The dynamical role of the circumplanetary disc in planetary migration

¹ Institut für Astronomie & Astrophysik, Abt. Computational Physics, Universität Tübingen, Auf der Morgenstelle 10, 72076 Tübingen, Germany e-mail: crida@tat.physik.uni-tuebingen.de
² Laboratoire AIM-UMR 7158, CEA/CNRS/Université Paris Diderot, IRFU/Service d'Astrophysique, CEA/Saclay, 91191 Gif-sur-Yvette Cedex, France
³ ICF-UNAM, Av. Universidad s/n, Cuernavaca, Morelos, CP 62210, Mexico

Received: 31 December 2008
Accepted: 21 April 2009

Abstract

Context. Numerical simulations of planets embedded in protoplanetary gaseous discs are a precious tool for studying the planetary migration; however, some approximations have to be made. Most often, the selfgravity of the gas is neglected. In that case, it is not clear in the literature how the material inside the Roche lobe of the planet should be taken into account.

Aims. Here, we want to address this issue by studying the influence of various methods so far used by different authors on the migration rate.

Methods. We performed high-resolution numerical simulations of giant planets embedded in discs. We compared the migration rates with and without gas selfgravity, testing various ways of taking the circum-planetary disc (CPD) into account.

Results. Different methods lead to significantly different migration rates. Adding the mass of the CPD to the perturbing mass of the planet accelerates the migration. Excluding a part of the Hill sphere is a very touchy parameter that may lead to an artificial suppression of the type III, runaway migration. In fact, the CPD is smaller than the Hill sphere. We recommend excluding no more than a 0.6 Hill radius and using a smooth filter. Alternatively, the CPD can be given the acceleration felt by the planet from the rest of the protoplanetary disc.

Conclusions. The gas inside the Roche lobe of the planet should be very carefully taken into account in numerical simulations without any selfgravity of the gas. The entire Hill sphere should not be excluded. The method used should be explicitly given. However, no method is equivalent to computing the full selfgravity of the gas.