Abstract
Purpose Advanced remote sensing and imagery technology help to estimate variability in grass ley plant coverage (PC). Adjusted manure and fertiliser application rates can be derived according to this variability, by means of machine learning and advanced image processing. This study aimed to determine the effects of variable nitrogen (N) rate from manure and synthetic fertiliser application on a grass ley field experiment in southwestern Norway, thereby generating N rate recommendations. The effects on dry matter yield, N use efficiency, and forage nutritive value were determined. Methods A field experiment was conducted in 2022-2023 and repeated in 2023-2024, estimating PC using digital processing of autumn and spring aerial images to determine fertiliser rates. Three fixed and two variable manure and mineral N rates were applied in early spring and after the first and second cuts. Forage dry matter yield (FDMY) and agronomic N use efficiecy (AgNUE) were evaluated over two seasons. Results A low or variable N rate based on spring coverage led to FDMY and AgNUE comparable to high N rates. Spring and autumn coverage during the second season improved slurry application decisions, offering a valuable tool for grassland management. The N rate-response model effectively represented the nonlinear behaviour of FDMY, revealing a strong concave response to N rates, significant seasonal variations, and a notable flattening of the response in 2024. Predicted curves indicated that the most beneficial N application occurs in earlier cuts, as late-season applications showed diminished yield leverage under 2024 conditions. Conclusion Image analysis can effectively support variable-rate fertiliser recommendations for perennial grasslands, although such approaches only improved N usage in one of two years. Whilst variable-rate application (VRA) is resilient during constrained regrowth years, interannual weather variability and seasonal conditions significantly influenced N responsiveness, indicating the necessity for calibrating cover-based models to enhance nutrient management efficiency under varying climate conditions.