Using machine learning and remote sensing to track land use/land cover changes due to armed conflict

Area of Science:

• Environmental science
• Geoscience
• Conflict studies

Background:

• Armed conflicts significantly impact land systems, affecting food security and resource management.
• Understanding the complex dynamics between Land Use/Land Cover (LULC) and conflict is crucial for sustainable development in affected regions.
• The Orontes River Basin (ORB) in Syria has experienced conflict since 2013, with significant refugee influxes into Lebanon and Turkey.

Purpose of the Study:

• To develop and apply a novel methodology for analyzing spatio-temporal LULC changes in the ORB.
• To quantify LULC shifts in conflict-affected and refugee-hosting areas between 2004 and 2022.
• To address the challenge of limited ground truth data in geoscientific investigations.

Main Methods:

• Combined remote sensing products with machine learning techniques.
• Employed quasi-experimental statistical analysis for LULC change detection.
• Developed a flexible, cloud-based approach for analyzing annual LULC maps.

Main Results:

• Cropland decreased by 21-24% in Syrian conflict hotspots post-2013.
• Cropland area increased 3.4-fold in Lebanon.
• Identified distinct LULC patterns associated with refugee settlements in Lebanon and Syrian/Turkish border areas.

Conclusions:

• Spatio-temporal LULC changes are heterogeneous and significantly influenced by conflict and refugee dynamics.
• The developed methodology provides a robust framework for LULC research in conflict settings.
• Findings underscore the need for integrated approaches to resource management in conflict-affected regions.