The opening of the South Atlantic, between the Florianopolis and Agulhas-Falkland Fracture Zones (FZs), has often been presented as a case study of continental breakup across a mantle plume (Courtillot et al., 1999). The rising Tristan hot plume is proposed to have impacted the western Gondwana lithosphere (triggering the Paraná-Etendeka continental flood basalts), weakened and thinning it, and led to formation of volcanic rifted margins, breakup, and seafloor spreading (Beccaluva et al., 2020). However, seafloor spreading started > 2000 km south of the Paraná-Etendeka igneous province and propagated north (Franke, 2013), reaching the Walvis Ridge, the inferred plume tail, in Aptian times (Stica et al., 2014). This was well after the last continental flood basalt events at ~120 Ma (Beccaluva et al., 2020). This northward unzipping of the South Atlantic, starting far from the inferred plume head, led to questioning of the triggering role of Tristan plume impingement (Foulger, 2017; Franke, 2013).
Volcanic rifted margins typically contain a thick magmatic layer of seaward dipping reflectors (SDRs). These large melt volumes have often been attributed to melting of anomalously hot sublithospheric material (Morgan et al., 2020; White & McKenzie, 1989). Hot-plume-sourced mantle is conventionally considered to be the key factor creating SDRs during rifting and anomalously thick oceanic crust subsequently. Magma volume was used early on to estimate the proposed mantle thermal anomaly (White & McKenzie, 1989). However, one of the main pitfalls with this approach lies in determining the true volume of magma at volcanic rifted margins (Tugend et al., 2020).
At volcanic rifted margins, continental crust might be highly intruded resulting in a hybrid crust within the ocean-continent transition (OCT) (Figure 1). Intrusions into thinned continental crust cannot be properly resolved using geophysical methods (e.g., seismic reflection or refraction). By contrast, the seismic crustal thickness within the oceanic domain can be used to estimate melt supply rate. This can be readily measured on seismic reflection profiles.
Figure 1: Cartoon highlighting the architecture of the transition between the continental domain and unambiguous oceanic crust at magma-rich margins. The nature of the crust in the OCT is debated. While SDRs (yellow lines) are easily recognized on seismic reflection images, the amount of magma (blue) in the OCT is unknown. Both fully magmatic crust (McDermott et al., 2018) and exhumed continental crust (Geoffroy et al., 2022) are inferred. Here we do not speculate on the nature of hybrid crust but define the landward limit of oceanic crust (LaLOC) where the top of the crust shallows and its base deepens. These two inflexion points mark departure from typical oceanic crust.
We use 24 high-quality deep-penetrating seismic profiles measured by ION GeophysicalTM, south of the Florianopolis FZ to investigate the melt production rate at the onset of seafloor spreading. This occurred after formation of the SDRs and complete rupturing of the continental lithosphere. The data set was derived from published seismic reflection profiles of the southernmost part of the South Atlantic (Figure 2).
Figure 2: Maps of the conjugate South American (left) and West African (right) volcanic rifted margins. Thin black lines indicate the seismic reflection profiles we used. Thicker lines indicate oceanic crust seaward of the landward limit of oceanic crust (LaLOC). Blue, red, purple and yellow lines are the M0.0, M2.0, M3.0 and M4.0 isochrones, respectively (Collier et al., 2017). Black dashed lines indicate fracture zones, black areas show Paranà-Etendka igneous province and green areas show SDR extent (Chauvet et al., 2020). Background: the Free Air gravity anomaly grid (mGal) from satellite altimetry data (Sandwell et al., 2014).
Measuring oceanic crustal thickness at the onset of seafloor spreading requires determining the landward limit of the oceanic crust (LaLOC). We base our work on the remarkably constant structure and thickness of oceanic crust formed by steady-state seafloor-spreading at mid-ocean ridges (Figure 1). From the oceanic domain toward the continents, the top of the basement shallows sharply as the Moho strongly deepens creating a crustal taper and sudden crustal thickening. We define the first continent-ward occurrence of such inflexion points at both top and base of the crust as the LaLOC. We argue that any significant departure of the common geometry of ocean crust marks a rupture in the stable magma production that characterizes steady-state, classical seafloor spreading (Figure 3). We focus on the unambiguous oceanic crust, which starts at the LaLOC, and confidently indicates steady-state seafloor spreading at the initial South Atlantic mid-oceanic ridge.
Figure 3: Crustal thickness variations across the LaLOC for a seismic reflection profile at the West African margin. Vertical exaggeration is ~2x. Thick white lines indicate the top of the basement and Moho. Thin white lines show SDRs. Courtesy of ION Geophysical. (Sauter et al., 2023).
The oceanic Moho is at 2.08 ± 0.15 s TWTT on average beneath the top of the basement at the LaLOC corresponding to a 6.65 ± 0.47 km mean crustal thickness. This is similar to the 6.62 ± 0.86 km average calculated from the worldwide compilation of Christeson et al. (2019) for >100 Ma old oceanic crust away from hotspots. Most of the southernmost Atlantic Ocean opened without anomalously hot or fusible mantle, high magma supply being restricted to the Walvis Ridge area, the inferred plume tail of the Tristan hot spot (Figure 4).
Figure 4: Crustal thickness variation along the M3.0 isochrone along both conjugate margins relative to the Florianopolis FZ. Purple circles: South American plate, dark blue diamonds: African plate. Black dashed line: best polynomial fit. Blue lines: crustal thickness variations from seismic strike lines along the Namibian margin close to the isochrone (< 20 km). Dashed horizontal grey line: 6 km crustal thickness. Grey shaded area: Walvis ridge width from the Free Air gravity anomaly map (satellite altimetry data).
Our results have important implications for the origin of the large melt volumes in proximal parts of the South Atlantic margins and other volcanic rifted margins. Because there is no anomalously hot mantle beneath the initial South Atlantic spreading ridge away from the Walvis ridge area, the plume model requires that high mantle potential temperatures triggered large-volume melt production along the rifted margins, climaxing at the Paraná-Etendeka igneous province, but then returned to normal 8 Myr later at the onset of seafloor spreading. Such a mantle potential temperature fluctuation may be compared to the situation along the Mid-Atlantic ridge south of Iceland. There, mantle potential temperature variations inferred by the plume model to produce crustal thickness variations along the Reykjanes ridge occurred an order of magnitude more slowly. This suggests that the fast cooling required by a plume in the South Atlantic is rather unlikely.
Alternatively, we propose that:
either the magma volume along the South Atlantic volcanic margins must be re-evaluated downward;
and/or explanations other than hotter mantle prior to the onset of seafloor spreading must be favoured to explain massive magmatic production.
This is in line with recent numerical results that demonstrate that the large volumes of magma at volcanic rifted margins can be explained by depth dependent extension and much smaller temperature anomalies than previously suggested. We thus suggest that, contrary to conventional assumptions, hot, plume-sourced mantle is not responsible for large melt volumes during rifting and thicker than average oceanic crust during initial seafloor spreading.
Sauter, D., Manatschal, G., Kusznir, N., Masquelet, C., Werner, P., Ulrich, M., Bellingham, P., Franke, D., and Autin, J. (2023). Ignition of the southern Atlantic seafloor spreading machine without a hot-mantle booster: Scientific Reports13, 1195.
MECHANISMS 
