Get reliable science

Prediction of feed intake in beef enterprises is important for feed budgeting, productivity,

profitability and environmental outcomes (e.g. carbon and methane accounting). The objective of

this study was to develop empirical prediction models for feed intake of growing beef cattle fed a

range of tropical forages. Data were gathered from previous metabolism and pen trials (76 treatment

diets) having live weight (LW), dry matter intake (DMI) and diet characteristics including in vivo

DM digestibility (IVDMD), nitrogen (N), neutral (NDF) and acid (ADF) detergent fibre. Prediction

equations of DMI were derived using mixed-effects linear regression models with LW and diet

characteristic as independent variables and location of trial as a random factor. The models were

later validated with an independent dataset from published literature related to tropical grazing

trials. Results indicated that DMI could be predicted with similar accuracy using LW and any one

measure of diet characteristic (R2 from 73 to 81%) with the highest R2 from the equation based on

LW and ADF. However, validation against an independent dataset from grazing trials indicated that

DMI was more accurately predicted from LW and IVDMD (R2 = 75%), LW and N (R2 = 71%), LW

and NDF (R2 = 61%), and least with LW and ADF (R2 = 24%). The lower accuracy of the models

to predict DMI from grazing trial may be due to diet selection and method used to measure it. The

present models may be used with results from faecal NIRS as input to predict DMI more accurately.