Motion Between the Indian, African and Antarctic Plates in the Early Cenozoic

We used a three-plate, best-fit algorithm to calculate four sets of Euler rotations for India (Capricorn) - Africa (Somali), India (Capricorn)-Antarctic, and Africa (Somali)-Antarctic motion for twelve time intervals between Chrons 20 and 29 in the early Cenozoic. Each set of rotations had a different combination of data constraints. The first set of rotations used a basic set of magnetic anomaly picks on the Central Indian Ridge (CIR), Southeast Indian Ridge (SEIR) and Southwest Indian Ridge (SWIR), but did not incorporate data from the Carlsberg ridge and did not use fracture zones on the SWIR. The second set added fracture zone constraints from the region west of the Bain FZ on the SWIR and also included corrections for Nubia-Somalia and Lwandle-Somalia motion on the western and central SWIR, respectively. The third set of rotations used the basic constraints from the first rotation set and added data from the Carlsberg ridge. The fourth set of rotations combined both the additional SWIR constraints of the second data set and the Carlsberg ridge constraints of the third data set. Data on the Indian plate side of the Carlsberg ridge (Arabian Basin) were rotated to the Capricorn plate before being included in the constraints. We found that the rotations constrained by the Carlsberg ridge data set diverged from the other two sets of rotations prior to anomaly 22o. We concluded that, relative to the rest of the CIR, there is a progressively larger separation of anomalies on the Carlsberg ridge, starting at anomaly 22o and increasing to over 100 km for anomaly 26. These observations support two alternative interpretations. First, they are consistent with a distinct Seychelles microplate in the early Cenozoic. The sense of the misfit on the Carlsberg ridge is consistent with roughly 100 to 150 km of convergence across a boundary between the Seychelles microplate and Somali plate between Chrons 26 and 22 running from the Amirante Trench and extending north to the Carlsberg ridge axis. Alternatively, the misfit is consistent with convergent motion of the same magnitude between the Indian and a proto-Capricorn plate east of the CIR between Chrons 26 and 22. Our work also sharpens the dating of the two major Eocene events that Patriat and Achache (1984) recognized in the Indian Ocean: a large but gradual slowdown on the CIR and SEIR starting shortly after Chron 23o (51.9 Ma) and continuing until Chron 21y (45.3 Ma), a period of 6.6 Ma, followed two or three million years later by an abrupt change in spreading azimuth on the CIR and SEIR which occurred around Chron 20o (42.8) Ma and which was completed by Chron 20y (41.5 Ma). No change in spreading rate accompanied the change in spreading direction.