COUPLED ORBITAL AND THERMAL EVOLUTION OF GANYMEDE

We explore the hypothesis that passage through an eccentricity-pumping resonance could lead to the resurfacing of Ganymede. To do so, we cou ple R. Malhotra's (1991, Icarus 94, 399-412) orbital model for the tid al evolution of the Laplace resonance to an internal model of Ganymede . Our model explores the conditions under which Ganymede can undergo g lobal thermal runaway, assuming that the elk of Ganymede is strongly d ependent on internal temperature. (Here Q is the tidal dissipation fun ction and k is the second-degree Love number.) We allow the system to pass through the omega(1)/omega(2) approximate to 2 or omega(1)/omega( 2) approximate to 1/2 resonance, where omega(1) = 2n(2) - n(1), omega( 2) = 2n(3) - n(2), and n(1), n(2), and n(3) are the mean motions of Io , Europa, and Ganymede. If Ganymede's initial internal temperature is either ''too hot'' or ''too cold,'' no runaway occurs, while for inter mediate temperatures (similar to 200 K in the upper mantle), condition s are ''just right,'' and runaway occurs. The range of mantle temperat ures that allows runaway depends on the model for tidal Q; we use the Maxwell model, which ties Q to the creep viscosity of ice. Runaways ca n induce up to similar to 50-100 K warming and formation of a large in ternal ocean; they occur over a 10(7) to 10(8)-year period. Assuming c arbonaceous chondritic abundances of radionuclides in Ganymede's rocky portion, however, we find that the interior cannot cool to the initia l temperatures needed to allow large runaways. If our model is correct , large runaways cannot occur, although small runaways are still possi ble. Different formulations of tidal Q or convective cooling may allow large runaways. Large runaways are also possible if radionuclides are substantially depleted, although this is unlikely. We next consider t he consequences of a large runaway, assuming it can occur. Ganymede ca n undergo 0.5% thermal expansion (by volume) during the largest therma l runaways. Melting of the ice mantle provides up to 2% expansion desp ite the fact that contraction produced by melting ice I offsets expans ion produced by melting high-pressure ice phases. Solid-solid phase tr ansitions cause negligible satellite expansion. Lithospheric stresses caused by expansion of 2% over 10(7) to 10(8) years are similar to 10( 2) bars at the surface, and drop to a few bars at several kilometers d epth. Such stresses could cause cracking to depths of several kilomete rs. The cracking and near-surface production of warm or partially molt en ice make resurfacing a plausible outcome of a large thermal runaway . The tidal heating events proposed here may also be relevant for gene ration of Ganymede's modern-day magnetic field. (C) 1997 Academic Pres s.