Three-way debate – measurements, pedotransfer functions, and inverse modeling – input strategy for soil hydraulic parameters in wheat simulations

Abstract

Accurate characterization of soil hydraulic parameters is crucial for simulating soil-plant-water dynamics in agroecosystem models. Hydraulic parameters, specifically the field capacity and permanent wilting point, govern the soil water balance and define the limits of plant-available water. This study evaluated the impact of soil hydraulic parameters in the Decision Support System for Agrotechnology Transfer (DSSAT) by comparing simulated soil water dynamics, wheat growth, and yield with observed data. To this end, a detailed 3-yr field dataset (2022, 2023, and 2025) from South Germany was employed, distinguished by its integration of wheat growth with an extensive daily record of soil water content and matric potential measurements at multiple depths. Within the DSSAT shell, permanent wilting point and field capacity were derived using both the Saxton and Rawls pedotransfer functions (PTFs); these estimates were then evaluated against lab-measured and sensor data using retention curves and soil-water simulations. In particular, the Saxton PTF introduced higher uncertainty in soil-water simulations and was selected as the starting point for an inverse modeling approach using Bayesian and Grid search algorithms to optimize field capacity and permanent wilting point. Using soil water content, wheat growth, and biomass data as target variables in the optimization procedure significantly reduced the normalized Root Mean Square Error (nRMSE) for both optimization algorithms. Overall model performance improved significantly, as shown by a reduction in nRMSE from 20% of the Saxton PTF to ≤ 10% for Bayesian targeting of grain weight and Grid search targeting of above-ground biomass, thereby enhancing the accuracy of wheat growth and yield predictions. This study provides a realistic evaluation of how measured, estimated, and calibrated parameters affect the soil water balance, demonstrating that plant response data can serve as an effective target variable for optimizing field capacity and permanent wilting point.