Modeling and Predicting Storm Surges Using Machine Learning Methods Pubblico

Singhal, Arshia (Spring 2022)

Permanent URL: https://etd.library.emory.edu/concern/etds/1v53jz15v?locale=it
Published

Abstract

Machine learning methods offer significant potential to reduce computational cost in predictive modeling. Through a combination of unsupervised and supervised learning methods, key insights can be gleaned on past data and can be used to forecast future data. This is especially valuable in cases where the original model requires significant computational time and power, such as in the case of storm surge prediction. Although several such models exist to simulate and forecast storm surge, such as Sea, Land, and Overland Surges (SLOSH), Advanced Circulation (ADCIRC), and Delft3D, these models are often computationally expensive and time consuming due to the numerical techniques used to solve the associated partial differential equations. As a result, we aim to reduce this computational complexity by implementing k-means clustering, linear regression, decision tree, k-nearest neighbors, and artificial neural network machine learning techniques to predict storm surge based on storm characteristics.

Table of Contents

1 Introduction 1

1.1 Tropical Cyclones and Storm Surge......................................................... 1

1.2 Hydrodynamic Modeling........................................................................ 2

1.3 Background on Machine Learning Methods............................................. 6

1.4 Unsupervised Learning Methods............................................................. 7

1.5 Supervised Learning Methods................................................................. 8

1.6 Model Accuracy..................................................................................... 10

1.7 Applications to Storm Surge Modeling ................................................... 13

2 Methods 14

2.1 Data Curation....................................................................................... 14

2.2 Exploratory Data Analysis..................................................................... 16

2.3 Unsupervised Machine Learning Methods: K-Means Clustering................ 16

2.4 Supervised Machine Learning Methods................................................... 17

2.4.1 Linear Regression............................................................................... 17

2.4.2 Decision Tree..................................................................................... 18

2.4.3 K-Nearest Neighbors.......................................................................... 18

2.4.4 Artificial Neural Network.................................................................... 19

3 Results 20

3.1 K-Means Clustering............................................................................... 20

3.2 Linear Regression.................................................................................. 22

3.2.1 Using All Predictor Variables............................................................... 22

3.2.2 Using Subsets of Predictor Variables.................................................... 24

3.2.3 Separating Extreme Surges.................................................................. 26

3.3 Decision Tree........................................................................................ 28

3.4 K-Nearest Neighbors.............................................................................. 31

3.5 Artificial Neural Network....................................................................... 32

4 Discussion 35

5 Conclusion 39

Bibliography 42

About this Honors Thesis

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
School
Department
Degree
Submission
Language
  • English
Research Field
Parola chiave
Committee Chair / Thesis Advisor
Committee Members
Ultima modifica

Primary PDF

Supplemental Files