Convergence in simulating global soil organic carbon by structurally different models after data assimilation.

Area of Science:

• Environmental science
• Earth system science
• Computational modeling

Background:

• Biogeochemical models exhibit significant uncertainty in carbon-climate feedback projections.
• Model structural differences in simulating soil organic carbon (SOC) dynamics are a primary cause of this uncertainty.
• Model parameterization also contributes to uncertainties in simulating soil carbon cycle processes.

Purpose of the Study:

• To demonstrate the critical role of observational data in reducing model uncertainty for global SOC storage estimates.
• To investigate how data assimilation impacts SOC simulations from structurally different models.
• To assess the influence of parameter values versus model structure on SOC dynamics.

Main Methods:

• Employed a data assimilation approach using a common global SOC database.
• Utilized two structurally distinct biogeochemical models with different carbon pools and decomposition kinetics.
• Compared model simulations before and after data assimilation.

Main Results:

• Two different models, initially predicting opposite global SOC distributions, converged to similar results after data assimilation.
• Data assimilation led to comparable simulations of key components like carbon transfer efficiency and decomposition rates.
• Both first-order and Michaelis-Menten kinetics models showed effective SOC simulation at the global scale post-assimilation.

Conclusions:

• Observational data are crucial for informing model development and constraining predictions of global SOC dynamics.
• Data assimilation effectively reduces model uncertainty, regardless of model structure or kinetic approach.
• Further high-quality data, including microbial genomics, are needed to refine SOC models and reduce parameter uncertainties.