Evaluation du modèle nutritionnel InraPorc pour un système de finition de porcs lourds du sud du Brésil

To evaluate the utilization of the InraPorc model in a heavy pig system developing in the South Brazil Region, an experiment was performed with 12 barrows and 12 gilts (77.2 ± 3.7 kg initial body weight, BW), distributed in eight pens, half fed ad libitum (AL) and half restrictively fed (R). Feed intake (FI), BW and backfat thickness (BFT) were weekly measured for 12 weeks, until the average BW of 151.0 ± 7.8 kg. Weekly cumulated average data of BW and FI, and the ingredient composition and analysed nutritional feed composition were used to calibrate the model’s animal profile, using the gamma maintenance equation for FI. Linear regression analysis between predicted (X) and measured (Y = b0 + b1 X) data showed very good calibration for BW (0.23, 99%, 0.27, 0.99, < 0.01, respectively for RSD, r², b0, b1 and P-value), cumulated FI (0.77, 99%, 0.90, 0.99, < 0.01), and BFT for AL barrows (0.70, 98%, 2.00, 1.07, < 0.01). The predicted value for R barrows resulted in a small overestimation for BW (0.95, 99%, 7.68, 0.90, < 0.01) and BFT (0.56, 97%, 3.66, 0.49, < 0.01). For gilts, prediction of the tissue deposition dynamic was divergent from the biological kinetic, due to the fluctuation in FI and BW gain of many females that entered on puberty during the trial. Using a measurement of feed wastage during the calibration trial and a BFT equation corrected for the Brazilian measurement method could be useful to improve the prediction capacity of the InraPorc calibration to the local system.