Can a mathematical model accurately predict intake of grazing
animals? Testing the Q-Graze model
S. J. R. WOODWARD, M. G. LAMBERT, A. J. LITHERLAND AND C. J.
BOOM
AgResearch Limited, Private Bag 3123, Hamilton
E-Mail:
Proceedings of the New Zealand Society of Animal Production 2001. 61:
4-7
A decision support model, Q-Graze, was developed to assist
farmers with grazing management decisions based on visual assessments of
pasture quality. To evaluate the model’s ability to predict dry matter
intake and diet composition, it was tested against pre- and post-grazing
herbage mass and composition data from a trial conducted at the
Whatawhata Research Centre involving bulls and steers grazing
mixed-species hill pasture. Q-Graze was able to predict apparent intakes
of grass leaf, grass stem, legume, weed, dead material and total dry
matter with means and residual standard deviations 6.4 ± 1.5, 0.6
± 1.3, 1.3 ± 0.7, 1.2 ± 0.8, 1.5 ± 1.4 and
10.6 ± 2.3 kg DM per animal per day, respectively. These
predictions accounted for R 2 = 83%, 15%, 76%, 73%, 54% and 55%,
respectively of the variation in the apparent intakes calculated from
the data. This indicates that Q-Graze is able to predict herbage intake
and diet selection of cattle grazing mixed hill-country pastures to a
high degree of correlation. The limiting factor was the measurement
error of pre- and post-grazing herbage mass, rather than the model
design. Model testing for additional animal and pasture types is
currently being undertaken.
Keywords: NZSAPAB;
diet selection; rotational grazing; model calibration;
metabolisable energy; energy demand; cattle.
Last Updated 7/08/01
