Authors:

William Messman

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

The purpose of this project is to provide a data-based approach for determining between multiple competing models for atmospheric blocking via the use of sparse regression methods such as PySINDy and Subsampling Sparse Bayesian Regression (SubTSBR) for identifying the governing equations from the data. This project examined the effectiveness of such methods in correct identification of governing equations through their application to partial differential equations relevant to the current models of atmospheric blocking such as the Burgers equation and the nonlinear Schrodinger equation, both with and without the addition of noise to the system. We then apply the regression methods to both simulation and real atmospheric data for discovery of the underlying governing equation. We have found both the SubTSBR and PySINDy approaches to be effective on identification the Burgers equation for noisy data and have determined that the PySINDy approach is effective on the identification of the nonlinear Schrodinger equation with noisy data, providing confidence in the method's ability to accurately identify a governing equation for atmospheric blocking from noisy real world data. Atmospheric blocking is thought to be involved in extreme weather events such as heat waves, and so a better understanding of the driving equations behind this phenomenon will help to better understand, predict, and plan for such events in the future.

Source:

Purdue University / 2023

Topics:

No topics listed

Co-authors:

William Messman