AbstractPastures in animal production are widely used, but production is conditioned to several factors. The nutritional composition of forage can be altered by soil conditions, season, plant maturity and morphology, so it is important to monitor its quality through chemical analysis. To optimize this type of analysis and speed up decision‐making by farmers and technicians, the use of near‐infrared spectroscopy (NIRS) is a tool that has been successfully applied. This research aimed to develop predictive models for chemical components of Megathyrsus maximus cv. Tanzania using NIR spectroscopy. Laboratory determinations of ash, crude protein (CP), in vitro dry matter digestibility (IVDMD), neutral detergent fiber (NDF) and acid detergent fiber (ADF) of 345 forage samples were used as reference data and correlated with their NIRS spectra. To calibrate the models, principal component analysis and partial least squares regression were applied. The results indicated that the prediction models of the studied parameters presented a coefficient of determination (R2) equal to or greater than 0.90; residual predictive deviation rate (RPD) greater than 3.0; error interval ratio (RER) greater than 12; close mean square error values between calibration and validation; and optimal number of latent variables (LV) between seven and eight for model calibration. For CP and IVDMD prediction, the regions with the highest simultaneous contribution were 1,414, 1996 and 2,384 nm; while for NDF and ADF, 1714, 1784, 1786, 2,160, 2,320 and 2,450 nm. The success in the development of predictive models by NIR spectroscopy to evaluate dry matter digestibility and other main chemical attributes of M. maximus cv. Tanzania shows that the quality of the models developed in this study enables them to be used alternatively in routine laboratory analysis in a quick, reliable and accurate way.
Read full abstract