Superparameterised

Superparameterised is a term used in scientific contexts, particularly in climate modeling and atmospheric research, to describe a computational approach where a complex physical process is represented by a much simpler, yet highly detailed, model. This simpler model, often referred to as a "superparameterization," is embedded within a larger-scale model. The key idea is to allow the sub-grid scale physics, which cannot be explicitly resolved by the main model's grid resolution, to be simulated with a high degree of fidelity by this specialized, embedded model.

The purpose of superparameterization is to improve the representation of phenomena that occur at scales smaller than the grid cells of the main model, such as clouds, convection, and turbulence. Instead of using simplified bulk parameterizations that approximate these processes, superparameterization employs a small, two-dimensional or three-dimensional model that runs at a much finer resolution within each grid column of the host model. This embedded model simulates the detailed physics of the sub-grid scale process, and its output is then used to inform the larger-scale model.

This technique offers a compromise between the computational expense of a very high-resolution global model and the potential inaccuracies of traditional parameterizations. By "parameterizing" the sub-grid processes using a high-fidelity model rather than empirical formulas, superparameterization aims to provide more realistic simulations of weather and climate phenomena. It has been particularly effective in improving the representation of precipitation and cloud cover in global climate models.