Access to this full-text is provided by MDPI.
Content available from Animals
This content is subject to copyright.
animals
Review
Recent Advances on Early Detection of Heat Strain in Dairy
Cows Using Animal-Based Indicators: A Review
Hang Shu 1,2 , Wensheng Wang 1, *, Leifeng Guo 1, * and Jérôme Bindelle 2
Citation: Shu, H.; Wang, W.; Guo, L.;
Bindelle, J. Recent Advances on Early
Detection of Heat Strain in Dairy
Cows Using Animal-Based Indicators:
A Review. Animals 2021,11, 980.
https://doi.org/10.3390/ani11040980
Received: 22 February 2021
Accepted: 23 March 2021
Published: 1 April 2021
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2021 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
1Agricultural Information Institute, Chinese Academy of Agriculture Sciences, Beijing 100086, China;
shuhang@caas.cn
2AgroBioChem/TERRA, Precision Livestock and Nutrition Unit, Gembloux Agro-Bio Tech,
University of Liège, 5030 Gembloux, Belgium; jerome.bindelle@uliege.be
*Correspondence: wangwensheng@caas.cn (W.W.); guoleifeng@caas.cn (L.G.)
Simple Summary:
In the dairy industry, heat stress and its induced heat strain result in huge
economic loss every year. To better manage heat strain in dairy cows, it is more sensible to advance
the detection by using more sensitive indicators so that cooling measures can be implemented in
time. With the development of sensor technologies and wireless transmission technologies, body
surface temperature and respiration rate can be measured automatically through wearable devices.
Lots of efforts have been made recently to develop meaningful thresholds on both physiological
and environmental sides. However, the existing thresholds should be used carefully considering
the differences in experimental conditions and animal information. Further studies are required to
evaluate and customize thresholds based on different influencing factors. A comprehensive early
detection system for heat strain based on both animal- and environment-based indicators is expected.
Abstract:
In pursuit of precision livestock farming, the real-time measurement for heat strain-related
data has been more and more valued. Efforts have been made recently to use more sensitive
physiological indicators with the hope to better inform decision-making in heat abatement in dairy
farms. To get an insight into the early detection of heat strain in dairy cows, the present review
focuses on the recent efforts developing early detection methods of heat strain in dairy cows based on
body temperatures and respiratory dynamics. For every candidate animal-based indicator, state-of-
the-art measurement methods and existing thresholds were summarized. Body surface temperature
and respiration rate were concluded to be the best early indicators of heat strain due to their high
feasibility of measurement and sensitivity to heat stress. Future studies should customize heat strain
thresholds according to different internal and external factors that have an impact on the sensitivity
to heat stress. Wearable devices are most promising to achieve real-time measurement in practical
dairy farms. Combined with internet of things technologies, a comprehensive strategy based on both
animal- and environment-based indicators is expected to increase the precision of early detection of
heat strain in dairy cows.
Keywords: heat strain; heat stress; dairy cows; early detection; measurement; threshold
1. Introduction
Thermoregulation is of importance for homeotherms to maintain homeothermy and
homeostasis by keeping the balance between the production and loss of body heat [
1
].
Heat production in dairy cows mainly comes from basal metabolism, rumen fermentation,
nutrient absorption, growth, lactation, gestation, immunization, and exercise [
2
]. Heat
loss is mainly through heat transfer mechanisms: Conduction, convection, radiation, and
evaporation [1,2].
Cows can maintain a balance of heat production and heat dissipation in the ther-
moneutral zone (TNZ), in which cows perform the lowest physiological and immune
Animals 2021,11, 980. https://doi.org/10.3390/ani11040980 https://www.mdpi.com/journal/animals
Animals 2021,11, 980 2 of 24
costs, and the highest productivity [
1
]. When the thermal environment exceeds TNZ and
suppresses the efficiency of non-evaporative heat loss, evaporative heat loss, or both [
3
,
4
],
various mechanisms are activated in dairy cows to dissipate excess heat and maintain
homeothermy [
5
]. In the present review, “heat stress” represents the sum of external
stressors, forcing the animal to exceed its TNZ. Environmental thermal stressors consist of
ambient temperature (AT), relative humidity (RH), wind speed, and radiation, where AT
plays the leading role [
6
]. “Heat strain” is used to represent the overall response resulting
from heat stress. These two terms are separated since many internal animal factors can
alter dairy cows’ sensitivity to heat exposure.
To address the question of how animals are able to cope with heat stress and their
induced heat strain, the use of animal-based indicators to detect heat strain directly is
of particular importance [
7
–
9
]. Unfortunately, the animal husbandry sector has long
been limited by the lack of labor- and time-saving measurement methods for animal-based
indicators [
10
]. Many thermal indices have been developed to reduce the multi-dimensional
thermal stressors to better depict the thermal environment, and have been connected to
animal-based indicators with the hope to assess heat strain indirectly with easily accessible
weather data. Temperature-humidity indices (THI)-a series of highly correlated indices
combining temperature and humidity-have been developed by different scientists and are
used most commonly [
11
]. Thermal indices have already been evaluated comprehensively
in other reviews [
12
,
13
]. Notably, environmental indicators have difficulties dealing with
individual differences in the onset of heat strain and showing whether animals are being
adequately cooled [7,14].
For decades, the onset of heat strain has long been associated with productive pa-
rameters [
15
]. However, production loss is more likely a result of heat strain than an
early indicator [
7
] and usually lags the changing environment by about two days [
16
,
17
].
In fact, heat strain includes both short- and long-term responses from vasodilation that
happens instantly to gradual physiological adaptation [
1
,
5
]. Recently, many efforts have
been made to use more sensitive physiological indicators to help detect heat strain in the
early stage [
18
,
19
]. Among all kinds of physiological indicators, body temperatures and
respiratory dynamics are discussed and used most commonly [7,18].
For dairy cows, the increased heat production due to the recent progression in pro-
ductivity coupled with their insufficient heat dissipation capacity makes them currently
more susceptible to heat strain [
1
,
7
]. Due to global warming, dairy cows are being exposed
to heat wave events more often, and the resulting economic losses are projected to increase
substantially [
20
,
21
]. All this information emphasizes the importance of improving the
early detection of heat strain in dairy cows to avoid the subsequent reduction in production
and welfare, and the increase in disorders and mortality, etc.
Precision livestock farming (PLF) aims to provide real-time monitoring for significant
issues so that interventions can be implemented immediately [
22
]. In light of this, the
demand for non-invasive, continuous, remote, accurate, real-time, and ultimately automatic
measurements for physiological indicators of dairy cows has been more and more valued
for heat strain detection [
5
,
7
,
9
,
14
,
23
]. Recently, many new methods have been developed to
measure heat strain-related data [
10
,
24
]. On the basis of measurement, a series of thresholds
are naturally developed to quantify and assess heat strain.
To get an insight into the early detection of heat strain in dairy cows, the present
review focuses on the recent efforts developing early detection methods of heat strain
in dairy cows based on body temperatures and respiratory dynamics. The primary ob-
jective was to summarize the progress on the measurement methods, determining the
most feasible method. The secondary objective was to summarize the existing thresholds
from both physiological and environmental viewpoints, aiming to find the most sensitive
indicators for the future development of early detection methods in the setting of practical
dairy farms.
Animals 2021,11, 980 3 of 24
2. Detection of Heat Strain Based on the Most Commonly Used Animal-Based Indicators
The methods for obtaining heat strain-related data can be divided into two categories,
i.e., direct measurements [
9
,
10
] and predictive modeling [
11
]. The predictive modeling
is not the focus of this review, but still, it is mentioned when describing the relationship
between candidate indicators. To assess and quantify heat strain, critical thresholds on
both environment and animal sides have been developed. Therefore, for every candidate
indicator in this section, the latest measurement methods and thresholds were discussed.
2.1. Core Body Temperature
Core body temperature (CBT) is the most commonly used animal-based indicator
of heat strain in dairy cows, in which rectal temperature (RT) plays a leading role [
18
].
Normally, the range of CBT of homeotherms is very narrow with a circadian rhythm due
to the effort of the thermoregulation [
25
]. When cows are heat-stressed, their CBT increases
abnormally if they fail to dissipate excess body heat. The latest measurement methods and
thresholds of CBT are listed in Tables 1and 2, respectively.
Table 1. Summary of the state-of-the-art measurement methods for body temperatures.
Indicator Technology Automatic Continuous Real-Time Accuracy Reference
Rectal
temperature
Digital thermometer No No No ±0.1 ◦CGarner et al. [26]
A thermistor probe attached to a
recorder Yes Yes No Unknown Spiers et al. [27]
An indwelling data logger
supported by a customized tail
harness or piping
Yes Yes No ±0.2 to ±0.5 ◦CReuter et al. [28];
Lees et al. [29]
Radiofrequency-based digital
thermometer Yes Yes Yes ±0.5 ◦CDebnath et al. [30]
Vaginal
temperature
Temperature probe with a
long-finger anchor Yes Yes No ±0.2 ◦CLee et al. [31]
Temperature data loggers and
modified vaginal controlled
internal drug release
Yes Yes No ±0.1 to ±1◦C
Kendall et al. [32];
Vickers et al. [33];
Burfeind et al. [34];
Burdick et al. [35];
Tresoldi et al. [36];
Kaufman et al. [37];
Garner et al. [26]
A wireless vaginal temperature
device Yes Yes Yes Unknown Sakatani et al. [38]
An indwelling device equipped
with temperature sensors, a data
collector, and a computer system
Yes Yes Yes
Mean difference of
0.02 ◦C with a 95%
confidence interval:
−0.23 to 0.26 ◦C
Wang et al. [39]
Subcutaneous
temperature
Implantable wireless
thermometers based on Bluetooth Yes Yes Yes ±0.05 ◦CIwasaki et al. [40]
Implantable radiofrequency
identification biosensors and
wearable scanners
Yes Yes Yes Unknown Chung et al. [41]
Ruminal
temperature Radiotelemetric ruminal and
reticular bolus Yes Yes Yes Unknown Nogami et al. [42]
Tympanic
temperature
Continuous monitoring
temperature probes Yes Yes Yes ±0.1% Mahendran et al. [43]
Infrared ear thermometer Yes No No ±0.2 ◦CWoodrum Setser et al. [44]
Body surface
temperature
Handheld infrared camera No No No ±1% Peng et al. [45]
Handheld infrared gun No No No ±1.5 ◦CKaufman et al. [37]
Fixed infrared camera Yes Yes Yes ±2% Jorquera-Chavez et al. [46]
Animals 2021,11, 980 4 of 24
Table 2. Summary of the thresholds of body temperatures.
Indicator
Threshold for
Animal-Based
Indicators
Threshold for
Environment-Based
Indicators
Cow Information Thermal Condition Reference
Rectal temperature
38.4 ◦CTHI a70
High-producing
Holstein-Friesian dairy
cows
A continental climate
with a THI range of
17.8 to 85.1 Pinto et al. [19]
38.6 ◦CTHI a70 High-producing Holstein
dairy cows
Hot climate with an AT
range of 9.5 to 30.8 ◦CLi et al. [47]
38.55 ◦CTHI a74.1 High-producing Holstein
dairy cows
Warm temperate
semi-humid
continental monsoon
climate with a THI
range of 58 to 84
Peng et al. [45]
Unknown BGHI 76.44 1⁄2Holstein dairy cows BGHI stayed higher
than 72 Dalcin et al. [48]
Unknown BGHI 73.51 3⁄4Holstein dairy cows BGHI stayed higher
than 72 Dalcin et al. [48]
Unknown THI a67 Preweaning dairy calves
Averaged THI of 78
with shade only in a
subtropical climate
Dado–Senn et al.
[49]
Unknown THI b88.1 Preweaning Holstein bull
calves
Hot climate with a THI
range from 70.3 to 94 Kovács et al. [50]
Vaginal temperature
Unknown THI c69
Multiparous
nonpregnant
Holstein-Friesian dairy
cows
THI ranged from 55.8
to 79.9 Nabenishi et al. [51]
38.9 ◦CUnknown Dairy cows Hot–humid climate
with an average THI of
82.4 Hillman et al. [52]
Unknown THI a70 Lactating Holstein cows Unknown Atkins et al. [53]
Ruminal temperature 39.3 ◦CTHI a65
Lactating
Holstein-Friesian dairy
cows
Temperate climate with
THI values varied from
9.6 to 84.9
Ammer et al. [54]
Milk temperature Unknown
Dynamic thresholds
using a decision tree
model
Lactating Holstein cows
A subtropical climate
with an ambient
temperature range of
2.0 to 38.0 ◦C
Ji et al. [55]
Body surface
temperatures: Eyes,
hindquarters, nose, part
of horns, and ears
Unknown THI a65 (eyes) to 70
(hindquarters) Hanwoo heifers THI ranged from 75.1
to 84.7 Kim et al. [56]
Body surface
temperatures: Forehead
(mean, maximum)
30.05, 30.34 ◦CTHI a71.4, 66.8 High-producing Holstein
dairy cows
Warm temperate
semi-humid
continental monsoon
climate with a THI
range of 58 to 84
Peng et al. [45]
Body surface
temperature: Ear Unknown THI b83.0 Preweaning Holstein bull
calves
Hot climate with a THI
range from 70.3 to 94 Kovács et al. [50]
THI = temperature–humidity index; BGHI = black globe-humidity index.
a
THI equation from National Research Council [
57
];
b
THI
equation from Bianca et al. [58]; cTHI equation from Mader et al. [59].
2.1.1. Rectal Temperature (RT)
Traditionally, RT is measured manually with a digital thermometer inserted 9 cm
into the rectum for calves and 15 cm for the cows [
60
]. This procedure is both time- and
labor-consuming and cannot achieve a measurement in real-time [
34
]. More importantly,
this invasive operation would cause additional stress to cattle, biasing the results. Gaughan
et al. [
61
] used a temperature data logger (Smart Reader 8, ARC Systems, Brisbane) to con-
tinuously measure temperature rectally. Spiers et al. [
27
] used a stainless-steel thermistor
probe attached to a recorder (Cole Parmer Instruments, Chicago, IL, USA) to measure RT.
Reuter et al. [
28
] designed a device consisting of an indwelling data logger supported by a
custom-fabricated aluminum tail harness to achieve an automated measurement for RT.
However, due to the potential damage to the tail and the influence of fecal temperature,
this device is not widely used [
9
,
35
]. Debnath et al. [
30
] achieved a real-time measurement
for RT in dairy cows using a radiofrequency-based digital thermometer. Lees et al. [
29
]
Animals 2021,11, 980 5 of 24
developed an intra-rectal device to measure RT continuously for 23 h in grazing heifers.
This device consists of a data logger (iButton DS1922L, Maxim Integrated, San Jose, CA,
USA) and a soft polyethylene piping. As summarized by Burfeind et al. [
62
], the measure-
ment of RT could be biased by the type of thermometer, the depth into the rectum, and the
operation itself.
A daily temperature circadian rhythm in cattle older than two months was observed
with RT at dusk up to 1.4
◦
C higher than that in dawn [
60
]. It is noted by Tresoldi
et al. [
36
] that RT measured a few times a day may result in a relatively lower daily average,
especially when measuring only once in the morning. Except for circadian rhythm, many
heat-producing events (e.g., feeding, exercise, estrus) would lead to an increased RT [
30
,
60
].
Therefore, it is of importance to avoid potential bias from these interfering events when
determining the onset of heat strain.
Many studies treated environmental indicators (e.g., THI) as a categorical variable
to test the RT differences among different levels of heat exposure [
26
,
63
]. Other studies
adopted similar methods by separating the records into different levels according to climate
conditions (e.g., hot versus moderate, summer versus spring) [
34
,
64
]. These methods could
find a strong association between RT and environmental parameters, but could not obtain
accurate environmental and RT thresholds.
In a study using 45 high-producing Holstein cows, Li et al. [
47
] applied a broken-line
model and determined the threshold of RT as 38.6
◦
C when heat strain was triggered at a
THI of 70. Similarly, Pinto et al. [
19
] demonstrated a slightly lower threshold of RT (38.4
◦
C)
at the same THI of 70 using 139 multiparous lactating Holstein-Friesian cows. However,
these two studies were inconsistent in the hypothesis of normal RT. The former assumed
normal RT as a plateau (i.e., a constant), while the latter used a line with a slight slope to
fit normal RT. In theory, it is more reasonable to regard normal RT as a range normally
distributed around a constant due to the existence of the thermoregulatory system.
Dalcin et al. [
48
] also used broken-line models with plateaus representing normal
conditions to explore thresholds in 38 lactating cows from various genetic groups (
1⁄2
,
3⁄4
,
and pure-bred Holstein). The black globe-humidity index (BGHI) was used to describe
the thermal environment. The BGHI thresholds 76.44 and 73.51 were determined for
1⁄2
and
3⁄4
Holstein, respectively. Corresponding RT thresholds were around 38.2 and 37.8
◦
C.
No threshold was derived for pure-bred Holstein due to the high linear correlation with
BGHI. As the BGHI stayed higher than 72 during the entire experiment, it can be inferred
that the BGHI threshold for pure-bred Holstein could be somewhere lower than 72. These
results indicate that
1⁄2
Holstein cows were most heat-tolerant, while pure-bred Holstein
cows were most vulnerable. The plateaus were also applied by Dado–Senn et al. [
49
]
when determining thresholds of various physiological indicators in preweaning dairy
calves. Limited attention has been paid to heat abatement in dairy calves due to their larger
surface-to-mass ratio. Unexpectedly, a THI threshold (67) for RT was determined for calves
under shade without fans.
2.1.2. Vaginal Temperature
Vaginal temperature (VT) is another representative CBT. The traditional measurement
of VT is very similar to that of RT and many efforts have been made to achieve automated
measurements. Various temperature data loggers used in conjunction with support devices
have been fixed into cows’ vaginas. A plastic anchor was first introduced as a support
device by Hillman et al. [
65
] and further validated by Lee et al. [
31
]. It is demonstrated
that long-finger plastic anchors were superior due to better stability when inserted in the
vagina [
31
]. However, the application of these anchors may be difficult due to inaccessibil-
ity [
35
]. The most popular way is to use a modified vaginal controlled internal drug release
insert, in which various data loggers are placed [
32
,
33
,
35
]. The prices of data loggers vary
distinctly. When considering elevated VT thresholds, inexpensive data loggers failed to
correctly classify animals compared with an expensive and high-accurate data logger [
36
].
Animals 2021,11, 980 6 of 24
Notably, all these abovementioned methods cannot monitor VT in real-time due to the lack
of wireless transmission [35].
Sakatani et al. [
38
] achieved wireless transmission of VT data using a temperature
sensor (Gyuonkei, Remote Inc., Oita, Japan) in conjunction with Wi-Fi technology. More
recently, Wang et al. [
39
] designed a wireless measuring system consisting of an indwelling
device equipped with temperature sensors (ADT7320, Analog Devices Inc., Norwood, MA,
USA), a data collector, and a computer system.
All the indwelling devices for measuring VT can only be placed in the vagina for
up to 24 days (around a week in most applications) [
31
,
36
] due to the increased risk of
irritation and infection [
35
]. Thus, long-term monitoring cannot be carried out and only
short-term monitoring is meaningful for research purposes. In addition, the temperatures
recorded vaginally may be affected by various factors, e.g., logger displacement, lying [
33
].
In addition, the requirement of an insert may hurt cow welfare and pregnancy, which is
still under assessment.
Due to close anatomical proximity between the vagina and the rectum [
66
], VT has
been regarded to be as reliable as RT in measuring the severity of heat strain [
33
]. VT is
highly correlated with RT with a correlation coefficient up to 0.97 and the measured values
are very close, especially when measuring with the same temperature loggers
[33,35,67,68]
.
This correlation is highly biased by the accuracy of the devices, as the digital thermometers
used for RT measurement are supposed to have higher accuracy than most of the data
loggers used for measuring VT [
33
,
36
]. Mean VT was found higher than mean RT which
might be due to the fact that the data logger measuring VT was inserted into the vagina
deeper than the thermometer measuring RT [
34
]. Due to the access to continuous measure-
ment, VT is better at describing circadian rhythm than RT [
33
]. To reflect daily average
CBT accurately, VT should be measured every 120 min or more [36].
The relationship between VT and the thermal environment has been demonstrated
by many studies that VT was higher during the high THI period [
26
,
34
,
37
,
51
]. More
importantly, Kaufman et al. [
37
] reported that VT showed stronger relationships with THI
compared with RT, and therefore may be a better indicator determining heat strain in dairy
cows. The vagina has abundant blood flow, making VT more sensitive to the changes
in CBT than RT [
35
]. Nabenishi et al. [
51
] reported that the normal mean daily VT of
11 healthy multiparous nonpregnant Holstein-Friesian cows was stable around 38.5
◦
C
and began to rise when THI reached 69. Furthermore, a threshold for VT (38.9
◦
C) was
observed by Hillman et al. [
52
] according to the critical point where dairy cows began
to stand and seek cooling. Using a broken-line model, a THI threshold (70) for VT was
determined for eight lactating Holstein cows by Atkins et al. [53].
2.1.3. Ruminal and Reticular Temperature
The rumen and the reticulum are also locations of interest to measure CBT. Many
boluses have been designed and applied to measure ruminal and reticular parameters (e.g.,
temperature, pH) and many of them have already been commercially available [
24
,
69
–
72
].
A rumen or reticulum bolus is a configured pill placed in the reticulum or the junction
between the rumen and reticulum, consisting of a temperature sensor, a telemetry system,
and a battery [
9
]. Hicks et al. [
69
] first measured the temperature ruminally using the
CorTemp sensors (HTI Technologies, Inc., Palmetto, FL, USA) in three cows. Initially, this
method required a fistula or an implant surgery, which compromised animal welfare. At
present, boluses are administered orally via a customized bolus applicator or a balling
gun [68,73,74]. Rumen boluses have recently been customized for dairy calves to monitor
infections, but whether these devices can be used for heat strain detection in dairy calves
remains unknown [
75
]. With the development of wireless and battery technology, the new
minimized bolus can work for more than 600 days [42].
Rumen temperature (RUT) and reticular temperature (RET) have been measured with
the hope to develop a continuous and less invasive method for monitoring CBT than RT or
VT [
24
,
69
,
71
]. The correlation coefficients between RUT and RT vary greatly among studies
Animals 2021,11, 980 7 of 24
from 0.34 [
70
] to 0.92 [
76
]. The correlation coefficients between RET and RT range from
0.50 [77] to 0.645 [71].
The temperatures recorded in the rumen and reticulum are greater than RT by ap-
proximately 0.5
◦
C due to the heat produced by rumen microorganisms [
71
]. A drinking
bout can also decrease RET by 9.2
◦
C and it would take up to 3.5 h to return to the original
temperature before drinking [
54
,
71
]. Given that the reestablishing time depends on both
temperature and amount of water drunk, it is possible to use it to predict water intake,
which also plays a part in heat strain in cattle [
78
]. To mitigate the impact of water intake
on RET, Ammer et al. [
68
] calculated median and mean RET for 2 h preceding RT and VT
measurement and demonstrated that the median RET was more highly correlated to RT
and VT (r = 0.48, 0.53) compared with mean RET (r = 0.43, 0.46) and single RET (r = 0.40,
0.48). Since calves drink less cool water than adult cattle and the milk or milk supplements
are preheated, the intake of water may not bias RET as much as for adults [75].
Many previous studies demonstrated a correlation between RUT, RET, and ambient
thermal environment [
54
,
71
,
79
,
80
]. Notably, Prendiville et al. [
70
] found that AT correlated
more with RUT (r
2
= 0.55) than with RT (r
2
= 0.29), indicating RUT was more sensitive to
thermal changes. Liang et al. [
79
] reported that daily RET remained stable around 40
◦
C
until daily AT exceeded the upper critical temperature, and the increasing rate of daily
RET increased with AT. However, no threshold was given in their study. Ammer et al. [
54
]
divided THI into 4 classes: <60, 60 to 65, 65 to 70, and
≥
70, and reported that the daily
median RET of 28 Holstein-Friesian dairy cows with an average milk yield of 35.2 kg/d
remained unchanged at 39.2
◦
C when THI was lower than 65 and increased to 39.3
◦
C
(
p< 0.001
) when THI was between 65 and 70, and further increased to 39.4
◦
C (p< 0.001)
when THI
≥
70. These thresholds are not accurate since THI was treated as a categorical
variable. Studies on the heat strain threshold of RUT, RET, as well as the corresponding
environmental indicators, are very limited.
2.1.4. Tympanic Temperature
Tympanic temperatures (TT) were measured in beef cattle through data loggers
placed in the ear canal [
81
–
83
]. This method was also used in dairy cows to measure
heat strain
[84,85]
. However, several limitations may affect its further application in prac-
tice. Firstly, cows behave uncomfortably to these foreign objects and potential ear infections
may occur following the installation, especially over a long-term use [
66
]. Secondly, the
probes need to be properly located near the tympanic membrane, and any dislocation
could result in inaccurate readings [
9
,
70
,
86
]. To date, wireless temperature sensors that can
be fixed in the ear canal have already been commercialized [
86
,
87
]. It has been reported
that the latest model of these temperature-sensing ear tags has an accuracy of
±
0.1% in
◦
F [
43
]. In addition, for instant use, an infrared ear thermometer provides a reading of TT
in one second [44].
Many efforts have been made to use TT as an indicator for heat strain in beef cat-
tle
[83,88]
. Bergen and Kennedy [
66
] found that TTs were consistently lower (p< 0.01) but
highly correlated (r = 0.77) with VTs during the test days in nine beef heifers. Prendiville
et al. [
70
] reported that the temperatures recorded in the rectum through a digital ther-
mometer were not statistically different from those recorded wirelessly in the ear canal
(
p> 0.05
). In another beef cattle study, Richeson et al. [
87
] used a heat strain threshold
of 39.8
◦
C for TT. However, this threshold was directly obtained from the preexisting
threshold for RT. Very few studies focus on TT in dairy cows. Myers and Henderson [
89
]
found that TT readings were lower than RT consistently in both Holstein calves and cows.
Jara et al. [
85
] reported that the TT of 15 multiparous Holstein-Friesian dairy cows was
highly associated with thermal indices, especially with the comprehensive climate index
(CCI). However, no threshold was given in this study.
Due to the anatomical proximity to the hypothalamus [
66
], some researchers argued
that TT is more sensitive and responds earlier by as fast as 18 min to changing AT compared
with RT or VT [8]. Notably, TT is the only validated indicator of CBT in calves [44].
Animals 2021,11, 980 8 of 24
2.1.5. Subcutaneous Temperature
Subcutaneous temperature (ST) is also a candidate to represent CBT since heat-stressed
cows would drive more blood from the core to the peripheral body parts [
1
,
41
]. In gen-
eral, measuring temperatures subcutaneously requires the use of implantable sensors and
wireless transmission. The ST can vary among different implantation locations and many
efforts have been made to determine the best locations for measuring accurate CBT in
cattle [
40
,
90
,
91
]. Reid et al. [
91
] used implantable radiofrequency identification (RFID)
microchips measuring temperature in three sites, subcutaneously at the base of the ear, pos-
terior to the poll, and beneath the umbilical fold. They demonstrated that the temperatures
at the base of the ear correlated most with RT with coefficients of 0.30 during heat stress
and 0.20 during TNZ. Similarly, Chung et al. [
41
] designed a real-time measurement for ear
base temperature by using implantable sensors and wearable RFID scanners in conjunction.
In addition, they utilized RFID technology, long-range wireless technology, and Wi-Fi
technology to develop a wireless communication network aiming to increase efficiency and
lower human involvement. Furthermore, a long short-term memory network was applied
to predict VT by using THI and ST. Their results show that the correlation coefficients
between temperatures recorded under the ear base and in the vagina of three dairy cows
ranged from 0.58 to 0.85. The selected machine learning algorithm could predict CBT better
than linear regression with a lower root mean square error (0.081 ◦C versus 0.137 ◦C).
Lee et al. [
90
] measured ST in seven Holstein steers using button-shaped data loggers
(iButton DS1922L, Maxim Integrated, San Jose, CA, USA) in three sites (lateral neck, upper,
and lower scapula). Due to the lack of wireless transmission, temperatures recorded were
not available until recovered from an explant surgery. Their long-term results (during
a whole summer and winter) show that STs were 1.39 to 1.65
◦
C lower than the RT and
might have been influenced by AT and animal stillness; among three test locations, the
ST recorded near the upper scapula had the smallest mean difference from RT. In a recent
study of dairy calves [
44
], temperatures recorded subcutaneously in three sites (the ear
scutulum, the upper scapula, and intramuscularly in the trapezius muscle of the neck)
were found very different from RT. Iwasaki et al. [
40
] implanted wireless thermistor sensors
based on Bluetooth
TM
4.0 in more locations to find the best location for measuring CBT.
Specifically, three endoceliac and seven subcutaneous locations were selected, and machine
learning was further applied to predict CBT from STs. The results show that ST under
the tail base was a good indicator due to its high correlation with RT (r = 0.62) and the
correlation coefficient could be further increased to 0.93 by introducing a back-propagation
type neural network with the input of seven STs and AT.
Basically, previous studies in ST aim to predict gold standard CBT (i.e., RT and VT),
and some locations very related to AT were considered to interfere with the prediction.
In fact, the STs of the peripheral parts respond to the thermal changes earlier than RT
or VT. To date, very few studies have attempted to use ST directly as a more sensitive
indicator of heat strain. In addition, although adverse events were not reported by almost
all relative studies, very few of them observed the impact of implantation on health over a
long-term period. It is also unclear whether the implanted sensors can be recycled from
slaughterhouses or have any impact on food safety. The cost and complexity limit their
further applications to research instead of practical production.
2.1.6. Other Core Temperatures
Milk temperature (MT) has been proposed as a potential indicator for heat strain in
dairy cows. With the development and application of automatic milking systems, the MT
per milking per cow can be collected and stored automatically.
MT was reported to be correlated with VT in 31 dairy cows during a week (r = 0.52) [
92
].
Igono et al. [
93
] reported that MT measured in the clawpiece at 1600 pm correlated most
with RT measured at 1630 pm right after the milking in Holstein cows housed in shade
(
r = 0.875
) and shade plus water spray and fan (r = 0.894). In addition, the MT was reported
to be significantly influenced by the maximum AT [
94
] and THI [
95
]. All of these contribute
Animals 2021,11, 980 9 of 24
that MT is a good indicator of heat strain in dairy cows. More importantly, the effect of
THI on MT differs by production level [
95
]. Specifically, MT increased with the increase of
THI and the effect was more obvious for high-producing cows, indicating the importance
to group the cows by their productivity. Ji et al. [
55
] first divided the cows into different
age, body mass, and day in milk subgroups according to production level, respectively,
and then used segmented regression models to obtain a series of thresholds. In a study
where the fever was defined as VT higher than 39.5
◦
C, a threshold for MT (38.7
◦
C) had
the highest combination of sensitivity (0.77) and specificity (0.66) [92].
Although MT is a time- and labor-saving way to detect heat strain in dairy cows, it
is only applicable to lactating cows [
90
]. At the same time, MT may be less effective to
inform cooling decisions in the early stage due to less frequent measurements. Instead, MT
is more like a quality controller to assure that the intensive cooling protocol implemented
before milking truly works. In addition, bringing cows to the milking parlor could cause
additional stress resulting in an increased body temperature [96].
Finally, the nasal submucosal temperature measured through biothermal sensors was
highly correlated with THI (p< 0.01) and increased more rapidly than RT [
97
]. However,
no more efforts were found, which may have been due to the huge fluctuation.
2.2. Body Surface Temperature
When dairy cows first feel discomfort during heat exposure, vasodilation allows more
blood to divert from the core to the skin, which increases the body surface temperature
(BST), allowing for increased heat exchange with the environment [
41
]. Although CBT
seems to be the most valuable parameter for assessing heat strain and only 6% of the
studies considered BST [
18
], the application of BST has been increasingly mentioned in
recent years as its non-invasive measurement is more in line with the requirements of
animal welfare and PLF [
98
–
100
]. The latest measurement methods and thresholds of BST
are listed in Tables 1and 2, respectively.
There are multiple ways to measure BST, among which infrared thermometry (IRT)
is the most popular one due to its low cost and the fact that it does not require direct
contact [
101
]. Infrared thermometry can be used in both portable and fixed ways. For
portable use, an infrared camera or infrared gun can be held to measure data according
to users’ needs [
37
,
45
,
102
,
103
]. However, manual operation limits the frequency of data
collection, and therefore may be not applicable to achieve real-time measurement in dairy
farms. For fixed use, an infrared camera can be fixed at a specific location (e.g., milking
parlor, calf feeder, water station) [
46
,
104
,
105
]. In the study by Hoffmann et al. [
105
],
the regions of interest were manually defined. The introduction of the feature tracking
technique helped automatically locate the regions of interest, and therefore increased the
efficiency [
46
]. Using RFID technology, an automated and wireless collection of IRT data
from cattle was successfully achieved [
104
]. Generally, high-accuracy infrared cameras are
relatively expensive, whereas affordable infrared guns may be not reliable in results. In
addition, the results of IRT are affected by the dirt on the surface, the skin fold, the direct
sun exposure, the distance and angle to the object animal, etc. [
7
]. It has been reported
that BSTs from IRT measured at eyes (r = 0.58) [
106
] and forehead (r = 0.557) [
45
] have the
highest correlation with RT. More importantly, the presence of liquid (e.g., sweat) along
the surface of the skin will significantly underestimate the real BST [
107
], indicating that
additional image processing is needed when animals are under heat stress.
In addition to IRT, BST can also be measured through wearable devices. Wearable
devices equipped with wireless transmission and positioning technologies are of particular
importance in the setting of grazing pastures due to the ability to remotely monitor the
thermal comfort of individual animals or the entire herd. Data loggers (iButton DS1922L,
Maxim Integrated, San Jose, CA, USA) were placed on the skin above the tail vein opposing
the rectum using flexible tape [
108
]. Thermocouples and thermistors were also fixed to the
skin to measure BST [
109
–
111
]. The results of the thermistors (EU-UU-VL5–0, Grant Instru-
ments, Cambridge, UK) and data loggers (iButton DS1922L, Maxim Integrated, Sunnyvale,
Animals 2021,11, 980 10 of 24
CA, USA) were demonstrated to have a very good agreement when measuring BST [
107
].
However, wearable devices are difficult to fix on the skin surface and prone to displacement,
and the results recorded were easily affected by the hair and air movement [
40
]. To mini-
mize the influence of wind and hair, Kou et al. [
110
] designed an automated measurement
for BST by using a shell shape to fix the thermistor appropriate to the metatarsus of the
hind leg with less hair.
To further estimate the relationship between BST and other physiological indicators,
some studies have intended to predict standard indicators (i.e., RT, respiration rate (RR))
based on BST [
110
]. Recently, machine learning models (e.g., artificial neural networks
(ANNs)) have been used to enhance the predictive ability from BST to RT and RR in
beef cattle and dairy cattle [
112
,
113
]. Sousa et al. [
112
] yielded a better coefficient of
determination of 0.72 from an ANN model compared with a linear regression model (0.57)
when predicting RT. Similarly, Pacheco et al. [
113
] also found that ANN-based models
yielded higher coefficients of determination (0.71 for RT and 0.74 for RR) than the linear
regression models (0.55 for RT and 0.69 for RR). These efforts provide powerful tools
to predict core parameters which are important physiological indicators not only in the
detection of heat strain.
To better and earlier manage heat strain, some studies have attempted to develop
thresholds based on BST directly. Unlike CBT, BSTs vary apparently with the ambient
environment and no plateau stage exists. Therefore, attention should be paid to the
point where BST begins to rise more rapidly. Peng et al. [
45
] used segmented regressions
developing thresholds of THI for forehead temperature. Mean and maximum forehead
temperatures began to increase rapidly when THI reached 71.4 and 66.8, respectively, which
were lower than the THI threshold for RT (74.1). The corresponding thresholds of mean
and maximum forehead temperatures were 30.05 and 30.34
◦
C, respectively. Kim et al. [
56
]
developed a series of THI thresholds for five body regions (eyes, hindquarters, nose, part
of horns, and ears) based on data from eight Hanwoo heifers. The THI thresholds ranging
from 65 (eyes) to 70 (hindquarters) were obtained as the point where BSTs surpassed CBT.
Dado–Senn et al. [
49
] also utilized a segmented regression to develop the threshold of BST
in preweaning dairy calves. However, no breakpoints were found due to the strong linear
relationship to THI and the authors suggested that the breakpoint of BST could exist in a
much lower range of THI.
2.3. Respiration-Based Indicators
Respiration-based indicators (i.e., RR, panting score (PS)) are considered to be relevant
indicators for assessing heat strain in dairy cows just like RT. Indeed, panting, along
with sweating, contributes to increased evaporative heat losses that are used to maintain
homeothermy in cattle. More importantly, the respiratory system starts to help dissipate
heat before CBTs measured rectally or vaginally begin to rise [
8
,
61
,
114
,
115
]. Respiration-
based indicators may be the most appropriate indicators to early monitor heat strain in
dairy cows due to high sensitivity to stressors and cost-effective measurement [
115
]. The
latest measurement methods and thresholds of respiration-based indicators are listed in
Tables 3and 4, respectively.
Animals 2021,11, 980 11 of 24
Table 3. Summary of the state-of-the-art measurement methods for respiration-related data.
Indicator Methodology Technology Automatic Continuous Real-Time Accuracy Reference
Respiration rate
Temperature changes
around nostrils during
breathing
Infrared thermography Yes Yes Yes 8.4 ±3.4 (mean ±
SD) BPM lower
Jorquera–Chavez
et al. [46]
Thermistor sensor Yes Yes No ±2 BPM for 80% of
the times Milan et al. [116]
Chest and abdominal
expansion associated with
breathing
Pressure sensor Yes Yes No ±2–3 BPM Eigenberg et al.
[117]
±0.47 BPM Atkins et al. [53]
Pressure changes around
nostrils during breathing Pressure sensor Yes Yes No
Mean difference
(BPM): −0.2 when
dozing; 0.2 when
lying; 1.4 when
standing
Strutzke et al.
[118]
Breathing sounds Acoustic processing Yes Yes No Mean bias: 2.75
BPM de Carvalho et al.
[119]
Distance change of
abdominal movement
during breathing
Laser distance sensor Yes Yes No
Mean difference of
6 BPM when
measurement rate
is 100 Hz
Pastell et al. [120]
Body movements due to
respiration Accelerometer-based
neck collar Yes Yes No Unknown Davison et al.
[121]
The change in respiration
volume Spirometer Yes Yes Yes Unknown de Melo Costa
et al. [122]
Panting
Panting scoring system Visual scoring No No No Not applicable
Mader et al. [59];
Gaughan et al.
[88]; Lees et al.
[123]
Body movements due to
panting Accelerometer-based
ear tag and neck collar Yes Yes Yes Positive predictive
value = 79%
Bar et al. [8];
Islam et al. [124]
SD = standard deviation; BPM = breaths per minute; Hz = Hertz.
Table 4. Summary of the thresholds of respiration-related data.
Indicator Threshold for
Animal-Based Indicators
Threshold for
Environment-Based
Indicators
Cow Information Thermal Condition Reference
Respiration rate
48 BPM THI a70 High-producing
Holstein dairy cows
Hot climate with a Ta
range of 9.5 to 30.8 ◦CLi et al. [47]
37, 39 BPM THI a70, 65
Standing and lying
high-producing
Holstein-Friesian dairy
cows
Continental climate
with a THI range of 17.8
to 85.1 Pinto et al. [19]
60 BPM Unknown
Lactating
Holstein-Friesian dairy
cows
Three summers and two
winters in Queensland,
Australia Lees et al. [123]
30 BPM BGHI 73.61 1⁄2Holstein dairy cows BGHI stayed higher
than 72 Dalcin et al. [48]
45 BPM BGHI 72.29 3⁄4Holstein dairy cows BGHI stayed higher
than 72 Dalcin et al. [48]
Unknown THI a70 Lactating Holstein cows Unknown Atkins et al. [53]
61 BPM Unknown Dry cows Unknown Toledo et al. [125]
Unknown THI a65 Preweaning dairy
calves
Averaged THI of 78
with shade only in a
subtropical climate
Dado–Senn et al. [49]
Unknown THI a69 Preweaning dairy
calves
Averaged THI of 78.25
with shade plus fans in
a subtropical climate
Dado–Senn et al. [49]
Unknown THI b82.4 Preweaning Holstein
bull calves
Hot climate with a THI
range from 70.3 to 94 Kovács et al. [50]
BPM = breaths per minute; THI = temperature–humidity index; T
a
= ambient temperature (
◦
C); BGHI = black globe-humidity index.
a
THI
equation from National Research Council [57]; bTHI equation from Bianca et al. [58].
2.3.1. Respiration Rate
The traditional method to measure RR is to count the movements of the flank manually
and convert it into breaths per minute (BPM). Obviously, such a visual observation is both
time- and labor-consuming [
118
]. In addition, the interaction between people and cows
when measuring may bias the result [
116
]. Many efforts have been made recently with the
Animals 2021,11, 980 12 of 24
hope to design automatic measurement methods for RR. Generally, these new methods can
be categorized into wearable and non-wearable devices.
Most of the automatic measurement methods are wearable devices, which are mounted
to a specific location of the cow (e.g., face, flank, chest, neck). Respiration-related electrical
signals are collected based on different methodologies and further converted into RR.
Eigenberg et al. [
117
] developed a thoracic belt for cows based on a commercially available
device for humans. The device consists of a force transducer deriving electrical signals
from rib cage movement, and a small battery-powered micro-computer further converting
the signal into RR. The result of this device differed from the result of manual counting
by 2–3 BPM, probably due to shallow breathing and other confounding movements. Its
memory could collect RR data for four consecutive days under the condition of collecting
1 min every 15 min. In addition, the device was reported to slip off frequently [
118
]. More
recently, Atkins et al. [
53
] designed a device for research based on a similar methodology as
Eigenberg et al. [
117
]. Sensors were equipped on eight lactating Holstein cows to detect the
breathing-related abdominal expansion. Two methods (peak-to-peak and fast Fourier trans-
form) were used to derive RR from the signal. Milan et al. [
116
] developed a halter based
on thermistor to measure RR according to the temperature changes around nostrils during
breathing and found that the results of the device were slightly lower than that of manual
counting in four out of five cows. However, they did not validate its reliability when cows
were heat-stressed, nor did they achieve a real-time measurement. It can be inferred that
this method may be of limited use when the AT is close to exhaled air temperature since no
pulses in temperature would then be detected. Strutzke et al. [
118
] also designed a halter
that can automatically detect the RR of cows based on the pressure difference between in-
halation and exhalation of air. This method was also expected to provide information about
breath depth. A wireless local area network was applied to achieve wireless transmission.
The results show no significant difference (p> 0.05) between the manual counting and test
device when cows were dozing or lying. However, the device tended to overestimate the
RR with a mean difference of 1.4 BPM when the cows were standing (p< 0.05). The authors
explained that this could be due to incomplete manual counting. Their device also had
many limitations, e.g., short battery life, loss of devices, and the unknown reliability of the
device in hot conditions as well. In addition, although there was no adverse event observed
during the study, the need to insert a flexible tube 10 cm deep into the nasal cavity may
cause some problems since cattle like to lick noses [
126
]. A facial mask equipped with an
indirect calorimetry system is very important for research purposes due to the ability to
measure comprehensive respiratory characteristics, including RR, tidal volume, and the
composition of exhaled breath [
122
]. Recently, an accelerometer-based neck-mounted collar
was further developed to measure RR [
121
]. Using Fourier transform, the respiration could
be distinguished from rumination, and RR could be obtained directly from the frequency
domain. In addition, the amplitude may be indicative of breath depth. More recently, an
acoustic method was developed by using an MP3 audio recorder mounted to the cow’s
halter [
119
]. No significant difference (p> 0.05) was found between RR measured by the
acoustic and visual method in both heifers and lactating cows. However, the RR was
retrospectively analyzed after data collection and the battery could only last up to two days
of continuous work. To sum up, considering that almost no wearable device was reported
to interfere with cows’ normal behavior, wearable devices are expected to continuously
measure RR in real-time in dairy production if battery life, wearing stability, and remote
wireless transmission can be further improved.
Among the non-wearable devices of measuring RR, the temperature changes around
nostrils during breathing can also be captured by hand-held IRT cameras and this method
has already been validated in both calves and adult dairy cattle [
127
,
128
]. The mean
difference between IRT and manual counting was 0.83 BPM in adult cattle and 0.02 BPM in
calves. More recently, Jorquera–Chavez et al. [
46
] used fixed thermal camera and computer
vision algorithms to calculate RR based on the changes in pixel intensity within regions of
interest. Again, the systemic defect that the AT and exhaled air temperature were difficult
Animals 2021,11, 980 13 of 24
to distinguish in a hot environment makes this method only suitable for specific cool indoor
areas, such as milking parlors. Pastell et al. [
120
] designed a novel method based on a laser
distance sensor. The laser rangefinder (L-Gage LT3, Banner, Minneapolis, MN, USA) was
fixed on the side of the cow and the RR was calculated from the change in the distance to
the flank detected when the cow was breathing. The mean difference was 6 BPM when the
detection frequency was set to 100 Hz. To sum up, non-wearable devices measuring RR are
not mobile, and therefore would be better to use in the scenarios where cows are separated
and fixed in specific areas (e.g., milking parlors, calf feeder).
A non-heat-stressed cow has a RR range of 26–50 BPM [
121
] and heat stress would
significantly increase RR [
26
,
129
]. Eigenberg et al. [
117
] reported that RR fluctuated less
significantly during the thermoneutral period (18
±
7
◦
C) (45
±
16 BPM) compared with the
heat stress period (>25
◦
C) (78
±
26 BPM), indicating a nonlinear response of RR to AT. A
consistent trend was found by Li et al. [
47
] in a field study with 45 healthy high-producing
Holstein cows in Beijing where the AT averaged 22.0
◦
C (9.5–30.8
◦
C) and RH averaged
61.8% (37.9–82.2%). They found a greater R
2
(0.916) after introducing a quadratic term
of AT. A threshold of RR (48 BPM) where cows began to enter heat strain was further
determined and the corresponding AT was 20.4
◦
C. Pinto et al. [
19
] also found RR increased
faster with a higher positive slope as THI increased. Broken-line models were applied and
RR thresholds of 37 BPM for standing cows and 39 BPM for lying cows were determined
by choosing the models with the lowest Akaike information criterion. The corresponding
THI thresholds were 70 and 65 for standing and lying cows, respectively. A lower RR
threshold (60 BPM) was determined for lactating dairy cows based on the occurrence of
characteristic panting behaviors (slight panting, apparent chest movement, no drool, and
closed mouth) [
123
]. Using segmented regression models, Dado–Senn et al. [
49
] determined
THI thresholds (65, 69) for RR in preweaning dairy calves under shade only and shade
plus fans, respectively. In the study by Dalcin et al. [
48
] where similar broken-line models
were used, BGHI thresholds (73.61, 72.29) for RR were determined for
1⁄2
and
3⁄4
Holstein
cows, respectively. A recent study by Toledo et al. [
125
] summarized 6 different dry-period
experiments and concluded that RR over 61 BPM can be used as a threshold to indicate
heat strain in dry cows.
An RR ceiling was reported by many studies, which could be due to a shift in the
respiratory pattern from rapid and shallow breathing (reduced tidal volume) to slower and
deeper breathing (increased tidal volume) [
5
,
61
,
130
]. This process is necessary to exhaust
vaporized moisture in the lung and cope with respiratory alkalosis following prolonged
panting [
8
]. This biphasic pattern is also supported by the panting scoring system adapted
by Lees et al. [
123
] where RR may decrease during the most severe situation of panting.
The critical CBT threshold for the shift in the respiratory pattern was estimated between
40.9 and 41.4
◦
C for Angus steers [
130
]. A close threshold (CBT of 40.6
◦
C) was found for
Holstein cows where the probability of heavy breathing began to decrease [
8
]. To detect
this change in respiratory pattern, it is of importance to measure both the speed (RR) and
width (tidal volume) of breathing. In addition, RR must be measured every 30 min to
provide an accurate reflection of heat strain [131].
2.3.2. Panting Score
Panting-as an important physiological pathway of evaporative heat loss—has long
been proposed to describe heat strain in dairy cows [
132
]. The main panting characteris-
tics include increased RR, chest movement, drooling, mouth opened, tongue protruding,
and neck extended [
123
]. With the accumulation of heat load in dairy cattle (manifested
as a rising RR), cows first exhibited drooling, following by open mouth and protruding
tongue [
131
]. PS has been developed and refined as a numerical scale to grade the obvi-
ous respiratory dynamics and behaviors [
59
,
88
,
123
,
124
]. A recent study utilized a score
consisting of RR and various panting characteristics to assess heat strain in grazing dairy
cows [
133
]. This score is very similar to PS, using a 0 to 4 scale, where 0 represents that
the cow is under no heat strain and 4 represents that the cow is moribund. The PS per
Animals 2021,11, 980 14 of 24
se, as a scoring system, has a good ability to assess heat strain in cattle. Existing panting
scoring systems often provide corresponding RR thresholds, e.g., PS 1 (slight panting)
corresponding to a RR of 60 BPM [
123
]. Thermal indices (heat load index (HLI) and dairy
heat load index (DHLI)) have been developed based on PS due to its high association with
the thermal environment [88,123].
To measure PS, trained observers would rate a cow’s respiratory dynamics according
to the numerical scale, and the process would take about 10 s per observation per cow.
Although measuring PS is relatively easier than RR, the inter-observer difference can be
an important source of error. Accelerometer-based ear tags and neck collars have been
designed to measure a series of behaviors including “heavy breathing”. The capacity of a
commercially available collar has been validated to accurately reflect heat strain in dairy
cows [
8
]. “Heavy breathing” detected by this tag is defined by characteristic movements
(e.g., forward-backward heaving, increased RR, increased chest movement). Their results
show that the proportion of cows breathing heavily was consistent with the change of
mean CBT in both dry and lactating cows. An accelerometer-based ear tag from the
same manufacturer was also validated in steers [
124
]. The visual measurement of PS and
sensor-recorded “heavy breathing” were compared. The sensor was designed to give a
single output within seven behaviors for every single minute, probably leading to a high
false-negative rate by rating breathing heavily cows as other more prominent behaviors.
Although data processing was conducted to mitigate this systematic underestimation,
low sensitivities were still yielded for all PS data, PS1, and PS2 (59%, 54%, and 82%,
respectively). The higher sensitivity for PS2 indicates that mild symptoms were hard to
detect. In fact, the determination of “heavy breathing” in these devices is most likely to
be based on a default threshold of the frequency of respiratory movements. Therefore,
the original output of these devices is probably much closer to RR, and “heavy breathing”
can be seen as the extended results of RR. As proposed by both Bar et al. [
8
] and Islam
et al. [
124
], these accelerometer-based sensors have difficulties in detecting slight panting
(PS1) and only give a binary result. Although the detection of panting or not is enough
for the start of heat abatement protocol, more sophisticated classification would be more
informative for customized interventions. Further studies can therefore focus on increasing
the sensitivity and classification of the sensor.
3. Comparison and Future Development of Detection Methods
This section first compares candidate animal-based indicators from the perspectives
of measurement feasibility and sensitivity to heat stress. Next, the issues that need careful
attention in the process of threshold establishment are discussed and summarized. Last, a
comprehensive assessment method to help with the precision management of heat strain
is proposed.
3.1. Measurement Feasibility of Different Animal-Based Indicators
In pursuit of PLF and animal welfare, an ideal measurement method for heat strain-
related data should be automatic, accurate, continuous, remote, non-invasive, low-cost,
and real-time [
14
,
99
,
100
]. Non-contact measurements based on infrared thermometry are
non-invasive to dairy cows but difficult to achieve real-time measurement at the herd level.
They are more suitable for monitoring at specific locations such as calf islands, calf feeders,
and milking parlors.
Indwelling devices measuring CBT rectally and vaginally cannot work for a long time
due to the interference with normal physiological activities, making them more appropriate
for short-term temperature monitoring. Although implantable devices can measure CBT
subcutaneously, their cost and long-term safety should be further examined. Notably, ear
canal sensors and rumen boluses are more appropriate for being less invasive. However,
commercially available products should be further validated on the function of detecting
heat strain in dairy cows.
Animals 2021,11, 980 15 of 24
Wearable devices are most promising since they can measure BST and RR non-
invasively in real-time, and can further forecast CBT through machine learning algorithms.
BST is more difficult to measure relative to RR due to various influencing factors. Cur-
rently, few wearable devices on the market can measure BST and RR. Some commercially
available accelerometer-based ear tags and collars can measure “heavy breathing” [
8
,
124
].
In practical farms, there is a need for measuring comprehensive respiratory dynamics
including both frequency and depth. These functions should preferably be integrated
into existing commercial devices. In the future, the battery life, biocompatibility, stability,
remote transmission, and accuracy of devices should be improved.
3.2. Sensitivity of Different Animal-Based Indicators to Heat Stress
As discussed above, many candidate physiological indicators may reflect heat strain in
dairy cows. However, these animal-based indicators play different roles in different stages
of coping with heat stress. Some of them (e.g., BST and RR) represent the effort made by the
thermoregulatory mechanism while CBT is more like the result of thermoregulation [
134
].
From a thermodynamics point of view, homeotherms take two ways to dissipate body
heat, which are sensible heat loss (non-evaporation) and latent heat loss (evaporation). The
former consists of conduction, convection, and radiation. The latter consists of respiratory
evaporation and cutaneous evaporation, where cutaneous evaporation accounts for up to
80% of total latent heat loss in lactating Holstein cows raised in a tropical environment [
4
].
However, cutaneous evaporation is more difficult to measure in practice than respiratory
evaporation. When AT increases beyond the upper critical temperature of TNZ, the non-
evaporative ways of heat loss appear to be far less efficient [
1
]. When the AT is close to or
higher than 32
◦
C, Holstein cows even start to gain heat from the ambient environment
through sensible heat transfer [
3
]. As the thermal environment continues to deteriorate,
cows can succumb to hyperthermia with an increased CBT at some point, where they fail
to maintain thermoneutrality [1].
From a physiological point of view, when the AT rises in the first place, homeotherm
will take three main physiological procedures to increase heat dissipation, i.e., vasodilata-
tion, sweating, and panting. The first one is to drive more blood from the core to the skin
and peripheral to increase the BST so that promote sensible heat loss. The remaining two
procedures are to promote evaporation [
1
,
5
]. Only when all these preliminary efforts fail to
dissipate excess body heat would CBT increase abnormally [
114
,
125
]. When the increase in
BST and RR is effective to keep thermoneutrality, RT can remain unchanged [135].
This temporal sequence responding to heat stress is supported by many studies that
BST and RR increased at lower THI threshold relative to CBT, and BST responded even
before RR due to direct contact with the ambient environment [
19
,
45
,
48
,
49
]. In addition, RR
started to increase approximately an hour earlier than RT [
61
]. Panting responded about
15 min earlier than CBT both when increasing and decreasing [
8
]. This evidence supports
the increase in CBT as a consequence of failed thermoregulation.
Indeed, the increased energy demanded by the preliminary physiological responses
(e.g., panting, sweating) in addition to the reduction in energy intake (decreased feed intake)
lead to a decrease in the energy used for production [
136
,
137
]. Cooling measures are used
to reduce the extra energy used by cows to maintain thermoneutrality, thereby consistently
maximizing production performance. Since an increased CBT already represents failed
thermoregulation, it might be reasonable to advance the detection by using more sensitive
indicators so that cooling measures can be taken as early as possible. Providing cooling
supports when the cow is still able to dissipate heat through its thermoregulation system
may be more efficient. It has been reported that providing convective cooling when the
temperature gradient from the skin to the surrounding environment was larger could better
reduce heat strain and yield higher milk production [138].
Recently, there has been a trend to use more sensitive animal-based parameters as
indicators [
19
,
49
]. A study by Dalcin et al. [
48
] argued that RR was most suitable to indicate
heat strain in dairy cattle compared with RT and heart rate due to the highest sensitivity
Animals 2021,11, 980 16 of 24
to the change in the thermal index. Amamou et al. [
129
] also proposed that BST and RR
might be better early predictors for heat strain due to more significant variability with THI
compared with RT and production traits. Since the BST is highly linearly related to AT, it
may be difficult to find a valuable inflection point. Future studies should compare BST
with RR to determine the best point to implement interventions.
3.3. Selection of Thermal Indices Suitable for Specific Environments
Although AT showed close or even better correlations with animal-based indicators
than thermal indices in some studies [
11
,
47
,
121
,
139
,
140
], a comprehensive index is still
indispensable to represent the thermal environment to which animals are exposed for most
of the time when developing thresholds. Thermal indices mentioned in this review were
summarized in Table 5. Obviously, their formulas vary from simple indices like THI to
complex indices like CCI. For more information on thermal indices, see [12,13].
Table 5. Summary of the thermal indices mentioned in this review.
Thermal Index Formula Resource
Temperature–humidity
index (THI)
THI = (1.8 ×Ta+ 32) −(0.55 −0.0055 ×RH) ×(1.8 ×Ta−26.8) National Research
Council [57]
THI = (0.35 ×Ta+ 0.65 ×Tw)×1.8 + 32 Bianca et al. [58]
THI = 0.8 ×Ta+ (RH/100) ×(Ta−14.4) + 46.4 Mader et al. [59]
Black globe-humidity
index (BGHI) BGHI = Tbg + 0.36 ×Tdp + 41.5 Buffington et al. [141]
Comprehensive climate
index (CCI)
CCI =Ta+Equation (1)+Equation (2)+Equation (3)
Equation (1)=e(0.00182×RH+1.8×10−5×Ta×RH)×0.000054 ×Ta2+0.00192 ×Ta−0.0246×(RH −30)
Equation (2)=−6.65
e
(1
(2.26×WS+0.23)0.45×(2.9+1.14×10−6×WS2.5 −log0.3 (2.26×WS+0.33)−2))−0.00566 ×WS2+3.33
Equation (3)=0.0076 ×RAD −0.00002 ×RAD ×Ta+0.00005 ×Ta2×√RAD +0.1 ×Ta−2
Mader et al. [142]
Heat load index (HLI) HLI (Tbg < 25) = 10.66 + 0.28 ×RH + 1.9 ×Tbg −WS
HLITbg >25=8.62 +0.38 ×RH +1.55 ×Tbg −0.5 ×WS +e2.4−WS Gaughan et al. [88]
Dairy heat load index
(DHLI) DHLI =1.681813 ×1+e−(−8.50749+0.206159×Tbg+4.088399 ×RH)−1Lees et al. [123]
T
a
= ambient temperature (
◦
C); RH = relative humidity; T
w
= wet bulb temperature (
◦
C); T
bg
= black globe temperature (
◦
C);
Tdp = dewpoint temperature (◦C); WS = wind speed (m/s); RAD = radiation (W/m2).
As pointed out in other reviews, there is no recognized best index at present
[12,13]
.
In a recent study by Yan et al. [
143
], nine typical cattle-related thermal indices have been
compared using 273 lactating Holstein-Friesian dairy cows in a naturally ventilated com-
mercial barn in Jiangsu, China. The results show that some indices could not even reflect
the response of physiological indicators (RR, VT) to heat stress, and CCI was determined
as the best thermal index. Similarly, CCI was found to have the highest correlation with
CBT by Jara et al. [
85
]. Kaufman et al. [
37
] found simple thermal indices like THI only had
a weak correlation with physiological indicators which could be due to the lack of wind
speed and other management factors. Dado-Senn et al. [
49
] reported that wind speed was
an important predicting factor of animal-based indicators in calves exposed to fans. In
addition, THI with larger weights on humidity were more suitable for use in areas with
higher humidity, and vice versa [
144
]. Generally, different indices are suitable for specific
environments, especially in places similar to where the indices were originally established.
Inappropriate use of thermal indices may obscure real physiological responses. Therefore,
future studies should comprehensively compare and select the most appropriate thermal
index in specific environments.
An important reason for poor extrapolation is due to the fact that the thermal indices
of dairy cows are mostly modeled by empirical statistical models. The lack of consideration
of the heat transfer mechanism between dairy cows and the thermal environment results
in insufficient performance when the thermal indices are extrapolated to a non-modeled
environment [
145
]. The thermal index based on the rational model may have better
Animals 2021,11, 980 17 of 24
adaptability and extrapolation due to the comprehensive consideration of the heat balance
equation. However, THI are still the most welcomed thermal indices in both experimental
and practical conditions probably due to easy access and acceptable performance.
3.4. Development of Critical Thresholds
The most commonly used models to describe the relationship between environmental
indicators and animal-based physiological indicators are piecewise regression models and
polynomial models. Although polynomial models or other nonparametric models (e.g.,
generalized additive model) may bring better fitting results, they cannot directly provide
easy-to-interpret slope estimates and thresholds. In contrast, the result of the piecewise
function is more explanatory and can provide more useful information, including the
threshold, the slope before the threshold, and the slope after the threshold. By introducing
more phases into the piecewise model, more thresholds corresponding to different severity
of heat strain could be identified, providing more options for cooling.
In the previous studies, the consideration of the plateau period is quite different when
using piecewise functions. The assumption of the plateau could have a great impact on
the results. Li et al. [
47
] and Peng et al. [
45
] collected data from the same dairy farm and
both used segmented regression to obtain heat strain thresholds for RT in dairy cows. The
former assumed the existence of a plateau while the latter did not. The THI thresholds were
70 and 74.1, respectively. As the experimental condition and animal status were almost
identical, it can be inferred that the introduction of the plateau period could lower the
THI threshold to a certain extent. In fact, the decision should base on the changing law of
specific physiological indicators. For example, it is suitable to use a plateau to fit CBT as
the normal CBT of homeotherms fluctuates within a narrow range. Other physiological
indicators that are more sensitive to the thermal environment, such as BST and RR, may
not have a significant plateau.
In addition, the thermal condition (THI range) during the testing days is of key
importance to the results. For example, Dado-Senn et al. [
49
] and Kovács et al. [
50
] both
used piecewise regression models to determine the THI thresholds for RT, RR, and BST in
dairy calves. However, the THI thresholds of the former were between 65 and 69, while
the THI thresholds of the latter were between 82.4 and 88.1. It is mostly because that the
former was conducted under a broader range of thermal exposures (THI 60–85), while the
environment during the latter study was much hotter (the average THI was greater than
80 for three out of four days). Therefore, the thresholds obtained by the latter may be the
points at which heat strain was aggravated rather than triggered.
Many studies only measured heat strain-related data once or twice a day, which
may affect the accuracy of the threshold. If the physiological indicators change more
frequently than the frequency of measurements, it is difficult to observe the real effect [
28
].
Continuous measurement is not only beneficial for threshold acquisition but also beneficial
for early detection of heat strain. A broader range of thermal exposure and more frequent
measurements are therefore needed for future studies on early detection of heat strain.
Various factors have an impact on the sensitivity to heat stress, including internal
animal factors (e.g., age, breed, productivity, lactation stage, parity, body posture, circadian
rhythm, previous thermal exposure, acclimation) [
1
,
19
,
27
] and external management factors
(e.g., milking, cooling, nutrition, shading) [
32
,
49
,
146
]. Therefore, it is difficult to assess the
heat strain status of a herd based on a generic threshold. To deal with these dilemmas,
efforts have been made to adjust or customize the thresholds according to various animal
and management factors. The HLI was developed for feedlot beef cattle by Gaughan
et al. [
88
]. Genotype, coat color, health status, acclimatization, shade, days on feed, manure
management, and drinking water temperature were considered when determining the
HLI threshold for reference animals and further threshold adjustments. Other studies also
tried to customize heat strain threshold according to productivity [
55
], posture [
19
], or
cooling strategy [
49
]. Notably, the difference in sensitivity to heat stress manifests both
in environmental and physiological levels. For instance, high-producing cows may enter
Animals 2021,11, 980 18 of 24
heat strain in a lower THI and a higher RT [
55
,
95
]. Further studies are required to evaluate
and customize heat strain thresholds based on different internal and external influencing
factors that have an impact on the sensitivity to heat stress.
Most of the present critical thresholds assume that physiological indicators change
simultaneously with the thermal environment and do not consider the cumulative effect
of heat exposure. It is generally assumed that this accumulative effect only makes sense
after the trigger point of heat strain. For the early detection of heat strain, due to the
sensitivity of the selected indicators, the effect of time may be negligible. However, for the
prediction of long-term effects such as milk production, the duration of heat exposure is an
indispensable factor [
17
]. In addition, when dairy cows have long been exposed to heat
stress, some of them may adapt to heat stress and can better dissipate excess body heat,
manifesting as increased RR and BST, and restored milk yield [
129
]. Therefore, existing RR
and BST thresholds may be exaggerated to some extent if lots of thermotolerant data were
used for thresholds development.
3.5. Individual Monitoring to Help Precision Management
After obtaining the thresholds on both physiological and environmental dimensions,
it is important to assess how to use them in practice. Currently, the most popular method
is to evaluate the heat strain level of the entire herd indirectly through the thermal index
calculated from easily accessible meteorological data, such as data from the nearby weather
station, or the average value from temperature and humidity sensors placed in the barn.
However, given the large differences in the microclimate to which individual dairy cows
are exposed [
147
], a more accurate way is to measure and evaluate the microclimate by
placing more sensors and dividing the barn into smaller measuring and controlling units.
Through the internet of things (IoT) technology, the cooling facilities in a certain unit can
be monitored continuously and the heat strain of the herd can be managed more effectively
and efficiently.
Environmental indicators neither truly reflect how animals respond to the changes in
the thermal environment nor show whether animals are being adequately cooled [
14
]. It has
been reported that both CBT and RR remained high even when THI was low at night [
53
].
In addition, there are large differences in the sensitivity to the thermal environment among
individual animals [
1
,
27
], even within the same subgroup [
124
]. A large amount of energy
used for cooling measures may be wasted on low-risk animals. Therefore, direct monitoring
of animal-based physiological indicators at the individual level should help better manage
heat strain. At present, many large-scale commercial dairy farms have used smart ear
tags or collars to monitor multiple physiological behaviors of individual cows. Combined
with positioning technology (e.g., RFID technology), the nearby cooling facilities can be
turned on automatically when a large proportion of cows breathing heavily in a small area.
Such an IoT system is very promising not only because it could detect heat strain more
accurately and earlier based on animal’s feeling, but also because it could save energy. In
addition, the heat strain screening system in the passage to the parlor could achieve an
automatic separation of heat-stressed cows [148].
Combining environmental indicators and various physiological indicators to evaluate
heat strain from multiple perspectives also has a good prospect. Due to different physio-
logical dynamics, it is possible to distinguish heat strain from other events that may lead to
increased body temperature and accelerated breathing, including exercise [
30
], diseases
(e.g., infectious disease [
91
]), physiological processes (e.g., estrous [
149
]), and management
factors (e.g., bringing cows to the milking parlor [96]).
4. Conclusions
In consideration of the measurement feasibility and the sensitivity to heat stress, we
concluded that BST and RR are the most appropriate indicators for the early detection
of heat strain in dairy cows. Although the thresholds of both physiological and environ-
mental indicators can be obtained through regression models, many studies did not give
Animals 2021,11, 980 19 of 24
corresponding thresholds for physiological indicators probably due to the lack of real-time
measurement for physiological data in actual production. The existing thresholds should
be used carefully due to the differences in experimental conditions and animal information.
Future studies should customize heat strain thresholds according to different internal and
external factors that have an impact on the sensitivity to heat stress. With the development
of wireless transmission technology, sensor technology, and battery technology, wearable
devices are most promising to achieve real-time measurements in practical dairy farms.
Combined with IoT technologies, a comprehensive strategy based on both animal- and
environment-based indicators is expected to increase the precision of early detection of
heat strain.
Author Contributions:
H.S., W.W., and L.G. discussed the topic and generated the outline. H.S.
wrote the original manuscript. W.W. and L.G. reviewed and edited the manuscript. J.B. reviewed
and corrected the final version. All authors have read and agreed to the published version of
the manuscript.
Funding:
This research was funded by Major Science and Technology Program of Inner Mongolia
Autonomous Region, grant number 2020ZD0004; Key Research and Development Plan of Hebei
Province, grant number 20327202D; and Key Research and Development Plan of Hebei Province,
grant number 20326602D.
Institutional Review Board Statement: Not applicable.
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
Kadzere, C.T.; Murphy, M.R.; Silanikove, N.; Maltz, E. Heat stress in lactating dairy cows: A review. Livest. Prod. Sci.
2002
,77,
59–91. [CrossRef]
2.
Ouellet, V.; Cabrera, V.E.; Fadul-Pacheco, L.; Charbonneau, É. The relationship between the number of consecutive days with
heat stress and milk production of Holstein dairy cows raised in a humid continental climate. J. Dairy Sci.
2019
,102, 8537–8545.
[CrossRef] [PubMed]
3.
Maia, A.S.C.; daSilva, R.G.; Loureiro, C.M.B. Sensible and latent heat loss from the body surface of Holstein cows in a tropical
environment. Int. J. Biometeorol. 2005,50, 17–22. [CrossRef] [PubMed]
4.
Maia, A.S.C.; Silva, R.G.d.; Loureiro, C.M.B. Latent heat loss of Holstein cows in a tropical environment: A prediction model. Rev.
Bras. Zootec. 2008,37, 1837–1843. [CrossRef]
5.
Becker, C.A.; Collier, R.J.; Stone, A.E. Invited review: Physiological and behavioral effects of heat stress in dairy cows. J. Dairy Sci.
2020,103, 6751–6770. [CrossRef]
6.
Gorczyca, M.T.; Gebremedhin, K.G. Ranking of environmental heat stressors for dairy cows using machine learning algorithms.
Comput. Electron. Agric. 2020,168, 105124. [CrossRef]
7.
Hoffmann, G.; Herbut, P.; Pinto, S.; Heinicke, J.; Kuhla, B.; Amon, T. Animal-related, non-invasive indicators for determining heat
stress in dairy cows. Biosys. Eng. 2020,199, 83–96. [CrossRef]
8.
Bar, D.; Kaim, M.; Flamenbaum, I.; Hanochi, B.; Toaff-Rosenstein, R.L. Technical note: Accelerometer-based recording of heavy
breathing in lactating and dry cows as an automated measure of heat load. J. Dairy Sci.
2019
,102, 3480–3486. [CrossRef] [PubMed]
9.
Koltes, J.E.; Koltes, D.A.; Mote, B.E.; Tucker, J.; Hubbell, D.S. Automated collection of heat stress data in livestock: New
technologies and opportunities. Transl. Anim. Sci. 2018,2, 319–323. [CrossRef]
10. Neethirajan, S. Transforming the Adaptation Physiology of Farm Animals through Sensors. Animals 2020,10, 1512. [CrossRef]
11.
Dikmen, S.; Hansen, P.J. Is the temperature-humidity index the best indicator of heat stress in lactating dairy cows in a subtropical
environment? J. Dairy Sci. 2009,92, 109–116. [CrossRef] [PubMed]
12.
Wang, X.; Bjerg, B.S.; Choi, C.Y.; Zong, C.; Zhang, G. A review and quantitative assessment of cattle-related thermal indices.
J. Therm. Biol. 2018,77, 24–37. [CrossRef] [PubMed]
13.
Herbut, P.; Angrecka, S.; Walczak, J. Environmental parameters to assessing of heat stress in dairy cattle—A review.
Int. J. Biometeorol. 2018,62, 2089–2097. [CrossRef] [PubMed]
14.
Polsky, L.; Von Keyserlingk, M.A.G. Invited review: Effects of heat stress on dairy cattle welfare. J. Dairy Sci.
2017
,100, 8645–8657.
[CrossRef]
15.
Collier, R.J.; Hall, L.W.; Rungruang, S.; Zimbleman, R.B. Quantifying Heat Stress and Its Impact on Metabolism and Performance.
In Proceedings of the MidSouth Ruminant Nutrition Conference, Gainesville, FL, USA, 1 February 2012; pp. 74–84.
16.
West, J.W.; Mullinix, B.G.; Bernard, J.K. Effects of Hot, Humid Weather on Milk Temperature, Dry Matter Intake, and Milk Yield
of Lactating Dairy Cows. J. Dairy Sci. 2003,86, 232–242. [CrossRef]
17.
Li, G.; Chen, J.; Peng, D.; Gu, X. Short communication: The lag response of daily milk yield to heat stress in dairy cows.
J. Dairy Sci.
2021,104, 981–988. [CrossRef]
Animals 2021,11, 980 20 of 24
18.
Galán, E.; Llonch, P.; Villagrá, A.; Levit, H.; Pinto, S.; Del Prado, A. A systematic review of non-productivity-related animal-based
indicators of heat stress resilience in dairy cattle. PLoS ONE 2018,13, e0206520. [CrossRef]
19.
Pinto, S.; Hoffmann, G.; Ammon, C.; Amon, T. Critical THI thresholds based on the physiological parameters of lactating dairy
cows. J. Therm. Biol. 2020,88, 102523. [CrossRef]
20.
St-Pierre, N.R.; Cobanov, B.; Schnitkey, G. Economic Losses from Heat Stress by US Livestock Industries. J. Dairy Sci.
2003
,86,
52–77. [CrossRef]
21.
Key, N.; Sneeringer, S. Potential Effects of Climate Change on the Productivity of U.S. Dairies. Am. J. Agric. Econ.
2014
,96,
1136–1156. [CrossRef]
22. Berckmans, D. General introduction to precision livestock farming. Anim. Front. 2017,7, 6–11. [CrossRef]
23.
Liu, J.; Li, L.; Chen, X.; Lu, Y.; Wang, D. Effects of heat stress on body temperature, milk production, and reproduction in dairy
cows: A novel idea for monitoring and evaluation of heat stress—A review. Asian Australas. J. Anim. Sci.
2019
,32, 1332–1339.
[CrossRef]
24.
Alzahal, O.; Alzahal, H.; Steele, M.A.; Van Schaik, M.; Kyriazakis, I.; Duffield, T.F.; McBride, B.W. The use of a radiotelemetric
ruminal bolus to detect body temperature changes in lactating dairy cattle. J. Dairy Sci.
2011
,94, 3568–3574. [CrossRef] [PubMed]
25.
Berman, A.; Folman, Y.; Kaim, M.; Mamen, M.; Herz, Z.; Wolfenson, D.; Arieli, A.; Graber, Y. Upper Critical Temperatures and
Forced Ventilation Effects for High-Yielding Dairy Cows in a Subtropical Climate. J. Dairy Sci. 1985,68, 1488–1495. [CrossRef]
26.
Garner, J.B.; Douglas, M.; Williams, S.R.O.; Wales, W.J.; Marett, L.C.; DiGiacomo, K.; Leury, B.J.; Hayes, B.J. Responses of dairy
cows to short-term heat stress in controlled-climate chambers. Anim. Prod. Sci. 2017,57, 1233–1241. [CrossRef]
27.
Spiers, D.E.; Spain, J.N.; Sampson, J.D.; Rhoads, R.P. Use of physiological parameters to predict milk yield and feed intake in
heat-stressed dairy cows. J. Therm. Biol. 2004,29, 759–764. [CrossRef]
28.
Reuter, R.; Carroll, J.; Hulbert, L.; Dailey, J.; Galyean, M. Technical note: Development of a self-contained, indwelling rectal
temperature probe for cattle research. J. Anim. Sci. 2010,88, 3291–3295. [CrossRef]
29.
Lees, A.M.; Lea, J.M.; Salvin, H.E.; Cafe, L.M.; Colditz, I.G.; Lee, C. Relationship between Rectal Temperature and Vaginal
Temperature in Grazing Bos taurus Heifers. Animals 2018,8, 156. [CrossRef]
30.
Debnath, T.; Bera, S.; Pal, P.; Debbarma, N.; Haldar, A. Application of radio frequency based digital thermometer for real-time
monitoring of dairy cattle rectal temperature. Vet. World 2017,10, 1052–1056. [CrossRef]
31.
Lee, C.; Gebremedhin, K.; Parkhurst, A.; Hillman, P. Placement of temperature probe in bovine vagina for continuous measure-
ment of core-body temperature. Int. J. Biometeorol. 2015,59, 1201–1205. [CrossRef]
32.
Kendall, P.E.; Tucker, C.B.; Dalley, D.E.; Clark, D.A.; Webster, J.R. Milking frequency affects the circadian body temperature
rhythm in dairy cows. Livest. Sci. 2008,117, 130–138. [CrossRef]
33.
Vickers, L.A.; Burfeind, O.; Von Keyserlingk, M.A.G.; Veira, D.M.; Weary, D.M.; Heuwieser, W. Technical note: Comparison of
rectal and vaginal temperatures in lactating dairy cows. J. Dairy Sci. 2010,93, 5246–5251. [CrossRef]
34.
Burfeind, O.; Suthar, V.S.; Heuwieser, W. Effect of heat stress on body temperature in healthy early postpartum dairy cows.
Theriogenology 2012,78, 2031–2038. [CrossRef] [PubMed]
35.
Burdick, N.C.; Carroll, J.A.; Dailey, J.W.; Randel, R.D.; Falkenberg, S.M.; Schmidt, T.B. Development of a self-contained, indwelling
vaginal temperature probe for use in cattle research. J. Therm. Biol. 2012,37, 339–343. [CrossRef]
36.
Tresoldi, G.; Schütz, K.E.; Tucker, C.B. Sampling strategy and measurement device affect vaginal temperature outcomes in
lactating dairy cattle. J. Dairy Sci. 2020,103, 5414–5421. [CrossRef] [PubMed]
37.
Kaufman, J.D.; Saxton, A.M.; Ríus, A.G. Short communication: Relationships among temperature-humidity index with rectal,
udder surface, and vaginal temperatures in lactating dairy cows experiencing heat stress. J. Dairy Sci.
2018
,101, 6424–6429.
[CrossRef] [PubMed]
38.
Sakatani, M.; Sugano, T.; Higo, A.; Naotsuka, K.; Hojo, T.; Gessei, S.; Uehara, H.; Takenouchi, N. Vaginal temperature measurement
by a wireless sensor for predicting the onset of calving in Japanese Black cows. Theriogenology
2018
,111, 19–24. [CrossRef]
[PubMed]
39.
Wang, S.; Zhang, H.; Tian, H.; Chen, X.; Li, S.; Lu, Y.; Li, L.; Wang, D. Alterations in vaginal temperature during the estrous cycle
in dairy cows detected by a new intravaginal device—A pilot study. Trop. Anim. Health Prod. 2020,52, 2265–2271. [CrossRef]
40.
Iwasaki, W.; Ishida, S.; Kondo, D.; Ito, Y.; Tateno, J.; Tomioka, M. Monitoring of the core body temperature of cows using
implantable wireless thermometers. Comput. Electron. Agric. 2019,163, 104849. [CrossRef]
41.
Chung, H.; Li, J.; Kim, Y.; Van Os, J.M.C.; Brounts, S.H.; Choi, C.Y. Using implantable biosensors and wearable scanners to
monitor dairy cattle’s core body temperature in real-time. Comput. Electron. Agric. 2020,174, 105453. [CrossRef]
42.
Nogami, H.; Arai, S.; Okada, H.; Zhan, L.; Itoh, T. Minimized Bolus-Type Wireless Sensor Node with a Built-In Three-Axis
Acceleration Meter for Monitoring a Cow’s Rumen Conditions. Sensors 2017,17, 687. [CrossRef] [PubMed]
43.
Mahendran, S.A.; Booth, R.; Burge, M.; Bell, N.J. Randomised positive control trial of NSAID and antimicrobial treatment for calf
fever caused by pneumonia. Vet. Rec. 2017,181, 45. [CrossRef] [PubMed]
44.
Setser, M.M.W.; Cantor, M.C.; Costa, J.H.C. A comprehensive evaluation of microchips to measure temperature in dairy calves.
J. Dairy Sci. 2020,103, 9290–9300. [CrossRef] [PubMed]
45.
Peng, D.; Chen, S.; Li, G.; Chen, J.; Wang, J.; Gu, X. Infrared thermography measured body surface temperature and its relationship
with rectal temperature in dairy cows under different temperature-humidity indexes. Int. J. Biometeorol.
2019
,63, 327–336.
[CrossRef] [PubMed]
Animals 2021,11, 980 21 of 24
46.
Jorquera-Chavez, M.; Fuentes, S.; Dunshea, F.R.; Warner, R.D.; Poblete, T.; Jongman, E.C. Modelling and Validation of Computer
Vision Techniques to Assess Heart Rate, Eye Temperature, Ear-Base Temperature and Respiration Rate in Cattle. Animals
2019
,9, 1089.
[CrossRef] [PubMed]
47.
Li, G.; Chen, S.; Chen, J.; Peng, D.; Gu, X. Predicting rectal temperature and respiration rate responses in lactating dairy cows
exposed to heat stress. J. Dairy Sci. 2020,103, 5466–5484. [CrossRef]
48.
Dalcin, V.C.; Fischer, V.; dos Santos Daltro, D.; Alfonzo, E.P.M.; Stumpf, M.T.; Kolling, G.J.; da Silva, M.V.G.B.; McManus, C.
Physiological parameters for thermal stress in dairy cattle. Rev. Bras. Zootec. 2016,45, 458–465. [CrossRef]
49.
Dado-Senn, B.; Ouellet, V.; Dahl, G.E.; Laporta, J. Methods for assessing heat stress in preweaned dairy calves exposed to chronic
heat stress or continuous cooling. J. Dairy Sci. 2020,103, 8587–8600. [CrossRef]
50.
Kovács, L.; Kézér, F.L.; Póti, P.; Boros, N.; Nagy, K. Short communication: Upper critical temperature-humidity index for dairy
calves based on physiological stress variables. J. Dairy Sci. 2020,103, 2707–2710. [CrossRef] [PubMed]
51.
Nabenishi, H.; Ohta, H.; Nishimoto, T.; Morita, T.; Ashizawa, K.; Tsuzuki, Y. Effect of the Temperature-Humidity Index on Body
Temperature and Conception Rate of Lactating Dairy Cows in Southwestern Japan. J. Reprod. Dev.
2011
,57, 450–456. [CrossRef]
52.
Hillman, P.E.; Lee, C.N.; Willard, S.T. Body Temperature Versus Microclimate Selection in Heat Stressed Dairy Cows.
Trans. ASABE
2005,48, 795–801.
53.
Atkins, I.K.; Cook, N.B.; Mondaca, M.R.; Choi, C.Y. Continuous Respiration Rate Measurement of Heat-Stressed Dairy Cows and
Relation to Environment, Body Temperature, and Lying Time. Trans. ASABE 2018,61, 1475–1485. [CrossRef]
54.
Ammer, S.; Lambertz, C.; Gauly, M. Is reticular temperature a useful indicator of heat stress in dairy cattle? J. Dairy Sci.
2016
,99,
10067–10076. [CrossRef] [PubMed]
55.
Ji, B.; Banhazi, T.; Ghahramani, A.; Bowtell, L.; Wang, C.; Li, B. Modelling of heat stress in a robotic dairy farm. Part 2: Identifying
the specific thresholds with production factors. Biosys. Eng. 2020,199, 43–57. [CrossRef]
56.
Kim, N.Y.; Moon, S.H.; Kim, S.J.; Kim, E.K.; Oh, M.; Tang, Y.; Jang, S.Y. Summer season temperature-humidity index threshold for
infrared thermography in Hanwoo (Bos taurus coreanae) heifers. Asian Australas. J. Anim. Sci.
2020
,33, 1691–1698. [CrossRef]
[PubMed]
57.
National Research Council. A Guide to Environmental Research on Animals; National Academy of Sciences: Washington, DC, USA, 1971.
58.
Bianca, W. Relative Importance of Dry- and Wet-Bulb Temperatures in Causing Heat Stress in Cattle. Nature
1962
,195, 251–252.
[CrossRef] [PubMed]
59.
Mader, T.; Davis, M.; Brown-Brandl, T. Environmental factors influencing heat stress in feedlot cattle. J. Anim. Sci.
2006
,84,
712–719. [CrossRef]
60.
Piccione, G.; Caola, G.; Refinetti, R. Daily and estrous rhythmicity of body temperature in domestic cattle. BMC Physiol.
2003
,3, 7.
[CrossRef] [PubMed]
61.
Gaughan, J.B.; Holt, S.M.; Hahn, G.L.; Mader, T.L.; Eigenberg, R.A. Respiration rate—Is it a good measure of heat stress in cattle?
Asian-Australas. J. Anim. Sci. 2000,13, 329–332.
62.
Burfeind, O.; Von Keyserlingk, M.A.G.; Weary, D.M.; Veira, D.M.; Heuwieser, W. Short communication: Repeatability of measures
of rectal temperature in dairy cows. J. Dairy Sci. 2010,93, 624–627. [CrossRef]
63.
Ekine-Dzivenu, C.C.; Mrode, R.; Oyieng, E.; Komwihangilo, D.; Lyatuu, E.; Msuta, G.; Ojango, J.M.K.; Okeyo, A.M. Evaluating
the impact of heat stress as measured by temperature-humidity index (THI) on test-day milk yield of small holder dairy cattle in
a sub-Sahara African climate. Livest. Sci. 2020,242, 104314. [CrossRef]
64.
Rejeb, M.; Sadraoui, R.; Najar, T.; M’rad, M. A Complex Interrelationship between Rectal Temperature and Dairy Cows’
Performance under Heat Stress Conditions. Open J. Anim. Sci. 2016,6, 24–30. [CrossRef]
65.
Hillman, P.E.; Gebremedhin, K.G.; Willard, S.T.; Lee, C.N.; Kennedy, A.D. Continuous Measurements of Vaginal Temperature of
Female Cattle Using A Data Logger Encased in a Plastic Anchor. Appl. Eng. Agric. 2009,25, 291–296. [CrossRef]
66.
Bergen, R.D.; Kennedy, A.D. Relationship between vaginal and tympanic membrane temperature in beef heifers.
Can. J. Anim. Sci.
2000,80, 515–518. [CrossRef]
67.
Suthar, V.; Burfeind, O.; Maeder, B.; Heuwieser, W. Agreement between rectal and vaginal temperature measured with temperature
loggers in dairy cows. J. Dairy Res. 2013,80, 240–245. [CrossRef] [PubMed]
68.
Ammer, S.; Lambertz, C.; Gauly, M. Comparison of different measuring methods for body temperature in lactating cows under
different climatic conditions. J. Dairy Res. 2016,83, 165–172. [CrossRef] [PubMed]
69.
Hicks, L.C.; Hicks, W.S.; Bucklin, R.A.; Shearer, J.K.; Bray, D.R.; Soto, P.; Carvalho, V. Comparison of Methods of Measuring Deep
Body Temperatures of Dairy Cows. In Proceedings of the 6th International Symposium, Louisville, KY, USA, 21–23 May 2001;
pp. 432–438.
70.
Prendiville, D.J.; Lowe, J.; Earley, B.; Spahr, C.; Kettlewell, P. Radiotelemetry Systems for Measuring Body Temperature; Grange
Research Centre: Tegasc, Ireland, 2002.
71.
Bewley, J.; Einstein, M.; Grott, M.; Schutz, M. Comparison of Reticular and Rectal Core Body Temperatures in Lactating Dairy
Cows. J. Dairy Sci. 2008,91, 4661–4672. [CrossRef]
72.
Ipema, A.H.; Goense, D.; Hogewerf, P.H.; Houwers, H.W.J.; van Roest, H. Pilot study to monitor body temperature of dairy cows
with a rumen bolus. Comput. Electron. Agric. 2008,64, 49–52. [CrossRef]
73.
Lees, A.M.; Lees, J.C.; Lisle, A.T.; Sullivan, M.L.; Gaughan, J.B. Effect of heat stress on rumen temperature of three breeds of cattle.
Int. J. Biometeorol. 2018,62, 207–215. [CrossRef]
Animals 2021,11, 980 22 of 24
74.
Jonsson, N.N.; Kleen, J.L.; Wallace, R.J.; Andonovic, I.; Michie, C.; Farish, M.; Mitchell, M.; Duthie, C.-A.; Jensen, D.B.;
Denwood, M.J.
Evaluation of reticuloruminal pH measurements from individual cattle: Sampling strategies for the assessment of
herd status. Vet. J. 2019,243, 26–32. [CrossRef]
75.
Knauer, W.A.; Godden, S.M.; McDonald, N. Technical note: Preliminary evaluation of an automated indwelling rumen tem-
perature bolus measurement system to detect pyrexia in preweaned dairy calves. J. Dairy Sci.
2016
,99, 9925–9930. [CrossRef]
[PubMed]
76.
Sievers, A.K.; Kristensen, N.B.; Laue, H.-J.; Wolffram, S. Development of an intraruminal device for data sampling and transmis-
sion. J. Anim. Feed Sci. 2004,13, 207–210. [CrossRef]
77.
Burns, P.D.; Wailes, W.R.; Baker, P.B. Changes in reticular and rectal temperature during the periestrous period in cows.
J. Anim. Sci. 2002,80, 128.
78.
Cantor, M.; Costa, J.; Bewley, J. Impact of Observed and Controlled Water Intake on Reticulorumen Temperature in Lactating
Dairy Cattle. Animals 2018,8, 194. [CrossRef]
79.
Liang, D.; Wood, C.L.; McQuerry, K.J.; Ray, D.L.; Clark, J.D.; Bewley, J.M. Influence of breed, milk production, season, and
ambient temperature on dairy cow reticulorumen temperature. J. Dairy Sci. 2013,96, 5072–5081. [CrossRef] [PubMed]
80.
Stone, A.E.; Jones, B.W.; Becker, C.A.; Bewley, J.M. Influence of breed, milk yield, and temperature-humidity index on dairy cow
lying time, neck activity, reticulorumen temperature, and rumination behavior. J. Dairy Sci.
2017
,100, 2395–2403. [CrossRef]
[PubMed]
81.
Hahn, G.L.; Eigenberg, R.A.; Nienaber, J.A.; Littledike, E.T. Measuring physiological responses of animals to environmental
stressors using a microcomputer-based portable datalogger. J. Anim. Sci. 1990,68, 2658–2665. [CrossRef]
82.
Mader, T.; Holt, S.; Hahn, G.; Davis, M.; Spiers, D. Feeding strategies for managing heat load in feedlot cattle. J. Anim. Sci.
2002
,
80, 2373–2382. [CrossRef]
83.
Davis, M.S.; Mader, T.L.; Holt, S.M.; Parkhurst, A.M. Strategies to reduce feedlot cattle heat stress: Effects on tympanic temperature.
J. Anim. Sci. 2003,81, 649–661. [CrossRef]
84.
Arias, R.A.; Mader, T.L.; Escobar, P.C. Climatic factors affecting cattle performance in dairy and beef farms. Arch. Med. Vet.
2008
,
40, 7–22.
85.
Jara, I.E.; Keim, J.P.; Arias, R.A. Behaviour, tympanic temperature and performance of dairy cows during summer season in
southern Chile. Arch. Med. Vet. 2016,48, 113–118. [CrossRef]
86.
McCorkell, R.; Wynne-Edwards, K.; Windeyer, C.; Schaefer, A. Limited efficacy of Fever Tag(
®
) temperature sensing ear tags in
calves with naturally occurring bovine respiratory disease or induced bovine viral diarrhea virus infection. Can. Vet. J.
2014
,55,
688–690.
87.
Richeson, J.T.; Powell, J.G.; Kegley, E.B.; Hornsby, J.A. Evaluation of an Ear-Mounted Tympanic Thermometer Device for Bovine
Respiratory Disease Diagnosis; Arkansas Agricultural Experiment Station Division of Agriculture University of Arkansas System
Fayetteville: Little Rock, AR, USA, 2012; pp. 40–42.
88.
Gaughan, J.B.; Mader, T.L.; Holt, S.M.; Lisle, A. A new heat load index for feedlot cattle. J. Anim. Sci.
2008
,86, 226–234. [CrossRef]
89.
Myers, M.J.; Henderson, M. Assessment of two devices for measuring tympanic membrane temperature in swine, dairy cattle,
and dairy calves. J. Am. Vet. Med. Assoc. 1996,208, 1700–1701.
90.
Lee, Y.; Bok, J.D.; Lee, H.J.; Lee, H.G.; Kim, D.; Lee, I.; Kang, S.K.; Choi, Y.J. Body Temperature Monitoring Using Subcutaneously
Implanted Thermo-loggers from Holstein Steers. Asian Australas. J. Anim. Sci. 2016,29, 299–306. [CrossRef] [PubMed]
91.
Reid, E.D.; Fried, K.; Velasco, J.M.; Dahl, G.E. Correlation of rectal temperature and peripheral temperature from implantable
radio-frequency microchips in Holstein steers challenged with lipopolysaccharide under thermoneutral and high ambient
temperatures. J. Anim. Sci. 2012,90, 4788–4794. [CrossRef] [PubMed]
92.
Pohl, A.; Heuwieser, W.; Burfeind, O. Technical note: Assessment of milk temperature measured by automatic milking systems as
an indicator of body temperature and fever in dairy cows. J. Dairy Sci. 2014,97, 4333–4339. [CrossRef] [PubMed]
93.
Igono, M.O.; Johnson, H.D.; Steevens, B.J.; Krause, G.F.; Shanklin, M.D. Physiological, Productive, and Economic Benefits of
Shade, Spray, and Fan System Versus Shade for Holstein Cows During Summer Heat. J. Dairy Sci.
1987
,70, 1069–1079. [CrossRef]
94.
Chaudhari, B.J.; Singh, M. Relationship between udder, skin and milk temperature in lactating Murrah buffaloes during the
hot-humid season. Buffalo Bull. 2015,34, 181–188.
95.
Igono, M.; Johnson, H.; Steevens, B.; Hainen, W.; Shanklin, M. Effect of season on milk temperature, milk growth hormone,
prolactin, and somatic cell counts of lactating cattle. Int. J. Biometeorol. 1988,32, 194–200. [CrossRef] [PubMed]
96.
Du Preez, J.H. Parameters for the determination and evaluation of heat stress in dairy cattle in South Africa.
Onderstepoort J. Vet. Res.
2000,67, 263–271.
97.
Theurer, M.E.; Anderson, D.E.; White, B.J.; Miesner, M.D.; Larson, R.L. Effects of weather variables on thermoregulation of calves
during periods of extreme heat. Am. J. Vet. Res. 2014,75, 296–300. [CrossRef]
98.
Poikalainen, V.; Praks, J.; Veermäe, I.; Kokin, E. Infrared temperature patterns of cow’s body as an indicator for health control at
precision cattle farming. Agron. Res. 2012,10, 187–194.
99.
Berckmans, D. Precision livestock farming technologies for welfare management in intensive livestock systems. Sci. Tech. Rev. Off.
Int. Epizoot. 2014,33, 189–196. [CrossRef]
100.
Norton, T.; Berckmans, D. Developing precision livestock farming tools for precision dairy farming. Anim. Front.
2017
,7, 18–23.
[CrossRef]
Animals 2021,11, 980 23 of 24
101.
Gody´n, D.; Herbut, P.; Angrecka, S. Measurements of peripheral and deep body temperature in cattle—A review. J. Therm. Biol.
2019,79, 42–49. [CrossRef]
102.
Salles, M.S.V.; da Silva, S.C.; Salles, F.A.; Roma, L.C.; El Faro, L.; Mac Lean, P.A.B.; de Oliveira, C.E.L.; Martello, L.S. Mapping the
body surface temperature of cattle by infrared thermography. J. Therm. Biol. 2016,62, 63–69. [CrossRef] [PubMed]
103.
Vogel, B.; Wagner, H.; Gmoser, J.; Wörner, A.; Löschberger, A.; Peters, L.; Frey, A.; Hofmann, U.; Frantz, S. Touch-free measurement
of body temperature using close-up thermography of the ocular surface. MethodsX 2016,3, 407–416. [CrossRef] [PubMed]
104.
Schaefer, A.L.; Cook, N.J.; Bench, C.; Chabot, J.B.; Colyn, J.; Liu, T.; Okine, E.K.; Stewart, M.; Webster, J.R. The non-invasive and
automated detection of bovine respiratory disease onset in receiver calves using infrared thermography. Res. Vet. Sci.
2012
,93,
928–935. [CrossRef] [PubMed]
105.
Hoffmann, G.; Schmidt, M.; Ammon, C. First investigations to refine video-based IR thermography as a non-invasive tool to
monitor the body temperature of calves. Animal 2016,10, 1542–1546. [CrossRef] [PubMed]
106.
George, W.D.; Godfrey, R.W.; Ketring, R.C.; Vinson, M.C.; Willard, S.T. Relationship among eye and muzzle temperatures
measured using digital infrared thermal imaging and vaginal and rectal temperatures in hair sheep and cattle. J. Anim. Sci.
2014
,
92, 4949–4955. [CrossRef]
107.
Bach, A.J.E.; Stewart, I.B.; Disher, A.E.; Costello, J.T. A Comparison between Conductive and Infrared Devices for Measuring
Mean Skin Temperature at Rest, during Exercise in the Heat, and Recovery. PLoS ONE
2015
,10, e0117907. [CrossRef] [PubMed]
108.
Hill, T.M.; Bateman, H.G.; Suarez-Mena, F.X.; Dennis, T.S.; Schlotterbeck, R.L. Short communication: Changes in body temperature
of calves up to 2 months of age as affected by time of day, age, and ambient temperature. J. Dairy Sci.
2016
,99, 8867–8870.
[CrossRef] [PubMed]
109.
Legrand, A.; Schütz, K.E.; Tucker, C.B. Using water to cool cattle: Behavioral and physiological changes associated with voluntary
use of cow showers. J. Dairy Sci. 2011,94, 3376–3386. [CrossRef]
110.
Kou, H.; Zhao, Y.; Ren, K.; Chen, X.; Lu, Y.; Wang, D. Automated measurement of cattle surface temperature and its correlation
with rectal temperature. PLoS ONE 2017,12, e0175377. [CrossRef]
111.
Sellier, N.; Guettier, E.; Staub, C. A Review of Methods to Measure Animal Body Temperature in Precision Farming. Am. J. Agric.
Sci. Technol. 2014,2, 74–99. [CrossRef]
112.
Sousa, R.V.d.; Rodrigues, A.V.d.S.; Abreu, M.G.d.; Tabile, R.A.; Martello, L.S. Predictive model based on artificial neural network
for assessing beef cattle thermal stress using weather and physiological variables. Comput. Electron. Agric.
2018
,144, 37–43.
[CrossRef]
113.
Pacheco, V.M.; Sousa, R.V.d.; Rodrigues, A.V.D.S.; Sardinha, E.J.d.S.; Martello, L.S. Thermal imaging combined with predictive
machine learning based model for the development of thermal stress level classifiers. Livest. Sci. 2020,241, 104244. [CrossRef]
114.
McArthur, A.J. Thermal interaction between animal and microclimate: A comprehensive model. J. Theor. Biol.
1987
,126, 203–238.
[CrossRef]
115.
Brown-Brandl, T.M.; Eigenberg, R.A.; Nienaber, J.A.; Hahn, G.L. Dynamic Response Indicators of Heat Stress in Shaded and
Non-shaded Feedlot Cattle, Part 1: Analyses of Indicators. Biosys. Eng. 2005,90, 451–462. [CrossRef]
116.
Milan, H.F.M.; Maia, A.S.C.; Gebremedhin, K.G. Technical note: Device for measuring respiration rate of cattle under field
conditions. J. Anim. Sci. 2016,94, 5434–5438. [CrossRef] [PubMed]
117.
Eigenberg, R.A.; Hahn, G.L.; Nienaber, J.A.; Brown-Brandl, T.M.; Spiers, D.E. Development of a new respiration rate monitor for
cattle. Trans. ASAE 2000,43, 723–728. [CrossRef]
118.
Strutzke, S.; Fiske, D.; Hoffmann, G.; Ammon, C.; Heuwieser, W.; Amon, T. Technical note: Development of a noninvasive
respiration rate sensor for cattle. J. Dairy Sci. 2019,102, 690–695. [CrossRef] [PubMed]
119.
De Carvalho, G.A.; Salman, A.K.D.; da Cruz, P.G.; de Souza, E.C.; da Silva, F.R.F.; Schmitt, E. Technical note: An acoustic method
for assessing the respiration rate of free-grazing dairy cattle. Livest. Sci. 2020,241, 104270. [CrossRef]
120.
Pastell, M.; Aisla, A.M.; Hautala, M.; Poikalainen, V.; Praks, J.; Veermäe, I.; Ahokas, J. Contactless measurement of cow behavior
in a milking robot. Behav. Res. Methods 2006,38, 479–486. [CrossRef]
121.
Davison, C.; Michie, C.; Hamilton, A.; Tachtatzis, C.; Andonovic, I.; Gilroy, M. Detecting Heat Stress in Dairy Cattle Using
Neck-Mounted Activity Collars. Agriculture 2020,10, 210. [CrossRef]
122.
De Melo Costa, C.C.; de Melo Costa, C.C.; Maia, A.S.C.; Maia, A.S.C.; Nascimento, S.T.; Nascimento, S.T.; Nascimento, C.C.N.;
Nascimento, C.C.N.; Neto, M.C.; Neto, M.C.; et al. Thermal balance of Nellore cattle. Int. J. Biometeorol.
2018
,62, 723–731.
[CrossRef]
123.
Lees, J.C.; Lees, A.M.; Gaughan, J.B. Developing a heat load index for lactating dairy cows. Anim. Prod. Sci.
2018
,58, 1387.
[CrossRef]
124.
Islam, M.A.; Lomax, S.; Doughty, A.K.; Islam, M.R.; Clark, C.E.F. Automated Monitoring of Panting for Feedlot Cattle: Sensor
System Accuracy and Individual Variability. Animals 2020,10, 1518. [CrossRef] [PubMed]
125.
Toledo, I.M.; Fabris, T.F.; Tao, S.; Dahl, G.E. When do dry cows get heat stressed? Correlations of rectal temperature, respiration
rate, and performance. JDS Commun. 2020,1, 21–24. [CrossRef]
126.
Laister, S.; Stockinger, B.; Regner, A.-M.; Zenger, K.; Knierim, U.; Winckler, C. Social licking in dairy cattle—Effects on heart rate
in performers and receivers. Appl. Anim. Behav. Sci. 2011,130, 81–90. [CrossRef]
127.
Stewart, M.; Wilson, M.T.; Schaefer, A.L.; Huddart, F.; Sutherland, M.A. The use of infrared thermography and accelerometers for
remote monitoring of dairy cow health and welfare. J. Dairy Sci. 2017,100, 3893–3901. [CrossRef]
Animals 2021,11, 980 24 of 24
128.
Lowe, G.; Sutherland, M.; Waas, J.; Schaefer, A.; Cox, N.; Stewart, M. Infrared Thermography—A Non-Invasive Method of
Measuring Respiration Rate in Calves. Animals 2019,9, 535. [CrossRef]
129.
Amamou, H.; Beckers, Y.; Mahouachi, M.; Hammami, H. Thermotolerance indicators related to production and physiological
responses to heat stress of holstein cows. J. Therm. Biol. 2019,82, 90–98. [CrossRef]
130.
Gaughan, J.B.; Mader, T.L. Body temperature and respiratory dynamics in un-shaded beef cattle. Int. J. Biometeorol.
2014
,58,
1443–1450. [CrossRef]
131.
Tresoldi, G.; Schütz, K.E.; Tucker, C.B. Assessing heat load in drylot dairy cattle: Refining on-farm sampling methodology.
J. Dairy Sci. 2016,99, 8970–8980. [CrossRef]
132.
Silanikove, N. Effects of heat stress on the welfare of extensively managed domestic ruminants. Livest. Prod. Sci.
2000
,67, 1–18.
[CrossRef]
133.
Becker, C.A.; Aghalari, A.; Marufuzzaman, M.; Stone, A.E. Predicting dairy cattle heat stress using machine learning techniques.
J. Dairy Sci. 2021,104, 501–524. [CrossRef]
134.
Kabuga, J.D. The influence of thermal conditions on rectal temperature, respiration rate and pulse rate of lactating Holstein-
Friesian cows in the humid tropics. Int. J. Biometeorol. 1992,36, 146–150. [CrossRef]
135.
Martello, L.; Junior, H.S.; Silva, S.; Balieiro, J. Alternative body sites for heat stress measurement in milking cows under tropical
conditions and their relationship to the thermal discomfort of the animals. Int. J. Biometeorol. 2010,54, 647–652. [CrossRef]
136.
Sammad, A.; Wang, Y.J.; Umer, S.; Lirong, H.; Khan, I.; Khan, A.; Ahmad, B.; Wang, Y. Nutritional Physiology and Biochemistry
of Dairy Cattle under the Influence of Heat Stress: Consequences and Opportunities. Animals 2020,10, 793. [CrossRef]
137. Fuquay, J.W. Heat Stress as it Affects Animal Production. J. Anim. Sci. 1981,52, 164–174. [CrossRef]
138.
Spiers, D.E.; Spain, J.N.; Ellersieck, M.R.; Lucy, M.C. Strategic application of convective cooling to maximize the thermal gradient
and reduce heat stress response in dairy cows. J. Dairy Sci. 2018,101, 8269–8283. [CrossRef]
139.
Ji, B.; Banhazi, T.; Ghahramani, A.; Bowtell, L.; Wang, C.; Li, B. Modelling of heat stress in a robotic dairy farm. Part 1:
Thermal comfort indices as the indicators of production loss. Biosys. Eng. 2020,199, 27–42. [CrossRef]
140.
Kovacs, L.; Kezer, F.L.; Ruff, F.; Jurkovich, V.; Szenci, O. Assessment of heat stress in 7-week old dairy calves with non-invasive
physiological parameters in different thermal environments. PLoS ONE 2018,13, e0200622. [CrossRef]
141.
Buffington, D.E.; Collazo-Arocho, A.; Canton, G.H.; Pitt, D.; Thatcher, W.W.; Collier, R.J. Black Globe-Humidity Index (BGHI) as
Comfort Equation for Dairy Cows. Trans. ASAE 1981,24, 711–714. [CrossRef]
142.
Mader, T.L.; Johnson, L.J.; Gaughanf, J.B. A comprehensive index for assessing environmental stress in animals. J. Anim. Sci.
2010
,
88, 2153–2165. [CrossRef]
143.
Yan, G.; Li, H.; Zhao, W.; Shi, Z. Evaluation of thermal indices based on their relationships with some physiological responses of
housed lactating cows under heat stress. Int. J. Biometeorol. 2020,64, 2077–2091. [CrossRef]
144.
Bohmanova, J.; Misztal, I.; Cole, J.B. Temperature-Humidity Indices as Indicators of Milk Production Losses due to Heat Stress.
J. Dairy Sci. 2007,90, 1947–1956. [CrossRef]
145.
Hernández-Julio, Y.F.; Yanagi, T.; Pires, M.d.F.Á.; Aurélio Lopes, M.; Ribeiro de Lima, R. Models for Prediction of Physiological
Responses of Holstein Dairy Cows. Appl. Artif. Intell. 2014,28, 766–792. [CrossRef]
146.
Schütz, K.E.; Rogers, A.R.; Poulouin, Y.A.; Cox, N.R.; Tucker, C.B. The amount of shade influences the behavior and physiology of
dairy cattle. J. Dairy Sci. 2010,93, 125–133. [CrossRef]
147.
Mondaca, M.R.; Choi, C.Y.; Cook, N.B. Understanding microenvironments within tunnel-ventilated dairy cow freestall facilities:
Examination using computational fluid dynamics and experimental validation. Biosys. Eng. 2019,183, 70–84. [CrossRef]
148.
Fuentes, S.; Viejo, C.G.; Cullen, B.; Tongson, E.; Chauhan, S.S.; Dunshea, F.R. Artificial Intelligence Applied to a Robotic Dairy
Farm to Model Milk Productivity and Quality based on Cow Data and Daily Environmental Parameters. Sensors
2020
,20, 2975.
[CrossRef]
149.
Miura, R.; Yoshioka, K.; Miyamoto, T.; Nogami, H.; Okada, H.; Itoh, T. Estrous detection by monitoring ventral tail base surface
temperature using a wearable wireless sensor in cattle. Anim. Reprod. Sci. 2017,180, 50–57. [CrossRef]
Available via license: CC BY 4.0
Content may be subject to copyright.
