New Global Maps of Soil Radiocarbon and Carbon Mean Age Indicate Turnover is Too Rapid in Models
Researchers from the University of California Irvine, with collaborators from the Max Planck Institute for Biogeochemistry, Lawrence Berkeley National Laboratory, and the National Center for Atmospheric Research, created global maps of soil radiocarbon and carbon mean age and compared our estimates with two vertically resolved global land models. The comparisons show that the two models simulated younger soil carbon and more rapid carbon turnover.
Our findings suggest that soils will accumulate less carbon than predicted by current Earth system models over the 21st century. Our global maps of soil radiocarbon and carbon mean age can serve as critical, independent benchmarks that will improve predictions of soil carbon turnover and storage as climate changes.
Based on the relationships with environmental drivers in our random forest model, we scaled up ∆14C measurements from individual soil profiles to create global maps. Strong patterns of variation are evident across space and depth. The carbon-weighted ∆14C was −244 ± 48‰ globally, with values of −97 ± 24‰ in surface soil (0–30 cm) and −391 ± 56‰ in subsurface soil (30–100 cm). Globally, the carbon-weighted mean age of mineral soil carbon was 4830 ± 1730 years between 0 and 100 cm depth, with a younger mean age (1390 ± 310 years) in surface soils than subsurface soils (8280 ± 2820 years). Compared to the gridded dataset, the land surface models overestimated ∆14C in both surface and subsurface soil layers. In surface soils, over 60% of carbon in each of the models had positive ∆14C values compared to only about 14% of carbon in the global gridded dataset. The two models also predicted that about 50% of subsurface soil carbon had ∆14C more positive than −200‰, whereas this amount was less than 10% in the data-derived product.