Modelos matemáticos para tomada de decisão na produção de ovinos em ambiente pastoril

Abstract

The objective of this thesis was to develop and evaluate mathematical models to assist in decision making in the production of sheep under grazing. In Chapter 1, a modeling study was developed to test the hypothesis that the current published equations to predict the dry matter intake (DMI) of beef sheep are not applicable to animals raised on pastures in a tropical climate. The objective was to evaluate whether the DMI prediction equations for beef sheep are valid for sheep raised exclusively on tropical pastures, and to propose a new equation to predict the DMI of sheep kept on tropical pastures. The data used came from an experiment conducted with Santa Inês sheep kept in Massai grass pastures. It was observed that the null hypothesis was rejected, that is, the equations in the literature generated different predictions (β0 ≠ 0 and β1 ≠ 1) for the CMS observed in practical conditions of feeding sheep on grazing. Thus, the following equation was proposed: DMI (% PV) = 7.16545 (±0.76522) - 0.21799 (±0.01812) × LW + 0.00273 (±0.00034) × LW2 - 0 .00688 (±0.00299) × GT + 0.000007 (±0.000002) × GT2 + 0.00271 (±0.00108) × FFA, where LW is live weight (kg); GT is the grazing time (min/day) and FFA is fresh forage allowance (kg DM/100kg BW). In chapter 2, the hypothesis that biometric measurements can be used to predict the live weight of crossbred lambs kept in tropical pastures was tested. The objective was to propose a mathematical model to predict the body weight of lambs kept in tropical pastures based on biometric measurements. For this study, data from lambs with a genetic composition of at least 50% of the Santa Inês breed were used. The measurements of withers height (WH), rump height (RH), body length (BL), chest width (CW), rump width (RW), heart girth (HG) and abdominal circumference (AC) were used as input variables in the model. The model to predict the BW of lambs was: Weight (kg) = 0.4455 × HG - 0.5794 × AC + 0.0019 × RH2 + 0.0053 × AC2. In chapter 3 was aimed at predicting the carcass traits of Santa Inês lambs finished on tropical pastures through biometric measurements. The measurements of WH, CW, HG, RW, BL, leg circumference, leg length and slaughter weight explained most of the variation in carcass weight and major cuts. Therefore, modeling is an important tool to predict variables that are difficult to obtain in experimental conditions and production systems. The use of this tool facilitates planning and decision making.