Protracted late Neoproterozic – early Palaeozoic deformation and cooling history of Sør Rondane, East Antarctica, from 40Ar/39Ar and U–Pb geochronology

Abstract 40Ar/39Ar and U–Pb data from five structural domains constrain the late Neoproterozoic – early Palaeozoic tectonothermal history of the eastern part of the East African–Antarctic Orogen in Sør Rondane. A total of 27 new Ar/Ar ages span 570–474 Ma, roughly corresponding to the age range of three generations of syn- to post-tectonic granitoids. The ages are distinct for the five structural domains. The oldest cooling ages come from the weakly deformed southern part of the SW Terrane of Sør Rondane (SW Terrane S), a sliver of a Tonian island arc, which escaped much of the late Neoproterozoic accretionary deformation. This terrane was intruded by the oldest and largest granitoid complex at c. 640–620 Ma. The oldest Ar/Ar amphibole and biotite ages of 570–524 Ma are from the Main Shear Zone, along the northern margin of the SW Terrane S sliver. It hosts granites of age c. 584–570 Ma strung out along the shear zone. Two younger granitoid phases are recorded in the adjacent four terranes to the west, north and east of the SW Terrane S, and correlate with the younger group of Ar/Ar biotite ages spanning 513–474 Ma. We interpret the magmatic and cooling history of duration > 150 Ma to reflect repeated phases of accretion, magmatism and reactivation, that is, collage-style tectonism, partly pre-dating the incorporation of Sør Rondane into Gondwana. The study area first accreted to the cryptic Valkyrie Craton in Tonian times, was then ‘sandwiched’ between the Kalahari and Indo-Antarctica cratons, and experienced extensional tectonics and elevated heat flux due to lithospheric delamination, which resulted in slow cooling during the Pan-African Orogeny.


Introduction
Sør Rondane in Dronning Maud Land (DML) is located in a key region of East Antarctica and is crucial for understanding the final amalgamation of Gondwana. Two models aim to explain the final collision. Jacobs & Thomas (2004) proposed a Himalayan-style continental collision that resulted in the formation of the late Neoproterozoicearly Palaeozoic East African-Antarctic Orogen (EAAO) (Fig. 1a, c). In the EAAO, remnants of the Mozambique Ocean, that is, the Tonian Oceanic Arc Super Terrane (TOAST; Jacobs et al. 2015), are sandwiched between the Kalahari craton (Namaqua-Natal Belt, Proto-Kalahari, Maud Belt, Nampula Complex) and Indo-Antarctica (Indian cratons, Napier Craton, Lützow-Holm Belt, Rayner Belt, Madagascar, Sri Lanka, Ruker Craton). The eastern margin of the Kalahari craton and the western boundary of the TOAST are delineated by the Forster Magnetic Anomaly (Riedel et al. 2013;Jacobs et al. 2015Jacobs et al. , 2020, whereas the eastern boundary of the TOAST is transitional in character and is located east of the Yamato Mountains with its continuation found in Sri Lanka and Madagaskar (Boger et al. 2015;Ruppel et al., 2018). An alternative model interprets the Sør Rondane region as part of the E-W-trending Kuunga Orogen (e.g. Meert, 2003), in which large parts of DML, including Sør Rondane, constitute a mega-nappe rooting in northern Mozambique (e.g. Meert, 2003;Grantham et al. 2013) (Fig. 1b).
Recent geological and geophysical studies have provided insights into the Meso-to early Neoproterozoic history of the region Elburg et al. 2015;Jacobs et al. 2015;Ruppel et al. 2018), the intrusive activity during Gondwana assembly (Elburg et al. 2016), the structural and tectonic framework Owada et al. 2013;Ruppel et al. 2015) and the Gondwana disintegration (Krohne, 2017). To better understand the final assembly of Gondwana during the late Neoproterozoicearly Palaeozoic, we investigated the late-tectonic igneous and cooling history of a broad suite of samples from the Sør Rondane region to trace the geodynamic evolution during and after the formation of the EAAO across DML. In addition, we aim to test the hypothesis of collage-style accretion tectonics of the TOAST Elburg et al. 2016). We present Ar/Ar ages from five distinct terranes of Sør Rondane, including the little-studied region between Sør Rondane and central DML (western SR), and the remote eastern areas of Sør Rondane (Fig. 2). In addition, new U-Pb zircon ages from syn-to post-tectonic granitoids constrain the timing of deformation along the Main Shear Zone (MSZ), the boundary between the northern and southern parts of the SW Terrane of Sør Rondane (SW Terrane N and SW Terrane S, respectively).

Geological background and previous Rb-Sr, K-Ar and Ar/Ar thermochronology of Sør Rondane
Sør Rondane's late Mesoproterozoicearly Neoproterozoic (c. 1100-900 Ma) juvenile basement rocks lack significant remnants of Archaean or Palaeoproterozoic continental crust. Since late Mesoproterozoic time, the region has undergone oceanic arc and/or terrane accretion, probably related to tectonics outboard of Rodinia (e.g. Shiraishi et al. 2008;Jacobs et al. 2015;Elburg et al. 2016). The spatial extent of this so-called Tonian Oceanic Arc Super Terrane (TOAST) was estimated as c. 500 000 km 2 , including the area to the south and SE of central DML and Sør Rondane; it likely continues to the Belgica Mountains ( Fig. 1; Takigami & Funaki, 1991;Ruppel et al. 2018). The TOAST experienced medium-to high-grade metamorphism from c. 650 to 580 Ma, resulting from ocean closure and the final amalgamation of Gondwana. There is only minor evidence for Tonian metamorphism in the rocks of the TOAST (Jacobs et al. 2017).
The SW Terrane S and a few scattered nunataks to the west represent the type locality of the TOAST ( Fig. 2; Jacobs et al. 2017). It consists mainly of a gabbro-tonalite-trondhjemitegranodiorite complex (GTTG), dated at c. 1000-900 Ma; Nd whole-rock and Hf-in-zircon isotopic signatures indicate that this complex represents a juvenile oceanic arc Elburg et al. 2015). The Main Shear Zone delimits the SW Terrane S from the SW Terrane N (Fig. 2b); the latter is dominated by greenschist-to granulite-facies metasedimentary and metaigneous rocks, with the youngest detrital zircon ages of c. 700 Ma (Otsuji et al. 2013;Owada et al. 2013). Peak metamorphic conditions, part of an anticlockwise pressure-temperature (P-T) path, were dated at c. 640-600 Ma Osanai et al. 2013).
To the NE, the SW Terrane is separated from the amphiboliteto granulite-facies metamorphic NE Terrane by the Main Tectonic Boundary (MTB) (Fig. 2b;Osanai et al. 2013). The NE Terrane comprises metasupracrustal rocks with the youngest detrital zircon ages at c. 750 Ma; metamorphism followed a clockwise P-T path and peaked at c. 640-600 Ma . The two contrasting P-T paths on either side of the MTB were interpreted to result from thrusting of the NE Terrane over the SW Terrane during the Pan-African Orogeny Osanai et al. 2013). The former Sør Rondane Suture is interpreted here as part of the frontal thrust zone of the MTB rather than a suture zone between the NE and SW Terranes. So far, pre-750 Ma basement has not been recognized in the NE Terrane; it therefore remains unclear whether the NE Terrane is part of the TOAST or is a separate tectonic domain.
Intrusive rocks record at least four thermal pulses between c. 650 and 500 Ma in Sør Rondane, suggesting protracted accretion and collision in this part of the EAAO (e.g. Elburg et al. 2016). One particular igneous group is represented by granites of the Vengen type (Vengen: locality name) of age c. 560 Ma. They intruded along the MSZ and are weakly foliated, indicating a late syn-to posttectonic emplacement Ruppel et al. 2015;Elburg et al. 2016). Metamorphic U-Pb zircon ages of c. 590-530 Ma were reported across the entire region and interpreted as recording post-collisional retrograde metamorphism and/or hydration after peak amphibolite-facies metamorphism at c. 590-570 Ma Osanai et al. 2013;Elburg et al. 2016).
Previous Rb-Sr, K-Ar and Ar/Ar thermochronology data span c. 660-420 Ma  and references therein). The distribution of K-Ar and Ar/Ar data shows an age trend across Sør Rondane, generally becoming younger from south to north. The oldest K-Ar cooling age of c. 660 Ma is from a gneiss from Nils Larsenfjellet in the SW Terrane S. Most K-Ar and Ar/Ar data scatter between 500 and 480 Ma and are from granitic and syenitic rocks at Lunkeryggen and a gneiss at Otto Borchgrevinkfjellet. The SW Terrane N provided one K-Ar biotite age of c. 500 Ma at Pingvinane (Takigami & Funaki, 1991). Rb-Sr cooling ages of c. 510-460 Ma are from granites and a gneiss collected from the central Sør Rondane corridor (CSRC, Fig. 2) (Deutsch et al. 1961;. The NE Terrane yielded the youngest cooling ages: K-Ar biotite ages of c. 470 Ma at Austkampane; and K-Ar whole-rock ages of 430-420 Ma and K-Ar and Ar/Ar biotite ages of 490-440 Ma at Utnibba (Takigami et al. 1987). Osanai et al. (2013) interpreted the majority of the Rb-Sr, K-Ar and Ar/Ar ages within a range of c. 490-420 Ma as cooling ages, post-dating metamorphism.  suggested a refinement of the tectonic subdivision of Sør Rondane by integrating new aeromagnetic data with the geological framework. The general subdivision into the NE and SW Terranes remained, whereas major lineaments were confirmed or redefined (e.g. MSZ, MTB, Sør Rondane Suture). The magnetic anomaly pattern prompted a subdivision of the SW Terrane S into a western, more tholeiitic, and an eastern, more calc-alkaline, portion of the GTTG complex . Furthermore, the CSRC, which is also characterized by more enriched Hf-isotopic signatures (Elburg et al. 2016), was introduced as a distinct tectonic domain, separating Sør Rondane in an eastern and western part (Fig. 2b). The CSRC was interpreted as being related to late Pan-African extensional tectonics .

3.a. Samples
In total, 28 new samples were analysed, including 25 samples for Ar/Ar (biotite and amphible) and 3 for U-Pb zircon dating (see Section 4.b below). Samples were collected during the 'Geodynamic Evolution of East Antarctica' (GEA) expeditions I and II in the austral summers of 2010/2011 and 2011/2012. They represent all major terranes and/or units previously identified by integrated geological-aerogeophysical studies. From SW to NE, our sampling for Ar/Ar dating included the following (Fig. 3).

638
AS Ruppel et al.

3.b. 40 Ar/ 39 Ar analyses
Ar/Ar analyses were carried out at the Argonlab Freiberg, TU Bergakademie Freiberg, Germany. Mineral processing used the electical-discharge SELFRAG® facility that yields grain sizes corresponding closely to the true crystal size; it also minimizes the proportion of broken grains . We used the largest grain-size fraction from which it was possible to hand-pick optically inclusion-free crystals. The mineral separates were washed repeatedly in deionized water using an ultrasonic bath. After drying, the mineral separates were wrapped in Al foil and loaded into wells on Al discs (33 mm diameter) for irradiation. Irradiation was carried out for 3.8 hours without Cd-shielding at the LVR-15 research reactor of the Nuclear Research Institute in Řež, Czech Republic. The thermal neutron fluence was in the order of 6 × 10 13 neutrons per (cm 2 s) at a thermal to fast neutron ratio of c. 1.1 (c. 7.6 MW reactor power; for detailed irradiation geometry and reactor specifics see Rutte et al. 2015). Irradiated micas were unwrapped and loaded into 3 × 1 mm (diameter × depth) wells on oxygen-free copper discs (30 samples per disc). Hornblende separates were loaded into Mo-crucibles for furnace degassing.
Stepwise heating of micas was performed using a powercontrolled, floating, 30 W New Wave CO 2 laser system, with a defocused beam with a 3 mm diameter. The heating time was 5 minutes per step. Stepwise heating of hornblende separates was performed using a Createc high-temperature cell (HTC; for details see Pfänder et al. 2014), with a heating time of 10 minutes per step. Two GP50 getter pumps, one at room temperature and one at 400°C, achieved gas purification. Cleaning time per step was 5 minutes for laser degassing and 10 minutes for furnace degassing. Argon isotope compositions were measured in static mode on a GV Instruments ARGUS noble gas mass spectrometer equipped with five Faraday cups, and 10 12 Ohm resistors on mass positions 36-39 and a 10 11 Ohm resistor on mass position 40. Typical blank levels were 2.5 × 10 −16 mol 40 Ar and 8.1 × 10 −18 mol 36 Ar. Measurement time was 7.5 minutes per step, acquiring 45 scans at 10 seconds integration time each. Mass bias was corrected assuming linear mass-dependent fractionation and using an atmospheric 40 Ar/ 36 Ar ratio of 298.6 ± 0.3 (Lee et al. 2006;Mark et al. 2011). For raw data reduction and time-zero intercept calculation, we used a Matlab® toolbox developed in-house. Inverse isochron and weighted mean ages were calculated using ISOPLOT 4.15 (Ludwig, 2008). All ages were calculated using Drachenfels sanidine (DRF1) as fluence monitor (25.682 ± 0.030 Ma). This in-house standard was calibrated against Fish Canyon Tuff sanidine, applying an age of 28.305 ± 0.036 Ma and using the decay constants given in Renne et al. (2010). Interference correction factors are given in online Supplementary Table S1 (available at http://journals.cambridge.org/geo) along with the raw data of the Ar isotope measurements. Ar/Ar age spectra and inverse isochron diagrams of all analysed samples are provided in online Supplementary Figure S1.

3.c. U-Pb zircon analyses
Zircon concentrates were separated from 1-2 kg sample material using standard methods. The final selection of the zircon grains for U-Pb dating was achieved by hand-picking under a binocular microscope. Zircon grains of all sizes and morphological types were selected, mounted in resin, and polished to approximately half their diameters. Zircons were analysed for U, Th and Pb isotopes by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) using a Thermo-Scientific Element 2 XR sector field ICP-MS coupled to a NewWave UP-193 Excimer Laser System at the Senckenberg Naturhistorische Sammlungen Dresden, Germany. Cathodoluminescence imagery guided the selection of the analysed spots and was used to study the zircon textures, aiding the interpretation of their geological history (e.g. Corfu et al. 2003). A teardrop-shaped, low-volume laser cell constructed by Ben Jähne (Dresden) and Axel Gerdes (Frankfurt am Main) was used to enable sequential sampling of heterogeneous grains (e.g. growth zones) during time-resolved data acquisition. Each analysis consisted of approximately 15 seconds background acquisition followed by 30 seconds of data acquisition, using a laser spot size of 25 or 35 μm. A common-Pb correction, based on the interference-corrected and background-corrected 204 Pb signal and a model Pb composition (Stacey & Kramers, 1975), was applied if necessary (i.e. if the corrected 207 Pb/ 206 Pb lay outside of the internal errors of the measured ratios). Raw data were corrected for background signal, common Pb, laser-induced elemental fractionation, instrumental mass discrimination and time-dependent elemental fractionation of Pb/Th and Pb/U using an Excel® spreadsheet program developed by Axel Gerdes. Reported uncertainties were propagated by quadratic addition of the external reproducibility obtained from the standard zircon GJ-1 (c. 0.6 and 0.5-1.0% for 207 Pb/ 206 Pb and 206 Pb/ 238 U, respectively) during individual analytical sessions, and the within-run precision of each analysis. Concordia diagrams (2σ error ellipses) and concordia and weighted mean ages (95% confidence level) were calculated using ISOPLOT 4.15 (Ludwig, 2012). To ensure reproducibility of the reference material, we report all calculated 2σ crystallization-age uncertainties < 2% as 2% in parentheses. For further details on the analytical protocol and data processing, see Gerdes & Zeh (2006) and Frei & Gerdes (2009). Cathodoluminescence images are provided in online Supplementary Figure S2 (available at http://journals.cambridge.org/geo) and the complete LA-ICP-MS zircon data are presented in online Supplementary Table S2.

4.a. Ar/Ar dating of biotite and amphibole
The results are presented and described in regional order from SW to NE and grouped according to their geographic position within the distinct terranes (online Supplementary Fig. S1, available at http://journals.cambridge.org/geo). Table 1 summarizes the analytical results. In many samples, most of the steps are highly radiogenic (% 40 Ar* > 99); most data points therefore cluster close to the 39 Ar/ 40 Ar axis in the inverse isochron plots and provide poor fits to the 36 Ar/ 40 Ar versus 39 Ar/ 40 Ar data. Most biotite separates yielded flat sections in the age versus cumulative 39 Ar diagrams, with weighted mean ages (WMA) with mean square of weighted deviates (MSWD) around 1 and probabilities of fit > 5%; however, the amphibole spectra are commonly U-shaped and do not permit the calculation of a WMA for the majority of analyses. Our preferred age interpretation includes age uncertainties of ±2σ and ≥1% of the age to account for sample complexities, in particular, the uncertainities in the calculated initial 40 Ar/ 36 Ar.  age of the pegmatite . A total of 87.4% of the released 39 Ar yielded an Ar/Ar biotite WMA of 501.8 ± 1.3 Ma (MSWD = 1.2) and an IIA of 499.3 ± 2.7 Ma (MSWD = 0.71), the latter with an imprecise, non-atmospheric initial 40 Ar/ 36 Ar value of 943 ± 290. Our age interpretation is 500 ± 5 Ma.
4.a.1.4. J1213D, biotite, pegmatitic granitoid, Van Autenboerfjellet (72.362983°S; 20.244717°E). This pegmatite is from a coarse-grained granite sheet intruding into migmatic gneisses. The sample consists of quartz, plagioclase, K-feldspar and biotite. Zircons gave a 206 Pb/ 238 U concordia age of 582 ± 5 Ma (MSWD = 2.25), interpreted as the crystallization age . A total of 91.5% of the released 39 Ar yielded a biotite WMA of 513.0 ± 1.0 Ma (MSWD = 0.74); this sample did not allow an IIA calculation. The hump in the central part of the spectrum indicates that these steps incorporated minor extraneous Ar. Our preferred age is 513 ± 5 Ma.
4.a.1.6. Sample J1221D_1, biotite, grey gneiss, moraine sample (72.33105°S; 23.25415°E). This gneiss from a moraine between Swaabsteinen and Causinknappen (Figs 2, 3) consists mainly of quartz, plagioclase, microcline, myrmekite and biotite. Quartz shows undulose extinction and subgrain-rotation and grainboundary-migration recrystallization. Zircon cores provided a weighted mean 206 Pb/ 238 U age of 1087 ± 28 Ma (MSWD = 2.5), while metamorphic rims yielded a weighted mean 206 Pb/ 238 age of 542 ± 5 Ma (MSWD = 2.8) (Jacobs et al. 2017). Dark brown, columnar biotite shows no preferred orientation. A total of 80.4% of the released 39 Ar yielded a biotite WMA of 505.3 ± 1.5 Ma (MSWD = 0.86) and a IIA at 501.8 ± 4.7 Ma (MSWD = 2.3) with a 40 Ar/ 36 Ar value of 904 ± 440; this is consistent with the hump in the central part of the spectrum, which indicates that these steps incorporated extraneous Ar. Our preferred age is 503 ± 10 Ma.  11°S; 23.1987°E). This banded gneiss, composed of layers of recrystallized feldspar and quartz, was sampled at the western side of Ketelersbreen. Quartz shows both subgrain-rotation and grain-boundary-migration recrystallization. Plagioclase, K-feldspar and hornblende form porphyroclasts with asymmetric pressure-shadow tails of recrystallized quartz, hornblende and biotite. Dark brown biotite and greenish-blue hornblende occur within the foliation. Epidote and opaque minerals are accessories. A total of 56.3% of the released 39 Ar yielded a biotite WMA of 534.6 ± 3.3 Ma; the high MSWD of 4.1 reflects the hump-shape of the spectrum. The IIA of 523.2 ± 4.1 Ma (MSWD = 1.07) is younger with an initial 40 Ar/ 36 Ar value of 1145 ± 230. Our preferred age interpretation of 524 ± 10 Ma emphasizes the youngest high-temperature steps and follows the IIA. A total of 96.4% of the released 39 Ar yielded an amphibole WMA of 561.5 ± 2.8 Ma (MSWD = 0.25). This date is based on only two steps; the IIA of all five steps (the small amount of material did not allow more steps to be measured) at 560.2 ± 4.1 Ma (MSWD = 1.6; 40 Ar/ 36 Ar = 8349 ± 5500) may indicate that this date has geological meaning. Our preferred age is 561 ± 20 Ma.
Quartz occurs in bands as coarse-grained porphyroclasts and small recrystallized grains. Plagioclase porphyroclasts showing bulging recrystallization and strong sericitization are partly replaced by quartz, and record dextral shear sense by asymmetric, biotite-rich pressure-shadow tails. Biotite, chlorite and muscovite form the foliation. The mineral association, with quartz and plagioclase, and minor amounts of chlorite, biotite, muscovite and amphibole, suggests that this schist formed due to localized shearing of a tonalitic host rock under retrograde, greenschist-facies conditions. A total of 78.3% of the released 39 Ar yielded a biotite WMA of 526.0 ± 1.3 Ma (MSWD = 0.77) and an IIA of 522 ± 6.3 (MSWD = 1.5). Again, the high radiogenic yield results in an ill-defined initial 40 Ar/ 36 Ar value of 900 ± 470; the hump-shaped spectrum reflects the extraneous Ar in this sample. Our preferred age is 525 ± 5 Ma.

642
AS Ruppel et al. ). This mylonitic sample from the Vengen/Vikinghøgda granite was taken close to the MSZ. It contains quartz, plagioclase, and K-feldspar as major, and biotite, muscovite, chlorite, titanite and magnetite as minor components (Fig. 4c). Asymmetric plagioclase and K-feldspar porphyroclasts, indicating a dextral sense of shear, are surrounded by recrystallized plagioclase and bands of elongated quartz (Fig. 5b). Dark brown biotite occurs along the foliation and in the pressure shadows of plagioclase porphyroclasts. The Vengen/Vikinghøgda granitoids yielded crystallization ages of 562 ± 7 Ma ) and 551 ± 8 Ma (Elburg et al. 2016 (Fig. 7a) and is mainly composed of quartz, K-feldspar and plagioclase, with minor amounts of titanite and secondary scapolite and epidote. Quartz shows grain-boundary-migration recrystallization. K-feldspar is recrystallized and contains deformation twins. The large plagioclase grains show sericitization. The foliation is co-linear with the mylonitic foliation of the MSZ.
The sample contains small (up to 150 μm), mainly prismatic, euhedral zircons (online Supplementary Fig. S2a). A total of 106 spots were measured with 94 concordant analyses; the mostly younger, discordant analyses imply lead loss from c. 580 Ma (Fig. 6a). The 95-105% concordant spots yieded a population of normal distributed ages for which we obtained an age of 570 þ 5/ − 2 Ma (coherent group of 38 out of 63 analyses) with the TuffZirc algorithm of Ludwig & Mundil (2002), which minimizes both positive (from xenocrysts or cores) and negative (from Pb loss) age biases (Fig. 6b). Most of the 38 analyses were conducted on oscillatory zoned zircon cores with sometimes weakly luminescent rims (online Supplementary Fig. S2a). These spots yielded a concordia age of 570.0 ± 1.8 (12) Ma (Fig. 6c). The Th/U ratios span 0.19-0.80 with a median of 0.37 typical for igneous origin. The concordia age of 570 ± 12 Ma is interpreted as the crystallization age of the monzogranitic dyke.
4.b.2. Sample 1217A1, syenite, Lunkeryggen E (72.03301°S; 24.59775°E) This syenite is from the eastern flank of Lunkeryggen along the easternmost part of the known extent of the MSZ (Fig. 7b). The most common mineral -K-feldsparshows oscillatory Deformation and cooling history of Sør Rondane Fig. 6. (Colour online) U-Pb zircon data. Sample 1205A5: (a) Terra-Wasserburg diagram for all data; (b) age distribution of the 95-105% concordant ages with age range defined by the TuffZirc algorithm (in red); and (c) concordia age calculated from the red data in (b). Sample 1217A1: (d) Terra-Wasserburg diagram for all data; (e) age distribution of 95-105% concordant dates and age range defined by the TuffZirc algorithm (in red); the two dates marked in blue may be geologically meaningless or may represent late-stage metamorphic-hydrothermal zircon grains; (f) concordia age calculated from the red data in (e). Sample 1219B1: (g) Terra-Wasserburg diagrams for ages with a level of concordance of 90-110%; upper and lower intercepts through the main age clusters; (h, i) concordia ages for the age groups > 900 Ma; (j) age range defined by the TuffZirc algorithm (in red); the four dates marked in blue may be geologically meaningful, delineating inherited zircons; (k, l) concordia ages for the age groups < 600 Ma.

646
AS Ruppel et al.
The zircons dated here are mainly subhedral and round; some are anhedral or occur as fragments. All zircons are < 200 μm in size. Twenty-nine analyses on 25 crystals were performed (Fig. 6d). Analyses younger than c. 570 Ma are mostly discordant as a result of Pb loss. The mostly discordant young grains have low Th/U ratios (0.03-0.11) and are dark and featureless in cathodoluminescence (online Supplementary Fig. S2b). The 95-105% concordant spots yielded a coherent group of 11 dates (out of 16) at 566 þ 13/ − 2 Ma with the TuffZirc algorithm (Fig. 6e). These 11 spots yielded a concordia age of 570.6 ± 3.6 (12) Ma (Fig. 6f). They comprise zircons with partly blured oscillatory and sector zoning. Their Th/U ratios span 0.2-0.7, typical for igneous zircons. Within the 95-105% concordant spots a younger population of ages with two concordant (99 and 100%) outliers yielded a WMA of 514 ± 13 Ma and a Th/U ratio of 0.06. We interpret the concordia age of 571 ± 12 Ma as the syenite crystallization age. The c. 514 Ma grains may represent a metamorphic-hydrothermal overprint.  This dyke intruded a grey gneiss and has a mylonitic texture that is parallel to the MSZ (Fig. 7c). The main constituents are quartz, K-feldspar and plagioclase. Quartz occurs in foliation-parallel layers and shows grain-boundary-migration recrystallization. Feldspar forms porphyroclasts with asymmetric pressure shadows of muscovite and minor biotite. Chlorite, illmenite, magnetite and titanite are accessories. Zircons are subhedral, mainly long prismatic, sometimes round, and reach up to 200 μm in length. Under CL, many zircons show bright and oscillatorily zoned cores surrounded by dark rims, which are either oscillatory zoned or structureless (online Supplementary Fig. S2c). We measured 119 spots that yielded Tonian and Ediacaran age groups (Fig. 6g). The youngest dates generally have higher U contents, probably indicating Pb loss due to metamictization. Two subclusters of the Tonian 90-110% concordant dates yielded concordia ages of 987.2 ± 9.1 (20) and 968.1 ± 7.7 (20) Ma; the respective Th/U ratios are 0.14-0.36 and 0.14-0.59, typical for igneous zircons (Fig. 6h, i).
We interpret the c. 585 Ma age as the crystallization age of the dyke; the c. 987 and 648 Ma zircon grains as likely inherited, and the c. 553 Ma age may represent a metamorphic-hydrothermal overprint. We discard the c. 968 Ma age due to the high MSWD of concordance (7.1), which implies that this population also comprises normal discordant grains due to the Ediacaran overprint.

4.c. Cooling rates
Detailed cooling rate determinations are premature as too few thermochronometers have been employed so far. To obtain first-order cooling-rate estimates, we took the following approach. First, the new and published U-Pb zircon ages from our samples are from late-stage, small volume intrusions, that is, pegmatitic granitoids intruding as dykes and sheets. These rocks likely crystallized at relatively low temperatures; we therefore assumed a 650 ± 25°C crystallization temperature. We then recognized that the Ar/Ar ages are generally much younger than the zircon ages, implying slow cooling. We first assigned classical closure temperatures to the biotite (c. 300°C) and amphibole (c. 525°C) dates and calculated cooling rates from the zircon, amphibole and biotite data for several samples and sample groups. The obtained rates were low, scattering around 10°C Ma -1 . We then recalculated the closure temperatures using the mean of the grain-size range as the effective dimension of the diffusion domain and a range of cooling rates (5-30°C Ma -1 ) with the CLOSURE program of Brandon et al. (1998). This program applies the parameters given in Grove & Harrison (1996; for biotite) and Harrison (1981; for amphibole). Next, we calculated the final closure temperatures using a cooling rate of 10°C Ma -1 and assigned uncertainites of ±15°C, the latter approximating the possible range of values in Sør Rondane (5-30°C Ma -1 ). Finally, we calculated error-weighted regressions through dates from single samples and sample groups.
A few first-order features can be derived from these calculations (Fig. 8). The cooling rates through c. 700-250°C are low, around 5-10°C Ma -1 in western Sør Rondane and the northern rim of the SW Terrane S along the MSZ where the database allows calculations; they are probably also low in the NE Terrane. No cooling rate calculations are possible for the CSRC. The biotite ages of the SW Terrane N are distinctly younger than those of the SW Terrane S (with one exception); this may imply an even lower cooling rate for this terrane. The cooling path for the SW Terrane S along the MSZ passes through the P-T space of retrograde metamorphic stage M3S (c. 400-500°C at 530-590 Ma; Osanai et al. 2013). The cooling path touches the P-T space of retrograde metamorphic stage M3N (c. 400-550°C at 530-590 Ma; Osanai et al. 2013) in the NE Terrane.

4.d. Summary of new Ar/Ar mineral and U-Pb zircon dating
Twenty-two new Ar/Ar biotite and five new amphibole Ar/Ar ages from five different structural domains span c. 570-470 Ma and fall into two main age groups. The older Ar/Ar age group (c. 570-525 Ma) is from the MSZ along the northern rim of the SW Terrane S. The younger age group (c. 515-470 Ma) is widespread across large parts of the study area but particulary in the SW Terrane N, the NE Terrane and the CSRC. U-Pb zircon data of three granitoids give igneous crystallization ages of c. 585, 571 and 570 Ma, with one sample recording inheritance at c. 987 and 648 Ma. Two samples record a metamorphic and/or hydrothermal overprint at c. 553 and probably 514 Ma. Together with previously published dates for western Sør Rondane , our samples span a crystallization age range of 597-528 Ma. Slow cooling rates of 5-10°C Ma -1 characterize Sør Rondane.

Discussion
Sør Rondane is mainly underlain by rocks of the TOAST that include GTTG granitoids and their supracrustal cover. There is little evidence for metamorphic overprint before the main granulite-facies metamorphism at c. 650-600 Ma. Prior to this high-grade metamorphic overprint, the Tonian oceanic arcs were probably an integral part of the Mozambique Ocean, or had undergone accretion without records of high-grade metamorphism (e.g. Jacobs et al. 2015). Granulite-facies metamorphism in the NE Terrane at c. 650-600 Ma is thought to relate to overthrusting of the NE Terrane over the SW Terrane along the MTB; this is interpreted from the contrasting P-T-t paths on either side of the MTB , although the MTB itself is largely unexposed. After granulite-facies metamorphism, Sør Rondane underwent amphibolite-to greenschist-facies tectonothermal reworking at c. 590-530 Ma (e.g. Shiraishi et al. 2008;Osanai et al. 2013). With the exception of the MSZ, large parts of the SW Terrane S were only weakly affected by this event (online Supplementary Fig. S3).
Sør Rondane shows pulses of syn-to late-tectonic magmatism at > 600, 580-550, c. 530 and 510-500 Ma ('phase one to four'); some of the pulses are unique to one or several of the five distinct tectonic domains (e.g. Shiraishi et al. 2008;Elburg et al. 2016) (Fig. 9). The interior of the SW Terrane S is only affected by the early, 'phase one' granitoid magmatism, represented by the 648 AS Ruppel et al. voluminous arc-type Dufek granitoid dated at c. 640-620 Ma (Li et al. 2006;Elburg et al. 2016). The SW Terrane S appears as a rheological stiff tectonic sliver that is surrounded by tectonic domains with significant proportions of rheological weaker metasedimentary rocks; there, 'phase two to four' granitoids occur in significant quantities (Fig. 9). On the one hand, the spatial distribution of our new, as well as the published, Ar/Ar and K-Ar mineral ages appear to reflect the contrasting rheology of the different tectonic domains; on the other hand, they appear to correlate with the spatial distribution of the syn-to late-tectonic granitoid magmatism (Figs 9, S3).

5.a. Cooling history of Sør Rondane
Most Ar/Ar mineral ages of our study post-date both the main granulite-facies metamorphic event at c. 650-600 Ma and, to a large extent, also the later amphibolite-to greenschist facies metamorphism at c. 590-550 Ma. We interpret the bulk of our Ar/Ar ages to document cooling to below the mineral-specific closure temperatures of c. 520-500°C for amphibole of c. 160 μm size and c. 340-310°C for biotite of 250 μm size and a cooling rate of c. 10°C Ma -1 , typical for Sør Rondane (see Section 4.c above). Our two main Ar/Ar ages groups of c. 570-525 and 515-475 Ma show a clear spatial relation to the five different tectonic domains and the regional distribution of the four phases of granitoids in the region. The older Ar/Ar age group of c. 570-525 Ma is related to the highly sheared northern contact of the SW Terrane S and the MSZ, the tectonic domain where in a previous study a K-Ar mineral age of c. 660 Ma was reported (Takigami & Funaki, 1991). The northern margin of the SW Terrane S records significant volumes of 'phase two' granitoid magmatism (c. 580-550 Ma), but lacks the two younger granitoid phases. Our new U-Pb zircon crystallization ages from two deformed monzogranitic dykes and one syenite from the MSZ span c. 585-570 Ma and correspond to the 'phase two' granitoid magmatism. The late syn-to posttectonic nature of these intrusions and the metamorphic and/or hydrothermal zircons as young as c. 550 Ma imply that the major deformation in the central part of the MSZ was older than c. 550 Ma. The rather weak NE-striking foliation of the SW Terrane S that bends into the approximately E-striking MSZ therefore probably pre-dates the 'phase 2' granitoid magmatism (Fig. 9). The new Ar/Ar amphibole (570-550 Ma) and biotite ages (547-524 Ma) of this tectonic domain indicate cooling to below the mineral specific closure temperature after granitoid magmatism, and are therefore interpreted as dating cooling during and after late-stage ductile shearing along the MSZ.
The SW Terrane S is surrounded by four distinct tectonic domains to the north, east and SW, each with contrasting orientations of the main foliation; the orientations differ from those of the SW Terrane S (Fig. 9). All these domains record significantly younger Ar/Ar cooling ages than the SW Terrane S and are dominated by significant volumes of the two later phases of granitoid magmatism at c. 530 and 510-500 Ma. 'Phase three' granitoids crop out frequently as small-to moderate-sized bodies throughout these domains; 'phase four' magmatism has so far only been documented in the NE Terrane and west of our study area at Steingarden (Fig. 2; Jacobs et al. 2015;Elburg et al. 2016). In western Sør Rondane, which is part of the magnetically distinct SE-DML province, the foliation is SE-striking, following the main trend of the dominant magnetic anomalies . Our five new Ar/Ar biotite ages from this region span 513-500 Ma (Fig. 9), consistent with cooling after the intrusion of 'phase three' granitoids at c. 530 Ma. A sixth sample, with an Ar/Ar biotite age of c. 503 Ma, was sampled from a moraine (J1221D-1, Jacobs et al. 2017), indicating that similar young cooling ages may occur further inland.
The SW Terrane N has an overall E-striking foliation, co-linear with the MSZ (Fig. 9). This region records amphibolite-to greenschist-facies metamorphism at 590-530 Ma, which is also documented in the CSRC and in the NE Terrane ). It has a major, 'phase three' granitoid at Utsteinen, dated at c. 528 Ma (Elburg et al. 2016);. Three of our Ar/Ar biotite cooling ages derive from this domain and span 490-493Ma, clearly post-dating the 'phase three' granitoid magmatism and possibly indicating post-intrusion cooling in the SW Terrane N.
The NE Terrane is dominated by a mainly E-striking foliation and records large-scale folding, which is related to a compressional phase at c. 610 Ma, followed by a transpressional phase and associated dextral shearing . Further, the Balchen extensional detachment (Fig. 9) possibly indicates largescale extension between 600 and 550 Ma . The NE Terrane exposes only small granitoids that are mostly undated. One granitic dyke at Balchenfjella crystallized at c. 549 Ma online Supplementary Fig. S3d, available at http://journals.cambridge.org/geo); at Vesthaugen, an undeformed granitoid yielded a crystallization age of c. 509 Ma (Elburg et al. 2016). Our two new Ar/Ar cooling ages of c. 488 and 476 Ma from the eastern part of the NE Terrane are the youngest ages of this study (Fig. 9). These young ages may either account for the thermal effect of the so-far unobserved 'phase four' granitoids in the region or record cooling related to extension after metamorphism in the NE Terrane (Fig. 8c).
The CSRC to the east of both the SW Terrane S and SW Terrane N (east of the Gjelbreen Lineament) has highly variable structural trends and shows large-scale folding .
Magnetic anomaly data indicate that the CSRC may disrupt the MTB into western and eastern parts ( Fig. 9; . The CSRS has therefore been interpreted as a domain of major late-tectonic extension ). Two small granitoid bodies within the CSRC crystallized at c. 527 and 532 Ma (Elburg et al. 2016). Our four Ar/Ar biotite data scatter widely over 474-539Ma. The oldest Ar/Ar age pre-dates 'phase three' granitoid magmatism in the CSRC, whereas the remaining ages mainly post-date 'phase four' granitoid magmatism. The widely scattered Ar/Ar ages may be best explained by the complex extensional character of this domain, in which different tectonic levels were juxtaposed.
Each of the five tectonic domains appear to have unique Ar/Ar mineral cooling age signatures that together span c. 100 Ma. While the earlier cooling history, as recorded in the SW Terrane S, probably relates to collision of the TOAST with the Kalahari craton and associated magmatic activity and deformation along the MSZ, the later cooling history, as recorded in the NE Terrane and specifically in the CSRC, mostly records extension following orogenic thickening.
In summary, the apparent continuous cooling implied by our data (Fig. 8) for west Sør Rondane and the MSZ along the northern margin of the SW Terrane S supports the suggestion of Adachi et al. (2013) that the cooling history is a result of the cooling of the c. 570 Ma intrusions and associated regional amphibolite-to greenschist-facies metamorphism. Extrapolation of our cooling rates also yields intersections with the P-T space of the granulite-facies metamorphism in the NE Terrane; however, this is very speculative . The most conservative interpretation is that our cooling paths describe a part of the exhumation

650
AS Ruppel et al. history of a long-lasting orogeny, including post-contractional erosion and extension. This is supported by the wide age range (597-528 Ma) of the concordant U-Pb zircon dates in our samples that back up the inference that these volumetrically variable and partly regionally confined intrusions appear to have influenced the distinct cooling pattern that we observe in the Sør Rondane region.
5.b. Regional comparison of mineral cooling ages across DML Comparing our new Ar/Ar mineral cooling ages with published dates across DML, we recognize a number of first-order similarities and differences. As in Sør Rondane, most regions in western and central DML also show Ar/Ar mineral cooling ages of < 530 Ma, specifically in regions characterized by extensive volumes of late-tectonic granitoid magmatism (Fig. 10). The latter has previously been related to delamination tectonics, subsequent to orogenic collapse and an overall associated increased heat flow (e.g. Jacobs et al. 2008). Significant older Ar/Ar ages (c. 800-1200 Ma) only occur to the west of the western orogenic front of the EAAO, that is, west of the Heimefront Shear Zone (Jacobs et al. 1995) and in Kirwanveggen (Kleinschmidt et al. 2000;Grantham et al. 2019).
As for the SW Terrane S in Sør Rondane, a few other tectonic domains also appear to have escaped the regional collisional tectonometamorphism at c. 590-500 Ma in central DML. These domains include the c. 600 Ma Gruber anorthosite (Fig. 10), the core of which appears largely undeformed (Jacobs et al. 1998) and which has been interpreted as another mega-boudin or tectonic sliver at the boundary between the eastern margin of Kalahari craton and the TOAST (Bauer et al. 2004). The Schirmacher tectonic domain also shows little evidence of post-600 Ma tectonothermal overprint, and widely lacks post-600 Ma intrusions (e.g. Henjes-Kunst, 2004;Jacobs et al. 2020). It therefore also shows slightly older Ar/Ar cooling ages than the surrounding regions.
In general, Ar/Ar ages of < 580 Ma occur throughout DML and exhibit a heterogeneous age pattern with an overall trend from younger ages in central DML to older ages in its western and eastern parts (Fig. 10). The slightly older Ar/Ar mineral ages along the westernmost EAAO in eastern Heimefrontfjella and in Kirwanveggen coincide with the lack of syn-to post-tectonic granitoid intrusions in the region. In this respect, the western part of the EAAO has similarities with the main part of the SW Terrane S, which similarly lacks syn-to post-tectonic granitoid intrusions and which therefore has comparatively old Ar/Ar ages. As suggested by the data of Adachi et al. (2013) and supported by our data (see Section 5.a), the late-tectonic intrusions likely kept the thermal gradient elevated, perhaps supporting ductile extension and keeping the Ar/Ar system open for a prolonged time. This resulted in slow cooling in some parts of central DML (e.g. Markl & Henjes-Kunst, 2004;Hendriks et al. 2013). The close relationship of late-to post-tectonic granitoids, comparatively young Ar/Ar ages, high thermal gradients and slow cooling has also been noted in a study from northern Mozambique (Ueda et al. 2012), which was positioned adjacent to DML in Gondwana.
The large spread in the Ar/Ar mineral ages and the protracted structural and metamorphic evolution of Sør Rondane and other parts of DML are best explained by a protracted accretionary and collision history that led to the assembly of Gondwana. It is unlikely that a single tectonic mega-event (e.g. Grantham et al. 2008Grantham et al. , 2013 can explain the age pattern that we record in this study. The suggested mega-nappe thrust model with top-to-SW thrusting would require a reverse distribution of cooling ages, that is, predominantly older ages in the NE Terrane, representing a nappe/klippe, and younger ages farther to the south or SW in the SW Terrane. This is in contrast to the age zonation in Sør Rondane (Figs 8, S3c).

Summary and conclusions
Our new Ar/Ar mineral ages from Sør Rondane span c. 570-470 Ma. The data record a protracted cooling history of > 100 Ma that can best be explained by a long-lasting orogeny. The major stages of interlinked tectonic and magmatic processes are as follows.
1. Accretionary tectonics resulted in the formation of the TOAST by the collision of a number of oceanic arc terranes of the Mozambique Ocean, probably outboard the Valkyrie Craton. Two different rheological domains of the TOAST had formed and can now be differentiated: (a) domains predominated by relatively competent igneous rocks (SW Terrane S); and (b) domains dominated by a mixture of rheologically weaker meta-supracrustal cover and arc rocks (remaining areas in the Sør Rondane region). The oldest cooling ages are found in the more competent domains, while the younger age group is found in domains dominated by the meta-supracrustal rocks. 2. Late Neoproterozoicearly Palaeozoic tectonothermal reworking of the TOAST was associated with distinct phases and variable volumes of granitoid intrusions, spanning c. 150 Ma. The distribution of our Ar/Ar mineral cooling ages is closely related to these intrusions. The northern margin of the SW Terrane S, the Main Shear Zone, hosts granitoids with three of them newly dated in this study at 584-570 Ma. The new crystallization ages coincide with the second main magmatic period in Sør Rondane that spanned c. 580-550 Ma; they provide an age estimate for major deformation along the Main Shear Zone, that is, older than c. 550 Ma. Our new Ar/Ar ages (570-524 Ma) record cooling during and after late-stage ductile shearing. This second main magmatic period and deformation along the northern margin of the SW Terrane S may be related to collision of the TOAST with the Kalahari craton. The two younger magmatic events at c. 530 Ma and 510-500 Ma are mainly recorded in domains dominated by a mixture of rheologically weaker meta-supracrustal rocks intruded by granitoids; these domains also have the youngest cooling ages. Ages of < 500 Ma predominantly occur in the NE Terrane and in the central Sør Rondane corridor, which are both dominated by late extensional tectonics. 3. The complex, differential and protracted cooling pattern recognized in the five distinct tectonic domains in Sør Rondane are best explained by a protracted collisional history during the final amalgamation of Gondwana. It is unlikely that the structure and cooling pattern seen in the Sør Rondane region and elsewhere in Dronning Maud Land can be resolved by a single large tectonic event such as a mega-nappe thrust model.