Modelling the circumnuclear coma of comets: objectives, methods and recent results

We investigate whether the modelling of the immediate vicinity of an active nucleus-currently unobservable-can, as the modelling of the outer, observable coma, be based on unrealistic simple assumptions such as those of nucleus and dust grains sphericity. We point out the inconsistency of models based on such assumptions, which, to manage compatibility with the observations, have to introduce additional assumptions that conflict with the previous ones, such as the existence of active areas of the nucleus. We argue that, while the outer coma models being phenomenological in nature, can perhaps tolerate such inconsistencies, the circumnuclear coma models must be predictive, having to obviate the lack of observational data, and therefore must exclude implausible and ad hoc assumptions, and advocate only well-understood physical processes and duly validated modelling methods. We describe the first steps of development of a predictive circumnuclear coma model, and present a set of results obtained with parameters fitted to comet P/Wirtanen, the target of the 'Rosetta' mission, but of a quite general significance. Considering, first an inhomogeneous spherical nucleus with spherical dust grains, and then an aspherical homogeneous nucleus with spherical dust grains, we show that, in both cases (1) the surface temperature and initial gas parameters differ considerably from the Hertz-Knudsen values; (2) the near-surface gas and dust flows are not in general vertical, (3) the gas and dust density do not always monotonically decrease outwards, (4) the gas and dust velocity vary strongly from point to point, (5) shock structures are formed, which result in the formation of pseudo-jets of dust grains originating from various points of the surface. No simple method to distinguish between dust structures created by the surface inhomogeneity and by the surface orography is found. We show, for the first time, the deformation of the near-nucleus dust coma during a full rotation of an homogeneous, aspherical nucleus. We also show that identical active regions located at different points of an inhomogeneous spherical nucleus produce very different dust distributions, suggesting that the dust distribution is also strongly deformed during the rotation of such a nucleus. Finally, we consider, for the first time, a spherical homogeneous nucleus emitting aspherical dust grains. We show that, in such a case, the terminal grain velocity depends upon the shape, initial position, and even possibly upon the initial orientation of the grain at the surface, so that there cannot exist a precise relation between terminal velocity and dust grain mass. We conclude that, far from giving an approximate or average representation of the circumnuclear coma, the classical modelling approach yields in this region predictions that are in total conflict with the real behaviour of the gas and dust. As a most dramatic consequence, the use of this classical approach may have obscured completely the significance of the few direct and of the many indirect informations acquired hitherto on the nucleus activity. (C) 1999 Elsevier Science Ltd. All rights reserved.