A scoping review of side-dress nitrogen recommendation systems and their perspectives in precision agriculture

A scoping review of the relevant literature was carried out to identify existing N recommendation systems, their temporal and geographical diffusion, knowledge gaps. In total, 151 studies were identified categorised. Seventy-six percent systems are empirical based on spatialised vegetation indices (73% them); 21% mechanistic crop simulation models with limited use spatialized data (26% 3% machine learning techniques integration non-spatialised data. Recommendation appeared worldwide in 2000; they often applied exact location where calibration had been out. Thirty advanced techniques, such as sensor/approach fusion (44%), algorithm add-ons (30%), estimation environmental benefits (13%), multi-objective decisions (13%). However, some limitations have identified. For example, need specific calibrations for each site, species, sensor, rarely using soil, vegetation, weather together, while large input sets, non-spatialised. We conclude that can be improved by better algorithms. Highlights- main side-dress nitrogen recommendations systems.- Empirical most common but difficult generalize.- Mechanistic consider spatial variability.- Advanced solutions propose data/algorithm study outcomes.- Future research must maximize high-resolution monitoring