An emerging consensus in the field of soil science is highlighting the critical role of microbe-mineral interactions in determining and understanding soil organic matter (SOM) cycling. This change has been reflected in the emergence of a new family of SOM models that explicitly simulate interactions among microbial SOM degradation and mineral SOM stabilization processes. While microbe-mineral models allow us to examine and predict SOM cycling processes in more mechanistic ways than previous models, uncertainties in key processes can drive strong contrasts in predictions and behaviors among models. We used one microbe-mineral model to investigate global patterns of physically protected and unprotected SOM driven by climate, litter properties, and soil properties. We then used a comparison of five SOM models incorporating different sets of assumptions along with a meta-analysis of warming and litter addition experiments to quantify divergence among different models and identify key mechanistic uncertainties that drive diverging model predictions.
Results/Conclusions
Global SOM model simulations showed strong climate-driven patterns, with mineral-associated SOM comprising a greater proportion of total SOM in tropical regions and particulate (unprotected) SOM comprising a higher fraction in temperate and boreal regions. Comparison to a meta-analysis showed that variations in litter quality and soil mineral properties also contributed to variations in protected fraction across sites. In the multi-model comparison, we found strong diverging patterns among model predictions of changes in SOM stocks and CO2 fluxes under both litter addition and warming manipulations. These differences were driven by key uncertainties in microbial growth kinetics and accessibility of mineral-associated SOM to decomposers. Divergence among models had the same range of variability as variation in observed treatment effects across the meta-analysis, preventing us from validating or eliminating any of the models included in the comparison. This highlights the need for more measurements of physico-chemically protected and unprotected SOM fractions in manipulative experiments to facilitate better evaluation of models.