Even Newer Dynamics for General atmospheric modelling of the environment (ENDGame)

ENDGame is an evolution of the current dynamical core, the New Dynamics, and is based on a semi-implicit semi-lagrangian discretization of the governing equations.

In common with the New DynamicsENDGame is a finite-difference model discretised on a latitude-longitude grid and is based on the fully compressible, nonhydrostatic Euler equations. However, distinct from the New DynamicsENDGame has been designed to allow the code to be switchable between various options: a nonhydrostatic and a hydrostatic formulation; deep-atmosphere and shallow-atmosphere formulations, and use of spheroidal, spherical or Cartesian co-ordinates (as appropriate).

Exact mass conservation can be critical for long-term climate simulations. In the New Dynamics this is achieved by the application of a flux form Eulerian scheme for dry density. The use of an Eulerian scheme within an otherwise semi-Lagrangian model is undesirable. Therefore, a novel mass-conserving, semi-Lagrangian transport scheme (SLICE) has been developed that will be available within ENDGame for transport of dry density and other scalar variables that need exact conservation.

An important aspect of ENDGame is that it is designed around an iterative approach to solving the semi-implicit aspects of the scheme. This permits more accurate coupling of the scheme to the physics parametrizations. It also produces a simpler form of the Helmholtz problem that arises from the semi-implicit discretization. Building on recent research into the dispersion properties of Rossby waves on C-grids ENDGame has improved handling of these meteorologically significant waves whilst retaining the same optimal properties. This has required a modification to the staggering of variables with respect to the poles. These two aspects, the different Helmholtz problem and the changes to the polar discretization, are expected to improve the scaling of the model on massively parallel computer architectures.

Key aims

  • To improve the robustness, accuracy and efficiency of the dynamical core of the UM.
  • To maintain at least the functionality of the current dynamical core.
  • To develop and maintain a suite of test environments: a shallow-water version; a Cartesian vertical slice version; and the full 3D dynamical core version on a sphere.

Last updated: 30 April 2014