To determine the current motion between the Australian, Antarctica, and African plates, and to test whether this plate circuit obeys the closure, all plate motion data available along the Southeast, Southwest, and Central Indian ridges are analyzed and reduced to 67 spreading rates, 38 transform fault azimuths, and 135 earthquake slip vectors. Carlsberg Ridge data were excluded because they record India-Africa not Australia-Africa motion. New data include 10 new transform azimuths along the Southwest Indian Ridge, many slip vectors, and a dense aeromagnetic survey along the Southeast Indian Ridge. All published and many unpublished magnetic profiles are modeled to determine rates consistently over a 3-m.y. time-averaging interval for three reasons: (1) magnetic profiles from the Southeast and Central Indian ridges suggest recent spreading rate changes, (2) some published rates differ for identical profiles, and (3) prior studies have not used identical criteria and time scales to determine spreading rates. The new rates differ from published rates by as much as 5 mm/yr. Transform fault azimuths are estimated from bathymetry, Seasat altimetric data, epicenter distributions, and offsets of magnetic lineations.

Azimuths of plate motion are also estimated from the horizontal projections of slip vectors from the centroid-moment tensor solutions and other focal mechanism studies. Our new Australia-Antarctica and Australia-Africa Euler vectors differ from all prior Euler vectors at the 95% confidence level. Along the Southeast and Central Indian ridges, our model gives rates differing by 4--7 mm/yr from those prior models. From a systematic analysis of the plate motion data, we find no evidence for a Nubia-Somalia-Antarctica triple junction along the Southwest Indian Ridge. We also find no evidence for a triple junction previously proposed to be at ~80¿E along the Southeast Indian Ridge; any deformation within this seismically active region of the Australian plate adds up to less than a few millimeters per year.

Along each boundary, data are fit well by a single Euler vector, except for a ~5¿ systematic misfit to azimuthal data along the eastern Southeast Indian Ridge, near a seismically active region of the Australian plate south of Tasmania. Although semismicity and the is azimuthal misfit suggest the Australian plate deforming measurably, possibly by distributed deformation or by the westward motion of a small microplate southeast of Tasmania, the significance of the misfit is marginal and is small enough that systematic errors may cause it. These alternative explanations could be distinguished by surveying these transform faults with modern seafloor mapping techniques. Prior studies found that enforcing plate circuit closure causes systematic misfits to rates along the Southeast Indian Ridge. Here we find that enforcing closure causes no systematic misfits to any data. Moreover, only significant nonclosure is found by an F-ratio test, which numerical experiment suggest could detect deformation exceeding ~2 mm/yr. We conclude that Indian Ocean plate circuit nonclosure and the deformation it suggests are much smaller than thought before. The absence of significant nonclosure argues against the usefulness of a model of deformation distributed throughout an Indo-Australian plate, but favors a model in which the significant deformation occurs in a diffuse plate boundary along the equatorial Indian Ocean between the Central Indian Ridge and the Sumatra Trench. ¿ American Geophysical Union 1988