GIS, Remote Sensing and Machine Learning: Data Integration to Support the Management of Coastal Island Ecosystems

Area of Science:

• Ecology
• Geospatial Science
• Climate Change Science

Background:

• Coastal islands are vital ecosystems but often poorly understood due to isolation.
• Assessing the ecological status of numerous islands presents significant logistical challenges.

Purpose of the Study:

• To evaluate the vulnerability of Nova Scotia's coastal islands using advanced geospatial techniques.
• To develop an integrated Ecosystem Stress Index (ESI) for region-wide assessment.

Main Methods:

• Utilized geographic information systems (GIS), remote sensing (RS), and machine learning (ML) for island analysis.
• Employed 1m resolution LiDAR for topographic classification and elevation assessment.
• Developed a random forest ML model incorporating environmental factors and tree mortality data.

Main Results:

• Approximately 70% of Nova Scotia's islands have an average elevation below 2m, indicating high flood vulnerability.
• Vegetation cover is strongly correlated with island topography, influencing habitat types.
• Significant sea surface temperature warming was observed, particularly in the Gulf of St. Lawrence, alongside pervasive marine traffic impacts.

Conclusions:

• Geospatial data and ML provide powerful tools for understanding and monitoring coastal island vulnerability.
• The developed Ecosystem Stress Index (ESI) offers a framework for prioritizing conservation efforts.
• Findings highlight the urgent need for adaptive management strategies for vulnerable island ecosystems.