January 2012 LIP of the Month

The Southeast African Large Igneous Province: a model of its crustal growth and plate-kinematic dispersal

Karsten Gohl¹ and Gabriele Uenzelmann-Neben²

¹Alfred Wegener Institute for Polar and Marine Research, Dept. of Geosciences, Am Alten Hafen 26, D-27568 Bremerhaven, Germany. e-mail: karsten.gohl@awi.de

²Alfred Wegener Institute for Polar and Marine Research, Dept. of Geosciences, Am Alten Hafen 26, D-27568 Bremerhaven, Germany. e-mail: gabriele.uenzelmann-neben@awi.de

Introduction

Two prominent submarine plateaus and ridges exist in the Indian Ocean off southeastern Africa (Fig. 1). To the west, the Agulhas Plateau (AP) is morphologically isolated from the continent by the Agulhas Passage. East of it, the Mozambique Ridge (MOZR) has a morphological connection to the African continent. The AP and MOZR are separated by the Transkei Basin (TB) and a corridor of oceanic crust which is bathymetrically shallower by about 500-1000 m from the surrounding Indian Ocean crust of 5000 m water-depth but also shallower by a few hundred meters than the seafloor of the TB. The fact that this corridor of shallower and/or thicker crust – which we named Transkei Rise (TR) in Gohl et al. (2011) – cannot be dated by magnetic spreading anomalies can be used to argue for its age to be that of the magnetically quiet Cretaceous period (120-80 Ma). Although various authors have discussed the crustal origin and nature of the AP and the MOZR (e.g. Ben-Avraham et al., 1995; Gohl and Uenzelmann-Neben, 2001; Tikku et al., 2002; König and Jokat, 2006; Parsiegla et al., 2008), any relationship between these two plateaus with regard to a potentially common evolutionary history has not been developed. Uenzelmann-Neben et al. (1999), Gohl and Uenzelmann-Neben (2001), and Parsiegla et al. (2008) found geophysical evidence for the AP to be regarded an oceanic Large Igneous Province (LIP), consisting of over-thickened oceanic crust formed at 110-95 Ma. Small continental fragments, which may explain the few dredged rock samples of felsic composition (Tucholke et al., 1981, Ben-Avraham et al., 1995), cannot be entirely excluded, though. Previous geoscientific studies of the MOZR (e.g. Ben-Avraham et al., 1995) failed to reveal its deep crustal nature and origin. Some scattered dredged rock samples (Ben-Avraham et al., 1995) imply a continental affinity of the MOZR. Although plate-kinematic studies adopted a continental MOZR and placed the MOZR adjacent to the Astrid Ridge of Dronning Maud Land in Antarctica (Tikku et al., 2002; König and Jokat, 2006), a continental block of substantial size, as the entire MOZR would be, does not allow a close-fit in reconstructing the pre-breakup position of Africa to Antarctica. König and Jokat (2010) and Leinweber and Jokat (2011) analyzed new densely spaced magnetic survey data across the MOZR and the Mozambique Basin and excluded the presence of major continental blocks throughout the MOZR.

Deep crustal seismic refraction and wide-angle reflection data, using ocean-bottom seismometers (OBS), and multichannel seismic reflection data were collected from the AP and southern MOZR during two research cruises, one with MV Petr Kottsov in 1998 (Uenzelmann-Neben et al., 1999; Gohl and Uenzelmann-Neben, 2001) and the other one with RV Sonne in 2005 (Schlüter and Uenzelmann-Neben, 2007; Parsiegla et al., 2008, 2009). The data analyses provide evidence for the crustal structure and composition of the AP and MOZR (Parsiegla et al., 2008, 2009; Gohl et al., 2011).

Overthickened oceanic crust and strong magmatism

For OBS system description, acquisition and processing parameters of the seismic refraction and reflection data collected on AP and MOZR as well as modeling procedures and results, we refer to Schlüter and Uenzelmann-Neben (2008), Parsiegla et al. (2008, 2009) and Gohl et al. (2011).

The vertical and horizontal seismic velocity distribution and crustal thickness of the southern MOZR is analogue to that observed for the AP (Figs. 2 and 3). Both the AP and the MOZR are characterized by over-thickened, seismically homogeneous, lower crustal units making up between half to two-thirds of the crustal column with P-wave velocities of more than 7.0 km/s and increasing to 7.5-7.6 km/s at the crustal base (Parsiegla et al., 2008; Gohl et al., 2011). These very high velocities imply that the lower crust was thickened by the addition of large volumes of mantle-derived magma. Large quantities of uprising hot peridoditic melts can be generated by adiabatic decompression.  Trumbull et al. (2002) have shown for the Namibian volcanic margin that mantle temperatures of 1450-1560 °C can generate basaltic melts of 14-18% MgO content which results in P-wave velocities between 7.25 and 7.4 km/s. A similar relationship can be expected for the lower crust of the MOZR.

Beneath thin sediments, subparallel stratified wedges which in places overlap can be observed in seismic reflection data from MOZR (Fig. 4) (Gohl et al., 2011). Similar structures have been observed on the AP where they were interpreted as volcanic flows emerging from extrusion centers (Uenzelmann-Neben et al., 1999; Gohl and Uenzelmann-Neben, 2001; Parsiegla et al., 2008). The wedges of inclined reflectors resemble seaward dipping reflectors (SDR) associated with volcanic passive margins (e.g. Hinz, 1981; Franke et al., 2007) and flood basalt volcanic fields. In contrast to the AP, we do not observe extrusion centers on the southern MOZR. Furthermore, with a few exceptions, the flow-like reflections are inclined southwards. Interpreting the flow-like reflections as SDR implies that the seafloor spreading center creating those features must have been active in the south. Seismic data from the Astrid Ridge (AR), the conjugate structure at the Antarctic continental margin, clearly show SDR wedges (Hinz et al., 2004). This strongly resembles our observation from the southern MOZR. Following the formation of SDRs, the southern MOZR was subject to faulting, which may be considered as evidence for the extension during the separation from the AR and AP. In this, the southern MOZR resembles the AP (Parsiegla et al., 2008) where volcanic flows and faults bear witness to extrusive volcanism followed by extension during the fragmentation of the larger LIP into its provinces consisting of the present-day AP, the Northeast Georgia Rise off the Falkland Plateau, and Maud Rise off East Antarctica (Gohl et al., 2011).

Timing and extent of LIP formation

The plate-kinematic reconstructions provide relative constraints for the timing of formation of the LIP components off southeastern Africa. In most recent reconstructions (e.g. by Marks and Tikku, 2001; König and Jokat, 2006), the MOZR was treated as a continental microplate which had an independent motion between M11 and M2 (133-124 Ma) (Marks and Tikku, 2001) and which was located in close vicinity to an extinct spreading center. König and Jokat (2010) and Leinweber and Jokat (2011) present new magnetic data from the MOZR and the Mozambique Basin to its northeast, which indicate volcanic formation of the rise between 140 and 122 Ma. In our model (Fig. 5), the MOZR must have had its own maximum extent by about 120 Ma while the Antarctic-African spreading zone ran underneath its eastern flank. Parts of the Astrid Ridge were possibly attached and developed as part of the same magmatic process. The main volcanic formation of the greater AP in unity with the Northeast Georgia Rise (NEGR) and Maud Rise (MR) can be estimated to be around 100 Ma (Gohl and Uenzelmann-Neben, 2001; Parsiegla et al., 2008) when the region in the vicinity of the triple junction passed over the Bouvet hotspot. This main eruption phase lasted to about 94 Ma (Parsiegla et al., 2008), after which the NEGR and MR detached from the AP by continued regular seafloor spreading.

Both plateau and rise are connected by the TR, a crustal corridor bathymetrically elevated to 4500-3500 m. We can rule out thicker sedimentary cover because it has the same 1.5-1.7 km thickness just to the southwest of the MOZR as observed for the lower lying Transkei Basin (Schlüter and Uenzelmann-Neben, 2007). If normal oceanic spreading with normal oceanic thickness had occurred south and southwest of the MOZR in the Early Cretaceous, the seafloor would have subsided to a mean depth of around 5000 m, similar to the crust surrounding the MOZR-TR-AP province to the east, south, and west. We, therefore, imply a slightly over-thickened crystalline crust for the TR. It is reasonable to assume that there had been a continuous process of excess volcanic activity in the ~20 m.y. period between the main formation of the MOZR and that of the greater AP. The similarity of the crustal structure and seismic velocity-depth distribution between the MOZR and AP is striking and suggests that they developed through the same magmatic regime which produced accreted mantle-derived material of the same mafic to ultramafic composition. The distinctly thinner TR, however, implies that active LIP magmatism continued in between the main formation of the MOZR and that of the greater AP but with much less intensity of accreted mafic to ultramafic volume production.

Conclusions

Analyses of OBS data and seismic reflection profiles from the southern MOZR and Agulhas Plateau provide evidence for the formation of a greater Southeast African LIP. The velocity-depth distribution of the MOZR shows that the both the MOZR and the AP consist of lower crustal units making up between half to two-thirds of the crustal column with P-wave velocities of more than 7.0 km/s, increasing to 7.5-7.6 km/s at the crustal base. These velocities imply that the lower crust was accreted by large volumes of mantle-derived material to form an over-thickened equivalent of an oceanic layer 3. Similar to the AP, the southern MOZR must therefore be of predominantly oceanic origin.

The seismic reflection data reveal seaward-dipping reflector sequences similar to those observed on AP and Astrid Ridge. The southern MOZR resembles the AP. There, volcanic flows and faults bear witness to extrusive volcanism followed by extension during the fragmentation of the larger LIP into its provinces AP, Northeast Georgia Rise, and Maud Rise. On the MOZR, we only observe faults as a documentation of the separation of MOZR and Astrid Ridge. This may be an indication of modifications in magma supply.

The MOZR and AP are connected by a bathymetrically elevated corridor, Transkei Rise, which implies a continuous process of less extensive volcanic activity in the ~20 m.y. period between the main formation of the MOZR and that of the greater AP. A similarity can be observed between the timing, size and formation history of the Southeast African LIP and the Kerguelen-Heard Plateau, which provokes speculation about related mantle processes of periodic magma generation at that time.

References