Hamid Amanlou; Hadi Shafiei; Tahereh Amirabadi Farahani; Farshid Kheiri; Najme Eslamian Farsuni
Volume 18, Issue 1 , April 2016, , Pages 27-38
Abstract
In order to compare the precision of predictions of NRC (2001) and CNCPS models, 17 Holstein fresh cows were assigned in a randomized complete block design to a balanced diet with NRC (2001) and CNCPS software. The predicted DMI, energy allowable milk and protein allowable milk by both models were compared ...
Read More
In order to compare the precision of predictions of NRC (2001) and CNCPS models, 17 Holstein fresh cows were assigned in a randomized complete block design to a balanced diet with NRC (2001) and CNCPS software. The predicted DMI, energy allowable milk and protein allowable milk by both models were compared with observed values. The precision of the predictions of models was evaluated by calculation of R2, mean bias and root mean square prediction error (RMSPE). Both models predicted DMI lower than the observed values (P<0.01). The mean bias, RMSPE and R2 were 2.22, 2.84 and 0.48 for NRC (2001) and 3.18, 3.99 and 0.52 for CNCPS, respectively and CNCPS predicted the DMI 20 percent lower than the observed value. The observed milk production was 29.78 kg/d. predicted energy allowable milk by NRC (2001) and CNCPS were 25.68 and 22.32 kg/d, respectively. Predicted protein allowable milk with NRC (2001) and CNCPS were 25.80 and 28.37 kg/d, respectively and NRC (2001) predicted the milk production 14 percent lower than observed value. Given the importance of DMI in fresh cows seems NRC (2001) has more precision compared to CNCPS for estimating nutritional requirements.