Agroecosystem modeling and precision agriculture for sustainable nitrogen management

As a key nutrient for crop production, nitrogen (N) has long been used as fertilizer in agriculture, particularly after the advent of industrial N fixation in the first decade of the 20 th century. However, increasing N-based fertilizers in combination with low N use efficiency leads to environmental N losses. Reactive nitrogen (Nr) losses in the form of NH 3 , NO 3 - and N 2 O are a major threat to the environment and in need of mitigation. Due to the high spatial and temporal variability of Nr losses, quantifying the environmental impacts of agriculture is difficult. To reflect on methods used for quantifying and reducing N losses from agricultural systems that hold promise for large-scale application. This paper reviews precision agriculture (PA) practices and the representation of the N cycle in major crop system models (CSMs) to identify beneficial linkages between PA and CSMs to identify sustainable N management options. A combination of geographic information system (GIS), Remote Sensing (RS) and CSMs offers a promising approach to provide data at different spatial scale to predict environmental losses of N from agroecosystems. In this process, Artificial Intelligence (AI) can be used to process large datasets, improve and supplement the parameterization of CSMs, integrate sensor data, and support CSMs. Using a combination of GIS, RS and AI to empower CSMs to guide PA practices will allow sustainable N management to be implemented at scale. • 26 N models were reviewed and classified by submodule structure • Linkages between PA, CSMs, and GIS-based integration approaches were reflected • Challenges of spatializing and calibrating crop models limit their use in PA • RS and AI were identified as promising tools to enhance CSMs