APPLICATION OF REMOTE SENSING FOR THE YIELD AND QUALITATIVE IMPROVEMENT OF CEREAL CULTIVATION FOR ADVANCED SUPPLY CHAINS

This PhD thesis investigates the application of remote sensing (RS) and proximal sensing technologies for improving nitrogen (N) management in wheat and maize cultivation. The general objective is to optimize yield and grain quality while reducing environmental impact, supporting sustainable precision agriculture (PA) practices. Chapter 1 introduces the potential of RS in agriculture, explaining how vegetation indices (VIs) derived from both proximal and satellite sensors can support N management through crop status monitoring. Sentinel-2 imagery, due to its high temporal resolution and free access, is highlighted as a valuable tool for large-scale monitoring. Chapter 2 analyses spectral phenotyping in bread and durum wheat under different N rates, sources, and environments. NDVI measurements at various stages were linked to agronomic and NUE traits. Results showed that NDVI at booting predicted grain yield and N uptake effectively, while later stages correlated with grain protein content. Strong genotype-environment-management (GxExM) interactions emphasized the need for locally adapted varieties. Chapter 3 represents a central part of the thesis. A site-specific N management strategy was developed for winter wheat using canopy N status indicators (CNSIs) collected by RapidSCAN© proximal sensors. VIs (NDVI, NDRE) were analysed in relation to different N rates, allowing real-time assessment of crop N status and development of a predictive model to determine the optimal in-season N rate (N31) to meet yield targets. Validation confirmed the approach’s ability to optimize yield, with potential environmental benefits via reduced over-fertilization. The study highlights the importance of proximal sensing for dynamic N recommendations in wheat, although further work is needed to integrate grain quality aspects into the model. Chapter 4 compares Sentinel-2 spectral data with RapidSCAN© sensor data in winter wheat fields, showing that Sentinel-2 provides reliable estimates for monitoring large areas, though limited by spatial resolution in fragmented landscapes. Chapter 5 proposes a new method for comparing proximal and satellite data with different spatial resolutions. Using geostatistical tools, the method quantifies spatial variability and enables scaling of plot-level findings to larger field scales, facilitating operational use of RS in PA. Chapter 6 is another key contribution, addressing maize N management through simulation modelling, as VIs tend to saturate early in maize. Using the SALUS model, 43 years of maize growth simulations were conducted under varying weather, irrigation, and fertilization scenarios. The study classified years based on drought severity and determined the economically optimal N rates based on weather probabilities. Results showed that current farmer practices often exceed optimal N rates. The model identified that in wet years, 220 kg N ha⁻¹ maximized profitability, while lower doses were preferable in dry seasons, avoiding N losses and improving NUE. This modelling approach thus provides a tactical tool for pre-season fertilization planning in maize, compensating for RS limitations. Chapter 7 highlights that combining RS data with modelling can overcome respective limitations, supporting site-specific, sustainable, and profitable N management strategies across cereal cropping systems. In conclusion, this thesis demonstrates that integrating proximal sensing, satellite imagery, and crop modelling offers an effective and scalable framework for improving N management in cereals. Real-time proximal sensing enables dynamic N applications in wheat, while modelling supports adaptive maize fertilization under variable climate conditions. Together, these approaches enhance resource use efficiency and environmental sustainability in precision agriculture