Mineralogical Proxies Constrain Turnover and Warming Responses at Regional Scale

Modeling future soil organic carbon (SOC) dynamics is subject to significant uncertainty due to oversimplified representations of our mechanistic knowledge on C cycling across pedogenetically distinct soil types. To address this issue, we investigated how different calibration strategies affect modeled SOC stocks and turnover times across a pedo-climatic soil gradient in Chile. First, we performed a calibration for each site to obtain site-specific parameters and to derive empirical relationships between mineral associated organic carbon (MAOC)-related model parameters and soil mineralogical properties. Second, we performed a multi-site calibration testing (i) different site-selection strategies and (ii) the use of model parameters calculated based on the obtained relationships with mineralogical properties. Finally, we simulated an intermediate CMIP6-warming scenario (SSP2-4.5) to quantify relative changes in SOC stocks and how they related to the simulation using site-specific parameter sets. Our results show that considering soil heterogeneity through relating soil mineralogical properties to model parameters is a promising way to tackle the common oversimplification of soil landscapes in current modelling frameworks. Multi-site calibrations disregarding established empirical relationships failed to reproduce overall SOC stocks and MAOC turnover times regardless of calibration site selection. Capturing heterogeneity of MAOC turnover times was key to reflect the response of SOC to warming. We conclude that the relation to climatic and soil mineralogical properties and pedogenetic modification of MAOC turnover time, is crucial to improve the simulation of SOC stocks and its future responses to warming at larger scales.