NIRS for predicting Brown Swiss heifer diet composition on mixed pastures in the Amazon region

This study evaluated the use of near-infrared spectroscopy (NIRS) to predict the chemical composition of diets consumed by heifers grazing on mixed pastures. A total of 96 diet samples were collected from eight Brown Swiss heifers, dried at 60°C for 48 hours, and analysed for crude protein (CP), ash, neutral detergent fiber (NDF), acid detergent fiber (ADF), and in vitro dry matter digestibility (IVDMD). Samples were scanned using a Unity Scientific near-infrared spectrometer over the 1100–2500 nm wavelength range at 1 nm resolution. Prediction models were developed using partial least squares regression in UCAL software. Excellent calibration results were obtained for CP and NDF, with determination coefficients (\(\:{R}_{c}^{2}\)) of 0.99 and 0.94, respectively. Ash and ADF showed good predictive accuracy (\(\:{R}_{c}^{2}\) = 0.85 and 0.86), while IVDMD predictions were moderate (\(\:{R}_{c}^{2}\) = 0.74). These findings demonstrate that NIRS is a rapid, precise, and reliable tool for estimating key nutritional parameters in heifers’ mixed pasture diets, supporting its use for efficient forage quality monitoring.