Ocean vortex mixed non-locality prediction method based on random forest model

Abstract

The invention discloses an ocean vortex mixed non-locality prediction method based on a random forest model. In order to solve the problem that in an existing method, research on the nonlinear relation between Lagrange equilibrium time and the mixed non-local degree and between the vortex speed and the mixed non-local degree is insufficient, according to the method, by calculating the vortex diffusivity, the vortex diffusivity error value, the Lagrange equilibrium time, the mixed non-local ellipse and the vortex velocity at all the positions of the adaptive box based on the standard deviation method, the Lagrange equilibrium time and the vortex velocity are used as input samples, and the mixed non-local degree is used as an output sample, and a random forest model is constructed so as to obtain an ocean mixed non-locality degree prediction value. According to the method, the important degree of vortex mixing non-locality in ocean and climate simulation and forecasting can be correctly known, a new thought is provided for improving an existing vortex mixing parameterization scheme, and a direction is pointed out for further improving the accuracy of mode simulation and forecasting.

Description

technical field [0001] The invention relates to the field of non-local prediction of ocean eddy mixing, in particular to the prediction of the degree of non-locality of vortex mixing by using a machine learning method. Background technique [0002] Some subgrid processes in existing ocean climate models need to be parameterized. Among them, the parameterization of ocean eddy mixing is an important part of ocean climate model simulation. The vortex mixing parameterization assumes that the vortex flux can be represented by the diffusivity coefficient multiplied by the local tracer gradient, but the tracer flux is extended by the Green's function method to a form that clearly shows that at a given point The tracer flux on depends on both local and nonlocal large-scale tracer gradients [1], and the Lagrangian eddy diffusivity coefficient is also nonlocal in nature. However, the understanding and application of the non-locality of eddy mixing in existing climate models is not e...

AI Technical Summary

Problems solved by technology

There is a complex nonlinear relationship between the two characteristic parameters of Lagrangian equilibrium time and vortex velocity and the degree of mixing nonlocality, but the existing methods cannot describe this nonlinear relationship

Images