High-Resolution Mapping of Soil Organic Carbon Stocks Using Machine and Deep Learning Approaches Across Mediterranean Land Uses

Summary

This 2025 study presents a computational framework for high-resolution digital soil mapping of organic carbon stocks across Mediterranean farming systems, comparing machine learning and deep learning approaches. The work addresses the challenge of scaling soil carbon assessment in regions characterised by complex topography and heterogeneous land management. As suggested by the title and journal scope, the research demonstrates that these algorithmic methods can generate actionable soil carbon maps at finer spatial resolution than conventional approaches, with implications for land managers and soil monitoring practitioners.

UK applicability

Whilst this study focuses on Mediterranean contexts, the machine learning and deep learning methodologies for digital soil mapping are transferable to UK farming systems. However, application would require recalibration with UK soil data, climate variables, and land use patterns; Mediterranean-specific environmental conditions may limit direct predictive transfer to temperate British contexts.

Key measures

Soil organic carbon stocks (spatial resolution and mapping accuracy); algorithmic performance metrics for machine learning and deep learning models across different Mediterranean land use types

Outcomes reported

The study developed and compared machine learning and deep learning algorithmic approaches for high-resolution digital soil mapping of organic carbon stocks across heterogeneous Mediterranean land uses. The work evaluated the performance and practical utility of these computational methods for scaling soil carbon assessment in complex topographical regions.

Topic tags