Quantifying management-driven changes in organic matter turnover in an agricultural soil: An inverse modeling approach using historical data and a surrogate CENTURY-type model

In order to facilitate inverse modeling of time-series data and/or parameter estimation for process-based models of soil organic carbon (SOC) dynamics, we developed and applied a new surrogate CENTURY SOC model (SCSOC). While SCSOC has mass balance and decomposition kinetics equations for residue and soil organic matter (SOM) pools identical to those of CENTURY, it differs by being: (1) decoupled from models of plant growth, nutrient cycling, and hydrologic processes; (2) capable of employing daily, monthly, or annual timesteps; (3) solvable using widely available non-linear regression software. Here, we describe SCSOC and demonstrate that its numerical solution to the CENTURY SOC equations is highly accurate for both daily and monthly timesteps, provided the monthly physical forcing function inputs are appropriately derived from daily-resolution hydrologic model results.As an example of the proper application of the tool to obtain site-specific parameters and to investigate hypotheses using historical data, we analyzed SOC data from a single subplot of the Morrow Plots (Urbana, Illinois, U.S.A.) that has been cropped continuously in corn since 1876 using management practices that were common for east-central Illinois. The impact of different approaches typically used during calibration were examined by objectively estimating a variety of parameter sets that both govern SOM decomposition kinetics and define the initial mass and fractionation of the SOM. The results suggest that mean SOM turnover times were circa 5-fold longer during the period 1876–1954 relative to 1955–1995, when modern management practices were employed, and significantly different from default CENTURY predictions during both.Modified versions of SCSOC were also constructed in order to (1) show that uncertainty about historical soil erosion does not confound this interpretation of the data and (2) explore alternative assumptions about the structure of the model. One such alternate model showed that unexpectedly rapid decomposition of the corn-derived SOM combined with depletion of slowly-decomposing prairie-derived SOM could explain the post-1955 acceleration in decomposition.