Present-day thermal field and Mesozoic-Cenozoic thermal evolution of the Western Bredasdorp Basin (South Africa): An integrated 3D numerical forward modelling approach

Accepted Manuscript Present-day thermal field and Mesozoic-Cenozoic thermal evolution of the Western Bredasdorp Basin (South Africa): An integrated 3D numerical forward modelling approach W.A. Sonibare, J. Sippel, R. di Primio, Z. Anka, M.Scheck Wenderoth, D. Mikeš PII:

S0264-8172(18)30074-6

DOI:

10.1016/j.marpetgeo.2018.02.028

Reference:

JMPG 3256

To appear in:

Marine and Petroleum Geology

Received Date: 1 November 2017 Accepted Date: 19 February 2018

Please cite this article as: Sonibare, W.A., Sippel, J., di Primio, R., Anka, Z., Wenderoth, M.S., Mikeš, D., Present-day thermal field and Mesozoic-Cenozoic thermal evolution of the Western Bredasdorp Basin (South Africa): An integrated 3D numerical forward modelling approach, Marine and Petroleum Geology (2018), doi: 10.1016/j.marpetgeo.2018.02.028. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

ACCEPTED MANUSCRIPT Journal: Marine and Petroleum Geology Title: Present-day thermal field and Mesozoic-Cenozoic thermal evolution of the Western Bredasdorp Basin (South Africa): an integrated 3D numerical forward modelling approach Authors: W. A. Sonibare 1,2(*), J. Sippel 2, R. di Primio 3, Z. Anka 4, M. Scheck-Wenderoth 2, and D. Mikeš 5 1

2

RI PT

Department of Earth Sciences, Stellenbosch University, Private Bag X1, Matieland 7602, Stellenbosch, South Africa Helmholtz Centre Potsdam, GFZ German Research Centre for Geosciences, Telegrafenberg, 14473 Potsdam, Germany Lundin Norway AS, Strandveien 4, N-1366 Lysaker, Norway

4

Total Exploration – New Ventures, 92078 Paris – La Défense 6, France

SC

3

5

M AN U

Department of Geosciences, Nelson Mandela Metropolitan University, PO Box 77000, Port Elizabeth 6031, South Africa (*) now at: Schlumberger Limited, Lagos, Nigeria

TE D

Corresponding author at: Department of Earth Sciences, Stellenbosch University, Private Bag X1, Matieland 7602, Stellenbosch, South Africa. Email: [email protected]; [email protected] Keywords: Bredasdorp Basin, 3D structural model, numerical modelling, crustal thermal field, thermal evolution

EP

ABSTRACT

AC C

We analyse a crust-scale 3D structural model of the Western Bredasdorp Basin (WBB) located offshore southern South Africa to assess both the spatial and temporal development of geothermal gradients in the Southern African ‘passive-transform’ continental margin. The results of simulations of the present-day conductive thermal field were used to constrain the heat flow evolution during 3D numerical modelling of the paleo-thermal regime. Besides conforming to the regional geotectonic framework, the ensuing models are largely consistent with measured temperatures. The largest control on present-day temperatures is exerted by the basin’s present-day geometry and distribution of thermal properties which, in turn, trace back to the earliest tectonic event involving crustal thinning and significant syn-rift sedimentation. Implying from the modelled shallow and deep geotherms, we find that the variations of present-day surface heat fluxes in the WBB are much more dictated by the long-term crustal radiogenic heat production in concert with the interaction of low thermally conductive sediments and high thermally conductive crustal rocks rather than by the present-day mantle heat flow. In addition, model calibration with measured vitrinite reflectance trends suggests that three major phases of thermal disequilibrium resulting in high heat flow during syn-rift and post-rift times characterise the distribution of maximum paleo-temperatures and paleo-

2

ACCEPTED MANUSCRIPT

RI PT

geotherms. In terms of subsidence, the syn-rift tectonic event plays the most significant role in basin evolution. The height and width of the syn-rift heat flow anomalies and their lateral variations reflect a direct consequence of instantaneous crustal stretching leading to rapid burial and lateral differences in crustal thicknesses and amounts of radiogenic heat produced. The post-rift heat flow peaks are interpreted to be related to a Late Cretaceous mantle-related hotspot event and a late Miocene episode of margin uplift. As indicated by validated vitrinite trends and the modelled distribution of eroded sediments, the hotspot event represents a more localised thermal event, while the late Cenozoic erosion and inferred uplift depicts a regional character. The modelled evolution of deep heat flow further suggests that ancient thermal perturbations have little influence on the present-day thermal field.

SC

1. Introduction

AC C

EP

TE D

M AN U

Sedimentary basins documenting signs of both crustal extension and strike-slip faulting, such as those bordering the Agulhas Falkland Fracture Zone (AFFZ) (Fig. 1), are well known from intracontinental and continental margin environments worldwide (e.g. Madon and Watts, 1998; Allen and Allen, 2005; Xie and Heller, 2009). Due to their complex tectonic and structural histories, the understanding of possible causes and processes of thermal-related basin mechanisms are often problematic to reconstruct. Located on the southern shelf of South Africa, the initiation of the Western Bredasdorp Basin (WBB; Fig. 1a-b) was associated with the mid-late Jurassic lithospheric separation of west Gondwana into South America and Africa (Fig. 2). This continental break-up and the subsequent thermal relaxation resulted in the development of syn-rift graben and half-graben structures being superposed by post-rift sedimentary successions (e.g. Dingle et al., 1983; Thomson, 1998; McMillan et al., 1997; Broad et al., 2006). Thus, the resultant effect of these complex geo-tectonic processes is the accumulation of sediments (locally reaching ~7 km in the WBB) that are not only of high significance for hydrocarbon prospectivity but also providing an archive of basin evolution. The attendant thermal disequilibrium and erosion activities of such complex rift-transformpassive margin process are still poorly understood in southern Africa (e.g. Dingle et al., 1983; Nyblade, 2003; Johnson et al., 2006; de Wit, 2007; Kounov et al., 2007; Tinker et al., 2008). Consistent with well, seismic and gravity data, Sonibare et al. (2014) present a crust-scale 3D geological model of the WBB that enables direct correlation of the deep crust and shallow sedimentary structure, typically showing characteristics of passive margin settings, i.e. synrift faulted sediments overlying a thinned crust (Figs. 3 and 4). However, differing directions of extension as inferred from (i) syn-rift normal faults (striking predominantly NW-SE) and (ii) thinned sub-sedimentary crust (with W-E striking Moho high) suggests a crustal thinning mechanism that is difficult to explain just by a single-phased, pure, and margin-perpendicular extension. Besides, the Moho high is very narrow spanning only ~ 60-70 km. Sonibare et al. (2014) therefore interpret the WBB as having been initiated under a transtensional setting with the corresponding component of strike-slip being related to movements along the AFFZ. To further deepen our understanding of the dominant basin-forming mechanisms in this domain of the Southern African continental margin, this study explores how the WBB has evolved thermally since the rifting phase. Over the last three decades, the study of past and present-day thermal conditions of sedimentary basins has significantly advanced through the application of numerical basin

3

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

modelling techniques, involving iterative simulations of physical processes and related lithology-dependent parameters within the limitations of specific geologic boundary conditions (Yükler et al., 1979; Tissot and Welte, 1984; Ungerer et al., 1990; Hermanrud, 1993; Poelchau et al., 1997; Bayer et al., 1997; Yalçin et al., 1997; Hantschel and Kauerauf, 2009; Maystrenko et al., 2013). A great challenge in performing such thermal modelling studies for a specific region is to reasonably parameterise the modelling domain and choose appropriate boundary conditions to ensure that the resulting present-day and past thermal fields are consistent with observations such as measured temperatures and vitrinite reflectance data, respectively. Therefore, in this study we have integrated the abundance of exploration data (resulting from intense oil and gas prospecting that dated back to the early 1970s in South Africa; Fig. 1b) and the existing 3D model (Sonibare et al., 2014) into a three-dimensional modelling workflow. The Petroleum Agency of South Africa (PASA) and the Petroleum Corporation of South Africa (PetroSA) provided all seismic and well data used for this study. Firstly, we calculated the present-day crust-scale 3D conductive thermal field as is consistent with measured borehole temperatures. Thereafter, we used the data-consistent present-day thermal field as a major constraint for 3D forward modelling of the paleo-temperature evolution since rifting times. The forward modelling approach numerically simulates deposition, nondeposition and erosion in time while considering evolving sediment compaction trends and thermal conditions. Throughout the modelling procedure, heat generation by radioactive decay is fully taken into account. Model outputs in terms of present-day temperature distribution as well as paleo-heat flow and erosion predictions are calibrated with measured present-day temperatures and vitrinite reflectance data, respectively. The pattern of the past and present-day thermal field and the responsible dominating factors for these variable distributions are assessed. In addition, the modelling approach allows us to establish the linkage between crust-scale boundary conditions and the thermal evolution within the basin. Thus, the results of our studies offer new insights into: (1) the distribution of present-day surface heat flow and geothermal gradient in relation to the possible contributing factors (i.e. understanding the interplay of deep crustal and shallow sedimentary influences), (2) the role of syn-rift tectonics and post-rift thermal perturbations in the distribution of maximum paleotemperatures and paleo-heat flow, and (3) the timing and amount of eroded sedimentary strata including their lateral distribution. 2. Geological and tectonic setting 2.1. Geodynamic and stratigraphic evolution The deepest parts of the sub-sedimentary crust of southern South Africa consist of intensely deformed Proterozoic to Late Paleozoic sequences (consisting of metamorphic, metasedimentary and igneous rocks; Johnson et al., 2006; Fig. 1a). Unconformably overlying these deepest crustal rocks are the Ordovician-Devonian thick siliciclastic sequences of the Cape Supergroup, which were deposited during an intermittent phase of pre-rift transgressionregression cycles (Johnson et al., 2006) occasioned by continental collision and accretion in Southern Africa. During late Carboniferous to early Jurassic times, the Karoo Basin was formed as a foreland system with sedimentary infill dominantly composed of clastic sequences (e.g. Lock, 1980; Veevers et al., 1994; Johnson et al., 2006). Subsequently, a phase

4

ACCEPTED MANUSCRIPT

M AN U

SC

RI PT

of Permo-Triassic crustal shortening caused folding of the Cape Supergroup and possibly the earliest Karoo Supergroup sequences, to form the Cape Fold Belt as a major component of the Gondwanide orogeny (e.g. de Wit and Ransome, 1992; Visser and Praekelt, 1996). We combined the described knowledge about the sub-sedimentary crust with empirical relationships between seismic velocities and densities (as presented in Christensen and Mooney, 1995) to infer prevailing lithologies for the corresponding geological units (which provides the base for the parameterization of the thermal models, see below). Seismic velocities of the deep crust are indicated by the Agulhas-Karoo deep seismic profile of Parsiegla et al. (2009) located some 110 km east of the study area. Deep crustal densities (Fig. 4e) are derived from the gravity-constrained 3D model of Sonibare et al. (2014). Accordingly, a lower crustal unit located mostly at depths of >25 km is of felsic to intermediate composition, consisting of granodioritic-dioritic rocks as indicated by average P-wave velocities of 6.3-6.8 km/s and a modelled density of 2800 kg/m3. With average P-wave velocities of 5.7-6.2 km/s and a density of 2700 kg/m3 mostly located at 10-15 km depths, granitic-granodioritic rocks most likely dominate the upper crust. As documented by offshore wells, the shallower basement of the Bredasdorp Basin is formed by the Ordovician to Devonian Cape Supergroup metaclastics (Bokkeveld shales and slates) and Table Mountain metasandstones (McMillan et al., 1997; Johnson et al., 2006; Figs. 2 and 4). Average P-wave velocities of 5.0-5.5 km/s, a density of 2600 kg/m3, and a depth range of 5-10 km suggest that this pre-rift metasedimentary layer might be dominated by metagreywacke (Christensen and Mooney, 1995). The absence of Karoo Supergroup rocks in the study area may reflect either non-deposition (e.g. Biddle et al., 1986) or erosion (Cole, 1992).

AC C

EP

TE D

2.1.1. Syn-rift evolution and products In mid-late Jurassic times, the Gondwanaland rifting processes propagated from south (at ~136 Ma; Martin and Hartnady, 1986; Thomson, 1998) to north (at ~130 Ma; Broad et al., 2006). The complete lithospheric separation in southern South Africa was accompanied by dextral strike-slip movements along the AFFZ (Fig. 1) in the Early Cretaceous. These strikeslip movements overprinted the earlier, purely extensional traces to form the southern South African transform margin (Figs. 1 and 4a-b; e.g. McMillan et al., 1997; Broad et al., 2006). These rift-transform complexes comprise five W-E trending sub-basins that are separated by faults and basement highs and are collectively known as the Outeniqua Basin, and its deepwater southern extension, the Southern Outeniqua Basin (Fig. 1). The syn-rift successions comprise mainly continentally-derived fluvial and lacustrine sediments (claystones, sandstones and conglomerates) with intercalated shallow-marine glauconitic fossiliferous sandstones (McMillan et al., 1997; Roux, 2007). Following upward in the sequence, high-gamma ray deep-marine shales predominate (Fig. 3-4; Table 1; Sonibare et al, 2014). Important unconformities within the syn-rift interval are the Valanginian (1At) break-up unconformity, that stratigraphically demarcates the syn-rift infills into syn-rift 1 (Basement-1At) and syn-rift 2 (1At-5At and 5At-6At) phases, and the late Hauterivian (6At) unconformity, which overlies syn-rift sequences and marks the onset of post-rift sedimentation (Figs. 2, 3 and 4e). 2.1.2. Post-rift evolution and products The post break-up margin evolution in Southern Africa is a multiphase process. Firstly, two major mantle hotspots, i.e. the Bouvet and Shona hotspots traversed the Southern African

5

ACCEPTED MANUSCRIPT

M AN U

SC

RI PT

continental margin at ~80-60 Ma (e.g. Duncan, 1981; Martin, 1987; Gohl and Uenzelmann, 2001). Duncan (1981) provided a track of one of these hotspots beneath the Bredasdorp subbasin in late Cretaceous to early Cenozoic times with postulation of the inter-hotspot motion to be less than 5 mm/yr over the last 100 my. Previous studies (Duncan, 1981; Dingle et al., 1983; Martin, 1987; Davies, 1997b; Thomson, 1999; Gohl and Uenzelmann, 2001; Broad et al., 2006; Tinker et al., 2008) have variedly linked regional episodes of uplift, thermal disequilibrium as well as emplacements of alkaline intrusive bodies during latest Cretaceous and early Cenozoic times to these hotspot events. Several authors (e.g. Dingle et al., 1983; McMillan et al., 1997; Broad et al., 2006) have reported prevalence of regional erosion during the early Palaeocene, late Oligocene and late Miocene in the Bredasdorp sub-basin which they tentatively attributed to local tectonics and eustatic sea-level changes. Burden (1992), Davies (1997b), Roux (2007) and Sonibare et al. (2014) also provided evidence for the removal of a substantial amount of sediments from the western Bredasdorp Basin in late Cenozoic times. The sedimentary successions formed during post-rift times are characterised by deltaic and shallow to deep marine siliciclastics (siltyshaly sandstones) with minor proportions of igneous intrusions (Dingle et al., 1983; McMillan et al., 1997; Broad et al., 2006; Roux, 2007) (Fig. 2). Long-term drowning of the system during different progradational episodes resulted in the development of coast-parallel continental shelf and slope systems above the Early Aptian (13At) unconformity (Figs. 4b and 4e). The post-13At sequences, in turn, are characterised by the occurrence of upward coarsening sediment stacking patterns indicating progradation of deltaic sedimentary successions (Sonibare et al., 2014).

TE D

2.2. Present-day crust-scale configuration of the basin

AC C

EP

The 3D model of Sonibare et al. (2014) offers insights into the present-day structure and density distribution of the WBB (Fig. 4). The sedimentary infill of the model is structurally separated by basin-wide unconformities (Fig. 3; Brown et al., 1995; Broad et al., 2006) and major faults (Figs. 2, 3 and 4) the geometries of which were derived from a total length of 2100 km of 2D seismic reflection profiles (Figs. 1b and 3). Age and depth constraints of these unconformities and faults were achieved through integration of the seismic data with the analysis of data and well completion reports (WCR) from 47 offshore wells (Fig. 1b). Brown et al. (1995), McMillan et al. (1997) and Broad et al. (2006) described the chrono- and biostratigraphic features of the mapped unconformities (Fig. 3) in detail. Sonibare et al. (2014) provide details about (i) time-to-depth conversion of interpreted horizons and faults using available formation tops (in metres) and checkshot data (i.e. time-to-depth relations from individual wells), and (ii) the deep crustal reconstruction procedure using a combined isostatic and 3D gravity modelling approach. The crust-scale 3D model stratigraphically differentiates three syn-rift and three post-rift sedimentary units of up to ~5-8 km cumulative thickness overlying a tripartite crystalline continental crust with different average densities [ρpre-rift meta-sediments = 2650 kg/m3, ρupper crust = 2700 kg/m3 and ρlower crust = 2800 kg/m3] (Fig. 4a-e). The pre-rift meta-sedimentary sequence (Fig. 4e) locally reaches maximum thicknesses of ~3000 to 5500 m. The cumulative thickness of the metasedimentary and crystalline crust ranges from 20-35 km due to crustal thinning along a W(NW)-E(SE) trending axis (Fig. 4c), whilst the crust-mantle boundary (Moho) is

6

ACCEPTED MANUSCRIPT shallowest in a W-E striking domain (Fig. 4d-e). In line with the syn-rift fault traces, domains of increased syn-rift thicknesses are extended in NW-SE direction. On the contrary, the composite thickness distribution of overlying post-rift sequences reveals a W-E striking domain of maximum thicknesses locally reaching >3000 m (Fig. 4b). 3. Modelling the present-day thermal field

RI PT

3.1. Database and model setup

M AN U

SC

We employed the crust-scale 3D geological model (Sonibare et al., 2014; Fig. 4e) as the basis for numerical simulations of the present-day conductive heat transport and related subsurface temperature. The 3D model extends 135.3 km in the east-west and 110 km in north-south directions (UTM Zone 34S). The horizontal resolution of this model is represented by a regular grid spacing of 275 m in both the northing and easting directions, while its vertical resolution is defined by the number and thicknesses of the resolved geological layers extending down to the Moho (Figs. 3 and 4a-e). We assessed the present-day thermal field of the WBB (Fig. 1a-b) using equation (1) assuming steady-state conditions: ∇ ∙ λ ∇T = −S

(1) o

φ

∗λ

φ

AC C

λ =λ

EP

TE D

where λb is the bulk (whole-rock) thermal conductivity [W/(m*K)], T is the temperature [ C] and S is the radiogenic heat production [W/m3]. Thermal properties were assigned to the respective modelled units considering their dominant lithologies and corresponding literature values (Fig. 4e; Table 1). We utilised well data consisting of gamma ray log, sonic log, density log, cores, lithologic and biostratigraphic information, formation tops and checkshot data as constraints for the lithology of the sedimentary layers. For sedimentary units, the bulk thermal conductivity is calculated from the lithologydependent matrix (solid) thermal conductivity of the sediment grains, λm, the depth-dependent porosity, φz, and a thermal conductivity of water of λw=0.6 W/(m*K) assuming water-filled pores (Fig. 5; Table 1). Therefore, we used the geometric mean equation: (2)

Following the study of Sonibare et al. (2014) and thus Athy’s (1930) exponential law of compaction, we defined for each sedimentary unit a surface porosity at the top of the model (φ0) and a compaction coefficient (c) to calculate the depth-dependent porosity (φz): φz = φ0*e-cz

(3)

The matrix thermal conductivities presented in Table 1 represent standard values for crustal rocks given in Clauser and Huenges (1995), Midttømme and Roaldset (1999) and Allen and Allen (2005). Based on the modelled lithology- and depth-dependent porosities, the 3D variation in bulk thermal conductivity was calculated (Equation 2). The radiogenic heat production of sediments and deep crustal rocks were derived after Wollenberg and Smith (1987) and Vilà et al. (2010). Apart from the assigned lithology-dependent thermal properties (Table 1), solving equation (1) depends on the definition of thermal boundary conditions for the model. For the

7

ACCEPTED MANUSCRIPT

q = − λ ∗ dT/dz

TE D

M AN U

SC

RI PT

upper boundary, a constant temperature of 18oC was assigned following the reported average seafloor temperature across the southern margin of South Africa (Goutorbe et al., 2008). Since the thermal budget within a given sedimentary basin also depends on the amount of heat entering the system from the mantle and the crystalline crust (e.g. Allen and Allen, 2005), the model incorporates the whole crust and thus its lower boundary condition was set at the Moho. To validate the model against measured temperatures, we tested a range of values for heat flux at the Moho (i.e. q = 30, 35, 38 and 40 mW/m2). The tested lower boundary conditions are all consistent with the estimated range of mantle heat flow for the Archean and Proterozoic belts of the Southern African continental interior (Goutorbe et al., 2008). Finally, we assumed the lateral model boundaries to be closed for heat flow. To validate predicted temperature distributions, we used measured present-day borehole temperatures. Out of the provided offshore wells, measured borehole temperatures were only available for 24 wells as two types of temperature records: static bottom-hole temperatures (SBHTs) and fluid temperatures acquired during drillstem tests (DSTs). While SBHTs are available for all these wells, DSTs are only available for some of them. Being direct temperature measurements on formation fluids during on-site reservoir flow testing, DSTs are generally considered as the best representation of the formation temperatures. Since SBHTs do not represent true formation temperature because of the cooling effect of the drilling mud, they are usually corrected to reflect static formation temperature (e.g. Allen and Allen, 2005). All SBHTs used in this study were already corrected following the “Horner”-type plot as described by Dowdle and Cobb (1975). The available borehole temperature dataset consisted of 41 corrected SBHTs and 16 DSTs. Based on the modelled and validated 3D conductive thermal field, we calculated the crustscale heat flow distribution (Fig. 6) by solving Fourier’s equation assuming only vertical heat flow: (4)

AC C

EP

where q is the heat flux [mW/m2], λ is the bulk thermal conductivity [W/(m*K)], and o dT/dz is the geothermal gradient [ C/km]. Thus, this heat flux calculation utilised the established geothermal gradients from the calculated 3D conductive temperature distribution. We performed the calculations of the present-day temperature and heat flow distributions with the software package GMS (GeoModelling System, an in-house geological modelling software developed at the Helmholtz Centre Potsdam, GFZ German Research Centre for Geosciences and described in detail by Bayer et al., 1997) that implements the Finite Element Method (FEM). 3.2. Results

Since modelled thermal properties depend on the predominant lithologies (Fig. 2; Table 1), and the upper boundary condition (seafloor temperature) is data-based (Goutorbe et al., 2008), the lower boundary condition, i.e. the heat flux at the Moho, therefore constitutes the only free parameter that was to be varied to match the measured borehole temperatures. As such, we find that a conductive thermal field (Fig. 5-6) modelled with a heat flux of 38 mW/m2 at the Moho reliably fits the temperature data as can be seen from the depth plots of temperatures and frequency distributions of calculated temperature misfits (e.g. cc=0.90; Fig. 7-8). The histogram of the general trend of temperature deviation (with ∆T representing modelled

8

ACCEPTED MANUSCRIPT o

AC C

EP

TE D

M AN U

SC

RI PT

minus measured temperatures) and particularly the median of model misfits of -0.57 C (see Figure 8) depict the overall consistency between the final thermal model and measured temperatures. Furthermore, a very good agreement between DSTs and corrected SBHTs for wells having both temperature datasets (Fig. 8b) greatly increases the level of confidence in the available measured temperature datasets. Cross-checking our model with all available temperature data results in 68% of the modelled temperatures to deviate by only ±5 °C from the measured data (Fig. 8a). This amount of best fit is reduced to 66% (fitting by ±5 °C) if the model is compared only to SBHTs. Considering only wells with both DSTs and SBHTs gives o a better calibration with 94% of the modelled temperatures deviating by only ±5 C from the measured data (Fig. 7b). Thus, the remaining temperature misfits may partly be ascribed to limitations in SBHT corrections. Testing different values for the Moho heat flux depicts how sensitive the modelled temperature distribution is to this parameter. Initially setting the lower boundary condition at a constant heat flux of q=30 mW/m2 at the Moho and employing the lithology-dependent thermal properties as given in Table 1, resulted in a conductive 3D thermal model being significantly too cold. This thereby led to 87.5% of the DST-data and 60% of the SBHT-data o o being underestimated by -10 to -20 C and -10 to -25 C, respectively. Similarly, for q=40 mW/m2 at the Moho, the model is overall too warm, particularly for the SBHTs. The achievement of a physically consistent thermal field model, exhibiting a reasonable fit with the measured temperatures, allows us to analyse the relative importance of the main temperature-controlling factors in the WBB. Figures 5 and 6a-c show calculated properties and temperature distributions for selected depth levels and at the Moho. Generally, the temperatures vary strongly both laterally and vertically and with different wavelengths. As depicted at all depth levels (Figs. 5 and 6a-c), long-wavelength anomalies characterise the distribution of modelled temperatures and properties between the basin margins and the sediment-filled basin centre. Lowest temperatures spatially correlate with the basin margins, where the basement is relatively shallow, i.e. covered only by a thin sedimentary layer (of low thermal conductivity) underlain by thick crystalline crust (of high thermal conductivity; Fig. 4a-e; Table 1). Apart from the 2 km depth level, high temperatures are confined around the NW-SE trending basin axis correlating spatially with the largest thicknesses of low-conductive sediments (Figs. 4 and 5ad). o For the Moho, lowest temperatures (~550-670 C) are predicted around the overall W-E shallowing trend of the Moho (Figs. 4d and 6c). The regions of high temperatures (reaching o up to 690-770 C) are generally restricted to the north and south of the Moho high, i.e. where the crust is thickest and the Moho deepest (Fig. 4c-d). At the 2 km depth level (Fig. 5a-d), the distribution of short-wavelength temperature variations related to differences in the properties (i.e. lithology, porosity, thermal conductivity; Table 1) of differing model units (upper crust, pre-rift meta-sediments and basin-fill sediments; Fig. 4) are clearly revealed. Most importantly, we find a clear correlation between high temperatures (Fig. 5d) and low bulk thermal conductivities (Fig. 5c). For instance, the disparities in bulk conductivities between the two shale-dominated sequences (i.e. Unit 5 and Unit 4) as well as between the Unit 3 and Unit 2 sequences are governed by variations in porosities and matrix conductivities (Fig. 4a-d). Although the Unit 6 and Unit 2

9

ACCEPTED MANUSCRIPT

TE D

M AN U

SC

RI PT

sequences show the same matrix conductivity ( max=2.7 Wm-1K-1; Table 1), the Unit 6 being more porous characteristically shows lower bulk conductivity ( bulk=~1.7 Wm-1K-1) at this depth than the younger Unit 2 (Fig. 5b-c). Figure 6d-f shows the calculated heat flux distribution for selected depth levels, namely for the seafloor (Fig. 6f), for the top of the crystalline basement (Fig. 6e), and for a constant depth of 10 km (Fig. 6d). From the 38 mW/m2 fixed heat flux at the Moho, heat flow increases upwards to the crystalline basement (with heat fluxes between ~55-74 mW/m2) and gradually to the seafloor (with heat fluxes ranging from 63 to ~80 mW/m2). Common to the crystalline basement and the seafloor is the distribution of low and high heat fluxes corresponding, respectively, to the sediment-filled basin centre and the basin margins. This spatial trend of heat fluxes conforms to the established lateral and vertical variations of geothermal gradients (Figs. 5d, 5f, 6a-c and 6g-h) related to high radiogenic heat production in the crystalline crust and low thermal conductivity (thermal blanketing) of the sediments. Accordingly, high geothermal gradients characterise the sediment-filled basin axis while low geotherms mark the shallow basement at the basin margins (Fig. 6g-h). At the seafloor and at 2 km below the seafloor, geothermal gradients along the NW-SEo o trending basin axis range approximately between 35 Ckm-1 and 45 Ckm-1 (Fig. 6g-h). Going downwards into the deep crust, we observe a gradual decrease in geothermal gradient as the system transits from the sedimentary cover into the underlying sub-sedimentary crystalline crust (see Fig. 6g). In line with the observed geotherms, high heat fluxes are restricted to the basin margins where the crystalline basement with high thermal conductivities is at shallow depths. In contrast, low heat fluxes are confined to the NW-SE trending axis of the lowthermally-conductive basin-fill sequences. 4. Modelling the thermal evolution

4.1. 3D forward modelling: model setup and thermal calibration

AC C

EP

To investigate the Mesozoic-Cenozoic thermal evolution incorporating the effects of steadystate and transient heat transport over geological times, we applied a three-dimensional forward modelling approach using the software PetroMod v. 2014.1 (© by Schlumberger). The 3D forward modelling approach simulates the burial and thermal history based on the principles of basin modelling techniques described by, e.g. Poelchau et al. (1997) and in detail by Hantschel and Kauerauf (2009). These principles define certain sedimentation intervals (including non-deposition and erosion) of known ages (Fig. 9) in terms of interrelated processes such as (i) porosity loss (due to burial and compaction), (ii) isostatic re-adjustments (due to changing loads), and (iii) conductive heat transport (in relation to changing thermomechanical properties and boundary conditions). The modelling procedure first requires setting up an initial (starting) model involving the definition of the following parameters: 1. 2. 3. 4. 5.

present-day 3D geometry of the basin, lithofacies and source rocks definition tectono-stratigraphic events (i.e. deposition, non-deposition and erosion), paleo-geometries of water depth and eroded material, boundary conditions (upper and lower thermal boundaries).

10

ACCEPTED MANUSCRIPT

RI PT

By performing a series of simulation runs and modifying certain parameters (see below), model predictions can iteratively be adjusted to observed temperature and vitrinite reflectance (VR) data (see Fig. 10). Vitrinite reflectance is the most widely used paleo-geothermal indicator, which measures characteristic changes in organic matter reflection properties under evolving burial and thermal conditions of a sedimentary basin through time (e.g. Poelchau et al., 1997; Allen and Allen, 2005; Hantschel and Kauerauf, 2009). We calculated the increase of vitrinite reflectance as a function of time and temperature using the EASY% Ro chemical kinetic equations of Sweeney and Burnham (1990). In our case, vitrinite reflectance data collected from rock cuttings, sidewall cores and cores were available from 17 wells (Figs. 1b and 10), totalling 836 sample points (source: internal reports).

M AN U

SC

4.1.1. Present-day geometry The present-day geometry was defined based on the crust-scale 3D structural model of the WBB described in sub-sections 2.2 and 3.1. We further refined the vertical resolution of the stratigraphic units that were previously modelled as composite units. This led to the insertion of an Albian (14At) unconformity, and Santonian (K) and Campanian (17At) unconformities (Brown et al., 1995; McMillan et al., 1997; Figs. 3, 4e and 9). Below the ten unconformitybounded Mesozoic-Cenozoic sedimentary units, a lower boundary for the 3D model was imposed at a constant depth of 10 km, which is located below the pre-rift metasediments and cuts through the upper crustal unit (Fig. 9).

AC C

EP

TE D

4.1.2. Lithofacies and organic source rocks definition In addition to refining the vertical resolution of the stratigraphy (Fig. 9), we performed well correlation using gamma ray (GR) logs to lithologically separate the stratigraphic units according to vertical and lateral variations in lithofacies. Based on this well correlation and published litho-stratigraphic variations in the Bredasdorp Basin (McMillan et al., 1997; Davies, 1997a; Roux, 2007; PASA, 2012), we developed lithofacies maps for each stratigraphic interval (see Figure 9). The lithofacies definition also utilised the corresponding PetroMod standard values for the thermo-mechanical parameters of thermal conductivity, radiogenic heat production, specific heat capacity, porosity, and density (Table 2) guided by the established lithologies and thermal properties stated in Table 1. Six source rock intervals consisting of shale to sandy shale (sh60ss40; Fig. 9) (i.e. Valanginian, Mid Hauterivian, Late Hauterivian, Barremian, Aptian, and Turonian; see Figures 2 and 9) were delineated from available pyrolysis and vitrinite reflectance data. All the source rocks, except from the Aptian source, show evidence of mixed Type II and Type III organic matters. The Aptian source is Type II of shallow to deep marine deposits with average Total Organic Carbon and Hydrogen Index of 4 wt.% and 400, respectively. 4.1.3 Age definition of stratigraphic events We defined stratigraphic events (Figs. 2 and 9) based on the geological ages of the modelled units and related timing of deposition, non-deposition (hiatus) and erosion as indicated by WCR and published data (Brown et al., 1995; Broad et al., 2006; PASA, 2012). These ages also correlate with both the local and regional geological and tectonic history of the study area (e.g. Dingle et al., 1983; McMillan et al., 1997; Thomson, 1998; Johnson et al., 2006; Fig. 2).

11

ACCEPTED MANUSCRIPT

M AN U

SC

RI PT

4.1.4. Paleo-geometries Paleo-geometries implemented as maps in our model were of two types: paleo-water depths and erosion magnitudes. We gained approximate paleo-water depths from WCR biostratigraphic information, facies relationships from cored wells and knowledge of the inferred depositional environments for the Bredasdorp sedimentary successions as provided by McMillan et al. (1997), Broad et al. (2006) and PASA (2012). The assigned paleo-water depths remain relatively shallow over the past and up to present-day, ranging between ~40 and 300 m with maximum water-depths occurring during Hauterivian-Aptian times. Because of limitations in the availability of investigated data (i.e. seismo-stratigraphic information, wells and WCR biostratigraphic well-correlation panels), it was possible to derive direct estimates on erosion amounts only for the Hauterivian event (134-133 Ma) and the late Miocene erosion event (16-11 Ma). Hauterivian amounts of erosion were mainly mapped on seismic profiles as an erosional truncation where the Valanginian (1At) unconformity is truncated by the Mid-Hauterivian (5At) unconformity. The related eroded thicknesses are limited to the NW proximal parts of the basin ranging from 40-240 m (Fig. 11a). For the late Miocene event, we find indications for ~1000 m of eroded material in the NW proximal parts of the basin (see Figs. 9 and 11b) from where they decrease towards the east. For the starting model (later to be stepwise modified, see below), we used the results of Davies (1997a) to additionally develop preliminary erosion distribution maps for the erosion phases at 84-75 Ma and 43-38 Ma (Fig. 11c-d).

AC C

EP

TE D

4.1.5. Boundary conditions The upper boundary conditions for the forward modelling process are controlled by varying paleo-climate and temperature at the sediment-water interface (SWIT; Francu et al., 2002; Hantschel and Kauerauf, 2009). Thus, at the upper boundary we assigned the surface temperature through time following the approach of Wygrala (1989), incorporating paleolatitude variations (Hantschel and Kauerauf, 2009) for our study locality. This surface o temperature profile by Wygrala (1989) gives a present-day SWIT as warm as 18 C, which corresponds well with the seafloor temperature used for modelling the present-day thermal field (see above). The lower boundary condition is the inflow of heat at the base of the sediments for different time intervals (e.g. Hantschel and Kauerauf, 2009). In our case, these basal heat flow values correspond to maps of laterally varying heat flux (ranging between 44 and 62 mW/m2; Fig. 6d) at present-day 10 km depth. For the initial simulation, the basal heat flux distribution was directly extracted from the previously established present-day crust-scale 3D thermal model (Figs. 4e, 5 and 6). This definition holds the advantage that the thermal-evolutionmodelling procedure directly considers what we already know about present-day geometry and thermal state such as the Moho depth and heat flux and deep crustal heat generation. Since the crust-scale conductive thermal model is consistent with observed present-day temperatures (Fig. 7-8), this holds also for the extracted 10 km-depth heat flux distribution (Fig. 6d) thereby representing an ideal constraint for the final, present-day stage of the forward modelling procedure. Clearly, the amount of heat entering the shallow parts of the crust has not remained constant through syn- and post-rift times, particularly considering the complex tectonic history of the basin. Such information, however, does not exist for the WBB so far. Therefore,

12

ACCEPTED MANUSCRIPT for the initial simulation we assumed a constant-through-time basal heat flux distribution equal to the one calculated for the present-day (Fig. 6d). Based on the related (in-) consistencies with vitrinite reflectance data, we improve model predictions through stepwise modification of the basal heat flux maps until reaching an acceptable correlation between modelled and measured values (see below).

AC C

EP

TE D

M AN U

SC

RI PT

4.1.6. Thermal calibration Figure 12 summarises the complete model calibration workflow showing that we firstly calibrated the model against present-day measured temperatures (i.e. DST and SBHT; Fig. 78) and then reconstructed the paleo-temperature distributions using VR data (Fig. 10). As anticipated, we achieved a straightforward calibration of the modelled present-day temperature using the predicted heat flow at 10 km depth level from the best-fit crust-scale conductive model (Figs. 5 and 6) as the lower boundary condition of the initial model (i.e. 1st model run, Fig. 12). Thus, this calibration step confirms the consistency of the paleo-thermal model with the previously established present-day geothermal field. However, at this stage, the measured and simulated VR trends were largely not in good agreement. Based on the detailed analysis and interpretation of the available VR data, we find that the modelled area reveals three sub-domains: proximal, central and distal domains as shown in Figure 10. In these domains, VR data range from 0.35-1.27 VRo% (proximal) to 0.47-2.16 VRo% (central) and 0.35-1.39 VRo% (distal). These VR profiles are generally sub-linear (on a logarithmic plot), with the profile of the proximal domain depicting a bulge towards higher VR at intermediate depths (Fig. 10a). Moreover, in the upper 200-2400 m sequence of the proximal and central domains, two separate VR profiles are observed (Fig. 10a-b) which are indicative of paleo-thermal events and erosion (e.g. Allen and Allen, 2005; Hantschel and Kauerauf, 2009). In the proximal part, the profile with higher VR values runs sub-parallel to the lower VR profile (Fig. 10a), whereas in the central domain, the higher VR profile gives an obvious bulge (Fig. 10b). As VR trends depend on the maximum paleo-temperature, we analysed the fit between modelled and measured VR trends at single wells (Fig. 13) for different model scenarios. Thus, adjusting modelled VR trends to observations requires changing the parameterisation of paleo-temperature models. We restricted any subsequent steps of model adjustments to modifications of two free parameters: erosion magnitude and basal heat flow, i.e. those two parameters that are mostly unknown and very vital for any reconstruction of past geothermal gradients. In a first VR calibration “loop” (Fig. 12), we have adjusted paleo-erosion maps. We did not, however, change the amount of Hauterivian (134-133 Ma) erosion (Figs. 9 and 11a) as it is well constrained directly from seismic interpretation. Instead, erosion magnitudes for the late Cretaceous (84-75 Ma), early Tertiary (43-38 Ma) and late Miocene (16-11 Ma) times (Fig. 9) were adjusted. Thereby, local (well-based) adjustments followed the described proximal to distal trends of VR profiles, particularly in the upper 2.5 km of the sequence (Fig. 10a-b). The assessment of late Miocene erosion (Fig. 11d) was guided by available information from seismic interpretations, well data and predicted cumulative late Tertiary erosion (Sonibare et al., 2014). We also inferred erosion magnitudes using published studies on erosion and sedimentation rates in the Outeniqua Basin (e.g. McMillan et al., 1997; Davies, 1997a; Broad et al., 2006). By implementing the aforementioned erosion adjustments, a reasonable fit of measured and simulated vitrinite trends was achieved for the distal and

13

ACCEPTED MANUSCRIPT

M AN U

SC

RI PT

most of the central domains of the model. In the proximal domain, a significant misfit still prevailed, as reflected particularly by the VR profiles of wells D-A1 and D-B1 (Figs. 1b and 13). In general, such misfits might be either related to underestimation of erosion amounts or to the paleo-heat flux at the predefined base of the model having been higher during the past than at present-day. To correct the remaining misfits by increasing erosion magnitudes for the respective erosive events would require several kilometres of eroded sediment thicknesses, which we believe is not geologically possible as inferred from previous studies (e.g. Roux, 2007) as well as seismic and well data observations. For this reason, a second loop for thermal calibration implemented basal heat flow adjustment in order to improve the fit between modelled and measured VR data. Modifying the heat flux distribution through time is even more realistic as the study area was subjected to rifting and complex temperature perturbations due to recurrent tectonic activity and hotspot transient events (e.g. Duncan, 1981; Davies, 1997a; Broad et al., 2006). Hence, we implemented three periods of possible thermal perturbation: (1) Upper Jurassic-Late Cretaceous lithospheric thinning and faulting during syn-rift sedimentation, assuming a rift-type heat flow peak (135-125 Ma; e.g. McKenzie, 1978), (2) Late Cretaceous hotspot transient event (80-70 Ma; Duncan, 1981), and (3) Cenozoic margin uplift event (12-8 Ma; Nyblade and Robinson, 1994; Hartnady and Partridge, 1995). We stepwise increased the basal heat flow at the 10 km depth level for the respective peak periods following the proximal to distal trend of the measured VR profiles (Fig. 10a-d) until the modelled paleo-3D thermal field satisfactorily matched all available VR data.

TE D

4.2. Results: Thermal development through time

AC C

EP

The VR calibration procedure (Fig. 12) provided thickness distributions of eroded material for the different erosional phases (Figs. 9 and 11). Of all the spatially reconstructed amounts of erosion, the late Miocene erosion has the largest magnitude varying from 980-1240 m in the proximal NW to 760-960 m and 600-720 m, respectively, in the central and distal SE domains of the WBB (Fig. 11d). The late Cretaceous (Campanian) lateral variations of erosion are in between 400-600 m (proximal), 280-350 m (central), 220-260 m (distal), and the early Tertiary (Paleogene) variations of erosion amount to 550-620 m (proximal), 460-500 m (central) and 420-440 m (distal; Fig. 11b-c). Implementing the adjusted lower boundary heat flux for the three peak events (syn-rift phase, post-rift hotspot event and post-rift margin uplift; Figs. 14 and 15) resulted in a satisfying calibration of the measured and simulated present-day temperatures and vitrinite reflectance trends as presented in Figures 13. Larger deviations are only shown by well EAH1 at depths of 2500-3500 m (Fig. 13). However, we considered this deviation to be within acceptable limits as the wells located nearby show good calibration results (Figs. 2 and 13) and any attempt to reduce the modelled VR values for this depth interval caused the modelled present-day temperatures to underestimate measured temperatures. Of all the VR data delineated domains (Fig. 10d), the proximal domain reveals highest basal heat fluxes for the three thermal events (Figs. 14 and 15). Near the roughly NW striking half-graben faults and basement highs of the proximal domain (Figs. 1 and 4), heat fluxes at 10 km depth are in the range of ~60-80 mW/m2 for the syn-rift event, ~75-90 mW/m2 for the late Cretaceous hotspot event and ~55-73 mW/m2 for the late Miocene thermal peak (see Fig.

14

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

15 for an exemplary well D-A1). For comparison, in both the central and distal domains, these 10- km-depth heat fluxes are in the range of ~40-70 mW/m2, ~40-63 mW/m2 and ~38-65 mW/m2 for the syn-rift, late Cretaceous hotspot and late Miocene thermal events, respectively (see Fig. 15). The burial plots for exemplarily chosen wells in the proximal (Fig. 16a-b), central (Fig. 16c-d) and distal domains (Fig. 16e-f) reveal the temporal trends of thermal evolution. The evolving depths of stratigraphic and lithofacies boundaries are shown superimposed by the modelled paleo-temperature distribution and dictated paleo-water depths. Thus, subsidence and uplift history as well as sedimentation rates can directly be related to the thermal evolution. The rifting phase (~160-130 Ma; Broad et al., 2006; PASA, 2012) is characterised by upward displacements of isotherms at the three selected wells (Fig. 15-16). The magnitude of this upward bulge of isotherms increases with increasing burial depth. This rift-related temperature increase also corresponds to the period of rapid burial rates at ~135-136 Ma. During the rifting phase, the most deeply buried sediments within half-grabens (Figs. 3 and o 4a), such as in wells D-A1, E-BD2 and E-AA1, reach temperatures of ~97-110 C, ~136-145 o o C and ~110-128 C, respectively (Fig. 15a-17a). In the proximal domain, temperature isolines occur at shallower depths (Fig. 16) than the same temperatures in the central and distal domains (Fig. 16-17), which demonstrates that temperature gradients generally decrease from the W(NW) to the E(SE). The base-of-sediment heat flow map at ~131.5 Ma depicts a similar W(NW) - E(SE) decreasing trend (see Fig. 14a). The post-rift phase (after ~125 Ma) depicts long-term thermal cooling interrupted by three short-lived periods of heat perturbation (Fig. 14-17) correlating temporally with erosion and thermal events (Figs. 15 and 16). Generally, the vertical distances between isotherms are larger during post-rift than the syn-rift times (Figs. 16a, 16c and 16e). Considering the amount of heat flux increment, the early Paleogene erosion is the least significant, raising the longterm post-rift mean heat flow by only ~4-6%. However, in the distal domain (Figs. 15 and 16e-f), the Paleogene thermal anomaly is hardly less significant in magnitude than the late Cretaceous and Neogene thermal anomalies. The predicted base-of-sediment heat fluxes at 75 Ma range from ~60 to 97 mW/m2 and depict a W(NW)-E(SE) decreasing trend (Fig. 14b) conforming to the measured VR trends along the three delineated sub-domains (Fig. 10). The heat fluxes for the late Cretaceous thermal anomaly range from ~90-94 mW/m2 (proximal domain, e.g. well D-A1) to ~62-70 mW/m2 (central domain, e.g. well E-BD2) and ~58-63 mW/m2 (distal domain, e.g. well E-AA1). Whereas for the late Miocene thermal peak, predicted ranges of heat flux for the respective domains are: ~88-94 mW/m2 (proximal), ~6578 mW/m2 (central) and ~64-73 mW/m2 (distal). 5. Discussion 5.1. Modelling approach We evaluated the thermal evolution at the southern South African continental margin using a crustal model that incorporates the regional geotectonic framework and generally shows consistency with measured temperature and vitrinite reflectance data (Figs. 5-8 and 13). The ability to reliably predict the present-day thermal field also for sub-areas away from borehole temperatures allows the definition of well-constrained lower boundary conditions for the 3D

15

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

numerical simulations of paleo-thermal fields (Figs. 9 and 14-16). In this way, our findings about present-day heat-related processes such as Moho heat flux and deep crustal heat production (see Table 1) contributed decisively to a better reconstruction of thermal field development in the WBB. Through a stepwise validation workflow (Figs. 7-8 and 12-13), it becomes relatively easy to handle complex interactions between modelled processes and physical parameters. This holds for both reproducing the present-day thermal field by varying the Moho heat flux, and for modelling the paleo-thermal development by adjusting erosion magnitudes and subsedimentary heat flow. By doing so, we gain insights not only into the associated uncertainties of the underlying geological assumptions but also regarding the dominating controls on the present and past thermal fields. For instance, knowing the sensitivity of the present-day thermal model to Moho heat flow enables us to delimit the associated uncertainties regarding the lower boundary condition. This is important as for the Southern African continental interior, mantle heat flow estimates have been noted to vary strongly across different geological provinces (e.g. Goutorbe et al., 2008). As reflected by the best-fit thermal model, the results of performing a sensitivity analysis on a range of fixed mantle heat fluxes (i.e. 30, 35, 38 and 40 mW/m2) suggests a mean present-day Moho heat flow of about 38 mW/m2 (Figs. 4d-e and 6). In general, the remaining model misfits (regarding temperature and VR data) largely cluster within a narrow domain in the centre of the model (Figs. 7-8, 10 and 13). These misfits may be attributable to uncertainties, which can be linked to: (i) bad corrections of SBHTs especially for wells without corresponding DSTs (see above and Fig. 7-8); (ii) not considering the effects of convective heat transport either resulting from mantle dynamics (Goutorbe et al., 2008) or fluid flow and hydrothermal activity (e.g. Davies, 1997a); (iii) assumptions regarding the population of the model units with lithology-dependent thermo-mechanical properties and (iv) not considering the radiogenic heat production of the crystalline crust as age-dependent (Vilà et al., 2010). However, the overall consistency of the final thermal models, as indicated by cc=0.90 and ∆T median=-0.57 for the present-day thermal field (Fig. 7-8) and as shown in Figure 13 for the modelled paleo-thermal field, indicates that our approach provides a good approximation of the overall past and present heat budget (i.e. heat entering and leaving the basin system) within the WBB. Consequently, thermal diffusion is identified as the dominant basin-wide heat transport mechanism in the study area, which is in general agreement with findings in other comparable sedimentary basin settings (i.e. Noack et al., 2010; Scheck-Wenderoth and Maystrenko, 2013; Maystrenko et al., 2013). 5.2. Present-day thermal field The spatial distributions of heat flow and geothermal gradients across the modelled area (Fig. 6d-h) suggests that the interplay of shallow and deep temperature-controlling factors, i.e. thermal conductivity and radiogenic heat production constitutes the primary driver for the present-day thermal budget in the WBB. This inference becomes obvious at the 5 km, 10 km and 15 km depth levels where high and low temperatures correlate to the main basin-fill axis and the shallow basement at the margins (Figs. 4, 5e-f and 6a-b), respectively. The distribution of highest temperatures around the basin axis indicates a phenomenon whereby the crystalline crust is thick enough to produce considerable amounts of heat and the sediments (being generally of lower thermal conductivities than the deep crust) are thick

16

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

enough to effectively invoke thermal blanketing (e.g. Wangen, 1995; Allen and Allen, 2005, Hantschel and Kauerauf, 2009; Vilà et al., 2010). Our crust-scale model satisfies the structural configuration for considerable heat storage within the overlying sediments as the underlying sub-sedimentary crystalline rocks are dominated by granitic-granodioritic rocks (see Fig. 4e, Table 1 and Table 3), which are known to have higher radiogenic-heat-producing capacity than basic rocks (e.g. Bjørlykke, 2010; Vilà et al., 2010). Within the shallow sedimentary layers, this blanketing effect is localised as short-wavelength thermal anomalies mimicking the lateral variations in lithofacies and depth-dependent bulk thermal conductivities (see Fig. 5a-d). The established spatial variations of shallow and deep geotherms of the WBB (Fig. 6) therefore, corroborates previous studies (Bayer et al., 1997; Ondrak et al., 1998; Allen and Allen, 2005; Hantschel and Kauerauf, 2009; Scheck-Wenderoth and Maystrenko, 2013). The predicted surface heat flow largely agrees with the previous estimates for the southern offshore of South Africa by Goutorbe et al. (2008) based on borehole temperatures and thermal conductivities derived from geophysical well logs. Some local highs [ =~71-73 mW/m2], particularly in the northwestern domain are closely related to local structural highs (i.e. shallow basement) where the crust is thick and overlain by relatively thin sediments (Fig. 6f). The depth distribution of the Moho (Fig. 4d) and the thickness of the crust (Fig. 4c) provide evidence for the effect of crustal thinning, a process that to some extent certainly involved lithospheric thinning and is dated to commence in the southern South Atlantic Margin around ~136 Ma (Martin and Hartnady, 1986; Thomson, 1998). The present-day thermal field is implicitly affected by this event as the geometry of the crystalline crust was preserved since then. Consequently, the Moho temperatures decrease from the margins to the central parts of the basin (Figs. 4d and 6c). Furthermore, the thinning trend of the crust along the (N)W-(S)E basin axis directly suggests a reduced amount of radiogenic heat and thus the crust adds to the basin’s heat budget less in the centre than at the margins. Goutorbe et al. (2008) causally relate an increase in surface heat flow across the southern South African margin (from 17-30 mW/m² within the continental interiors to ~60 mW/m² across the margin (and to even higher values in the oceanic domain) to variations in mantle heat flow. However, our model suggests that the heat input from the mantle is relatively uniform across the (allowedly smaller-scale) WBB (38 mW/m²). Furthermore, the Lithosphere-AsthenosphereBoundary (LAB) has been found to be relatively flat at depths that vary only between 134 and 145 km across the WBB (Fishwick, 2010). Assuming the LAB to represent a thermal boundary (i.e. a temperature of ~1300°C at which mantle material starts partial melting), the small variations in depth of the LAB confirm small variations of heat flow at the Moho. Moreover, the relatively large depths of the LAB found even beneath thinned continental crust argue for a thermally re-equilibrated system at present-day. 5.2.1. Comparison with the adjacent southwest Atlantic Margin The importance of the crustal configuration for the thermal field becomes evident by a comparison with the Orange Basin located in the adjacent southwest Atlantic Margin. In the WBB, predicted surface heat flow values vary from ~63 to 70 mW/m2, while in the Orange Basin surface heat fluxes are estimated to be between 50 and 60 mW/m2 (Goutorbe et al. 2008; Hirsch et al. 2010). The southwestern margin represents a magmatic Atlantic-type passive margin (e.g. Gladczenko et al., 1998; Bauer et al., 2000; Serrane and Anka, 2005; Broad et al., 2006; Hirsch et al., 2007) while the southern margin is a non-magmatic

17

ACCEPTED MANUSCRIPT

TE D

M AN U

SC

RI PT

transform-passive margin (Scrutton and Dingle, 1976; Parsiegla et al., 2007, 2009; Sonibare et al., 2014). Here, the related key question is: why is the surface heat flow lower in the magmatic margin area than in the non-magmatic WBB? The isothermal (1300°C) LAB depth is deeper in the WBB (see above) than in the Orange Basin (with a depth profile ranging between 110 and 120 km) (Fishwick, 2010). Ideally, the shallower LAB for the Orange Basin would mean a higher Moho heat flow than beneath the WBB. The modelled WBB, however, shows higher heat flow than the Orange Basin, thereby suggesting the significance of the crustal contribution which is larger from the thicker crust beneath the WBB. Since the sediment thicknesses and lithologies (including thermomechanical properties) are quite similar for the compared sedimentary basins, their radiogenic heat contribution to the surface heat flow in the South Atlantic Margin is found to be relatively similar, i.e. mostly less than ~5 mW/m2 (Goutorbe et al., 2008; Hirsch et al., 2010; Table 3). Besides, radiogenic heat production of the upper crust and lower crust (i.e. 1.45 µW/m3 and 0.8 µW/m3 respectively) in the Orange Basin (Maystrenko et al., 2013) are similar on average to values in the WBB (upper crust=2.0 µW/m3, lower crust=0.5 µW/m3; see Table 3). However, cumulative crystalline crustal thicknesses are different and thus decisive for the two adjacent basins. In the Orange Basin, mean crystalline crustal thickness ranges from 1015 km (Maystrenko et al., 2013), while in the WBB, crystalline crustal thickness is much thicker having 20-30 km thickness for ~70% of the modelled area. Apart from the effect of absolute heat generation due to differing crustal thicknesses, internal lithological controls may play additional role. If the postulated magmatic underplated bodies of the southwest Atlantic Margin (e.g. Bauer et al., 2000; Serrane and Anka, 2005; Hirsch et al., 2007) are dominated by mafic rocks, the overall crystalline crustal heat production would be additionally reduced. In the case of the modelled WBB, the deep crust, as confirmed by 3D gravity modelling and deep seismic studies is non-magmatic and compositionally dominated by granitic-granodioritic rocks (see sub-chapter 2.1), which are more efficient producers of radiogenic heat than mafic rocks.

EP

5.3. Mesozoic to Cenozoic thermal development

AC C

The VR profiles for the proximal, central and distal domains along the NW-SE basin axis (Fig. 10) directly point to possible vertical and lateral disturbance of paleo-geothermal gradients. This paleo-thermal disequilibrium becomes most evident in the proximal and central domains, especially in the upper 200-2400 m sequence, as indicated by two separate VR sublinear patterns (Fig. 10). These interpreted NW-SE trends of measured VR profiles corroborate the results of the thermal calibration workflow (see Fig. 12), thereby suggesting that the paleo-geothermal gradients in the modelled WBB have never been constant neither spatially nor through geological time. Generally (e.g. Allen and Allen, 2005), the pattern of measured vitrinite reflectance trends may reflect cumulative effects of consecutive phases of subsidence, uplift (and thus erosion) and thermal perturbation within the basin. Furthermore, for a uniformly stretched basin (e.g. McKenzie, 1978) the post-rift basal heat flow is expected to decrease gradually as the basin cools passively while achieving a state of thermal equilibrium. Contrastingly, our results suggest that this is not the case for the modelled WBB. 5.3.1. Role of syn-rift crustal stretching and rapid subsidence

18

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

As depicted by the burial history diagrams of the WBB (Fig. 16), sediments were rapidly buried to higher temperatures during the syn-rift phase (at ~150-130 Ma; Broad et al., 2006; PASA, 2012). Rapid and especially short-lived syn-rift subsidence has widely been linked to the formation of transform margins and strike-slip related subsidence, especially for the pullapart basin style (e.g. Cochran, 1983; Pitman and Andrews, 1985; Allen and Allen, 2005; Xie and Heller, 2009). Consistent with well, seismic and gravity data, Sonibare et al. (2014) propose a transtensional setting for the Bredasdorp sub-basin, with the corresponding component of strike-slip being related to movements along the AFFZ, as the observed crustal thinning is difficult to explain by a single-phased, pure, and margin-perpendicular extension. Moreover, we find that vertical distances between temperature isolines are smaller and generally at shallower depths for the syn-rift phase than the post-rift phase, thereby indicating that paleo-geothermal gradients are higher during syn-rift times due to rapid subsidence and sedimentation rates provoked by crustal thinning (White and McKenzie, 1989). If we assume that the contribution of the deeper crust to the basal heat flow of the model (Fig. 9) has not changed since break-up (i.e. constant thickness, neglecting the effect of decreasing radiogenic productivity), the contribution of mantle heat flow to the syn-rift thermal anomaly becomes evident by comparing the basal heat flow at present-day (Fig. 6f) with the modelled basal heat flow at 130 Ma (see Figs. 14a and 15). Given the difference in basal heat flow at present-day (Fig. 6f) and at syn-rift times (Fig. 14a), the mantle heat flow contribution during the syn-rift phase differs from the present-day (background) mantle heat flow (~38 mW/m2) by about 20-30 mW/m2 over about 75% of the modelled area. The corresponding thermal event may therefore, be linked to the Karoo plume (postulated to be centred in the region of Mozambique) which is associated with volcanism and the formation of the Karoo Large Igneous Province in southern Africa during the break-up of Africa from Antarctica (Burke and Dewey, 1972; White and McKenzie, 1989; Cole, 1992; Cox, 1992; Thomson, 1999). Apart from the thermal anomaly directly associated with the compensating rise of the asthenosphere, Jackson and Pollack (1984) suggested that rapid sedimentation during syn-rift subsidence, as indicated by the occurrence of listric half-graben faults in the WBB, may additionally have contributed to syn-rift thermal disequilibrium as heat transfer is inhibited by insulating sediments. Another interesting observation is the lateral decrease in syn-rift heat flow by ~10-15 mW/m2 (Figs. 14a and 15) from the proximal (W)NW to the distal (E)SE of the modelled area. The highest values of modelled heat flow, however, do not correspond with the largest thinning of the crystalline crust (Fig. 4c-d). As the crustal thickness and sediment distribution are symmetric with respect to the basin axis, there must be a different reason for this syn-rift heat flow pattern (that shows the largest lateral gradients in (S)W-(N)E direction; Fig. 14a). A possible explanation would be the strike-slip nature of the basin-forming mechanism that may have resulted in locally different stretching factors of the crust and the lithospheric mantle (e.g. Royden and Keen, 1980; Beaumont et al., 1982; Madon and Watts, 1998). 5.3.2. Role of post-rift mantle dynamics: hotspot mechanism and margin uplift The calibration of maximum paleo-heat fluxes against measured vitrinite reflectance suggests that the long-term, basin-scale post-rift geothermal development is complex. As commonly inferred for rift basins, post-rift thermal anomalies may be a result of either margin uplift activities or local magmatic and fluid flow effects (e.g. Allen and Allen, 2005, Hantschel and

19

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Kauerauf, 2009). In southern Africa, the post-rift thermal reconstruction becomes even more problematic as there is no consensus regarding the timing, duration and processes of mantlerelated epeirogenic uplift (e.g. Nyblade, 2003; de Wit, 2007). Our approach of thermal calibration (see Fig. 12) therefore becomes an important aspect of this study as it enables us to test the integration of the two major products of possible post-rift mantle dynamics, i.e. erosion and thermal perturbation. The good agreement between measured and simulated vitrinite reflectance trends (Fig. 13) enables us to further corroborate the validity of the various geologic assumptions regarding the timing as well as the variations of post-rift paleo-thermal anomalies in the WBB. The variation of paleo-heat flow from ~90-94 mW/m2 (proximal) to about ~58-63 mW/m2 (distal) during the late Cretaceous peak event at ~75 Ma (Figs. 14b and 15) supports the idea of a thermal anomaly which may have been linked to spatially confined mantle-related activities (hotspots; e.g. Duncan, 1981; Dingle et al., 1983; Davies, 1997a; Roux, 2007). It is therefore likely that the Santonian-Early Campanian (84-75 Ma) and Paleogene (43-38 Ma) erosion might be related to intrusions (including the emplacement of kimberlites and related rocks; Dingle et al., 1983; Smith et al., 1985; Tinker et al., 2008) belonging to the late Cretaceous to early Tertiary hotspot events in southern Africa. The youngest erosion (between 16 and 11 Ma), representing the most prominent one based on seismic-stratigraphy (Fig. 9), well correlation and measured vitrinite reflectance data, is considered as being related to the late Tertiary margin uplift event (at ~late Miocene; Dingle et al., 1983; Nyblade and Robinson, 1994; Hartnady and Partridge, 1995; Davies, 1997a, 1997b; Nyblade, 2003). The popular explanation for the late Miocene epeirogenic uplift in southern African is related to the so-called “African Superswell” mantle plume head (e.g. Nyblade and Robinson, 1994; Hartnady and Partridge, 1995; Nyblade, 2003). This African Superswell is estimated to be responsible for topography and bathymetry anomalies of about 500 m wavelengths with many superimposed shorter-wavelengths features (Nyblade and Robinson, 1994; Nyblade, 2003). Consistent with this hypothesis, Hirsch et al. (2010) also suggested a renewed sub-crustal thinning to account for the late Miocene uplift and erosion in their tectonic evolution models of the Orange Basin (southwestern South African margin). Other authors (e.g. Nyblade, 2003; de Wit, 2007; Tinker et al., 2008) have considered as equally important the possible contribution from long-lived Mesozoic periods of denudation and uplift scenarios that may particularly be linked to the Late Jurassic-Cretaceous volcanism in southern Africa. 5.3.3. Implications for petroleum prospectivity of the basin Active petroleum systems require the availability of good-quality source rocks that have attained sufficient thermal maturity (e.g. Magoon and Dow, 1994). Our data-constrained present-day and past thermal fields allow a critical evaluation of the thermal maturity trends of the intersected source rocks in the WBB. The maturity of the Valanginian and MidHauterivian source rocks (Fig. 16a-f) is mostly controlled by significant burial and increased heat flow associated with syn-rift tectonism. Whereas, the increased sedimentation rates during late Albian to early Cenomanian times aided the maturity of Late Hauterivain, Barremian and Aptian source rocks (see Figs. 2 and 16a-f). The Aptian source maturity trend suggests additional contribution from the hotspot-related heat flow peak at ~75 Ma. Despite

20

ACCEPTED MANUSCRIPT

SC

RI PT

the significant contribution of post-rift burial and heat flow events on thermal maturity trends in the WBB, the Turonian source is largely immature at present-day. The overall timing of thermal maturation, which started as early as the syn-rift phase, seems very favourable for successful charge of potential traps in the basin. The evidence of ‘gas window’ maturity (i.e. 1.3 – 2.5 %Ro) recorded by the syn-rift source rocks in the Late Cretaceous at ~110 Ma particularly in the deeper sections of the WBB (see Well E-BD2 and E-AA1; Fig. 16c-f), is consistent with the significant gas discoveries along the northern flank of the Bredasdorp Basin by Soekor (now PetroSA) in the 1980s. However, the complex postrift tectonics leading to margin uplift and fault rejuvenation (e.g. Dingle et al., 1983; Nyblade, 2003; de Wit, 2007; Roux, 2007) makes a case for either hydrocarbon loss or re-migration of previously trapped hydrocarbons into adjacent sub-basins in the Outeniqua Basin. This is probably responsible for the dry holes encountered during the early 1980s exploration campaigns in the Bredasdorp Basin.

M AN U

6. Conclusions

EP

TE D

A seismic and gravity consistent crust-scale 3D structural model of the WBB was utilised as a basis to forwardly model the present-day- and paleo-thermal field regimes in the southern South Africa continental margin. The modelling approach implements validation and calibration workflows that allow the assignment of well-constrained boundary conditions and thermal properties, which has led us to investigate the thermal evolution of the WBB since rifting times. An important advantage of our approach is that the present-day crustal thermal model readily provides first order constraints for simulation of basin-scale thermal evolution. In general, the resultant thermal models exhibit a very reasonable agreement with measured borehole temperatures (i.e. overall correlation coefficient of cc=0.90 and an average deviation o of -0.57 C) and vitrinite reflectance data. In summary, our findings on the controlling mechanisms governing the distribution of present-day and past thermal field regimes in the WBB are the following: Present-day thermal field:

AC C

1. Major controls on the present-day thermal field are exerted by: (a) the interplay of shallow and deep geotherms as governed primarily by the geometries of the heatproducing crystalline crust and sediments in conjunction with the stratigraphic units’ thermal properties and (b) basal heat flow from the mantle. 2. The partitioning into colder and hotter regions, at the basin’s margins and NW-SE axis, respectively, reflects the superposition of low thermally conductive sediments over high thermally conductive and high radiogenic heat producing deeper crust (particularly from the granitic-dominated upper crust). 3. The overall crustal thermal field suggests that the predicted variations in present-day surface heat flow in the Southern African continental margin are much more controlled by crustal heat generation, resulting from variations in crustal thicknesses, rather than by the present-day mantle heat flow. This inference becomes readily more elucidative when the surface heat fluxes and crustal thickness variations of the WBB

21

ACCEPTED MANUSCRIPT are compared with those of the Orange Basin in the adjacent southwestern South Atlantic Margin. Mesozoic-Cenozoic thermal evolution:

RI PT

1. The results of thermal calibration workflows indicate that the classic passively cooling post-rift phase subsequent to syn-rift thermal anomaly fails to reproduce all available measured vitrinite reflectance data, thereby suggesting non-constant evolution of paleo-geothermal gradients in the WBB. The best-fit forward model suggests that the basin is characterised by three major phases of thermal disequilibrium related to mantle dynamics, firstly starting with a rifting-type anomaly and later followed by two additional episodes during post-rift times.

M AN U

SC

2. The height and width of the syn-rift thermal anomaly suggests that its mantle heat flow contribution exerts the strongest control on the long-term basin thermal evolution, with the variations of its heat fluxes and geothermal gradients reflecting a direct consequence of instantaneous crustal stretching and rapid syn-rift burial rates. The lateral variations of syn-rift heat flow also suggest a differential thinning of the lithosphere whereby the crust might have been significantly more thinned than the lithospheric mantle, partly in response to the strike-slip basin-forming mechanism.

TE D

3. The main post-rift heat flow peaks relate to uplift and erosion associated to a Late Cretaceous mantle-related hotspot event and another major episode of margin uplift during late Miocene times. As indicated by measured vitrinite reflectance trends and reconstructed erosion scenarios, the hotspot event represents a more localised and transient thermal event, while the late Miocene event recording the highest amounts of post-rift erosion represents a seemingly much more regional post-rift thermal event. Additional minor contribution is provided by the Paleogene erosion which may speculatively reflect additional hotspot-related magmatism during late Cretaceous to early Tertiary times.

EP

4. Our results support the hypothesis of post-rift mantle dynamics as influencing the long-term paleo-geothermal gradients in the WBB even though the precise mechanism for these mantle processes is still problematic to establish.

AC C

The presented 3D thermal model currently represents the most advanced insights on the thermal evolution of the WBB, and therefore serves as robust boundary conditions for constraining the timing of petroleum generation, migration, accumulation as well as possible past and present-day leakage episodes in the study area. The data-integrative thermal calibration workflow implemented in this study may also offer useful guide for other crust-scale thermal history studies. Acknowledgements This research project was conducted at the Helmholtz Centre Potsdam, GFZ German Research Centre for Geosciences, Potsdam under the framework of Inkaba yeAfrica, a collaborative research cooperation program between Germany and South Africa. The authors wish to thank the Petroleum Agency of South Africa and the Petroleum Corporation of South

22

ACCEPTED MANUSCRIPT Africa (PetroSA) for provision of exploration data and permission to publish our findings. The first author is grateful to the colleagues at the Section 4.4 “Basin Analysis” and Section 4.3 “Organic Geochemistry” for providing the necessary logistical supports during his research visit at the GFZ Potsdam, and particularly, to Mr. Björn Lewerenz for assisting with the usage of the GMS (GeoModelling System) software for present-day thermal modelling. This is Inkaba yeAfrica contribution 84.

AC C

EP

TE D

M AN U

SC

RI PT

References Allen, P.A., and J.R. Allen (2005), Basin Analysis: Principles and Applications, 2nd edn. Blackwell Publishing, p. 549. Amante, C., and B.W. Eakins (2009), ETOPO1 1 arc-Minute global relief model: procedures, data sources and analysis. NOAA Technical Memorandum NESDIS NGDC-24, 19. Andersen, N.J. B. and M.A.G. Andreoli (1990), The structural evolution of the coastal area between Danger Point and Struisbaai in the southern Cape fold belt, with implications for the siting of a nuclear power station. South African Journal of Science, 86, 499 – 511. Athy, L.F. (1930), Density, porosity, and compaction of sedimentary rocks: Am. Assoc. Pet. Geol. Bull., 14(1), 1-24. Bauer, K., Neben, S., Schreckenberger, B., Emmermann, R., Hinz, K., Fechner, N., Gohl, K., Schulze, A., Trumbull, R. and Weber, K. (2000), Deep structure of the Namibia continental margin as derived from integrated geophysical studies. J. Geophys. Res. 105 (B11), 25829–25853. Bayer, U., M. Scheck, and M. Köhler (1997), 3D modelling of the thermal field in the Northeast German basin. Geol. Rundsch 86, 241-251. Beaumont, C., C.E. Keen, and R. Boutilier (1982), On the evolution of rifted continental margins; comparison of models and observations for Nova Scotian margin. Geophysical Journal of R. Astr. Society, 70, 667-715. Ben-Avraham, Z., C.J.H. Hartnady, and J.A. Malan (1993), Early tectonic extension between the Agulhas Bank and the Falkland Plateau due to the rotation of the Lafonia micro-plate. Earth Planet. Sci. Lett., 117, 43–58, doi:10.1016/0012-821X(93)90116-Q. Ben-Avraham, Z., C.J.H. Hartnady, and K.A. Kitchin (1997), Structure and tectonics of the Agulhas-Falkland fracture zone, Tectonophysics, 282, 83–98, doi:10.1016/S00401951(97)00213-8. Ben-Avraham, Z. and Zoback, M.D. (1992), Transform-normal extension and asymmetric basins: an alternative to pull-apart models. Geology, 20, 423-426. Biddle, K.T., Uliana, M.A., Mitchum, R.M., Fitzgerald, M.G. and Wright, R.C. (1986), The stratigraphic and structural evolution of the central and eastern Magellanes Basin, southern South America. In: Foreland Basins (Ed. By P.A. Allen and P. Homewood), International Association of Sedimetologists Special Publication No. 8, 41-61. Bjørlykke, K. (2010), Heat Transport in Sedimentary Basins. Petroleum Geoscience, Springer Berlin Heidelberg, 2010, pp. 253-257. Broad, D.S., E.H.A. Jungslager, I.R. Mclachlan, and J. Roux, (2006), Geology of the offshore Mesozoic basins. In: The Geology of South Africa (Ed. by M. R. Johnson et al.) pp. 553– 571, Geol. Soc. South Africa, Pretoria.

23

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Brown Jr., L.F., J.M. Benson, G.J. Brink, S. Doherty, A. Jollands, E.H.A. Jungslager, J.H.G. Keenen, A. Muntingh, and N.J.S. Van Wyk, (1995), Sequence stratigraphy in offshore South African divergent basins. An atlas on exploration for Cretaceous lowstand traps. Soekor (Pty) Ltd. Studies in Geology, Am. Assoc. Pet. Geol., 41, 184. Burden, P.L.A. (1992), Soekor, partners explore possibilities in Bredasdorp Basin off South Africa. Oil and Gas Journal, 90, 109-112. Burke, K. and Dewey, J.F. (1972), Plume-generated triple junctions: key indicators in applying plate tectonics to old rocks. Journal of Geology, 81, 403-433. Christensen, N.I., and W.D. Mooney (1995), Seismic velocity structure and composition of the continental crust: a global view. J. Geophys. Res., 100, 9761–9788. Clauser, C. and Huenges, E. (1995), Thermal conductivity of rocks and minerals. In: Rock Physics & Phase Relations: A Handbook of Physical Constants, (Ed. By T.J. Ahrens), American Geophysical Union, 105-126. Cochran, J.R. (1983), Effects of finite rifting times on the development of sedimentary basins: Earth Planet. Sci. Lett., 66, 289–302, doi: 10.1016/0012-821X(83)90142-5. Cole, D.I. (1992), Evolution and development of the Karoo Basin. In: Inversion Tectonics of the Cape Fold Belt, Karoo and Cretaceous Basins of Southern Africa (Ed. by M.J. de Wit and I.G.D. Ransome), Balkema, Rotterdam, pp. 87-99. Cox, K.G. (1992), Karoo igneous activity\ and early stages of the break-up of Gondwanaland. In: Magmatism and the Causes of Continental Break-up (Ed. By B.C. Storey, T. Alabaster and R.J. Pankhurst), Journal of the Geological Society, London, Special Publication Vol. 68, 137-148. Davies, C.P.N. (1997a), Hydrocarbon evolution of the Bredasdorp Basin, offshore South Africa: from source to reservoir. Unpublished PhD Thesis, University of Stellenbosch 286p. Davies, C.P.N. (1997b), Unusual biomarker maturation ratio changes through the oil window, a consequence of varied thermal history. Org. Geochem., 27(7/8), 537-560. deWit, M.J., (2007), The Kalahari Epeirogeny and climate change: differentiating cause and effect from Core to Space. S. Afr. J. Geol., 101, 367–392. de Wit, M.J., and I.G.D. Ransome (1992), Regional inversion tectonics along the southern margin of Gondwana. In: Inversion Tectonics of the Cape Fold Belt, Karoo and Cretaceous Basins of Southern Africa (Ed. by M.J. de Wit and I.G.D. Ransome), Balkema, Rotterdam, pp. 15–21. Dowdle,W.L., Cobb,W., 1975. Static formation temperature from well logs- an empirical method. J. Petr. Tech. 1326–1330. Dingle, R.V., W.G. Siesser, and A.R. Newton, (1983), Mesozoic and Tertiary Geology of Southern Africa. A.A. Balkema, Rotterdam, p. 375. Duncan, R. A. (1981), Hotspots in the Southern Ocean- an absolute frame of reference for motion of the Gondwana continent. Tectonophysics, 74, 29-42. Francu, E., J. Francu, J. Kalvoda, H. S. Poelchau, and J. Otava (2002), Burial and uplift history of the Palaeozoic Flysch in the Variscan foreland basin (SE Bohemian Massif, Czech Republic). EGU Stephan Mueller Special Publication Series, 1, 167–179. Fishwick, S., 2010. Surface wave tomography: imaging of the lithosphere–asthenosphere boundary beneath central and southern Africa. Lithos, 120, 63–73.

24

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Gladczenko, T.P., Skogseid, J. and Eldholm, O. (1998), Namibia volcanic margin. Marine Geophysical Researches 20 (4), 313–341. Gohl, K. and G. Uenzelmann-Neben, (2001), The crustal role of the Agulhas Plateau, southwest Indian Ocean: Evidence from seismic profiling. Geophys. J. Int., 144, 632–646, doi:10.1046/j.1365-246x.2001.01368.x. Goutorbe, B., F. Lucazeau, and A. Bonneville (2008), The thermal regime of South African continental margins. Earth Planet. Sci. Lett., 267, 256–265. Guillou, L., and C. Jaupart (1995), On the effect of continents on mantle convection. J. Geophys. Res. 100 (B12), 24217–24238. Guiraud, R., and W. Bosworth (1997), Phanerozoic geodynamic evolution of northeastern Africa and the northwestern Arabian platform. Tectonophysics 315, 73–108. Hantschel, T., and A.I. Kauerauf AI (2009), Fundamentals of Basin and Petroleum Systems Modeling. Springer-Verleg Berlin Heidelberg, p. 476. Hartnady, C. J. H., and T.C. Partridge (1995), Neotectonic uplift in Southern Africa: a brief review and geodynamic conjecture. In Centennial Geocongress (1995): Extended Abstracts, Vol. 11, pp. 456-459. Geology Society of South Africa. Hermanrud, C. (1993), Basin modelling techniques: an overview. In: Basin modelling: advances and applications (Ed. by A.G. Doré, J.H. Augustson, C. Hermanrud, D.J. Steward and É. Sylta), NFP Spec Publ 3. Elsevier, Amsterdam, pp 1–34. Hirsch, K.K., M. Scheck-Wenderoth, D.A. Paton, and K. Bauer (2007), Crustal structure beneath the Orange Basin, South Africa. S. Afr. J. Geol., 110, 249-260. Hirsch, K.K., Scheck-Wenderoth, M., van Wees, J.-D., Kuhlmann, G. (2010), Tectonic subsidence history and thermal evolution of the Orange Basin. Marine and Petroleum Geology, 27 (3), 565–584. Jackson, M.J., and H.N. Pollack (1984), On the sensitivity of parameterized convection to the rate of decay of internal heat sources. J. Geophy. Res., 89, (B12), 10,103-10,108. Johnson, M.R., C.R. Anheausser and R.J. Thomas (2006), The geology of South Africa. Geological Society of South Africa. Johannesburg, Council of Geoscience, Pretoria. P. 691. Kounov, A., S. Niedermann, M.J. de Wit, G. Viola, M. Andreoli, and J. Erzinger (2007), Present day denudation rates bearing on erosion processes along selected west- and southfacing sections of the South African Great Escarpment and its interior derived fromin situ produced cosmogenic 3He and 21Ne. S. Afr. J. Geol., 110, 235–248. Lock, B.E. (1980), Flat-plate subduction and the Cape Fold Belt of South Africa. Geology, 8, 35-39. Ludwig, W.J., J.E. Nafe, and C.L. Drake (1970), Seismic refraction. In: The Sea (Ed. By A.E. Maxwell), Wiley-Interscience, New York, pp. 53-84. Maclennan, J., D. McKenzie, and K. Gronvöld (2001), Plume-driven upwelling under central Iceland, Earth Planet. Sci. Lett., 194, 67–82. Madon, M.B., and A.B. Watts (1998), Gravity anomalies, subsidence history and the tectonic evolution of the Malay and Penyu Basins (offshore Peninsular Malaysia). Basin Res., 10, 375-392. Magoon, L.B. and W.G. Dow (1994), The petroleum system. In: The Petroleum SystemdFrom Source to Trap (Ed. By L.B. Magoon, and W.G. Dow), AAPG Memoir, vol. 60, pp. 3–24.

25

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Martin, A. K. (1987), Plate reorganizations around Southern Africa, hotspots and extinctions. Tectonophysics, 142, 309-316. Martin, A.K., and C.J.H. Hartnady (1986), Plate tectonic development of the south west Indian Ocean: a revised reconstruction of east Antarctica and Africa. J. Geophys. Res., 91 (B5), 4767–4786. Maystrenko, Y., M. Scheck-Wenderoth, A. Hartwig, Z. Anka, A.B. Watts, K.K. Hirsch, and S. Fishwick (2013). Structural features of the Southwest African continental margin according to results of lithosphere-scale 3D gravity and thermal modelling. Tectonophysics, 604, 104-121. McKenzie, D. (1978), Some remarks on the development of sedimentary basins. Earth Planet. Sci. Lett., 40, 25-32. Mcmillan, I.K., G.I. Brink, D.S. Broad, and J.J. Maier (1997), Late Mesozoic basins off the south coast of South African. In: African Basins (Ed. by R. C. Selley), pp. 319-376, Elsevier, Amsterdam. Midttømme, K. and Roaldset, E. (1999), Thermal Conductivity of Sedimentary Rocks: Uncertainties in Measurement and Modelling. In: Muds and Mudstones: Physical and Fluid Flow Properties (Ed. by A. C. Aplin, A. J. Flees and J. H. S. Macquaker), 158, 45-60. Geological Society, London, Special Publications London. Noack, V., Y. Cherubini, M. Scheck-Wenderoth, B. Lewerenz, T. Höding, A. Simon, and I. Moeck (2010), Assessment of the Present-Day Thermal Field (NE German Basin) Inferences from 3D Modelling. Chemie der Erde - Geochem, 70, 47-62. Nyblade, A.A., and S.W. Robinson (1994), The African superswell. Geophy. Res. Lett. 21 (9), 765–768. Nyblade, A.A. (2003), The origin of the African superswell. American Geophysical Union, Fall Meeting 2003, abstract #S51F-01. Ondrak, R., Wenderoth, F., Scheck, M. and Bayer, U. (1998), Integrated geothermal modelling on different scales in the Northeast German basin. - Geologische Rundschau, 87, 1, 32-42. Parsiegla, N., K. Gohl, and G. Uenzelmann-Neben (2007), Deep crustal structure of the sheared South African continental margin: First results of the Agulhas-Karoo Geoscience Transect. S. Afr. J. Geol., 110, 393–406, doi:10.2113/gssajg.110.2-3.393. Parsiegla, N., J. Stankiewicz,. K. Gohl, T. Ryberg, and G. Uenzelmann-Neben (2009), Southern African continental margin: Dynamic processes of a transform margin. Geochem. Geophys. Geosyst., 10, Q03007, doi:10.1029/2008GC002196. Petroleum Agency South Africa - PASA (2012) Petroleum exploration in South Africa: information and opportunities. PASA, Cape Town, p. 30. Pitman, W.C., III, and J.A. Andrews, (1985), Subsidence and thermal history of small pullapart basins: Special Publication, Society of Economic Paleontologists and Mineralogists, 37, 45-119. Reid, D.L., Cooper, A.F., Rex, D.C., and Harmer, R.E. (1990), Timing of post-Karoo alkaline volcanism in southern Namibia. Geological Magazine, 016, 316_322. Poelchau, H.S., D.R. Baker, Th. Hantschel, B. Horsfield, and B. Wygrala (1997), Basin simulation and the design of the conceptual basin model. In: Petroleum and basin evolution (Ed. by D.H. Welte, B. Horsfield and D.R. Baker), Springer, Berlin - Heidelberg - New York, pp. 5-70, 1997.

26

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Roux, J. (2007), Republic of South Africa 2007 license round- Area C: Proximal Bredasdorp Basin. Petroleum Agency South Africa, p. 26. Royden, L., and C.E. Keen (1980), Rifting processes and thermal evolution of the continental margin of eastern Canada determined fron subsidence curves. Earth and Planetary Science Letters, 51, 714-717. Scheck-Wenderoth, M., and Y. Maystrenko (2013), Deep control on shallow heat in sedimentary basins. Energy Procedia, 40, 266-275. Scrutton, R., and R. Dingle (1976), Observations on the processes of sedimentary basin formation at the margins of southern Africa. Tectonophysics, 36(1–3), 143-156. Seranne, M. and Anka, Z., (2005), South Atlantic continental margins of Africa: a comparison of the tectonic vs climate interplay on the evolution of equatorial west Africa and SW Africa margins. Journal of African Earth Sciences, 43 (1–3), 283–300. Sonibare, W.A., J. Sippel, M. Scheck-Wenderoth, and D. Mikeš (2014), Crust-scale 3D model of the Western Bredasdorp Basin (Southern South Africa): data-based insights from combined isostatic and 3D gravity modelling. Bas. Res., doi: 10.1111/bre.12064. Sweeney, J.J., and A.K. Burnham (1990), Evaluation of a simple model of vitrinite reflectance based on chemical kinetics, Am. Assoc. Pet. Geol. Bull., 74, 1559–1570. Tankard, A. J., M.P.A. Jackson, K.A. Eriksson, D.K. Hobday, D.R. Hunter, and W.E.L. Minter (1982), Crustal Evolution of Southern Africa, 523 pp., Springer, New York. Thomson, K. (1998), When did the Falklands rotate? Mar. Pet. Geol., 15, 723-736. Thomson, K. (1999), Role of continental break-up, mantle plume development and fault reactivation in the evolution of the Gamtoos Basin, South Africa. Marine and Petroleum Geology, 16, 409-429. Tinker, J., M.J. De Wit, and R. Brown (2008), Linking source and sink: evaluating the balance between onshore erosion and offshore sediment accumulation since Gondwana break-up, South Africa. Tectonophysics, 455, 94–103. Tissot, B.P., and D.H. Welte (1984), Petroleum formation and occurrence. Springer-Verleg Berlin Heidelberg, p. 679. Ungerer, P., J. Burrus, B. Doligez, P.Y. Chenet, and F. Bessis (1990), Basin evaluation by integrated two-dimensional modeling of heat transfer, fluid flow, hydrocarbon generation and migration. Am. Assoc. Pet. Geol. Bull., 74, 309-335. Van Der Merwe, R., and J. Fouché (1992), Inversion tectonics in the Bredasdorp Basin, offshore South Africa. In: Inversion tectonics of the Cape Fold Belt, Karoo and Cretaceous basins of Southern Africa (Ed. by M.J. De Wit & I.G.D. Ransome), A.A. Balkema, Rotterdam, 49-59. Van Wyk, N.J.S., J. Roux, D.S. Broad, V.H. Valicenti, and G.J. Brink (1994), Correlation of old and new horizons, South Coast Blocks 7 to 12. SOEKOR unpublished rept., SOEEXP-RPT-240, p. 7. Veevers, J.J., D.I. Cole, and E.J. Cowan (1994). Southern Africa: Karoo basin and Cape Fold Belt. In: Permian-Triassic Pangean Basins and Foldbelts along the Panthalassan Margin of Gondwanaland (Ed. by J.J. Veevers and C.McA. Powell), Geological Society of America, Memoir, 184, 223-279, Boulder, Colorado. Vilà, M., M. Fernández, and I. Jiménez-Munt (2010), Radiogenic heat production variability of some common lithological groups and its significance to lithospheric thermal modelling. Tectonophysics, 490, 152–164, doi:10.1016/j.tecto.2010.05.003.

27

ACCEPTED MANUSCRIPT

EP

FIGURE CAPTIONS

TE D

M AN U

SC

RI PT

Visser, J.N.J., and H.E. Praekelt (1996), Subduction, mega-shear systems and Late Paleozoic basin development of Gondwana. Geologische Rundschau, 86, 632-646. Wollenberg, H.A. and Smith, A.R. (1987), Radiogenic heat production of crustal rocks: An assessment based on geochemical data. Geophysical Research Letters 14, doi: 10.1029/GL014i003p00295. Walker, J.D., and J.W. Geissman (compilers, 2009), Geologic Time Scale. Geol. Soc. America, doi:10.1130/2009.CTS004R2C. Wangen, M. (1995), The blanketting effect in sedimentary basins. Bas. Res.,7, 283-298. White, R. S. (1997), Rift-plume interaction in the North Atlantic, Philos. Trans. R. Soc. London, Ser. A, 355, 319–339. White, R.S., and D. Mckenzie (1989), Magmatism at rift zones: the generation of volcanic continental margins and flood basalts. J. Geophy. Res., 94, 7685-7729. Wu, J. E., K. Mcclay, P. Whitehouse, and T. Dooley (2009), 4D analogue modelling of transtensional pull-apart basins. Mar. Pet. Geol., 26, 1608-1623, doi:10.1016/j.marpetgeo.2008.06.007. Wygrala, B.P. (1989), Integrated study of an oil field in the southern Po Basin, Northern Italy. PhD thesis, University of Cologne, Germany. Xie, X., and P.L. Heller (2009): Plate tectonics and basin subsidence history. GSA Bulletin, 121 (1/2), 55–64, doi: 10.1130/B26398.1. Yalçin, M.N., R. Littke, and R.F. Sachsenhofer (1997), Thermal history of sedimentary basins. In: Petroleum and basin evolution (Ed. by D.H. Welte, B. Horsfield and D.R. Baker), Springer, Berlin - Heidelberg - New York, pp. 71-167. Yükler, M.A., C. Cornford, and D. Welte. (1979), Simulation of geologic, hydrodynamic, and thermodynamic development of a sediment basin – a quantitative approach. In: Initial Reports of the Deep Sea Drilling Project (Ed. by U. von Rad, W. B. F. Ryan, and al.), 761–771.

AC C

Figure 1. (a) Location of the Western Bredasdorp Basin (WBB) in the southern offshore of South Africa. The extent of onshore and offshore geological and tectonic provinces is modified after Johnson et al. (2006), Broad et al. (2006) and PASA (2012). Location of the Agulhas-Karoo deep seismic transect is modified after Parsiegla et al. (2009). Also shown are the structural configuration and major fault trends of the southern offshore Mesozoic basins (i.e. Bredasdorp, Infanta, Pletmos, Gamtoos, Algoa and Southern Outeniqua). AGA - Agulhas Arch; IFA – Infanta Arch; DMR – Diaz Marginal Ridge; COB – Continent-Ocean Boundary; AFFZ – Agulhas-Falkland Fracture Zone. (b) Bathymetric map of the modelled WBB detailing the integrated dataset for the 3D structural and thermal model. Bathymetry is constructed from ETOPO1 global relief model (Amante and Eakins, 2009) and seismic profiles. AGA - Agulhas Arch; IFA – Infanta Arch; m.a.s.l. – metre above sea-level. Figure 2. Chrono-stratigraphic chart of the Western Bredasdorp Basin showing key unconformities and tectonic stages in relation to geodynamic events.

28

ACCEPTED MANUSCRIPT Figure 3. Interpreted NW-SE seismic profile in two-travel time along the main basin axis showing the mapped key basin-wide unconformities and major faults.

RI PT

Figure 4. Structure and stratigraphy of the WBB (after Sonibare et al., 2014). (a) and (b) cumulative thicknesses and major fault trends of the syn-rift and post-rift sequences, respectively; (c) cumulative deep crustal thickness underlying the basin-fill; (d) crust-mantle boundary (Moho) depth trend; (e) W-E cross sectional profile through the WBB structural model; location shown in (c) and (d). UTM Zone 34S.

SC

Figure 5. Results: present-day thermal field. (a) geological units at 2 km depth; (b) porosities at 2 km depth; (c) bulk thermal conductivities at 2 km depth; (d) temperature distribution at 2 km depth; (e) bulk thermal conductivities at 5 km depth; (f) temperature distribution at 5 km depth level; numbers in (a), (b), (c) and (e) refer to geological units in Fig. 4e. Plus sign = wells with measured present-day temperature; indicated well names represent wells given in Figure 7. UTM Zone 34S.

TE D

M AN U

Figure 6. Results (continued): present-day thermal field. (a) temperature distribution at 10 km depth; (b) temperature distribution at 15 km depth; (c) temperatures at the Moho; (d) heat flow at 10 km depth; (e) heat flow at the top of the crystalline basement; (f) heat flow at the top of the model, i.e. at the seafloor; Present-day geothermal gradients calculated for specific depth intervals below the seafloor: (g) seafloor to 2 km b.sf; (h) 2 km b.sf to 5 km b.sf; b.sf – below the seafloor. Plus sign = wells with measured present-day temperature; indicated well names represent wells given in Figure 7. UTM Zone 34S.

EP

Figure 7. Depth plots of temperatures for all wells (upper left plot) and exemplarily chosen wells showing the general consistency of the final thermal model with measured temperatures; solid blue line is the modelled present-day geothermal gradient from the numerically simulated thermal evolution models; well locations are indicated in Figs. 1b, 5 and 6; Temp – temperature. Figure 8. Model validation for the best-fit present-day thermal model using a lower boundary 2

AC C

condition of 38 mW/m heat flux at the Moho. (a) histogram of the differences between modelled and measured temperatures for the complete temperature data set (i.e. for all wells); (b) histogram of the differences between modelled and measured temperatures for 16 wells at which DSTs and SBHTs are available (wells with highest confidence); n – number of temperature data; cc – correlation coefficient (cc=0 for no correlation, and cc=1 for perfect correlation); temperature difference ∆T = modelled minus measured. Figure 9. Model setup for the Mesozoic-Cenozoic thermal evolution showing the defined stratigraphic events (as a series of deposition, non-deposition (hiatus) and erosion), major unconformities (Fig. 3) and vertical and lateral variations of lithofacies. Figure 10. Calibration data for the Mesozoic-Cenozoic thermal evolution showing measured vitrinite reflectance (%VRo) plots for the observed three paleo-geothermal gradient subdomains, i.e. (a) “proximal”, (b) “central” and (c) “distal” along the NW-SE trending axis

29

ACCEPTED MANUSCRIPT of the WBB; red oval shape indicates the occurrence of VR bulges/separate profiles at intermediate depths; also note the decrease in sublinearity of the VR trends from proximal to distal; (d) bathymetric map indicating VR subdomains and the location of wells and NW-SE profile. Shown well names are wells given in Figures 15-16.

SC

RI PT

Figure 11. Mesozoic-Cenozoic distribution of amount of eroded sediments. (a) amount of Hauterivian (134-133 Ma) erosion as derived from seismic interpretation; (b) amount of late Cretaceous (84-75 Ma) erosion; (c) amount of early Tertiary (43-38 Ma) erosion; (d) amount of late Miocene (16-11 Ma) erosion; (b) and (c) initial values after Davies (1997a); (d) initial values from seismic interpretation, well correlation and Sonibare et al. (2014); the overall spatial distributions of (b), (c) and (d) are calibrated against measured vitrinite reflectance. Plus sign = wells with measured present-day temperatures and vitrinite reflectance. Shown well names are wells given in Figures 15 and 16. UTM Zone 34S. Figure 12. Integrated thermal calibration workflow developed for this study.

M AN U

Figure 13. Calibration of modelled vitrinite trends (%VRo) against measured vitrinite reflectance data (see Figs. 1b and 10 for well location); full thermal calibration is achieved when three paleo-heat flow peaks and four erosion scenarios were considered in the final thermal evolution model.

TE D

Figure 14. Predicted paleo-heat flow. (a) base of sediment flow at 131.5 Ma; (b) base of sediment heat flow at 75 Ma; Plus sign = wells with measured present-day temperatures and vitrinite reflectance. Shown well names are wells given in Figures 15 and 16. UTM Zone 34S.

EP

Figure 15. Heat flow through geological time for selected wells i.e. D-A1, E-BD2 and E-AA1 (see Figs. 11 and 14 for well location); solid lines = predicted heat flow for pre-rift, syn-rift and post-rift successions (see Fig. 4 for colour scheme); dashed black line = calibrated basal heat flow trend at 10 km constant depth.

AC C

Figure 16. Burial history with tempature and vitrinite reflectance overlays for selected wells (i.e. (a) and (b) = D-A1; (c) and (d) = E-BD2; (e) and (f) = E-AA1; see Figs. 11 and 14 for well location) showing the thermal maturity trends of intersected source rocks in the WBB. TABLE CAPTIONS

Table 1. Input thermal properties for 3D present-day thermal modelling. Table 2. Input petrophysical properties for 3D modelling of paleo-thermal evolution. Table 3. Contribution of cumulative sediment and crustal thicknesses to heat flow.

ACCEPTED MANUSCRIPT

Table 1. Input thermal properties for 3D present-day thermal modelling Sand:Shale ratio [%]

85:15 75:25 65:35 10:90 15:85 70:30

0.6 2.8 2.7 2.5 1.9 2.1 2.7 2.6 2.7 2.4

AC C

EP

TE D

M AN U

Water Upper Cenomanian-Cenozoic (15At-Sf) Sand and silt with clay Early Aptian-Upper Cenomanian (13At-15At) Sand with silt and shale Late Hauterivian-Early Aptian (6At1-13At) Sand with silt and shale Mid-Late Hauterivian (5At-6At) Shale Valanginian-Mid Hauterivian (1At-5At) Shale Upper Jurassic -Valanginian (Basement-1At) Sand with conglomerate and clay Pre-rift (Pre-Mesozoic) Meta-sediments Upper crust Granite to granodiorite Lower crust Granodiorite to diorite

Matrix thermal conductivities [W/(m*K)]

RI PT

Principal lithology

SC

Modelled stratigraphic unit

Radiogenic heat production [W/m3]

1.15e-06 1.18e-06 1.21e-06 1.37e-06 1.36e-06 1.19e-06 1.80e-06 2.00e-06 5.00e-07

ACCEPTED MANUSCRIPT Table 2. Input petrophysical properties for 3D modelling of paleo-thermal evolution Stratigraphic unit

Lithofacies

Thermo-mechanical properties

17At - Seafloor

Sandstone (typical) Shaly Sandstone (ss60sh40) Sandy Shale (sh50ss50)

Thermal conductivity [W/(m*K)] 20oC / 100oC / 200oC 3.35 / 2.95 / 2.63 2.78 / 2.53 / 2.33 2.55 / 2.36 / 2.21

K – 17At

Sandstone (ss75sh25) Shaly Sandstone (ss60sh40)

2.80 / 2.54 / 2.34 2.78 / 2.53 / 2.33

0.21 / 0.24 / 0.27 0.20 / 0.24 / 0.27

1.07 / 0.80 0.99 / 0.74

15At – K

Sandstone (typical) Sandy Shale (sh50ss50) Shale (typical)

3.35 / 2.95 / 2.63 2.55 / 2.36 / 2.21 1.64 / 1.69 / 1.73

0.21 / 0.24 / 0.27 0.20 / 0.24 / 0.27 0.21 / 0.24 / 0.27

0.88 / 0.66 1.10 / 0.82 1.63 / 1.22

14At – 15At

Sandstone (typical) Sandstone (ss75sh25) Shaly Sandstone (ss60sh40)

3.35 / 2.95 / 2.63 2.80 / 2.54 / 2.34 2.78 / 2.53 / 2.33

13At – 14At

Sandstone (ss75sh25) Shaly Sandstone (ss60sh40) Sandy Shale (sh50ss50) Shale (typical)

2.80 / 2.54 / 2.34 2.78 / 2.53 / 2.33 2.55 / 2.36 / 2.21 1.64 / 1.69 / 1.73

0.21 / 0.24 / 0.27 0.20 / 0.24 / 0.27 0.20 / 0.24 / 0.27 0.21 / 0.24 / 0.27

1.07 / 0.80 0.99 / 0.74 1.10 / 0.82 1.63 / 1.22

6At – 13At

Sandstone (typical) Shaly Sandstone (ss60sh40) Shale (typical)

3.35 / 2.95 / 2.63 2.78 / 2.53 / 2.33 1.64 / 1.69 / 1.73

0.21 / 0.24 / 0.27 0.20 / 0.24 / 0.27 0.21 / 0.24 / 0.27

0.88 / 0.66 0.99 / 0.74 1.63 / 1.22

5At – 6At

Sandstone (ss75sh25) Shaly Sandstone (ss60sh40) Shale (typical)

2.80 / 2.54 / 2.34 2.78 / 2.53 / 2.33 1.64 / 1.69 / 1.73

0.21 / 0.24 / 0.27 0.20 / 0.24 / 0.27 0.21 / 0.24 / 0.27

1.07 / 0.80 0.99 / 0.74 1.63 / 1.22

1At – 5At

Sandstone (ss75sh25) Sandy Shale (sh50ss50) Shale (typical)

2.80 / 2.54 / 2.34 2.55 / 2.36 / 2.21 1.64 / 1.69 / 1.73

0.21 / 0.24 / 0.27 0.20 / 0.24 / 0.27 0.21 / 0.24 / 0.27

1.07 / 0.80 1.10 / 0.82 1.63 / 1.22

Basement – 1At

Conglomerate Sandstone (typical) Sandstone (ss90sh10) Sandy Shale (sh60ss40) Shale (typical)

2.30 / 2.18 / 2.45 3.35 / 2.95 / 2.63 3.12 / 2.78 / 2.51 2.33 / 2.20 / 2.10 1.64 / 1.69 / 1.73

0.20 / 0.23 / 0.26 0.21 / 0.24 / 0.27 0.21 / 0.24 / 0.27 0.20 / 0.24 / 0.27 0.21 / 0.24 / 0.27

0.68 / 0.51 0.88 / 0.66 0.96 / 0.72 1.20 / 0.90 1.63 / 1.22

Metasediments

2.79 / 2.54 / 2.34

0.20 / 0.23 / 0.26

1.07 / 0.80

Granite

2.6 / 2.4 / 2.25

0.19 / 0.22 / 0.25

2.66 / 1.99

Upper crust ϕ = porosity

SC

RI PT

Radiogenic heat [µW/m3] 20%ϕ / 40%ϕ 0.88 / 0.66 0.99 / 0.74 1.10 / 0.82

0.21 / 0.24 / 0.27 0.21 / 0.24 / 0.27 0.20 / 0.24 / 0.27

M AN U

TE D

EP

AC C

Pre-rift metasediments

Heat capacity kcal/kg/K 20oC / 100oC / 200oC 0.21 / 0.24 / 0.27 0.20 / 0.24 / 0.27 0.20 / 0.24 / 0.27

0.88 / 0.66 1.07 / 0.80 0.99 / 0.74

ACCEPTED MANUSCRIPT Table 3. Contribution of cumulative sediment and crustal thicknesses to heat flow Thickness range [km]

Mean radiogenic heat production [µW/m3]

Post-rift Syn-rift Crystalline Crust 1. Pre-rift metasediments 2. Upper crust 3. Lower crust

0.6 – 3.0 0.5 – 4.5 18.0 – 32.0 1.25 – 5.0 5.0 – 10.0 12.0 – 18.0

1.18 1.31 1.43 1.8 2.0 0.5

SC M AN U TE D EP AC C

Heat flow at top of unit [mW/m2] 0.71 – 3.54 0.65 – 5.88 25.8 – 45.9 2.25 – 9.0 10.0 – 20.0 6.0 – 9.0

RI PT

Composite unit

Figure 1 (a)

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

(b)

Figure 10

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

Proximal

Central

Distal

Figure 11

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 12

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 13

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 14 Base of sediment heat flow at 131.5 Ma ACCEPTED MANUSCRIPT

IFA

RI PT

(a)

Base of sediment heat flow at 75 Ma

IFA

TE D

M AN U

(b)

SC

AGA

AGA

AC C

EP

Heat flow [mW/m2] (shared legend)

Figure 15 Heat ACCEPTED flow evolution at MANUSCRIPT well D-A1 (proximal)

D B

C

Heat flow evolution at well E-BD2 (central)

B

C

M AN U

A

SC

D

RI PT

A

Heat flow evolution at well E-AA1 (distal)

TE D

D

AC C

EP

A

B

C

Figure 16

(c)

EP

(f)

AC C

(e)

TE D

M AN U

SC

(d)

RI PT

ACCEPTED MANUSCRIPT (b)

(a)

(shared legend)

Figure 2

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Onset of rifting in west Gondwana(?)

Figure 3

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 4

(8) ρ = 2700 kg/m3

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

(1)

EP

(7) ρ = 2650 kg/m3

(2) (3)

AC C

(9) ρ = 2800 kg/m3

(1) = geological units in Figure 5

(4) (5) (6)

Figure 5 2 km depth level

2 km depth level

RI PT

ACCEPTED MANUSCRIPT

2 km depth level

AC C

EP

5 km depth level

TE D

M AN U

SC

2 km depth level

5 km depth level

Figure 6

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 7

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 8

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 9

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT