Nudging

Nudging is a technique for relaxing a free-running simulation towards an external reference state. It can be targeted at a chosen range of horizontal scales and a chosen range of vertical levels. This allows the large-scale state of the simulation to be kept close to the reference state, while still permitting the model to generate its own small-scale variability. Two approaches are available for this nudging: selecting only certain horizontal scales (known as spectral nudging), or relaxing the model state directly towards the reference state at every horizontal scale (known as Newtonian relaxation).

This section describes the scientific formulation of nudging in GungHo. It begins with an overview of the method and the relaxation increment, before describing how the large scales can be selected using a convolution, how this may be computed efficiently on a coarse mesh, and finally how the reference data is prepared in the vertical. For details of how to configure and enable nudging, see the Nudging section of the User Guide.

• Overview
• Formulation
• Filtering by Convolution
• Multi-resolution Nudging
• Vertical Treatment