Determining Asteroid Spin States Using Radar Speckles

Michael W. Busch, Shrinivas R. Kulkarni, Walter Brisken, Steven J. Ostro, Lance A. M. Benner Jon D. Giorgini, and Michael C. Nolan.

Icarus 209, 535-541 (2010)


Knowing the shapes and spin states of near-Earth asteroids is essential to understanding their dynamical evolution because of the Yarkovsky and YORP effects. Delay-Doppler radar imaging is the most powerful ground-based technique for imaging near-Earth asteroids and can obtain spatial resolution of less than 10 m, but frequently produces ambiguous pole direction solutions. A radar echo from an asteroid consists of a pattern of speckles caused by the interference of reflections from different parts of the surface. It is possible to determine an asteroid's pole direction by tracking the motion of the radar speckle pattern. Speckle tracking can potentially measure the poles of at least several radar targets each year, rapidly increasing the available sample of NEA pole directions. We observed the near-Earth asteroid 2008 EV5 with the Arecibo planetary radar and the Very Long Baseline Array in December 2008. By tracking the speckles moving from the Pie Town to Los Alamos VLBA stations, we have shown that 2008 EV5 rotates retrograde. This is the first speckle detection of a near-Earth asteroid.

A companion website describing the radar observations and shape modeling of 2008 EV5 is available here.

Fig. 1. Schematic of a radar speckle pattern (not to scale). Top. The reflected light from each point on the target's surface forms a wavefront and these interfere constructively or destructively, producing the random pattern of bright and dark speckles at the Earth, intercepted by a pair of antennas with some baseline. As the asteroid rotates, the phase at each point on the surface changes, moving the speckles in the same direction as the surface. Bottom. A single bright speckle passing over two antennas with baseline B < LSpeckle and the echo power at each as a function of time. The average difference in arrival time over many speckles is equal to tlag. Figure based on that in Green (1968).

Fig. 2. 2008 EV5 radar echo power received by Green Bank as a function of time on 2008 December 23, with 0.025 s resolution, showing speckles moving over the station. Given a speckle scale of 900 km, the average speckle duration of 0.65 ± 0.05 s gives a speckle v elocity of 1350 ± 150 km/s, as expected from EV5's distance and rotation rate. The point-to-point variations in echo power are self-noise due to the small number of Fast Fourier Transforms in each measurement. The data were processed using the special purpose software correlator written for this work. Time resolution is 0.025 s; the echo arrived at the telescope starting at 08:25:01.30 UT.

Fig. 3. Cross-correlation of the echo power received at four closest pairs of VLBA stations as a function of relative time lag. Labels give stations and baseline lengths (PT = Pie Town, LA = Los Alamos, FD = Fort Davis, KP = Kitt Peak). tlag was sampled at 0.01 s intervals. The peaks in correlation amplitude at negative tlag indicate retrograde rotation, and are consistent with EV5's rotation period and the angles between the spin axis and the baselines. The KP-PT baseline was aligned almost at right angles to the speckle motion, so that tlag is indistinguishable from zero (see text). The secondary peaks are due to speckles moving into and out of correlation with each other, and have width approximately equal to the speckle duration (0.65 s). The uncertainties in our estimates of the position of the peaks include upper limits on systematic errors. For LA-FD the correlation peak is too weak for a fit to be meaningful.

Fig. 4. Principal axis views of our 2008 EV5 shape model. The model is viewed from six orthogonal directions, along its principal axes. Rotation is around the z-axis, with +z in the direction of the angular momentum vector. Yellow-shaded regions were seen only at incidence angles >45° or not seen at all.

Last update: 2011 October 06