Tectono-Thermal Evolution and Morphodynamics of the Central Dronning Maud Land Mountains, East Antarctica, Based on New Thermochronological Data

The lack of preserved Mesozoic–Cenozoic sediments and structures in central Dronning Maud Land has so far limited our understanding of the post-Pan-African evolution of this important part of East Antarctica. In order to investigate the thermal evolution of the basement rocks and place constraints on landscape evolution, we present new low-temperature thermochronological data from 34 samples. Apatite fission track ages range from 280–85 Ma, while single-grain (U-Th)/He ages from apatite and zircon range from 305–15 and 420–340 Ma, respectively. Our preferred thermal history models suggest late Paleozoic–early Mesozoic peneplanation and subsequent burial by 3–6 km of Beacon sediments. The samples experienced no additional burial in the Jurassic, thus the once voluminous continental flood basalts of western Dronning Maud Land did not reach central Dronning Maud Land. Mesozoic–early Cenozoic cooling of the samples was slow. Contrary to western Dronning Maud Land, central Dronning Maud Land lacks a mid-Cretaceous cooling phase. We therefore suggest that the mid-Cretaceous cooling of western Dronning Maud Land should be attributed to the proximity to the collapse of the orogenic plateau at the Panthalassic margin of Gondwana. Cooling rates accelerated considerably with the onset of glaciation at 34 Ma, due to climate deterioration and glacial denudation of up to 2 km.


Introduction
The Dronning Maud Land Mountains in East Antarctica form an impressive, largely coast-parallel mountain range with a total relief exceeding 5 km, extending for c. 1500 km in length. The highest peaks reach elevations up to c. 3200 m a.s.l., while the deepest incisions are 2 km below sea level. The Dronning Maud Land Mountains originated as a continental margin escarpment following the break-up of Gondwana in Jurassic times. This escarpment is now located c. 200 km inland of the continental margin. Here, we focus on a c. 350 km long section in central Dronning Maud Land from Mühlig-Hofmannfjella in the west to Wohlthatmassivet in the east (Figures 1 and 2). The exposed basement comprises Mesoproterozoic-Early Paleozoic high-grade rocks, documenting a protracted older geodynamic evolution, including the formation and destruction of two supercontinents-Rodinia and Gondwana, e.g., [1,2].  [3], König and Jokat [4], Boger [5], Slater et al. [6] and Stone [7].
Applying low-temperature thermochronology on central Dronning Maud Land basement rocks allows us to trace the tectonic and morphodynamic evolution during pre-, syn-and post-break-up times in a region that lacks marker horizons, such as unconformable overlying sedimentary rocks. Previously published thermochronological data from the area are interpreted to reflect long-lasting monotonic cooling of the basement since the last orogeny in Early Paleozoic times [8][9][10]. Similar datasets from western and eastern Dronning Maud Land have revealed a much more complex thermal evolution of the mountain range, e.g., [11][12][13][14], including Late Paleozoic-Early Mesozoic peneplanation and syn-rift reburial, either due to sedimentary basins or by emplacement of Jurassic continental flood basalts. This apparently contrasting long-term tectonic and morphodynamic evolution is puzzling, and has raised several questions: Was the basement in central Dronning Maud Land also exposed to the surface during the Late Paleozoic-Early Mesozoic? Were the central Dronning Maud Land Mountains covered by Jurassic continental flood basalts or Mesozoic sedimentary basins? What did the central Dronning Maud Land Mountains look like at the onset of the glaciation?
In order to trace the evolution of the central Dronning Maud Land Mountains since the Early Paleozoic, we have revisited the area with modern fission track dating and the addition of new apatite and zircon (U-Th)/He analyses (AHe and ZHe, respectively). Meier [8] and Meier et al. [9] still used non-standard etching techniques for their apatite fission track analyses and also used a somewhat unconventional zeta calibration approach. We have therefore re-analyzed many of Meier's samples to bring them up to the standard required for state-of-the-art thermochronological modelling. Additionally, new samples have been added from Schirmacheroasen (Figures 2 and 3). Based on this  [3], König and Jokat [4], Boger [5], Slater et al. [6] and Stone [7].
Applying low-temperature thermochronology on central Dronning Maud Land basement rocks allows us to trace the tectonic and morphodynamic evolution during pre-, syn-and post-break-up times in a region that lacks marker horizons, such as unconformable overlying sedimentary rocks. Previously published thermochronological data from the area are interpreted to reflect long-lasting monotonic cooling of the basement since the last orogeny in Early Paleozoic times [8][9][10]. Similar datasets from western and eastern Dronning Maud Land have revealed a much more complex thermal evolution of the mountain range, e.g., [11][12][13][14], including Late Paleozoic-Early Mesozoic peneplanation and syn-rift reburial, either due to sedimentary basins or by emplacement of Jurassic continental flood basalts. This apparently contrasting long-term tectonic and morphodynamic evolution is puzzling, and has raised several questions: Was the basement in central Dronning Maud Land also exposed to the surface during the Late Paleozoic-Early Mesozoic? Were the central Dronning Maud Land Mountains covered by Jurassic continental flood basalts or Mesozoic sedimentary basins? What did the central Dronning Maud Land Mountains look like at the onset of the glaciation?
In order to trace the evolution of the central Dronning Maud Land Mountains since the Early Paleozoic, we have revisited the area with modern fission track dating and the addition of new apatite and zircon (U-Th)/He analyses (AHe and ZHe, respectively). Meier [8] and Meier et al. [9] still used non-standard etching techniques for their apatite fission track analyses and also used a somewhat unconventional zeta calibration approach. We have therefore re-analyzed many of Meier's samples to bring them up to the standard required for state-of-the-art thermochronological modelling. Additionally, new samples have been added from Schirmacheroasen (Figures 2 and 3). Based on this dataset, we present new thermal history models from 34 samples that shed light on the sedimentation, erosion and exhumation history of the central Dronning Maud Land Mountains.

Post-Pan-African Evolution
The post-Pan-African evolution of Dronning Maud Land is difficult to resolve due to the scarcity of preserved onshore Phanerozoic sediments; in central Dronning Maud Land, no Phanerozoic sedimentary rocks are exposed at all. When Gondwana had formed, two major active continental margin systems evolved on its northern and southern sides ( Figure 1). The southern Panthalassic margin evolved into a long-lasting active continental margin that significantly influenced the tectonic evolution of East Antarctica from Paleozoic-Mesozoic times, e.g., [37] and led to the formation of large sedimentary basins ( Figure 1). In western Dronning Maud Land, Phanerozoic sediments include only up to c. 160 m of Late Carboniferous-Middle Permian Beacon sediments [38][39][40][41][42]. Within the Transantarctic Mountains, on the other hand, a much more complete sedimentary succession is preserved, comprising up to c. 3 km of Devonian to Triassic siliciclastic Beacon sediments [43][44][45]. Equivalents of the Beacon sediments are widespread all-over southeastern Africa, along the conjugate margin to Dronning Maud Land, where sediments of the Karoo Supergroup are preserved in cumulative thicknesses of up to c. 12 km (e.g., [46]; Figure 1).
In western Dronning Maud Land and within the Transantarctic Mountains, the Beacon sediments unconformably overlie Mesoproterozoic and Cambrian-Ordovician basement rocks, forming a distinct paleosurface, e.g., [42,45]. Relics of a paleosurface, interpreted to be of similar age, have also been identified at Jutulsessen in Gjelsvikfjella [47]. In Sør-Rondane (eastern Dronning Maud Land), low-relief surfaces have been identified, but their age is still debated. These surfaces are mainly composed of glacially abraded bedrock without any preserved glacial deposits and it has therefore been suggested that they formed during an older glaciation [48,49], making them potentially of the same age as the paleosurfaces further west.

Post-Pan-African Evolution
The post-Pan-African evolution of Dronning Maud Land is difficult to resolve due to the scarcity of preserved onshore Phanerozoic sediments; in central Dronning Maud Land, no Phanerozoic sedimentary rocks are exposed at all. When Gondwana had formed, two major active continental margin systems evolved on its northern and southern sides ( Figure 1). The southern Panthalassic margin evolved into a long-lasting active continental margin that significantly influenced the tectonic evolution of East Antarctica from Paleozoic-Mesozoic times, e.g., [37] and led to the formation of large sedimentary basins ( Figure 1). In western Dronning Maud Land, Phanerozoic sediments include only up to c. 160 m of Late Carboniferous-Middle Permian Beacon sediments [38][39][40][41][42]. Within the Transantarctic Mountains, on the other hand, a much more complete sedimentary succession is preserved, comprising up to c. 3 km of Devonian to Triassic siliciclastic Beacon sediments [43][44][45]. Equivalents of the Beacon sediments are widespread all-over southeastern Africa, along the conjugate margin to Dronning Maud Land, where sediments of the Karoo Supergroup are preserved in cumulative thicknesses of up to c. 12 km (e.g., [46]; Figure 1).
In western Dronning Maud Land and within the Transantarctic Mountains, the Beacon sediments unconformably overlie Mesoproterozoic and Cambrian-Ordovician basement rocks, forming a distinct paleosurface, e.g., [42,45]. Relics of a paleosurface, interpreted to be of similar age, have also been identified at Jutulsessen in Gjelsvikfjella [47]. In Sør-Rondane (eastern Dronning Maud Land), low-relief surfaces have been identified, but their age is still debated. These surfaces are mainly composed of glacially abraded bedrock without any preserved glacial deposits and it has therefore been suggested that they formed during an older glaciation [48,49], making them potentially of the same age as the paleosurfaces further west. During the early stages of Gondwana fragmentation, large amounts of continental flood basalts associated with the Karoo mantle plume were emplaced at c. 183 Ma [50][51][52][53][54]. Today, these are exposed in thicknesses up to c. 5 km within the Lebombo Monocline in southeastern Africa, whereas they are only preserved in thicknesses up to c. 1 km in Vestfjella and up to c. 400 m within the southwestern Maud Belt in western Dronning Maud Land [51,[55][56][57][58][59]. Based on thermochronological data, however, it has been suggested that western Dronning Maud Land was covered by up to 2 km of Jurassic continental flood basalts, extending at least as far east as Hochlinfjellet at c. 4 • E [11,14]. Exposures of continental flood basalts have not been reported further east in central Dronning Maud Land, but Jurassic mafic dykes in Petermannkjedene and at Schirmacheroasen have been associated with the Karoo magmatism [60,61].
The initial fragmentation of Gondwana started in Early Jurassic times with the separation of Eastand West Gondwana along the Davie Fracture zone and other related transform faults, e.g., [62,63]. Continued rifting eventually led to the Middle Jurassic separation of Antarctica and Africa, resulting in the opening of pull-apart-type basins, such as the western Riiser-Larsen Sea and the Mozambique Basin, e.g., [5,62,63]. The oldest sea floor anomalies within the western Riiser-Larsen Sea have been dated to c. 155 Ma, and thus represent the latest possible age of the onset of the separation [62,64]. The fragmentation continued with the separation of India from both Antarctica and Australia at c. 135 Ma, e.g., [65], the opening of the eastern Riiser-Larsen Sea at c. 124 Ma [62], and the separation of Australia from Antarctica at c. 95 Ma [66]. It has furthermore been suggested that during the Cretaceous, a high orogenic plateau formed along the Panthalassic continental margin. This plateau began to collapse at ca. 105 Ma, leading to the formation of the West Antarctic Rift System, e.g., [67].
The post-rift evolution of Dronning Maud Land has mainly been inferred from offshore data. Seaward-dipping reflectors (Early Jurassic volcanics; [68,69]) overlain by c. 0.7-2.9 km of Cenozoic sediments have been documented between 25 • W and 9 • E. During ODP (Ocean Drilling Program) leg 113, five sites (sites 689-693) from the Maud Rise, north of central Dronning Maud Land, and on the eastern Weddell Sea margin were drilled [70,71]. These cores comprise uppermost Cretaceous-Cenozoic sediments, providing detailed information of the paleoclimate, including the growth of the East Antarctic ice sheet. Additionally, Quaternary sediments have been drilled at the Riiser-Larsen Ice Shelf, eastern Weddell Sea. These sediments are characterized by detritus from the Jurassic continental flood basalts and subordinate plutonic and/or metamorphic basement rocks [72].
The sediments from the ODP drill sites indicate that Dronning Maud Land was characterized by a temperate-cool subtropical climate during latest Cretaceous times, followed by generally increasing temperatures throughout the Late Paleocene, reaching a climax in Middle Eocene times. These ODP data also indicate a progressively cooler climate later throughout the late Eocene, eventually leading to formation and major growth of the East Antarctic ice sheet in late Eocene-Early Oligocene times [70].

Previous Thermochronological Results
Central Dronning Maud Land has been targeted by one quite extensive fission track study in the 1990s [8,9]. It has to be pointed out, however, that Meier [8] and Meier et al. [9] used non-standard etching techniques for their apatite fission track analyses, as well as an unconventional zeta calibration approach. This limits the usefulness of their track lengths measurements for thermochronological modelling, and also their fission track ages should be regarded with some caution. Titanite yielded fission track ages from Late Ordovician to Early Permian times [8], whereas zircon fission track ages are Late Devonian to Middle Triassic [8,9]. Early-Middle Jurassic apatite fission track ages clearly dominate the sample suite, although AFT ages as old as Middle Carboniferous and as young as Late Cretaceous are also present [8,9]. Emmel et al. [10] report the only available (U-Th)/He analyses for central Dronning Maud Land; apatite (U-Th)/He ages from six samples range from Late Carboniferous to mid-Cretaceous. These published datasets from central Dronning Maud Land have been interpreted to record cooling since the Pan-African Orogeny until the present day, with periods of slow cooling interrupted by several episodes of faster cooling [8][9][10]. The three main phases of accelerated cooling reported in central Dronning Maud Land have been related to Early Jurassic Gondwana rifting and the evolution of the passive margin [8][9][10], the opening of the western Riiser-Larsen Sea and the detachment of India from Antarctica during the Late Jurassic, and the mid-Cretaceous accelerated northward drift of India [8,9], respectively.
To the west of our study area, low-temperature thermochronological data have been reported from Ahlmannryggen and Annandagstoppane within the Archean Grunehogna Craton [14], and the Maud Belt between Heimefrontfjella in the southwest and Hochlinfjellet in the northeast [11,12,14,73]. Zircon (U-Th)/He analyses gave Neoproterozoic to Triassic single-grain ages, although most ages are either Permian or Ordovician, indicating that most of the samples had reached upper crustal depths either shortly after the Pan-African Orogeny or during the late Paleozoic peneplanation of western Dronning Maud Land [14].
The apatite fission track ages from western Dronning Maud Land range from Carboniferous to mid-Cretaceous. The dataset is dominated by Cretaceous and Triassic ages, with the oldest ages found at high elevations (such as Kirwanveggen), but the AFT ages also generally increase towards Hochlinfjellet in the east [11,12,14]. The apatite (U-Th)/He analyses yielded a wide range of single-grain ages, spanning from Early Ordovician to Eocene. Late Jurassic to Cretaceous ages are, however, predominant in western Dronning Maud Land [14,73]. Based on the combined apatite fission track and (U-Th)/He dataset, it has been suggested that western Dronning Maud Land was buried under up to 2 km of Jurassic continental flood basalts during Gondwana rifting [11,14]. This was followed by Late Jurassic-Cretaceous cooling related to rifting of East-and West Gondwana, subsequent opening of the South Atlantic, major plate reorganization and enhanced chemical weathering [11,14]. A final phase of Early-Middle Cenozoic cooling is recorded, associated with the onset of the glaciation, either as a result of increased glacial erosion [14], or due to differential exhumation and flexural isostatic rebound due to the load of the developing ice sheet [73].

Samples and Analytical Methods
Thirty-four samples from basement rocks of the Pan-African Maud Belt and the Tonian Oceanic Arc Super Terrane in central Dronning Maud Land are included in this study (Table 1). Twenty-eight of the samples have previously been analyzed by Meier [8], while six new samples are also included. The samples come from the mountain crest itself or from its northern side in areas between c. 7 • E (Mühlig-Hofmannfjella) and c. 16 • E (eastern Wohlthatmassivet) ( Figure 3). Most of the samples come from elevations between c. 1000 m a.s.l. and c. 3000 m a.s.l., except for five samples from c. 50-150 m a.s.l. at Schirmacheroasen, close to the coastline. The samples have been selected according to their localities and elevations in order to analyze a representative sample suite from the Dronning Maud Land escarpment. The apatite and zircon concentrates were extracted using standard mineral separation techniques, including a shaking table, followed by magnetic-and heavy-liquid separation.
By combining apatite fission track analyses with apatite and zircon (U-Th)/He analyses, a wide temperature range is covered. The partial annealing zone (PAZ) of the fission track system ranges from c. 60 • C to c. 120 • C, e.g., [74]. The partial retention zones (PRZ) for apatites and zircons with low degrees of metamictization are c. 35-70 • C and c. 150-230 • C, respectively, e.g., [75,76], although more recent studies have proven that radiation damage strongly affects the closure temperatures of the different (U-Th)/He systems [77,78].

Apatite Fission Track Analyses
Apatites from 31 samples were analyzed by the fission track technique (AFT), applying the external detector method [79]. The apatite crystals were first mounted in epoxy, then ground and polished to expose an internal crystal surface. The grain mounts were etched in 5 M nitric acid for 20 s at 20 ± 0.5 • C in order to reveal the spontaneous fission tracks. External mica detectors were placed on top of the grain mounts, and the sample packages were irradiated at the FRM II research reactor at the Technical University of Munich (Germany), using a thermal neutron flux of 1 × 10 16 neutrons/cm 2 . The neutron flux was monitored by using the dosimeter glasses IRMM-540R. The mica detectors were then etched for 20 min in 40% hydrofluoric acid at room temperature in order to reveal the induced tracks.
The fission track analyses were conducted on an Olympus BX51 optical microscope equipped with a computer-driven stage and the FT-Stage software [80] at the Department of Earth Science, University of Bergen, Norway. For fission track counting, a magnification of 1250× was used. The AFT central ages were calculated by the TrackKey software [81], applying the zeta calibration approach [82] with a zeta calibration factor of 214 ± 5 (H. Sirevaag).
Etch-pit diameters (Dpar; [83]) and confined track lengths were measured using a magnification of 2000×. As Dpar can be used as a proxy for apatite annealing kinetics, five Dpars were measured for each grain that was counted, and three Dpars were measured for each measured confined track length. Only track-in-tracks (TinTs) were considered for the track length measurements. If possible, 100 TinTs were measured for each sample.

(U-Th)/He Analyses
(U-Th)/He analyses were conducted on apatite crystals from 33 samples and zircon crystals from two samples. The grains were analyzed at the GÖochron Laboratories, Geoscience Center, University of Göttingen, Germany. The individual grains were carefully evaluated and handpicked under binocular and petrographic microscopes in order to avoid fractures and mineral-and fluid-inclusions as far as possible. The grains were photographed for determining crystal dimensions and then packed individually in platinum capsules. The 4 He content was determined by degassing under high vacuum by heating with an infrared diode laser for 2 min. The extracted gas was purified with a SAES Ti-Zr getter at 450 • C and analyzed with a Hiden triple-filter quadrupole mass spectrometer, equipped with a positive ion-counting detector. During sequential reheating and He measurements, the crystals were checked for complete degassing of He (re-extract).
The platinum capsules were removed after the He analyses in order to measure the U, Th and Sm contents. The apatites were dissolved in a 4% HNO 3 + 0.05% HF acid mixture in Savillex teflon vials, and the zircons were dissolved in a pressurized teflon bomb with a mixture of double-distilled 48% HF and 65% HNO 3 . The dissolved crystals were spiked with calibrated 230 Th and 233 U solutions and analyzed by the isotope dilution method either on a Perkin Elmer Elan DRC II, or by a Thermo iCAP Q ICP-MS, equipped with an APEX micro-flow nebulizer. Alpha-ejection correction (F T -correction) was applied to the raw (U-Th)/He ages after the procedures described by Farley et al. [84] and Hourigan et al. [85].
The (U-Th)/He dataset was carefully evaluated. Grains with a He re-extract > 5%, total analytical uncertainty > 10%, or grains that were statistical outliers compared to the remaining single-grain analyses in the same sample according to the Grubbs and Dixon tests [86][87][88] were excluded. Excluded grains are not included in any figures, calculations or thermal models, but are reported together with the remaining dataset in italic.

Modelling of the Thermal History
The software HeFTy v. 1.9.1 [89] was used for the thermal history modelling. For the AFT data, we applied the annealing model of Ketcham et al. [90], as well as c-axis projection of the confined track lengths according to Ketcham et al. [91]. The radiation damage accumulation and annealing model of Flowers et al. [92] and the model of Guenthner et al. [78] were used for modelling apatite and zircon (U-Th)/He data, respectively. As most samples include more than one thermochronometer, all thermochronometers in a sample were generally modelled together. In cases where it was not possible to model all (U-Th)/He data and AFT data together, the "problematic" analyses were excluded from the model. "Problematic" analyses have been identified by testing various combinations of AFT data and single-grain AHe analyses and, if necessary, excluding single-grain AHe analyses that are incompatible with the remaining AHe analyses and/or the fission track analyses from the same sample (i.e., prevent the model from finding any cooling paths). The thermochronometers that were used are specified for each model.
During the thermal modelling, 100,000 random paths were tested. In the following sections, we use the best-fit and weighted mean paths of the acceptable (goodness-of-fit ≥ 0.05) and good (goodness-of-fit ≥0.5) paths for comparisons of the models. The goodness-of-fit is calculated by using Kuiper's statistics in HeFTy [89].
In order to model the thermal histories, external t-T constraints must be applied, based on pre-existing thermochronological and geological evidence. For all thermal models, similar start and end constraints were applied. As a starting constraint, we used biotite 40 Ar- 39 Ar ages of c. 450 Ma from Filchnerfjella (westernmost Orvinfjella) in the westernmost part of our study area [93], recording temperatures of c. 300-350 • C. The present-day surface temperature (−25 • C) is used as the end constraint. We have tested different geological scenarios, which will be discussed in more detail below. Briefly, three sets of models have been run: (1) simple cooling with only start and end constraints.
(2) late Paleozoic peneplanation with subsequent late Paleozoic-early Mesozoic burial beneath sediments. For these models, we inferred the position of the late Paleozoic peneplain by connecting mountain peaks from the escarpment to the hinterland. These peaks define a gently southward sloping surface, which we interpreted as the last remnant of the peneplain. Based on glacial deposits within both the Beacon sediments in Heimefrontfjella and the Karoo Supergroup in south-central Africa, e.g., [42,94,95], we have applied a surface temperature of −10 • C and corrected the temperature for each sample based on their vertical distance to the extrapolated peneplain, assuming a geothermal gradient of 25 • C/km (this has been done for all following steps as well). This approach works well for the main mountain range but not for samples from the coast (e.g., Schirmacheroasen)-they were generally given a lot more freedom during the modelling to account for the uncertainty in their position with regard to the extrapolated peneplain. A second constraint box allows for Permo-Triassic sedimentary burial; Triassic surface temperatures have been set to 25 • C [96]. (3) late Paleozoic peneplanation with subsequent late Paleozoic-early Mesozoic burial beneath sediments, followed by rapid cooling in the Late Triassic-Early Jurassic and renewed burial in the Jurassic (as suggested by studies from adjacent regions by e.g., Krohne [13] and Sirevaag et al. [14]). This has been implemented by forcing the samples to the surface in the Early Jurassic, assuming a surface temperature of 25 • C, corrected for each sample's depth below the extrapolated peneplain, followed by a second constraint box to allow for re-burial.

Results
The fission track and (U-Th)/He analyses are summarized in Appendices A and B, as well as  Table S1, while track length data are reported in Supplementary Material Table S2 and Figure S1.

Apatite Fission Track Results
Thirty-one samples were selected for apatite fission track analyses. The uranium concentrations range from c. 5 to 98 ppm, with most apatites showing relatively homogeneous uranium distribution. The average measured Dpars, which are used as a proxy for apatite annealing kinetics, range from c. 1.2 to 1.8 µm. Most of the Dpars are lower than 1.6 µm, representing fast-annealing near-end member fluorapatites [97]. There are no observed correlations of age to either Dpar or to uranium concentration within the dataset.
The AFT ages range from c. 280 to 85 Ma, although the majority of the ages are either Cretaceous (n = 14) or Jurassic (n = 10). Measured confined track lengths (MTL) span from c. 10.3 to 13.4 µm, with all but six samples recording MTLs longer than 11.3 µm.
The AFT ages younger than c. 165 Ma are mainly derived from elevations below c. 1500 m a.s.l., whereas ages older than c. 165 Ma are mostly found at higher elevations ( Figure 4A). This results in a moderate correlation (R 2 = 0.5) between age and elevation for all samples excluding the samples from Schirmacheroasen. Additionally, a regional trend is observed when plotting AFT data against latitude ( Figure 4B), with younger ages close to the coast and older ages towards the great escarpment. Since elevations also increase towards the inland, it is difficult to determine whether elevation or distance from the coast, or both, control the distribution of fission track ages.
The age-MTL plot ( Figure 4C) shows mainly two groups of samples. The majority of samples have mean track lengths above 11.3 µm and cover the entire range of ages, whereas six samples with mean track lengths below 11.3 µm mostly cluster around 150 Ma. The samples from this latter group were all collected below c. 1800 m a.s.l. and south of c. 71 • S ( Figure 4D).

(U-Th)/He Results
(U-Th)/He analyses were conducted on 132 apatite crystals from 33 samples. The selected crystals were mostly euhedral and of good quality. Some of the selected crystals contained small inclusions and fractures even though considerable effort was put into selecting the best possible grains for analyses. The calculated sphere radii are between 31 and 122 µm, although most grains are

(U-Th)/He Results
(U-Th)/He analyses were conducted on 132 apatite crystals from 33 samples. The selected crystals were mostly euhedral and of good quality. Some of the selected crystals contained small inclusions and fractures even though considerable effort was put into selecting the best possible grains for analyses. The calculated sphere radii are between 31 and 122 µm, although most grains are smaller than c. 100 µm. Most of the samples have effective uranium (eU) concentrations up to c. 240 ppm. However, extremely high eU concentrations are recorded in all five grains from sample JJ1875 (Madsensåta), with eU concentrations ranging from 780 to 1812 ppm.
Fourteen of the 132 single-grain analyses were excluded according to the filtering criteria described in Section 4.2. Eleven single-grain ages were statistical outliers compared to the rest of the single-grain analyses of the specific sample, and were thus excluded. Two analyses were excluded due to large error, and one analysis was excluded because the age was significantly older than all AFT and ZHe ages of the study area.
The remaining apatite (U-Th)/He analyses gave single-grain ages between c. 305 and 15 Ma. Similar to the AFT data, the AHe ages correlate moderately well (R 2 = 0.5) with elevation ( Figure 5A) and are generally younger towards the coast to the north ( Figure 5B). The remaining apatite (U-Th)/He analyses gave single-grain ages between c. 305 and 15 Ma. Similar to the AFT data, the AHe ages correlate moderately well (R 2 = 0.5) with elevation ( Figure 5A) and are generally younger towards the coast to the north ( Figure 5B). The (U-Th)/He data have been evaluated based on the scatter of single-grain ages within a sample. The standard deviation of the single-grain ages is below 10% for 15 of our samples and between 10 and 20% for eight samples; ten samples have standard deviations exceeding 20%. Some scatter in single-grain ages can be expected, even though the individual grains in a sample have experienced similar t-T histories. To some extent, this can be explained by grain characteristics, such as grain size, radiation damage, eU zonation, U-Th-rich micro-inclusions and He-implantation from neighboring grains [98][99][100][101]. Large apatites are expected to yield older ages than smaller apatites since the larger apatites have greater effective diffusion dimensions, and thus higher He retentivities [100]. Also, as the accumulation of radiation damage reduces the diffusivity of He in apatites, it is expected that high-eU apatites will yield older ages than low-eU apatites [77]. Thus, a positive relationship for both age-eU and age-grain size can explain the scatter of single-grain ages within a sample. For samples where a 1σ standard deviation above 20% cannot be explained by either age-eU or age-grain size relationships, Flowers and Kelley [101] suggest to exclude the apatite (U-Th)/He data from the thermal modelling. From the ten samples with a standard deviation above 20%, six samples show a moderate to very good correlation (R 2 > 0.6) between either age-eU, age-grain size or both, and are therefore included in the models. In the remaining four samples (JJ1677, JJ1875, JJ1924 and JJ1974) ages do not correlate with eU or grain size (R 2 < 0.2), and the samples are thus excluded from modelling.
Zircons from two samples, both from Schirmacheroasen, were analyzed by the (U-Th)/He method. This resulted in six single-grain ages ranging from c. 420 to 340 Ma. The calculated sphere radii range from 47 to 55 µm, while measured eU concentrations are between 130 and 339 ppm. None of the zircon (U-Th)/He analyses were rejected. The single-grain zircon (U-Th)/He ages show a The (U-Th)/He data have been evaluated based on the scatter of single-grain ages within a sample. The standard deviation of the single-grain ages is below 10% for 15 of our samples and between 10 and 20% for eight samples; ten samples have standard deviations exceeding 20%. Some scatter in single-grain ages can be expected, even though the individual grains in a sample have experienced similar t-T histories. To some extent, this can be explained by grain characteristics, such as grain size, radiation damage, eU zonation, U-Th-rich micro-inclusions and He-implantation from neighboring grains [98][99][100][101]. Large apatites are expected to yield older ages than smaller apatites since the larger apatites have greater effective diffusion dimensions, and thus higher He retentivities [100]. Also, as the accumulation of radiation damage reduces the diffusivity of He in apatites, it is expected that high-eU apatites will yield older ages than low-eU apatites [77]. Thus, a positive relationship for both age-eU and age-grain size can explain the scatter of single-grain ages within a sample. For samples where a 1σ standard deviation above 20% cannot be explained by either age-eU or age-grain size relationships, Flowers and Kelley [101] suggest to exclude the apatite (U-Th)/He data from the thermal modelling. From the ten samples with a standard deviation above 20%, six samples show a moderate to very good correlation (R 2 > 0.6) between either age-eU, age-grain size or both, and are therefore included in the models. In the remaining four samples (JJ1677, JJ1875, JJ1924 and JJ1974) ages do not correlate with eU or grain size (R 2 < 0.2), and the samples are thus excluded from modelling.
Zircons from two samples, both from Schirmacheroasen, were analyzed by the (U-Th)/He method. This resulted in six single-grain ages ranging from c. 420 to 340 Ma. The calculated sphere radii range from 47 to 55 µm, while measured eU concentrations are between 130 and 339 ppm. None of the zircon (U-Th)/He analyses were rejected. The single-grain zircon (U-Th)/He ages show a standard deviation of 2 and 7% for the two different samples.
By combining apatite fission track with apatite and zircon (U-Th)/He data, the thermal evolution of a sample can be traced through a wide temperature range (c. 230-35 • C). The obtained thermochronological age corresponds to the cooling through the partial annealing/retention zones of the different thermochronometers. Based on the partial annealing/retention zones of the different systems, we would expect the ZHe age to be oldest (PRZ: c. 230-150 • C; [76]), followed by the AFT age (PAZ: c. 120-60 • C; [74]) and the AHe age (PRZ: c. 70-35 • C; [75]). For the 31 multidated samples in this study, the ZHe age is the oldest in both cases. While the AHe ages of most samples are younger than the corresponding AFT ages, five samples yielded similar unweighted mean apatite (U-Th)/He ages and apatite fission track ages, meaning that the ages either overlap within their uncertainties or the fission track age overlaps with the range of single-grain (U-Th)/He ages. In three other samples, all apatite (U-Th)/He single-grain ages are older than the fission track age. These samples are all older than 160 Ma (AFT age) and have U concentrations above 30 ppm (Appendix A). Samples with similarly high (or higher) U but younger AFT ages do not show a crossover in age, thus we consider this to be an effect of radiation damage accumulation over time.

Tectono-Thermal Evolution
During the Pan-African orogeny, various parts of East-and West Gondwana collided to form the Himalaya-scale East African-Antarctic Orogen and the supercontinent Gondwana, e.g., [31]. The collapse of this orogen and eventual breakup of Gondwana in Jurassic-Cretaceous times, gave birth to our modern oceans. Central Dronning Maud Land was in the thick of these momentous events and preserves excellent evidence of the collisional history and early orogenic collapse, e.g., [29]. The geological record of rifting, continental breakup and the transformation into a passive continental margin, on the other hand, has been lost to subsequent erosion, and the only post-Pan-African rocks exposed in central Dronning Maud Land are a few mafic dykes associated with the Jurassic Karoo flood basalt province [60,61]. In the absence of geological evidence to the contrary, previous authors have interpreted thermochronological data from central Dronning Maud Land in the context of a relatively simple history of protracted cooling from Late Cambrian times until today, e.g., [8][9][10]. Significantly more complex thermal histories have since been proposed for areas east and west of central Dronning Maud Land and cast doubt on this simple model [12][13][14]. Testing such a simple cooling model against our data has failed to produce any acceptable cooling paths in about half the samples (Appendix C). Samples that are incompatible with protracted cooling can be found all over the study area, confirming that this scenario is likely too simple and more complex thermal histories should be investigated (Figures 6-8). Considering the scarcity of a post-Pan-African geological record in central Dronning Maud Land, we will have to look towards adjacent regions in western and eastern Dronning Maud Land and towards South Africa-Mozambique, the conjugate margin to central Dronning Maud Land, to develop reasonable geological scenarios that can be tested against our data.  [42]. Additionally, remnants of a similar paleosurface have been reported from Gjelsvikfjella (western Dronning Maud Land; [47]) and Sør-Rondane (eastern Dronning Maud Land; [13,48,49]). Though lacking the overlying sediments, these paleosurfaces have been correlated with the peneplain exposed in western Dronning Maud Land. In the Transantarctic Mountains, c. 800 km west of Dronning Maud Land, the Beacon sediments (Devonian-Triassic) also cover an erosional unconformity, the Kukri erosional surface [45]. This distinct Devonian-Carboniferous paleosurface or peneplain below the Beacon sediments can therefore be traced over hundreds of kilometers, from the Transantarctic Mountains far to the west of our study area to Sør-Rondane to the east of our study area. While this paleosurface is less well preserved in central Dronning Maud Land, peneplains are large-scale features [102] and we consider it likely that the same erosional surface was once present in our study area as well; it must today mostly lie above the present-day erosional level. Mountain tops from the crest of the escarpment to the hinterland can be linked by an enveloping surface with a gentle southward dip of c. 1 • , and we speculate that this enveloping surface represents the last remnant of the once continuous Devonian-Carboniferous peneplain that truncated all of Dronning Maud Land.
Peneplanation was followed by subsidence and sedimentation. Only erosional remnants of the Beacon sediments are preserved in western Dronning Maud Land, but a more complete succession of 2.5-3 km thick Devonian-Triassic Beacon sediments can be found in the Transantarctic Mountains [43][44][45]. In Mozambique and South Africa, the upper Carboniferous-Lower Jurassic Karoo sediments are considered to be equivalent to the Beacon Sediments and are still preserved in thicknesses of 4-4.5 km and 12 km, respectively [46,[103][104][105]. Western Dronning Maud Land was thus surrounded by upper Paleozoic-lower Mesozoic sedimentary basins, suggesting that equivalent sediments might have been deposited in our study area as well. Indeed, all apatite (U-Th)/He and fission track ages are Permian or Mesozoic, thus a late Paleozoic peneplanation, bringing the samples close to the surface, must have been followed by reburial to account for the younger ages.
Most of our samples are Jurassic and younger and do not record the late Paleozoic-early Mesozoic history. Thus, we have selected three of the oldest samples, yielding Triassic ages (JJ1720, JJ1736 and JJ1746), and two samples from Schirmacheroasen, for which ZHe ages are also available (J03.02./1 and S30.1), to investigate potential peneplanation and subsequent re-burial. All five samples allow late Paleozoic cooling to (near) surface temperatures, followed by re-heating due to sedimentary burial. Temperatures during burial must have reached at least 80 • C during Permo-Triassic times (Figures 6 and 8). The maximum temperatures are not well constrained; most good-fit paths suggest temperatures between c. 80-140 • C, corresponding to at least c. 3 km, but possibly up to c. 6 km of sediments (assuming a geothermal gradient of c. 25 • C). These estimates, though vague, agree with recorded Beacon and Karoo sediment thicknesses in the Transantarctic Mountains, Mozambique and South Africa. The depth of the samples from Schirmacheroasen below the late Paleozoic peneplain is uncertain. These samples come from coastal outcrops far north of the escarpment and the peneplain might have been kilometers above the present-day erosional level or might have been downfaulted or downwarped during rifting [106]. The two models from Schirmacheroasen indicate similar temperatures during peneplanation and burial than samples from the main mountain range and thus might have been at a similar crustal depth in the late Paleozoic-early Mesozoic. With thousands of meters difference in elevation today, down-to-the-north faulting along margin parallel normal faults during rifting seems the most likely possibility.
The samples with Jurassic and younger AFT and AHe ages have lost all information about the late Paleozoic-early Mesozoic thermal history. We assume, however, that they have experienced peneplanation and subsequent sedimentation as well, as these are not local, but usually rather widespread phenomena that should affect the entire study area. Thus, we have applied the information on peneplanation and sedimentary burial that we have gained from the five samples discussed above as external constraints to all the models for younger samples. We would like to point out that all samples are compatible with late Paleozoic-early Mesozoic peneplanation and burial (Appendix C), thus we prefer this scenario over a simple cooling history as discussed above, which failed to produce even acceptable-fit cooling paths in half the samples.      In Heimefrontfjella (western Dronning Maud Land), the Beacon sediments have been eroded down to a few meters, before being covered by up to 1.5-2 km of Jurassic continental flood basalts, indicating that the basement rocks exposed today were already close to the surface twice before, first during the Carboniferous-Permian deposition of the Beacon sediments and then again during the Early Jurassic emplacement of the continental flood basalts [14,42]. These flood basalts were associated with early Gondwana rifting and extended at least as far east as Hochlinfjellet in western Dronning Maud Land. Apatite fission track data from that region indicate diminishing basaltic thicknesses away from the emplacement zone in Jutulstraumen [14]. Today, the only evidence for the Karoo magmatism in central Dronning Maud Land are Jurassic mafic dykes in Petermannkjedene [60] and Schirmacheroasen [61]. In order to further pinpoint the extent of the continental flood basalts, we tested the westernmost of our samples from the main mountain range (JJ1700, JJ1742 and JJ1768) for Early Jurassic surface exposure and rapid reheating due to burial beneath flood basalts. While all three samples might have cooled to (near) surface temperatures already in Early Jurassic times, none of the models showed signs of Jurassic reheating (Appendix C). The models were not forced to higher temperatures in the Jurassic, but a relatively large constraint box was used, allowing samples to either stay at (near) surface temperatures or be reheated due to burial. The absence of Jurassic reheating in these samples leads us to suggest that the eastern margin of the Jurassic continental flood basalts was located somewhere between Hochlinfjellet (4° E) and Mühlig-Hofmannfjella (7° E).
Other areas in East Antarctica have been buried beneath Jurassic sedimentary basins (Sør-Rondane: [13]; Victoria Basin: [111]; Shackleton Range: [112]). We have thus tested a similar scenario (late Paleozoic peneplanation, burial beneath the Beacon sediments, erosion of Beacon sediments, burial beneath Jurassic basin) against our data. However, the majority of our samples either do not favor Jurassic reheating (significantly fewer good-or acceptable-fit paths are found than for models without Jurassic reheating, e.g., samples JJ1796, JJ1812, SG-25; Appendix C)) or do not show any significant increase in temperature during the Jurassic (e.g., samples JJ1700, JJ1736, JJ1730; Appendix C). Several models allow for Jurassic reheating, simply because the thermochronological ages are Cretaceous and the samples essentially preserve little to no information on the Jurassic thermal history (e.g., samples JJ1886, JJ1931, JJ1875, JJ1731). None of these models actually requires Jurassic-Cretaceous reheating and they all work equally well without it. Our modelling results might In Heimefrontfjella (western Dronning Maud Land), the Beacon sediments have been eroded down to a few meters, before being covered by up to 1.5-2 km of Jurassic continental flood basalts, indicating that the basement rocks exposed today were already close to the surface twice before, first during the Carboniferous-Permian deposition of the Beacon sediments and then again during the Early Jurassic emplacement of the continental flood basalts [14,42]. These flood basalts were associated with early Gondwana rifting and extended at least as far east as Hochlinfjellet in western Dronning Maud Land. Apatite fission track data from that region indicate diminishing basaltic thicknesses away from the emplacement zone in Jutulstraumen [14]. Today, the only evidence for the Karoo magmatism in central Dronning Maud Land are Jurassic mafic dykes in Petermannkjedene [60] and Schirmacheroasen [61]. In order to further pinpoint the extent of the continental flood basalts, we tested the westernmost of our samples from the main mountain range (JJ1700, JJ1742 and JJ1768) for Early Jurassic surface exposure and rapid reheating due to burial beneath flood basalts. While all three samples might have cooled to (near) surface temperatures already in Early Jurassic times, none of the models showed signs of Jurassic reheating (Appendix C). The models were not forced to higher temperatures in the Jurassic, but a relatively large constraint box was used, allowing samples to either stay at (near) surface temperatures or be reheated due to burial. The absence of Jurassic reheating in these samples leads us to suggest that the eastern margin of the Jurassic continental flood basalts was located somewhere between Hochlinfjellet (4 • E) and Mühlig-Hofmannfjella (7 • E).
Other areas in East Antarctica have been buried beneath Jurassic sedimentary basins (Sør-Rondane: [13]; Victoria Basin: [111]; Shackleton Range: [112]). We have thus tested a similar scenario (late Paleozoic peneplanation, burial beneath the Beacon sediments, erosion of Beacon sediments, burial beneath Jurassic basin) against our data. However, the majority of our samples either do not favor Jurassic reheating (significantly fewer good-or acceptable-fit paths are found than for models without Jurassic reheating, e.g., samples JJ1796, JJ1812, SG-25; Appendix C)) or do not show any significant increase in temperature during the Jurassic (e.g., samples JJ1700, JJ1736, JJ1730; Appendix C). Several models allow for Jurassic reheating, simply because the thermochronological ages are Cretaceous and the samples essentially preserve little to no information on the Jurassic thermal history (e.g., samples JJ1886, JJ1931, JJ1875, JJ1731). None of these models actually requires Jurassic-Cretaceous reheating and they all work equally well without it. Our modelling results might therefore allow for a few relatively local, shallow sedimentary basins, but we do not believe that entire central Dronning Maud Land was covered by a thick section of Jurassic or younger sediments.

Mesozoic-Cenozoic Cooling
The Mesozoic-Cenozoic thermal evolution of central Dronning Maud Land is generally characterized by relatively slow cooling until the onset of glaciation (Figures 6-8). During Jurassic to early Paleogene times, average cooling rates were well below 1.0 • C/Myr for most samples. A significant change in cooling rates can be observed at the Eocene-Oligocene boundary, with average cooling rates of c. 0.7-2.6 • C/Myr being recorded during the last 34 Ma. Temporarily, cooling rates reached up to c. 7.5 • C/Myr, e.g., in Orvinfjella. The cooling rates since Oligocene times are similar to the cooling rates recorded in western Dronning Maud Land during the glaciation [14].
Throughout the Paleogene, the southward drift of Antarctica resulted in a progressively more isolated position of the continent. This led to the transition from a temperate-cool subtropical climate during the Early Eocene, to cooling and eventually glaciation around the Eocene-Oligocene boundary [70,[113][114][115]. The dramatic decrease in temperature due to the changing climate (from 25 • C surface temperatures in the Late Cretaceous to −25 • C surface temperature today) is large enough to be recorded in low-temperature thermochronological data. At the same time, the deteriorating climate and eventual onset of glaciation marked a significant change in weathering and erosion conditions, enhancing denudation and cooling. To differentiate between the two, we compare the modelled temperatures of our samples with the surface temperature curve since Late Jurassic times. Interestingly, several samples from Orvinfjella, Wohlthatmassivet and Weyprechtfjella reached surface temperatures already in the Late Jurassic (JJ1736 and JJ1924) or during the Cretaceous (JJ1700, JJ1720, JJ1746, JJ1766, JJ1768 and JJ1812). The cooling histories of these particular samples more or less follow the surface temperature curve throughout the Cenozoic, indicating that the samples were sitting close to the surface and apparently experienced no significant Cenozoic erosion; the Cenozoic cooling in these models is entirely attributed to climate cooling. All other samples, on the other hand, show greater Cenozoic cooling than can be explained by the changing climate and we attribute this cooling to erosion of the overlying rock column. Freeze-thaw processes, especially during the early stages of glaciation, are a highly effective mechanism for physical weathering [116] and the offshore sedimentary record, which is dominated by Cenozoic sediments, also supports this interpretation. Additionally, Cenozoic faulting might have contributed to Cenozoic cooling by tectonic denudation. However, the continental margin has been a passive-margin since Cretaceous rifting and we expect fault activity to have played a minor role at most. Comparing the modelled temperatures of samples at the onset of glaciation to the surface temperatures and applying a geothermal gradient of 25 • C/km, we can estimate the paleodepth of the samples and hence the eroded rock column since the onset of glaciation. Many samples from the continental wedge and Orvinfjella were located at relatively shallow crustal depths at the onset of the glaciation (< 500 m), indicating that most of the erosion must have taken place prior to 34 Ma. Several samples from the main mountain range (e.g., JJ1673, JJ1796, JJ1931), on the other hand, experienced up to 2 km of erosion in the last 34 Ma. Erosion was thus strongly focused, which is to be expected from glacial erosion processes.

Regional Thermochronological Age Distribution and Tectonic Implications
A comparison of the published apatite fission track ages from Dronning Maud Land reveals important differences between western, central and eastern Dronning Maud Land (Figure 9). While peaks in the age distribution should be regarded with caution, since fission track ages in slowly cooled samples do not necessarily correspond to any particular geological event, the differences are also evident in age-elevation plots and modelled cooling histories. As a first observation, the proportion of ages pre-dating the Jurassic onset of rifting seems to increase from west to east. Jurassic rifting was associated with the eruption of voluminous continental flood basalts in western Dronning Maud Land and burial beneath these basalts reset most of the low-temperature thermochronometers. Older ages only survived in areas where the basaltic cover was relatively thin, e.g., in easternmost western Dronning Maud Land [14]. Central and eastern Dronning Maud Land were most likely never covered by basalts and thus preserve better evidence of the pre-rift history.
In eastern Dronning Maud Land, the Late Triassic-Early Jurassic and Late Jurassic-Early Cretaceous age peaks generally correspond to periods of cooling recorded in the thermal history models from that area, and have been linked to early rift processes in Gondwana and rifting in the Riiser-Larsen Sea, respectively [13]. In central Dronning Maud Land, the Jurassic and Cretaceous age peaks do not correlate with any significant cooling episodes in the thermal models, suggesting that these ages might not record any particular geological event but rather reflect relatively slow cooling through the partial annealing zone. This is also indicated by the age-elevation distribution in central Dronning Maud Land, where Permian to Cretaceous fission track ages are clearly correlated with elevation, with a gentle slope indicative of slow Mesozoic cooling (excepting samples from Schirmacheroasen, which was most likely downfaulted with respect to the main mountain range). This is a marked contrast to western Dronning Maud Land, where mid-Cretaceous ages clearly dominate the age spectrum and occur over a wide range of elevations (c. 900-2400 m a.s.l.), suggesting rapid cooling at this time. The latter is reflected in increased Cretaceous cooling rates recorded in many of the thermal history models from western Dronning Maud Land. western Dronning Maud Land [14]. Central and eastern Dronning Maud Land were most likely never covered by basalts and thus preserve better evidence of the pre-rift history.
In eastern Dronning Maud Land, the Late Triassic-Early Jurassic and Late Jurassic-Early Cretaceous age peaks generally correspond to periods of cooling recorded in the thermal history models from that area, and have been linked to early rift processes in Gondwana and rifting in the Riiser-Larsen Sea, respectively [13]. In central Dronning Maud Land, the Jurassic and Cretaceous age peaks do not correlate with any significant cooling episodes in the thermal models, suggesting that these ages might not record any particular geological event but rather reflect relatively slow cooling through the partial annealing zone. This is also indicated by the age-elevation distribution in central Dronning Maud Land, where Permian to Cretaceous fission track ages are clearly correlated with elevation, with a gentle slope indicative of slow Mesozoic cooling (excepting samples from Schirmacheroasen, which was most likely downfaulted with respect to the main mountain range). This is a marked contrast to western Dronning Maud Land, where mid-Cretaceous ages clearly dominate the age spectrum and occur over a wide range of elevations (c. 900-2400 m a.s.l.), suggesting rapid cooling at this time. The latter is reflected in increased Cretaceous cooling rates recorded in many of the thermal history models from western Dronning Maud Land. Figure 9. Probability density plot generated from available apatite fission track central ages from Dronning Maud Land. The time range of the main events that potentially affected the region are indicated in the background. The AFT ages from western Dronning Maud Land (DML) include data from Jacobs and Lisker [11], Emmel et al. [12] and Sirevaag et al. [14]. AFT ages from central Dronning Maud Land include the data from the present study together with the data from Emmel et al. [10]. The AFT ages from eastern Dronning Maud Land are solely based on data from Krohne [13]. Timing of the different events are from Plumstead [38], Lindström [40,41], Isbell [45], Duncan et al. [50], Roeser et al. [62], Bialas et al. [67] and Ingólfsson [114]. Figure 9. Probability density plot generated from available apatite fission track central ages from Dronning Maud Land. The time range of the main events that potentially affected the region are indicated in the background. The AFT ages from western Dronning Maud Land (DML) include data from Jacobs and Lisker [11], Emmel et al. [12] and Sirevaag et al. [14]. AFT ages from central Dronning Maud Land include the data from the present study together with the data from Emmel et al. [10]. The AFT ages from eastern Dronning Maud Land are solely based on data from Krohne [13]. Timing of the different events are from Plumstead [38], Lindström [40,41], Isbell [45], Duncan et al. [50], Roeser et al. [62], Bialas et al. [67] and Ingólfsson [114].
Comparing the thermal history models from western and central Dronning Maud Land, we suggest that samples from both areas were most likely close to the surface during the late Paleozoic peneplanation and were subsequently buried beneath upper Paleozoic-lower Mesozoic sedimentary basins. Inversion of these basins and erosion of the sediments occurred relatively quickly during Early Jurassic times in western Dronning Maud Land and much slower during Jurassic-Cretaceous times in central Dronning Maud Land. This difference might be explained by dynamic uplift above the Karoo mantle plume starting at ca. 200 Ma, which affected western Dronning Maud Land, but did not extend as far east as central Dronning Maud Land. The eruption of a thick layer of Jurassic flood basalts associated with this plume then preserved the remnants of Carboniferous-Permian Beacon sediments that can still be found in western Dronning Maud Land today, while the unshielded sediments in central Dronning Maud Land have since been completely stripped by erosion.
The pronounced Cretaceous cooling in western Dronning Maud Land has been linked to rift processes in the South Atlantic and Riiser-Larsen Sea, passive continental margin development and climate change [14]. Its absence in central Dronning Maud Land, however, suggests that it might have more local sources. To the west, in the area of the West Antarctic Rift System, an orogenic plateau with strongly thickened crust had developed during long-lasting subduction along the Panthalassic margin of Gondwana [67,117,118]. This plateau existed until mid-Cretaceous times, when the subduction on the Pacific margin ceased ( [119]; and references therein), leading to plateau collapse at c. 105 Ma and the development of the West Antarctic Rift System and the present-day Transantarctic Mountains [67]. The age of plateau collapse coincides well with the main age peak recorded in western Dronning Maud Land and the relative proximity of western Dronning Maud Land to the West Antarctic Rift System might explain the pronounced mid-Cretaceous cooling recorded here and lack thereof in areas further east.

Summary and Conclusions
We have applied a combination of low-temperature thermochronological methods (i.e., apatite fission track, apatite and zircon (U-Th)/He) to gain insights into the post-Pan-African evolution of central Dronning Maud Land. While previous studies from central Dronning Maud Land have interpreted thermochronological data in terms of simple, monotonic post-Pan-African cooling, our thermochronological data suggest a more complex tectono-thermal history. We have tested models including late Paleozoic peneplanation and subsequent re-burial beneath the Beacon sediments, as well as models that additionally include Early Jurassic erosion followed by a second period of Mesozoic re-burial. Whereas a second period of surface exposure and subsequent burial is favored in western and eastern Dronning Maud Land, the new data are most consistent with a scenario that only includes late Paleozoic (near) surface exposure and late Paleozoic-early Mesozoic re-burial by c. 3-6 km of Beacon sediments. Also, as the westernmost samples do not show signs of Jurassic re-burial beneath continental flood basalts, the eastern boundary of the Jurassic continental flood basalts can be pinpointed to the area between Hochlinfjellet (4 • E) and Mühlig-Hofmannfjella (7 • E).
Contrary to eastern and western Dronning Maud Land, the AFT age peaks in central Dronning Maud Land do not correlate to any particular cooling phases in the thermal models. Combined with a gently sloping AFT age-elevation correlation for all central Dronning Maud Land samples, except for Schirmacheroasen (which was probably downfaulted), this indicates slow cooling throughout the Mesozoic (≤1 • C/Myr). The lack of a pronounced Cretaceous cooling phase as observed in western Dronning Maud Land, suggests that the mid-Cretaceous cooling of western Dronning Maud Land [14] could be attributed to the proximity to the collapse of the orogenic plateau at the Panthalassic Gondwana margin at c. 105 Ma, and the following formation of the West Antarctic Rift System, rather than to rift processes in the South Atlantic and the Riiser-Larsen Sea. Since the Paleogene, increased cooling rates are attributed to climate deterioration and glacial denudation of up to 2 km.

Supplementary Materials:
The following are available online at http://www.mdpi.com/2076-3263/8/11/390/s1. Table S1: Apatite fission track single-grain ages. Table S2: Individual track lengths used for thermal modelling. Figure S1: Confined track length distributions. Acknowledgments: Judit Dunklné-Nagy and Irina Maria Dumitru are thanked for help with sample preparation and Karen Tellefsen for help with editing text and figures. We thank the Ph.D. program at the University of Bergen and the University of Bergen Meltzer fund for funding. We appreciate the reviews from Maria Laura Balestrieri and Ratislav Vojtko, and thank the editors Allen Wang and Xiaozhen Han for handling this paper.

Conflicts of Interest:
The authors declare no conflict of interest.       [85]). Uncertainties of helium and the radioactive element contents are given as 1σ in relative error %. When radioactive element concentrations were close to, or below, the detection limit, uncertainties were not calculated (n.c.). Uncertainty of the single-grain age is given as 1σ in Ma and it includes both the analytical uncertainty and the estimated uncertainty of the F T . Uncertainty of the sample average age is 1σ in Ma.