Which plant traits increase soil carbon sequestration? Empirical evidence from a long-term poplar genetic diversity trial

Abstract

Breeding or engineering crops to increase soil organic carbon (SOC) storage is a potential route to land-based carbon dioxide removal on working agricultural lands. However, due to limited observational datasets plus shifting paradigms of SOC stabilization, there is a lack of understanding of which plant traits to target for SOC enhancement or the ultimate sequestration potential of such measures. Existing long-term common gardens of genetically diverse plant populations may provide an opportunity to evaluate biological controls on SOC outcomes, separate from environmental or management variability. Here we report on soil carbon and root chemical data collected for 24 genotypes within a 13-year-old common garden in northwestern Oregon planted with over a thousand natural variants of Populus trichocarpa. Fractionating surface soil (0-15 cm) revealed substantial variation in stocks of mineral-associated organic matter (MAOM; 18-67 tonnes C/ha) and particulate organic matter (POM; 2-22 tonnes C/ha). Tree genotype explained 24 and 26% of the MAOM and POM stock variability, respectively, after controlling for background variability. We found minimal association between SOC concentration and either aboveground tree productivity or conventional metrics of root biomass recalcitrance (C/N ratios and lignin content). However, root elemental content appeared influential for MAOM-C, which showed a strong positive association with root aluminum (Al), and strong negative association with root boron (B) and magnesium (Mg). Furthermore, root concentrations of these elements were highly heritable (57-78%) and not simply a reflection of background variation in bulk soil elemental concentrations. We estimate that surface SOC stocks under these 24 genotypes have diverged at rates of up to 1.2-4.3 tonnes C/ha/yr. These results suggest that genetic diversity trials have value for elucidating biological controls on SOC dynamics, and that traits associated with root elemental content may be an important target for enhancing SOC.