Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
1

ARTICLE
Using unmanned aerial vehicles to record behavioral and
physiological indicators of heat stress in cattle on feedlot
and pasture
Can. J. Anim. Sci. Downloaded from cdnsciencepub.com by Dr. John Church on 11/08/21
For personal use only.

J.T. Mufford, M.W. Reudink, M. Rakobowchuk, C.N. Carlyle, and J.S. Church

Abstract: Physiological and behavioral indicators of heat stress in cattle are time- and labor-intensive to measure,
and difficult to observe in extensive feedlot and pasture settings. We proposed to record respiration rate and standing behavior using unmanned aerial vehicles. Videos were recorded above steers on feedlot in the morning
(0830–1130) and afternoon (1400–1700) over 10 d between 25 July and 10 August and cows on pasture over 9 d
between 19 and 29 August In the feedlot, video recordings on 925 individuals (264 black coated, 413 red, and 248
white) were obtained, varying in breed which included Black Angus, Hereford, Charolais, Canadian Speckle Park,
and Simmental. On pasture, video recordings on 267 individuals (116 Black Angus and 151 Hereford) were obtained.
Observer software was used to analyze videos. Respiration rate in feedlot cattle was the highest in black cattle,
followed by red cattle, then white cattle. Coat color did not affect respiration rate in cows on pasture; temperatures
on pasture were lower than in feedlots and the effect of coat color may not manifest until a certain heat load threshold. The probability that cattle would be standing increased with heat load index in feedlot and pasture settings.
Key words: beef cattle, coat color, heat stress, respiration rate, unmanned aerial vehicles.
Résumé : Les indicateurs physiologiques et comportementaux du stress thermique chez les bovins sont longs et
laborieux à mesurer, et difficiles à observer dans les environnements extensifs de parcs d’engraissement et de
pâturage. Nous avons proposé d’enregistrer le taux de respiration et le comportement en position debout à l’aide
de véhicules aériens sans pilote. Les vidéos ont été enregistrées au-dessus des bouvillons dans le parc d’engraissement les matins (0830 à 1130) et les après-midis (1400 à 1700) pendant dix jours entre le 25 juillet et le 10 août,
et au-dessus des vaches en pâturage pendant neuf jours entre le 19 et le 29 août. Dans le parc d’engraissement,
des enregistrements vidéos de 925 individus (264 à pelage noir, 413 à pelage roux, et 248 à pelage blanc) ont été
obtenus, variant selon les races, qui comprenaient Black Angus, Hereford, Charolais, Canadian Speckle Park,
et Simmental. En pâturage, des enregistrements vidéos de 267 individus (116 Black Angus et 151 Hereford) ont
été obtenus. Le logiciel Observer a été utilisé afin d’analyser les vidéos. Le taux de respiration dans les bovins du
parc d’engraissement était le plus élevé chez les bovins à pelage noir, suivi de ceux à pelage roux, puis ceux à
pelage blanc. La couleur du pelage n’a pas eu d’effet sur le taux de respiration chez les vaches en pâturage; les
températures au pâturage étaient plus faibles que dans le parc d’engraissement et l’effet de la couleur du pelage
pourrait ne pas se manifester tant qu’un seuil de charge thermique soit atteint. La probabilité que les bovins soient
en position debout augmentait avec l’indice de charge thermique dans les environnements de parc d’engraissement et de pâturage. [Traduit par la Rédaction]
Mots-clés : bovin de boucherie, couleur du pelage, stress thermique, taux de respiration, véhicule aérien sans pilote.

Introduction
Heat stress is an emerging problem for both animal
welfare and production in cattle in temperate climates.
Heat stress in cattle (Bos taurus) adversely affects growth,

feed conversion efficiency, and reproductive performance (Bernabucci et al. 2019; Lees et al. 2019). In addition, heat waves can cause mortalities which result in
economic losses (Lees et al. 2019). Climate change models

Received 27 July 2020. Accepted 8 June 2021.
J.T. Mufford, M.W. Reudink, M. Rakobowchuk, and J.S. Church. Thompson Rivers University, 805 TRU Way, Kamloops, BC V2C 0C8,
Canada.
C.N. Carlyle. University of Alberta, Edmonton, AB T6G 2P5, Canada.
Corresponding author: J.S. Church (email: jchurch@tru.ca).
Copyright remains with the author(s) or their institution(s). Permission for reuse (free in most cases) can be obtained from
copyright.com.
Can. J. Anim. Sci. 00: 1–8 (0000) dx.doi.org/10.1139/cjas-2020-0125

Published at www.nrcresearchpress.com/cjas on 25 June 2021.

Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)

Can. J. Anim. Sci. Downloaded from cdnsciencepub.com by Dr. John Church on 11/08/21
For personal use only.

2

predict that cattle will experience heat stress on a
greater number of days (Reeves and Bagne 2016) as the
average summer temperatures and the frequency
and magnitude of heat waves are forecasted to increase
(Coumou and Rahmstorf 2012; Pasqui and Di
Giuseppe 2019).
Measuring indicators of heat stress in cow–calf operations situated on pasture or rangelands is logistically
challenging (Godyn et al. 2019). Monitoring physiological
indicators of heat stress such as body temperature
require invasive procedures and (or) wearable devices
that are reliable (Godyn et al. 2019) but may be cost prohibitive, especially in large-scale studies (Koltes et al.
2018; Carabaño et al. 2019). Given these challenges and
limitations, there is a growing interest in developing
more effective tools to measure indicators of heat stress
in cattle (Koltes et al. 2018; Carabaño et al. 2019; Lowe
et al. 2019). For example, respiration rate is a reliable
indicator of heat stress that can be measured through
observation without invasive surgical procedures (Lowe
et al. 2019). However, respiration rate is time- and laborintensive to measure in the field (Gaughan et al. 2008;
Gaughan et al. 2010) as it requires the observer to
approach, at a close distance, individual cattle spread
across a large feedlot or pasture.
Unmanned aerial vehicles (UAV) offer a non-invasive
and practical approach to studying physiological (respiration rate) and behavioral (standing behavior) indicators of heat stress in cattle in both large-scale feedlots
and pasture conditions. The battery life, affordability,
and data-collection capability of consumer-grade
UAVs have substantially improved in the last decade
(Whitehead and Hugenholtz 2014a; Whitehead et al.
2014b), and they have potential for use in cattle production and behavioral studies. Unmanned aerial vehicles
have been used for identification (Andrew et al. 2017),
enumeration (Whitehead et al. 2014b; Shao et al. 2020),
monitoring feed intake (Nyamuryekung’e et al. 2016),
and studying social behavior in cows (Mufford
et al. 2019).
In response to the worsening problem of heat stress,
there is interest in determining factors associated with
heat stress susceptibility (Brown-Brandl 2013).
Identifying these factors may be useful for mitigating
production loss and improving animal welfare. Animals
known to be susceptible to heat stress can be selectively
managed; this can be more efficient than applying the
same heat stress management procedure to every animal (Brown-Brandl and Jones 2011). Furthermore, determining cattle traits that either increase or reduce their
susceptibility to heat stress could inform trait selection
(Carabaño et al. 2019).
One important factor that affects heat stress susceptibility is coat color. Darker coats, having a lower albedo,
absorb more solar radiation than lighter coats (Finch
et al. 1986; Hillman et al. 2005). The impact of coat color
on heat stress has been well established in feedlot cattle

Can. J. Anim. Sci. Vol. 00, 0000

(Brown-Brandl 2013), but little work has been conducted
in cattle on pasture. Furthermore, little work has been
done in Canada even though heat-stress mortality has
been known to occur in Canadian production settings
(Bishop-Williams et al. 2015; Bishop-Williams et al. 2016).
The primary objective of this study was to determine
whether UAV could be used to record indicators of heat
stress, respiration rate, and standing behavior. In addition, we chose to do so in feedlot and pasture cattle to
test the effectiveness of this approach in different production settings. We chose to examine respiration rate
and standing behavior specifically because (a) respiration rate increases with heat load (Brown-Brandl et al.
2006; Johnson et al. 2012; Veissier et al. 2018), and
(b) standing behavior is associated with heat stress as
the time spent standing increases with increasing heat
loads (Brown-Brandl et al. 2006; Tucker et al. 2008;
Tucker et al. 2008; Provolo and Riva 2009). The secondary
objective of this study was to examine the effect of coat
color on heat stress in cattle on pasture.

Materials and Methods
All the procedures used in our experiments were conducted in accordance with Canadian Council of Animal
Care guidelines (CCAC 2009), and it was pre-approved
by the Animal Care Committee of Thompson Rivers
University (Kamloops, BC, Canada) (file number: 101909).
Site 1: Feedlot

The first study site was a feedlot operated by Kasko
Cattle Company (Ltd.), located near Purple Springs, AB,
Canada (49°50′38.2″N, 111°58′39.8″W). This feedlot contained 66 pens, each containing 100–200 beef cattle. In
total, there were roughly 9000 steers throughout the
feedlot. The average weight at arrival ranged between
450 and 700 kg, and all individuals were kept in the feedlot for approximately 3 mo. Pens had a soil surface and
were 50 m × 60 m; the feeding bunks faced an east/west
orientation. In addition, pens were adjacent to each
other, separated by 2.5 m fencing, and there were six
rows of pens. Each pen contained a variety of breeds
including, but not limited to, Black Angus, Hereford,
Charolais, Canadian Speckle Park, Simmental, and various crosses. Cattle that were recently treated for the disease were identified by ear tag and excluded from the
study. Grain feed was provided by truck once in the
morning at 0800–1000 in a feed bunk along the width
of each pen, which was freely accessible. Each pen contained a water trough that enabled ad libitum water
intake. There were no artificial shade structures, but
fencing provided some shade for a few cattle depending
on the time of day. Cattle along the shaded fence line
were not included in the study.
Site 2: Pasture

The second study site was the University of Alberta
Mattheis Research Ranch (50°53′41.8″N, 111°57′00.4″W).
Published by NRC Research Press

Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)

Can. J. Anim. Sci. Downloaded from cdnsciencepub.com by Dr. John Church on 11/08/21
For personal use only.

Mufford et al.

Two cow–calf herds in different pastures were included
in the study. The first herd consisted of approximately
175 Black Angus cow–calf pairs and 15 Hereford cow–calf
pairs; the age of the cows ranged from 5 to 10 yr old. This
herd was in a native grassland dominated by needle-andthread grass (Hesperostipa comata, Trin. and Rup.) and
blue grama grass (Bouteloua gracilis, Willd ex Kunth), with
sand reed grass [Calamovilfa longifolia (Hook.) Scribn.],
June grass [Koeleria macrantha (Ledeb.) Schult.], and
western wheatgrass [Pascopyrum smithii (Rydb.) Á. Löve]
as common subdominants. The second herd consisted
of approximately 350 Hereford cow–calf pairs and 50
Black Angus cow–calf pairs; there was a wide range in
age of the cows, 3–14 yr old. This herd was in an irrigated
field dominated by perennial agronomic plants, primarily smooth brome (Bromus inermis Leyss), alfalfa (medicago sativa L.), and cicer milkvetch (Astragalus cicer L.).
Only cows were included in the study. Each pasture was
approximately 300 ha of flat grassland with no shade
from trees or artificial covers. Water was available in
each pasture from natural sources or provided by truck
to a watering trough on a consistent basis to ensure ad
libitum water intake.
Data collection

We used a DJI Mavic Pro quadcopter (Dà-Jiāng
Innovations Science and Technology Co., Ltd.,
Shenzhen, China) to record video of cattle at an altitude
of 8–10 m in the feedlot setting and 5–10 m in the pasture
setting. Because we were unable to identify individuals,
we ran the risk of pseudo-replication in sampling.
Videos were recorded for 3 min at a time to maximize
the number of samples that were obtained on a single
battery charge. The UAV was returned to the home point
between recordings to exchange batteries when
necessary.
At the feedlot, data collection occurred over 10 d
between 25 July and 2 Aug. and between 8 and 10 Aug.
2018, during a morning period (0830–1130) and during
an afternoon period (1400–1700). To minimize the potential effects of pseudo-replication, during each data collection period, we flew the UAV over randomly selected
pens to record video. Multiple videos were recorded per
pen, each video recording different cattle within each
pen. It is possible that the same individual may have
been pseudo-replicated between each video if an individual moved across the pen between video recordings.
However, we generally observed that within the time
frame of the three recordings, most cattle did not move
locations within the pen. Furthermore, the observer
was able to keep track of movement throughout most
of the pen through real-time video streaming between
the UAV and the controller.
At the research pasture, data collection occurred over
9 d between 19 and 29 Aug. 2018. Each day, we collected
data during the morning period, 0830–1130, and the
afternoon period, 1400–1700. The two herds studied were

3

separated into different pastures spaced far enough
apart that it was not logistically possible to collect data
on both herds during the same period. On the first day,
we collected data on one herd for both collection periods; on the second day, we collected data on the other
herd for both collection periods, and we continued alternating herds each day. During the collection period, we
recorded videos of as many cattle as possible. We manually flew the UAV but moved in a grid pattern to minimize the risk of sampling the same cattle.
Prior to data collection at the pasture and the feedlot,
cattle were given a week to habituate to the UAV. On
the first day of exposure, we flew the UAV over the cattle
at an altitude of 100 m and gradually descended to 80 m,
hovered stationary over the cattle and flew in various
directions haphazardly above them. We descended
20 m lower each subsequent day and repeated this process each day until we reached 10 m. At 10 m, most cattle
did not react to the UAV, but some showed behavioral
responses, including sudden changes in position
(i.e., lying to standing or standing to a fast walking pace),
rapid head turns, and frequent tail flicking. Any cattle
exhibiting these behaviors during the data collection
period were not included in the study. Cattle that did
not show a behavioral response were considered habituated and were included in the study.
Environmental data

A Kestrel 5400AG portable weather station (NielsenKellerman Company, Boothwyn, PA, USA) was used
throughout all data collection periods at both sites to
measure wind speed, black globe temperature, ambient
air temperature, and relative humidity. These variables
were used to determine the heat load index (HLI), which
was calculated as follows, originally described by
Gaughan et al. (2008):
If Ta > 25°C, then HLI = 8.62 + ð0.38 × RHÞ
+ ð1.55 × TbgÞ − ð0.5 × WS + eð2.4 − WSÞÞ
If Ta < 25°C, then HLI = 10.66 + ð0.28 × RHÞ
+ ð1.3 × TbgÞ − WS
where Ta is the ambient air temperature (°C), RH the relative humidity (%), Tbg the black globe temperature (°C),
and WS the wind speed (m·s−1).
Heat load index is predictive of heat stress behavior in
cattle (Gaughan et al. 2008; Brown-Brandl 2013). These
conditions were measured and automatically recorded
every 10 min. Each individual was assigned the HLI value
that was measured closest to the time that respiration
rate was recorded, as all video recordings were time
stamped. The portable weather station was mounted on
a tripod within 1 km of the study animals at the feedlot
and within 3 km of the study animals at the research
ranch.
Published by NRC Research Press

Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
4

Can. J. Anim. Sci. Vol. 00, 0000

Table 1. The mean ± standard error of mean (SEM) and range of weather conditions
during feedlot observations and pasture observations.

Feedlot

Can. J. Anim. Sci. Downloaded from cdnsciencepub.com by Dr. John Church on 11/08/21
For personal use only.

Ambient temperature (°C)
BG temperature (°C)
Rh (%)
Wind speed (m·s−1)
Heat load index

Pasture

Range

Mean ± SEM

Range

Mean ± SEM

18–38.9
27.3–51.5
17.3–70.5
0–3.1
75.6–102.2

28.9 ± 0.2
41.2 ± 0.2
36 ± 0.4
1.1 ± 0.02
91.3 ± 0.2

10–33.2
15–40.8
19.4–60
1.3–5.7
50.9–91.7

21.9 ± 0.4
32.1 ± 0.5
42.6 ± 1.2
1.9 ± 0.1
73 ± 0.7

Table 2. Ethogram of behaviors that were recorded in Observer XT.

Behavior

Definition

—Inhale
—Regurgitate
—Skin twitch
—Groom
—Change positions
Standing
Lying
Walking

Flanks expand and collapse at a steady rate
Flanks abruptly expand and collapse after chewing momentarily stops
The skin shakes and obscures view of inhale events
Head is turned to the side to lick the coat
Adjusts position of the flank and (or) legs while lying
All legs are in an upright position
The flank is in contact with the ground
Both the front and back legs are stepping

Note: Behaviors denoted with an em dash (—) were recorded as events; all other behaviors
were recorded as durations.

The range of the HLI was 75.6–102.2 during feedlot
observations and was 50.9–91.7 during pasture observations. The range of the ambient air temperature, black
globe temperature, relative humidity, and wind speed is
summarized in Table 1.
Data acquisition

Videos of cattle captured by the UAV were processed
in Observer XT Software (Noldus, Information
Technology, Wageningen, The Netherlands) to quantify
respiration rate and behavior (see Table 2 for ethogram).
Multiple animals were captured in each video, so each
animal was analyzed individually. Respiration rate was
quantified by counting flank movements for 3 min; each
flank movement was recorded and time stamped as a
behavioral event. Within those 3 min, any behavior that
obscured flank movement was also recorded and time
stamped as a behavioral event. These behaviors included
changing positions, skin twitching, grooming, and
regurgitating. Lying, walking, and standing were
recorded as duration behaviors. The observer coding system was configured such that behavioral events can be
recorded at the same time that standing, lying, and walking were recorded. Only observations in which the flank
movements were observable for at least 2 min were
included in the dataset.
The duration of a behavioral event that obscured flank
movement was determined by recording the time that
elapsed between the flank movement that occurred
before and after the behavioral event. We determined

the time during which flank movements were observable by subtracting the total observation time by the
total time spent exhibiting behaviors that obscured
flank movement. Respiration rate was calculated by
dividing the total flank movements by the time (seconds)
during which flank movements were observable. This
value was multiplied by 60 to obtain breaths per minute
(BPM).
Statistical analysis

Analysis of videos was randomly divided between
three observers. The intra- and inter-reliability was determined by comparing BPM scores. Each observer randomly selected 25 cattle and quantified BPM on each
individual twice. The intra-observer reliability measured
by linear regression was 0.70, 0.80, and 0.98 for the three
observers. To determine inter-observer reliability, each
observer quantified BPM of the same 40 cattle, which
were randomly selected. The inter-observer reliability,
measured by linear regression for each pair of observers
was high (r2 = 0.89, 0.90, and 0.91). Both the intra- and
inter-reliability scores were comparable to other behavioral studies (Schütz et al. 2011; Gutmann et al. 2015;
Vogt et al. 2017).
We first examined the factors associated with respiration rate in feedlot cattle and pasture cattle, separately.
We used a linear mixed model in R version 3.4.3 statistical software (R Core Team 2017), using the stats v3.6.2
package. Coat color (red, black, or white), HLI, and an
HLI × coat color interaction term were treated as fixed
Published by NRC Research Press

Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)

Can. J. Anim. Sci. Downloaded from cdnsciencepub.com by Dr. John Church on 11/08/21
For personal use only.

Mufford et al.

5

effects. Because sampling sites were repeated, we treated
pen (i.e., which pen or which herd) as a random effect.
The alpha level was set at 0.05. Day and time of day were
not included in the model because they are associated
with HLI. We did not include individual in the model
because we were not able to distinguish unique individual cattle.
We also wished to determine which factors were associated with posture (i.e., standing or lying treated as a
binomial response) in feedlot cattle and pasture cattle,
separately. To do so, we conducted a generalized linear
mixed model with binomial error distribution and logit
link function in R statistical software (R Core Team
2017). Coat color, and HLI were treated as fixed effects.
The HLI × coat color interaction was also treated as a
fixed effect. Pen was treated as a random effect. The
alpha level was set at 0.05.
For posture behavior, very few cattle spent time walking during an observation, so any walking durations that
were scored were later converted to standing durations.
The time spent for standing and lying was first calculated as a proportion relative to the total time of the
observation. Proportions were then converted into a categorical response (i.e., standing or lying). Standing ≥50%
of the observation time was categorized as standing.
Lying more than half of the observation time was categorized as lying. We categorized posture as binomial
responses because only a small fraction of the cattle
spent time both lying and standing within an observation (i.e., within 3 min). Most of the cattle were either
standing during the entire observation or lying during
the entire observation.

Fig. 1. Respiration rate responses [breaths per minute (BPM)]
to increasing heat load index (HLI) in feedlot steers (top) and
lactating beef cows on pasture (bottom). [Colour online.]

Results

posture in feedlot cattle (P > 0.13) nor pasture
cattle (P > 0.08).

In the feedlot, a total of 925 individuals were video
recorded: 264 back-coated, 413 red-coated, and 248
white-coated cattle. Coat color affected respiration rate
(F[2, 923] = 20.69, P < 0.001). The mean and standard error
of respiration rates (BPM) in feedlot cattle were as follows: black, 91 ± 1.5; red, 86 ± 1.3; and white, 75 ± 1.5. The
rate of respiration increased with HLI (F[1, 924] = 207.5,
P < 0.001; Fig. 1). The HLI × coat color interaction was
not significant (F[2, 923] = 0.025, P = 0.98).
In the pasture, a total of 267 individuals were video
recorded: 116 black-coated and 151 red-coated cattle.
Respiration rate increased with HLI (F [1, 266] = 88.71,
P < 0.001) (Fig. 1). Coat color did not influence respiration
rate (F[1, 266] = 1.53, P = 0.22) nor was the interaction term
significant (HLI × coat color: F[1, 266] = 0.033, P = 0.85). The
mean and standard error of respiration rates (BPM) in
cattle on pasture were as follows: black, 48 ± 1.3;
red, 48 ± 1.3.
The probability that cattle would be standing, instead
of lying, increased with HLI for both cattle on feedlot
(P = 0.03) and pasture (P < 0.01) (Fig. 2). Coat color, coat
color × HLI interaction, and pen did not determine

Discussion
We successfully used UAVs to measure behavioral and
physiological indicators of heat stress in a large-scale
feedlot and pasture. In feedlot cattle, the respiration rate
was the highest in black cattle (91 BPM), followed by red
(85 BPM) cattle, then white cattle (75 BPM), across all
weather conditions (HLI), and the magnitude of these
differences was the same across all weather conditions.
These findings are consistent with similar feedlot
studies investigating indicators of stress (i.e., panting
and (or) respiration rate) (Brown-Brandl et al. 2005;
Brown-Brandl et al. 2006; Gaughan et al. 2010).
As mentioned above, the HLI × coat color interaction
was not significant. Based on previous work (BrownBrandl et al. 2006), we suspected that the interaction
may have been significant if we had observed cattle in
cooler conditions which would have caused a convergence of breathing rates in cooler ambient conditions.
All observations of feedlot cattle took place above an
HLI of 75, and an HLI above 70 is considered to be above
Published by NRC Research Press

Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
6

Can. J. Anim. Sci. Downloaded from cdnsciencepub.com by Dr. John Church on 11/08/21
For personal use only.

Fig. 2. Effect of heat load index (HLI) on the probability of
standing for feedlot steers (top) and lactating beef cows on
pasture cattle (bottom). Individual cattle standing ≥50% of
the observation are represented by black dots at the top
and <50% at the bottom.

the thermal neutral zone (TNZ) for feedlot cattle
(Gaughan et al. 2010). Other than individual variation,
the respiration rate should not differ between cattle
when they are in their TNZ. Thus, the respiration rate
responsiveness (i.e., the rate of change of respiration
rate with respect to HLI) between black, red, and white
cattle may differ as the HLI increases above their TNZ.
In contrast to the feedlot, cows on pasture with different coat colors did not differ in terms of respiration rate.
Pasture cattle were observed within an HLI range of
50.9–91.7, part of the range being above their TNZ. The
lack of difference shows that coat color does not have a
significant impact on heat stress in grazing cows in this
HLI range. The effect of coat color may not manifest
unless exposed to conditions at a higher heat load
threshold. Further research should make this comparison on days with a higher HLI.
It is possible that feedlot steers are more susceptible
to heat stress compared with cows on pasture. The soilsurfaced pens may have been hotter, on average, compared with the surface on pasture. The feedlot cattle
included in this study may have had, on average, a
higher body condition score (i.e., more subcutaneous

Can. J. Anim. Sci. Vol. 00, 0000

fat) than pasture cattle. Cattle with higher condition
scores have higher respiration responses to high heat
load (Brown-Brandl et al. 2006) as fat cover affects heat
dissipation (Brown-Brandl and Jones 2011). Generally,
feedlot cattle close to their finishing weight have high
condition scores (Brown-Brandl et al. 2006; Gaughan
et al. 2008) compared with cows in cow–calf operations
(Nephawe et al. 2004) that need to be in moderate condition for optimal reproductive performance (Diskin and
Kenny 2016). Feedlot steers may have been heavier on
average than cows; heavier cattle are more susceptible
to heat stress (Brown-Brandl and Jones 2011). The range
of arrival weight of feedlot steers was 450–700 kg; there
were no available data on the weight of cows in this
study, but the average mature weight (measured at
4 yr old) of beef cows is approximately 520 kg
(Nephawe et al. 2004; Bao et al. 2019). The effect of sex
may also explain differences in heat stress susceptibility
between feedlot steers and pasture cows; we are unable
to separate the effects of sex from animal factors (body
size and fat cover) or from the operational context
(pasture vs feedlot).
In the pilot work for this study, we found that over a
3 min period, the respiration rate within a 1 min time
interval can change by 40 BPM in the subsequent 1 min
time interval. Therefore, despite the inter- and intrareliability error, taking the average respiration rate over
a 3 min period was more accurate than extrapolating
respiration rate from a short time interval sample.
We also sought to determine if this method could be
used to observe posture (i.e., lying/standing), a behavioral indicator of heat stress. As HLI increases, cattle are
more likely to stand, regardless of coat color, in both pasture and feedlot cattle. This is likely because more surface area is exposed while standing which facilitates
greater convective heat loss. Other studies have shown
that standing increases as HLI (Tucker et al. 2008), temperature–humidity index (Provolo and Riva 2009), and
dry bulb temperature (Brown-Brandl et al. 2006)
increase.
This study has demonstrated that consumer-grade
UAVs can be used as an effective tool for measuring heat
stress indicators of cattle in large-scale feedlot and pasture operations. Future research should further improve
the efficacy of UAVs as a tool for measuring indicators of
heat stress. For example, camera lenses with optical
zoom are now available on consumer grade UAVs such
as the one in this study; zoom lenses would make it possible to identify individual cattle within an extensive
feedlot/pasture by reading dangle ear tags. Identifying
cattle would be necessary for relating heat stress
responses to biomarkers of heat stress (e.g., blood
parameters) (Carabaño et al. 2019). Furthermore, identification would be useful for determining how individuals
acclimate to hot environments over time (Bernabucci
et al. 2010). Zoom lenses would eliminate or minimize
the effect of pseudo-replication in future studies similar
Published by NRC Research Press

Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
Mufford et al.

Can. J. Anim. Sci. Downloaded from cdnsciencepub.com by Dr. John Church on 11/08/21
For personal use only.

to this. Other behaviors associated with heat stress can
also be identified by aerial-based video; for example,
panting is a severe sign of heat stress and identifying this
behavior would be useful from a management perspective. Potentially, quantifying respiration rate could be
automated using machine learning, which would substantially decrease time and labor for large-scale studies
(Koltes et al. 2018; Lowe et al. 2019). Regular video cameras on UAVs can also record simultaneously with thermal sensors, which may prove helpful in detecting body
temperature and identifying severe signs of heat stress.

Acknowledgements
This research was supported by the Natural Science
and Engineering Research Council (NSERC) and the
University of Alberta Rangeland Research Institute
(RRI). We thank the Kasko Cattle Company (Ltd.) for
access to their research herd. We also thank staff at the
University of Alberta Mattheis Research Ranch for support with our research. We also acknowledge Edwin
and Ruth Mattheis for making it possible to conduct
research at said Ranch. Finally, two anonymous
reviewers provided valuable comments that improved
our manuscript.

References
Andrew, W., Greatwood, C., and Burghardt, T. 2017. Visual localisation and individual identification of Holstein Friesian
cattle via deep learning. IEEE International Conference on
Computer Vision Workshops.
Bao, T., Carlyle, C.N., Bork, E.W., Becker, M., Alexander, M.J.,
DeMaere, C., et al. 2019. Survey of cattle and pasture management practices on focal pastures in Alberta. Can. J. Anim. Sci.
99(4): 955–961. doi:10.1139/cjas-2018-0110.
Bernabucci, U. 2019. Climate change: impact on livestock and
how can we adapt. Anim. Front. 9(1): 3–5. doi:10.1093/af/
vfy039. PMID:32002232.
Bernabucci, U., Lacetera, N., Baumgard, L.H., Rhoads, R.P.,
Ronchi, B., and Nardone, A. 2010. Metabolic and hormonal
acclimation to heat stress in domesticated ruminants.
Animal, 4: 1167–1183. doi:10.1017/S175173111000090X.
PMID:22444615.
Bishop-Williams, K.E., Berke, O., Pearl, D.L., and Kelton, D.F.
2016. Mapping rural community and dairy cow heat stress
in Southern Ontario: a common geographic pattern from
2010 to 2012. Arch. Environ. Occup. Health, 71(4): 199–207.
doi:10.1080/19338244.2015.1058235. PMID:26067385.
Bishop-Williams, K.E., Berke, O., Pearl, D.L., Hand, K., and
Kelton, D.F. 2015. Heat stress related dairy cow mortality during heat waves and control periods in rural Southern Ontario
from 2010–2012. BMC Vet. Res. 11(1): 291. doi:10.1186/s12917015-0607-2.
Brown-Brandl, T.M. 2013. Managing thermal stress in feedlot
cattle: environment, animal susceptibility and management
options from a US perspective. Pages 722–740 in A. Aland
and T. Banhazi, eds. Livestock housing: modern management to ensure optimal health and welfare of farm animals,
Wageningen Academic Publishers.
Brown-Brandl, T., and Jones, D.D. 2011. Feedlot cattle susceptibility to heat stress: an animal-specific model. Trans. Am.
Soc. Agric. Biol. Eng. 54(2): 583–598.

7

Brown-Brandl, T.M., Eigenberg, R.A., and Nienaber, J.A. 2006.
Heat stress risk factors of feedlot heifers. Livest. Sci.
105(1–3): 57–68. doi:10.1016/j.livsci.2006.04.025.
Brown-Brandl, T.M., Eigenberg, R.A., Nienaber, J.A., and Hahn,
G.L. 2005. Dynamic response indicators of heat stress
in shaded and non-shaded feedlot cattle, part 1: analyses
of indicators. Biosyst. Eng. 90(4): 451–462. doi:10.1016/
j.biosystemseng.2004.12.006.
Canadian Council on Animal Care (CCAC). 2009. Canadian
Council on Animal Care. Guidelines on: the care and use of
farm animals in research, teaching and testing. CCAC,
Ottawa, ON, Canada.
Carabaño, M.J., Ramón, M., Menéndez-Buxadera, A., Molina, A.,
and Díaz, C. 2019. Selecting for heat tolerance. Anim. Front.
9(1): 62–68. doi:10.1093/af/vfy033. PMID:32002241.
Coumou, D., and Rahmstorf, S. 2012. A decade of weather
extremes. Nat. Clim. Change, 2(7): 491–496. doi:10.1038/
nclimate1452.
Diskin, M.G., and Kenny, D.A. 2016. Managing the reproductive
performance of beef cows. Theriogenology, 86(1): 379–387.
doi:10.1016/j.theriogenology.2016.04.052. PMID:27180327.
Finch, V.A. 1986. Body temperature in beef cattle: its control
and relevance to production in the tropics. J. Anim. Sci.
62(2): 531–542. doi:10.2527/jas1986.622531x.
Gaughan, J.B., Mader, T.L., Holt, S.M., Sullivan, M.L., and Hahn,
G.L. 2010. Assessing the heat tolerance of 17 beef cattle genotypes. Int. J. Biometeorol. 54: 617–627. doi:10.1007/s00484-0090233-4. PMID:19458966.
Gaughan, J.B., Mader, T.L., Holt, S.M., and Lisle, A. 2008. A new
heat load index for feedlot cattle. J. Anim. Sci. 86(1): 226–234.
doi:10.2527/jas.2007-0305. PMID:17911236.
Godyn, D., Herbut, P., and Angrecka, S. 2019. Measurements of
peripheral and deep body temperature in cattle — a review.
J. Therm. Biol. 79: 42–49. doi:10.1016/j.jtherbio.2018.11.011.
PMID:30612684.
Gutmann, A.K., Špinka, M., and Winckler, C. 2015. Long-term
familiarity creates preferred social partners in dairy cows.
Appl. Anim. Behav. Sci. 169: 1–8. doi:10.1016/j.applanim.
2015.05.007.
Hillman, P.E., Gebremedhin, K.G., Brown-Brandl, T.M., and Lee,
C.N. 2005. Thermal analysis and behavioral activity of heifers
in shade or sunlight. Livestock Environment VII —
Proceedings of the Seventh International Symposium.
Johnson, J.S., Scharf, B., Weaber, R.L., Eichen, P.A., and Spiers,
D.E. 2012. Patterns of heat response and adaptation on
summer pasture: a comparison of heat-sensitive (Angus)
and tolerant (Romosinuano) cattle. J. Therm. Biol. 37(4):
344–350. doi:10.1016/j.jtherbio.2011.10.014.
Koltes, J.E., Koltes, D.A., Mote, B.E., Tucker, J., and Hubbell, D.S.
2018. Automated collection of heat stress data in livestock:
new technologies and opportunities. Transl. Anim. Sci. 2(3):
319–323. doi:10.1093/tas/txy061. PMID:32704715.
Lees, A.M., Sejian, V., Wallage, A.L., Steel, C.C., Mader, T.L., Lees,
J.C., and Gaughan, J.B. 2019. The impact of heat load on cattle.
Animals, 9(6): 322. doi:10.3390/ani9060322. PMID:31174286.
Lowe, G., Sutherland, M., Waas, J., Schaefer, A., Cox, N., and
Stewart, M. 2019. Infrared thermography—A non-invasive
method of measuring respiration rate in calves. Animals,
9(8): 535. doi:10.3390/ani9080535. PMID:31394802.
Mufford, J.T., Hill, D.J., Flood, N.J., and Church, J.S. 2019. Use of
unmanned aerial vehicles (UAVs) and photogrammetric
image analysis to quantify spatial proximity in beef cattle.
J. Unmanned Veh. Syst. 7(3): 194–206. doi:10.1139/juvs-20180025.
Nephawe, K.A., Cundiff, L.V., Dikeman, M.E., Crouse, J.D., and
Van Vleck, L.D. 2004. Genetic relationships between sex-specific traits in beef cattle: mature weight, weight adjusted for
body condition score, height and body condition score of
Published by NRC Research Press

Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)

Can. J. Anim. Sci. Downloaded from cdnsciencepub.com by Dr. John Church on 11/08/21
For personal use only.

8

cows, and carcass traits of their steer relatives. J. Anim. Sci.
82(3): 647–653. doi:10.2527/2004.823647x. PMID:15032420.
Nyamuryekung’e, S., Cibils, A.F., Estell, R.E., and Gonzalez, A.L.
2016. Use of an unmanned aerial vehicle — mounted video
camera to assess feeding behavior of raramuri criollo cows.
Rangel. Ecol. Manag. 69(5): 386–389. doi:10.1016/j.rama.
2016.04.005.
Pasqui, M., and Di Giuseppe, E. 2019. Climate change, future
warming, and adaptation in Europe. Anim. Front. 9(1): 6–11.
doi:10.1093/af/vfy036. PMID:32002233.
Provolo, G., and Riva, E. 2009. One year study of lying and standing behaviour of dairy cows in a frestall barn in Italy. J. Agric.
Eng. Res. 40(2): 27. doi:10.4081/jae.2009.2.27.
R Core Team. 2017. R: a language and environment for statistical computing. R Foundation for Statistical Computing,
Vienna, Austria.
Reeves, M.C., and Bagne, K.E. 2016. Vulnerability of cattle production to climate change on U.S. rangelands. United States
Department of Agriculture General Technical Report.
Schütz, K.E., Rogers, A.R., Cox, N.R., Webster, J.R., and Tucker,
C.B. 2011. Dairy cattle prefer shade over sprinklers: Effects
on behavior and physiology. J. Dairy Sci. 94(1): 273–283.
doi:10.3168/jds.2010-3608. PMID:21183037.
Shao, W., Kawakami, R., Yoshihashi, R., You, S., Kawase, H., and
Naemura, T. 2020. Cattle detection and counting in UAV

Can. J. Anim. Sci. Vol. 00, 0000

images based on convolutional neural networks. Int. J.
Remote Sens. 41(1): 31–52. doi:10.1080/01431161.2019.1624858.
Tucker, C.B., Rogers, A.R., and Schu, K.E. 2008. Effect of solar
radiation on dairy cattle behaviour, use of shade and body
temperature in a pasture-based system. Appl. Anim. Behav.
Sci. 109(2–4): 141–154. doi:10.1016/j.applanim.2007.03.015.
Veissier, I., Palme, R., Moons, C.P., Ampe, B., Sonck, B.,
Andanson, S., and Tuyttens, F.A. 2018. Heat stress in cows at
pasture and benefit of shade in a temperate climate region.
Int. J. Biometeorol. 62(4): 585–595. doi:10.1007/s00484-0171468-0. PMID:29150763.
Vogt, A., Aditia, E.L., Schlechter, I., Schütze, S., Geburt, K.,
Gauly, M., and König von Borstel, U. 2017. Inter- and intraobserver reliability of different methods for recording temperament in beef and dairy calves. Appl. Anim. Behav. Sci.
195: 15–23. doi:10.1016/j.applanim.2017.06.008.
Whitehead, K., and Hugenholtz, C.H. 2014a. Remote sensing of
the environment with small unmanned aircraft systems
(UASs), part 1: scientific and commercial applications.
Unmanned Veh. Syst. 2(3): 69–85. doi:10.1139/juvs-2014-0006.
Whitehead, K., Hugenholtz, C.H., Myshak, S., Brown, O.,
LeClair, A., Tamminga, A., et al. 2014b. Remote sensing of
the environment with small unmanned aircraft systems
(UASs), part 2: a review of progress and challenges.
Unmanned Veh. Syst. 2(3): 86–102. doi:10.1139/juvs-2014-0007.

Published by NRC Research Press