Much of our work centres on modelling quantitatively the behaviour of mid-ocean ridges. In large areas of the world's oceans, ridges are typically stable and predictably mid-way between conjugate margins (see Animation H). Where three ridges meet at a triple junction, such as the Bouvet triple junction, instability occurs rather more often, leading to a complex tectonic history. We offer here a working model with a limited number of ridge jumps to explain the observed complexity in the ocean floor topography and the origin of the three submarine edifices - the Agulhas Plateau, the NE Georgia Bank and the Maud Rise - that were created at the triple junction during Cretaceous time. The region shown is defined by the Falkland-Agulhas Fracture Zone (FAFZ) to the NW and the Africa-Antarctica corridor (AAC) to the east. Eight changes in ridge function and position merit special mention:

1. At the end of Jurassic time (142.3 Ma), the long offset separating East and West Gondwana off SE Africa (highlighted in yellow) turned into a rapid spreading axis in the Weddell Sea, shown by extensive M-series marine magnetic anomalies mapped there. Further north, the offset was initially a dextral strike-slip fault between the Maurice Ewing Bank (MEB - dislodged from Africa by about 170 km along the Agulhas fault since 155 Ma) and Limpopia (still fixed to Antarctica). Both attachments progressively ended in the following 20 My. Further north, in southernmost Mozambique, rifts such as Xai-Xai indicate E-W extension at this time.

2. By 121.4 Ma (M0, start Aptian), MEB had moved SW along the Agulhas FZ by about 800 km and Limpopia (now the Mozambique Ridge) had rifted off Antarctica and become firmly attached to the Africa plate. The velocity at which (a) Antarctica was leaving Africa and (b) MEB (not yet attached to South America) was moving along the Agulhas fault were closely similar for the next 40 My. The mid-ocean ridges to these two systems were joined and increased in length to the NW and S respectively. Ocean growth between Antarctica, on the one hand, and South America and the MEB on the other slowed (vide Anomaly-T in the Weddell Sea) from this time. The MEB became firmly attached to South America from about 113 Ma.

3. At about 105 Ma (Albian) magmatism above the Bouvet triple junction produced the NE Gerogia Bank and the Maud Rise. The latter was subsequently carried away from the former on the Antarctic plate.

4. At about 98 Ma (Cenomanian) an eastward jump of the ridge separating the Africa and South America plates left the NE Georgia Bank firmly within the South America plate. The mid-ocean ridge between Africa and Antarctica continued separating the Maud Rise from its earlier conjugate, SW of Limpopia.

5. At 95 ± 3 Ma creation of the Agulhas plateau commenced at the margins of the South America and Africa plates, a large part of which was a 500 km leaky transform, perpetuating the ealier transform between Limpopia and MEB. Eruption of the Agulhas Plateau was completed by about 90 Ma. The outline of Iceland is superposed for scale.

6. About 86 Ma (Santonian) the South America-Africa ridge jumped again, this time to a more south-directed path, leaving the Agulhas plateau stranded on the Africa plate. As with previous jumps, the ridge position immediately south of the Agulhas FZ/FAFZ changed but little while its orientation further south (towards the triple junction) changed markedly. The new ridge orientation is preserved in the C34 marine magnetic anomalies off Africa and Antarctica that mark the end of the Cretaceous Quiet Zone (121.4-83.64 Ma). Note that the new ridge orientation leaves only a triangular area of SAM-AFR ocean created in the interval 98-86 Ma attached to the South America plate. The remainder became part of the Africa plate. Marine magnetic anomalies C34 and C33 are left on the eastern margin of this trinagular fragment.

7. About 79 Ma (Campanian) the mid-ocean ridge between Africa and South America switched to a slightly more southwesterly path, closer to the aforementioned triangular fragment. The new ridge orientation was active for at least 14 My, even penetrating slightly N of the FAFZ before meeting its demise. This is well supported by marine magnetic anomalies with these dates.

8. At about 65 Ma (end Maastrichtian) the active Africa-South America ridge made its biggest jump, to the west of the trinagular fragment. The long (1100 km) offset on the FAFZ was finally abandoned after more than 60 My. The mid-Atlantic ridge south of the FAFZ now lay only 250 km NE of the NE Georgia Bank with only a short transform offest across the FAFZ. Remarkably, the initial trace of the new ridge appears to replicate the form and location on the South America plate of the margin first adopted there at 98 Ma. This earlier alignment is now preserved on the Africa plate as the NE 'escarpment' of the Agulhas plateau. The tectonic complexity of ocean growth in the interval 99 to 65 Ma is preserved in about 2000 km of ocean crust on the Africa plate, east of where the triangular fragment sits today. Almost 2500 km of new ocean crust has been created between the triangular fragment and South America since the present alignment of the Mid-Atlantic Ridge was established at about 65 Ma.

Note in the animation how all these changes occur closely within the vicinity of the Bouvet plume (orange star).

Regional geophysical interpretation of an area has always been about honouring what is known of the geology and making credible predictions about what may exist where the geology is unknown. The ocean created by the Bouvet triple junction covers some 12 miilion km2, more than one third the area of Africa. There is precious little geological information to constrain an intepretation of its tectonic history. In practice, a conservative plate tectonic interpretation has proven exceedingly dificult to find, consistent with the record of ocean-floor topography and the rules of global tectonics. There is already enough data to paint the interpreter into a corner, as it were. The model presented here will certainly contain errors in detail but here is a working model that is worthy of consideration at a broad scale that may be refined later. Refinement may prove more productive than trying to start over and eliminating again all the untenable possibilities I have tried while developing what is shown here. Making the supporting evidence explicit is beyond the scope of this web page but I am always pleased to hear of anything I may have missed that could impact the solution.