A soil matrix capacity index to predict mineral-associated but not particulate organic carbon across a range of climate and soil pH

Understanding controls on soil organic carbon (SOC) will be crucial to managing soils for climate change mitigation and food security. Climate exerts an overarching influence on SOC, affecting both carbon (C) inputs to soil and soil physicochemical properties participating in C retention. To test our hypothesis that climate, C inputs, and soil properties would differently affect particulate organic carbon (POC) and mineral-associated organic carbon (MAOC), we sampled 16 agricultural sites (n?=?124 plots) in the United States, ranging in climate (mean annual precipitation (MAP)—potential evapotranspiration (PET; MAP-PET)), soil pH (5.8–7.9), and soil texture (silt?+?clay?=?13–96%). As MAP-PET increased, soils increased in oxalate-extractable iron (Fe_O) and aluminum (Al_O), decreased in exchangeable calcium (Ca_ex) and magnesium (Mg_ex), and received greater C inputs. Soil physicochemical properties did not strongly predict POC, confirming the relative independence of this SOC fraction from the soil matrix. In contrast, MAOC was well predicted by combining Al_O?+?[1/2]Fe_O with Ca_ex?+?Mg_ex in a ‘matrix capacity index’, which performed better than individual soil physicochemical properties across all pH levels (r?>?0.79). Structural equation modeling indicated a similar total effect of MAP-PET on MAOC and POC, which was mediated by total C inputs and the matrix capacity index for MAOC but not POC. Our results emphasize the need to separately conceptualize controls on MAOC and POC and justify the use of a unified soil matrix capacity index for predicting soil MAOC storage.