Thomas H. Fleming

Synchronous emplacement of Ferrar and Karoo dolerites and the early breakup of Gondwana

Constraints on the timing of Karoo and Ferrar continental flood-basalt magmatism in Africa and Antarctica, respectively, are critical to understanding the relationship of the Karoo and Ferrar to mantle plumes, subduction, and the initial breakup of Gondwana. Although recent work has shown that Ferrar magmas were emplaced over a short interval (<1 m.y.), the timing of magmatism within the Karoo and its relationship to the Ferrar have been problematic. New zircon and baddeleyite U-Pb ages on Ferrar (183.6 ± 1.0 Ma) and southern Karoo (183.7 ± 0.6 Ma) dolerites demonstrate that part of Karoo magmatism occurred during the rapid emplacement of Ferrar magmas. A mantle plume is thought to have been important in the genesis of the Karoo province, whereas lithospheric extension, perhaps related to subduction, has been invoked for Ferrar magmatism. The new ages now suggest that Ferrar and southern Karoo magmatism were related to a single mantle thermal anomaly and rifting event. This event may have produced local rift basins and caused rotation of microblocks in west Antarctica several million years before the breakup of east and west Gondwana.

A short interval of Jurassic continental flood basalt volcanism in Antarctica as demonstrated by 40Ar39Ar geochronology

A continental flood basalt province, the Ferrar Group (Kirkpatrick Basalt and Ferrar Dolerite), crops out along 3000 km of the Transantarctic Mountains in Antarctica and is temporally related to the break-up of Gondwanaland. Although a wide range of dates, between 90 and 193 Ma, have been published for the Kirkpatrick Basalt, it is now recognized that the young dates reflect non-ideal behavior of Ar in the matrix. In order to refine the geochronology, feldspar separates have been analyzed by the 40Ar39Ar incremental heating method. The main objectives are to constrain the duration of extrusive activity and the timing of volcanism along the outcrop belt. Basalt samples have been studied from the three principal outcrop areas, yielding the following apparent ages: central Transantarctic Mountains 176.8 ± 0.5 Ma; south Victoria Land 176.4 ± 0.4 Ma; north Victoria Land 176.6 ± 0.7 Ma. Ages from different stratigraphic levels within each area and from the three different areas are not analytically distinct. The data imply that the eruptive activity which produced the Kirkpatrick Basalt occurred within a short interval of less than about 1 m.y. at 176.6 ± 1.8 Ma, over an area which included more than 1200 km of the Transantarctic Mountains. The Jurassic volcanism in Antarctica represents a short episode of magmatism, comparable in duration with other well dated continental flood basalt provinces. The linearly extensive outcrop of the Ferrar Province and the rapid eruption of the lavas suggests that lithospheric stretching exerted a major control on magmatism. The poorly constrained age of the Bajocian-Bathonian boundary makes the previously suggested connection between Ferrar volcanism and an extinction event at that boundary uncertain.

Occurrence and Dispersal of Magmas in the Jurassic Ferrar Large Igneous Province, Antarctica

Tholeiitic rocks of the Ferrar Large Igneous Province (FLIP) occur in a linear belt from the Theron Mountains to Horn Bluff in the Transantarctic Mountains and extend into southeastern Australasia. The FLIP was emplaced during the initial stages of Gondwana break-up from a source suggested to be in the proto-Weddell Sea region. Magma transport from its source (Weddell triple junction) was controlled by an Early Jurassic zone of extension. The FLIP comprises the Dufek intrusion, Ferrar Dolerite sills and dykes (sheet intrusions), and extrusive rocks consisting of pyroclastic strata overlain by Kirkpatrick Basalt lavas. The Dufek intrusion occurs in deformed supracrustal rocks of the foldbelt along the paleo-Pacific Gondwana margin. A few sills were emplaced in basement rocks, but the majority of the sheet intrusions occur in flat-lying Devonian to Triassic Beacon strata. Only in the central Transantarctic Mountains (CTM) and south and north Victoria Land (SVL, NVL) are extrusive rocks preserved overlying Beacon strata. The greatest cumulative thicknesses of magmatic rocks (ca. 2 km) occur in areas where lavas are preserved (CTM and SVL). Sheet intrusions have complex relationships. Dyke swarms (sensu stricto) are unknown and dykes cutting basement rocks are uncommon. Nevertheless, these dykes, including a 30-m-wide dyke in SVL, suggest that some magmas locally migrated up through basement rocks. In CTM and NVL the outcrop belt has a width of about 160 km. Sills originally extended farther toward the plate margin but have been cut out by erosion and Cenozoic faulting, most clearly in CTM; geophysical data suggest extension under the East Antarctic ice sheet for at least 100 km. Although Early Jurassic extension is documented in CTM, major rift-bounding faults have not been observed. Models for magma emplacement include transport along the axis of the Transantarctic Mountains and off-axis transport from major rift-bounding faults. Contrasts in geochemistry between lavas of NVL (MgO=67%) and CTM (MgO=24%) and the presence of massive dolerite bodies (CTM, SVL) suggest discrete episodes and locations of magma emplacement, and that there was no long range interconnection along the mountain range in supracrustal rocks.

Long-distance transport of magmas in the Jurassic Ferrar Large Igneous Province, Antarctica

Abstract The youngest preserved lava flows of the Jurassic Kirkpatrick Basalt of the Ferrar Large Igneous Province are unusually distinctive. These flows, recognized in three geographic areas spread over 1600 km, have identical stratigraphic positions and geochemical characteristics. The lavas have an evolved, Fe-rich tholeiitic composition with lithospheric trace element ratios and enriched 87Sr/86Sr and 143Nd/144Nd isotope ratios. Except for the highly mobile elements, the major, trace and isotopic compositions of lavas from all these localities lie within, or near, analytical precision of each other, and are distinct from other Ferrar rocks. Moreover, the ages of the flows at all localities are indistinguishable. The unique characteristics of these capping lavas suggest that they were derived from a single batch of magma. Magma dispersal from a single reservoir through dike swarms at middle to upper crustal levels is considered the most probable mechanism for large-scale transport that extended for more than 3000 km.

Weddell triple junction: The principal focus of Ferrar and Karoo magmatism during initial breakup of Gondwana

Middle Jurassic Ferrar tholeiites of Antarctica were emplaced during a short time interval (<1 m.y.) into an active rift system initiated in the Early Jurassic. Ferrar magmas were dispersed into the Antarctic sector of Gondwana from a source in the Weddell Sea region, within the thermal anomaly that also gave rise to the Karoo basaltic rocks of southern Africa. The Golden Gate lavas, part of the Karoo central area low-Ti tholeiites, show geochemical similarities to Ferrar rocks. Tectonic and geochemical relationships of the Ferrar and Karoo low-Ti magmas, which constitute a major part of the Jurassic Gondwana large igneous province, suggest that they were derived from a single source associated with a triple junction in the proto-Weddell Sea region.

40Ar/39Ar geochronology of Ferrar Dolerite sills from the Transantarctic Mountains, Antarctica: Implications for the age and origin of the Ferrar magmatic province

The Ferrar Dolerite constitutes the hypabyssal phase of the tholeiitic Ferrar Group of Antarctica. Sills with compositions representing most of the range of geochemical variation of the Ferrar Dolerite, and separated by distances of as much as 1400 km, have been analyzed by the 40 Ar/ 39 Ar method on feldspar and biotite separates. The 40 Ar/ 39 Ar ages for five individual sills range from 176.2 to 177.2 Ma and show no significant difference. These ages reflect crystallization at 176.7 ± 1.8 Ma (where the uncertainty includes provision for systematic uncertainty in the age of the neutron-fluence monitor calibrated relative to MMhb-1 at 513.5 Ma). Combining data from these sills with previous determinations on coeval lavas and underlying pyroclastic units indicates an age of 176.6 ± 1.8 Ma for the Ferrar tholeiitic rocks as a whole. The duration of magmatic activity was less than approximately 1 m.y. By extension, other rocks in the Ferrar magmatic province, which occur from southeastern Australia, along the Transantarctic Mountains to the Theron Mountains, are inferred to have this age. The short duration of magmatic activity as well as the consistent pattern of geochemical variation and distinctiveness of the Ferrar rocks suggest that magmas were transported laterally by an extensive dike swarm which is inferred to have originated in the Weddell Sea sector of the province.

Potassium-argon dating of fine-grained basalts with massive Ar loss: Application of the 40Ar39Ar technique to plagioclase and glass from the Kirkpatrick Basalt, Antarctica

40Ar39Ar incremental-heating measurements are reported for separated plagioclase and matrix fractions and a glass sample of the Kirkpatrick Basalt from Victoria Land, Antarctica. These results are used to address: the limitations of whole-rock analyses of glassy or very fine-grained basalts; the use of plagioclase for determining ages of such rocks; the Ar behavior of matrix, glass and feldspar components; the use of K/Cl ratios in interpretation of step-heating results; and the age of the Kirkpatrick Basalt. The precise age of the Kirkpatrick Basalt has been a long standing problem; the wide range of reported K-Ar dates has been attributed to loss of Ar. Despite massive Ar loss from the matrix, plagioclase is shown to yield good 40Ar39Ar plateaus. Plateaus are observed for increments above 700°C and are defined not only by consistent apparent ages but also appropriately low K/Ca and high K/Cl ratios. Cl relations help identify Ar released from small amounts of matrix which can seriously contaminate a plagioclase separate. An age of 176.8 ± 1.8 Ma defines the time of formation of the youngest lava at Mt. Frustum in north Victoria Land. A glass sample from south Victoria Land gives the same age, indicating that terrestrial basaltic glasses as old as Jurassic do not inherently lose Ar. The step-heating profile of glass with very fine crystals shows discordance which is an artifact of Ar recoil during irradiation.

Isotopic and chemical constraints on the crustal evolution and source signature of Ferrar magmas, north Victoria Land, Antarctica

Isotopic (Nd and Sr) and chemical compositions of the 177 Ma Kirkpatrick Basalt and Ferrar Dolerite from north Victoria Land, Antarctica, are examined in order to address the role of crustal assimilation and the characteristics of their mantle source. Results for the Scarab Peak chemical type (SPCT) that constitutes the flow unit capping the lava sequence [Mg-number, Mg/(Mg+Fe+2)=24, MgO=2.4%, SiO2=57.1%, initial87Sr/86Sr=0.7087–0.7097, (εNd=−4.3) conform previous reports that attribute variations in the concentrations of the more mobile elements and calculated initial87Sr/86Sr to mid-Cretaceous alteration and elevated δ18O to low-temperature interaction with meteoric water. The underlying lavas and the sills that are of the Mt. Fazio chemical type (MFCT) display a much wider range of both chemical and isotopic compositions (Mg-number=40–65, MgO=3.7 7.5%; SiO2=52.6–58.3%, initial87Sr/86Sr=0.7087–0.7117, εNd=−5.6 to −4.8). The effects of rock alteration on apparent initial87Sr/86Sr are demonstrated by large differences between the initial ratio of mineral separates or leached fractions and whole rocks. Cretaceous alteration produced Rb and Sr redistribution within the lava sequence that results in erroneous calculated initial87Sr/86Sr ratios. These effects are responsible for the large initial87Sr/86Sr variations previousl7 proposed which, combined with the large range in whole-rock δ18O, were purported to show very large degrees of crustal assimilation. The variations in εNd are restricted and indicate much smaller degrees of assimilation. The least altered of the MFCT rocks show good chemical and isotopic correlations that can be integrated into a model involving fractionation of pyroxene and plagioclase coupled with assimilation of material similar to early Paleozoic basement. The lower87Sr/86Sr and higher εNd of the SPCT suggest that they were derived by extensive fractionation of a more primitive, less contaminated, precursor of the MFCT. The most isotopically primitive Ferrar rocks from the region still have a high initial87Sr/86Sr and low initial143Nd/144Nd; this may reflect either earlier assimilation or an enriched source. The chemical and isotopic similarities, as well as the close geographic correspondence of the Ferrar Group to granitoids produced during the early Paleozoic Ross Orogeny suggest that in either case Ross-type material may have been involved in the development of the enriched isotopic signature.

Chemical and isotopic variations in an iron-rich lava flow from the Kirkpatrick Basalt, north Victoria Land, Antarctica: implications for low-temperature alteration

Chemical and isotopic (Sr, O, H) variations have been examined in an iron-rich lava flow of the Kirkpatrick Basalt from the Mesa Range in north Victoria Land, Antarctica. The flow is homogeneous with respect to the less mobile elements, whereas variations observed in K, Na, Si, Fe, and Rb result largely from alteration of glassy matrix material. Whole-rock Rb−Sr isotope data fall along a poorly-defined 103 Ma array attributed to secondary mobilization of Rb during the mid-Cretaceous. Alteration at that time is suggested by paleomagnetic data and would also account for discordant K−Ar dates. Whole-rock δ18O values vary from +5.8 to +8.2‰ and a plagioclase separate has a δ18O value of +5.6‰, reflecting the original composition of the magma. The range of δ18O values for the whole-rock samples results from low-temperature alteration occurring primarily in the Jurassic and/or mid-Cretaceous. Whole-rock δD values (-201 to -243‰) are markedly depleted, approaching equilibrium with modern meteoric water. In light of these data, variable Sr and O isotopic ratios in the underlying sequence of flows, previously interpreted in terms of an assimilation-fractionation model, may largely reflect post-magmatic alteration.

An early Cretaceous volcanic arc/marginal basin transition zone, Peninsula Hardy, southernmost Chile

Abstract The Hardy Formation represents a latest Jurassic-Early Cretaceous volcanic arc that was located along the Pacific margin of southern South America. It was separated from the continent by a marginal basin floored by portions of an ophiolite sequence (the Rocas Verdes ophiolites). The transition between the arc and marginal basin occurs on Peninsula Hardy, southernmost Chile, where there is a lateral facies transition from arc deposits of the Hardy Formation into proximal marginal basin fill of the Yahgan Formation. Interfingering of arc and marginal basin sequences demonstrates that subduction-related arc magmatism was concurrent with marginal basin formation. The lateral facies transition is reflected in the geochemistry of volcanic rocks from the Hardy and Yahgan formations. Basalts, andesites and dacites of the arc sequence follow a calc-alkaline differentiation trend whereas basalts from the marginal basin follow a tholeiitic differentiation trend. Estimates of temperature and oxygen fugacity for crystallization of the arc andesites are similar to values reported for other calc-alkaline andesites. It is suggested that water activity influenced the early or late crystallization of Ti-magnetite and this controlled the style of differentiation of the magmas erupted on Peninsula Hardy. Magmas with high water contents evolved along the calc-alkaline differentiation trend whereas those with low water contents evolved along the tholeiitic differentiation trend. Some rhyolites are differentiated from the calc-alkaline andesites and dacites, but most appear to be the products of crustal anatexis on the basis of trace-element evidence. The arc basalts and some marginal basin basalts show relative enrichment in LILE, relative depletion in HFSE, and enrichment in LREE. Other marginal basin basalts are LREE depleted and show small relative depletions in HFSE. Basalts with both calc-alkaline and tholeiitic affinities can also be recognized in the Rocas Verdes ophiolites, where their eruption is correlated with temporal changes in the evolution of the marginal basin. Geochemical data are consistent with derivation of all basalts from a mantle source that is a mixture of MORB-like mantle and variable proportions of a subduction-related component. Basalts with calc-alkaline affinities were erupted initially in the marginal basin, when the locus of back-arc magmatism was near the active arc and the upper mantle source region was strongly influenced by components released from the subducted slab. MORB-like tholeiitic basalts were erupted at a later stage in the wider parts of the basin, when the angle of subduction had increased, the trench had rolled back, and the basin had widened. These MORB-like tholeiitic magmas reflect partial melting of mantle dominated by upwelling asthenosphere containing only a minor component, if any, derived from the subducted slab.