Estimation of Rectal Temperature of Goats Based on Surface Temperature

Infrared thermography (IR) is a non-invasive tool with potential to indicate changes in the animal's thermal conditions in response to the thermally stressful environment. The objective of this study was to evaluate the application of IR to estimate the rectal temperature of crossbred goats of the Boer breed. Six male crossbred goats of the Boer breed were distributed in a completely randomized design and submitted to temperatures of 26, 30 and 34 °C and 68% relative humidity. Rectal temperature (RT) and thermograms data were collected from animals at each air temperature (AT) evaluated. In the thermograms, the temperatures of the ocular globe (PT), head (HT), shoulder (ST), hindquarter (HQ) and maximum infrared (IR Max ) temperatures of the animals' surfaces were collected. The correlation of PT, HT, ST, HQ and IR Max data with RT was evaluated through the Pearson coecient analysis and the concordance using Bland-Altman diagrams. With the exception of the IR Max surface temperature, the others were adequate for the accurate estimation of RT, with PT standing out for presenting the highest correlation coecient with RT (r = 0.951) and estimation errors varying in the range of ± 0.27 °C.


Introduction
The exploration of beef goats is a potential activity, especially in arid and semi-arid regions, because they are rustic animals and adapted to the speci c climatic conditions of these regions. However, when they are exposed to high air temperatures, their production potential is reduced due to stimuli of the peripheral receptors and the corporeal nucleus that are sent to the speci c centers of the hypothalamus for the activation of evaporative and non-evaporative cooling systems altering the operation of the appetite control center. The suppressive impulses transmitted to the center of the appetite cause a decrease in food intake. Thus, fewer substrates are available for enzymatic activities, hormone synthesis and heat production which helps to cool the body thus reducing animal production e ciency (Marai and Haeeb, 2010).
The evaluation of the thermal state of animals destined for commercial exploration is usually done by rectal thermometry, where it is adopted as the representative temperature of the body core and its increase indicates that the thermoregulatory mechanisms are not being e cient for the dissipation of metabolic heat produced, this method in most cases requires that the animals be handled directly by the experimenter. This method is laborious and may in uence in the behavior animals, which in turn affects the thermoregulatory responses Kammersgaard et al. (2011). According to Kammersgaard et al. (2013) there are non-invasive methods that can be used to evaluate the thermal state of the animals, especially infrared (IR) thermography that has good precision and dispenses direct contact with the individuals, thus presenting a high potential both for the development of research as for the monitoring of animals at the farm.
The IR measures the thermal radiation of the animal's surface and translates this to surface temperature.
Comparing the rectal and IR thermometry methods, the two are distinct not only in the technique, where the rectal thermometry is based on the transfer of conductive heat to the sensor, while the thermographic equipment measures the radiation. In addition, thermography measures the temperature at the surface of the body, which is constantly involved in heat exchanges with the environment, while temperature of the rectal cavity depends on the thermal situation of the body core George et al. (2014).
The objective of this study was to evaluate the application of IR to estimate the rectal temperature of crossbred ¾ Boer + ¼ goats with no de ned racial pattern.

Experimental design
The animals were distributed in a completely randomized design with three treatments (air temperatures) and six replicates (animals). The air temperatures used in the experiment were determined based on the thermal comfort zone (ZCT) for goats mentioned by Souza et al. (2008), which is between 20 and 30°C, with relative humidity of between 50 and 70%. Thus, the animals were submitted to the three different average temperatures controlled: T26 = 26 ºC (thermal comfort zone), T30 = 30 ºC (temperature limit between comfort zone and thermal stress) and T34 = 34 ºC (above ZCT), with relative humidity and wind speed averages of 68% ± 4% and 1 m/s, respectively.

Experimental procedures
For each thermal condition studied, a period of ve days was used to adapt the animals to the controlled environment, as well as handling and feeding. Data were collected within 10 days for each treatment. In the interval between treatments, the animals were exposed to the temperature and relative humidity of the ambient air (with the open chamber) for the restoration of their physiological functions, for ve days.

Data collect
The thermograms and rectal temperature (RT) were acquired on the third, sixth and ninth days after the start of each experimental phase. The RT was taken manually by inserting a digital thermometer (Incoterm, Med ex, Digital Thermometer, Rio Grande do Sul, Brazil, variation of 32 ~ 43°C and accuracy of ± 0,20°C) ~ 2 cm in the rectum of each animal.
For the acquisition of the thermograms (Fig. 1A), a thermographic camera model Ti55FT (60 Hz, rmware version v.1.22, 320x240, accuracy ± 2°C, Fluke®, Washington, USA) was used. In the thermograms, was collected the mean infrared temperature of the ocular globe (PT) and of the skin surface were evaluated in the regions of the head (HT), shoulder (ST) and hindquarters (HQ) that were previously depilated, as shown in Fig. 1B. In addition to the mentioned temperatures, the maximum infrared temperature (IR Max ) of the animals' surface, identi ed in the lower region of the eye mucosa, was also evaluated.
Three thermograms of each animal were selected, obtained at each air temperature tested for the analysis of the regions under study. Emissivity was adjusted to 0.98, based on data for humans (Steketee, 1973), taking into account that the evaluated regions were absent of hair and that the characteristics of the skin are similar between humans and the animals evaluated. The distance to the targets was ~ 2 m, allowing a complete view of the animals, from the snout to the tail.

Statistical analysis
The data were presented in box diagrams (boxplot). The normality of the data was veri ed using the Shapiro-Wilk test and, subsequently, the data were analyzed using the analysis of variance (ANOVA) and F test, using the ExpDes.pt package (Ferreira et al., 2013) of statistical software R version 3.4.1. The Tukey test was used to compare the averages, assuming a probability of error of 5% (P < 0.05).
The correlations between PT, HT, ST, HQ and IR Max data with RT were analyzed using simple linear regressions, with a 95% con dence interval and by assessing the degree of elevation of Pearson's correlation coe cient.
The occurrence of agreement between the evaluated methods was veri ed using the t-test for paired samples (one sample t-test) (P < 0.05), applied in the differences between of the thermogram and RT measurements. In addition, Bland-Altman diagrams were generated in order to assess the differences between the compared methods, according to the increase in air temperature. The agreement limits were determined by calculating the average of the differences (bias, ) and their standard deviation (Sd), these limits being calculated as follows: . The possibility of occurrence of systematic and random errors in the prediction of RT by the IR method was evaluated, for this, the veri cation of the occurrence or not of signi cant bias correlation with the mean of their respective measures was performed.

Physiological responses
The effect of increasing air temperature on the RT, PT, HT, ST, HQ and IR Max responses, respectively, is shown in Fig. 2. It can be observed a signi cant increase (P < 0.05) in all physiological responses as a function of elevation of the air temperature from 26 to 30 and subsequently to 34°C.

Correlation between of the methods
The correlations between the physiological response RT and the temperatures PT, HT, ST, HQ and IR Max , according to the elevation of the air temperature are presented in Fig. 3 with a con dence interval of 95%. When analyzing the agreement between RT and IR Max , it is found that among the evaluated agreements, this is the one that presents the bias (difference between the measures) closest to zero (0.75), but it is also veri ed that the bias is correlated signi cantly (P < 0.05) with the averages of the measurements, thus, the difference observed between the values of RT and IR Max , depends on the amount of data collected, which may lead to the occurrence of random errors and inaccuracies in the estimate of RT.
The Fig. 5 shows the analysis of agreement between the RT values and the PT, HT, ST and HQ data, added with their respective biases. It can be noted that the methods have strong concordances, with biases values statistically equal to zero (P = 0.11, 0.37, 0.75 and 0.46, for PT, HT, ST and HQ, respectively) between the evaluated methods.

Discussion
When subjected to environments with air temperatures above the thermoneutral zone, homeothermic animals employ several thermoregulatory mechanisms to compensate for heat gain, per equivalent loss, and maintain internal body temperature within narrow limits of variation and achieve thermal equilibrium (McKinley et al., 2017). The present study shows that PT, HT, ST, HQ and IR Max increased (P < 0.05) according to the elevation of the air temperature, suggesting that this fact is due to the signi cant increase (P < 0.05) in the body core representative temperature (RT) as a function of the heat generation due to the metabolic reactions and the reduction of the thermal gradient between the animal and the environment, which reduces the ability of the animals to dissipate heat in a sensitive way, being this a primary physiological mechanism responsible for the dissipation of body core heat through the

Conclusions
It is possible to estimate the rectal temperature (RT) of crossbred Boer goats accurately, based on the surface temperatures of the ocular globe (PT), head (HT), shoulder (ST) and hindquarter (HQ) of the animals, being the region most suitable for this, according to the results of the present research, the region of the ocular globe that presented Pearson's correlation coe cient of 0.956. Another fact that makes this estimate feasible is that although none of the surface temperatures have shown agreement with the RT, the error values in the estimates showed a statistically constant behavior (with the exception of IR Max ), as the air temperature increased. Thus, it was possible to accurately estimate the RT values through the values of PT, HT, ST and HQ added to their respective biases, with PT also standing out in this criterion for presenting lower errors (± 0.27°C) in the estimates.