Introduction
Deep beneath the Arabian Sea, the 1,100-km-long Owen fracture zone marks the boundary between the Indian and Arabian tectonic plates (Fig. 1). Both plates are colliding with the southern edge of Eurasia but the Arabian plate is generally considered to be moving northeastward slightly faster than the Indian plate, and it is this difference in motion that is accommodated by the Owen fracture zone. The rate of differential motion along this fracture zone has long been known as one of the slowest among major plate boundaries, but until recently this area has not been surveyed with modern geophysical equipment, leaving important gaps in our understanding of plate tectonics in this region. On page 54 of this issue, Fournier and co-authors1 describe a detailed survey of the southern end of the Owen fracture zone near the Aden–Owen–Carlsberg triple junction, including direct evidence for the direction and velocity of movement along the Owen fracture zone. The evidence suggests that a significant wedge of the Arabian plate was scavenged by the Indian plate some time between 
10 Myr ago and today.
Figure 1: A satellite-derived gravity anomaly map of the Owen fracture zone region in the Arabian Sea, as a proxy for seafloor bathymetry.
In general, blue areas (negative gravity anomalies) are deeper and red areas (positive gravity anomalies) are shallower. Earthquakes between 1963 and 2007 near the southern Owen fracture zone are indicated by red dots. Arrows show the directions of spreading of the Sheba and Carlsberg mid-ocean ridges and movement along the Owen fracture zone and Owen transform fault. The white box shows the extent of the study area, and the orange-shaded region shows the approximate extent of the wedge-shaped area of the Arabian plate referred to in the text.
Full size image (74 KB)The principal challenge in studying the Owen fracture zone is that, until now, there has been nearly a complete absence of relevant information about its geometry and structure. Along the fracture zone, the Indian plate appears to move southward with respect to the Arabian plate at a rate of about 3 mm per year as a result of the different rates of northward motions of the two plates, but this value is based on indirect estimates of motion, derived, for example, from small earthquakes along the fracture zone. As the Indian and Arabian plates move along the fracture zone, new oceanic crust is created by the seafloor spreading along the Sheba and Carlsberg ridges that terminate the southern end of the fracture zone. The differences in recent rates of seafloor spreading across these mid-ocean ridges2, 3 also provide an estimate for plate motion. However, these observations impose only broad constraints.
A key observation revealed by the undersea imaging conducted by Fournier and co-authors1 is that the underwater Owen Ridge, adjacent to the Owen fracture zone, is offset by 12 km to the south by a prominent vertical fault that had not previously been seen. The direction — or trend — of this fault, as determined by the researchers, is inconsistent with the trend of the Owen fracture zone suggested in previous models2, 3, but its existence confirms the southward motion of the Indian plate relative to the Arabian plate predicted by those models.
The direction of the relative motion of the Indian and Arabian plates predicted from velocities taken from a handful of scattered global positioning system stations agrees with the trend of the newly discovered plate boundary fault, and the available observations regarding regional geology and motions of the Earth's plates are consistent with 3
2 mm of motion per year between Arabia and India, parallel to the Owen fracture zone. The good agreement suggests that the relative motion between the Arabian and Indian plates is now well constrained.
Fournier and co-authors also show that the southern end of the newly imaged fault terminates in a deep basin that extends 50 km or farther to the west, into what was previously believed to be the Arabian plate interior. Seismic data presented by the researchers show that layers of young sediments within this basin are offset by extension-related faults, along which earthquakes occur, demonstrating that the basin is actively extending. Diffuse earthquakes in and to the west of the basin delineate a wedge-shaped area that was initially part of the Arabian plate, but has completely or partly transferred to the Indian plate.
Such an alternative configuration of present plate boundary faults in this region had recently been suggested4. Fournier and co-authors estimate that motion along the underwater fault they discovered started between 3 and 8 Myr ago, which would coincide with the transfer of the wedge-shaped region of the Arabian plate to the Indian plate. They suggest that this time marks the beginning of motion between the Indian and Arabian plates.
Relative to an observer in Eurasia, the differential motion of the Indian and Arabian plates is only 3 mm per year. This observation is rarely noticed but is remarkable, given the different characteristics of Indian and Arabian plate motion: India collides with Eurasia along the Himalayas, whereas the Arabian–Eurasian collision buckles up the Zagros Mountains. Furthermore, unlike Arabia, the Indian plate is connected to subducting oceanic crust, and different seafloor spreading centres — the Sheba and Carlsberg ridges, respectively — are located along the plates' southern boundaries.
Why the Indian and Arabian plates should move in such a similar way, given the seemingly different balances of forces that drive and resist their motion, is unknown and unstudied. Fournier and co-authors' mapping work is perhaps a first step toward unravelling the role of the Owen fracture zone in the answer to this question.
