Microbe - driven turnover o sets mineral - mediated storage of soil carbon under elevated CO 2

The sensitivity of soil organic carbon (SOC) to changing environmental conditions represents a critical uncertainty in coupledcarboncycle–climatemodels1.Muchof thisuncertainty arises from our limited understanding of the extent to which root–microbe interactions induce SOC losses (through accelerated decomposition or ‘priming’2) or indirectly promote SOC gains (via ‘protection’ through interactions with mineral particles3,4). We developed a new SOC model to examine priming and protection responses to rising atmospheric CO2. The model captured disparate SOC responses at two temperate free-air CO2 enrichment (FACE) experiments. We show that stabilization of ‘new’ carbon in protected SOC pools may equal or exceed microbial priming of ‘old’ SOC in ecosystems with readily decomposable litter and high clay content (for example, Oak Ridge5). In contrast, carbon losses induced through priming dominate the net SOC response in ecosystems with more resistant litters and lower clay content (for example, Duke6). The SOC model was fully integrated into a global terrestrial carbon cycle model to run global simulationsofelevatedCO2 e ects.Althoughprotectedcarbon provides an important constraint on priming e ects, priming nonetheless reduced SOC storage in the majority of terrestrial areas, partially counterbalancing SOC gains from enhanced ecosystem productivity. Soils contain more carbon (C) than plant biomass and the atmosphere combined7. Although a large body of literature has explored the effects of elevated CO2 on plant growth8, there is considerable uncertainty as to how such changes will affect SOC stocks5,9. A large fraction of this uncertainty is due to the complexity of soil processes and structure: an enormous variety of chemical compounds and a diverse community of bacteria and fungi in the soil respond in complex ways to changes in temperature, moisture and inputs of fresh plant C (refs 2,10). Furthermore, recent advances in isotopic, genomic and spectroscopic tools have revealed that a suite of physical, chemical and biological factors control not only SOC decomposition, but also SOC formation and stabilization4,11. Current global-scale land surface models represent SOC decay as a first-order process that depends only on abiotic factors such as temperature and moisture1, with limited representations of root and microbial influences on SOC. Whereas microbial models have been applied at global scales12, rhizosphere processes and microbial influences on SOC stabilization and mineralization have not previously been integrated into global land surface models. Hence, the development of global-scale SOC models that represent essential processes and interactions while remaining tractable for parameterization in Earth system models (ESMs) remains a major challenge. There is now substantial evidence from both empirical andmodelling studies that inputs of simple, readily assimilatedC compounds such as glucose and amino acids (hereafter referred to as ‘simple C’) can accelerate the decomposition of complex organic compounds2. Such ‘priming effects’ are likely to have important consequences for global SOC stocks. Rising atmospheric CO2 concentrations generally increase the inputs of simple C to soils through greater leaf and root production13 and enhanced root exudation14. Such increases have been identified as responsible for accelerated losses of SOC in multiple CO2-enrichment experiments6,9,15,16, as well as in a broad synthesis of ecosystem responses to elevated CO2 (ref. 17). The importance of priming is further supported by modelling efforts, as ecosystem models based on first-order decomposition have been unable to explain observed changes in C and N cycling under elevated CO2 (refs 18,19), and land surface models that include coarse representations of priming have produced more accurate maps of global SOC stocks20. Although priming effects are critically important and globally significant, an important constraint on their impact is that plantderived inputs can also lead to the formation of SOC that is protected frommicrobial degradation. There is now increasing evidence that a significant proportion of stable SOC is derived from simple C rather than chemically resistant compounds3, and that the long-term preservation of SOC depends more on its accessibility to microbial decomposers than on its chemical complexity4,11. Such ‘protection’ of SOC results from physical occlusion inmicroaggregates or chemical sorption in organo-mineral complexes. Notably, protected SOCmay also have lower temperature sensitivity than chemically resistant SOC, meaning that its response to future climate change could be different even if its contemporary turnover rate is the same21. To investigate the global consequences of SOC priming and protection, we developed a new SOC model, Carbon, Organisms, Rhizosphere, and Protection in the Soil Environment (CORPSE), that represents protected and unprotected SOC pools, and uses a dynamicmicrobial biomass pool to control SOC transformation and decomposition (Fig. 1). Both protected and unprotected SOC pools contain a combination of compounds with different decomposition rates and parameters. This approach is more flexible than previous formulations that represented only bulk SOC and dissolved organic C, and did not include a separate protected C pool12. Furthermore, CORPSE is novel in that it links microbial turnover to protected