Inferência exploratória do comportamento de suínos em ambiente climatizado

Abstract

The Brazilian swine industry suffers numerous economic losses related to heat stress, and it is necessary to seek adequate solutions and mechanisms to control the production environment. The objective of this research was to carry out image monitoring and infer exploratory analysis of abiotic factors, physiological responses and behavioral indicators of swine in the growing and finishing phase, submitted to supplementary lighting programs in air-conditioned environments. The data used in this study came from an experiment carried out at the Experimental Animal Facility with Swine of the Serra Talhada Academic Unit (BES-UAST/UFRPE), located in the Sertão of the state of Pernambuco, Brazil. Twenty-seven swine (males and females) in the growing phase, of commercial lineage 3/4 Duroc and 1/4 Pietrain, mated by male Duroc (Pure of Origin) were used. The experiment was carried out from August to November 2018, totaling 66 days. The animals were submitted to pens without air conditioning (BS), pens with forced ventilation (BV) and pens equipped with an adiabatic evaporative cooling system (BR), associated with different light supplementation programs: 12 h of natural light (L12), L12 + 4 h of artificial light (L16) and L12 + 6 h of artificial light (L18). The meteorological variables inside each pen, in the external environment to the facilities, the physiological and behavioral responses of the animals were recorded. The behavior of the pigs was evaluated through images (640 × 480 pixels) obtained once a week, during 24 hours, during the entire experimental period. The images were analyzed at 10 min intervals. The experimental design used was completely randomized, in a 3×3 factorial scheme, randomly distributed in nine pens, with three replications. Behavioral analysis was performed by determining the frequency of events of each behavior and determining the probability by the Chi-Square (X2) test. The data analysis used was multivariate, using the grouping method (tree diagram) and the principal components technique (PCA), to indicate the main predictors of pig ingestive behavior and the development of multiple linear regression models. The accumulated variance for, between PC1 and PC2, was higher than 69.70% in both stages of development. The analysis of hierarchical components (HCA) showed the relationship between the variables studied and the ingestive behavior, in addition, it allowed the characterization of the comfort of the animals to their respective treatments. The analysis of principal components evidenced the relationship between abiotic and biotic factors with the ingestive behavior of the animals, allowing to establish regression models using the variables that expressed greater influence with the ingestive behavior.