Estimation of thermal comfort indexes for production animals using multiple linear regression models

Abstract

Enthalpy, physical quantity indicating the amount of thermal energy in the medium, is used by many researchers as an indicator of thermal comfort for humans and production animals. This physical quantity has as input variables the dry bulb temperature, the relative humidity of the air and the local barometric pressure. According to consolidated information of temperature and relative humidity related to the animal homeostasis, it was possible to establish enthalpy ranges for thermal comfort of swine, poultry and cattle, considering the local barometric pressure and its variations, which is not easily accessible in situations of field. Thus, the present study aimed to use multiple linear regression models to estimate enthalpy values by means of easily accessible variables (dry and wet bulb temperatures and relative humidity) which can be obtained by means of psychrometers or even by means of low-cost sensors, currently accessible. Meteorological data from three cities of the Brazilian territory, each representing an animal production system (poultry, swine and cattle) were accessed from the National Institute of Meteorology (INMET) database. According to the analysis of the prediction quality verification indices, the obtained models are efficient in predicting enthalpy values with the use of dry bulb temperature and relative humidity.