The effects of dynamic soil moisture-dependent parameters on runoff and crop yield estimation for the APEX model

Authors: PALLARDY, QUINN; BAFFAUT, CLAIRE

Submitted to: Agricultural Water Management
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/12/2025
Publication Date: N/A
Citation: N/A

Interpretive Summary: Agricultural models can be used to simulate runoff and crop yields from farms. These models typically use parameters that determine how the simulated biological and physical processes function. For the Agricultural Policy Environmental Extender agricultural model, these parameters are static and do not change throughout the model simulation. This study assessed the validity of this assumption by evaluating a version of the model that allowed for dynamic parameters that changed based on soil moisture conditions. The results indicated that dynamic parameters resulted in improvements in model performance, and that optimal parameter values changed in dry and wet soil conditions. These results have implications for both accuracy and uncertainty in Agricultural Policy Environmental Extender model outcomes for current and future climate scenarios and may guide future research efforts; agricultural modelers should be aware of the limitations imposed by static parameters when evaluating the reliability of model outcomes.

Technical Abstract: The APEX, or Agricultural Policy/Environmental Extender Model, simulates many of the bio-physical processes that take place on a farm or small watershed. The parameters APEX uses to detail how these processes are simulated are typically determined through calibration and validation procedures based on a comparison of model output to observational data. However, traditional methods of parameterization assume a static set of optimal parameters for the full simulation. To investigate the validity of the assumption of stationarity in optimal parameter values, this study assessed whether the introduction of parameters allowed to vary based off soil moisture conditions could offer improved runoff and crop yield estimation for the APEX model. The null hypothesis examined was that optimal parameters under dry conditions were similar to those under wet conditions. Results indicated that dynamic parameters resulted in improvements in model performance for an objective function that combined mean absolute error performance of modeled output water yields and crop yields under calibration scenarios. There was also evidence of improvements in validation performance with less magnitude. The calibration process for several dynamic parameters indicated changes in optimal values based on soil moisture levels. These results have potential implications for both the accuracy and confidence in APEX outcomes under current scenarios and for APEX outcomes under changing environmental conditions.