The use of infrared reflectance spectroscopy to predict the dry matter intake of lactating grazing dairy cows

Monitoring the feed efficiency of grazing dairy cows is currently restricted by lack of a quick, cost effective and accurate method to estimate dry matter intake (DMI). The aim of this study was to compare Near Infrared Reflectance Spectroscopy (NIRS) analysis of faeces and Mid Infrared Reflectance Spectroscopy (MIRS) analysis of milk to predict the DMI of lactating grazing dairy cows. Faecal samples were obtained from four different grazing experiments where DMI had been estimated using the n-alkane technique resulting in a dataset of 1,083 cows. Equations were developed using the following variables; 1) milk yield (MY), fat %, protein %, body weight (BW), stage of lactation (SOL) and parity, which was used as a benchmark to which the following could be compared; 2) MIRS wavelengths, 3) MIRS wavelengths, MY, fat %, protein %, BW, SOL and parity, 4) NIRS wavelengths, 5) NIRS wavelengths, MY, fat %, protein %, BW, SOL and parity. The benchmark equation was more accurate (coefficient of determination (R²⁾ = 0.61; root mean square error (RMSE) = 1.68 kg) than MIRS wavelengths alone (R² = 0.28; RMSE = 2.27 kg) or NIRS wavelengths alone (R² = 0.15; RMSE = 2.45 kg). The combination of the benchmark equation with MIRS wavelengths (R² = 0.64; RMSE = 1.58 kg) resulted in slightly more accurate predictions than the benchmark equation alone. The combination of the benchmark equation with NIRS wavelengths (R² = 0.59; RMSE = 1.48 kg) did not result in more accurate predictions than the benchmark equation.