Simulating impacts of silage maize (Zea mays) in monoculture and undersown with annual grass (Lolium perenne L.) on the soil water balance in a sandy-humic soil in Northwest Germany

Abstract

This study was focused on modeling soil water, carbon (C), and nitrogen (N) dynamics in soil and crop emphasizing uncertainties in model parameterization and the evaluation of potential water stress for silage maize cultivations on a drained field. The CoupModel was applied on different management systems for silage maize (Zea mays) in monoculture and undersown with grass (Lolium perenne L.) on a sandy-humic soil. Four different fertilization levels with 0, 150kg of mineral-N, 40m3 of cattle slurry (72–148kgNha−1 year−1), and combined slurry/mineral-N (222–298kgNha−1 year−1) were simulated over five years. Results were based on most plausible parameter combinations regarding simulated biomass obtained from 10,000 runs by the Generalized Likelihood Uncertainty Estimation (GLUE) approach. The uncertainty in model parameterization was reduced significantly by limiting the number of simulations for each treatment sequentially resulting in quartile coefficients of variation (CV*)<25% for 26% and 36% of selected input parameters in bi-cropping and monoculture systems, respectively. Average soil temperatures in upper soil depths, the groundwater level, water potentials, and water contents between 10 and 80cm of depth were reproduced plausibly with the model as well as plant C and N contents. The CV* values of evapotranspiration and total runoff ranged between 0 and 26% and 8–21%, respectively, on half-yearly basis. Significant differences between the cropping systems were found, even though the soil water balance was positive for all systems, and the potential water stress was only minor in bi-cropping systems.