Late Neoproterozoic-to-recent long-term t–T-evolution of the Kaoko and Damara belts in NW Namibia

Abstract

This research aims to reconstruct the Late Neoproterozoic-to-recent long-term time–temperature-evolution of the NW Namibian Kaoko and Damara belts combining numerical modeling of new thermochronological data with previously published geochronological data, i.e., U–Pb, Sm–Nd, and Rb–Sr analyses, and K/Ar, ⁴⁰Ar/³⁹Ar low-temperature thermochronology. Consequently, we retrieve a coherent long-term time–temperature-evolution of the NW Namibian Neoproterozoic basement rocks including rates of exhumation and subsidence periods over the last ~ 500 Myr. Neoproterozoic basement rocks indicate fast post-Pan African/Brasiliano cooling and exhumation, reheating, or rather subsidence during the development of the Paleozoic-to-Mesozoic SW Gondwana intraplate environment and a significant thermal overprint of the rocks during South Atlantic syn- to post-rift processes, and therefore, resemble the opponent SE Brazilian time–temperature-evolution. We provide an overview of thermochronological data including new apatite and zircon fission-track data derived from Neoproterozoic, Late Paleozoic, and Lower Cretaceous rocks. Apatite fission-track ages range from 390.9 ± 17.9 Ma to 80.8 ± 6.0 Ma in the NW Kaoko Belt with youngest ages confined to the coastal area and significant age increase towards the inland. New zircon apatite fission-track data reveal ages between 429.5 ± 47.8 and 313.9 ± 53.4 Ma for the rocks of the Kaoko Belt. In the central Damara Belt, new apatite fission-track ages range between 138.5 ± 25.3 Ma to 63.8 ± 4.8 Ma. Combined apatite fission-track age distributions from Angola to Namibia and SE Brazil correlate for both sides of the South Atlantic passive continental margin and the reset AFT ages overlap with the lateral Paraná–Etendeka dike swarm distribution.

Introduction

Precambrian metamorphic and magmatic rocks are important archives that have stored information about long-term landscape forming processes, e.g., extensive and localized tectonic and volcanic activity, surface uplift and erosion (exhumation), and crustal scale subsidence. These long-term landscape forming processes are related to upper mantle and crustal tectono-thermal dynamics, which can be reconstructed from the long-term time (t)–temperature (T)-evolution of rocks in specific geological environments, i.e., along passive continental margins around the world (e.g., Brown et al. 2014; Green et al. 2015, 2018; Japsen et al. 2012, 2014; Braun 2018).

Until recently, most of the research along the South Atlantic passive continental margin (SAPCM) in NW Namibia was either focused on the Precambrian deformation, the Paleozoic-to-Mesozoic intraplate development, or the syn- to post-rift evolution (e.g., Goscombe et al. 2003a, b, 2007, 2017; Foster et al. 2009; Raab et al. 2002, 2005; Catuneaunu et al. 2005; Miller 2008; Brown et al. 2014; Milani et al. 2015). Moreover, this area provides a unique natural laboratory to study in detail sole stages of the Wilson cycle (Wilson 1963, 1965; Stanistreet et al. 1991) and beyond that allows connecting those individual geological stages to retrieve a coherent long-term t–T-evolution, i.e., the development from an orogeny to a passive continental margin environment.

In a recent study, Krob et al. (2019) published a coherent long-term t–T-evolution including rates of exhumation and subsidence periods over the last ~ 500 Myr of the western side of the SAPCM, the SE Brazilian passive continental margin. Consequently, this research focusses on the long-term t–T-evolution of the eastern counterpart, the NW Namibian passive continental margin. We present a large number of thermochronological data including new apatite and zircon fission-track (AFT, ZFT) data, integrate them to published AFT data sets along the Angolan-to-Namibian SAPCM (Silva et al. 2019; Green and Machado 2015; Raab et al. 2002, 2005; Brown et al. 2014), and provide a comparison of AFT data sets from both the Angolan-to-Namibian, and SE Brazilian SAPCM.

We combine numerical modeling of our data set with geo- and thermochronological data (U–Pb, Sm–Nd and Rb–Sr, and K/Ar, ⁴⁰Ar/³⁹Ar, AFT) from previous work (e.g., Seth et al. 1998; Goscombe et al. 2003a, b, 2005a, b; Kröner et al. 2004; Gray et al. 2006; Foster et al. 2009; Milani et al. 2015) to reconstruct the coherent long-term t–T-evolution of the Neoproterozoic basement rocks along the NW Namibian Kaoko and Damara belts (Fig. 1). Through this reconstruction, we retrieve the timing and rates of (a) the syn- to post-Pan African cooling and exhumation, (b) the heating, or rather subsidence during the intraplate basin formation of the Karoo Supergroup, and (c) the thermal influence of the Lower Cretaceous Paraná–Etendeka Large Igneous Province (LIP) event, and the SAPCM syn- to post-rift evolution.

Fig. 1

Schematic workflow showing the approach of this research study. Sample preparation of Neoproterozoic rock samples from the NW Namibian Kaoko and Damara belts were performed at the institute of Earth sciences, Heidelberg University (right panel). Fission-track dating on apatite and zircon provided apatite and zircon single grain ages, apatite confined fission-track length, and etch pit size (Dpar^®) data (Eldracher, in prep.). We used the published geological record to generate a numerical modeling set-up by transferring single geological events into time (t)–temperature (T)-constrained boxes (left panel). By testing our thermochronological data against the modeling set-up, we derived continuous t–T-paths that were combined with previous geo- and thermochronological data to reconstruct a possible coherent long-term t–T-evolution for the Neoproterozoic basement rocks in NW Namibia. Thereby, we are able to make estimations on exhumation and subsidence rates and generate interpolation maps of the Neoproterozoic basement thermal structure. Colors of boxes: yellow: data compilation; blue: methods and approach; green: results

Geological setting

Late Neoproterozoic-to-early Paleozoic evolution—Pan African/Brasiliano orogeny

From the Late Neoproterozoic to Early Paleozoic (ca. 650–490 Ma), the Pan African/Brasiliano Orogeny took place during the large-scale consolidation of the Gondwana supercontinent (Fig. 2) (e.g., Stanistreet et al. 1991; Goscombe and Gray 2008; Foster et al. 2009; Frimmel et al. 2011; Nascimento et al. 2016). Following an early phase of collisional events in the northern parts of the Gondwana Assembly (Amazonia, Paranapanema, Goías, São Francisco, and Congo cratons), continuous movement of the cratonic blocks (Rio de la Plata, Luis Alves, Angola, and Kalahari cratons) in the southern part led to the amalgamation of the SE Brazilian Dom Feliciano, and the NW Namibian Kaoko and Damara Neoproterozoic mobile belts (ca. 560–490 Ma) (e.g., Goscombe et al. 2003a, b, 2007; Gray et al. 2006; Miller 2008; Oyhantcabal et al. 2011; Schmitt et al. 2012).

Fig. 2

Left: Neoproterozoic mobile belts and cratons in South America and Africa (SW Gondwana) modified after Heilbron et al. (2008). Red rectangle shows study area. Neoproterozoic belts: 1: Brasilia Belt; 2: Araçuaí Belt and Central Ribeira Belt; 3: Southern Ribeira Belt; 4: Dom Feliciano Belt; 5: West Congo Belt; 6: Kaoko Belt; 7: Damara Belt. Major Cratons: AM Amazonia, SF São Francisco, LA Luis Alves, RP Rio de la Plata, WA West Africa, CO Congo, ANG Angola, KA Kalahari, PP Paranapanema. Right: Tectonic units of the NW Namibian Kaoko and Damara belts (modified after Goscombe et al. 2003a, b; Hoffmann et al. 2004; Miller 2008; Jung et al. 2014): a Bangweulo shield; b Mesozoic sedimentary and volcanic rocks and Cenozoic cover; Kaoko Belt: Western Kaoko Zone: c Coastal Terrane; d Orogen Core; Central Kaoko Zone: e Pre-Damara Congo Craton; f Central Kaoko Zone; Eastern Kaoko Zone: g Northern Platform; Southern Kaoko Belt: h Ugab Zone. Damara Belt: i Outjo Zone (Northern Zone); j Swakop Zone (Northern and Southern Central Zone); k Khomas Zone (Southern Zone); l Hakos Zone (Southern Margin Zone); m Pre-Damara Kalahari Craton; n Southern Nama Foreland. AS Angola shield, EMC Epupa Metamorphic Complex, E Etendeka flood basalts, KI Kamanjab Inlier, W Windhoek, WB walvis bay

The coastal, NW-trending Kaoko Belt extends from northern Namibia to the Huab Basin in the south (Fig. 3). The basement consists of the Archean and Paleo- to Mesoproterozoic rocks of the Congo Craton that are overlain by more than 1000 m of Neoproterozoic para-metamorphic (sedimentation age: 770–600 Ma) and igneous rocks of the Damara Sequence (Miller 1983; Frimmel et al. 1998, 2011; Hoffman et al. 1998; Goscombe et al. 2003a, b, 2007; Goscombe and Gray 2008; Nascimento et al. 2016). The Kaoko Belt is subdivided into three individual zones that comprise several tectonic domains. The Western Kaoko Zone (WKZ) involves the Coastal Terrane representing a transpressive magmatic arc, and the Orogen Core (Fig. 2c, d) that both obducted onto the Congo Craton (Fig. 2e, f; Central Kaoko Zone, CKZ). The Eastern Kaoko Zone (EKZ) contains the Northern Platform and the Epupa Metamorphic Complex in the north (Fig. 2g, EMC) (Goscombe et al. 2007). Finally, the Ugab Zone (Fig. 2h) as part of the Orogen Core represents the most southern part of the Kaoko Belt (Goscombe et al. 2007; Miller 2008, 2013).

Fig. 3

Simplified geological map of the study area in NW Namibia with the corresponding sample locations and numbers (NA11-xx) (modified after Frimmel et al. 2011; locations of shear zones after Foster et al. 2009; Salomon 2015). WKZ Western Kaoko Zone, CKZ Central Kaoko Zone, EKZ Eastern Kaoko Zone, ST Sesfontein thrust, PMZ Purros mylonite zone, TPSZ Tree palms shear zones, OT Otjohorongo thrust, OmL Omaruru lineament, OL Okahandja lineament. For further information, see supplementary material, table S1

The EW-trending Damara Belt stretches from the coast to the Waterberg Basin further inland and comprises Precambrian basement rocks of the Kalahari Craton also overlain by Neoproterozoic para-metamorphic and igneous rocks of the Damara Sequence (Fig. 3) (Miller 1983). Several tectonic zones subdivide the Damara Belt in the Northern (NZ, Outjo Zone), Northern and Southern Central (NCZ and SCZ, Swakop Zone), and Southern (SZ, Khomas Zone) zones (Fig. 2). In the south, the Southern Margin Zone (Hakos Zone) and the Southern Nama Foreland border the southern Damara Belt and represent the transition towards the southern Kalahari Craton (Miller 2008).

Late Paleozoic-to-Jurassic evolution—SW Gondwana intraplate formation of sedimentary basins

Following the Pan African Orogeny, the SW Gondwana intraplate environment experienced a long period of predominantly surface uplift, and erosion caused by post-orogenic denudation of the Neoproterozoic mobile belts during the Early-to-Middle Paleozoic, and Carboniferous-to-Permian syn- to post-Gondwana Ice Age processes (Frimmel et al. 1998; Goscombe et al. 2003a, b; Goscombe and Gray 2008; Miller 2008). Subsequently, geological conditions changed towards the Late Paleozoic and renewed transport and deposition of sediments led to the development of the major SW Gondwana intraplate basins, e.g., the southern African Karoo and the SE Brazilian Paraná basins (Stollhofen et al. 1999).

In the NW Namibian Huab (southern Kaoko Belt) and the Waterberg (eastern Damara Belt) basins, localized sedimentation of the Karoo Supergroup started during the Late Carboniferous discordantly onto Damara orogen rocks (Fig. 4). Siliciclastic rocks of the Carboniferous-to-Permian Dwyka Group were deposited during the Gondwana Ice Age, whereas Ecca and Drakensberg sedimentary and volcanic rocks indicate that the climate became progressively more arid changing from tropical and subtropical to more desert-like depositional environments (Miller 2008). In the Huab Basin, post-glacial sediments (Ecca and Beaufort Group) consist of siliciclastic units intercalated with coal seams and carbonate rocks. In contrast, these sedimentary sections are lacking within the Waterberg Basin where sedimentation resumed with mainly siliciclastic rocks (Dwyka-to-Stromberg Group) with only minor intercalation of carbonate units (Ecca Group) until the Early Jurassic. These sedimentary rocks (< 750 m) are overlain by a siliciclastic section (Etjo Fm) interfingering with basaltic flows (5 m, Rooival Fm) that are correlated to the Karoo LIP event of Early Jurassic time (~ 183 ± 1 Ma, Johnson et al. 1996; Duncan et al, 1997; Catuneaunu et al. 2005).

Fig. 4

Simplified stratigraphic W–E transect through the NW Namibian sedimentary Huab and Waterberg basins with corresponding chronostratigraphy, lithostratigraphy, max. column deposited, lithology, depositional environment, facies (F.), and events (after Johnson et al. 1996; Stollhofen et al. 1999; Catuneaunu et al. 2005). For locations of the basins, see Fig. 3

Late Jurassic-to-Cenozoic evolution-break-up, and syn- to post-rift processes.

The Late Jurassic-to-Early Cretaceous evolution of the Namibian SAPCM was dominated by continuous pre- to post-rift processes, e.g., tectonic and surface uplift, erosion, subsidence, and sedimentation, as well as the emplacement of dike swarms and flood basalts caused by the Paraná–Etendeka plume movement (e.g., Courtillot et al. 1999; Sengör 2001; Ernst and Buchan 2001a; Trumbull et al. 2007; Hu et al. 2018).

In the western Huab Basin, this geological evolution is recorded by Etendeka volcanic rocks overlying Upper Jurassic-to-Lower Cretaceous eolian sediments (Twyfelfontein Formation) (Fig. 4) (Miller 2008). These sedimentary rocks interfinger with the overlying volcanic rocks (Jerram et al. 1999; Salomon et al. 2016), and both sections correlate with its Brazilian counterparts, the sedimentary Botucatu, and volcanic Serra Geral formations (Milner 1986; Stollhofen et al. 2000; Miller 2008). Together with these Brazilian Paraná basin volcanic rocks, the Etendeka flood basalts form the Lower Cretaceous Paraná–Etendeka LIP that emplaced during the South Atlantic syn-rift stage. In Africa, the Etendeka volcanic suite covers more than 78,000 km² (Fig. 3), and maximum preserved thicknesses reach ~ 900 m (Milner et al. 1992; Peate 1997; Stollhofen et al. 1999). However, Etendeka dike swarms indicate a wider area of originally lava covered area reaching as far north as the Angolan border and as far west as the southern Waterberg Basin area within the Damara Belt (Peate 1997; Trumbull et al. 2004, 2007; Torsvik et al. 2009, 2014). The dike swarms have most likely acted as feeder channels for the volcanic rocks and follow the Kaoko (~ margin parallel) and the Damara (~ margin perpendicular) belt structures (Trumbull et al. 2004, 2007; Salomon et al. 2017). Radiometric ages for the volcanic rocks range from ~ 135 to ~ 127 Ma (Trumbull et al. 2004, 2007; Baksi 2018) indicating ongoing magmatic activities along the Namibian SAPCM for approximately up to 10 Myr (Milner et al. 1995).

Following the Paraná–Etendeka magmatic activity, the NW Namibian SAPCM rocks experienced post-rift cooling and exhumation. In the NW Kaoko Belt, published AFT ages range from ~ 80 to ~ 50 Ma along the coast and from ~ 330 to ~ 260 Ma further inland (Brown et al. 2014). In the Damara Belt, published AFT data revealed age from ~ 120 to ~ 70 Ma. Further south (Windhoek area), AFT ages range between ~ 550 and  ~ 200 Ma (Raab 2001; Raab et al. 2002, 2005; Brown et al. 2014). Brown et al. (1990) proposed cooling of the coastal rocks during the Early Cretaceous-to-Cenozoic post-rift processes due to erosion of up to 3.5 km of sediments and transport to the offshore basins.

Tectonic structures

In the Kaoko Belt, predominant tectonic zones are of Neoproterozoic age [Sesfontein Thrust (ST), Purros Mylonite Zone (PMZ), and Tree Palms Shear Zone (TPSZ)], and follow the ~ NW trend dividing the mobile belt into the longitudinal Eastern, Central, and Western (Kaoko) Zones (Fig. 3) (Goscombe et al. 2003a, b). Salomon et al. (2015) assume partial reactivation of the tectonic structures during the South Atlantic opening, with estimated normal offsets < 1200 m. Kukulus (2004) also assumes major post-flood basalt deposition tectonic activity due to the offset of tilted basal Etendeka volcanic rocks.

South of the Kaoko Belt (Fig. 3), major margin perpendicular lineaments [Otjohorongo Thrust (OT), Omaruru Lineament (OmL), and Okahandja Lineament (OL)] extend across the EW-trending Damara Belt following the pre-Damara metamorphic basement structures of the intracontinental branch and also have probably been reactivated during the break-up processes (Raab et al. 2002, 2005; Brown et al. 2014; Salomon et al. 2015, 2017).

Methods

Thermochronology

Thermochronology is based on the accumulation and thermally controlled retention of isotopic daughter products and linear crystal defects produced during the radioactive decay of a parent. Due to the temperature sensitivity of thermochronometers, ages provide information about the thermal history of the rock, rather than mineral crystallization ages (although in fast cooled volcanic rocks, they do record crystallization ages).

For this research, apatite and zircon fission-track (AFT and ZFT, respectively) analyses were performed on 89 samples of Neoproterozoic metamorphic, and magmatic, Carboniferous-to-Permian sedimentary, and Lower Cretaceous volcanic rocks along the NW Namibian Kaoko and Damara belts (Fig. 3, Tab. S1). Sample preparation and analyses followed the same protocol as, e.g., Karl et al. (2013). Whenever possible, two thermochronometry dating techniques were applied to the samples, allowing a more robust evaluation of the spatial and temporal cooling of the sampled rocks. However, only a few samples (25) contained apatite and zircon grains in quantities, to allow AFT and ZFT analyses.

Fission-track dating techniques correspond to a specific total annealing temperature [T (an)]. When temperatures exceed T (an) over a certain time, the chronometer is totally reset. Partial annealing occurs when temperatures or required time did not fulfil conditions for total annealing. For further interpretations, this research uses temperatures between 110 °C/10 Myr and 60 °C/10 Myr for the partial annealing zone (PAZ) for the AFT analysis (Gleadow and Duddy 1981), and between 330 °C/10 Myr and 190 °C/10 Myr for ZFT analysis (Garver and Kamp 2002; Garver 2003; Hurford 1986; Rahn et al. 2004; Reiners and Brandon 2006).

Numerical modeling of continuous t–T-paths (t–T-evolution)

Thermal modeling of thermochronological data was used to determine continuous cooling and heating histories (t–T-evolution) of individual samples (Fig. 1). The numerical modeling software code HeFTy^® (v.1.9.3.) (Ketcham 2005; Ketcham et al. 2007a, b, 2009,2017) allows for reconstruction of the t–T-evolution of the Neoproterozoic basement to quantify the exhumation and subsidence history of several crustal segments of the upper crust.

Therefore, we used the existing information derived from the published geological record, such as metamorphic ages, magmatic events, and lithostratigraphic data that were transferred into single constrained t–T-boxes and embedded into the modeling software (Fig. 5). The t–T-constraints differ across sample locations within the key area. However, the same t–T-coordinates were applied strictly to samples of the same geological area to guarantee a better comparability of the t–T-histories of the modeled samples in their respective regions.

Fig. 5

Numerical modeling set-up for samples of the Namibian Damara Sequence located in the Kaoko Belt (upper), and Damara Belt (lower). The set-up was generated from previous work and uses field observations and published stratigraphic (Johnson et al. 1996; Duncan et al. 1997; Catuneaunu et al. 2005; Miller 2008), paleo-geographic (Scotese et al. 1992, 1999; Torsvik et al. 2009, 2014), and -climatic knowledge (Wygrala 1989; Brown et al. 1990). We took single geological events from the published geological record, i.e., metamorphic ages, magmatic events, and lithostratigraphic data and transferred the given information into specific t–T-constraints applied to the numerical modeling software code HeFTy^® (v.1.9.3.) (Ketcham 2005; Ketcham et al. 2007a, b, 2009,2017). The connected tables below the t–T-evolution show further explanations for the applied t–T-constraints

Then, HeFTy^® performs a so-called inverse model and seeks paths that fit through the t–T-constraints considering the input thermochronological data. Thereby, we test our thermochronological data against specific t–T-constraints derived from the published geological record to retrieve continuous long-term t–T-paths since the Late Paleozoic. The thermochronological data sets used for the numerical modelling are: AFT single grain ages, confined spontaneous fission-track length distribution (> 50 individual length) corrected for c-axis related angle (Donelick et al. 1999; Ketcham et al. 2007a, b, 2009), etch pit size (D_par^®), and annealing kinetics of Ketcham et al. (2007a, b).

Reconstruction of the t–T-paths and calculation of rates

Thermochronological data presented in this study only cover the last ~ 400–300 Ma and we, therefore, integrated published geo- and thermochronological data (U–Pb, Sm–Nd and Rb–Sr, and K/Ar, ⁴⁰Ar/³⁹Ar, AFT) to our own data set to reconstruct a coherent t–T-path over more than 500 Ma. Most of the geo- and thermochronological data are confined to specific geological regions or only describe local geological phenomena, i.e., intrusions or contact metamorphism that do not provide information of the t–T-evolution of entire crustal blocks and segments. Therefore, we chose published data, which we consider relevant for our study locations. Consequently, the published geochronological data reflect the main intrusion, crystallization, metamorphic, and cooling processes, and constrain the presented t–T-evolution. Published ages were transferred to specific temperature areas according to their effective closure, and/ or annealing temperatures including wide ranges of errors, both in time, and temperature (Fig. S1; Pollard et al. 2002). In a further step, we drew average t–T-paths connecting the existing data taking into account that uncertainties related to the time, temperature, locations, and the probability of the reconstructed cooling paths often entail the generation of error strings, and should not be disregarded.

This critical view also holds for the calculation of exhumation and subsidence rate that directly base on the course of the differentiated t–T-path and do not reflect the timing of geological events, e.g., single intrusion and/or metamorphism ages. The intervals were also constrained by the different cooling paths and describe the distance from tipping point to tipping point. The change in temperature (ΔT) over time (Δt) was used to determine the cooling or heating rate (ΔT/Δt) using the “weighted-mean” paths of the models t–T-histories. The cooling and heating rates were divided by an assumed geothermal gradient to obtain an exhumation and subsidence rate, respectively.

Geothermal gradients depend on the predominant average heat flow over long-term periods including the heat generation and thermal conductivity of rocks, metamorphic reactions, and descendent waters (Barker et al. 1986; Rybach et al. 1988; Yalçin et al. 1997; Burov 2009). There exist short-term periods with anomalously high or low geothermal gradients, e.g., in direct surrounding of intrusions or during Ice House periods. However, quantitative research of the thermal influence in relation to distance or depth (e.g. intrusions, dikes, or ice shields), and the effect and duration of those anomalous periods (e.g., intrusion or magmatic activity in LIP environments) is lacking (Barker et al. 1986; Rybach et al. 1988; Yalçin et al. 1997; Burov 2009). Therefore, we decided to work with a homogenous and arithmetical mean paleo-geothermal gradient of 30 °C/km for the entire modeled time (Brown et al. 1990). The paleo-geographical reconstruction of the research area was taken from Scotese et al. (1992,1999) and leads to paleo-surface temperatures of an average of 10 °C for Precambrian-to-Late Paleozoic time, and an mean of 25 °C since Mesozoic time (Wygrala 1989). All rates are expressed as positive rates and are related to exhumation or subsidence according to the decrease and increase of temperature over time.

Results

Apatite fission-track data

Kaoko belt

For the Kaoko belt igneous and sedimentary rocks, altogether 72 samples, revealed at least between 10 and 29 grains that could be dated with AFT analysis (Figs. 6; S1; Table S2). Six samples revealed only between 3 and 9 grains, and therefore, their age information remain less certain. Except for 8 samples (all 0%), all single grain age distributions (of 64 samples) passed the Chi-square test with values > 5% indicating a homogeneous distribution with respect to 1σ-error of the single grain ages (Galbraith 1981).

Fig. 6

Simplified geological map of the study area in NW Namibia showing the regional apatite fission-track (AFT) age distribution in the Kaoko and Damara belts (modified after Frimmel et al. 2011; locations of shear zones after Foster et al. 2009; Salomon 2015). Dots: Neoproterozoic samples; rectangles: Paleozoic-to-Mesozoic Karoo samples; triangles: Late Jurassic-to-Early Cretaceous Etendeka samples. WKZ Western Kaoko Zone, CKZ Central Kaoko Zone, EKZ Eastern Kaoko Zone, ST Sesfontein thrust, PMZ Purros mylonite zone, TPSZ Tree palms shear zones, OT Otjohorongo thrust, OmL Omaruru lineament, OL Okahandja lineament. For further information, see supplementary material, tables S1 and S2

AFT ages of the Precambrian Damara Sequence (58 samples) range between 390.9 ± 17.9 Ma and 81.3 ± 4.2 Ma. In general, AFT ages are younger along the coast and ages increase systematically inland within the Central and Eastern zones (Figs. 6, 7; S1). Sedimentary rocks of the Karoo Supergroup (9) revealed ages between 342.2 ± 18.3 Ma and 96.5 ± 8.4 Ma. Magmatic rocks of the Etendeka Group (5) provided ages between 125.6 ± 23.6 Ma and 103.5 ± 4.9 Ma.

Fig. 7

Detailed interpolation map showing a possible modeled projection of the regional AFT age distribution of the NW Namibian Kaoko and Damara belts. Samples of this study (white) are integrated to published AFT data by Brown et al. (2014) (black), and Raab et al. (2002, 2005) (purple) using the Golden Software Surfer^®. Whereas yellow-to-red colors indicate younger ages, green-to-blue colors imply older ages. Isolines show age in [Ma] without error. Colored dots (Neoproterozoic Damara basement rocks), rectangles (Late Paleozoic-to-Mesozoic Karoo sedimentary rocks), and triangles (Lower Cretaceous Etendeka sedimentary and volcanic rocks) show sample locations according to the references. Isolines show age in [Ma] without error. ST Sesfontein thrust, PMZ Purros mylonite zone, TPSZ Tree palms shear zones, OT Otjohorongo thrust, OmL Omaruru lineament, OL Okahandja lineament. Dashed box: locations of ZFT data see Figs. 8 and 9

Confined spontaneous fission-track length and etch pit size (D_par^®) data were acquired for all 72 Kaoko Belt samples (Table S3) showing D_par^® values between 0.8 and 1.8 µm (average: 1.2 µm). This range indicates a relatively inhomogeneous composition concerning the fluorine and chlorine content of apatite. We measured altogether 2529 lengths of confined spontaneous fission tracks for all 72 Kaoko Belt samples. For 24 samples, between 50 and 152 confined spontaneous fission-track lengths could be measured and were integrated to the software code HeFTy^® to model the t–T-evolution. Mean track lengths vary between 7.9 and 14.8 µm. The c-axes correction of the confined track lengths led to a length distribution between 9.8 and 15.4 µm.

Damara Belt

For 17 Damara Belt rock samples (Figs. 6; S2; Table S2), at least 15 grains could be dated with AFT analysis except for NA11-69B (five grains). All single grain age distributions passed the Chi-square test with values > 5% indicating a homogeneous distribution with respect to 1σ-error of the single grain ages (Galbraith 1981).

AFT ages of the Precambrian Damara Sequence (15) range between 138.5 ± 25.3 and 63.8 ± 4.8 Ma. In general, AFT ages are younger along the coast and ages increase systematically inland towards the Waterberg Basin (Figs. 6, 7; S2). AFT analysis revealed only one age for a Karoo conglomerate (88.4 ± 7.2 Ma) and Etendeka granite (71.6 ± 5.2 Ma).

Confined spontaneous fission-track length and etch pit size (D_par^®) data were acquired for all Damara Belt samples (Table S3). We measured 1963 etch pit sizes for all apatite grains. The mean D_par^® value for each sample ranges between 1.1 and 1.6 µm with an average D_par^® value of 1.3 µm. This narrow range indicates a homogeneous composition concerning the fluorine and chlorine content of apatite. We measured 736 confined spontaneous fission-track lengths for all Damara Belt samples. Eight samples obtained between 51 and 111 measured confined spontaneous fission-track lengths that were applied to the software code HeFty^® for the numerical modeling of the t–T-evolution. Mean track lengths vary between 8.6 and 13.5 µm. The c-axes correction of the confined track lengths led to a distribution between 11.9 and 14.5 µm.

Zircon fission-track ages

ZFT analyses were performed on 29 rock samples of the Kaoko Belt (Figs. 8; S3; Tab. S4). All ages obtained for the Precambrian Damara Sequence (26 samples) are younger than their stratigraphic age. ZFT ages range between 429.5 ± 67.8 and 313.9 ± 53.4 Ma. Only two samples, NA11-89G (118.2 ± 25.1 Ma) and NA11-91 (233.5 ± 43.5 Ma) in the north, show younger ages and do not lie in the general age trend (Fig. 9). However, both samples revealed only two grains to perform FT analysis and therefore, the reliability of the age information remains uncertain. Furthermore, both samples have abnormal high uranium contents in relation to the other samples. This also holds for only three measured grains of the gneiss sample NA11-126 (316.3 ± 70.7 Ma). Nevertheless, all single grain age distributions fulfilled the requirement of the χ²-test with values > 5% indicating a homogenous distribution with respect to 1σ-error of the single grain ages (Galbraith 1981) with the exception of NA11-106 (436.8 ± 45.9 Ma) that fails the χ²-test (0.01%).

Fig. 8

Simplified geological map of the study area in NW Namibia showing the regional zircon fission-track (ZFT) age distribution in the Kaoko and Damara belts (modified after Frimmel et al. 2011; locations of shear zones after Foster et al. 2009; Salomon 2015). Dots: Neoproterozoic samples; rectangles: Paleozoic-to-Mesozoic Karoo samples; triangles: Late Jurassic-to-Early Cretaceous Etendeka samples. WKZ Western Kaoko Zone, CKZ Central Kaoko Zone, EKZ Eastern Kaoko Zone, ST Sesfontein thrust, PMZ Purros mylonite zone, TPSZ Tree palms shear zones, OT Otjohorongo thrust, OmL Omaruru lineament, OL Okahandja lineament. For further information, see supplementary material, tables S1 and S4

Fig. 9

Detailed interpolation map showing a possible modeled projection of the regional ZFT age distribution of the NW Namibian Kaoko Belt using the Golden Software Surfer^®. For orientation, see Fig. 7 (dashed box). Whereas yellow-to-red colors indicate younger ages, green-to-blue colors imply older ages. White dots show sample locations of the Neoproterozoic Damara basement rocks. Isolines show age in [Ma] without error. ST Sesfontein thrust, PMZ Purros mylonite zone, TPSZ Tree palms shear zones, OT Otjohorongo thrust, OmL Omaruru lineament, OL Okahandja lineament

Three sedimentary rock samples (NA11-118, NA11-135GO, NA11-135R) of the Karoo Supergroup revealed ZFT ages between 365.6 ± 30.6 Ma and 377.2 ± 30.4 Ma and, therefore, represent detrital ages.

Numerical modeling of the t–T-evolution

Using the software code HeFty^® (v.1.9.3.; Ketcham 2005; Ketcham et al. 2007a, b, 2009, 2017), the inverse numerical modeling of the t–T-evolution of the Neoproterozoic basement rocks within the NW Namibian Kaoko and Damara belts was applied to test the thermochronological data set against single geological events (t–T-constraints) described in the published geological record (Fig. 5).

For 29 samples (Tab. S3), a high number of measured fission tracks in apatite (> 51) allow robust constraints on the t–T-path through 60–110 °C. In general, 50 000 t–T-paths have been tested against the data set and yielded a high goodness of fit (G.O.F.) (Figs. S4.1–S4.8) (Ketcham et al. 2007a, b). However, to provide comparability of the modeled samples, lower G.O.F. values were also accepted without omitting t–T-constraints to gain better results. Therefore, t–T-models passed our requirements when good fits were found. Altogether, t–T-models reveal between 1 and 519 good, and between 3 and 2681 acceptable t–T-paths, while 34% of the models show > 100 good paths, and 58% > 30. More than 1000 acceptable paths could be reproduced by 38% of the models, and > 100 by 93%. The AFT age and confined fission-track length distribution for 86%, or rather 55% of the samples were reproduced with an excellent G.O.F. of > 80%. Therefore, we find sufficient t–T-paths consistent with both the thermochronological data and geological history.

The sole numerical modeling of our thermochronological data focusses on the t–T-evolution since Carboniferous time over temperatures below 300 °C (Fig. 10). The published record suggests that Neoproterozoic basement rocks had reached the surface before the the onset of the Karoo sedimentation during the Early-to-Late Paleozoic SW Gondwana intraplate environment (Johnson et al. 1996; Stollhofen et al. 2000; Catuneaunu et al. 2005; Milani et al. 2007). During the Triassic, the models show coherent heating of the samples showing the ongoing subsidence of the Paleozoic-to-Mesozoic basins (Huab and Waterberg). From Early Jurassic, t–T-models allow cooling inferred from possible initial plume movement during pre-rift-processes (Şengör 2001). All models show a possible cooling to temperatures near or at the Earth’s surface. The evolution is followed by a reheating during the Early Cretaceous that reflects the thermal influence caused by the emplacement of dike swarms and accompanied volcanic activity, and the deposition of the Etendeka flood basalts. Maximum temperatures are reached especially in the southern Damara Belt (~ 120 °C). Then, renewed cooling sets in during the Early Cretaceous and continues until the Cenozoic.

Fig. 10

Simplified geological map of the study area in NW Namibia showing the results of the inverse numerical modeling in different areas of the Kaoko and Damara belts (modified after Frimmel et al. 2011; locations of shear zones after Foster et al. 2009, Salomon 2015). Windows show the continuous “weighted-mean” t–T-paths of the samples since the Late Carboniferous modeled with the software code HeFty^® (Ketcham et al. 2007a, b; 2017). Samples were subdivided in individual areas of the northern area of the Western Kaoko Zone (WKZ, green), the northern central area of the Central Kaoko Zone (CKZ, blue), the central area of the Western and Central Kaoko Zone (orange), and the Ugab Zone of the Southern Kaoko Belt (SKB, yellow) within the Kaoko Belt, and the northern central and southern zones (red) of the Damara Belt. WKZ Western Kaoko Zone, CKZ Central Kaoko Zone, EKZ Eastern Kaoko Zone, ST Sesfontein thrust, PMZ Purros mylonite zone, TPSZ Tree palms shear zones, OT Otjohorongo thrust, OmL Omaruru lineament, OL Okahandja lineament. For further information, see supplementary material figures S6.1–S6.8

Interpretation and discussion

For discussion and interpretation of the proposed t–T-evolution of the Neoproterozoic basement, published geo- and thermochronological data (Tables S5.1–S5.6) were directly incorporated into the modeled t–T-histories to retrieve a coherent syn- to post-Late Neoproterozoic evolution of the NW Namibian Kaoko and Damara belts (Fig. 11). The projected t–T-paths were used to make estimates on exhumation and subsidence rates (Fig. 12a, b) and reflect an average and/or “weighted”-mean t–T-path of the modeled t–T-histories including wide error ranges for temperatures as well as on the temporal scale related to each dating method. We are well aware of the uncertainties and the generated error string entailed by the calculation of long-term exhumation and subsidence rates covering large areas over long-time periods. Nevertheless, we would like to propose these long-term rates as an approximation approaching the long-term temperature processes. Finally, temperature interpolation maps of the Neoproterozoic basement surface were used to approach a visualization of the t–T-evolution of the SW Gondwana intraplate thermal structure (Fig. 13).

Fig. 11

Coherent long-term time (t)–temperature (T)-evolution of the individual areas of the Kaoko and Damara belts in NW Namibia. Published geochronological data were directly incorporated into the “weighted-mean” t–T-paths (Fig. 10) of the samples modeled with the software code HeFty^® (Ketcham et al. 2007a, b) to reconstruct the entire syn-to post-Late Neoproterozoic t–T-evolution of the SAPCM in NW Namibia. Dashed blue lines show possible t–T-paths for the published geochronological data. Numbers show the differentiated t–T-paths for the individual area that were used for the calculation of cooling and heating, and exhumation and subsidence rates, respectively. Grey bar: Pan African/Brasiliano orogeny (e.g., Goscombe and Gray 2008; Foster et al. 2009; Frimmel et al. 2011; Nascimento et al. 2016); red line: Paraná–Etendeka event Trumbull et al. 2004, 2007; Baksi 2018). For further information of the geochronological data, see supplementary material, tables S5.1–S5.6 and figures S7.1–7.6

Fig. 12

a Calculated cooling and heating rates, derived from combined geochronological and thermochronologial t–T-modeling of the thermal evolution in the individual areas of the Western-to-Eastern Kaoko Zones. The differentiated cooling paths refer to the individual t–T-paths of Fig. 11. For further information, see text, chapter 3.3. Blue and red bars indicate cooling and heating [°C/Ma]. The thickness of the bars implies higher/lower rates. Black numbers [mm/a] indicate calculated exhumation and subsidence rates derived from the cooling and heating rates. b Calculated cooling and heating rates, derived from combined geochronological and thermochronologial t–T-modeling of the thermal evolution in the individual areas of the Southern Kaoko and Damara belts. The differentiated cooling paths refer to the individual t–T-paths of Fig. 11. For further information, see text, chapter 3.3. Blue and red bars indicate cooling and heating [°C/Ma]. The thickness of the bars implies higher/lower rates. Black numbers [mm/a] indicate calculated exhumation and subsidence rates derived from the cooling and heating rates

Fig. 13

Coherent Late Neoproterozoic-to-recent t–T-evolution of the SAPCM in NW Namibia showing the interpolated thermal structure of the Neoproterozoic basement surface at specific time steps (a–o) based on the t–T-paths obtained from modeling of our own thermochronological (black line) and published geochronological data (dashed black line). Red colors signify that the Neoproterozoic basement was in depth at higher temperatures, whereas lighter colors indicate lower temperatures. No data were available for the white areas

Late Neoproterozoic-to-Early Paleozoic—syn-Pan-to-post-Pan African cooling and exhumation

The Late Neoproterozoic-to-Early Paleozoic is characterized by post-Pan African cooling and exhumation of the Neoproterozoic basement rocks. The published geochronological data indicate a very complex and distinct geological evolution of the different crustal segments (Figs. 10, 11). We assume individual cooling paths within each crustal segment based on the published data. Therefore, we propose cooling paths reflecting the cooling of pre-Pan African deformation intrusion and crystallization of rocks, the evolution most likely representing the syn-Pan African intrusion/crystallization, main metamorphism and subsequent cooling of the area, as well as the cooling of post-Pan African deformation intrusion and crystallization of rocks. These cooling paths connect the individual published data that often stem from different locations and, thus, are not linked to each other. Nevertheless, we consider the projected cooling paths to represent a possible t–T-evolution of the crustal segments along the contemporary SAPCM of NW Namibia.

In the northern Kaoko Belt, geochronological data from the northern WKZ (Figs. 10, 11, green (1); Table S5.1) show U–Pb crystallizations and metamorphic ages of the Neoproterozoic basement rocks between ~ 575 and ~ 540 Ma (Goscombe et al. 2005a, b). We suggest three individual cooling paths for the Orogen Core (Fig. 11, dark blue dashed lines) in the northern area of the WKZ: the first and second cooling path describes syn-orogenic cooling either connecting the U–Pb ages and ⁴⁰Ar/³⁹Ar hornblende cooling ages around 540 Ma (Gray et al. 2006), or the Sm–Nd metamorphic age of ~ 506 Ma by Goscombe et al. (2005a, b). We propose relatively fast cooling with exhumation rates between ~ 0.3 mm/a (path 1) and ~ 0.1 mm/a followed by very fast exhumation with ~ 1.17 mm/a (path 2) (Fig. 12a, WKZ, Orogen Core). Both t–T-paths most likely show homogeneous cooling from ~ 500 Ma on (⁴⁰Ar/³⁹Ar, Gray et al. 2006), and therefore, we propose ongoing cooling connecting ZFT ages of this study of either Early Devonian, or Carboniferous age. Consequently, post-orogenic cooling initially decreased showing exhumation rates between ~ 0.02 and ~ 0.01 mm/a, and therefore, we assume that rocks have been kept at middle-to-upper crustal levels at temperatures of about ~ 300–200 °C. However, we assume both scenarios probable to reflect the syn- to post-orogenic cooling based on the most constrained cooling path for the northern Orogen Core.

A third possible path (3) describes syn- to post-orogenic cooling from ~ 540 Ma onwards connecting U–Pb, Sm–Nd (Goscombe et al. 2005a, b; Gray et al. 2006) and Carboniferous ZFT ages of this study showing exhumation rates of ~ 0.08 mm/a until the Early Carboniferous. There exists one younger U–Pb post-kinematic age by Goscombe et al. (2005a, b) around ~ 505 Ma that reflects differentiated late- or post-deformation cooling from Early Paleozoic to Carboniferous, and we propose a homogeneous long-term exhumation rate of ~ 0.3 mm/a due to the lack of further data constraining this post-deformation cooling path.

Towards the east, the northern central area of the CKZ (Figs. 10, 11, blue (2); Table S5.2) shows a different syn-to post-orogenic cooling. However, less data were available within this area leaving the possible cooling less constrained (Table S5.2). Kröner et al. (2010) describe old U–Pb ages between ~ 1800 and ~ 1200 Ma that represent pre-Pan African deformation crystallization of the Pre-Damara rocks of the Congo Craton. However, our proposed cooling path sets in during syn-Pan African cooling recorded by a metamorphic Sm–Nd age by Goscombe et al. (2003b) followed by an ⁴⁰Ar/³⁹Ar age of ~ 470 Ma (Gray et al. 2006). Therefore, we present only one average cooling path including the intrusion, metamorphism, and post-orogenic cooling for the northern central area of the CKZ. Consequently, exhumation initially occurred from Late Neoproterozoic with rates of ~ 0.06 mm/a, and probably increased during the Ordovician/Silurian to 0.2 mm/a (Fig. 12a, CKZ, northern central area). The post-Pan African cooling can be better constrained by ZFT ages of this study, and thus, we assume differentiated cooling from the Silurian to Carboniferous. Whilst a ZFT range of ~ 429 Ma that allows a relatively homogeneous exhumation (0.1 mm/a) directly to near surface temperatures (~ 15 °C), we propose individual cooling paths including the ZFT data of Devonian to Carboniferous age with possible exhumation rates between ~ 0.3 and ~ 0.1 mm/a. Therefore, post-Pan African cooling of the northern central area of the CKZ permits a similar stable phase showing very slow cooling of < 0.5 °C/Ma until the Late Devonian/Early Carboniferous.

Further south in the central area of the WKZ and CKZ [Figs. 10, 11, orange (3); Table S5.3], the geological record comprises a large set of solid geochronological data that allow constraining the differentiated cooling paths for the individual crustal segments. Based on the data, we suggest four individual cooling paths for the western Coastal Terrane during the syn-to-post-Pan African t–T-evolution (Fig. 11, orange (3), sky blue lines) that join together from the Late Cambrian on for the post-Pan African exhumation. U–Pb ages between ~ 705 and ~ 690 Ma indicate pre-Pan African deformation crystallization (Seth et al. 1998; Kröner et al. 2004) that are combined with ⁴⁰Ar/³⁹Ar ages between ~ 540 and ~ 520 Ma (Foster et al. (2009) and constrain this separate pre-deformation cooling path (1) from ~ 700 to ~ 520 Ma. Due to the lack of further data, we assume a homogeneous cooling with exhumation rates of ~ 0.1 mm/a (1) (Fig. 12a, WKZ and CKZ, Coastal Terrane). Younger U–Pb zircon and monazite ages indicate syn-Pan African crystallization between ~ 660 and ~ 620 Ma, and from ~ 580 to ~ 510 Ma (Seth et al. 1998; Goscombe et al. 2005a, b; Konopásek et al. 2008) that either can be combined with a Sm–Nd age of ~ 534 Ma (path 2) (Jung et al. 2014), or an ⁴⁰Ar/³⁹Ar hornblende age of ~ 546 Ma (path 3) (Foster et al. 2009) to provide two distinct cooling paths. In both cases, the cooling shows a relatively slow initial cooling with exhumation rates of ~ 0.04 mm/a that is followed by faster exhumation with rates between ~ 0.3 and ~ 0.2 mm/a until the Early Paleozoic (Fig. 12a, WKZ and CKZ, Coastal Terrane, 2, 3, and 4).

In the Orogen Core (Figs. 10, 11, orange (3), dark blue lines), syn-Pan African crystallization is recorded by U–Pb ages ranging between ~ 580 and ~ 540 Ma (Seth et al. 1998; Goscombe et al. 2005a, b) that correlate with Sm–Nd metamorphic ages between ~ 580 and ~ 570 Ma (Goscombe et al. 2003a, b). Syn-Pan African cooling set in with Sm–Nd cooling ages between ~ 550 and ~ 500 Ma (Goscombe et al. 2005a, b; Foster et al. 2009; Jung et al. 2014) followed by ⁴⁰Ar/³⁹Ar ages on hornblende, and on muscovite and biotite around 530 Ma, and between ~ 530 and ~ 500 Ma, respectively (Goscombe et al. 2005a, b; Gray et al. 2006; Foster et al. 2009). Therefore, we suggest two relatively homogeneous t–T-paths including the syn-Pan African intrusion, metamorphism, and cooling of the rocks within the Orogen Core from Late Neoproterozoic to Early Cambrian showing both an average exhumation rate of ~ 0.3 mm/a (Fig. 12a, WKZ and CKZ, Orogen Core cooling path).

In the Escape Zone [Figs. 10, 11, orange (3), purple line], only one Sm–Nd age of ~ 574 Ma indicates syn-Pan African metamorphism in the CKZ (Goscombe et al. 2003b). The cooling of the Escape zone is recorded by ⁴⁰Ar/³⁹Ar hornblende data between ~ 530 and ~ 520 Ma (Foster et al. 2009) correlating with the data of the Orogen Core that also yields for the ⁴⁰Ar/³⁹Ar data on muscovite and biotite (~ 527–519 Ma, Gray et al. 2006). Consequently, we suggest a single average cooling path representing the syn-Pan African metamorphism and cooling of the Escape Zone rocks from the Late Neoproterozoic to Late Cambrian (0.1 mm/a).

For all three crustal segments (Coastal Terrane, Orogen Core, Escape Zone), ⁴⁰Ar/³⁹Ar data on muscovite and biotite show Cambrian Age, and therefore, we assume relatively simultaneous post-Pan African cooling from the Late Cambrian to Early Carboniferous for the central area of the WKZ and CKZ (orange). We suggest the cooling paths for the crustal segments connecting either ZFT data > 380 Ma, or between ~ 360 and ~ 310 Ma, while segments of the WKZ generally tend to be older. Consequently, the cooling paths connecting ages > 380 Ma allow a more direct exhumation to low-temperature areas with rates of ~ 0.05 mm/a. When combined with the younger ZFT ages, the individual cooling paths rather correlate with the northern areas of the WKZ and CKZ (green and blue), also indicating a stable phase at temperatures between 300–200 °C with exhumation rates of < 0.01 mm/a.

In the SKB [Figs. 10, 11, yellow (4), Table S5.4], U–Pb crystallization ages on zircon and monazite are recorded between ~ 550 and ~ 500 Ma (Van de Flierdt et al. 2003; Schmitt et al. 2012; Milani et al. 2015; Goscombe et al. 2017) followed by Sm–Nd metamorphic ages around ~ 491 Ma that can be connected to K/Ar and ⁴⁰Ar/³⁹Ar data between ~ 515 and ~ 490 Ma (Gray et al. 2006; Goscombe et al. 2007, 2017). We suggest a t–T-path including intrusion/crystallization, metamorphism, and cooling of the Orogen Core rocks in the Ugab Zone. Syn-Pan African cooling set in from Late Neoproterozoic/Early Cambrian with rates of ~ 0.1 mm/a. Then, exhumation rates increased to ~ 1.2 mm/a during the Late Cambrian, and dropped again to 0.06 mm/a during the Early Ordovician (Fig. 12b, SKB, Ugab Zone, Intr.-met.-cooling path). Seth et al. (2000) recorded a single U–Pb around ~ 456 Ma of a Voetspoor syenite that represents a post-Pan African deformation intrusion within the Ugab Zone. This age can be connected to ⁴⁰Ar/³⁹Ar ages of ~ 496 Ma and ~ 438 Ma for the same location (Voetspoor contact aureole schist, Gray et al. 2006). Therefore, we suggest a post-deformation t–T-path for the Voetspoor intrusion with cooling rates > 20 °C/Ma for the first ~ 20–30 Ma followed by slower cooling with exhumation rates of ~ 0.07 mm/a (Fig. 12b, SKB, Ugab Zone, post-def. path). Nevertheless, we assume the first intrusion-metamorphism-cooling path as more probable representing the t–T-evolution of the Ugab Zone. We propose exhumation rates of ~ 0.06 mm/a, and 0.07 mm/a (post-def. path) that indicate relatively simultaneous cooling of the entire area from Silurian to Late Carboniferous.

In the NCZ of the southern Damara Belt, crystallization and cooling of rocks are recorded for three individual locations (Figs. 10, 11, dark red (5); Tab. S5.5). At the Salem granitic suite, Miller and Burger (1983) record a U–Pb crystallization and metamorphic age for the Oetmoed granite around ~ 590 Ma (Milani et al. 2015). U–Pb crystallization and metamorphic ages on monazite range between ~ 538 and ~ 470 Ma, and is backed with Sm–Nd metamorphic ages of ~ 510 Ma and ~ 473 Ma (Jung et al. 2000a, b). Unfortunately, there do not exist any ZFT ages and the existing AFT data are thermally overprinted by younger events. We suggest two separate intrusion–metamorphism-cooling paths (Fig. 12b, Damara Belt, NCZ, intrusion-met.-cooling path) for the entire syn- and post-Pan African period until the final exhumation to the surface during the Late Carboniferous (Miller 2008). In both cases, the cooling path sets in during the Late Neoproterozoic with rates of ~ 0.07 mm/a (1) and ~ 0.05 mm/a (2), and rocks experienced the final exhumation with rates of ~ 0.1 mm/a until the Late Carboniferous.

A second intrusion is recorded for the Okombahe granite showing Rb–Sr ages between ~ 553 and ~ 514 Ma, and from ~ 495 to ~ 465 Ma (Haack and Martin 1983). Unfortunately, there are no data to further constrain the Okombahe granite cooling path, and therefore, we propose average cooling paths beginning either with the Rb–Sr ages > 514 Ma, or the Rb–Sr ages < 495 Ma and present long-time cooling to surface near areas during the Late Carboniferous with long-term exhumation rates between ~ 01 and ~ 0.08 mm/a [Fig. 12b, Damara Belt, Okombahe granite cooling path (1 and 2)].

For our samples close to the Waterberg Basin east of Otjiwarongo (NA11-09, -10, -11), we would like to introduce an Otjiwarongo area cooling path that is only recorded by a U–Pb age of ~ 550 Ma (Lobo-Guerrero Sanz 2005), and, therefore, is poorly constrained. However, we present a long-term exhumation rate from the Late Neoproterozoic to the Late Carboniferous with an average rate of ~ 0.1 mm/a (Fig. 12b, Damara Belt, NCZ, Otjiwarongo cooling path).

In the SZ of the Damara Belt (Figs. 10, 11, red (6), Tab. S5.6), published U–Pb crystallization and metamorphic ages on monazite range between ~ 527 and ~ 505 Ma (Kukla et al. 1991) followed either by syn-Pan African Sm–Nd cooling ages between ~ 545 to ~ 511 Ma, or late- to post-Pan African cooling Sm–Nd and Rb–Sr ages between ~ 500 and 488 Ma, and ~ 490 and ~ 470 Ma, respectively (Jung and Mezger 2001). The cooling path is continued by ⁴⁰Ar/³⁹Ar ages on hornblende (~ 505 Ma), and muscovite and biotite (~ 496–482 Ma, Gray et al. 2006). Therefore, we suggest three individual syn-to post-Pan African cooling paths from the Late Neoproterozoic to Late Carboniferous. The first path includes the first Sm–Nd ages (~ 545–511 Ma) and connects to the ⁴⁰Ar/³⁹Ar ages showing an exhumation rate until the Late Cambrian of initially ~ 0.2 mm/a, and then 0.05 mm/a for the post-orogenic period [Fig. 12b, Damara Belt, SZ, (1)]. A second possible cooling path connects to the group of younger Sm–Nd ages (< 500 Ma), and combines the Rb–Sr cooling ages with the ⁴⁰Ar/³⁹Ar ages that result in a differentiated cooling with exhumation rates of ~ 0.6 mm/a followed by post-orogenic exhumation with rates of ~ 0.06 mm/a [Fig. 12b, Damara Belt, SZ (2)]. The third possible t–T-path follows the post-Pan African Rb–Sr cooling ages showing exhumation rates of initially ~ 0.02 mm/a, and then continues with ~ 0.1 mm/a [Fig. 12b, Damara Belt, SZ (3)].

In general, syn-to post-Pan African crustal cooling and exhumation correlates for the northern Kaoko Belt within in the WKZ and CKZ [Figs. 10, 11, green (1), orange(3)]. While intrusion and crystallization occurred earlier in the Coastal Terrane, syn-Pan African cooling coincides with cooling paths of the northern and central Orogen Core showing fast cooling until the end of the Cambrian. This also yields for the Ugab Zone [Fig. 11, yellow (4)], as part of the Orogen Core in the SKB where rocks experienced fast Early Cambrian syn-Pan African cooling to temperatures around 300–200 °C. Within the northern Kaoko Belt, ZFT data allow constraining the post-orogenic t–T-paths that show decreasing cooling rates from the Late Cambrian on indicating an almost thermally stable period during the Early Paleozoic (~ Late Cambrian to ~ Early Carboniferous). This correlates with the t–T-evolution for the SE Brazilian Florianopolis, Peruibe, Curitiba, and Ilha Comprida blocks. Krob et al. (2019) revealed high exhumation rates between ~ 0.4 and ~ 0.2 mm/a during the Late Neoproterozoic, followed by an almost stable phase with exhumation rates between ~ 0.02 and ~ 0.01 mm/a reaching from the Early Cambrian to the Devonian.

Unfortunately, ZFT data lack for the Ugab Zone, and therefore, the post-Pan African cooling during the Early-to-Late Paleozoic is less constrained. Nevertheless, a similar t–T-evolution showing an almost stable period without a decrease in temperature during the post-Pan African cooling as for the northern Kaoko Belt and in SE Brazil cannot be ruled out. However, we propose to enlarge the thermochronological data set, especially the ZFT data to better constrain the post-orogenic cooling as such data would provide the transition from high rock temperatures involved during syn-orogenic processes to lower temperature areas (< 300 °C) of the upper crust and near the Earth’s surface providing more information about the post-orogenic evolution.

Consequently, the t–T-evolution of the northern central CKZ [Figs. 10, 11, blue (2)] remains less constrained due to the lack of geochronological and thermochronological data. We present a poorly constrained cooling path that differs from the WKZ and CKZ in the northern Kaoko Belt indicating the main crustal cooling during the Early Paleozoic.

All locations within the southern Damara Belt show similar syn- and post-Pan African crustal cooling and exhumation. Generally, geo- and thermochronological ages are slightly younger as in the Kaoko Belt indicating that crustal cooling set in later, and therefore, rocks stayed longer in high temperature ranges (> 300 °C). This t–T-evolution is also illustrated by the interpolation maps that provide a possible visualization of the Neoproterozoic basement thermal structure. During the Ediacaran-to-Early Cambrian t–T-evolution, basement rocks show a relatively homogeneous thermal structure (Fig. 13b). Basement temperatures of the southern Damara Belt correlate with those of the Kaoko Belt showing high temperatures (~ 750–600 °C) for the entire study area. The visualized t–T-evolution of the thermal structure implies that basement rocks in the northern Kaoko Belt cooled down earlier than within the SKB and Damara Belt showing a trend of decreasing temperature within the study area beginning in the north and followed by the south (Fig. 13b–d). We assume that an inner core of the Damara Belt remained in high temperatures the longest (Fig. 13d), and then cooled down faster than the Kaoko Belt until the Devonian (Fig. 13e) showing a more inhomogeneous temperature distribution of the basement surface.

Late Paleozoic-to-Late Jurassic—SW Gondwana intraplate basin formation, subsidence, and South Atlantic pre- to syn-rift processes

Following the post-Pan African denudation processes rock cooling, and surface uplift and erosion continued during the Late Paleozoic-to-Mesozoic Gondwana intraplate environment. The t–T-paths show exhumation of all samples to areas close to the surface by the end of the Carboniferous. Therefore, temperatures of the basement thermal structure range in near surface areas between ~ 25 and 20 °C. We assume that all crustal segments of the Kaoko and Damara belts experienced a relatively homogeneous t–T-evolution from the Late Paleozoic, and, therefore, describe a continued t–T-evolution including exhumation and subsidence rates for the individual areas regardless the different Pan African crustal segments.

ZFT ages and old AFT ages of this study (Figs. 6, 8) range between ~ 430 and 310 Ma, and ~ 390 and 280 Ma, respectively, indicating significant tectonic and surface uplift, and erosion during the Devonian-to-Permian. We assume glacial erosion and transport of sedimentary material, and a temporally lower geothermal gradient during the Late Paleozoic Gondwana Ice House (Montañez and Poulsen 2013) to have triggered the final Paleozoic cooling and surface uplift of the Neoproterozoic basement resulting in the displayed ZFT and AFT age distribution. This correlates with the increase of sedimentation and continuous formation of the surrounding sedimentary basins. Montañez and Poulsen (2013) observed significant amounts of sediments transported towards the southern and the western parts of the Gondwana intraplate environment. Stratigraphic records of the major African and South American basins indicate deposition of Upper Paleozoic-to-Lower Mesozoic glacial and siliciclastic sedimentary rocks (Stollhofen et al. 1999; Catuneaunu et al. 2005; Milani et al. 2007; Uliana et al. 2014; Krob et al. 2019) supporting the removal of enormous amounts of sediments. Before the Gondwana Ice House, the thermal structure of the Neoproterozoic basement (Fig. 13e) reveals temperatures between ~ 180 and ~ 120 °C suggesting an overburden thickness between ~ 3.5 and ~ 5.5 km that would have been removed by the glacial erosion until the end of the Carboniferous. However, precise calculations of removed sediment thicknesses which might have been even larger due to the temporally decreased geothermal gradient during the glacial period are still lacking.

Nevertheless, most likely glacial driven cooling, surface uplift, and erosion are recorded by renewed fast exhumation rates between ~ 0.8 and ~ 0.2 mm/a in the WKZ and CKZ of the northern Kaoko Belt during the Early-to-Late Carboniferous (Fig. 12a, b). Thus, wide areas of the Kaoko Belt show a threefold division of Paleozoic exhumation that is also observed in the Late Neoproterozoic-to-Late Paleozoic t–T-evolution of the SE coastal region of Brazil (Krob et al. 2019).

Sedimentation and intraplate subsidence recommenced in the Kaoko and Damara belts during the Late Carboniferous to Late Permian with the formation of the major SW Gondwana intraplate basins (Stollhofen et al. 1999; Catuneanu et al. 2005; Milani et al. 2007). Increasing temperatures of the Neoproterozoic basement rocks reflect the deposition of sedimentary material (Fig. 13f–h). Slow subsidence occurred with rates between ~ 0.02 and ~ 0.01 mm/a (~ 330/310–195 Ma) until the Early Jurassic (Fig. 12a, b). The ongoing subsidence stopped during the Early Jurassic recorded by a hiatuses in the stratigraphic records. We assume predominant pre- to syn-rift, possibly plume-driven, surface uplift, and erosion during this period (Courtillot et al. 1999; Şengör 2001; Hu et al. 2018; Krob et al. 2019). Exhumation rates range between ~ 0.04 and ~ 0.02 mm/a (~ 155–135 Ma).

Late Jurassic to recent—SAPCM syn- to post-rift processes, and the thermal influence of the Paraná–Etendeka LIP

The emplacement of Paraná–Etendeka dike swarms, local volcanic activity, and the eruption of flood basalts dominate the Late Jurassic-to-Early Cretaceous t–T-evolution of the SW Gondwana intraplate environment. We assume that rocks were either locally influenced for a relatively short time by the convective heat flow (geothermal gradient) caused by the crosscutting dike swarms and accompanied local volcanic activity, and/or reheated due to the overburden thicknesses of the erupted flood basalts. Thus, the calculation of subsidence rates remains uncertain. However, Neoproterozoic rocks show heating rates between ~ 1.2 and ~ 4.5 °C/Ma (~ 155–135 Ma) during pre-to syn-rift processes indicating higher rates towards the south, and in direct proximity to contemporary Etendeka flood basalt depositions. The thermal basement structure reveals maximum temperatures of ~ 80–120 °C between ~ 135 and ~ 120 Ma reaching higher temperatures in the southern Damara Belt (Fig. 13j, k) (Brown et al. 2014). However, maximum temperatures for NW Namibia do not reach temperatures as high as on the SE Brazilian side where the basement rocks range between ~ 80 and 160 °C in the northern regions and over 200 °C in the southern region (Krob et al. 2019). Consequently, the Neoproterozoic basement shows relatively slow, post-eruption, thermal recovering, and rock cooling during the Early Cretaceous indicating a thermally reestablished upper crust that also has been observed in SE Brazil (Hu et al. 2018; Krob et al. 2019).

The ZFT and inland AFT ages (Fig. 14), although younger than their stratigraphic age, did not experience any thermal overprint caused by volcanic activity during the Early Cretaceous Paraná–Etendeka Large Igneous Province and, therefore, correlate with observations of earlier studies (Raab et al. 2002, 2005; Brown et al. 2014). Nevertheless, we assume that AFT ages of intermediate age (~ 300–200 Ma) might have been partially influenced by the Paraná–Etendeka event and probably would have been older originally (> 300 Ma). However, those ages reveal a medium track length distribution when plotted against their corrected mean confined track lengths (Fig. 15a, c) and accordingly indicate a mixture of pre- and post-cooling track accumulation (Green 1986; Lewis et al. 1992; Gallagher et al. 1997, 1998; Green et al. 1998). The youngest AFT ages along the coast (Fig. 15a, c), and within the Damara Belt (b, d) are fully reset showing relatively long mean track lengths and reflect a significant Paraná–Etendeka thermal influence (Fig. 15a, b). However, the so-called boomerang plots (Green 1986; Gallagher et al. 1998) do not confirm an idealized pattern and indicate that the rocks experienced different amounts of cooling and a denudation history more complex than a single-stage rift elevated event (Gallagher et al. 1998). Nevertheless, we are able to recognize a major post-Paraná–Etendeka cooling event during the Cretaceous, but also assume long and complex and/or sequential and transient denudation processes following the South Atlantic break-up due to a wide range of mean track lengths (Green 1986; Gallagher et al. 1998; Raab et al. 2002; Brown et al. 2014). On the SE Brazilian side, boomerang plots (Fig. S6) reveal results that allow assumptions of a similar complex multi-stage rift evolution for the Early-to-Middle Cretaceous (100–80 Ma) (Gallagher et al. 1988; Krob et al. 2019). Whereas in SE Brazil, regional tectonic and volcanic activity, and local graben sedimentation indicates that the basement rocks most likely reached the Earth’s surface at the end of the Cretaceous, evidence supporting a similar geological evolution lacks for the NW Namibian counterpart. Therefore, we only provide exhumation rates for the entire post-rift to recent phase with rates ranging between ~ 0.01 and ~ 0.03 mm/a (Fig. 12a, b).

Fig. 14

Detailed interpolation map showing a possible modeled projection of the major apatite fission-track (AFT) data sets from Angola to NW Namibia using the Golden Software Surfer^®. Whereas yellow-to-red colors indicate younger ages, green-to-blue colors imply older ages. Isolines show age in [Ma] without error. Colored dots (Archean-to-Early Paleozoic basement rocks), squares (Karoo sedimentary rocks), and triangles (Lower Cretaceous sedimentary and volcanic rocks) show sample locations colored according to the references. ST Sesfontein thrust, PMZ Purros mylonite zone, TPSZ Tree palms shear zones, OT Otjohorongo thrust, OmL Omaruru lineament, OL Okahandja lineament

Fig. 15

Boomerang plots showing (a, b) the apatite fission-track (AFT) data set of this research study plotted against their corrected mean confined track lengths (L_c) for the NW Namibian Kaoko and Damara belts, and c, d the integration of our AFT data to results of previous AFT data from NW Namibia by Raab et al. (2002, 2005) and Brown et al. (2014) (right). For the integration, we plotted the AFT ages against their (uncorrected) mean confined track lengths (L₀). While old ages with long track lengths most likely reflect “source” ages that have not reached high post-depositional annealing temperatures, young ages having long track lengths reflect totally reset cooling ages. Samples showing both intermediate AFT ages and mean confined track lengths possibly represent “mixed” ages having a mixture of pre- and post-cooling track lengths accumulation (Green 1986, Lewis et al. 1992; Gallagher et al. 1997, 1998)

Subsequently, results of this study add to previously published observations and contribute significantly to the general understanding of the AFT data showing most of the totally reset ages with long mean track lengths congruent with the Paraná–Etendeka dike swarm distribution in the southern Damara Belt (Figs. 14, 15c, d). Furthermore, results show that the projected temperature isolines reflect the major cratonic boundaries in NW Namibia. Therefore, we propose that the old cratonic structures and individual crustal segments are still stable. Whereas the isolines do not point out any differences across the major lineament zones and, therefore, do not allow any interpretation of regional movement or reactivation along the NE-SW trending shear zones in the Damara Belt, isolines in the Kaoko Belt follow significantly the NW–SE-trending shear zones, i.e., the ST, PMZ, and the TPSZ (Fig. 14, and Fig. 7 in detail). Moreover, the course of isolines illustrates the significant age increase towards the inland and different existing age spectra in the subdivided longitudinal Eastern, Central, and Western (Kaoko) zones. This supports the possible reactivation and movement of the Precambrian shear zones during the South Atlantic break-up, and syn-to post-rift stages (Brown et al. 2014; Salomon et al. 2015, 2017).

Concluding, the modeled projection of our AFT data combined with previous AFT studies along the Angolan-to-Namibian and SE Brazilian SAPCM (Fig. 16; for further details, see Fig. 14 and supplementary material, Fig. S7) outline areas showing high thermal overprint by the Paraná–Etendeka LIP (ages <  Paraná–Etendeka age) that coincide with the lateral Paraná–Etendeka dike swarm distribution indicating a highly asymmetric plume central area (Torsvik et al. 2009, 2014; Franke 2013) with possibly two different eruption sources (black circles): (1) the Ponta Grossa Arc (PGA) together with the area north of the NW Namibian Etendeka flood basalt depositions, and (2) the Rio Grande do Sul (RGdS) dikes together with Damara Belt dike swarms in the south. On the South American side, large parts of the flood basalts are still preserved. Where volcanic rocks are eroded, dike swarms appear crosscutting the basement and its overlying sedimentary rocks. Generally, the AFT ages reflect these geological observations on the Earth’s surface reaching further inland on the South American SAPCM. A wide area of young ages occurs north of Rio de Janeiro (RdJ) where no evidence of erupted flood basalts and/or dike swarms exists and rocks rather experienced a period of intensified post-rift surface uplift and erosion than a thermally induced overprint (Jelinek et al. 2014). Milani et al. (2007) also describe large amounts of sedimentary material within the offshore records that support periods of enormous exhumation for the northern SE Brazilian SAPCM. On the opposite continental margin, only a few published data exist along the central-to-northern Angolan continental margin, and volcanic remnants of the Paraná–Etendeka LIP are missing. Therefore, the contemporary data set does not allow further interpretation. In general, thermochronological data provide an excellent archive to better understand the thermal influence caused by the Paraná–Etendeka LIP. More data covering a wider range especially in the north (Angola), and the south would be needed to support assumptions of a greater lateral sphere of originally lava covered area (Jackson et al. 2000; Torsvik et al. 2009, 2014).

Fig. 16

Detailed view of the Lower-to–Upper Cretaceous SAPCM of Africa and South America (top left corner): a Comparison of the interpolation maps showing the modeled projection of the AFT age distribution of the major thermochronological data sets along the SAPACM in SE Brazil [Gallagher et al. 1994, 1995 (red); Tello Saenz et al. 2003 (blue); Hiruma et al. 2010 (turquoise); Franco Magalhaes et al. 2010 (brown); Jelinek et al. 2014 (light green); Engelmann de Oliveira et al. 2016a, b (anthracite); Karl et al. 2013; Krob et al. 2019 (yellow)]; Angola [Green and Machado 2015 (green); Silva et al. 2019 (light grey)), and Namibia (Raab et al. 2002, 2005 (purple); Brown et al. 2014 (black); and this study (white)]. Circles: Archean-to-Early Paleozoic basement rocks; squares: Upper Paleozoic-to-Lower Mesozoic sedimentary rocks; triangles: Lower Cretaceous sedimentary and volcanic rocks. Paleo-geographic reconstructions were taken from Torsvik et al. (2009, 2014). For a more detailed view of the separated margins and individual sample locations, see Fig. 14, and supplementary material, figure S3. The modeled interpolation maps were generated using the Golden Software Surfer^®. Whereas yellow-to-red colors indicate younger ages, green-to-blue colors imply older ages. Isolines show age in [Ma] without error. Colored dots show samples locations. b Regional Paraná–Etendeka related dike swarm distribution (white bars) and contemporary flood basalt depositions (white areas) (after Renne et al. 1996; Peate et al.1997; Marzioli et al. 1999; Thompson et al. 2001; Trumbull et al. 2004, 2007; Gibson et al. 2006; Franco Magalhaes et al. 2010; Travasso 2014; Florisbal et al. 2014; Guedes et al. 2016; Hartmann et al. 2016) combined with the AFT age distribution along the SAPCM in SE Brazil, and Angola to NW Namibia

Conclusions

This research study focused on the coherent long-term t–T-evolution of the Neoproterozoic basement rocks of the NW Namibian Kaoko and Damara belts. We presented an overview of new thermochronological data and combined them with the existing geochronological data, such as U–Pb, Sm–Nd, and Rb–Sr analyses, and low-temperature thermochronology (LTT) data from K/Ar, ⁴⁰Ar/³⁹Ar analysis. Thereby, we propose a possible coherent long-term t–T-evolution of the Neoproterozoic basement rocks compatible with both, the published geological record, i.e., stratigraphic sections and geochronological ages, and our new thermochronological data.

Modeled t–T-histories correlate with the t–T-evolution of the Neoproterozoic basement rocks of the SE coastal region in Brazil. Models show fast syn-to-post-Pan African orogenic cooling and exhumation of the basement rocks during the Late Neoproterozoic-to-Early Paleozoic. New ZFT data range between ~ 430 and ~ 310 Ma for the rocks of the Precambrian Kaoko Belt and likely indicate a period of intense erosion during the Late Paleozoic Gondwana glaciation. The Late Paleozoic-to-Late Jurassic t–T-evolution is characterized by major intraplate basin formation and the most likely plume interacted pre-to syn-rift, and South Atlantic break-up processes, such as rock and surface uplift, and erosion. During the Paraná–Etendeka volcanic activity, the Neoproterozoic basement in NW Namibia reached maximum temperatures between ~ 80 and ~ 120 °C, and, therefore, experienced less thermal influence as the SE Brazilian counterpart (up to > 200 °C). Following, the t–T-reconstruction indicates a relatively slow, post-eruption thermal recovering and rock cooling during the Lower Cretaceous.

AFT ages along the coast and within the Damara Belt are fully reset by the Paraná–Etendeka volcanic event and, mostly range between ~ 150 and ~ 60 Ma, which correlates with the previous studies showing a long and complex post-rift cooling phase. Further inland ages increase significantly, and older AFT data may represent partially reset ages. AFT age distributions extending from Angola to Namibia and from eastern to south-eastern Brazil correlate for the African and South American SAPCM and coincide with the lateral Paraná–Etendeka dike swarm distribution. Finally, we also suggest a greater lateral sphere of originally lava covered area. However, more data covering a wider range in the north, and the south would be needed to fully understand the thermal influence caused by the Paraná–Etendeka LIP.