Pedotransfer functions and their impact on soil water dynamics simulation and yield prediction: A HERMES model case study

Pedotransfer functions used to determine water retention parameters from more readily available soil attributes (texture, soil organic carbon) may have a big impact on soil water dynamics and yield simulations by crop models. In this study we used the model HERMES to quantify and understand the impact of seven pedotransfer functions and two pedotransfer function ensemble models on crop yield on sites with variable texture using water retention parameters (i.e. field capacity, wilting point and total pore space) derived from the pedotransfer functions. Additionally, a continuous synthetic soil database was created to assess the effect of gradual changes in soil texture on yield simulation. The accuracy of the soil water dynamics simulations with the different pedotransfer functions was determined using soil water measurements at the test sites. The impact of pedotransfer functions on yield simulations was assessed by quantifying the yield differences at the experimental sites and across the synthetic soil database. The choice of pedotransfer functions resulted in most cases in measurable differences in soil water content dynamics, often showing a direct relationship between field capacity and water content. Often, pedotransfer functions producing higher estimated field capacity also resulted in higher yield, indicating that model yield simulation is sensitive to soil water availability. Pedotransfer functions showed an increasing variability in water retention parameters estimation and yield simulation at specific points in the sand percent continuum of the synthetic soil database, indicating that these functions can be sensitive even to small changes in particle size distribution.