Emory environmentals scientist combined field-level data with machine-learning techniques to estimate soil organic carbon at the U.S. scale, another step toward providing more accurate baseline data to improve Earth system models for climate change.