Evaluation of the thermoregulatory response of goats naturally exposed to the tropical climate using longitudinal data

Rectal temperature (RT), heart rate (HR), and respiratory rate (RR), determined as repeated measurements over time in female goats, were used to identify covariance matrices that best fit the data for residual modeling on these three traits. Then, based on this result, the goats' responses to heat were evaluated. Five matrices were found with convergence for the three traits. The Heterogeneous Compound Symmetry matrix showed a good fit for modeling the residual associated with RT, whereas the Heterogeneous Autoregressive matrix had a better fit for RR and HR, according to the Akaike Information Criteria (AIC), corrected AIC (AICc), and Schwarz Bayesian Information Criterion (BIC) used. After adjusting the residual data for these three traits, a mixed-model analysis was used to evaluate collection period (3), physiological stage (3), and animal age (3) as fixed effects. Residual modeling interfered differently with the p-value associated with the fixed effects studied. Collection period and interactions did not influence the variation in RT (P>0.761), which was within the standard range for goats in the tropics, while the physiological stage of the goats affected it (P<0.05). Rectal temperature, HR, and RR tend to show covariance structures that can be modeled using specific residual covariance matrices, that is, the heterogeneous compound symmetry matrix best suits RT data, whereas the heterogeneous autoregressive matrix is better suited for HR and RR, which are usually correlated. The goats of the evaluated breed maintain RT within the range of variation displayed by breeds adapted to a hot environment, regardless of their physiological condition. Variations occur in RR and HR, without, however, exceeding the normal range for goats. Pregnancy causes goats to raise their RR in the rainy season of the year in the region in order to maintain RT within the normal range for the species.