Exploring the drivers of reef island shoreline change using machine learning models

Area of Science:

• Geoscience
• Climate Science
• Coastal Geomorphology

Background:

• Reef island shoreline change is highly variable across different time and space scales.
• Attributing the specific processes driving observed island changes remains a challenge.

Purpose of the Study:

• To develop machine-learning models for identifying drivers of reef island shoreline and positional change.
• To analyze the interactions between various environmental and morphological predictors of island change.

Main Methods:

• Utilized multi-decadal shoreline and island footprint change records from the western-central Pacific.
• Developed and applied machine-learning models to identify significant predictors of island change.

Main Results:

• Identified key predictors including oceanographic, climatic, and local island/reef morphological properties.
• Demonstrated that local factors like broader reef platforms and high vegetation density can mitigate erosion from sea-level rise.
• Highlighted complex interactions between multiple drivers influencing reef island dynamics.

Conclusions:

• Machine learning provides a novel approach to understanding reef island physical changes.
• Local-scale variabilities significantly influence island responses to sea-level rise, requiring nuanced adaptation strategies.
• Findings are crucial for attribution studies, developing small island vulnerability indices, and projecting future island changes under intensifying climate change.