Compilation of Apatite Fission-Track Data from the Northeast Atlantic Realm: A Jigsaw Puzzle with Missing Pieces

The northeast (NE) Atlantic is one of the best-studied realms of extensional tectonics in the world [1-6]. Decades of geological mapping, geophysical investigations, and hydrocarbon exploration have created a dense body of data and literature on the geological history of the region from the Archean to the present day (Figure 1). Despite such a large amount of available data, considerable debates prevail regarding first-order processes such as orogenesis, magmatism, rifting, and glaciation within the regional geological development, and the evolution of topography [7-12]. Understanding these large-scale geological processes is greatly improved by regional datasets that cover the entire realm [6, 13-15].

The study of the onshore evolution of the NE Atlantic has been greatly supported by the application of low-temperature thermochronology [16-19]. For decades, studies have incorporated a variety of methods (e.g., zircon and titanite/sphene fission-track analysis, zircon and apatite (U-Th-Sm)/He dating), however, few have been consistently applied across the whole NE Atlantic, while others, like apatite (U-Th-Sm)/He dating, is hampered by significant age dispersion making these data difficult to present and interpret [20]. Among this list of methods, only apatite fission-track analysis (AFT) has been widely used across the entire region and collectively covers much of the NE Atlantic (Figure 2). As such, over seventy AFT studies have been conducted along the NE Atlantic margins and hinterlands in the last 45 years producing >2000 AFT samples (Figure 2).

In this work, we present a compilation dataset of published AFT data from across the NE Atlantic, including Fennoscandia, the British Isles, East Greenland, and Svalbard. This work builds on a legacy of such similar reviews that helped to highlight the previous work in regions of interest and highlight unanswered questions [17, 19, 21]. The principal goal of this review is to provide an accessible overview of this dataset to nonspecialists who may wish to advance their understanding of the purposes of AFT in the region and the onshore histories that have been resolved from this substantial body of work. A review of this scale has never been completed for the NE Atlantic, providing the opportunity to identify and discuss the major geodynamic processes that have shaped the region. We do this by presenting the spatial distributions of AFT ages and mean track lengths (MTL) from across each area, attempting to resolve first-order controls on these data, contextualizing with interpretations from the literature, and highlighting under-investigated areas that could be targeted in the future to help further answer larger questions about the history of the NE Atlantic.

The general distribution of data varies greatly, consistent with the contrasting geological histories of each individual region. However, it is clear that first-order mechanisms have a strong influence on the overall distribution of AFT data. The multi-stage opening of the NE Atlantic is shown to have a principal control on the thermal evolution of the upper crust, however, inherited features following Paleozoic orogenesis, sedimentary burial, and localized volcanism have a formative influence on the observed age distribution. Collectively, this dataset facilitates a regionally holistic interpretation of the NE Atlantic’s onshore evolution and highlights areas where data are lacking.

Fission-track dating has been widely used across the globe for decades, though we recognize that some readers may not be fully aware of the analytical approach and data handling. As such, we provide an overview of the method covering the basic theory, sample preparation, data collection, data handling, and interpretation.

Fission tracks are produced by the spontaneous fission of numerous radioactive isotopes [22]. Nuclear fission creates two positively charged ions that repel one another passing through the surrounding mineral lattice, creating a damage trail [22-24]. Of the many isotopes that decay in this manner, only the decay of ²³⁸U creates a high enough density of tracks for use in fission-track analysis. Isotopes produced from decay vary (e.g., ⁹⁰Sr, ⁹⁵Nb, ⁹⁵Zr, ¹³¹I, and ¹³⁷Cs) and their abundance is so rare it is not possible to measure the concentration of these decay products directly, leaving the resultant fission track as the only measurable remnant of the fission event [22]. Furthermore, preservation of a track within the mineral lattice of apatite only occurs below ~120°C on geological timescales [22]. Above this temperature fission tracks will anneal rapidly, removing any evidence of the event, while below it the track anneals at a temperature-dependent rate within the Partial Annealing Zone (60–120°C). Finally, below ~60°C the annealing of fission tracks is negligible, even on geological timescales [22]. This annealing behavior makes AFT ages and lengths together a valuable tool in re-constructing the thermal history of rocks in the top 5 km of the upper crust [22].

Calculating a fission-track age requires the measurement of two key components: (i) the amount of fission decay that has occurred and (ii) the remaining abundance of ²³⁸U. This can be done using both multi-grain (population/subtraction method) and single-grain (external detector, re-etch/re-polish, laser ablation) methods [25, 26]. The external detector method has been the most widely used approach for decades, though the laser ablation method is a more recent alternative gaining wider use, primarily due to the wider range of data that can be collected and faster analytical turnover. These two approaches will be explained in greater detail, while other methodological options will not be due to their relatively rare application.

The external detector method involves etching the internal surfaces of polished apatite grains (Figure 3) to allow “spontaneous” tracks to be observed and counted. Subsequently, an external detector (mica) free of uranium is attached to each sample mount and irradiated with low-energy neutrons in a batch (Figure 3) (27). Irradiation stimulates fission decay within the remaining fraction of ²³⁵U and produces a set of “induced” tracks within the detector, representative of the remaining uranium concentration (Figure 3) (27). Low-energy neutrons must be used not to inadvertently induce fission decay from multiple sources, meaning the ²³⁸U concentration is derived from the remaining ²³⁵U content and determined through their near-constant ratio in nature [28]. Once both sets of tracks (spontaneous and induced) can be observed and counted the fission-track age equation can be applied:

where t is the AFT age, ρ_s is the spontaneous fission-track density, ρ_i is the induced fission-track density, ρ_d is the dosimeter track density (used to calibrate neutron fluence during irradiation from a known uniformly distributed amount of uranium), λ_d is the total decay constant for ²³⁸U, ζ is an individual calibration factor attributed to each fission-track observer (zeta calibration [29]); and g is a geometry factor for the grain (typically 0.5) [24].

In recent years, laser ablation fission-track analysis has become a popular alternative to the external detector method [26, 30]. This approach removes the requirement for an external mica detector and irradiation by measuring the ²³⁸U concentration using laser ablation ICP-MS (Figure 3) [30]. This approach has a number of advantages, including the ability to measure additional isotopic concentrations, lower analytical costs, and faster analytical turnover [31], though issues pertaining to laser spot size, location, and the handling of grain zonation remain [32].

This dating procedure follows the same sample preparation as the external detector method up to the point where mounts are etched. From here, spontaneous tracks are counted, either in an area similar to the laser spot size that will be used [32] or across the entire grain that is then ablated (33) (Figure 3). The laser procedure measures the ratio of ²³⁸U and ⁴³Ca and the single-grain age is calculated using:

where t is age, N_s is the number of spontaneous tracks counted, P is the ²³⁸U/⁴³Ca ratio, Ω is the area over which the tracks were counted, λ_d is the total decay constant for ²³⁸U and ζ_ICP is the “machine” zeta calibration calculated by measuring age standards during each session [30].

Both the external detector and laser ablation methods allow single-grain ages to be calculated. However, these ages are often accompanied by high errors (>30%), primarily due to the low number of tracks that are counted in each grain [32]. To overcome this, a sample age can be determined by combining measurements from multiple grains (~20 for bedrock samples). For both methods, sample ages are typically defined through two summary statistics, the Pooled Age, and the Central Age. For the external detector method, the Pooled Age is calculated using the total number of spontaneous and induced tracks from a sample and is used when single ages are derived from a homogenous population, inferred from the results of a chi-squared test [34]. For the laser ablation approach, a Pooled Age is calculated using the total number of spontaneous tracks counted and the sum of the counted area and measured ²³⁸U/⁴³Ca ratio from each grain [30]. The Central Age is a random effects model assuming all single-grains ages are from a normal distribution with a standard deviation defined by the sample’s age dispersion and is used when single ages are not derived from a homogenous population, also inferred from the results of a chi-squared test [32, 34].

To properly interpret the distribution of sample ages across the NE Atlantic, it is important to clarify what an AFT age represents. Unlike other diffusion-based thermochronological systems (e.g., U-Pb, ⁴⁰Ar/³⁹Ar, or (U-Th-Sm)/He), AFT ages should not be considered the time since a sample was last at a given “closure temperature” [35]. Instead, a sample’s AFT age (Pooled or Central) is a function of a rock’s thermal history within the partial annealing zone (60–120°C) [17].

In addition to an AFT age, the lengths of horizontal, confined fission tracks can be measured to infer thermal history information [36, 37]. Track length data are typically simplified to a single value, the MTL (Figure 4). A symmetrical length distribution with a relatively high MTL (~13–15 μm) is typical for rapid cooling from temperatures >120°C to near-surface temperatures (Figure 4(a)). An asymmetrical length distribution with an intermediate MTL (~11–13 μm) is typical for protracted cooling from temperatures >120°C to near-surface temperatures (Figure 4(b)). A bimodal track length distribution with a low MTL (~8–11 μm) indicates a thermal history with multiple episodes of heating and cooling, with the analyzed sample remaining in the partial annealing zone for a relatively long time (Figure 4(c)). These MTL ranges are essentially arbitrary and should not be treated as fixed, though, in this work, these MTL values will be used as an aid for discussing whether areas have experienced rapid, protracted, or more complex thermal histories.

In contemporary AFT literature, thermal history models are typically generated from the data [18, 38-40], using open-source software such as QTQt [41] or HeFTy [42]. These software employ empirical models of fission-track annealing behavior [43] to derive thermal histories from the initial metadata AFT data collected. This metadata typically includes measurements of individual apatite grains, track lengths, c-axis measurements, and compositional information that can be used to model multi-kinetic annealing behavior (e.g., Cl wt%, D_par; the etch pit diameter parallel to the c-axis). Unfortunately, the necessary meta-data required to construct thermal histories is rarely reported in the studies covered in this review, making any attempt to re-model this dataset highly limited or biased to specific localities. We will review thermal history modeling results from the literature for each region, though care must be taken as thermal history modeling approaches have varied significantly in the 45 years over which the data were collected.

Limited by the inability to produce thermal history models, we have chosen to employ two simplistic interpretative approaches: (i) boomerang plots and (ii) apparent exhumation estimates. The limitations of these approaches are outlined in the discussion.

Plots of AFT ages and MTL values, termed “boomerang plots” (Figure 5(b)), can be used to infer interpretations of an area’s thermal history, assuming the crustal block has experienced a similar thermal history [44]. Samples exhumed from different temperatures (i.e., depths) by a single cooling event will produce a concave shape reminiscent of a boomerang (37) (Figure 5). The younger ages in this plot are defined by samples that have been brought from temperatures ≥100°C and have, hence, experienced minimal track annealing during passage through the partial annealing zone (Figure 5). The central portion of the plot is defined by samples that initially resided between ~100°C and 80°C and have experienced the greatest proportion of annealing producing the lowest MTL values in the sequence (Figure 5). The older ages in the plot are defined by samples that were initially ~<80°C and have experienced only minimal annealing (Figure 5). Importantly, the age of younger samples, paired with intermediate to high MTL values, is indicative of the onset of a regional cooling/exhumation event (Figure 5), providing an interpretive value to the plot should this age correlate to events in the geological history [44]. Moreover, if no boomerang trend is present, an interpretation can still be derived as the region has likely experienced a thermal history that cannot be attributed to a single regional exhumation event [24].

The consistent inclusion of sample elevations in published data allows apparent exhumation rates to be calculated for given locations. As the upper crust is exhumed, samples at different depths will pass through the partial annealing zone at different times, creating a relationship between AFT age and sample elevation where older samples are found at high elevations and younger ages at lower elevations (Figure 6). The gradient of the slope created can be used to estimate an apparent exhumation rate (m/Myr), with higher gradients indicative of higher exhumation rates (Figure 6). It is common for age/elevation relationships to produce nonlinear trends, indicative of a change in exhumation rate (Figure 6), though defining changes in these trends can be subjective unless a very clear break in slope is apparent, and the number of samples is high enough to make this resolvable. Additionally, anomalous ages may be present in the elevation profiles that could heavily influence a linear regression model through the data points and create data spread that would produce coefficients of determination too low for consideration (<0.1; e.g., profile 21 in online Supplementary Data). Initial assessments of age/elevation profiles in this work found no clear and obvious breaks in slope and numerous profiles with anomalous ages (Supplementary Data), therefore, to allow for this spread in data we adopt a simplistic approach only using the highest sample and lowest sample from a profile to calculate an apparent exhumation rate (Figure 6). All elevation profiles were created using samples <10 km apart, except for profile 9 in Fennoscandia (Supplementary Data) where samples were <12 km apart, though these were treated as a single elevation profile in their original publication [16].

An overview of the dataset can be found in Table 1, while maps of all studies from which data are derived can be found in Figure 2. This is believed to be an up-to-date dataset at the time of submission.

The first-order control elevation has on an AFT sample can lead to varied AFT ages over small distances where elevation profiles or well data are displayed, something difficult to represent on maps. Previous compilations have limited datasets in which data are filtered by elevation to reduce the effect of elevation sampling [17]. Examination of our full dataset highlighted that elevation limiting similar to other work [17] (0 and 500 m) would remove >25% of all AFT ages and >27% of MTL values, hence, limiting the dataset by elevation was not performed as it was deemed as too restrictive. Interested readers are recommended to revisit the original publications before drawing sophisticated conclusions from any single sample in map view.

Distribution maps were produced by interpolating the data using the Inverse Distance Weighted (IDW) raster function from ArcMap 10.8.1, created using a variable search radius. The cell size of each raster was limited to 500 m, the smallest distance measured between any two samples to ensure multiple samples did not occupy a single cell. Data density across the NE Atlantic is highly inconsistent and rasters were clipped using a graphical buffer with a radius of 50 km from any datapoint to ensure areas of low sample density were not over-represented in finalized maps.

The geological history of the NE Atlantic is highly complex (Figure 1) and only a brief summary is provided here [6, 13, 15, 45]. The Archean and Proterozoic crystalline bedrock of the continental NE Atlantic amalgamated through a series of Proterozoic and Paleozoic orogenies, typically compiled into larger orogenic events for ease of explanation. During the Proterozoic, the Svecofennian (1.9 and 1.8 Ga) and Sveconorwegian–Grenvillian (1.1 and 0.9 Ga) orogenies occurred across modern-day Fennoscandia and Svalbard [46, 47], while the Trans-Hudson Orogeny (2.0 and 1.8 Ga) occurred across northwest Britain, Greenland, and North America [48]. Later, in the late Neoproterozoic, the Timanian Orogeny (610 and 560 Ma) was active across the Eurasian Arctic with fragments present across Svalbard, Western Fennoscandia, northern Greenland, and Arctic Canada, though the relationships between these fragments is a matter of some debate [49-52]. These separate regions were later amalgamated during the Paleozoic as part of the Caledonian Orogeny (470 and 360 Ma), and other related orogenies, linking Avalonia, Baltica, and Laurentia to create Laurussia [53-56]. Finally, the southernmost extent of the study area was also affected by the Variscan Orogeny, a major orogenic event that saw Laurussia, Gondwana, and peri-Gondwana terranes collide (420 and 275 Ma) [57-59].

Following the Caledonian Orogeny, the Caledonian Mountains experienced orogenic collapse following the Early Devonian, expressed in the Devonian sediments across the Caledonian foreland [60, 61] (Figure 1). Further extension continued during the assembly and break up of Pangea occurring in four periods of rifting: Late Carboniferous, Permian/Triassic, Late Jurassic/Early Cretaceous, and Late Cretaceous/Eocene [62]. Rifting is recorded by widespread sedimentary deposition, with Late Paleozoic and Mesozoic strata now observed sporadically across southern Scandinavia and covering much of Ireland, central and southern Britain, central East Greenland, and central and southern Svalbard [63-72]. An early Cenozoic transpressional phase occurred across northern Greenland and western Svalbard resulting in the Eurekan and West Spitsbergen orogenies (~65 Ma) [73-75], generating the Paleocene and Eocene strata that is now found across western Svalbard [72].

The breakup of Pangaea initiated the opening of the Central and Southern Atlantic throughout the Mesozoic and continued northwards, with ocean spreading initially active in the North–West Atlantic in the Paleocene, before essentially switching over to the NE Atlantic in the Eocene [6]. The transition from rifting to spreading witnessed the widespread volcanism of the North Atlantic Igneous Provence across much of East Greenland, Britain, and Ireland [12]. The extrusion of lava across East Greenland is believed to have started at ~61 Ma and continued to ~55 Ma producing a volcanic pile up to ~7 km thick [76], while volcanism across Britain and Ireland is believed to have started in ~63 Ma and continued to ~52 Ma [77, 78]. The timing of volcanism both pre-dates and post-dates the onset of spreading in the NE Atlantic, suggesting it may have assisted in this transition, though the origin of this volcanism is debated, with both extension-related convection and the interaction of a mantle plume being postulated [12, 79, 80].

The post-rift stage across the NE Atlantic began following the onset of ocean spreading in the early Eocene. Originally, three spreading ridges were active (Reykjanes, Ægir, and Mohns), while a fourth was initiated in the Late-Oligocene (Kolbeinsey), gradually replacing the Ægir on the opposing flank of Jan Mayan microcontinent [6, 13] (Figure 1).

Western Fennoscandia is dominated by Jurassic AFT ages, with Cretaceous ages locally prevalent in the Lofoten area (Figure 7) online supplementary Table S1. Samples from across much of the coastal Southern (62 and 63°N) and Northern Scandes (67 and 69°N) (Figure 1) provide Jurassic and Cretaceous ages (Figure 7) and smaller pockets of older Triassic ages present in specific areas (58 and 59°N; 64°N; 70–71°N; Figure 7(a)). Away from the Atlantic margin, a mixture of Paleozoic and Mesozoic ages cover much of Southern Sweden (Figure 7(a)), with the boundary between Triassic and Jurassic ages correlating well with the inferred trend of the Caledonian Orogeny’s deformation front (Figure 1). Further east, across much of central and northern Sweden, and northern Norway, ages are Paleozoic, with localized samples in Sweden and central Finland exhibiting Proterozoic ages and Triassic ages in northern Norway (Figure 7(a)). This distribution of ages across central Fennoscandia is well organized with the youngest ages observed <200 km away from the present-day coastline and increasing with distance from the Atlantic margin (Figure 8, sections B, C, and D). There are areas where this trend is not apparent, such as northern Fennoscandia and southern Sweden (Figure 7, section E), though these do not dramatically deviate from the broader trend. No Cenozoic AFT ages are present in surface samples from the Fennoscandian landmass.

Apparent exhumation rates are calculated from 20 elevation profiles from across Western Fennoscandia (Figure 9). Results display accelerated–protracted exhumation between the Permian and Early Cretaceous (304 and 105 Ma) with many elevation profiles incorporating AFT ages from the Mesozoic (Figure 9). Apparent exhumation rate varies between 72 and 0.3 m/Myr, with a mean of 25 ± 19 (1σ) m/Myr.

MTL across Fennoscandia generally decrease with distance from the Atlantic margin (Figure 7(b)). High-to-moderate MTLs are present along the Atlantic coastline (12 and 14 µm), while shorter MTLs are observed along the eastern coast of Sweden and Finland (9 and 13 µm). Localized groups of samples in southern (60 and 62°N) and central Norway (65°N) show slightly lower MTLs (11 and 13 µm) compared with the remaining values across the southern Scandes. In the absence of Cenozoic ages, intermediate MTL values in samples of Mesozoic age indicate limited amounts of cooling and exhumation during the Cenozoic (Figure 7).

Boomerang plots show a range of trends (Figure 8). The most northern transect is dominated by older ages (>200 Ma) with intermediate to high MTL values (13 and 15 µm) with no obvious trend present, likely due to the low sample density in the region (Figure 8(a)). Further south, the distribution is comparable to the younger and central portion of a boomerang plot (Figure 8(b)). A large population of younger AFT ages cluster on the left side of the plot (322 and 71 Ma), with the youngest sample paired with an intermediate/high MTL value (13.13 µm). Across central Fennoscandia, a clear boomerang plot is present with the expected trend visible across the younger, central, and older portions of the plot, with the youngest sample (71 Ma) paired with an intermediate MTL value (12.44 µm; Figure 8(c)). Data from the Southern Scandes shows ages decreasing with increasing MTL value, except for two outliers with young ages and small MTL values (Figure 8(d)). The most southern region of Fennoscandia shows no obvious trend with a cluster of Mesozoic AFT ages and a small number of Paleozoic ages (351 and 117 Ma) paired with intermediate MTL values (13.7 and 12.4 µm; Figure 8(e)).

From regional age trends alone, it may be inferred that Mesozoic rifting across western Fennoscandia is the principal cause for the distribution of AFT ages. The multiple phases of rifting in the Mesozoic likely caused rock uplift and exhumation of the western margin, while the hinterland remained largely unaffected (Figure 7(a)). This is evident in apparent exhumation rates, which all highlight a period of accelerated or protracted exhumation during this time (Figure 9), though these data are clearly biased due to sampling locations (Figure 9). Mesozoic rifting as the main driver of exhumation is further supported by the distribution of MTLs across the region which are consistent with either accelerated or protracted cooling through the partial annealing zone along the continental margin and more protracted and episodic thermal histories toward the cratonic center of Baltica (Figure 7(b)). Boomerang plots suggest a major cooling event in the Mesozoic has occurred, bringing samples from a range of initial temperatures to the surface (Figure 8). This is most clearly visible in central Fennoscandia (Figure 8(c)) where a typical boomerang trend is evident. Surrounding areas show a similar trend, but do not complete a full boomerang trend (Figure 8(b) and (d)) or lack the sample density to provide meaningful interpretations (Figure 8(a) and (e)). Though lacking in samples, these other areas do appear to fit with the broad trend present in central Fennoscandia and a more regional interpretation of major cooling between ~280 and 130 Ma is possible (Figure 9).

These broad interpretations based on the regional dataset are consistent with those from many studies across Fennoscandia. Many authors conclude AFT ages are derived either from rock uplift and erosion during NE Atlantic rifting or extended residence in the partial annealing zone east of the Scandes [81-86]. In Southern Norway, AFT ages and track length distributions were used to infer a significant phase of Permo-Triassic exhumation due to rock uplift and protracted cooling throughout the Cretaceous and Cenozoic [82]. Across the North Scandes, the trend of AFT ages and thermal history modeling suggests greater amounts of denudation occurred along the coastline, likely during the Cretaceous-Paleocene [83]. Samples collected in transects across Sweden, from the Eastern Scandes to the Baltic Sea clearly show ages decreasing toward the Scandes, while thermal history models suggest rapid exhumation in the Mesozoic (250 and 150 Ma) in the west and long-term burial and near-surface residence in the east [84].

Though the multi-stage opening of the NE Atlantic shows a clear first-order signal across western Fennoscandia, localized studies of coastal fault blocks across Scandinavia do imply a more multifaceted exhumation history, highlighting fault reactivation in the Cenozoic [38, 87, 88]. The difference in AFT ages across major fault systems in the Lofoten and Vesterålen Archipelago (Figure 1) appears to suggest Cenozoic activity along major rift-related normal faults [38], with cooling present in thermal history modeling inferred as an erosion response to further uplift in the Miocene [89]. Additionally, multiple studies present thermal history modeling results incorporating later periods of cooling during the late Cenozoic not resolvable in this dataset [16, 90-93]. Thermal history models from the southern Scandes highlight a major cooling phase that begins at 30 Ma, postulated to result from thermal erosion of the lower lithosphere by mantle convection and intraplate compression [16]. Late Cenozoic cooling is also observed in thermal histories from the Northern Scandes, where the mechanism inferred to increase exhumation of the surface is tentatively suggested to be intraplate stress deflection caused by spreading ridge forces [91]. These late Cenozoic cooling phases have also informed other work that incorporated landscape analysis and suggests the Southern Scandes and Southern Sweden experienced multiple periods of uplift in both the early Miocene and Pliocene [94, 95]. This uplift of the region is thought to have been driven by post-rift “doming” of the lithosphere [16], though a competing interpretation suggests the data could also be explained through the protracted exhumation of prerift Caledonian topography [96].

Older Paleozoic/Proterozoic ages are present further away from the Atlantic margin (Figure 7(a)) implying the bedrock here has been near surface since the Paleozoic and was unaffected by Mesozoic rifting in the west. The area has been interpreted to have been buried by Sveconorwegian and Caledonian foreland sediments stretching >600 km across modern-day Sweden and Finland [90, 97-99]. AFT data from central and southern Sweden suggest much of the area was covered by >2.5 to 1 km of sedimentary cover in the Upper Paleozoic that remained in place up until the Cenozoic [97, 98]. Though this body of work has helped to better understand the stability of the cratonic setting, detailed interpretations and thermal history modeling of these much older samples may be problematic. The quality of AFT data in such old rocks is believed to be prone to differing annealing behavior induced by chemical variation and radiation damage [100]. Ongoing radiation damage in apatite is suggested to enhance the annealing process, especially during long-term residence in the upper crust (>100 Myr), creating younger AFT ages in sample from across the Baltic Shield [100]. This effect is also thought to be reflected in AFT ages from cratonic rocks across Canada, where an increasing uranium concentration in apatite grains appears to produce younger AFT ages [101]. This interpretation has been challenged, as it is thought that the chlorine content of the samples on the Baltic Shield may have influenced the trends in ages, instead of radiation damage [102].

Along the Atlantic coastlines of Britain and Ireland, AFT ages are predominantly Mesozoic in age (Figure 10(a)) online supplementary Table S1, though much more varied compared to Fennoscandia (Figure 7(a)). Notably, across the island of Britain, a large band of Cretaceous ages between 54 and 55.5°N, including localized Paleogene ages, separates areas of older Jurassic, Triassic, and Palaeozoic ages to both the north and south (Figure 10(a)). Paleozoic ages are present in northern Britain (56, 58°N) and across the Southern Midlands (52.5°N), while a small, localized patch of Paleogene ages is found around the Cenozoic volcanic centers of northwest Scotland (57°N) (Figure 10(a)). The island of Ireland is dominated by Jurassic ages (64%), except for a handful of localized Cretaceous samples (and one Paleogene sample) on the eastern coastline, and two bands of Triassic, Paleozoic, and Proterozoic samples (northeast coastline and 10 km inland of eastern coast) (Figure 10(a)). The ages across the northeast coastline of Ireland are in stark contrast to samples on the British margin of the Irish Sea, which are younger and predominantly Cretaceous and Paleogene in age.

Northern and southern transects show ages slightly increasing with distance from the Atlantic margin (Figure 11(a) and (c)), apart from the younger Paleocene ages around 57°N. The central transect B across Britain and Ireland appear greatly influenced by the Irish Sea with older ages closest and furthest away from the Atlantic margin, while the region in between exhibits much younger ages indicative of more episodic thermal histories related to the sedimentary geology of central Britain (Figure 11).

Apparent exhumation rates are calculated from 17 elevation profiles from across the British Isles (Figure 12). Results define two periods of accelerated and protracted exhumation between the Permian and Early Cretaceous (280 and 111 Ma) and Late Cretaceous (78 and 47 Ma) (Figure 12). The first period of exhumation is very similar to the timing of accelerated exhumation across Fennoscandia (Figure 9) with many of the AFT ages from profiles being Mesozoic. The younger period of exhumation incorporates profiles with AFT ages that span across the Mesozoic–Cenozoic boundary implying an accelerated phase of exhumation in the early Cenozoic (Figure 12). Apparent exhumation rate varies between 128 and 4 m/Myr, with a mean of 38 ± 34 (1σ) m/Myr for the earlier Mesozoic period and 41–12 m/Myr, with a mean of 24 ± 12 (1σ) m/Myr for the later “Cenozoic” period.

The distribution of MTL values across Britain and Ireland is similarly complex (Figure 11(b)). Broadly, higher MTLs are located across central/southeast Ireland (52 and 53.5°N) and central west Britain (55°N; 12 and 14 µm). Moderate MTL values are found across northern and western Ireland, and across much of Britain (54 and 58.5°N, 52–53.5°N; 11 and 13 µm). Shorter MTL values are generally confined to Mesozoic samples from central Britain, older Paleozoic/Triassic samples from Britain, and a single sample in NE Ireland (10 and 12 µm).

Plots of AFT age against MTL values show a broad range of distributions. The northern transect across northern Britain shows AFT ages decreasing while MTL increases, with two age populations evident (353.6 and 169.9 Ma; 67.8–47.1 Ma; Figure 11(a)). The central transect has a wide distribution of ages and MTLs, though a tentative boomerang trend could be resolved if a number of outliers are ignored, notably samples with ages between 180 and 150 Ma paired with higher MTL values (>13 µm; Figure 11(b)). The southern transect across Ireland and Southern Britain shows no clear trend only one major age population is present (275 and 112.7 Ma; Figure 11(c)).

The regional distribution of AFT ages and MTLs across Britain and Ireland is highly complex and appears to be controlled by a wide array of geological processes (Figure 1). The predominantly Triassic/Jurassic and Paleozoic ages from across northern Britain and Ireland consistently correlate with the crystalline bedrock and Paleozoic sediments of the region, suggesting these ages likely result from exhumation linked to rock uplift during multi-stage rifting, similar to western Fennoscandia (Figure 7(a)). This is captured in the earlier Mesozoic exhumation phase evident from apparent exhumation rates across both Britain and Ireland (Figure 12) and in boomerang plots, where older age populations with primarily intermediate and higher MTL values (>12 µm) are dominant and no obvious boomerang trend is present (Figure 11(a) and (c)). These age populations are similar to the “younger” age populations in plots from Fennoscandia (Figure 11), suggesting they may represent samples brought to surface from higher temperatures (>120°C) during a major exhumation event in the Mesozoic (e.g., rifting).

Long-lived protracted or accelerated Mesozoic rift-related exhumation of the landscape is consistent with interpretations from a number of studies from the region [103-107]. AFT dating and thermal history models from Devonian sandstones across the North Britain inferred a protracted cooling history from peak temperatures in the Late Paleozoic–Early Mesozoic [103]. Paired with apatite (U-Th)/He data, thermal history models from across NW Britain resolved multiple phases of cooling throughout the Paleozoic and Mesozoic attributed to the collapse of the Caledonian Orogeny and onset of spreading in the Atlantic [107]. Further work from Ireland and western Britain infers late Permian to mid-Jurassic cooling in thermal history models is rift-related exhumation [106], while similar work from western Ireland resolved a significant period of cooling, suggested to result from ~2.5 km rift-related exhumation, occurred in the late Jurassic–Early Cretaceous [105].

Much of the remaining areas of Britain and Ireland, dominated by younger Cretaceous–Paleogene AFT ages, are underlain by Paleozoic and Mesozoic sediments (Figure 1), implying these data reflect burial and exhumation during the Cenozoic. This correlates to the later period of “Cenozoic” exhumation evident in apparent exhumation rates from profiles across Central Britain (Figure 12). Cenozoic cooling is apparent in various forms from thermal history models in the literature that incorporate sedimentary and crystalline bedrock samples across the region, highlighting significant denudation occurred in multiple phases throughout the Cenozoic [108-114]. Thermal histories derived from samples of Carboniferous and Triassic sandstones in Central Britain suggest greatest extent of burial occurred in the Early Paleocene followed by 1–2.2 km of denudation during two phases of uplift in the early and middle Cenozoic [108]. Across Ireland, integration of AFT data with vitrinite reflectance data resolved multiple phases of cooling and heating, with notable cooling phases in both the early and late Cenozoic [112]. Moreover, modeling of elevation profiles from locations across both Britain and Ireland concluded denudation in the Paleocene varied between 1 and 2.5 km, focused around the Irish Sea, waning to the south and west [40]. In NW Britain, a rapid phase of early Cenozoic denudation is inferred from modeling of AFT and apatite (U-Th)/He data [115], while other work estimates the total amount of Cenozoic exhumation here to be up to 2.5 km [116].

AFT ages across East Greenland range dramatically from Proterozoic to Neogene online supplementary Table S3. The southernmost coastline is dominated by Triassic ages, with a small group of Jurassic ages found at the coast and a single Paleogene sample found in the far south (62°N; Figure 13(a)). Further north (66°N), a transect of ages perpendicular to the coastline shows AFT age increase with distance from the coast, ranging from the Cretaceous to Proterozoic over ~100 km (Figure 13(a)). Around Kangerlussuaq Fjord (68°N) (Figure 1) ages range between Mesozoic in the west and Paleogene and Neogene in the east, with the majority of Neogene samples found along the coastline. In central East Greenland, a localized zone of Cenozoic ages is present (72°N, Jameson land, Traill Ø), surrounded by predominately Mesozoic samples. Further north, the coastline is dominated by Paleozoic/Triassic ages between 75 and 81°N, while across the northern margin of Greenland, Paleogene, and Cretaceous ages along the coast are separated from Jurassic, Triassic, and Paleozoic ages across the Trolle Land/Harder Fjord fault zones (Figures 1 and 13(a)).

The distribution of ages across East Greenland appears to broadly show AFT ages increasing with distance from the coastline in the NE and southern portions of the margin. The central portion of the margin (70 and 74°N) is more complex due to the presence of onshore rift basins but shows that samples derived from crystalline bedrock outside of these basins (west of the SAF/PDMF; Figure 1) are generally older (Figure 14(c)). Samples at 68°N are anomalously young (Paleogene and Neogene), compared to much of the remaining basement margin and are either sampled directly from Paleogene intrusive or have likely been heated by the Paleogene volcanic activity that is related to the massive flood basalts there [76] (Figure 14(d)). Samples from the northern coastline are widespread and show only a weak positive trend between age and distance from the coastline (Figure 14(a)).

Apparent exhumation rates are calculated from 26 elevation profiles from across East Greenland (Figure 15). Results define two periods of accelerated exhumation (>15 m/Myr) between 250–119 Ma and 77–15 Ma (Figure 15). The first period is again very similar to the timing of accelerated exhumation across Fennoscandia, Britain, and Ireland (Figure 15) with many of the AFT ages in profiles from the Mesozoic. The younger period appears similar to the “Cenozoic” phase of exhumation across Britain and Ireland as it also incorporates profiles with AFT ages that span across the Mesozoic–Cenozoic boundary (Figure 15). Apparent exhumation rates vary between 80 and 19 m/Myr, with a mean of 36 ± 19 (1σ) m/Myr for the earlier Mesozoic period and 278–17 m/Myr, with a mean of 99 ± 77 (1σ) m/Myr or the later Cenozoic period. Aside from these two periods of accelerated exhumation, a number of profiles produced lower exhumation rates (<15 m/Myr) spanning across the Mesozoic and Cenozoic, implying protracted exhumation of the landscape throughout these eras. From these profiles, exhumation rates range between 12 and 6 m/Myr with a mean of 9 ± 2 (1σ) m/Myr.

MTL across East Greenland vary considerably, though notable trends can be observed (Figure 13(b)). Longer MTLs (13 and 15 µm) are evident across the zone of Paleozoic/Triassic ages (75 and 81°N) and moderate track MTLs across the southern margin (11 and 14 µm). In the central part of the margin (70 and 73°N) moderate MTL (11 and 14 µm) in the samples of Cenozoic age are neighboring short MTL (9 and 11 µm) in samples of Mesozoic age (Jurassic/Cretaceous) (Figure 13).

Plots of AFT age against MTL show a range of trends across the margin (Figure 14). Across the northern coastline, age appears to decrease with increasing MTL, with the youngest AFT age (24.9 Ma) paired with a large MTL value (14.2 µm). Samples from the low-lying NE of Greenland show a trend of increasing age with higher MTL values, a nearly opposite trend to the northern coastline (Figure 14(b)). The youngest AFT age (93.9 Ma) is paired with an intermediate MTL value (13.4 µm), suggesting this does not represent the end of a cooling event. Across central East Greenland (Figure 14(c)) no obvious trend is visible, likely due to the complex geological history of the region and the sampling of both sedimentary and crystalline bedrock. When separated into samples east and west of the main basin fault (SAF/PDMF; Figure 1), crystalline bedrock samples west of the fault show no obvious trend, while samples east of the fault define a clear boomerang shape (Figure 14(c)). Further south, AFT age appears to decrease with increasing MTL values, similar to the northern coastline. In the SE of Greenland, plots of AFT age against MTL show no positive or negative trend, with the youngest AFT (172 Ma) paired with the shortest MTL (12.33 µm), though this is the area with the lowest sample density limiting the plot’s effectiveness (Figure 14(e)).

Collectively, the AFT data from East Greenland likely represent the product of numerous geological processes. Much of the ice-free crystalline metamorphic basement of East Greenland (e.g., southern, central, and NE), appears to have been affected by rifting in the Mesozoic and Cenozoic. This is consistent with studies that sampled the crystalline bedrock in these areas, interpreting protracted exhumation throughout the Paleozoic and Mesozoic as the region was rifting [117, 118]. AFT and apatite (U-Th)/He data from SE Greenland are consistent with slow cooling throughout much of the Paleozoic and exhumation in the Mesozoic (~250–200 Ma) (117). In the NE, the small number of younger AFT ages (<100 Ma) with intermediate or higher MTL values in this area (Figure 14(b)) implies much of the bedrock here was <120°C for an extended period of time and were likely exhumed during Mesozoic rifting, but not enough to bring samples from a greater depth to surface. Thermal histories from the region, informed by inverse geodynamic modeling, suggest the landscape likely experienced protracted slow exhumation through the Paleozoic and Mesozoic as rifting was focused in what is now the offshore region [118].

Across central East Greenland, apparent exhumation rates from elevation profiles taken from crystalline bedrock outside of central onshore sedimentary basins and the Kangerlussuaq volcanic center show both long-term protracted (<15 m/Myr) and accelerated (>15 m/Myr) exhumation in the Mesozoic (Figure 15). Crystalline bedrock samples from central Greenland (west of the PDMF; Figure 1) are primarily Mesozoic in age, though the spread of ages and MTL values suggest the multifaceted geological history of this portion of the margin makes it difficult to resolve a clear regional cooling history (Figure 14). AFT data paired with zircon and sphene fission-track data suggest the crystalline bedrock west of the PDMF has been uplifted >8 km since the Devonian, with an enhanced period of uplift after 55 Ma, thought to be driven by Cenozoic rifting and volcanic activity [119]. Additionally, studies that combine thermal history modeling and landscape analysis resolve two further regional uplift events in the late Cenozoic (Late Miocene and Pliocene) [120-122].

The remaining AFT data from central East Greenland appear to result from local basin formation. Similar to the sedimentary samples from across Britain and Ireland (Figure 10), the distribution of younger AFT ages and lower MTLs across the central onshore basins of East Greenland (east of the PDMF; Figure 1) is likely the result of reheating during burial, with multiple cooling events resolved from thermal history models: Early Jurassic–mid-Cretaceous, Neogene, Eocene–Oligocene, Oligocene, and Miocene [121-125]. The distribution of AFT ages and MTL values in these basins define a boomerang shape, consistent with a major exhumation event (Figure 14(c)). The youngest ages of the samples (20 and 9 Ma) are paired with a wide range of MTL values (14.7 and 10.6 µm) implying a wide range of thermal histories in the late Cenozoic (125) (Figure 14(c)). AFT data incorporated into basin history models help to resolve a peak burial in the Paleocene followed by 2–3 km of exhumation derived from 1 km of tectonic uplift, though a definitive mechanism could not be resolved [126]. Models from samples collected from basin sediment ranging from Devonian to Cretaceous isolated two periods of cooling in the Cenozoic (40 and 30 Ma; 10–5 Ma) [123]. The first is thought to result from exhumation or hydrothermal fluids in the basin induced by rift-related volcanism, while the latter is postulated to have been caused by uplift-induced exhumation related to spreading reorientation in the NE Atlantic [123, 125].

Finally, the much younger Paleogene/Neogene samples surrounding Kangerlussuaq Fjord (68°N) are attributable to the volcanic centers in the region either representing the cooling from emplacement or thermal overprinting [76, 127, 128]. The increase in temperature throughout the crust during the emplacement of magma appears to have overprinted any previous thermal history (present in gneiss samples >10 km away from the main intrusion) [76]. Additionally, a number of samples in this area were specifically sampled from intrusive volcanic units and their AFT age are likely derived from the conductive cooling of these instructions following emplacement [76, 127].

AFT ages decrease from north to south, with northern Svalbard characterized by predominantly Cretaceous ages and southern Svalbard dominated by Paleogene ages (Figure 16) online supplementary Table S4. Localized Jurassic ages are found around 80°N near the northern Arctic margin. Samples from Bear Island (Bjørnøya; Figure 1) to the South show a large age range from Paleogene to Triassic over a small geographical spread (25 km), with younger ages found on the west side of the island (Figure 16).

The distribution of MTL is highly variable across Svalbard and data are absent from Bear Island. Across northern Svalbard MTL values broadly increase eastward, with localized low values 79.5°N, while in the south shorter lengths are apparent, coupled with many of the Paleogene ages in the region (77.5 and 78.5°N).

The distribution of AFT ages and MTLs suggests Svalbard can be separated into 2 distinct tectonic domains. West of the Billefjorden Fault Zone (Figure 1), the West Spitzbergen Orogeny is likely the principal cause for the young AFT ages with the more complex Neogene evolution being responsible for the differences in MTL along the western coastline. Shorter MTL values associated with Cretaceous ages across the northwest of Svalbard indicate only limited Cenozoic cooling and exhumation compared to the south (Figure 16). East of the Billefjorden Fault Zone (Figure 1), the northern coastline is dominated by active rifting phases in the Mesozoic. This is consistent with previous studies from across Svalbard with thermal history modeling of samples from the west of Svalbard suggesting two phases of cooling occurred in the Eocene-Oligocene and Miocene related to the Eurekan (West Spitsbergen) Orogeny [129, 130], while the northern margin of Svalbard is thought to have been primarily exhumed during the Jurassic and Cretaceous [39]. An anomalously young Paleogene sample inland of the Arctic margin is thought to result from localized exhumation during the Late Cretaceous–Early Paleocene related to the West Spitzbergen Orogeny [39]. Ages from Bear Island appear to record uplift of the island during the Cenozoic during the opening of the NE Atlantic [131], with the Triassic sample considered anomalous as it is largely defined by a single, very old, apatite grain [131].

The wealth of AFT data that has been collected across the NE Atlantic over four decades has monumentally improved our understanding of the realm’s onshore thermo-tectonic evolution. Each study has provided important interpretations into their specific localities; however, the goal of this review is to elucidate the key first-order processes that are resolvable from this composite dataset across the region and highlight areas of low sample density that require further study.

The distribution of AFT ages and MTL values across the NE Atlantic is highly complex, yet clearly appears to correlate well to three major first-order controls: (1) the multi-stage rifting history, (2) Cenozoic volcanism and orogenesis, and (3) the prerift history of the region. These are explored in detail in the subsequent sections.

There is ample evidence from this dataset that the greatest extent of exhumation across the NE Atlantic occurred during Mesozoic multi-stage rifting. This is evident in from the ages and MTL values alone which are predominantly Mesozoic (62%) and intermediate/high (12 and 15 µm), implying a major exhumation event occurred at this time (Figure 17). This is also reflected in many of the apparent exhumation rates from across Western Fennoscandia, Britain, Ireland, and East Greenland where a phase of accelerated or protracted exhumation is evident and coeval during the Mesozoic (Figures 9, 12 and 15). Moreover, this is supported by boomerang plots which highlight a major population of Mesozoic ages (~250–100 Ma) paired with intermediate to high MTL values (12 and 15 µm), inferring a regional cooling event coeval to the latter two stages of multi-stage rifting across the NE Atlantic in the Late Jurassic–Early Cretaceous and Late Cretaceous–Paleogene (Figures 8, 11 and 14). These Mesozoic ages, MTL values, and exhumation rates suggest the multi-stage rifting across the NE Atlantic is the primary mechanism controlling the distribution of AFT ages (Figure 18), an explanation provided in numerous studies from across the region [82, 83, 85, 86, 92, 107, 117, 118, 123]. Notably, the lack of Cenozoic ages in areas of exposed crystalline bedrock may imply rock uplift and exhumation was greatest during rifting and diminished as extension migrated to the offshore domain. This interpretation appears consistent across much of the NE Atlantic (Figure 18), though specific age trends can be found and are discussed for certain areas.

Across western Fennoscandia, many of the youngest AFT ages are either found at the coastline in the north or ~200 km inland in the south, while MTL values remain uniform, suggesting some complexity to the exhumation of the margin (Figure 8). In the north, this trend is likely indicative of rock uplift and exhumation along the rift margin during the Mesozoic, consistent with the expected ages from a typical “scarp retreat” model [19]. In the south, the trend is less clear, with the distribution of ages indicative of “downwarp” of the lithosphere, implying broad downward flexure during rifting, generating older ages along the coast and younger ages inland (Figure 8) (132). However, the proposed downwarp model fails to account for the isostatic response of the lithosphere [24] and cannot be seen as a viable explanation. Some authors have suggested this trend represents mid-Cenozoic domal uplift due to thermal erosion of the lower lithosphere by mantle convection and intraplate compression [16], though conclusive evidence of this process is lacking. Instead, this trend in the south may represent the protracted exhumation of ancient prerift topography underlain by a crustal root, leading portions of the surface underlain by thicker crust to generate younger ages due to greater levels of exhumation [96]. The influence of Fennoscandia’s heterogeneous crust on the topography of the region has been highlighted previously, suggesting isostasy and later mantle-driven dynamic vertical motions are the principal sources of southern and central Fennoscandia’s topography [133, 134]. Additionally, the distribution of AFT ages across major rift-related faults on Fennoscandia’s Atlantic margin has been shown to represent Cenozoic reactivation of major rift-related faults [38, 87, 88], thought to have occurred due to the tapering of the margin during extension flexure of the lithosphere [135]. Collectively, this implies these AFT data trends across Western Fennoscandia may be explained by exhumation during rifting in the north, and the effects of lithospheric heterogeneities in the south, with further abnormalities caused by fault reactivations.

Across central Britain, a band of younger Cretaceous and Paleogene ages is evident, paired with varied MTL values and accelerated Cenozoic exhumation rates (Figures 11, 12 and 18). These young AFT ages and the modeled variability in timing and extent of exhumation across Britain and Ireland in the Cenozoic have prompted debate about the mechanism required to drive exhumation during this time. This may be influenced by the Paleozoic/Mesozoic rifting of the North Sea, which played an important role in the exhumation of Britain, greatly influencing AFT ages across the region [109]. Moreover, other authors suggest variations in exhumation estimates are linked to uplift induced by the emplacement of the North Atlantic Igneous Province [40, 107, 113, 114, 116], while some have concluded that variations in crustal thermal properties and sedimentary blanketing may enhance exhumation estimates [136]. The former interpretation argues that an epeirogenic mechanism is required to exhume the region in the Cenozoic and not localized mechanisms such as fault inversion caused by far-field stresses [137]. This epeirogenic mechanism is linked to volcanism of the North Atlantic Igneous Province, where ~5 km of crustal underplating is believed to have formed in the Paleocene, underneath the modern Irish Sea [138]. The latter interpretation suggests the variation in AFT ages across Northern Britain may not result from variable exhumation patterns, but rather from the heat production from Caledonia granites and the insulating effect of a since removed layer of sedimentary rocks [136]. This debate and the location of these Cretaceous/Paleogene AFT ages imply the tectonic and exhumation history of the British Isles is more regionally heterogeneous than that of other margins across the NE Atlantic and warrants further study.

Across the margin of East Greenland, much of the crystalline bedrock appears to have been exhumed during Mesozoic rifting (Figure 18). The only deviation from this general trend is between 75 and 81°N, where old Paleozoic–Triassic AFT ages are paired with intermediate–high MTL values along the coastline, while much younger Jurassic and Cretaceous ages are found ~100 km inland, paired with intermediate MTL values throughout (~13 μm). This trend is similar to the Cretaceous ages found inland across SW Fennoscandia (Figure 7(a)) and may imply a similar complex thermo-tectonic history is present across NE Greenland. These samples are located within the NE Greenland Eclogite Province [139], where ultra-high-pressure metamorphism is thought to have occurred 365–350 Ma based on analysis of basement zircons [140]. This is significant as AFT samples <3 km from the same zircon-bearing rocks produce central ages <80 Ma younger [118], implying a rapid rise through the crust to their present position. Similar analysis is yet to be completed on areas where Cretaceous AFT ages are found, but a relatively young eclogite at surface today does suggest the crust here may have experienced a more complex history of exhumation than other portions of the margin, driven by rapid extensional tectonics during orogenic collapse of the Caledonides. Aside from samples related to this specific area, age trends from the remaining samples across NE Greenland appear much older than the conjugate NW Fennoscandia (Figure 18). This implies that the exhumation experienced across Fennoscandian margin is far greater than East Greenland and correlates well with the proposed model of these margins separating asymmetrically [3]. Therefore, the distribution of AFT ages on both of these margins may easily be explained simply as the result of asymmetrical rifting, further highlighting the control first-order processes on AFT data. Such inferences also suggest a signal of asymmetric rifting is present across other parts of these margins, such as SE Greenland, however, many of these areas lacks good data coverage for further extrapolation.

Cenozoic ages are limited across the NE Atlantic (19%; Figure 17). We find accelerated exhumation rates during this time are evident across Central and NW Britain and across East Greenland (Figures 12 and 15), though rates can mostly be attributable to local geological processes such as volcanism, rifting, and orogenesis. Cenozoic ages from across central East Greenland are limited to onshore rift basins and areas affected by Paleogene volcanism [76, 103, 124, 126] (13,18,13,18,Figures 13(a) and 18). In East Greenland, burial and volcanism during the younger Cenozoic ages have previously been identified as key mechanisms. Basin modeling has indicated that uplift in this region occurred in the post-rift stage, driven by spreading reorientation or the broader impacts of North Atlantic Igneous Province volcanism [126, 141] (Figure 13(a)).

Cenozoic samples across south-eastern Greenland and Britain are adjacent to volcanic centers of the North Atlantic Igneous Province, implying the volcanism may have had a broader role in these areas (e.g., increase in geothermal gradient, burial under lavas), and the dating of emplacement-related cooling of volcanic units [76, 110, 127, 128]. Boomerang plots from areas affected by Cenozoic volcanism highlight a population of Cenozoic ages with intermediate to high MTL values (13 and 16 µm) 11,14,Figures 11(a) and 14(d)), indicative of either dated volcanic features or reheated host-rock that has cooled post emplacement. Finally, the Cenozoic ages from across NE Greenland and western Svalbard are spatially consistent with the transpression of the Spitzbergen and Eurekan orogenies (Figures 1 and 18). Uplift of bedrock and the subsequent exhumation of buried sediments related to transpression are viable mechanisms for generating the younger ages in these areas [121, 129], though an additional phase of uplift and erosion is also inferred from thermal history models [121].

The remaining samples from across the NE Atlantic are either Paleozoic or Proterozoic in age (19%; Figure 17). These samples are primarily located across eastern Scandinavia (Figure 7(a)), northern and southern Britain, NE Ireland (Figure 10(a)), or sporadically across East Greenland (Figure 13(a)). These much older samples suggest these areas have experienced very little exhumation or burial under now absent overburden (Figure 18). Samples from across eastern Fennoscandia record the protracted exhumation of the Svecofennian and Caledonian foreland sediments, with thermal history modeling suggesting numerous periods of reburial and exhumation throughout the Paleozoic, Mesozoic, and Cenozoic [84, 90, 97, 99]. This is most clear by boomerang plots from Fennoscandia where Paleozoic and Proterozoic ages are paired with intermediate or low MTL values, indicative of long-term residence at low temperatures (<80°C), or burial of samples within the partial annealing zone (Figure 8(b)–(d)).

The remaining Paleozoic and Proterozoic samples from across Britain, Ireland, and East Greenland, are deemed to result from cratonic stability, protracted exhumation, or from an early Devonian rifting phase following the end of the Caledonian Orogeny [76, 106-108, 110, 118]. Once again, this is supported by boomerang plot of this area, where these ages are paired with lower MTL values, suggesting long-term residence within or below the partial annealing zone (Figure 8(b); Figure 11(d)). There remains difficulty in explaining some of these older samples, as studies across the region have rarely focused on the Paleozoic and Proterozoic histories of the landscape [99, 142, 143]. Moreover, AFT data from older cratonic rocks may be influenced by radiation damage and other effects [100, 144], though the significance of this effect on calculated ages is debated [102].

These regional first-order interpretations based on AFT age and MTL distributions provided here are simplistic yet consistently correlate well with the conclusions of many thermal history models in studies across the NE Atlantic [83, 86, 105, 117, 118, 129, 145]. Our maps, apparent exhumation rates, and boomerang plots correlate spatially and temporally with important first-order geological processes including rifting, orogenesis, volcanism, and the rock type dominant in an area (e.g., crystalline bedrock or sedimentary strata). However, these interpretations differ significantly from a variety of other studies, especially those that highlight multiple episodes of rock uplift and burial across the NE Atlantic during the post-rift stage [16, 95, 108, 114, 120, 122, 146]. This set of interpretations is heavily underpinned by thermal history models which resolve multiple periods of cooling and heating interpreted as exhumation and burial of the landscape [95, 147]. It is important we state that AFT age and MTL values alone cannot be used to determine the number of cooling or heating phases a sample has experienced and, therefore, it is understandable that these differences are apparent.

Although these data cannot provide a detailed thermal history of a rock, the spatial and temporal consistency of AFT and MTL values and their correlations with first-order processes imply these areas have likely experienced a similar regional thermal history associated with a shared geological history. There are localities where specific geological phenomena dominate the thermal history of the upper crust (e.g., Kangerlussuaq, Lofoten, Northern, and Western Svalbard), though large portions of the crystalline bedrock across Fennoscandia, East Greenland, and Northern Britain consistently present of Mesozoic AFT ages and intermediate MTL values. From the presentation of literature in previous sections, it is clear there exist two prominent hypotheses to explain this consistency of data distribution: (i) protracted exhumation of bedrock during extension [96] and (ii) a multi-phase history of rock burial, rock uplift, and erosion throughout the Mesozoic and Cenozoic [147]. It is unlikely that both explanations can co-exist in much of the NE Atlantic, therefore both hypotheses are evaluated below.

The first of these hypotheses appears consistent with our current understanding of plate tectonic theory, including how the lithosphere responds to passive margin formation [148]. The thermal effects of rifting can create surface uplift along rift flanks, though once ocean spreading has commenced surface uplift cannot continue [149]. Topography formed in this tectonic setting can be underlain by a thick buoyant crust formed during previous periods of collision tectonics that provides isostatic support and can preserve a landscape over 10–100 Myr [96]. Without rock uplift to drive elevated erosion rates, they may only slowly erode over time meaning an elevated landscape formed during rifting (or present prior to rifting) may remain elevated to some degree for >100 Myr, as shown from landscape evolution models [150, 151]. This hypothesis for topographic formation and preservation is apparent in AFT literature from across Fennoscandia, East Greenland, Britain, and Ireland [83, 86, 105, 117, 118, 129, 145], and foundational in the “ICE (Isostasy, Climate, and Erosion) Hypothesis” [96].

The latter of these hypotheses requires epeirogenic movement of the crust before and after rifting [147]. These movements allow the topography of the landscape to be regularly rejuvenated following erosion to near-sea level [147]. Thermal history models from studies reviewed in this work highlight such movement, while supporting evidence has been found in sonic velocity anomalies offshore [152], large sedimentary wedges deposited over short timescales [153], and interpretations of low relief high elevation landscapes [154]. This hypothesis for topographic rejuvenation is also evident in AFT literature from across Fennoscandia, East Greenland, Britain, and Ireland [16, 95, 108, 114, 120, 122, 146] and synthesized in a recent review [155]

Though evidence of epeirogenic movements across the NE Atlantic is widely documented, other work has consistently been critical of the methodologies used or found other mechanisms that can explain their formation without the need for episodic surface uplift. Critiques include a sonic velocity method’s reliance on variables that can dramatically affect the final result (e.g., effective stress, temperature, and mineralogy) [156], global climate shifts increasing offshore sediment yields [157], and glacial erosion creating low relief high elevation landscapes [158]. Moreover, a definitive mechanism that can explain these results is yet to be determined. There are a number of possible mechanisms including lithospheric folding [159, 160], underplating of various kinds [39, 155, 161-164], crustal flow [165], mantle diapirism [166, 167], and variations in mantle convection [168], though each has their own spatial/vertical limitations or remains an unproven theory [96, 169].

Either way, the distribution of AFT ages and MTL across the NE Atlantic cannot be used to conclusively dictate which hypothesis best represents the post-rift history of the region, but it does imply a broadly shared geological history is present. From these two prominent hypotheses, the former appears more suitable as it is underpinned by mechanisms that align with our understanding of plate tectonics. The latter is supported by a range of contestable evidence and the lack of a conclusive mechanism makes it hard to firmly support.

Collectively, the NE Atlantic remains one of the best-studied geological regions in the world with much of our understanding of passive margins being developed in this region [9-12, 96, 170, 171]. However, this vast jigsaw puzzle of rift zones, orogenies, terranes, and volcanic centers has missing pieces. The final goal of this study is to highlight areas where thermochronological data are sparse and where we encourage future studies to investigate. We accept significant barriers may exist to the production of future studies in these areas (e.g., access to bedrock geology and apatite poor lithologies), though feel they should still be highlighted as areas in need of coverage.

The first of these areas is western and southern Britain, where fission-track data are sparse, and the distribution of analyzed samples is irregular (Figure 10). This is particularly interesting as this margin of the Irish Sea encompasses elevated topography and predominantly Jurassic and Triassic fission-track ages, compared with the Cretaceous and Jurassic age of the adjacent Irish margin [40, 136]. As mentioned, this area has been a focal point for research concerning potential plume-related volcanism and its representation on the Earth’s surface [40, 136]. The lack of sampling across western Britain (e.g., Wales) means that only the northern portion of the Irish Sea has good data coverage and appears to produce Cretaceous ages on both flanks, consistent with higher levels of exhumation across the region (Figure 10). The western flank of the southern portion (eastern Ireland) has been well covered, showing older Triassic/Jurassic ages (Figure 10), while the British side is only defined by 20 samples, varying from Cretaceous to Triassic and lacking any clear pattern (Figure 10(a)). It may be that the northern Irish Sea saw greater surface uplift during volcanism leading to a greater level of exhumation, though only further study of the southern Irish Sea could evaluate this trend. Future studies here could help explain the source of topography across the region while also providing significant insight into the evolution of the Irish Sea and the potential role of plume-related volcanism in the area.

Second, the SE coast of Greenland is poorly covered, compared to the rest of the margin (Figure 13). This is likely due to the difficult terrain and lack of population centers in the region, though the area may be of interest to the study of the southern extent of the NE Atlantic and its relationship with the British and Irish regions. The trend in AFT data outlining asymmetric rifting across NE Greenland and NW Fennoscandia may be present across SE Greenland and the British Isles, though the complicated distribution of data across the British Isles makes this less obvious. The distribution of data across the British Isles is clearly more complicated than either of the margins in the north, primarily due to numerous other rift centers (e.g., Irish Sea, North Sea), but fully understand how this rift evolved can only be studied when data from both margins is apparent. Establishing any similarity or difference between AFT data across both the southern and northern extent of the NE Atlantic would further our understanding of how the entire rift developed and the apparent segmentation of the margin is evident.

Third, the southern coast of Svalbard remains wholly underexplored, with much of the current data located across the northern and western margins of the region. These southern areas are dominated by Proterozoic, Mesozoic, and Cenozoic bedrock, uplifted during the West Spitzbergen Orogeny, with the SE dominated by Mesozoic sediments which we recognize may prove unsuitable for apatite thermochronology. The Sørkapp–Hornsund basement high is horst structure found in the southern extreme of Svalbard and has a complex tectonic history that has included multiple phases of Paleozoic–Mesozoic rifting and transpression, and Cenozoic orogenesis [172]. The timing and extent of major tectonic events remain unclear [172, 173] and could be investigated further with the application of AFT analysis. Additionally, The Triassic sediments on Edgeøya have been shown to have been buried to a maximum of ~140°C, and AFT may provide great insight into the inferred Cretaceous or Cenozoic uplift of the region and any links to the evolution of the Barent Sea [174]. Svalbard has experienced a complex geological history and AFT analysis could be a vital tool in establishing the timing and magnitude of some of these major tectonic events.

Fourth, the eastern extent of Fennoscandia (Sweden, Finland, and Russia) remains underexplored, compared to the western portion of the region. The history of the region resolvable with low-temperature thermochronology is inherently complex, though more data from the region could greatly improve our understanding of the near-surface evolution of the area following the Caledonian Orogeny. Such work across the Canadian Shield has proved highly insightful in helping to understand how cratonic regions evolve over lengthy geological time scales, though debates still remain [101, 175-179].

Finally, though beyond the scope of this review, we wish to highlight that the NW Atlantic (Labrador Sea, Davies Strait, and Baffin Bay) remains a poorly explored area as a whole. AFT studies are present in the region [180-186]; however, there is a distinct lack of spatial coverage when compared to the NE Atlantic, undoubtedly due to access issues. This region was specifically not included in this work due to poor coverage and difficulty forming regional interpretations and correlations across these conjugate margins. Study of this region is important for a complete understanding how the whole North Atlantic formed and future studies here would provide great insight into the evolution of the realm. Prior to break up in the NE Atlantic, ocean spreading was underway in the NW Atlantic, causing Greenland to drift away from Laurentia for the first time in its history [187]. This appears to stop and transfer across to the NE Atlantic coeval to the Eurekan Orogeny that occurred across Northern Greenland and Canada [187]. Further study of the thermo-tectonic evolution of the area would be advantageous to best understand the correlation between these events and their effects on each margin’s evolution. Additionally, the topography of West Greenland and Eastern Canada (Baffin Island and Labrador) notably contrast north and south of the Davis Strait [188]. Further work on these margins could further highlight asymmetry across these margins and the effect this has on exhumation [189]. Moreover, as topography plays an important role in ice sheet development, understanding how asymmetry has shaped the landscape could help delineate the source of the Greenland Ice Sheet [190].

The central interpretation of this work, that AFT data distributions are controlled by first-order processes across the NE Atlantic, shows that future work could be used to answer major questions regarding the evolution of the North Atlantic. Filling in the gaps in the enormous jigsaw puzzle may be vital in helping us fully understanding how one of the best-studied geological regions in the world came to be.

This compilation of data from across the NE Atlantic does provide great insight into the wider evolution of the NE Atlantic realm, however, limitations are evident. These limitations are not wholly unique to this dataset, and many are inherent in the fission-track method itself. As mentioned, the elevation of a sample can play a significant role in calculated AFT ages and lengths and, as such, the presentation of data on two-dimensional maps does not allow for this effect to be accounted for. Typically, an uplifted portion of the rock column will erode in such a way that older AFT ages will be found at higher elevations. Elevation sampling is an important strategy for resolving vertical motions with low-temperature thermochronology [44], but elevation profile sampling can influence the presentation of ages in map view (Figure 19). This is evident from examples in Central East Greenland, where the steep topography allows for elevation profiles with an age span from Triassic to Neogene over ~1500 m of elevation (Figure 19). Such age distributions may create abnormal geometries or averaging values that occurred in the same cell (<500 m apart) during interpolation in the map, making it difficult to clearly define the boundary between these two geological periods in the region.

The use of boomerang plots in this work attempts to provide a simple way of resolving region thermal histories from the large area covered in this work. The density of samples in some locations means these plots cannot provide any meaningful information (e.g.,Figures 8(e) and 14(e)), limiting their use. Moreover, the complexity of some regions’ thermal histories shows that these plots simply cannot provide any meaning information about the regional thermal history (e.g.,Figures 8(d), 10(c) and 14(b)). However, as stated above, the absence of such trends does help to identify areas where a regional thermal history is not present, and further investigation is merited.

The use of apparent exhumation rates derived from elevation profiles provides insight into phases of exhumation across the NE Atlantic. However, these estimates are simplistic and spatially biased. Apparent exhumation rates cannot be treated as accurate representations of exhumation in each area. As noted above, nonlinear trends are common in elevation profiles, indicative of changes in exhumation rate, and our approach fails to capture these multiple phases. Moreover, these rates are biased to areas of elevation where elevation profiles were collected. This means much of the NE Atlantic remains ignored and far more complex exhumation histories are likely.

Many of the samples covered in this work were rarely presented with comprehensive metadata in their respective publications. Thermal history modeling requires the raw data collected during fission-track analysis to be compiled, and in most cases, these data were unavailable, massively limiting the scope of this work. Additionally, many samples were derived from onshore sedimentary outcrops and the stratigraphic ages of these rocks could provide an excellent record of when apatites analyzed were deposited and if these samples had experienced burial temperatures high enough to reset the AFT system (depositional age > AFT age). Unfortunately, the recording of stratigraphic ages is sporadic and rare throughout the literature, while the studies that do provide such information vary in their descriptions from defined specific age ranges to geological periods. As such, the data could not be included. Therefore, we wish to highlight the importance of a standardized reporting structure for all future AFT data. Recent publications have laid out reporting guidelines for fission-track data and thermal history modeling (if applicable) under the FAIR data principles (Findable, Accessible, Interoperable, and Reusable) [191, 192]. We fully support these reporting guidelines and emphasize that although responsibility falls primarily on the authors of any work including AFT data, it is imperative that reviewers and editors ensure they are followed.

Single-grain AFT analysis’s typically high error requires multiple grains to be dated from a sample to produce a sample age. Typically, analysts will count ~20 grains per bedrock sample to produce a robust age for a rock sample, though this can be difficult when apatite yield is low. Given the time span of data acquisition in this compilation, AFT ages have been derived using various methodologies [25, 26] and in case of the external detector method used number of grains counted per sample are not always given. Samples with grain information and ≥10 grains per sample account for 74% of the dataset and filtering of age data by quality would limit the dataset significantly, especially across Britain. Therefore, care should be taken while inspecting these maps, especially in areas characterized by only a single datapoint.

Finally, the MTL metric is the average of all track lengths; however, it has been shown that the annealing of fission tracks is anisotropic and dependent on composition [193]. Tracks at a high angle to the crystallographic c-axis will anneal at a higher rate than those at a lower angle, biasing the track length distribution. To overcome this, a projection model based on track length and angle from the c-axis was created that generates a secondary c-axis corrected MTL [194]. Due to the temporal span of publication of data in this work, most publications do not include both uncorrected and corrected MTL values and as such, this dataset is complex. Moreover, the chemical composition of an apatite affects its annealing characteristics of fission tracks. As such, horizontal track measured in fluorapatite (Ca₅(PO₄)₃F), will anneal differently to those in chlorapatite (Ca₅(PO₄)₃Cl) and MTL values will vary from the same thermal history. In the literature, chemical composition of apatite is rarely defined, though Cl wt% and D_par are documented to account for variable annealing characteristics in thermal history models. Such data are sparse in our dataset and cannot be easily accounted for, hence the quality of MTL values may be affected by these chemical differences.

The distribution of AFT ages and MTL values across the NE Atlantic varies significantly. Further investigation of these data suggests that the Mesozoic extensional tectonics related to regional multi-stage rifting together with the underlying bedrock geology are central in the modern distributions of both values. Across Scandinavia, AFT age and MTL appear to be primarily controlled by the opening of the NE Atlantic, though ages across the Southern Scandes may imply the protracted erosion of older topography. Across Britain and Ireland, ages and MTL derived from older metamorphic bedrock appear consistent with protracted erosion during the opening of the Atlantic and, while data from sedimentary rocks show a more complex distribution, they are likely linked to varied burial histories and the effects of North Atlantic Igneous Province. This trend appears similar across East Greenland, where data derived from older metamorphic bedrock appears consistent with asymmetric rift-related exhumation, while those from sedimentary rocks or adjacent to large volcanic centers are greatly affected by varied burial patterns and reheating during volcanism. Finally, AFT ages and MTL values from across Svalbard are consistent with the contrasting geology histories of the region.

Alongside broad interpretations of the tectonic and exhumation histories of the region, this data compilation highlights major localized trends in fission-track ages and length values across each region, while also highlighting areas where sample density remains low. Areas that remain underexplored include southern Britain, NE Britain, the southeast coastline of Greenland, southern Svalbard, and eastern Fennoscandia. The NE Atlantic has been studied for centuries helping to greatly improve our understanding of extensional tectonics, passive margin development, orogenesis, and volcanism. Though this work has underpinned many advances in our field, important pieces are missing, and it is only when we stand back and assess the full picture that we can recognize those gaps.

All data mentioned in this work are available in the Supplementary Data. This will also be uploaded to FigShare on the paper’s acceptance.

Scott Jess postdoctoral fellowship is supported by funding provided by the University of Toronto Mississauga to Prof. Lindsay Schoenbohm. Heike R. Gröger is thankful to Prof. Em. Stefan M. Schmid for fruitful discussions. Alexander Peace acknowledges the Natural Sciences and Engineering Research Council (NSERC) Discovery Grant RGPIN-2021-04011 for research program financial support. Christian Schiffer is funded by the Swedish Research Council (Vetenskapsrådet, 2019-04843). Ruohong Jiao is thanked for their work as editor, while Paul Green, Peter Japsen, and two other anonymous reviewers are thanked for their constructive feedback that greatly improved the manuscript.

Supporting Information includes an Excel document with all data provided Supplementary Materials Available: All data mentioned in this work are available in the Supplementary Data. This will also be uploaded to FigShare on the paper’s acceptance.