Thermal balance of livestock 2. Applications of a parsimonious model

A mathematical model developed from heat transfer principles to predict the thermal status of a homeotherm was applied to sheep and cattle outdoors an d pigs and broiler chickens indoors. The climatological variables considere d in the model include air temperature, wind speed, vapour pressure and sol ar radiation. For sheep, the fleece depth varied seasonally and thermal bal ance was achieved by a metabolic response, vasodilation and panting. For ca ttle, the thermal responses included sweating and piloerection of the coat. The insulation provided by the pig's sparse hair coat was neglected, but t he increase in its body insulation with age and environmental conditions wa s included as a major determinant of heat loss. For chickens, the insulatio n provided by the body tissue and feathers was described by a single therma l resistance. Their thermal responses included feather fluffing,, vasomotor action in the combs and feet, and changes in respiration rate and body tem perature. The models were tested successfully for each species by simulating the expe rimental conditions used by previous workers and comparing the predictions with measured values of heat loss, skin and body temperature. The intercept ion of solar radiation by animals outdoors was also tested successfully for solar elevations up to 45 degrees. For sheep, the predicted heat loss agreed with measurements to within 10%. The onset of vasodilation for a shorn sheep on maintenance food intake was predicted successfully to occur at an air temperature of 25 degrees C, and the variation of skin temperature on the legs with air temperature was pred icted to within the uncertainty of the measurements. The model predicted th e heat loss from cattle in the cold with acceptable accuracy when the wind speed was low, but overestimated heat loss from calves by up to 30% in wind . in warm conditions, the evaporative heat loss from cattle as a consequenc e of sweating was predicted with acceptable accuracy. The errors incurred b y ignoring solar radiation penetration into the coat were acceptably small, given the associated reduction in model complexity. Sensitivity analysis s howed that the predictions of heat loss from sheep and cattle were sensitiv e to wind speed and coat length, especially when the coat is short. For bot h species, the level of stress was sensitive to ambient vapour pressure at high air temperatures. For a single new-born pig, the model underestimated heat loss at 30 degrees C with an overall error of -9% over the range of wind speeds likely to be experienced indoors. The model over-predicted heat loss by an average of 20 % at 20 degrees C, probably due to the absence in the model of a temperatur e-dependent huddling response. However, far a 25 kg pig exposed to air temp eratures from -5 to 35 degrees C, the model predicted the skin temperature on the trunk - a good indication of its thermal status - to within the Limi ts of the experimental uncertainty. The total heat loss from chickens expos ed to temperatures in the range 0-38 degrees C was predicted with an overal l error of 6%. In a separate test, the body core temperature of hens was pr edicted to within 0.3 degrees C on average for the same range of air temper ature, again within the limits of experimental uncertainty. Sensitivity ana lysis showed that the prediction of body temperature for chickens was most sensitive to ambient humidity at high air temperatures, and to body resista nce. The paper discusses the limitations of the models and the need for mor e measurements of heat losses from current breeds of livestock. (C) 2000 El sevier Science B.V. All rights reserved.