ABSTRACT Near‐infrared reflectance spectroscopy (NIRS) is an effective method for quantifying key plant constituents in dried and milled forage samples. However, the labour‐intensive nature of sample preparation for laboratory‐based NIRS limits its application for selection in routine forage breeding. The implementation of higher throughput analysis through in‐field NIRS measurements could enhance the selection process for improved quality attributes in perennial ryegrass. Perennial ryegrass samples were collected from breeding plots under evaluation in 2022 and 2023 and subjected to chemical analysis for dry matter digestibility (DMD) to form a calibration dataset. During plot harvest, samples were scanned using a Zeiss Corona Extreme at 2 nm intervals (950–1690 nm). Various mathematical spectral treatments were evaluated for each calibration, ranked according to the root mean squared error of prediction (RMSE) and the ratio of percent deviation (RPD). This resulted in the development of robust calibrations for DMD using harvester‐mounted NIRS on fresh samples. The most effective calibrations demonstrated moderate precision for DMD with R 2 = 0.76 and an RPD of 2.18. The utilisation of harvester‐mounted NIRS will facilitate the routine collection of forage digestibility data throughout the growing season, without additional labour requirements, thereby supporting the development of novel forage grasses with enhanced digestibility profiles.
Konkolewska et al. (Thu,) studied this question.