Preventive Veterinary Medicine 224 (2024) 106131

Contents lists available at ScienceDirect

Preventive Veterinary Medicine

journal homepage: www.elsevier.com/locate/prevetmed

Farmers´sense of the biological impact of extreme heat and seasonality on

Swedish high-yielding dairy cows – A mixed methods approach
Lena-Mari Tamminen a, *, Renée Båge a, Maria Åkerlind b, Gabriela Olmos Antillón a, *
a
Department of Clinical Sciences, Swedish University of Agricultural Sciences, Uppsala 75651, Sweden
b
Växa Sverige, Uppsala 751 05, Sweden

A R T I C L E I N F O A B S T R A C T

Keywords: Supporting dairy farmers in becoming resilient towards extreme weather requires a broad understanding of the
Cattle experiences and perceived risks associated with these events from those who undergo them. We used a mixed
Climate methods approach to explore national trends of biological consequences on dairy cow udder health and fertility,
Resilience
combined with in-depth farmer conversations around extreme weather events, focusing on heat. The aim is to
Preparedness
Fertility
provide a comprehensive picture of how dairy farmer perceptions, priorities and decision-making are related to
SCC the season and extreme weather to identify preventive pathways that can reduce biological costs of heat stress on
Swedish dairy cattle during summer. Data collected monthly at cow and farm level between 2016–2019 as part of
the Swedish milk and disease recording system confirm seasonal trends and show increased somatic cell counts
(SCC) and negatively impacted fertility during summers. In addition, transcriptions of 18 interviews with dairy
farmers across the country and seasonal variations of SCC and fertility were thematically analysed. The results
suggest that farmers have a broad definition of extreme weather and are aware of the negative impacts. Yet
handling of extreme weather events can mainly be classified as reactive. Nevertheless, there are long-term effects
on the farm economy, health and herd dynamics. Swedish dairy farmers are currently showing resilience, albeit a
fragile one. The capability to ensure sufficient feed production in extreme weather is critical for farm self-
perceived resilience. However, acknowledging the long-term biological costs related to fertility, currently not
perceived by farmers, has the potential to support proactive planning and improve farm resilience and
profitability.

1. Introduction immune response (Bagath et al., 2019; Turk et al., 2015), and increased
incidence of clinical mastitis has been observed during summers (Vitali
Extreme weather (EW) events are increasing globally because of et al., 2016). In addition, heat negatively impacts fertility, for example,
climate change, especially rain, droughts and hot spells. In Sweden, the through disrupted follicular development or embryonic losses, and
summer of 2018 saw weather conditions that were extreme, due to long indirectly disrupting oestrus cycles and reducing the expression of oes­
periods of unexpected high temperatures and little rain (Sjökvist et al., trus behaviour (Sammad et al., 2020).
2019). These events had significant effects on the overall society. For Temperature humidity index (THI) is a commonly used indicator for
instance, wells went dry, wildfires increased, the health of elderly peo­ when a cow can experience heat stress (Hoffmann et al., 2020). For
ple was impaired and farming was negatively affected (Johnsson et al., temperate regions, a THI of 65 (i.e. 25 ◦ C +20% humidity or 22 ◦ C +
2019). Heat is an important threat to dairy production as dairy cows are 50% humidity) is considered a critical threshold (Pinto et al., 2020).
highly sensitive to heat stress which has tremendous biological cost to Nonetheless, dairy cows change their behaviour already at a THI of 56
them (Fabris et al., 2019; Guzmán-Luna et al., 2022). Heat stress will not (Hut et al., 2022) (Pinto et al., 2020). In Sweden, the maximum average
only decrease milk production but also reduce health and welfare THI reaches well above 65 during ordinary summers and in the summer
(Polsky and von Keyserlingk, 2017). Heat stress cause general discom­ of 2018, the hottest summer so far recorded (since 1951), the maximum
fort and alterations in dairy cows’ physiological parameters, such as THI was between 78–80 in the south and 72–78 in the north (SMHI,

* Corresponding authors.
E-mail addresses: [email protected] (L.-M. Tamminen), [email protected] (R. Båge), [email protected] (M. Åkerlind), gabriela.olmos.antillon@
slu.se (G. Olmos Antillón).

https://doi.org/10.1016/j.prevetmed.2024.106131
Received 17 March 2023; Received in revised form 1 January 2024; Accepted 16 January 2024
Available online 22 January 2024
0167-5877/© 2024 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
L.-M. Tamminen et al. Preventive Veterinary Medicine 224 (2024) 106131

2021). In a recent study we have shown that temperatures from 20 ◦ C events to identify preventive pathways that can reduce the biological
are associated with reduced milk production and increased economic costs of heat stress on Swedish dairy cattle during summer.
costs to the Swedish dairy industry (Ahmed et al., 2022). While it is clear
that heat can pose problems for the high-yielding cows in Swedish dairy 2. Materials and methods
production, the biological costs related to fertility and SCC and how
these are perceived by the herd managers have been less explored. An Our study follows an explanatory mixed methodology design (Cres­
important difficulty for evaluating any connection between heat stress well and Clark, 2017). First quantitative data from the Swedish national
and biological costs in the Swedish dairy herds are the between- and milk and disease recording scheme (SMDRS) were analysed to identify
within-herd management practice changes occurring during the warm­ national trends of SCC and fertility variables under normal and
est months. For example, a key change across the country is the abnormal summer conditions (2018) and to identify cases (i.e. farms
mandatory provision of access to pasture (60–120 days) during that manage the summer period without changes in udder health and
spring/summer (SJVFS, 2019). This means that estimating the impact of fertility and farms that experience negative impact) for the qualitative
heat stress by correlating the differences observed with THI or phase. In-depth semi-structured interviews with the identified Swedish
comparing the summer and winter season, using for example summer to dairy farmers were then performed, and the results were integrated and
winter ratios (Flamenbaum and Ezra, 2007), will be confounded with summarized as themes. An overview of the different parts of the study is
the effect of seasonal changes in management. The close correlation presented in Fig. 1.
between heat (i.e. THI ≥65) and the summer months in Sweden means For the interaction, integration and synthesis of quantitative and
that separating the effects of heat and seasonal changes in a field study is qualitative phases we followed a meta-theoretical perspective of Scott´s
difficult. However, the warm temperatures during the summer of 2018 critical realism where ontology (the way things are) determines episte­
differed from other years and comparing this year to other years pro­ mology (the way things are known) leading to an epistemic relativism
vides an opportunity to study the effect of adding additional heat. This embraced by using an inductive/deductive analytical approach (Scott,
effect is important because the temperatures observed in 2018 match the 2007). A critical realism perspective allows combining quantitative and
predictions of what to expect of future summers in Sweden as climate qualitative data sets, methods and analytical frames at the ontological
change progresses (Sjökvist et al., 2019). However, simply estimating level (thought and reality are fused), allowing the exploration of diverse
the impact at a population level does not automatically lead to im­ perspectives and uncovering relationships that exist in multifaceted
provements in dairy farm resilience. While it is acknowledged that the research challenges. Here the qualitative component focuses on deep
identification of a problem is the first step towards acting, the performed and detailed descriptions of experiences, actions, practices, activities
action is a result of a series of decisions that are impacted by perceived and interpersonal interactions from fieldwork and it is analysed jointly
threats, barriers and susceptibility (Janz and Becker, 1984). Swedish with the national trends for the interpretation of the meaning of a sit­
dairy farmers’ sense making of EW events is currently undescribed, uation (i.e. EW events) from the decision maker (i.e. farmer).
although it is assumed that it is complex and dependant on associations
and experience-based knowledge (Asplund, 2014). Thus, while the 2.1. Ethical statement
extremely hot and dry summer of 2018 has steered the national debate
to EW events related to heat and droughts in Sweden, other EW events In consultation with the ethics and legal department at the Swedish
may also provide significant challenges to agriculture and animal pro­ University of Agricultural Sciences university, in agreement with the
duction and thus be perceived as of greater importance by farmers. Swedish Ethical Authority, the study did not require a special provision
Additionally, there are other recurring seasonal challenges associated or permit according to Swedish law (SFS 2003:460). Nonetheless, a strict
with the warm summer period that may take precedence over heat code of conduct as set out by the Swedish Research Council (Swedish
related EW events. Research Council, 2017) was followed; including gaining informed
In addition to the perception of the problem and the risk, the consent by all the participants and guaranteeing the pseudoanonym­
perceived efficacy of different measures, in combination with barriers isation of their responses and herd registry data. Furthermore, no sen­
and capability of the farmers, will steer if and which preventive mea­ sitive personal information was discussed nor collected during the
sures are implemented (Janz and Becker, 1984; Robert et al., 2016; process. No financial incentive was offered to farmers in exchange for
Svensson et al., 2019). However, the efficacy of specific measures is not their participation. Nonetheless, the results of the project (e.g. their
constant. For example, the most efficient measure for handling climate fertility variables and SCC) were presented at a later occasion and dis­
change and sudden stressors, like EW events, depends on whether the cussed with each farmer over the phone or on the farm. Additionally, a
measure is taken reactively, i.e. during or after an event, or proactively, paper copy was sent by email.
i.e. taken before the event (Stewart, 2009). In addition, the efficiency of
preventive actions can also differ depending on whether short or 2.2. Overarching study population and measures for descriptive and
long-term effects are considered and on which following decisions are quantitative analysis
made over time (Robert et al., 2016). This complexity should be
considered when exploring the decision-making process for handling This study includes data from the SMDRS from 2016 to 2019. Details
different types of extreme weather on farms. Yet, gaining knowledge of on data used can be found in Anonymous et al. (2022). Farms with a
such complexity demands a multidisciplinary approach. minimum of 50 calvings per study year, participating in the SMDRS,
To support dairy farmers in becoming resilient towards EW in gen­ were included in the study. Somatic cell counts (SCC) were retrieved
eral, a broad understanding of farmers perceptions is needed. But from monthly test-day milk records and summarised at herd level
currently little is known about how dairy farmers’ perceive EW events (Table 1). To study seasonal deviations during the years, the herds’
and the impact on their animals now and in a warmer future. In this average monthly SCC was calculated based on individual test results. In
study a mixed methods approach is used to jointly explore the national addition, the deviation and proportional deviation of monthly SCC to the
trends and farmers’ perceptions of the biological consequences on dairy yearly herd average SCC for each study year was also calculated.
cow udder health and fertility during a “normal” vs an extremely hot The SMDRS also collects fertility parameters like inseminations and
summer (2018). Moreover, we explore how such consequences are calvings. Firstly, average number of expected calvings per month,
perceived by farmers, with a contextualized focus on heat and its po­ assuming an evenly distributed calving pattern over the year, was
tential biological impacts to their herds. The aim is to provide a calculated by dividing the total number of calves born per year by 12.
comprehensive picture of how dairy farmer perceptions, priorities and Then the proportional deviation from herd average for each month was
decision-making are related to seasonal weather changes, heat and EW calculated. Secondly, the pregnancy rate within 30 days after the herd-

2
L.-M. Tamminen et al. Preventive Veterinary Medicine 224 (2024) 106131

The summer of 2018 was unusually hot and dry in Sweden

- an example of extreme weather (EW) events that are
predicted to increase with climate change

2016-2019
Dairy cows are particularly Heat stress = Biological costs to QUANtitative part of study:
sensitive to heat (THI>65) health, welfare and fertility QUAN
Data collection
Biological costs related to
udder health and fertility Map national seasonal trends in
overlooked in Nordic udder health (SCC) and fertility (PV30)
Farmer decisions and implementation conditions, like Sweden
of measures depend on perception QUAN
and recognition of problem Data collection Identify case farms
C
- I.e. farms impacted by summer
Heat mitigation measures can and farms that manage better
reduce biological costs
Case
selection Explore impact of an “extreme”
year (2018) compared to
QUALitative part of the study: other years

QUAL Integration and interpretation of results:

What are the farmers saying Data collection
about heat and EW in general?
How farmer perceive biological costs and
Which impacts do they seasonal trends observed in
perceive are a result of national records
heat and EW? QUAL
Are there other challenges related Data analysis How farmer relate trends and deviations in
to the warm summer season 2018 to heat, EW and other challenges
that are prioritised? What actions are taken Integration Relationship between actions described
to reduce the effect of EW, heat and and effects observed to enhance
QUAN+QUAL
seasonal variation? farm resilience to heat stress

Fig. 1. Visualisation of the quantitative and qualitative part of the study and how the results were integrated in the analysis.

Table 1 ggeffects (Lüdecke, 2018). Significance of seasonal trends in monthly

Summary statistics of herds participating in the monthly trial milking between average SCC, proportional deviation from expected number of calvings
2016–2019. SD=Standard deviation. as well as deviation of the farms’ average PV30 were explored using
2016 2017 2018 2019
generalized additive models for large datasets (Wood, 2011). Models
were fitted with fast REML computation and fitted with a “scat” distri­
Number of farms 1235 1251 1252 1258
bution developed for heavily tailed response variables (Wood et al.,
Average Somatic Cell count [x103 cells/mL 223.0 243.8 253.4 261.3
milk] (86.8) (82.1) (85.6) (89.9) 2016). Month was modelled with a cubic cyclic spline by year (allowing
[SD] different smooth terms for different years). Herd was included as a
Average kg milk/ cow and day 32.9 31.07 31.0 31.4 random effect in all models to account for clustering. When modelling
(SD) (3.8) (3.63) (3.8) (4.0) SCC the farms´monthly average days in milk, average milk production
(kg) per cow and average lactation number from each trial sampling
specific voluntary waiting period (PV30), a composite reproductive were included to account for confounding. Initially as non-parametric
performance indicator showing the proportion of eligible cows identi­ coefficients with smooths but after evaluation of linearity all except
fied in oestrus, being inseminated and becoming pregnant, was calcu­ average days in milk were included as parametric variables. Biologically
lated (Löf, 2012). PV30 includes both the animal’s biological ability to plausible interactions were evaluated using AIC. Normality, hetero­
get pregnant and the reproductive management, like oestrus detection scedasticity were evaluated using residual plots and model residuals
and insemination skills, on farm (Löf, 2012). The herd specific voluntary were evaluated for remaining autocorrelation (Van Rij et al., 2022).
waiting period (VWP) used in the calculations was estimated by iden­ Final model setups are described in Table 2.
tifying number of days after calving at which 10% of the animals had
been inseminated. PV30 on a herd level was calculated for each calendar 2.3. Qualitative component
month and the yearly farm´s averages. Then the monthly deviation of
each farm from this average was calculated. Successful pregnancies were 2.3.1. In-depth interviews participants´selection
identified as insemination periods followed by a calving or a confirmed, The study sought to attain a theoretical maximal variation of experi­
positive, pregnancy check. Data from 2019 was only used to identify ences on the impact of yearly heat events on SCC and fertility of a dairy
calvings following inseminations and not included in the analysis of herd. For that, we actively sought participants that showed an increase in
farm´s deviation in PV30. SCC, a decrease milk production or fertility. Equally, farms that did not
Descriptive and quantitative analysis was performed in R (R Core show such variation during summer months. Data from the overarching
Team, 2018) and visualised using ggplot2 (Wickham, 2016) and study population was used to estimate impact of summer on SCC and
fertility at farm level. This was estimated by calculating the average

3
L.-M. Tamminen et al. Preventive Veterinary Medicine 224 (2024) 106131

Table 2
Overview of models fitted to describe seasonal deviations in somatic cell counts and fertility (pregnant within voluntary waiting period plus 30 days and monthly
calvings). NP = Modelled with non-parametric thin plate regression spline, CC = Modelled with cyclical cubic spline, P = no smooth applied, F = Factor, ECM
= Energy corrected milk.
Outcome Explanatory variables Deviance explained

Somatic cell count (SCC)

Deviation of SCC from yearly herd average (%) Season: Month (CC), Year (F) 8.36%
Average of sampled cows: kg ECM/cow (NP), days in milk (NP), lactation number (NP)
Farm: Herd id (random effect), Average yearly SCC (NP)
Interactions: Month x year, Month x kg milk/cow, Month x Yearly SCC, Yearly SCC x kg milk/cow
Average SCC (1000 cells/mL) Season: Month (CC), Year (F) 32.4%
Average of sampled cows: kg ECM/cow (NP), days in milk (NP), lactation number (NP)
Farm: Herd id (random effect)
Interactions: Month x year, kg milk/cow x days in milk, days in milk x month, kg milk/cow x month
Fertility
Deviation from expected calvings (%) Season: Month (cc), Year (F) 9.1%
Farm: Herd id (random effect)
Interactions: Month x year
Monthly deviation from yearly herd average PV30 Season: Month (cc), Year (F) 3%
(%) Farm: Herd id (random effect)
Interactions: Month x year

summer deviation in SCC and fertility estimates (PV30) from the yearly identify differences in the experiences arising from farms with large and
farm average for 2016 to 2018. From this a list of 100 farms with the little impact. Informed consent was gained from all participating farms
smallest and largest deviation in SCC as well as the 100 farms with ahead of the interview process taking place. The characteristics of the
smallest and largest deviations in PV30 was created. Farms from such list farms participating in the in-depth interviews are presented in Table 3.
with more than 50 dairy cows were purposively recruited to achieve The selection consisted of 18 farms, including 6 cases and 9 controls for
variation in seasonal impact as well as geographical spread via the leading mastitis (Case = Large, consistent seasonal deviations in SCC, Control =
dairy farm advisory company VÄXA Sverige, which has a well-established Small seasonal deviations in SCC) and 7 cases and 3 controls for metritis
trust relationship with dairy farmers (Fig. 2). The purposive sampling (Case = large, consistent seasonal deviations in PV30, Control: Small
from the identified farms was performed with the aim of providing seasonal deviations in PV30). Several farms were combinations, for
contrast in the participating voices and increase the opportunity to example 3 were cases for both mastitis and fertility (Table 3).

Fig. 2. Geographical location of the 18 participating farms for in-depth interviews. Location is an approximation so to maintain participants’ privacy.

4
L.-M. Tamminen et al. Preventive Veterinary Medicine 224 (2024) 106131

Table 3
Description of farms participating in the in depth interviews and farmer’s perceived impact of heat on cows. AMS =Automatic milking system. ECM = Kg Energy
corrected milk. M Case = Large, consistent seasonal deviations in SCC, M Control = Small seasonal deviations in SCC, F Case = Large, consistent seasonal deviations in
PV30, F Control: Small seasonal deviations in PV30.
ID Average kg Average SCC Dairy Milking Barn type Perceived impact of heat on cows and the barn
ECM/cow and (x103 cells/ cows system
day (2018) mL milk) (no.)
(2018)

111223 M: 28.8 216.0 120 AMS Insulated new barn (2020) with fans and Less impact in new barn. Mainly relates problems
ControlF: adjustable walls. to ensuring good feed hygiene and fetching
Case animals on pasture (although not such a big
problem).
111509 30.2 362.0 130 AMS Insulated barn (2004) that opens at the gable. Experiences that animals gather in specific places
M: Control Plans to install fans the coming year. (with better airflow) during hot days. Experiences
reduced feed intake and reduced activity.
111733 29.7 344.0 60 AMS Insulated barn from 2014. Cannot open up Experiences that hot days are a problem because
F: Case sides of barn. cows reduce activity and feed intake.
111756 24.5 248.0 78 Milking Insulated barn from 2007. Gables can be Describes signs of severe heat stress and that
M: Case parlour opened to increase wind flow. animals are lying down during hot days. Uses a
F: Case hose to cool individuals in severe cases.
111842 29.8 268.0 220 Milking Insulated barn from 2008 with fans. Main problem is increased somatic cell counts and
M: Case parlour drop in milk production. Puts cows to pasture
during cooler hours to avoid heat stress.
112094 36.4 254.0 60 AMS Barn from 2001 with insulated roof. Only experiences problems with heat a few days
M: CaseF: per year. Connects reduced production and poor
Case health and fertility with pasture.
112116 26.9 197.0 120 AMS Older, rebuilt barn (1994) without insulation Mainly impacted by heat and drought in 2018.
M: Control and some opportunities to open up windows Traffic goes down every summer, relates this to
and doors. However, ventilation could be pasture. Animals change their daily pattern and
better according to farmer. eat during evening/early morning during hot
days.
112606 30.3 275.0 240 AMS Relatively new barn (2017) with good cooling Experiences high impact of heat, lost animals
M: Case opportunities. because of mastitis related to hot spells in 2018
and struggles with feed hygiene.
112700 27.1 350.0 65 AMS Insulated barn from 2009. Ventilation that Experience a drop in milk, increased cell counts
M: CaseF: adapts to wind and temperature by adjusting and that cows are inactive and panting during hot
Case wind flow. days. Tries to cool animals with water hose during
warm days.
113064 27.8 276.0 220 Milking Barn from 1999. Experimented with sprinkler Has less problems after installing fans 2019.
F: Control parlour cooling during 2018. Before animals would reduce activity during
warm days.
113198 22.0 293.0 130 AMS Barn from 2009 with insulated roof and Has experienced problems with droughts and
M: Control adjustable curtains on the side. poor feed availability as a result of weather.
113268 24.8 297.0 130 AMS Non-insulated old barn with high ceiling. Experiences negative impact shorter periods
M: Control Open sides and two fans over cubicles. Cows during summer. Cows drop in activity, milk and
feed outside under shade. more difficult to get pregnant.
114258 27.6 304.0 95 Milking Insulated barn from 2007. Extra fans after Experiences that cows reduce activity and feed
M: Control parlour summer of 2018. Experiments with pasture at intake during hot days.
F: Case night and a homemade sprinkler system.
114490 29.7 89.5 105 AMS Insulated barn built 2000. Windows along the Experiences that the barn remains cool during
F: Control sides that open and good natural ventilation. warm days. Allows cows to pasture at nights
during the warm periods to avoid exposing them
to heat.
114546 31.8 130.0 125 AMS Barn from 2011 with insulated roof, curtains Experiences that cows are suffering because of the
M: Control that can be opened and fans. heat and that mastitis is more common. Also that
fat content drops in milk.
115843 28.7 253.0 170 AMS Non insulated barn from 2012 with curtains Experiences problems with reduced cow activity
M: Case that open on sides. leading to longer milking intervals during
summer. Also problems with feed hygiene.
115939 23.6 289.0 58 AMS Non-insulated barn built 2008. Fans installed After fans were installed cows prefer to stay
M: Control after 2018. indoors in warm weather. Less problems with
F: Case reduced activity.
115997 26.5 228.0 243 AMS Insulated barn built 2011. Limited Experiences that heat causes problems in activity
M: Control opportunities to open up to ensure wind flow and cow traffic but also indirectly through
M: Control through barn. negative effects on feed, which in turn causes
health problems.

2.4. Interview process, qualitative analysis and mixed methods integration Lastly, d) they described the current/future perspectives on how to cope
with the challenges. Farm interviews were done in Swedish either on-
The interview guidelines were developed by the last (GOA) and first farm face-to-face (n = 11), via videoconference (n = 2) or over the
authors (LMT) in consultation with the co-authors (RB, MÅ) and other phone (n = 5) during spring 2021 by the first (LMT) and last author
field advisors working closely with Swedish dairy farmers during 2018 (GOA). When conducting these interviews, we adhered to an interview
events. Interviews were structured around four topics; a) the general agenda to ensure each of the topics was addressed. Yet, the pace and
health situation of the farm, b) their definition/experiences of EW, c) a flow of the interview was guided by what we as interviewers interpreted
description of any on-farm challenges associated with weather events as meaningful to the interviewee, thus weaving in and out of topics as
and specific challenges related to animals and heat/the summer season. necessary (i.e. semi-structured). The interviews lasted on average one

5
L.-M. Tamminen et al. Preventive Veterinary Medicine 224 (2024) 106131

hour and they ranged from 40 min to 1.5 h. The interviews were fol­ 2007) where synthesis and contextualisation of the data are reported as
lowed by a short survey section (data not shown here) and a walk on the a joint section (Braun and Clarke, 2022, 2019; Byrne, 2022). This is a
farm when interviews were done on site. At the end of each interview significant departure from the traditional reporting convention followed
LMT and GOA discussed their impressions and took reflective notes that by research with either a positivist or deductive-oriented thematic
were later revised when themes were constructed. All interviews were analysis stance inconsistent with the framework and methodology of our
voiced recorded and manually transcribed verbatim in Swedish. Only study.
representative phrases for exemplification of the identified themes were
translated into English by LMT. The transcriptions, field notes and post- 2.4.1. Authors’ positionality statement
interview notes (memoing),were open-coded with the aid of Dedoose All authors of this paper have experience in and connections to the
application for managing, analysing and presenting qualitative and dairy sector in Sweden and abroad. LMT, GOA and RB have a veterinary
mixed methods research data (Version 9.0.85, 2021; Los Angeles, CA: degree. MÅ works as an expert cattle nutritionist at Växa Sweden and
SocioCultural Research Consultants, LLC www.dedoose.com) by the last has a graduate degree in animal nutrition and management. RB holds a
and first authors. A Reflexive Thematic Analysis (Braun and Clarke, professorship in domestic animal reproduction and is a European
2022, 2019; Byrne, 2022) was used meaning the qualitative data set was Diplomate in animal reproduction, subspecialty ruminant reproduction
analysed predominantly inductively within one inductive/deductive and herd health, LMT has a graduate degree in epidemiology and GOA a
continuum where quantitative phase results were part of the analytical graduate degree in applied animal behaviour and welfare; all fields
process. Thus, the qualitative dataset was open-coded, and farmers’ traditionally quantitatively oriented. However, LMT and GOA regularly
meanings were emphasised around central themes arising from the data, conduct social enquiries or mixed methods research guided by different
curated according to relevance of the research questions. Coding went paradigms. Moreover, GOA holds additional training in social sciences
beyond the descriptive level of the data in an attempt to identify connecting traditional epidemiological research with qualitative ap­
meanings, underlying assumptions and ideas of farmers around EW, proaches to enhance understanding of stakeholders’ decision-making
heat and seasonal changes and its impact to their animals. For this related to animal welfare. Her broad aim is to bridge scientific under­
analytical process, the first and last authors familiarized themselves with standing of risk factors of poor health and welfare to contextualized
the field notes and transcribed text independently generating initial practical solutions by highlighting contested knowledge among animal
codes capturing the salient features of the dataset. Through jointly caretakers and a way forward. LMT is a postdoctoral fellow aiming to
interactive discussions that included the third author comments, initial improve the welfare of dairy cattle by conducting research with people
themes were drafted and later on revised for a final selection. who care for animals. RB acknowledges the need to investigate dairy
The final detailed themes were jointly constructed by LMT and GOA. producer and animal caretakers perspectives through interviews, to
Themes integrated the qualitative and quantitative phases/datasets of facilitate a better uptake of scientific evidence and reducing the practical
the study following a meta-theoretical critical realist stance (Scott, know-how gap in Sweden and internationally. All authors believe there

Heat on-farm is experienced in relation to other EW events within a production year

Heat is just one type of EW

that causes challenges to
farmers

Success in handling EW
challenges is measured in
ability to produce enough
feed for animals each
year

The biological consequences of heat are too discrete to trigger action

2016-2019
C
2016-2019

Warmer summers are Biological costs related to

associated with more welfare and fertility are
severe biological The effect of heat stress on not clearly perceived
consequences individual cow is not in by farmers
farmers’ focus
Actions performed during warm
Preventive measures and coping strategies periods aim to maintain
are overarchingly reactive production and milk quality
C

in the face of heat

Preventive actions are mainly reactive,
taken to reduce immediate effects of
heat, not long-term consequences

Increased perception of consequences Active cooling is an

on farm could support action and a shift overlooked, important
to preventive measures piece of the puzzle

Fig. 3. Thematic map visualising the three prominent themes derived from analytical integration of quantitative and qualitative results.

6
L.-M. Tamminen et al. Preventive Veterinary Medicine 224 (2024) 106131

are no irrational farmers but a lack of contextualized understanding of some point. But it doesn´t have to be good, but it will… yes, but somehow it
their needs and aim within their farm. will work” (F115939)

3. Results and discussion

3.1.2. EW impacts feed production but farmers experience a fragile
resilience
Our aim was to provide a comprehensive picture of how Swedish
While farmers recalled various type of EW events, droughts due to
dairy farmers’ experience extreme heat and other EW events. Moreover,
heat, like those recorded during the summer of 2018, stand out in
by exploring the unusually hot and dry year of 2018 in particular, we
farmers’ memory as extra challenging for them. These EW events are
aim was to identify if farmers are aware and if so enact any preventive
perceived as more severe or having more long-lasting consequences. But
pathways to reduce the biological cost that heat stress can cause to dairy
as described before, there appears to be an overarching confidence to­
cattle when faced with hot weather. From the analytical integration of
wards the described EW events, often referring to how everything
quantitative and qualitative results three prominent themes were con­
”worked out in the end” despite the circumstances observed. This in­
structed that summarise the main topics from the national statistics in
dicates that the farmers had been buffering their problems within their
light of farmers’ experiences from extreme weather events. The three
farm means. The recurring statements of confidence reflect a certain
themes are presented in Fig. 3 and revolve around how: 1) heat on-farm
positive resilience to what farmers perceived as EW, albeit a fragile
is experienced in relation to other EW events within a production year,
status. The fragility of such resilience is based on individual farm ability
2) the biological consequences of heat on the herd have too discrete
to produce feed (i.e. excess land) or buy it (i.e. savings) despite the EW
consequences to trigger action and 3) preventive measures and coping
demands. Thus, the farmer´s capacity to absorb the disturbance in feed
strategies are overarchingly reactive. We develop these ideas further in
production resulting from EW events was key to the perceived farm
the next paragraphs and discuss the consequences and way forward.
stability in the long term.
3.1. Heat on-farm is experienced in relation to other EW events within a “. There are consequences in the economy that you have to buy in feed.
production year You don’t have the room to not manoeuvre that. you might be able to buy
feed, but is it financially worth it? You judge. Of course it has conse­
Conversations highlighted that heat is not the only experienced EW quences and it has consequences for those who work here too [referring to
that has affected Swedish farms in recent years (recalled events between the cows]. It can be anything. If you have problems with feeding”
2014 to 2019 included). Although dairy production in Sweden is an all- (F115843)
year round feature, it was clear that participants discussed on-farm
“[referring to the summer 2018]. We were behind with feed, … So that’s
events like harvest, silage production, winter feed usage as a circle of
when you buy grass from everywhere of varying quality. then the third
yearly events (e.g. 2018 harvest/silage production vs. 2017) where each
crop was probably the best quality we got then, so I could give it to the
year needed to have certain success for the year to be recalled as good or
dairy cows, but by then there were other consequences. The animals had
poor (Holland and Kensin, 2010). Farmers’ memory of the last few years,
been a bit heat stressed and we had problems with pregnancies and things
including the dystopic summer of 2018, conveyed an overarching pos­
like that” (F111509)
itive resilience towards the EW events experienced based on their ability
to resolve difficulties faced and their prognosis to continue doing so That feed is a prioritised area for Swedish farmers has been previ­
individually or as a community. A farm´s ability to solve difficulties was ously suggested (Barth and Melin, 2018). Animal production in Sweden
linked to their capacity to feed their animals, a priority for maintaining has been classified as having a relatively low vulnerability to climate
milk production and financial stability. Thus, feed production on-farm change and Northern Europe has been identified as a region where crop
was a key priority over other events in the herd as well as a barometer yield may increase with climate change (EEA, 2017; Horn et al., 2022).
of success. The construction of the theme is supported by two key ideas: Yet, the farmers in the study were already expressing challenges with
securing animal feed and associate the shifting weather patterns with
3.1.1. Heat is just one type of EW that causes challenges to farmers problems. This suggests that the vulnerability of Swedish dairy pro­
Although farmers generally considered heat and drought as impor­ ducers may be underrated and more comparable to the medium
tant types of EW their definition of EW is considerably broader than this. vulnerability projected for production of grains (Horn et al., 2022).
Farmers often mentioned examples of how rain, heavy wind and snow
fall affected their farm routines within a year thus, also being considered 3.2. The biological consequences of heat on the herd are too discrete to
extreme events. Thus the EW experiences are marked by how they occur trigger action
in relation to other events in a year and across the years.
While farmers described how EW events, specially droughts due to
”Yes, there are different kinds of extreme weather. It can be drought and
heat, had negative consequences at farm level; there was little discussion
it can be rain, but also snow. Wind. We rarely have snow, we’re getting
over the biological cost these events may have on their animals. If so,
less and less problems with that. But wind, rain and drought are causing
this was stated in terms of productivity and milk quality (i.e. increased
the most problems for us” (F111509)
SCC) rather than physiological or welfare impacts to their animals (e.g.
“In 2014 …there was this big forest fire. As it often is when these things reproduction success). Nevertheless, national herd statistics confirms
happen, it doesn’t feel like it’s real in any way, it just happens…a that extreme heat (i.e. 2018 spring/summer) had an impact that can be
powerful experience, absolutely! But it went well. It’s extreme weather separated from cyclical management changes related to spring/summer
that has created these scenarios anyway. And then. well, when we get to months. This period was recognized by farmers as the busiest periods on-
the drought of 2018 that also has. it has long-term consequences, even if farm and they described trade-offs as well as in adverted consequences
you sort it out for the moment.” (F111509) to the herd. This theme acknowledges that such circumstances shade
farmers’ ability to promptly detect and act to mitigate the effects of heat
“…when you can´t even go out in the fields [referring to heavy rain], then I
on the herd. In particular fertility aspects that generate the negative
would say that´s extreme…if it´s so dry that there´s no grass to pick up when
consequences expressed by farmers (i.e. disrupted calving patterns and
you´re going to do the second harvest, then that´s also extreme, it seemed to
stressful workflow).
me… I guess it´s up to each person how sensitive they are to what they think
This theme is constructed around five subheadings illustrating SCC
is extreme…Now you know that it almost always works out somehow at
and fertility national trends and farmers interpretations of heat stress in
cows and milk quality and fertility consequences on their herds.

7
L.-M. Tamminen et al. Preventive Veterinary Medicine 224 (2024) 106131

350

325

Predicted average Somatic Cell Count

300 Year
2016
2017
2018
275 2019
(1000 cells/ml)

250

1 2 3 4 5 6 7 8 9 10 11 12

Month

Fig. 4. Predicted herd average somatic cell count by month and year according to Generalized additive model by month (p < 0.001). Shaded areas represent 95%
confidence intervals.
Predicted deviation from yearly farm average in PV30 (%)

0.05

0.00

Year
2016
−0.05
2017
2018

−0.10

1 2 3 4 5 6 7 8 9 10 11 12
Month

Fig. 5. Predicted seasonal deviation (%) from farm average of pregnancy rate within 30 days after the voluntary waiting (PV30) according to Generalized additive
model by month (p < 0.001). Shaded areas represent 95% confidence intervals.

3.2.1. National trends and impact of the summer of 2018 2018 compared to 2016 and 2017 (Fig. 5). It should be remembered that
While ensuring feed availability was the priority when facing EW- PV30 is a composite measure combining management and physiologic
events, farmers expressed that they perceived negative impacts of on factors. Thus, the deviation observed can be a result of either unsuc­
herd profitability during warm periods in general and 2018 in partic­ cessful inseminations, missed inseminations or inseminations at subop­
ular. Exploring the seasonal pattern in Sweden using national data from timal times, all to which heat stress can contribute (Rhoads, 2023). On
SMDRS confirms a recurring negative impact on udder health and the other hand, the difference between 2018 and the other years may not
fertility during summer. For SCC a seasonal trend where the average SCC only be related to heat. For example active decisions to not inseminate
is lowest in March to peak in August was found (Fig. 4). The proportional animals due to feed shortages may have contributed to the observed
deviation in SCC to yearly farm average followed the same pattern (S1, reduction in fertility. For SCC on the other hand it is likely that feed
Table 1). Modelling of seasonal deviation in the cows ability to conceive shortages would have led to increased culling of animals with poor
at the expected time point, using the key performance indicator PV30, udder health. This would decrease SCC compared to other years but
shows that a smaller proportion of animals that are inseminated during instead the SCC remain higher in 2019. As heat stress in late lactation
the summer months become pregnant within the herd specific voluntary can have carry over effects to the coming lactation, both in terms of
waiting period (Fig. 5). For both average SCC and deviation in PV30 productivity as well as health status, there is likely some lingering effects
there was a significant interaction between month and year (p < 0.001). from the summer of 2018 (Dado-Senn et al., 2019).
SCC was higher during the summers of 2018 and 2019 (Fig. 4) and Other interactions included in the final models on SCC suggested that
fertility was more severely impacted the unusually warm summer of farms with lower average production per cow and higher average of

8
L.-M. Tamminen et al. Preventive Veterinary Medicine 224 (2024) 106131

days in milk during summer generally had larger deviations compared to winter, while the distribution of farms with lower SCC in winter
other farms. Yet, these effects were small compared to the effect of appeared to experience smaller differences throughout the year (Fig. 6).
month and year (S1, Fig. 1). Model residuals showed no signs of auto­ This uneven impact on SCC indicates that there are farms that manage to
correlation when month and year were included and residuals were avoid negative consequences during summer and future studies should
normally distributed and heteroscedastic. However, the predictive ca­ look deeper into differences between these two types of farms. However,
pabilities of the models were relatively poor indicating that there is August 2018, an unusually hot month, stands out from the general
unexplained variation in the data (Table 2). As both fertility and udder pattern and a higher proportion of farms appear to have been affected by
health are complex and management-dependent it is not a surprise that higher SCC. Thus, farms that are currently managing may need to adapt
the models are not able to explain all variation. Thus, models should be to face increasing number of heat waves in the future.
interpreted with caution as they indicate seasonal trends in the popu­
lation and do not represent accurate predictions for individual farms. 3.2.3. Perception of biological consequences of heat related to fertility
Full details of the models outputs can be found in the supplementary Compared to the impact on SCC, where the link to an increase during
material (S1, Table 1-4). the summer season was quickly recognised by all farmers, negative
impacts on fertility often required targeted elicitation before being
3.2.2. Farmers clearly perceive the connection between heat and SCC acknowledged. Here farmers described and confirmed problems asso­
In discussion with the farmers, a clear link between summer and hot ciated with poor oestrus expression in cows, linked to difficulties in
spells on SCC was expressed. However, while all farmers recognised the getting cows pregnant during summer. The consequences of impaired
impact, their perceptions of control differed. On one hand we had fertility during hot spells/summer were associated with difficulties in
farmers that expressed the situation as an unavoidable effect of summer: recruiting enough animals and in some cases added costs of buying new
animals. In particular fertility problems and consequences were
“Yeah, It’s inevitable. it’s. summer cells [referring to SCC] - we’ve had
mentioned in relation to the summer of 2018, which also stands out as a
that as our big dilemma” (F111842)
deviating year according to national data (Fig. 6), but also in relation to
Others discussed an active engagement in the challenge of keeping warmer periods during ordinary summers.
cell-counts down by naming the many reasons why this could happen as
“The number of inseminations per pregnancy went up a lot [referring to
well as potential areas of action associated with varying degrees of
summer of 2018]. Then we had… Well, specifically heifers that did not
success.
show any signs of heat at all. But we bought some new material, then. You
“Interviewer: You say it’s hard to keep low cells during the summer and want to avoid that, but that’s what we did anyway” (F111509)
feed hygiene was one reason. What else? Farmer: There are a lot of small
“No, they don’t show any heat [referring to summer]. And it is harder to
streams. We get a little less cow traffic in the summer, always have.
get them pregnant when it is very hot” (F111223)
Usually drop down there to 2.3–2.4 milkings per day. And it goes without
saying, the [milking] interval gets longer …that’s where you get the big “All aspects of fertility are affected [referring to producing dairy cows].
problems [referring to SCC] or challenges with it” (F115843) The oestrus behaviour is weak and often they do not get pregnant during
the hot period. You can see that clearly nine months after that yes, that
“Interviewer: When you notice the reduced cow traffic to the milking
was the hot weeks in July. That’s why we do not have any calvings right
robot and reduced feed intake. What do you do? Farmer: If I do some­
now.” (F113268)
thing? Interviewer: Yes, to improve traffic… Farmer: No it is more about
making sure that they have fresh feed inside and not old, disgusting… Feed The difficulty to get cows pregnant in summer can be related to the
quickly becomes warm when it’s hot. So it is better to provide less more cows expressing poor oestrus behaviour and warm temperatures having
often. During winter we feed once a day, in summer you can give… If it is negative effects on the oocyte development, implantation as well as
warm and they do not want to eat outside they get fresh feed in the uterine environment (Sammad et al., 2020). Separating these effects is
morning and afternoon. And maybe you clean in between. During winter not possible in this study but the discussions with farmers suggest that
we clean every third or fourth day” (F112700) they face a combination of problems. However, previous literature
suggests that inseminations per cow start decreasing when daily average
Also on a national level a variation in impact on SCC across farms can
THI is above 57 while successful pregnancies per insemination decrease
be observed (Fig. 6). The negative effect of summer on SCC was mainly
at higher temperatures (THI 68) (Gernand et al., 2019). When daily
observed as a shift in the population of farms with higher SCC also in

Fig. 6. Distribution of average herd somatic cell counts from the 1262 farms participating in the Swedish Milk and Disease Recording System between 2016–2019.

9
L.-M. Tamminen et al. Preventive Veterinary Medicine 224 (2024) 106131

1.2

Predicted deviation in calvings (%)

1.0 Year
2016
2017
2018
2019

0.8

1 2 3 4 5 6 7 8 9 10 11 12
Month

Fig. 7. Predicted deviation (%) from expected average farm calvings recorded in the Swedish milk and disease recording scheme between January 2016 to
December 2019.

average THI is 73 the conception rate falls drastically (Schüller et al., Uneven calvings do not only impact the number of young calves on
2014). Thus, it is possible that the unusually warm months of 2018 the farm, it also has an impact on the entire cattle herd dynamics,
caused additional challenges beyond the difficulties of observing oestrus lactation patterns of the cows and the flow of milking animals in the
behaviour which led to the observed deviation. barn. Especially in farms with automatic milking systems (AMS) this
even flow is important.
3.2.4. The long term consequences of reduced fertility are recognised but
”In the conventional system, it did not matter so much if ten cows calved
not connected to heat
in a few days. But with the milking robots it much more sensitive and you
Looking at the deviation from expected number of calvings (based on
want to have an even flow and try to match the cows that calve with the
monthly average of total calvings) on a national level, there was a clear
number of cows that are dried off” (F113268)
seasonal trend where the farms have fewer calvings in spring, with some
variation between years (Fig. 7). One of the participants made a clear The seasonal shift, leading to a peak in calvings in late summer/early
connection on how the difficulties during summer lead to problems fall, means that a high proportion of animals are at the final stages of
down the line: pregnancy during the warmest months. Heat stress during this period is
associated with long term effects on both dam and calf, for example
“Interviewer: How did it work with pregnancies and inseminations during
reduced production in subsequent lactation and changes in immune
that period [referring to 2018]?
responses (Dado-Senn et al., 2019; Dahl et al., 2020, 2019). Thus, while
Farmer: Yes, well that is the problem then, that they don’t become… they farmers focused on the immediate, recoverable impacts of heat stress on
don’t show heat [referring to the animals]. And when we approach the profitability, the long term consequences of reduced fertility in summer
winter, or fall and winter, then they are all in heat so to speak. And this can impact farm profitability for many lactations and across generations.
leads to an uneven flow. But you don’t want to… prefer not to… When
you are on the borderline with numbers of animals you don’t want to lose 3.2.5. The challenges for fertility are multifactorial
a single one which means that you keep inseminating and it becomes a Farmers reflected on heifers in particular and as a separate entity
vicious circle… Then they all [referring to calves] come in the wrong from the multiparous cows, especially in relation to insemination and
month so to speak, all of them” (F112606) fertility routines during summer. This group of animals was associated
with additional practical challenges in addition to difficulties in
However, these type of observations that linked/discussed the long-
observing oestrus or achieving pregnancy. The farmers also reflect
term consequences of summer/heat on calving patterns were rare.
around unevenly distributed calving patterns and connect it to fewer
Nevertheless, calving patterns was a recurrent topic, where most farmers
heifers getting pregnant in summer.
mentioned how uneven calving patterns was a challenge to their farm in
general. “Yes it is a little easier [referring to fertility of dairy cows]. You see the
cows more than you see the heifers in summer” (F112700)
“No, I don’t experience any [referring to fertility/health issues] as
problem, but a little uneven calving. When there are 30–35 calving in a “You should have the same number of calvings every month all year
month, and then there are only 10 the month after. That’s a challenge. round. We don’t really have that… Well, it’s better to have few cows in the
Last autumn and up until the turn of the year here now, we had such a summer than in the winter. But the best thing would be to have full all year
huge peak with a lot of calvings. And then you get to a different infection round. But then it has to do with this then that you are inseminating less.
pressure in calves…” (F111223) Fewer heifers get pregnant from May to October or September ”
(F.115939)
“You get out of that swing [referring to calving pattern], you might only
have two or three calving one month, and then the next month you might “It is difficult to say but heat makes it more difficult [referring to in­
have 20. And that means we can’t have a full stocking, we get disruption. seminations]. That is clear. And then it does not help that we cannot keep
works much better with the flow, with calves and weaning everything an eye on them, since they are outside. [….] For the heifers we use a bull,
becomes much easier” (F111509) since we can only keep them inside for two weeks” (F112116)

10
L.-M. Tamminen et al. Preventive Veterinary Medicine 224 (2024) 106131

2016 2017 2018 2019

12000

9000

Number of calvings
6000 Cows
Heifers

3000

0
April - Aug April - Aug April - Aug April - Aug
Month

Fig. 8. Number of calvings recorded in the Swedish milk and disease recording scheme between January 2016 to December 2019.

Calving patterns on a national level confirm farmers’ experiences All farmers participating in the study stated that heat has had an
through a recurring pattern where the number of heifers calving de­ impact on their animals (Table 3). The majority discussed the effects as
creases from February and remains low until June (which corresponds to behavioural changes, especially related to reduced activity, while others
becoming pregnant between May and September) (Fig. 8). This pattern also described panting and other severe signs of heat stress. Biological
is highly consistent throughout the 4 years, which supports that there costs such as decreased production, decreased udder health and imme­
are recurring important management related factors unrelated to diate impact on fertility were recognised but sometimes connected to
weather that influences it. For the multiparous cows on the other hand, a management changes (pasture and feed hygiene). It is interesting to
less consistent pattern was observed. This suggests that there is more notice that two of the farmers (112700, 111756) described severe
variation between years for the multiparous cows, possibly a result of symptoms of heat stress and that they used a water hose to cool the cows
the variation in temperatures across years and the build-up of previous that were affected. By the signs described the cows on these farms
years that remain unseen/unremarked to farmers. experienced severe heat stress and farmer reactions were clear examples
of reactive handling of heat. However, there were also examples of
3.3. Preventive measures and coping strategies are overwhelmingly proactive handling. Farmer 113064 had experimented with sprinkler
reactive systems and purchased fans. Farmer 115939 made the decision to pur­
chase fans after the summer of 2018 and saw smaller effects of summer
Participants’ described actions to reduce the impact of heat (i.e. heat on his animals. Thus, from what the farmers describe, heat is impacting
stress) on farm are overwhelmingly a reaction to an immediate reduc­ dairy cows in Sweden on a regular basis although 2018 was more severe.
tion in profitability, reinforcing that the long-term biological costs are However, the large variation in what is perceived as problematic is
not visible or attracting attention. Furthermore, conversations also noteworthy and suggests that there is considerable variation in how heat
highlight some farmers with a self-believed inability to act on the impact stress is perceived among Swedish farmers.
of seasonal heat has on their herd. Altogether we characterize current
Swedish farmer coping strategies as reactive. Yet, pro-active changes are 3.3.2. Differences in perception of biological costs influence coping
slowly gaining traction derived by a collective memory of heat waves strategies and decision-making
and belief they will recur in the future. The question remains if such Ensuring feed was farmers’ first priority when facing heat waves and
changes are enough for the events to come. other EW events and in discussions farmers described a mix of both
reactive and proactive actions. Reactive measures were described
3.3.1. Current control of heat among participating farms especially in relation to 2018 and farmers described new, proactive
Although the number of participants in the qualitative part of the strategies for ensuring feed security in a long term perspective being
study is too small to draw quantitative conclusions, some interesting implemented after 2018. Some farmers also described how they actively
general trends were observed in the farm characteristics and farmers’ worked to try to prevent the seasonal increase in SCC. The actions
perception of the impact of heat (Table 3). The majority of the barns on described were mainly reactive, handling the situation as it arose, and
the participating farms were insulated (only roof or roof and walls) and were connected to what farmers perceived to be the underlying reason
heat was generally controlled by natural ventilation. However, the for the seasonal increase in SCC. For fertility few actions to prevent the
newer barns (e.g. 111223) were built with cooling in mind as they had negative effects were described, although farmers are aware of the im­
curtains that can be opened. In addition, the most modern barns had fans mediate negative impacts.
and some farmers with older barns had invested in fans after 2018. In According to the Health Belief model, adoption of proactive behav­
most of the barns, farmer stated that the airflow could be modified to iours is connected to the perceived threat and susceptibility as well as
some extent by opening windows or doors, but to which extent this was the perceived benefits and barriers (Janz and Becker, 1984). As the focus
possible varied. Selection of study participants was performed to achieve of farmers was securing feed for the animals it is not surprising that this
variation in the impact of summer/heat as well as geographical variation area was also connected with preventive behaviours. Regarding the
and not farm characteristics. Thus, it is positive to see that variation in biological costs the conversations suggest farmer perception of conse­
farm characteristics, reflecting the general large variation observed in quences of reduced fertility is more indistinct than for the SCC issue,
Swedish dairy production, was achieved in the study. possibly because of the time lag between heat event and consequences.

11
L.-M. Tamminen et al. Preventive Veterinary Medicine 224 (2024) 106131

Whereas the increase in SCC is relatively immediate, the consequences that cows should live in cooled environments. As seen in Table 3 all
on fertility do not materialise until long after summer. Therefore the farmers described episodes of heat stress, thus the necessity to preven­
consequences (and severity of the consequences) may be more difficult tively cool the cows also in the northern hemisphere, where the col­
to perceive. An inability to see the consequences may impact the lective norm that cows do not need cooling is stronger compared to
perceived benefits of acting to control a threat which, in combination tropical areas, should be recognized. As farmers describe heat stressed
with perceived barriers, are an important part of how effectiveness of animals in all kinds of barns it appears that what has been seen as
preventive measures are perceived (Janz and Becker, 1984; Svensson optimized design for Sweden may no longer be enough.
et al., 2019). In addition, the practical barriers associated with fertility The aim of this study was to increase understanding of how dairy
may also contribute to why farmers choose to focus on SCC instead. farmers perceive and act in the face of climate change, increasing risk of
Another possible explanation of the differentiation of perception of heat waves and other types of EW. To identify how farmers perceive
consequences and actions towards of SCC vs. fertility is related to the yearly as well as extreme challenges a mixed methodology where the
emotional content attachment (Holland and Kensin, 2010). Such individual farm and national patterns were integrated and understood in
attachment may influence the way in which the event is remembered in relation to each other was necessary. This approach complements
relation to the past and present goals. This influences the recall of the existing literature of the effects of heat stress in dairy cows with new
event (i.e. memory), the believed consequences and can thereby impact perspectives and insight into the needs to act also in areas where heat
future goals. SCC is linked with milk quality and it is a constantly has not been considered the main challenge historically. However, the
monitored key indicator of the industry. As such, SCC variations can study design also has limitations. For example the 18 farms participating
have direct short-term consequences on the financial stability of a farm. are too few for identifying successful coping strategies on a national
Thus variation SCC during a stressful year, such as 2018, could easily level. Future studies including a larger sample size should focus on for
affect a farmer’s ability to cope with other stressors (i.e. buying feed) example identifying which types of barns and cooling systems work best
and future reinvestment on-farm for self-improvement (i.e. a new barn). under Swedish conditions. In addition, the potential of adding water
This is similar to how the consequences of reduced feed availability was based cooling systems should be explored.
described. Variation in fertility on the other hand lacks direct industry
penalty and recognising the negative impact on the farm requires 4. Conclusion
follow-up of long-term consequences. Therefore, emotional attachment
is presumed higher for SCC variations compared to fertility variations in Overall, conversations with farmers indicate a broad definition of
a given year. extreme weather, where its handling is reactive and the memory of the
events do not trigger a big concern for the biological cost of the events to
3.4. Moving forward – coping strategies and decision-making needs to the cows. There is a perceived lack of control of the events, i.e. some­
include biological costs thing that happens and must be handled as it comes but cannot be
planned for. However, at a national level, long-term effects on the farm
Our assumption that fertility is linked with a reduced emotional economy, health and herd dynamics indicate that attention must be
attachment, and thus less prioritised, is substantiated in the fewer so­ given to heat waves normally seen during the spring and summer season
lutions to reduce negative effects described by the participants. Instead, in Sweden. While Swedish dairy farmers are aware of negative impacts
costs of additional recruitment of cows was discussed. Overall, this of heat in particular, mainly expressed in production and SCC, they show
suggests that fertility is currently the least prioritised area and to a large fragile resilience. Their resilience is built and expressed around their
extent handled reactively with little thought to long-term effects. This is capability to ensure sufficient feed when extreme weather events hit the
problematic as good reproductive performance does not only enhance farm. Yet, the currently unacknowledged long-term biological costs
management control, but is also an important factor for herd profit­ related to fertility has the potential to visualise the need for proactive
ability (Gröhn and Rajala-Schultz, 2000). planning and thus improve farm resilience and profitability based on
While the observed shift in calving intervals may increase risk of long efficient farm dynamics in addition to feed supply. This might imply a
term consequences for productivity and health there may be other ef­ re-design of current and future barns for better and active cooling.
fects that support heat mitigation in practice. For example, fewer calv­
ings in spring could lead to fewer animals in the barn which can decrease Funding source
the risk of heat stress the sequent summer (Noordhuizen and Bonnefoy,
2015). However, reduced animal density only occurs if animals in late This work was supported by the Swedish research council FORMAS
lactation are also removed and some participants suggested the oppo­ [grant number 2018–02815].
site, thereby also exacerbating the risk of increased SCC during summer
by retaining late lactation animals or animals with high cell counts. CRediT authorship contribution statement
Similarly a low number of recruitment heifers may lead to farmers
keeping older cows which are, according to national animal health Olmos Antillón Gabriela: Conceptualization, Data curation, Formal
statistics at higher risk of developing mastitis (van den Borne et al., analysis, Funding acquisition, Investigation, Methodology, Supervision,
2010). Thus, proactive handling of fertility and animal flow on farm Writing – original draft, Writing – review & editing. Åkerlind Maria:
could potentially also decrease the negative impact on SCC. Conceptualization, Funding acquisition, Resources, Writing – review &
With increasingly warm summers that pose challenges on main­ editing. Båge Renée: Conceptualization, Funding acquisition, Re­
taining dairy cow fertility, awareness of seasonal trends on the farm and sources, Writing – review & editing, Formal analysis. Tamminen Lena-
their long term consequences are important to highlight in order to Mari: Data curation, Formal analysis, Investigation, Methodology,
prevent further escalation as well as preventive work to minimise ef­ Project administration, Visualization, Writing – original draft, Writing –
fects. Therefore there needs to be an increased awareness of how an review & editing, Conceptualization.
uneven calving patterns develops on farm and the consequences it has.
Developing tools that aid farmers visualise long-term effects on fertility Declaration of Competing Interest
and herd dynamics as a result of heat stress, may aid to create the
necessary emotional attachment for farmers to re-evaluate the prioriti­ The authors declare no conflicting or competing interests.
sation of actions and promote proactive planning to enhance future on-
farm resilience by integrating fertility measures for decision making. We
also propose that fertility decision making should not be blind to the fact

12
L.-M. Tamminen et al. Preventive Veterinary Medicine 224 (2024) 106131

Data availability Hoffmann, G., Herbut, P., Pinto, S., Heinicke, J., Kuhla, B., Amon, T., 2020. Animal-
related, non-invasive indicators for determining heat stress in dairy cows. Biosyst.
Eng. 199, 83–96. https://doi.org/10.1016/j.biosystemseng.2019.10.017.
The datasets for this article are not publicly available because of Holland, A.C., Kensin, E.A., 2010. Emotion and autobiographical memory. Phys. Life
privacy and confidentiality reasons. Requests to access the datasets Rev. 7, 88–131. https://doi.org/10.1016/j.plrev.2010.01.006.
should be directed to the corresponding author and Växa Sverige Horn, B., Ferreira, C., Kalantari, Z., 2022. Links between food trade, climate change and
food security in developed countries: a case study of Sweden. AMBIO 51, 943–954.
([email protected]), respectively. https://doi.org/10.1007/s13280-021-01623-w.
Hut, P.R., Scheurwater, J., Nielen, M., van den Broek, J., Hostens, M.M., 2022. Heat
stress in a temperate climate leads to adapted sensor-based behavioral patterns of
Acknowledgements dairy cows. J. Dairy Sci. 105, 6909–6922. https://doi.org/10.3168/jds.2021-21756.
Janz, N.K., Becker, M.H., 1984. The health belief model: a decade later. Heal. Educ.
The authors would like to thank Prof. emeritus Ulf Emanuelson and Behav. 11, 1–47. https://doi.org/10.1177/109019818401100101.
Johnsson, B., Mattsson, K., Søyland, V., Krekling, M.F., Nemming, A., Vainio-Mattila, B.,
the Swedish research council FORMAS for providing funding of the Reskola, V.-P., Högman, S., Steffensen, L.L., Jóhannesson, T., Rönnberg, J.,
study and the farmers that agreed to participate. We would also like to Hansson, Per, Tunberg, M., Hansson, Philip, 2019. Det nordiska jordbruket –
thank the advisors at Växa Sverige for aiding in the recruitment of utmaningar i en framtid präglad av mer extremväder. Nordisk Ministerråd,
Copenhagen. https://doi.org/10.6027/TN2019-536.
farmers to the study. Löf, E., 2012. Epidemiological Studies of Reproductive Performance Indicators in
Swedish Dairy Cows. Swedish University of Agricultural Sciences.
Lüdecke, D., 2018. ggeffects: tidy data frames of marginal effects from regression models.
Appendix A. Supporting information
J. Open Source Softw. 3, 772. https://doi.org/10.21105/joss.00772.
Noordhuizen, J., Bonnefoy, J.M., 2015. Heat stress in dairy cattle: major effects and
Supplementary data associated with this article can be found in the practical management measures for prevention and control. SOJ Vet. Sci. 1, 1–7.
https://doi.org/10.15226/2381-2907/1/1/00103.
online version at doi:10.1016/j.prevetmed.2024.106131.
Pinto, S., Hoffmann, G., Ammon, C., Amon, T., 2020. Critical THI thresholds based on the
physiological parameters of lactating dairy cows. J. Therm. Biol. 88, 102523 https://
References doi.org/10.1016/j.jtherbio.2020.102523.
Polsky, L., von Keyserlingk, M.A.G., 2017. Invited review: effects of heat stress on dairy
cattle welfare. J. Dairy Sci. 100, 8645–8657. https://doi.org/10.3168/jds.2017-
Asplund, T., 2014. Climate change frames and frame formation: An analysis of climate 12651.
change communication in the Swedish agricultural sector. Clim. Chang. Fram. Fram. Rhoads, M.L., 2023. Review: reproductive consequences of whole-body adaptations of
Form. An Anal. Clim. Chang. Commun. Swedish Agric. Sect. Linköping University. dairy cattle to heat stress. Animal 17, 100847. https://doi.org/10.1016/j.
https://doi.org/10.3384/diss.diva-105997. animal.2023.100847.
Ahmed, H., Tamminen, L.-M., Emanuelson, U., 2022. Temperature, productivity, and Robert, M., Thomas, A., Bergez, J.E., 2016. Processes of adaptation in farm decision-
heat tolerance: Evidence from Swedish dairy production. Clim. Change 56, making models. A review. Agron. Sustain. Dev. 36 https://doi.org/10.1007/s13593-
1391–1412. https://doi.org/10.1007/s11135-021-01182-y. 016-0402-x.
Bagath, M., Krishnan, G., Devaraj, C., Rashamol, V.P., Pragna, P., Lees, A.M., Sejian, V., Sammad, A., Umer, S., Shi, R., Zhu, H., Zhao, X., Wang, Y., 2020. Dairy cow reproduction
2019. The impact of heat stress on the immune system in dairy cattle: a review. Res. under the influence of heat stress. J. Anim. Physiol. Anim. Nutr. 104, 978–986.
Vet. Sci. 126, 94–102. https://doi.org/10.1016/j.rvsc.2019.08.011. https://doi.org/10.1111/jpn.13257.
Barth, H., Melin, M., 2018. A Green Lean approach to global competition and climate Schüller, L.K., Burfeind, O., Heuwieser, W., 2014. Impact of heat stress on conception
change in the agricultural sector – a Swedish case study. J. Clean. Prod. 204, rate of dairy cows in the moderate climate considering different temperature-
183–192. https://doi.org/10.1016/j.jclepro.2018.09.021. humidity index thresholds, periods relative to breeding, and heat load indices.
van den Borne, B.H.P., van Schaik, G., Lam, T.J.G.M., Nielen, M., 2010. Variation in herd Theriogenology 81, 1050–1057. https://doi.org/10.1016/j.
level mastitis indicators between primi- and multiparae in Dutch dairy herds. Prev. theriogenology.2014.01.029.
Vet. Med. 96, 49–55. https://doi.org/10.1016/j.prevetmed.2010.05.010. David Scott, 2007. Resolving the quantitative–qualitative dilemma: a critical realist
Braun, V., Clarke, V., 2019. Reflecting on reflexive thematic analysis. Qual. Res. Sport. approach. Int. J. Res. Method Educ. 30, 3–17.SFS 2003:460, 2003. Lag om
Exerc. Heal. 11, 589–597. etikprövning av forskning som avser människor [WWW Document]. URL https://
Braun, V., Clarke, V., 2022. Conceptual and design thinking for thematic analysis. Qual. rkrattsbaser.gov.se/sfst?bet=2003:460 (accessed 8.11.21).
Psychol. 9, 3–26. https://doi.org/10.1037/qup0000196. Sjökvist, E., Abdoush, D., Axén, J., 2019. Sommaren 2018 - en glimt av framtiden?
Byrne, D., 2022. A worked example of Braun and Clarke’s approach to reflexive thematic Norrköping.
analysis. Qual. Quant. 56, 1391–1412. https://doi.org/10.1007/s11135-021-01182- SJVFS, 2019. Statens jordbruksverks föreskrifter och allmänna råd om
y. nötkreaturshållning inom lantbruket m.m. [WWW Document]. URL https://lagen.
R. Core Team, 2018. R: A Language and Environment for Statistical Computing. nu/sjvfs/2019:18.
Creswell, J.W., Clark, V.L.P., 2017. Designing and Conducting Mixed Methods Research, SMHI, 2021. Så påverkar värmen våra mjölkkor [WWW Document]. URL https://www.
third ed.,. SAGE, Los Angeles. smhi.se/kunskapsbanken/sa-paverkar-varmen-vara-mjolkkor-1.138977 (accessed
Dado-Senn, B., Skibiel, A.L., Fabris, T.F., Dahl, G.E., Laporta, J., 2019. Dry period heat 4.9.22).
stress induces microstructural changes in the lactating mammary gland. PLoS One Stewart, A.E., 2009. Psychological perspectives on adaptation to weather and climate. In:
14. https://doi.org/10.1371/journal.pone.0222120. Biometeorology for Adaptation to Climate Variability and Change. Biometeorology,
Dahl, G.E., Skibiel, A.L., Laporta, J., 2019. In utero heat stress programs reduced vol 1. Springer, Dordrecht. https://doi.org/10.1007/978-1-4020-8921-3_10.
performance and health in calves. Vet. Clin. North Am. Food Anim. Pract. 35, Svensson, C., Lind, N., Reyher, K.K., Bard, A.M., Emanuelson, U., 2019. Trust, feasibility,
343–353. https://doi.org/10.1016/j.cvfa.2019.02.005. and priorities influence Swedish dairy farmers’ adherence and nonadherence to
Dahl, G.E., Tao, S., Laporta, J., 2020. Heat stress impacts immune status in cows across veterinary advice. J. Dairy Sci. 102, 10360–10368. https://doi.org/10.3168/
the life cycle. Front. Vet. Sci. 7 https://doi.org/10.3389/fvets.2020.00116. jds.2019-16470.
EEA, 2017. Enhancing coherence of the knowledge base, policies and practices. Swedish Research Council, 2017. Good research practice.
Publications Office of the European Union, Luxembourg. https://doi.org/10.2800/ Turk, R., Podpečan, O., Mrkun, J., Flegar-Meštrić, Z., Perkov, S., Zrimšek, P., 2015. The
938195. effect of seasonal thermal stress on lipid mobilisation, antioxidant status and
Fabris, T.F., Laporta, J., Skibiel, A.L., Corra, F.N., Senn, B.D., Wohlgemuth, S.E., Dahl, G. reproductive performance in dairy cows. Reprod. Domest. Anim. 50, 595–603.
E., 2019. Effect of heat stress during early, late, and entire dry period on dairy cattle. https://doi.org/10.1111/rda.12534.
J. Dairy Sci. 102, 5647–5656. https://doi.org/10.3168/jds.2018-15721. Van Rij, J., Wieling, M., Harald Baayen, R., Van Rijn, H., Maintainer,], 2022. itsadug:
Flamenbaum, I., Ezra, E., 2007. The" Summer to Winter performance ratio" as a tool for Interpreting Time Series and Autocorrelated Data Using GAMMs.
evaluating heat stress relief efficiency of dairy herds. J. Dairy Sci. 90, 605–606. Vitali, A., Bernabucci, U., Nardone, A., Lacetera, N., 2016. Effect of season, month and
Gernand, E., König, S., Kipp, C., 2019. Influence of on-farm measurements for heat stress temperature humidity index on the occurrence of clinical mastitis in dairy heifers.
indicators on dairy cow productivity, female fertility, and health. J. Dairy Sci. 102, Adv. Anim. Biosci. 7, 250–252. https://doi.org/10.1017/s2040470016000315.
6660–6671. https://doi.org/10.3168/jds.2018-16011. Wickham, H., 2016. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag, New
Gröhn, Y.T., Rajala-Schultz, P.J., 2000. Epidemiology of reproductive performance in York.
dairy cows. Anim. Reprod. Sci. 60–61, 605–614. https://doi.org/10.1016/S0378- Wood, S., Pya, N., Säfken, B., 2016. Smoothing parameter and model selection for
4320(00)00085-3. general smooth models. J. Am. Stat. Assoc. 1–45.
Guzmán-Luna, P., Mauricio-Iglesias, M., Flysjö, A., Hospido, A., 2022. Analysing the Wood, S.N., 2011. Fast stable restricted maximum likelihood and marginal likelihood
interaction between the dairy sector and climate change from a life cycle estimation of semiparametric generalized linear models. J. R. Stat. Soc. Ser. B Stat.
perspective: a review. Trends Food Sci. Technol. 126, 168–179. https://doi.org/ Methodol. 73, 3–36. https://doi.org/10.1111/j.1467-9868.2010.00749.x.
10.1016/j.tifs.2021.09.001.

13