Soil is a significant carbon sink, containing about twice as much carbon as the atmosphere and plants combined. Managing soil carbon levels is essential for mitigating climate change and maintaining soil health and agricultural productivity. However, measuring soil carbon is a challenging and expensive process that involves sending samples to a lab for analysis, making large-scale measurements difficult. To address this issue, environmental scientists have integrated field-level data with machine-learning techniques to estimate soil organic carbon at the U.S. continental scale, providing a more accurate benchmark for policymakers and land managers.
Debjani Sihi, the senior author of the study, emphasizes the importance of investing in sustainable land management practices to build up soil organic carbon levels. Land is more efficient than the ocean at retaining carbon, offering a nature-based solution to help mitigate climate change. By creating conditions that favor soil to capture carbon dioxide from the atmosphere, carbon can be locked in the soil for extended periods, contributing to long-term climate change mitigation efforts. Sihi’s research focuses on environmental and sustainability issues related to soil and climate, highlighting the potential of soil as a carbon sink.
Soil organic carbon is composed of plant and animal matter in various stages of decomposition and plays a crucial role in soil biology and quality. The U.S. Department of Agriculture’s National Cooperative Soil Survey database contains valuable soil information collected over decades, including soil organic carbon data from core samples. Soil organic carbon estimates within databases such as the Harmonized World Soil Database and SoilGrids can be inconsistent. Sihi’s team aimed to address these inconsistencies by developing more effective methods to scale up soil-sampling data, focusing on the United States.
The researchers divided the United States into 20 regions and created machine-learning models for each region using nearly 50,000 soil samples. By incorporating 36 environmental variables into their models, including climate, topography, soil properties, and vegetation, they generated more accurate estimates of soil organic carbon. The new method revealed variations in the effects of environmental variables across regions, shedding light on the factors influencing soil organic carbon levels. The approach provides a better benchmark for modeling Earth systems, offering insights into soil dynamics and responses to changing environmental conditions.
Future applications of this method to other countries and continents with sufficient data could enhance global soil organic carbon estimates and guide future sampling efforts. Considering the flexibility of the new model in adapting to changing climate conditions, such as warming soils and altered rainfall patterns, is essential for understanding how soil carbon dynamics may evolve. Accurate baseline data on soil organic carbon and its influencing factors are crucial for improving Earth system models and predicting the impact of climate change on soil carbon levels. Collaborative efforts among researchers and institutions are essential for advancing soil carbon research and addressing climate change challenges.
The study’s authors, including co-authors from various research institutions, stress the importance of accurate soil organic carbon estimates for enhancing climate change models and informing policymakers and land managers. By leveraging machine-learning techniques and environmental data, researchers can improve soil carbon estimates and examine the impact of environmental variables on soil organic carbon levels. Ultimately, investing in sustainable land management practices and enhancing our understanding of soil carbon dynamics are essential components of climate change mitigation efforts. The work was supported by the National Science Foundation and the U.S. Department of Energy, highlighting the collaborative nature of soil carbon research and its significance in addressing climate change.