Microbial models with data-driven parameters predict stronger soil carbon responses to climate change.

Area of Science:

• Earth and Environmental Sciences
• Ecology
• Climate Science

Background:

• Long-term climate feedbacks depend on soil organic carbon (SOC) dynamics.
• Current models struggle to accurately simulate SOC pools and predict microbial responses to climate change.
• Conventional models implicitly represent microbes via decay rates, limiting predictive accuracy.

Purpose of the Study:

• To calibrate parameters for two soil microbial models using observational data.
• To evaluate improved model performance in predicting contemporary carbon stocks.
• To compare SOC responses to climate change between microbial and conventional models.

Main Methods:

• Data-constrained parameter estimation for two distinct soil microbial models.
• Assessment of model performance against global SOC databases.
• Simulation of 95-year climate change impacts using CMIP5 climate and carbon input data.

Main Results:

• Calibrated microbial models explained 51% of SOC variability, outperforming calibrated conventional models (41%).
• Microbial models projected significant soil carbon losses (8-11%) under climate change, contrasting with CMIP5 model variability (7% loss to 22.6% gain).
• Observed unrealistic SOC oscillations in a 2-pool microbial model, with less prominent, avoidable oscillations in a 4-pool model.

Conclusions:

• Explicitly modeling microbial biomass dynamics significantly improves SOC stock predictions and climate change response accuracy.
• Microbial models offer more robust projections of soil carbon's role in future climate feedbacks.
• A 4-pool microbial model demonstrates potential for realistic, stable SOC dynamics under climate change scenarios.