Constraints on spin state and Hapke parameters of asteroid 4769 Castalia using lightcurves and a radar-derived shape model

A 167-parameter, 3-D shape model of the Earth-crossing asteroid Castalia, obtained from inversion of delay-Doppler images (Hudson and Ostro, 1994, Science 263, 940-943) constrained the object's pole to lie on a cone of half angle 55 +/- 10 degrees centered on the radar line of sight (right ascension 0.3 hr, declination 25.4 degrees) at the time of observations (Aug. 22, 1989) but could not constrain the pole's azimuthal orientation or the sense of rotation. Here we fit lightcurves obtained at Table Mountain Observatory on Aug. 23-25 with lightcurves calculated from Castalia's shape and Hapke's photometric model (Hapke 1981, J. Geophys. Res. 86, 3039-3054; 1984, Icarus 59, 41-59; 1986, Icarus 67, 264-280). The fits strongly constrain Castalia's spin state to one of two possibilities, One has a north pole within similar to 13 degrees of lambda = 253 degrees, beta = 56 degrees and a sidereal spin period P = 4.089 +/- 0.001 hr. The other has a south pole within similar to 13 degrees of lambda = 242 degrees, beta = 7 degrees and a sidereal spin period P = 4.094 +/- 0.001 hr, The north-pole solution has global-average Hapke photometric parameters w = 0.38 +/- 0.07, g = -0.11 +/- 0.09, <(theta)over bar> = 46 degrees +/- 10 degrees while the south-pole solution has w = 0.24 +/- 0.07, g = -0.30 +/- 0.09, <(theta)over bar> = 25 degrees +/- 10 degrees. Due to the large solar phase angles (60 degrees-90 degrees) the fits are not sensitive to the opposition surge parameters h and B-0. At a more subtle level, the least-bifurcated shape (''lower-bound model'') allowed by the radar data (Hudson and Ostro, 1994, Science 263, 940-943) consistently gives statistically better fits than the more bifurcated possibilities, However, the differences are small enough that the possibility that they are due to unmodeled photometric effects cannot be discounted. The fits are improved by allowing the macroscopic roughness parameter <(theta)over bar> to be a function of surface position. Statistically, this process improves the north-pole solution more than the south-pole solution suggesting that the radar observations viewed a northern latitude, thereby indicating a resolution of the ''mirror ambiguity'' inherent in the single-date radar data set. (C) 1997 Academic Press.