Giving every soil carbon fraction its own birth certificate: radiocarbon toolkit reveals how soils lock carbon away

Ask a climate model how long carbon stays in a soil and it will answer confidently. Ask the soil, and until now there was no direct way to hear the reply. A study in The Innovation Geoscience introduces an experimental chain that makes different organic carbon fractions reveal their own ages — and in doing so, turns decades of theorized soil carbon mechanisms into observable patterns. Organic carbon entering soil faces two fates: microbes consume the labile, energy-rich molecules, while mineral particles can lock away what survives — or sometimes lock away carbon before microbes even touch it. Which route dominates at a given site has been almost impossible to observe directly, because bulk radiocarbon measurements average everything into one number. A team led by Hong Wang of Beijing Normal University and Peng Cheng of the Institute of Earth Environment, CAS — with academicians Zhisheng An and Weijian Zhou among the contributors — coupled oxygen-free (anoxic) pyrolysis with infrared gas analysis and Fourier-transform infrared spectroscopy, then dated each extracted fraction by accelerator mass spectrometry (AMS) 14C. The approach separates thermally labile, metabolically attractive carbon from the stable, stubborn remainder, and reads each one's radiocarbon age independently. The team applied the method across strikingly different archives: a black soil profile in northeast China; the Anthropocene sediments of Sihailongwan maar lake in Jilin; Yellow River delta riverine deposits; loess on the Chinese Loess Plateau; and rock varnish sequences on karst bedrock in the southwest. Data from US Critical Zone soils, karst cave flood deposits and a Japanese grassland profile were integrated for comparison. The fraction-specific age patterns spoke. Where labile carbon turned out as old as — or older than — the stable pool, carbon had been mineral-protected before microbes could fully process it: the protective preservation route models propose. Where labile fractions were younger and stable fractions older, the classic "decompose first, preserve what's left" selective preservation pathway was operating. And down soil profiles, fractions that share near-modern carbon signals at the surface diverge in age with depth — direct evidence that unprotected organic carbon can be gradually immobilized into the deep stable pool during soil formation. The authors note that the infrared and FTIR datasets underlying part of the interpretation are available only on request, and that the technique's reach — rivers, oceans, seismically disturbed landscapes — remains to be demonstrated. But the conceptual gain is immediate: preservation mechanisms that previously lived inside model assumptions can now be checked against dated, fraction-by-fraction field data. For a planet trying to budget its carbon, soils just became a more accountable accountant.