Nondestructive quantification of soil mineral associated organic carbon (MAOC) species and their distribution at global scale

Several traditional chemical methods have been widely used in the evaluation of soil mineral associated organic carbon (MAOC). Yet, these methods may under- or over-estimate the different MAOC species due to incomplete extraction or weak selectivity. Here, we establish a nondestructive technique to assess different MAOC species using SEM-EDS combined with a mineral fingerprint algorithm. We find that Fe oxyhydroxides (Fe-OC), phyllosilicates (Si-OC), Fe oxyhydroxides - Ca (Fe-Ca-OC) and Ca associated OC (Ca-OC) contribute to MAOC storage by 44.6±29.3%, 24.2±19.9%, 17.2±16.5% and 14.0±18.8%, respectively, along a latitudinal gradient (18 - 47 N°, 3600 km) in China. Using the Random Forest model, we estimate the contribution of the four MAOC species to OC preservation in global soils at 1 m depth to be 333±25 Pg, 191±36 Pg, 207±23 Pg and 194±25 Pg C, respectively. Our predictions challenge prevailing ideas on the preservation and distribution of MAOC species in soils. A holistic conceptual framework is proposed to describe the distribution pattern of MAOC species. Contrary to current understanding, Fe-Ca-OC and Si-OC may play a critical role in the accumulation of OC in alkaline and acidic soils, respectively.