Tropical Forage NIRS Database: Tropical Dairy Cattle Forage Accuracy Detection Using NIRS Dry and Wet Calibrations

Balqis Ath-Thifa, Laila Sari, Despal , Adrizal , Dwitami Anzhany and Heri Sukria

Background and Objective: Daily fluctuation in the supply of nutrients from fresh forage offered by dairy farmers should be detected and adjusted to guarantee a consistent supply of nutrients for dairy cows. Currently, available dried forage detection using NIRS requires sample preparation. This study aimed to develop a wet forage NIRS database and compare its accuracy with a dry database. Materials and Methods: A total of 133 NIRS spectra were collected for fresh and dried forage, including napier grass, natural grass, rice straw, corn stover and corn husk. Chemical analysis was conducted using proximate and Van Soest methods to analyse dry matter (DM), ash, crude protein (CP), crude fiber (CF), neutral (NDF) and acid (ADF) detergent fibers. The chemical data were used to calibrate the spectrums to produce the NIRS prediction model. Results: The wet spectrums varied considerably compared to dry spectrums. Higher reflectance of the dry spectrum showed higher nutrient density in the dried forage. All nutrient contents can be detected accurately using dry or wet NIRS database (R²C>0.5 and RPD>1.5). However, a dried database is still more accurate (R²C>0.78) than a wet database (R²C>0.63). However, external validation of the dry database showed a significant difference in CP and ADF with the chemical analysis (t-test<0.05). Conclusion: The wet and dry NIRS databases can predict daily variation of fresh forage nutrient content. For higher accuracy, the prediction model for CP and ADF of the dry database is suggested to be improved.

View Fulltext Back