The persistence of root carbon in soil: data and modeling gaps

Belowground carbon inputs are five times more likely to be stabilized as soil organic matter compared with aboveground inputs. Yet our understanding and modeling of soil carbon are disproportionately focused on aboveground plant dynamics. Historically, this has been because of limited data on root traits and their relationship to soil properties. Furthermore, it is difficult to generalize site-scale linkages between root traits and soil carbon to broader spatial scales. Recent efforts in syntheses of root trait (TRY and FRED databases) and soil data (ISCN3 database, WoSIS, etc) make it possible to test broad-scale hypotheses regarding mechanistic links between root traits and soil carbon storage. We review the literature on root trait linkages with soil carbon to 1) generate hypotheses that can be tested with existing global data and 2) highlight data and model gaps. Our hypotheses and model recommendations focus on root traits related to the allocation of carbon to living roots, their lifespan and turnover, and what happens to roots after they die. Of these, improved allocation regimes (relative allocation to above- versus belowground parts, as well as the relative amount of absorbing fine roots compared with roots whose primary role is resource transport) present the most promising avenues for immediate model improvement, especially in large-scale Earth System Models. The development of more mechanistic linkages between root traits such as morphology and chemistry with the transfer of carbon to the many different pools of soil carbon are a next step for improved model structure. Lastly, we prioritize data needs across the continuum of root traits to soil properties that control organic matter persistence.