Towards an improved representation of the relationship between root traits and nitrogen losses in process-based models

CONTEXT: Nitrogen (N) application to crops is crucial to feed an increasing world population. Yet, much of this N is not taken up by crops, initiating a cascade of N losses with dire environmental and economic consequences. There is, therefore, a need to develop crops with traits that make them use N more efficiently, thereby reducing N losses. Process-based models have been used to design in-silico crops with desirable traits to maximize yield and increase climate resiliency, but few have been used with the perspective of reducing N losses. OBJECTIVE: To examine the way process-based models capture interactions between root traits and N losses, and propose opportunities to improve model representation of observed relationships. METHODS: We synthesize the current knowledge on the relationships between plant traits and N losses based on experiments reported in the literature, conduct a survey of process-based models simulating crop growth and N losses, and run a sensitivity analysis with selected models (DSSAT, APSIM, DNDCvCAN, Daisy). RESULTS AND CONCLUSIONS: The results show that the relationships between root traits and N losses can be very strong in experiments, but model simulations do not capture the magnitude of these associations well. This is mainly due to the lack of a robust representation of the plant root mechanisms influencing N losses. Suggested model improvements include designing new functions to link root traits with key N-cycling processes supported by experimental evidence – such as root exudation of various compounds including biological nitrification inhibitors – and using easily observable morphological traits in process-based models as proxies to predict changes induced by plants on N-cycling by soil microbial communities. SIGNIFICANCE: This work represents a key step towards designing novel root function-based ideotypes adapted to reduced fertilizer inputs while maintaining the same level of yield, and that is, therefore, potentially less harmful to the environment.