Modelling nitrogen leaching: Are we getting the right answer for the right reason?

•Different approaches to simulate N leaching in crop models have been reviewed.•Impact of impeded leaching factor assessed for APSIM, DSSAT, SWB-Sci.•Current ability to simulate N leaching is not developed enough for accurate predictive modelling.•Combining measurement and modelling identified as important in future research. The complexities and challenges in quantifying N leaching have led to development of a range of measurement and modelling techniques, but none are widely applied. Observations that N moves more slowly than water through the soil profile has resulted in different approaches being used to simulate impeded N movement in crop models: (i) by accounting for nitrate NO3− adsorption to the soil, (ii) by considering incomplete mixing between resident and draining soil water fractions or (iii) a combination of both. We compare and discuss strengths and weaknesses of these approaches. Our inability to directly measure model parameters (especially with regards to simulating N dynamics), and the risk of compensating errors during model testing and calibration, often results in low confidence in simulated N leaching. We caution that our current ability to simulate N leaching is in most cases not yet well enough developed for reliable and accurate predictions. We recommend a more strategic approach involving better linking measurement and modelling to improve understanding of the critical soil processes that control N leaching as one way of further improving our understanding and quantification of N leaching.