) is confirmed by geophysical data (Leitchenkov and Masolov, 1997).

Fig. 9.
figure 9

Reconstruction of the Karoo–Maud plume presented for the time of beginning of the Gondwana breakup (180 Ma). Fields show the distribution of different geochemical rock types in the Karoo igneous province, which reflects plume zoning. Compiled using (Ellam and Cox, 1989; Ferraccioli et al., 2005; Heinonen et al., 2016; Kamenetsky et al., 2017; Luttinen, 2018).

Full size image

Melts derived from a pyroxenite source could reach surface along fracture system. During further plume evolution, they mixed with the higher degree melts from peridotite source. The heterogeneity of magmatism in the marginal parts of the ancient Grunehogna Craton reflects a rapid transportation of forming melts along fracture system. Unlike the Siberian plume, whose impact did not caused a breakup of the Siberian Craton, the Karoo plume initiated magmatism over a spacious continental area and led not only to lithosphere thinning and uplift in the central part of the Gondwana supercontinent, but also to the subsequent intracontinental rifting with abundant magmatism, break-up of continent, and separation of Africa from Antarctica (Hastie et al., 2014; Leitchenkov et al., 2016). The central part of the plume is also marked by the largest volume of erupted magmatic material restricted to the rift margin of the eastern Weddell Sea, also abundant eruptions in the Lazarev Sea are developed beyond its limits, up to the plume margin (Leitchenkov et al., 2016). The thickness of the igneous accretion layer in the lower crust (underplating), in the central part of the plume at the Queen Maud Land according to deep seismic sounding data reaches 17 km (Leitchenkov et al., 2016).

CONCLUSIONS

Flood basaltic magmatism related to the Karoo plume gained the wide distribution in the South Africa and eastern Antarctica. It can be considered as marking a unique plume activity in the Mesozoic time, which is reflected in the formation of spatially separated high- and low-Ti magma types, and the confinement of the high-Ti magmas to the central part of the plume.

The manifestation of the Karoo–Maud plume is peculiar in the presence of high-magnesium, Fe-rich magmas depleted in incompatible elements among a geochemical diversity of magma types. Such high-Mg ferropicrites are scarce in other plume-related igneous provinces and their formation is related to melting of a peculiar pyroxenite mantle source. Similar melts were found only at Ahlmannryggen and Vestfjella in Antarctica, and in the Letabo province in South Africa, are correlated to the central part of the plume and likely correspond to the earliest eruptions.

The geochemical study of Ahlmannryggen basalts showed that the melts derived from a pyroxenite source of the Karoo plume differ in a relatively smoothed lithophile element pattern (from Th to Er) and the lowered content of most compatible elements (Y, Yb, Lu), they have low 206Pb/204Pb – 17.33–17.37 and moderately radiogenic Nd compositions with 143Nd/144Nd from 0.51249 to 0.51259, which indicate a relative old age of the pyroxenite component. The compositions of all high-Ti basalts point to the admixture of this source, but its proportion could vary significantly.