Improved prediction of the performance of cattle in the tropics

Summary

Two DSSs were chosen to address these objectives on the basis that they were quite different in their underlying approach, were well supported by research groups and were commercially available. There were GrazFeed, developed by CSIRO Canberra, and the Cornell Net Carbohydrate and Protein System (CNCPS) developed at Cornell University, NY State, USA. The basic equations from the Australian feeding standards (SCA), which are incorporated in GrazFeed, were also evaluated. We used an extensive data set from experiments with Bos indicus steers carried out in previous MLA-supported studies for our investigations.

The major findings were:

The established principles of energy use for growth by cattle do apply with tropical diets in the same way they do with temperate diets, as evidenced by the close linear relationship between (metabolisable) energy intake and cattle growth rate. This finding endorses the future application of the feeding standards to tropical grazing systems.
The CNCPS and SCA gave good predictions of liveweight gain (LWG) when the diet composition was described in detail and intake was known. GrazFeed, by contrast, did not predict LWG well.
The different DSSs use different equations and approaches for determining aspects of energy utilisation but the finding that LWG was well predicted by SCA and the CNCPS suggests that the equations underpinning these DSSs were robust and sound.
Improvements in growth rate predictions were achieved with the SCA by changing components of the equation predicting the maintenance requirements of animals, in particular the breed effect and effect of metabolisable energy intake. Several changes have been recommended.
GrazFeed uses essentially the same equations as SCA and based on the good LWG predictions with SCA, these equations are sound.
The under-prediction of LWG with GrazFeed was related to the fact that when this model predicts a deficiency of rumen degradable protein (RDP) in the diet, it reduces intake and thus the estimate of LWG. At times, GrazFeed used a lower intake than was actually measured.
Estimated microbial crude protein (MCP) production is the major contributor to metabolisable protein supply for cattle on tropical forage diets and, whilst CNCPS predicted MCP production well, GrazFeed under-predicted it.
Based on the good predictions of LWG from known intake, both the CNCPS and SCA equations could be used to back-calculate intake from a measured or historical estimate of LWG, given limited description of the diet (e.g., from NIRS) and of the animals.
A spreadsheet intake calculator, “QuikIntake”, was developed to calculate intake from basic inputs describing the animal and with known digestibility of the diet from NIRS. This spreadsheet is available now for producers to use, for instance to determine appropriate stocking rates.
Neither GrazFeed nor the CNCPS provided an accurate prediction of intake when only diet and animal descriptors were used due to the generally poor relationships between intake and diet composition parameters.
​Based on the poor predictions of intake, the models would not give good predictions of LWG when intake is not known.
In their present form, both GrazFeed and the CNCPS require complex inputs on the pasture or diet parameters in order to predict animal performance.​