Evaluation of dry matter intake prediction equations in a dynamic model of energy balance in the dairy cow.

摘要

A simple dynamic model of net energy balance was constructed to select, out of some common intake prediction equations, the one that results in a minimum cumulative bias in bodyweight. All equations over-predicted bodyweight, with an increasing degree of over-prediction as lactation progressed. An under-prediction of maintenance energy expenditure calculation was identified, and least sum of squares analysis on 21 data sets resulted in a new maintenance factor of 0.096 Mcal/kg BW0.75. Average maintenance parameters plotted against week of lactation resulted in the equation: Maintenance=-0.0227(+/-0.0098)xWOL 2+1.352(+/-0.456)xWOL+78.09(+/-4.92) Mcal/kg BW0.75x10 -6. Re-evaluation of the model with the new descriptions of maintenance showed the NRC (2001) intake equation to result in the minimum cumulative bias in predicted bodyweight. A lipostatic feedback mechanism (kxWOL+k2)x [(energy stores-initial energy stores)/initial energy stores] was integrated into the NRC equation, and significantly improved both bodyweight and intake predictions by the model.