Prediction of Chemical Parameters in Maize Silage by near Infrared Reflectance Spectroscopy

Abstract

A total of 125 maize silage samples were used to evaluate the ability of near infrared (NIR) reflectance spectroscopy to predict chemical compositions. NIR calibrations were developed by means of partial least-square (PLS) regression. Results showed that NIR analysis of dried samples of maize silage could provide accurate predictions of dry matter (DM), crude protein (CP), neutral detergent fibre (NDF), acid detergent fibre (ADF), hemicellulose, ash, pH, lactic acid and butyric acid content with validation correlation coefficient of determination (r² _v ) and standard deviation/root mean square error of prediction (SD/RMSEP) of 0.88 (2.98), 0.89 (3.10), 0.86 (2.81), 0.87 (2.36), 0.81 (2.37), 0.80 (2.53), 0.93 (3.85), 0.81 (2.20) and 0.64 (2.26), respectively in g kg⁻¹ on a dry weight basis. The NIR technique also could be used to predict (with r² _v and SD/RMSEP) the DM, 0.90 (3.54), CP, 0.82 (2.36), NDF, 0.82 (2.34), ash, 0.70 (2.10), pH, 0.79 (2.21) and lactic acid, 0.62 (2.11) of fresh samples of maize silage in g kg⁻¹ on a dry weight basis.