D. G. Fox

A decision support system to improve individual cattle management. 1. A mechanistic, dynamic model for animal growth

Abstract A deterministic and mechanistic growth model was developed to dynamically predict growth rate, accumulated weight, days required to reach target body composition, carcass weight (CW) and composition of individual beef cattle for use in individual cattle management systems. The model can predict either average daily gain (ADG) when dry matter intake (DMI) is known or dry matter required (DMR) when ADG is known. For both scenarios, the following parameters are required: metabolizable energy of the diet and length of feeding period, animal characteristics [age, gender, breed, initial body weight (BW), body condition score, and adjusted final BW at 28% empty body fat (EBF)] and environmental information (temperature, humidity, hours of sunlight, wind speed, mud, hair depth, and hair coat). Two iterative methods based on gain composition were derived to compute the efficiency of metabolizable energy to net energy for growth (NE g ). This growth model was evaluated with data from 362 individually fed steers with measured body composition and feed energy values predicted with the NRC (2000). The iterative method that used a decay equation to adjust NE g based on the proportion of retained energy as protein showed the best prediction of ADG and final BW. When dry matter intake was known, the model accounted for 89% of the variation with bias of −2.6% in predicting individual animal ADG and explained 83% of the variation with bias of −1% in estimating the observed BW at the actual total days on feed. When ADG was known, the growth model predicted the dry matter required for that ADG with only 2% of bias and r 2 of 74%. A sub-model was developed to predict accumulated body fat (FAT) for use in predicting carcass quality and yield grades (YG) during growth. With the unadjusted NE g method, this sub-model explained 84% of the variation and had −14.3% of bias in actual body fat when animal ADG was known. Additionally, an equation developed with 407 animals to predict YG from EBF (%) had an r 2 of 0.49. Equations developed to predict CW from empty BW that adjust for stage of growth accounted for 89% of the variation with a 3 kg of bias. In conclusion, this dynamic growth model can predict animal performance and body composition within an acceptable degree of accuracy.

The development and evaluation of the Small Ruminant Nutrition System

A mechanistic model that predicts nutrient requirements and biological values of feeds for sheep and goats (Small Ruminant Nutrition System, SRNS) was developed based on the Cornell Net Carbohydrate and Protein System for sheep. The SRNS uses animal and environmental factors to predict metabolisable energy (ME) and protein requirements. This model has been subjected to an extensive evaluation. In particular, evaluation of the SRNS for sheep using published papers indicated good accuracy and precision in the prediction of organic matter and CP digestibility, while NDF digestibility was underpredicted. In addition, the SRNS accurately predicted daily ME intake (mean bias (MB) = 0.04 Mcal/d; root mean square error of prediction (RMSEP) = 0.24 Mcal/d; r2 = 0.99) of lactating goats and goat wethers. The SRNS also accurately predicted the ADG of lambs (n = 42; MB = 1 g/d; RMSEP = 37 g/d; r2= 0.84) and kids (n = 31; MB = −6.4 g/d; RMSEP = 32.5 g/d; r2= 0.85) and the gains and losses of shrunk body weight of adult sheep (MB = −5.8 g/d; RMSEP = 30 g/d; r2= 0.73) and the energy balance (n = 21; RMSEP = 0.20 Mcal/d; r2 = 0.87) of lactating goats and wethers. In conclusion, based on our accumulated evaluation of the SRNS with literature data, the SRNS accurately predicts nutrient supply and requirements of sheep and goats. Recent unpublished evaluations, however, suggested that the SRNS may underpredict ADG when compensatory growth occurs.