Phanerozoic Tectonic and Sedimentation History of the Arctic: Constraints From Deep‐Time Low‐Temperature Thermochronology Data of Ellesmere Island and Northwest Greenland

Rocks exposed along both sides of the Smith Sound in Ellesmere Island and NW Greenland record the tectono‐sedimentary evolution of the whole Phanerozoic, including two periods of mountain building—the Palaeozoic Ellesmerian Orogeny and the Palaeogene Eurekan Orogeny—and the formation of two major sedimentary basins, the Franklinian and the Sverdrup Basins. We used geo‐ and thermochronology and apatite chemistry data to unravel this evolution. Apatite fission track and (U‐Th)/He dates vary strongly from >600 to <100 Ma. We present internally consistent thermal history models, which allow to explain the data variations by a unitized exhumation and burial history. Our models suggest that the cratonic areas were buried beneath a several km‐thick succession of Franklinian Basin deposits. During the Ellesmerian Orogeny, the craton acted as sediment source, as also suggested by the composition of apatite and by U‐Pb ages of zircon contained in Devonian foreland sediments. The Ellesmerian foreland was buried by up to 4–5 km thick strata on top of the preserved sedimentary rocks. During the Triassic, the Sverdrup Basin strongly widened and extended at least ∼370 km further toward the east, as compared with previous reconstructions of the basin based on the preservation of Triassic deposits. Thermal history modeling suggests Late Cretaceous to early Cenozoic reheating, which may be caused by deposition associated with the Eurekan Orogeny and/or enhanced heat flow associated with continental breakup. Our data also show that low‐temperature thermochronology is not suitable for resolving potential strike‐slip movements along the Wegener Fault.

The mid-Neoproterozoic breakup of Rodinia resulted in the separation of Siberia and Laurentia, which in turn resulted in the formation of the Franklinian Basin, a large passive margin basin along the northern rim of Laurentia (Ernst et al., 2016;Evans et al., 2016). This Franklinian Basin was filled by Neoproterozoic to Devonian strata (Figure 4), reaching a thickness of up to 15 km (Embry et al., 2019). In Southeast Ellesmere Island and Northwest Greenland, the Franklinian deposits unconformably overly the cratonic rocks and the deposits of the Thule Basin. Anfinson et al. (2012) distinguished three major periods of clastic inflow to the Franklinian Basin; namely the Neoproterozoic to Cambrian, the Early Silurian to Early Devonian, and the Middle to Late Devonian ( Figure 4). The Middle to Late Devonian succession is also referred to as "Devonian clastic wedge" (Thorsteinsson & Tozer, 1970). Nearly the whole stratigraphic succession of the Franklinian Basin is exposed in our study area ( Figure 2).
During the Late Silurian, parts of the Franklinian Basin experienced local uplift (Boothia, Inglefield and Rens Fiord Uplifts, Trettin, 1991). Furthermore, during the Late Devonian to Early Carboniferous Ellesmerian Orogeny, the Franklinian strata were progressively deformed. Ellesmerian deformation resulted from docking of continental fragments along the northern margin of Laurentia: the composite Pearya terrane (Trettin, 1987), today exposed along the northern margin of Ellesmere Island, and potentially a subsequent continental landmass (Crockerland; Embry, 1993) that is presumably submerged under the Arctic Ocean today. Precise age, extent, and details on collisional tectonics, such as the relationship between Pearya and Crockerland, still remain enigmatic (e.g., Beauchamp et al., 2019;McClelland et al., 2023;Piepjohn et al., 2007). The Devonian clastic wedge represents the foreland of the advancing Ellesmerian front. During the Late Devonian to Early Carboniferous peak of the Ellesmerian Orogeny, it became deformed and uplifted (Embry et al., 2019), marking the end of Franklinian deposition. The front of Ellesmerian deformation is situated within the study area, west and north of the exposures of the Precambrian rocks (Piepjohn et al., 2015).
After Ellesmerian deformation ceased, the area underwent extension, which led to the development of the Sverdrup Basin from the Early Carboniferous onwards, situated in the core region of the former Ellesmerian Orogen (Thorsteinsson & Tozer, 1960). During much of the Carboniferous, deposition was carbonate-dominated, changing to chert-dominated in the Permian (Figure 4; Embry & Beauchamp, 2008;Embry et al., 2023). During the Triassic, the basin deepened and received a large influx of clastic material (Embry & Beauchamp, 2008). Clastic influx and basin subsidence became strongly reduced during the Jurassic, and again increased during the Early Cretaceous (Embry & Beauchamp, 2008).
During the Cretaceous, Greenland and the Canadian Arctic Archipelago started to separate, leading to continental extension and finally to sea floor spreading within Baffin Bay. Spreading may have started already in the Late Cretaceous at the time of magnetic chron 33 (Roest & Srivastava, 1989), that is, between 84 and 73.6 Ma, using  Gasser (2014) and Piepjohn et al. (2015). Stippled rectangle: Area shown in (b). (b) Simplified geological map of Ellesmere Island and NW Greenland, after Okulitch (1991), Neben et al. (2006), von Gosen et al. (2012), and Ehrhardt et al. (2016), along with locations of samples analyzed for this study (orange dots). The black line refers to the approximate location of the transect studied by Hansen et al. (2011), discussed in the text.  Okulitch (1991) and Harrison et al. (2016), showing sample locations along with AFT dates (black) and average AHe dates (blue, italic; cross-over ages are not taken into account). For errors and AHe single grain ages, we refer to Tables 1 and 2, and Tables  S4 and S5 in Supporting Information S1. Note that average AHe dates are meant as approximations. For all interpretations and thermal history inversions, we used individual single-grain data. Black dots refer to locations of samples studied by Grist and Zentilli (2005). trace. Moho depths and crustal velocities are continuous across the fault, arguing against a major plate boundary . On the other hand, dyke swarms are cut off across Nares Strait, arguing in favor of significant motions (Denyszyn et al., 2006), and seismic reflection data from the Northwater Basin south of Smith Sound show flower structures indicative for strike slip movements . Zircon U-Pb data tracing the border between Archean and Proterozoic rocks of the cratonic exposures of Greenland and on Ellesmere Island imply lateral movements of at least 215 km (Gilotti et al., 2018), whereas AFT ages show strong variations on both sides of the Nares Strait (Grist & Zentilli, 2005;Hansen et al., 2011). Subsequent to proposed Palaeocene movements along the Wegener Fault, the Nares Strait area was the site of Eocene Eurekan deformation, which may have overprinted some of the original Palaeocene structures, further complicating the recognition of a distinctive crustal boundary . Arne et al. (1998Arne et al. ( , 2002 combined AFT analyses with vitrinite reflectance data of surface outcrops and drill core samples of the Sverdrup Basin, west of our study area. These data mostly reflect the Cenozoic history, showing that rapid erosion associated with the inversion of the Sverdrup Basin initiated during the Palaeocene. Palaeogene cooling was also reported from the northern margin of Ellesmere Island (Pearya terrane), where  Dawes (2006), showing sample locations along with AFT dates (black) and average AHe dates (blue, italic; cross-over ages are not taken into account). For errors and AHe single grain ages, we refer to Tables 1 and 2, and Tables S4 and S5 in Supporting Information S1. Note that average AHe dates are meant as approximations. For all interpretations and thermal history inversions, we used individual single-grain data. Black dots refer to locations of samples studied by Grist and Zentilli (2005). (b) AFT ages plotted versus sample elevation. Upper part: Data from cratonic areas of Ellesmere Island (black symbols) and NW Greenland including the Thule Basin (purple symbols). Lower part: Data from sedimentary rocks of Ellesmere Island. The light-blue, light-red, and yellow-shaded areas indicate the deposition ages of the Eifelian Strathcona Fiord Formation, the Givetian to Frasnian Hecla Bay and Fram Formations, and the Early Triassic Bjorne Formation, respectively. Note that some sedimentary rocks were not fully thermally reset post-deposition.

Previous Thermochronology Data
10.1029/2022TC007579 6 of 27  Embry et al. (2023) and Dewing and Hadlari (2023). The red boxes mark the stratigraphic levels and the formation names sampled for this study. Note. AFT analyses were performed using the external detector method and the zeta calibration approach with a ζ of 346 ± 12 or 319 ± 9 (samples with *) for dosimeter glass CN5. ns/ρs-number and density of spontaneous tracks, ni/ρi-number and density of induced tracks; nd/ρd-number and density of tracks induced from dosimeter glass. P(X) 2 = Probability of obtaining X 2 value for n degrees of freedom, where n is (number of crystals)-1. Age data are highlighted in bold. MTL: mean track lengths, referring to non-c-axis corrected values, SD: standard deviation.    Figure 1). Their data show two dominant groups with late Proterozoic and late Palaeozoic AFT dates. Time-depth history models of these data suggest continuous erosion of the cratonic samples, partly for as long as 1,000 Ma, without intermittent periods of reburial. According to these models, erosion rates change through time, but do not follow any common spatial or structural pattern, so no joint erosion history for the whole study area was obtained. Instead, modeled erosion histories vary in space and time between the individual samples.

Material and Methods
The samples were collected during BGR expeditions CASE 11, CASE 12, CASE 16 and RV Polarstern cruise ARK-XXV/3. They are roughly orientated along two east-west profiles crossing the Wegener Fault, and along one north-south profile approximately along strike (Figures 1-3). The cratonic rocks mostly comprise granitoids and high-grade Archean and Palaeoproterozoic gneisses including one itabirite containing predominantly quartz, magnetite and grunerite (Table 1). Samples from the Thule Supergroup exposed in Greenland consisted of slightly metamorphosed clastic rocks. Furthermore, we analyzed one quartzitic sample from the Cambrian Dallas Bugt Formation, several sandstone samples from the Middle to Late Devonian Strathcona Fjord, Hecla Bay, and Fram Formations, and one sandstone sample each from the Early Triassic Bjorne Formation, the Maastrichtian to Palaeocene Mt Bell Formation and the late Palaeocene Mt Lawson Formation ( Figure 4).
All samples were studied by AFT and AHe thermochronology. Zircon U-Pb analyses were conducted on a subset of Devonian sedimentary rocks. AFT and AHe analyses are temperature-dependant radiogenic dating methods. In combination, they date thermal processes within the temperature range of ∼120° to 40°C (Wagner et al., 1989;Wolf et al., 1998), and thus exhumation (associated with cooling) and burial (associated with heating) of the shallow continental crust. Detailed time-temperature paths are extracted from the thermochronology data by thermal history inversions, where track shortening and He diffusion in apatite is simulated based on laboratory exper-  iments, extrapolated to geological time scales, and then compared to the data observed (e.g., Ketcham, 2005). Zircon U-Pb analysis dates high-temperature processes usually associated with rock formation. In clastic sediments, Zircon U-Pb dates reflect the age of the source area and thus the provenance of the sediment.
Preparation of heavy mineral fractions from the samples and separation of apatite grains as well as AFT and AHe analyses were carried out at the Department of Geosciences of the University of Bremen. For AFT analysis, we used the zeta calibration approach and the external detector method (Gleadow, 1981;Hurford & Green, 1983, respectively). AHe analysis was performed using the 3 He isotope dilution method. Analytical precision was better than 1% for most analyses. Measurements of U, Th, and Sm contents are required for calculating AHe dates, but we additionally exploit these analyses for provenance information by calculating the U : Th : Sm ratios of detrital apatites and comparing these to the U : Th : Sm ratios of apatite from potential source areas, such as the cratonic rocks exposed in Greenland and Ellesmere Island. U, Th and Sm contents were determined by ICP-MS analysis. Precision was mostly better than 2% and accuracy better than 5%. Analytical details for these measurements, and also for the other analyses, are provided in Supporting Information S1 (Tables S1-S7).
Thermal history modeling was performed using the software HeFTy Version 1.9.1 (Ketcham, 2005), which utilizes a Monte Carlo approach. Depending on the statistical fit, we calculated between 10,000 to 500,000 time-temperature paths for each model run. We included Dpar values as kinetic indicator (Donelick, 1993) and c-axis projection (Ketcham et al., 2007). We used the annealing model of Ketcham et al. (2007) and the diffusion models of Farley (2000) and of Flowers et al. (2009). According to the fit with the data observed, thermal histories were classified as "acceptable" (goodness of fit value >0.05) or "good" (goodness of fit value >0.5). Furthermore, mean paths weighted according to the goodness of fit-values were calculated. More information on the modeling strategy is provided in Section 5 and summarized in Supporting Information S1 (Table S1).
U-Pb dating of detrital zircon grains was performed on five Middle to Late Devonian sandstone samples from Ellesmere Island. Zircon separation from the residual heavy mineral fraction, mount preparation, cathodoluminescence imaging, and analysis of U-Th-Pb isotopes by laser ablation combined with inductively coupled plasma-mass spectrometry (LA-ICP-MS) techniques were carried out at the geochronology laboratories at the Goethe University Frankfurt, Germany (samples CXI-119, −120, −124, −125), and the Senckenberg Naturhistorische Sammlungen Dresden, Germany (sample CXI-126). Analytical details of the LA-ICP-MS method used in the two laboratories are described in Koglin et al. (2022;with references). Only analyses with a concordance of 100 ± 10% are considered for the interpretation.

Apatite Fission Track Thermochronology
Sixty eight samples were processed, of which we chose 34 samples for AFT analysis. Apatite yield and quality were low for most samples, so that we were not able to analyze the envisioned >20 grains and >100 track lengths for all samples (Table 1 and Table S4 in Supporting Information S1). In agreement with previous studies (Grist & Zentilli, 2005;Hansen et al., 2011), AFT dates show large variations (99-631 Ma) with no clear spatial or age-elevation trend (Figures 2 and 3). For the cratonic areas, the oldest AFT dates occur in the north of the study area (Figures 2 and 3).
AFT dates from the Middle to Late Devonian clastic wedge vary between ∼212 and 399 Ma (excluding the statistically poorly constrained date of C-XII-20), without clear age-elevation relationship (Figure 3b). AFT dates from sediments of the Middle Devonian Strathcona Fiord Formation are mostly younger than their stratigraphic ages (Figure 3b), suggesting thermal resetting of the AFT system post-deposition. By contrast, AFT dates of the stratigraphically younger Hecla Bay and Fram formations are often slightly older than their deposition ages (although overlapping within error limits), indicating that they did not experience full resetting post-deposition.
In addition to Devonian sediments, we dated one sample from the Early Triassic (AFT date younger than stratigraphic age, suggesting at least partly resetting of the AFT system post-deposition), one sample from the Maastrichtian to Palaeocene, and one sample from the late Palaeocene (Table 1). The latter two samples yielded AFT dates much older than their deposition age, indicating little thermal overprint after sedimentation.

Apatite (U-Th)/He Thermochronology
Altogether, 104 aliquots were dated for AHe thermochronology, 12 of these revealed ages older than their corresponding AFT dates within error limits of both dating methods (cross-over dates, Table 2 and Table S5 in Supporting Information S1). AHe dates seem to show a weak positive correlation with effective Uranium-contents (eU) of apatite ( Figure S1 in Supporting Information S1), particularly for eU < 60 ppm, suggesting that radiation damage influenced He diffusion kinetics in apatite (Shuster et al., 2006).
Most of the AHe dates are of Palaeozoic to early Mesozoic age (Table 2). A notable exception from the overall AHe age pattern is sample KP711, containing Late Cretaceous to Miocene dates. KP711 was collected directly south and in the footwall of the Parrish Glacier Thrust, and we assume that the relatively young dates reflect heating associated with fault movements. Unfortunately, sample quality did not allow for fission track dating, so that we could not apply thermal history modeling for obtaining more detailed time-temperature histories.
AHe dates from all Devonian sediments are younger than their corresponding stratigraphic ages. Combining AHe and AFT information thus shows that the Middle Devonian strata from the Strathcona Fiord Formation experienced heating to temperatures >∼120°C post-deposition for most samples (i.e., sufficient for fully resetting the AFT system), whereas most samples of the Late Devonian Fram Formation only experienced temperatures between ∼85 and 120°C, high enough to fully reset the AHe system, but too low for fully resetting the AFT system.

Chemical Composition of Apatite
Our data show that apatites from the cratonic rocks of Ellesmere Island and Greenland have high Sm/Th + U ratios ( Figure 5), a fact that was also described by Grist and Zentilli (2005). Apatites from the Proterozoic Thule Basin and the Cambrian Dallas Bugt Formation show similar compositions as the cratonic rocks. Middle Devonian sedimentary rocks also contain apatite with high Sm/Th + U ratios, but generally show higher variations, also containing apatites with high Th/U-Sm ratios. These Th-rich apatites became more abundant in the Late Devonian strata ( Figure 5). The (few) apatite data from Palaeogene strata show similar variations in terms of apatite composition as the Middle Devonian sedimentary rocks.

Zircon U-Pb Geochronology
U-Pb dating was applied to detrital zircon from strata of the Devonian clastic wedge (i.e., the early Middle Devonian Strathcona Fiord Formation, the late Middle Devonian Hecla Bay Formation, and the Late Devonian Fram Figure 5. Chemical composition of apatite from the Precambrian cratonic rocks exposed on Ellesmere Island and Northwest Greenland, as compared with clastic strata from the Proterozoic, Cambrian, Devonian and Palaeogene. Data are from Table 2. Formation; Figures 4 and 6). It revealed three main age groupings in all Devonian clastic wedge samples: (a) a dominant Mesoproterozoic to late Palaeoproterozoic group (∼2,000-1,000 Ma) with three pronounced subgroups of ∼2,000-1,850 Ma, 1,800-1,600 Ma, and ∼1,300-1,100 Ma; (b) an Archean group (∼3,000-2,500 Ma); (c) a latest Neoproterozoic to Palaeozoic group (∼700-400 Ma). This youngest group is only poorly present in most samples, but is more prominent in the Hecla Bay Formation (n = 9; rejecting the youngest age of 349 ± 117 Ma with Th/U = 0.06). The youngest grain in sample CXI-120 (Strathcona Fiord Formation) with a U-Pb age younger than the stratigraphic age (298 ± 10 Ma; Th/U = 1.1) was also rejected because it most likely represents a contamination. Details are given in Tables S6 and S7 in Supporting Information S1.
These results are similar to those obtained by Anfinson et al. (2012), who applied zircon U-Pb dating to the same stratigraphic layers of the Franklinan Basin as we did. Only the trend of increasing presence of the ∼700-400 Ma age group toward stratigraphically younger Devonian strata found by these authors is not clearly recorded in our samples. This can be due to the more southern position of our sample CXI-126 from the Fram Formation on Ellesmere Island compared to the samples from Anfinson et al. (2012). Furthermore, our sample yielded only 50 concordant ages (out of 120 analyses) with only two ages within this youngest group (644 and 595 Ma).

Provenance of Devonian (and Other) Clastic Strata
Knowledge on the provenance of the clastic strata deposited west of the cratonic area is important for better constraining the modeled thermal histories based on the new thermochronology data. Provenance information is derived from the U-Pb data of detrital zircons ( Figure 6) and from the comparison of apatite compositions from the cratonic areas of Greenland and Ellesmere Island with those of detrital apatite from different clastic strata ( Figure 5).

Anfinson et al. (2012) interpreted the uplifted foreland of the East Greenland Caledonian Mountains and the
Ellesmerian Mountains in the north resulting from the docking of the Pearya Terrane and Crockerland as the main sources for their detrital zircon data from Devonian strata. We generally agree with this interpretation, but also stress that the cratonic areas of Greenland and Ellesmere Island are composed of Archean and Palaeoproterozoic basement rocks and may thus provide an additional important source area, delivering -or contributing to -the early Palaeoproterozoic 2,000-1,850 Ma age group and the Archean age group contained in the Middle to Late Devonian sediments. This group is most pronounced in the Strathcona Fiord Formation, and its relative contribution decreases upsection in the Hecla Bay and Fram Formation, suggesting that (i) the cratonic areas of Ellesmere Island and Greenland were uplifting and eroding during the Ellesmerian Orogeny, and that (ii) erosion rates either slowed down or that other sources became more dominant during the course of the Ellesmerian Orogeny. This suggestion is in agreement with the apatite compositions of the Devonian rocks, containing Sm-rich apatites similar to the cratonic rocks in Middle Devonian strata, changing to more Th-rich compositions contained in Late Devonian rocks ( Figure 5). Likewise, apatite compositions suggest that the analyzed samples from the Thule Basin and from the Dallas Bugt Formation of the Franklinian Basin seem to be derived from erosion of the cratonic basement. The Palaeogene sediment sample from the Eureka Sound Formation of the Sverdrup Basin shows similar apatite compositions and a similar AFT age distribution as the Devonian strata, suggesting that it is predominantly sourced from recycling of Devonian sedimentary rocks.

General Approach
Time-temperature models from the study area published previously by Hansen et al. (2011) suggest different thermal histories for nearly each investigated sample. They all assume continuous erosion, in places for as long as 1,000 Ma, without any periods of burial and thus (re-)heating. Taking into account the tectonic history and particularly the sedimentation history of the study area, we consider such time-temperature paths as unrealistic.
The old AFT and AHe dates as well as the shortened mean track lengths together with high standard deviations (Tables 1 and 2) suggest that all samples have experienced a protracted residence within the shallow crust (<5 km). It needs to be kept in mind that for such a presumably complex and protracted low-temperature evolution, thermochronology-based thermal history inversions are non-unique, that is, that although the derived thermal histories are in agreement with the data observed, other thermal histories may also be supported by the same data. This non-uniqueness can be reduced by including independent geological constraints, by combining several thermochronometers (such as AFT and AHe), and by assuming that adjacent samples from the same tectonic entity should reveal similar or complementary patterns for modeled time-temperature paths (see, e.g., Flowers et al., 2015).

Constraints Integrated Into Thermal Histories
A recent study by Embry et al. (2019) provides a very high-resolution history of deposition and erosion for the Franklinian and the Sverdrup basins. We use this study for constraining our models, as described in the following sections. However, as thermochronology can only resolve "events" which are associated with significant temperature changes, we restrict the constraints to (what we believe are) large-scale trends of tectonics or sedimentation. These constraints are set over large temperature ranges, so that the full range of statistically possible solutions could be explored. All parameters and assumptions used for thermal history inversions are also summarized in Table S1 in Supporting Information S1, following suggestions of Flowers et al. (2015).
As a starting point, we chose 2,000 to 1,900 Ma and temperatures of ∼800°C, in agreement with Palaeoproterozoic zircon U-Pb dates and granulite-facies metamorphism (Frisch & Hunt, 1988;Gilotti et al., 2018). Presumably, the rocks then cooled relatively fast to near-surface depths, as suggested by the preservation of high-temperature mineral assemblages, and then remained at or close to the surface for a protracted time, as indicated by peneplain formation (Dawes, 2006). However, neither geological nor thermal data provide independent record for this time interval. After ∼1,270 Ma, sediments of the Thule Basin start being deposited on top of the cratonic peneplain (Dawes, 2006). We therefore assume that the present-day structural level of the basement was exposed and thus at or close to surface temperatures. The subsequent deposition of the Thule Basin strata resulted in an unknown but potentially substantial burial and heating of the underlying cratonic rocks, which is why we allow for heating between 1,270 and 800 Ma. Parts of the craton (including the Thule Basin) are directly overlain by Early Cambrian sediments (e.g., Peel et al., 1982;Thorsteinsson et al., 2009), so we set another constraint for temperatures at or close to the surface between 590 and 530 Ma.
Neither the original thickness of the early Palaeozoic sediments in the Franklinian Basin nor the duration of deposition on top of the craton is known from the geological record. Thus, we used again a large constraint allowing for heating between 550 and 430 Ma. After 430, the area first experienced local uplift (Inglefield Uplift; Trettin, 1991), and then presumably regional uplift associated with the Ellesmerian Orogeny, which is why we assume that the cratonic rocks resided at shallower crustal or surface levels between ∼430 and 320 Ma. Furthermore, the partly cratonic provenance of the Middle to Late Devonian strata (discussed in Section 5) suggest that the craton was eroded and was thus cooling at that time. During the Carboniferous and Permian, the Sverdrup Basin was relatively shallow and did not receive significant clastic influx (Embry & Beauchamp, 2008). We thus assume that the cratonic areas stayed at (near-)surface levels and were neither significantly exhumed nor buried.
During the Triassic, the Sverdrup Basin deepened, receiving increased clastic discharge (Embry & Beauchamp, 2008). Furthermore, the basin also widened, extending into the cratonic areas (Embry & Beauchamp, 2008). Only one Triassic sedimentary sample (CXII-06; table 1) could be dated, that suggests at least partial resetting of the AFT system post-deposition and thus a several km-thick overburden. Furthermore, the western side of the cratonic exposures on Ellesmere Island also partially shows relatively young, post-Triassic AFT dates. Hence, we tested the possibility of significant sediment deposition on top of the cratonic basement by including a constraint allowing for temperatures of 30-180°C between 250 and 220 Ma.
Jurassic and earliest Cretaceous deposits are lacking in the study area (Harrison et al., 2015;Thorsteinsson et al., 2009), and this time is generally described as a period of quiescence (Embry & Beauchamp, 2008). We thus assume non-deposition or minor erosion, and that the cratonic rocks resided in relatively shallow crustal levels. Deposition resumed during the Barremian, and (with a short interruption during the Cenomanian (Embry & Beauchamp, 2008), which we neglect for modeling), persisted into the Eocene. Marine-influenced Palaeogene strata (West et al., 1975) deposited on top of the cratonic rocks suggest that the area was subsiding and experiencing burial during the time of the Eurekan Orogeny, which is why we allow for heating between the Barremian (∼130 Ma) and the end of the Eurekan Orogeny in the latest Eocene. At some time between the early Oligocene and the present day, the craton has experienced uplift, because Palaeogene marine strata are today exposed at altitudes of ca. 250 masl (own field observations).
A large part of the samples from this study are derived from the Devonian clastic wedge, that is, the foreland of the Ellesmerian Orogen. Accordingly, we used slightly different constraints, as compared to the cratonic samples: the starting constraint is the deposition age, followed by burial between deposition time and the end of Devonian foreland basin deposition at ∼370-365 Ma (e.g., Dewing & Hadlari, 2023;Embry, 1988;Thorsteinsson & Tozer, 1970). In accordance with latest Famennian to Early Carboniferous uplift and deformation of the foreland (Trettin, 1991), we allowed for cooling between ∼370 and 325 Ma. Subsequent to the Ellesmerian Orogeny, we assume an analogue thermal evolution for the sediments as for the cratonic rocks. For Devonian sedimentary samples with AFT dates not fully reset post-deposition, we assume a pre-depositional thermal history analogue to that of the cratonic areas.
If no AHe data were obtained from a sample, we modeled the AFT data alone. If only AHe dates without corresponding AFT dates were available, we refrained from thermal history inversions. Figure 7 shows the mean paths of each model run, weighted according to the goodness-of-fit-values. Results of the single models, including envelopes containing all statistically good and acceptable solutions and inversion points, are shown in Supporting Information S1 ( Figure S2).

Treatment of Highly Dispersed AHe Data
We generally tried to integrate AFT data with all AHe dates obtained from the same sample, except for cross-over AHe dates (i.e., AHe dates that are within error limits older than their corresponding AFT dates). We always applied both, the diffusion models of Farley (2000) and of Flowers et al. (2009). For well-replicating AHe dates, thermal history inversions based on the Farley diffusion model yielded better statistical fits in terms of higher goodness-of-fit values. Stronger dispersed AHe dates with high kinetic variability are better reproduced by the Flowers et al. (2009) diffusion model, yielding more often time-temperature paths in good statistical agreement with the data observed (defined as showing a goodness-of-fit value >0.5). When both models yielded statistically good solutions, the obtained thermal histories were usually very similar to each other, as exemplary shown for sample GRÖ35 (Figure 7 and Figure S2 in Supporting Information S1).
In a first model run, we tried to model all available AHe dates simultaneously with each other and the corresponding AFT date. If this approach did not yield statistically good solutions, we integrated the AFT data individually with several AHe subgroups or single AHe dates of the same sample (Tables S1 and S2 in Supporting Information S1), and used the range of obtained weighted mean paths as proxy for estimating the uncertainty associated with the variability of diffusion properties of individual apatite grains (e.g., sample CXII-46 in Figure 7).

Burial Related to Phanerozoic Basin Formation
Based on the constraints outlined above, we obtained good statistical fits for all samples, suggesting that the variations in AFT and AHe dates can be explained solely by local differences of the sedimentary overburden through  Figure S2 and Table S2 in Supporting Information S1. Thick lines refer to model runs where at least one AHe date is integrated with AFT data, thin lines to models solely based on AFT data. Dashed time-temperature-paths were modeled using the diffusion model of Flowers et al. (2009), solid lines refer to the diffusion model of Farley (2000). Gray-shaded temperature range outlines the main sensitivity of the combined AFT and AHe systems. The squares are modeling constraints that allow including independent geological information, as explained in the text.
time. Uncertainties and alternative models to the ones suggested here are discussed at the end of this chapter. According to the models, Palaeozoic peak heating/burial of the cratonic rocks was achieved between 495 and 480 Ma for most samples (Figure 7). Assuming a conventional geothermal gradient of ∼25°C/km, the "Franklinian" sedimentary cover reached thicknesses between ∼2 and 3 km on top of the northern cratonic exposures, and between 4 and 5 km in the south and in Greenland (Table S3 in Supporting Information S1 and Figure 8a; note that monitoring sedimentary thicknesses >∼5 km is beyond the resolution of the AFT system).
Our models also show that the data of all samples are in agreement with substantial Triassic burial and thus sedimentation on top of the deformed Franklinian Basin as well as on top of the cratonic exposures (Figure 8a). For the majority of the samples, maximum Triassic heating/burial was reached between 242 and 237 Ma (Figure 7). Again for a geothermal gradient of ∼25°C/km, modeled overburden varies between ∼1 and 5 km, increasing from the north toward the southern part of the study area (Table S3 in Supporting Information S1 and Figure 8a), suggesting pre-existing relief (presumably inherited from the Ellesmerian Orogeny) and/or differential subsidence rates across the basin. The models imply that the Triassic Sverdrup Basin (or an unconnected (?) neighboring basin with contemporaneous sedimentation) extended at least ∼370 km further toward the east, as compared with previous reconstructions of the Sverdrup Basin based on the preservation of Triassic deposits (Embry & Beauchamp, 2008; see Figure 9).
Timing of the youngest reheating period is relatively poorly constrained. It coincides with Eocene Eurekan tectonics and the associated deposition of the Eureka Sound Formation. The models, however, are also in agreement with earlier, Late Cretaceous to Palaeocene heating, and for some samples, earlier heating even yields better statistical fits (e.g., C-XII-36, C-XII-46, GRÖ21, GRÖ36, GRÖ64; Figure S2 in Supporting Information S1). During the Late Cretaceous to Palaeocene, the area underwent lithospheric stretching, continental breakup between Greenland and Canada, and onset of sea floor spreading (Roest & Srivastava, 1989). These processes are frequently related to enhanced heat flow in the adjacent areas (e.g., Nirrengarten et al., 2020;Seiler et al., 2009). Hence, the modeled Late Cretaceous to early Cenozoic heating period may represent a combined effect of enhanced heat flow and deposition of the Eureka Sound Formation, or even two separate heating episodes not resolved by thermal history modeling. For better understanding the relation between Late Cretaceous to Palaeogene deposition history, heating, and heat flow variations, numerical basin models, preferentially also including vitrinite reflectance data, as, for example, applied by Dörr et al. (2019) would be required.
Maximum Cretaceous to Palaeogene overburden values, as illustrated in Figure 8a, are calculated for an enhanced geothermal gradient of ∼40°C/km (Table S3 in Supporting Information S1), but need to be viewed with caution, as the individual effects of enhanced heat flow and sediment deposition on the thermal history cannot be distinguished.
In any case, two conclusions regarding the Late Cretaceous to Cenozoic history can be drawn from the models: 1. The thermal effect associated with Palaeogene deposition and/or tectonics on Ellesmere Island was low, not exceeding 50°C of heating or cooling for any of the samples, and with even lower values for most samples. Even though the presence of elevated Palaeogene marine-influenced deposits suggests uplift during or after the Palaeogene, this uplift was not balanced by substantial erosion. 2. Our models suggest the strongest thermal effects, but also the strongest variations of the Cretaceous to post-Cretaceous thermal history, for the Northwest Greenland coast. This may be explained by the position of the samples close to the Baffin Bay coast and thus the stronger thermal effect of continental breakup and incipient seafloor spreading. Furthermore, crustal stretching caused half graben formation along and offshore the NW Greenland continental margin (e.g., Whittaker et al., 1997), causing dissection of the continental margin that may explain the variations observed.

Deposition and Erosion Associated With the Ellesmerian Orogeny
Nearly half of the investigated samples were taken from the Devonian clastic wedge. Our data help to quantify the depth of the Devonian foreland basin in Ellesmere Island (including its eroded part) as well as magnitude and timing of erosion related to Ellesmerian deformation.
For almost all samples, peak burial was reached between ∼372 and 379 Ma (Figure 7). Modeled maximum overburden was between ∼4 and 5 km for most samples (Figure 8b), and thus in the typical range of foreland basins, comparable, for example, to the sedimentary thickness of the Swiss Molasse Basin associated with the Alpine Orogeny (e.g., Pfiffner et al., 1997). Maximum burial was followed by relatively rapid cooling until 350 to 343 Ma (Figure 7), which presumably was related to erosion caused by deformation and uplift. Accordingly, the erosion rate for this time period was approximately 75 m/Ma, again comparable to the rate of post-orogenic erosion reported from the North Alpine foreland basin (Mazurek et al., 2006). According to our models, erosion of the Ellesmerian Orogen was thus most intense between the Frasnian and the Visean (∼379 and 343 Ma), lasting slightly longer than the Tournaisian age postulated as termination of the Ellesmerian Orogeny previously (Mayr et al., 1998;Piepjohn, 2000). Net erosion of the former orogenic foreland during this period was mostly around 2 km (Figure 8b).
The cratonic areas of Ellesmere Island and Greenland (including the Thule Basin) partially served as source areas for the clastic infill of the foreland, as suggested by our provenance data. According to our models, erosion rates of the craton during the time of foreland deposition, as well as during the peak of the Ellesmerian Orogeny were rather low, with less than 1 km of net erosion for most samples during the whole time of Ellesmerian foreland deposition and peak deformation. A notable exception from the overall low cratonic erosion rates is sample C-XII-25, which experienced erosion similar to the foreland during peak orogeny. This is probably due to its position close to the border between cratonic exposures and foreland basin ( Figure 2). All in all, however, subsidence and erosion related to the Ellesmerian Orogeny was focused in the area of the foreland basin whereas the adjacent cratonic areas were little affected.  (Figure 7). Note that for calculating sedimentary overburden, geothermal gradients had to be assumed. All calculations and assumptions are summarized in Table S2 Figure 10.  (2008), based on the preserved sedimentary record. Rose-shaded color indicates areas where the sedimentary cover was removed by erosion and reconstructed by thermochronology (this study). Pink circles refer to data points. Note that no data are available for the north-eastern part of the "extended" Sverdrup (or neighboring) Basin.  Hansen et al. (2011) used an AFT transect from Northwest Greenland to Southeast Ellesmere Island across the Smith Sound (Figure 1) to argue against the existence of the Wegener Fault. Their main argument was that the AFT dates form an "unbroken, harmonious pattern" of an "intact crustal block unaffected by lithospheric fracturing." Their AFT pattern comprises a symmetric age configuration, with early Mesozoic to late Palaeozoic AFT dates in the coastal areas of Ellesmere Island and Greenland, abruptly changing to "constant" early Palaeozoic to late Proterozoic dates about 50 km inland. Figure 10 shows our AFT and AHe dates projected onto two ∼east-west profile bands. Data points of the northern profile are situated between ∼77°55′N and 78°47′N, whereas the southern profile band is situated between ∼76°17′N and 77°35′N. The northern profile is situated within roughly the same area as the Hansen et al. (2011) transect (Figure 1), only that our Greenland samples were collected from slightly further south, and the Ellesmere Island side of our transect reaches further inland.

Implications for the Wegener Fault Controversy
Our profiles show the same age variations as the Hansen et al. (2011) transect, with AFT and AHe dates ranging from <200 Ma to >600 Ma. However, we cannot confirm the simple symmetric pattern across Smith Sound: (a) in the northern profile, old, early Palaeozoic AFT dates close to the coast and at elevations near sea level contradict the findings of Hansen et al. (2011); (b) particularly along the Ellesmere Island sides of both transects, a trend to younger AFT dates toward the inland disagrees with "constant old AFT dates" reported by Hansen et al. (2011).
Instead, our AFT dates from the southern profile may suggest slightly younger dates on the Greenland side as compared to the Ellesmere Island side ( Figure 10). However, age variations on Southeast Ellesmere Island and Northwest Greenland are generally high. Thus, finding systematic trends or differences between both areas along the profiles is difficult and may overstretch the data. We explain the age variations mostly by differential subsidence across the Franklinian and Sverdrup Basins through time. Both basins were very large (>1,000 km in length), and the Sverdrup Basin was characterized by fault-controlled half-graben formations (Harrison, 1995), which may have caused relatively small-scale differences in exhumation and subsidence. In such a setting, it seems questionable whether a lateral offset between ∼85 and 260 km, as proposed for the Wegener Fault, would be detected by low-temperature thermochronology, even if it was associated with a vertical component. Instead, we suggest that studies relying on high-temperature geochronology, such as the study by Gilotti et al. (2018), and comparisons of the metamorphic histories on both sides of Nares Strait/Smith Sound would be better suited to address the topic of the Wegener Fault.

Uncertainties and Alternative Models
Weighted mean time-temperature paths of all thermal history inversions are illustrated in Figure 8. However, as mentioned previously, apatite crystals separated from the samples varied in terms of quantity and quality, so that the quality of the input data for modeling also varies. For providing a guideline for the reliability of the individual models, we used a color-coded classification in Table S2 in Supporting Information S1.
Another topic that needs to be discussed is the (non-)uniqueness of the suggested models. For the Devonian sedimentary samples, rapid heating to temperatures ≥100°C, followed by relatively rapid cooling, is essentially required for reproducing the measured data. Also, some samples necessarily require post-Ellesmerian Mesozoic reheating to temperature ≥120°C for matching the data observed. Thus, modeled time-temperature histories do not allow many alternatives to the histories suggested for this study.
This is different for the cratonic samples, where other thermal histories than the ones suggested here are also in agreement with the data. For example, the modeled thermal history of sample C-XII-25 may suggest that the cratonic areas were more intensively involved into Ellesmerian Orogeny. As a test, we remodeled sample C-XII-36 and included a constraint forcing cooling between 375 and 345 Ma, analogue to sample C-XII-25 (see Figure S2 in Supporting Information S1). The resulting time-temperature paths were also in good statistical agreement with the data observed, but did not reveal significant changes in the order of magnitude of erosion associated with the Ellesmerian Orogeny.
The largest difference of our models as compared to previous interpretations of thermochronology data is that we suggest sediment deposition on top of the craton and thus reheating. Previous studies, by contrast, assume continuous cooling over several hundreds of million years. As a test, we also tried a "constant-cooling-model" for three selected cratonic samples (C-XII-13, C-XII-36, C-XII-37). For this, the only constraints we introduced were (a) resetting of the AFT system at any time between the early Mesoproterozoic and the Early Cambrian, (b) a position close to surface temperatures during the Early Cambrian, as remnants of Early Cambrian strata are preserved on top of the craton, and (c) reheating to temperatures >40°C and onset of continuous cooling at any time after the Cambrian.
Indeed, we obtained statistically good solutions for all samples. However, maximum Phanerozoic heating and onset of continuous cooling occurred at 490 Ma for sample C-XII-37; at 295 Ma for sample C-XII-36, and at 227 Ma for sample C-XII-13. Hansen et al. (2011) explained temporal variations of their modeled erosion histories by faulting. Faulting may explain different magnitudes of burial and exhumation, even in samples situated close to each other. However, it cannot explain an entirely different timing of exhumation and burial periods of adjacent samples, which is why we consider the "continuous cooling-model" as unrealistic.

Conclusions
For this study, we used geochronologic, compositional, and thermochronologic data for investigating the sedimentary provenance and thermotectonic evolution of Southeast Ellesmere Island and Northwest Greenland. We derive the following conclusions: -U-Pb data of detrital zircon and mineral compositions of detrital apatites are in agreement with provenance of part of the Middle Devonian and-to a lesser degree-Late Devonian clastic deposits of the Ellesmerian foreland basin from the cratonic exposures of SE Ellesmere Island and NW Greenland. -During Ellesmerian Orogeny, the present-day structural level of the foreland basin first experienced 4-5 km of burial, followed by ∼2 km of erosion. The cratonic areas experienced limited erosion during the Ellesmerian Orogeny (but, due to the large areal extent, may have nevertheless been an important source for clastic deposition in the foreland). -Modeled time-temperature histories based on the thermochronology data and independent geological constraints are in agreement with substantial burial of the cratonic areas of Ellesmere Island and Northwest Greenland (including the Thule Basin exposures) during the early Palaeozoic and during the Triassic, suggesting a much wider Triassic Sverdrup Basin than previously assumed (or, alternatively, the existence of a neighboring basin). -Our models also imply Late Cretaceous to Palaeocene heating, which may reflect enhanced heat flow associated with rifting and incipient spreading of Baffin Bay or sediment deposition associated with the Eurekan Orogeny. -Even though there seems to be a diffuse trend toward younger AFT dates on the Greenland side of Smith Sound, any systematic correlation is obscured by the strong age scatter on both sides of Smith Sound, so that higher-temperature methods seem to be more promising for addressing the question of the Wegener Fault.

Data Availability Statement
Details on the analytical procedures, the detailed U-Pb data, as well as further information on the thermochronology data, thermal history inversions, and calculations of overburden and erosion are available at https://doi. org/10.5281/zenodo.8120628.