A large-scale framework for estimating soil carbon, nitrogen, pH, and salinity dynamics for 1985-2023

Area of Science:

• Earth and Environmental Sciences
• Soil Science
• Data Science

Background:

• Soils are critical for ecosystem services, climate regulation, and global food security.
• Human activities and climate change have led to significant global soil degradation.
• Effective adaptation strategies require robust global soil mapping and monitoring.

Purpose of the Study:

• To develop HUMERIS, a global dataset of key soil properties from 1985-2023.
• To create a framework for predicting long-term soil dynamics across scales, time, and depth.
• To analyze trends in soil salinity (ECe), pH, nitrogen (N), and organic carbon (OC).

Main Methods:

• Leveraging machine learning and cloud computing for global soil data analysis.
• Developing the HUMERIS dataset covering natural, ice-free land surfaces.
• Analyzing soil property changes over time, by biome, and land cover.

Main Results:

• Topsoil shows a yearly increase in N (+0.4%) and OC (+0.5%), and a decrease in ECe (-0.2%), with stable pH.
• Colder regions exhibit higher predicted OC and N increases compared to warmer regions.
• Land conversion from natural to cropland correlates with a -0.2% decrease in soil properties.

Conclusions:

• Global soil properties are shifting, with distinct trends observed across different biomes and land-use types.
• Findings have significant implications for agroecological modeling and sustainable land management.
• HUMERIS provides a valuable resource for understanding and addressing global soil dynamics.