Loading...
Thumbnail Image
Publication

Applied topological data analysis to El-NiƱo Southern Oscillation

Shirazbakhti, Ghazaleh
Citations
Altmetric:
Other Names
Location
Time Period
Original Date
Digitization Date
Issue Date
2021-07
Type
Thesis
Genre
Keywords
Subjects (LCSH)
Electronic dissertations
Electronic dissertations
Research Projects
Organizational Units
Journal Issue
Citation
Abstract
The El NiƱo Southern Oscillation (ENSO) is one of the most powerful climate phenomena that can change global air circulation, affecting temperature and rainfall around the planet. ENSO has three phases: El NiƱo and La NiƱa are two extreme phases of the ENSO cycle, and ENSO neutral is a transitional period between El NiƱo and La NiƱa. Topological data analysis (TDA) is an innovative approach which focuses on a data set's "shape" or topological structures such loops, holes, and voids. We used TDA to describe the homology groups of the two-dimensional function determined by sea surface temperatures of tropical Pacific Ocean. The persistent homology of the function, which is determined by the monthly mean sea surface temperature in an expressed area of the Pacific, was used to identify ENSO phases. We classify using TDA summaries and compare the results to the existing NOAA classification utilizing topology information of sea surface temperatures of correctly specified ENSO phases. We compute the morse filtration's persistent homology for each month, then utilize the ECC to summarize the persistent homology. To reduce arbitrary errors and make the curve continuous, we utilized the smoothing function. We used LDA algorithm as a classification method and for dimension reduction, we used PCA transformation.
Table of Contents
Description
Thesis (M.S.)-- Wichita State University, College of Liberal Arts and Sciences, Dept. of Mathematics, Statistics, and Physics
Publisher
Wichita State University
Journal
Book Title
Series
Digital Collection
Finding Aid URL
Use and Reproduction
Ā© Copyright 2021 by Ghazaleh Shirazbakhti All Rights Reserved
Archival Collection
PubMed ID
DOI
ISSN
EISSN
Embedded videos