<p>To reduce nitrogen (N) leaching from agricultural sources, polluting marine and freshwater environments, evaluation of the effectiveness of management strategies to reduce these losses are needed. Agroecosystem models are useful tools for this, but calibrating and validating sub-modules is needed to ensure that model predictions are accurate and represents agricultural advances in e.g. crop genotypes. The objective of the current study was to calibrate and evaluate three cereal crop-modules and two catch-crop-modules in the agroecosystem model Daisy to ensure spatiotemporally robust and accurate simulations of crop- and soil-N dynamics, in order to compare N leaching from cereal-based crop rotations with and without catch-crops. The winter wheat, winter rye, spring barley, oilseed radish and ryegrass crop modules in Daisy were calibrated against data from two field experiments that ran from 2015 to 2022. The calibrated crop modules were evaluated against data from two independent field experiments (2015 to 2021). The crop module calibrations reduced model error across all measured variables for the calibration datasets and maintained or reduced model error for nearly all measured variables in the evaluation dataset, across N-fertilizer treatments. Mean annual N leaching was increased substantially in the calibrated crop rotations for all but two, demonstrating the importance of ensuring correct crop module calibration. Spring barley with a catch-crop had the lowest mean annual leaching compared to spring barley without a catch-crop and early- and normal sown winter wheat and winter rye. Spring barley with a catch-crop also resulted in the smallest decline in soil organic nitrogen.</p> Graphical abstract <p></p>