Direct statistical simulation of the Lorenz96 system in model reduction approaches

Area of Science:

• Computational Physics
• Applied Mathematics
• Dynamical Systems Theory

Background:

• Direct statistical simulation (DSS) offers an alternative to traditional numerical methods for nonlinear dynamical systems.
• A key challenge for DSS is the curse of dimensionality, where statistical properties have higher dimensions than the system's dynamical variables.

Purpose of the Study:

• To investigate methods for reducing the dimensionality of direct statistical simulation.
• To assess the computational efficiency and accuracy of these dimensionality reduction techniques.

Main Methods:

• The study employed approximate closures at second and third order in equal-time cumulants for DSS.
• Numerical experiments were conducted using the Lorenz96 dynamical system to illustrate the proposed methods.

Main Results:

• Two distinct approaches were investigated for dimensionality reduction in DSS.
• Significant reductions in computational effort were achieved without compromising the accuracy of the simulations.

Conclusions:

• The developed dimensionality reduction techniques effectively mitigate the curse of dimensionality in DSS.
• These methods are applicable to complex systems, including turbulent fluid and magnetohydrodynamical systems.