Dynamical Thin Disks

TL;DR

This paper develops a new 2.5-dimensional model for thin accretion disks that accurately incorporates vertical dynamics and turbulence, improving upon traditional vertically integrated equations in dynamical settings.

Contribution

It introduces a set of six evolution equations capturing vertical surface dynamics, enabling more accurate modeling of thin disks in dynamical environments.

Findings

Recovered missing viscous terms involving vertical variables.

Proposed a resummation method for the gravitational force similar to softening models.

Enabled efficient study of vertical oscillations in thin disks.

Abstract

Thin disk accretion is often modeled in highly dynamical settings using the two-dimensional equations of viscous hydrodynamics, with viscosity representing unresolved turbulence. These equations are supposed to arise after vertical integration of the full three-dimensional equations of hydrodynamics, under the assumption of a geometrically thin disk with mirror symmetry about the midplane. But in the dynamical context, vertical dynamics are neglected by incorrectly assuming instantaneous vertical hydrostatic equilibrium. The resulting errors in the local disk height couple to the horizontal dynamics through some viscosity prescriptions and gravitational softening models. Furthermore, the viscous terms in the horizontal equations are only complete if they are inserted after vertical integration, as if the system is actually two-dimensional. Since turbulence breaks mirror symmetry, it is more physically correct to insert a turbulence model at the three-dimensional level, and impose mirror symmetry only on average. Thus, some viscous terms are usually missing. We revisit the vertical integration procedure, restricting to the regime of a Newtonian, non-self-gravitating disk. We obtain six evolution equations with only horizontal dependence, which determine the local vertical position and velocity of the disk surface, in addition to the usual fluid variables. This "2.5-dimensional" formulation opens the door to efficiently study vertical oscillations of thin disks in dynamical settings, and to improve the treatment of unresolved turbulence. As a demonstration, by including viscous stress at the three-dimensional level, we recover missing viscous terms which involve the vertical variables. We also propose a resummation of the vertically integrated gravitational force, which has a strikingly similar form to a gravitational softening model advocated for in protoplanetary disk studies.