- Email: [email protected]
Influence of space availability and weather conditions on shelter use by beef cattle during winter
Applied Animal Behaviour Science xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Applied Animal Behaviour Science journal homepage: www.elsevier.com/locate/applanim
Influence of space availability and weather conditions on shelter use by beef cattle during winter Katrine K. Fogsgaard, Janne W. Christensen
⁎
Department of Animal Science, Aarhus University, Blichers Allé 20, DK-8830 Tjele, Denmark
A R T I C LE I N FO
A B S T R A C T
Keywords: Animal welfare Behaviour Cold stress Out-wintering Shelter Weather
The objective of this experiment was to evaluate shelter use by beef cattle in relation to space allowance per individual and weather conditions. Nine groups of Angus cattle (n = 35 in total, 3–6/group) were kept on paddocks with a squared shelter (5 × 10 m) with an open long side. In a 3 × 3 crossover design, three experimental treatments were tested based on national recommendations: 1) 100% of the recommended m2/individual (4 m2 per adult), 2) 150% (6 m2 per adult) and 3) 200% (8 m2 per adult). The shelter area was fenced off according to treatment and the number and size of animals in each group. Shelter use was estimated from pictures taken every 15 min with infrared trail cameras placed in all shelters. When the available space per individual was 100% of the recommended space, the shelters were used less compared to when 150% and 200% of the recommended space was available (e.g. percentage of pictures where all animals were inside the shelter (%all_animals): P < 0.001). There was a significant effect of weather conditions on shelter use (e.g. %all_animals; chill factor index: P = 0.03, and precipitation: P = 0.006), i.e. the shelters were used more with decreasing chill factor index and with increased precipitation. In conclusion, beef cattle increased their use of the shelters when the space allowance per individual increased with 50% or 100% compared to the current, national recommendations; e.g. simultaneous use by a whole group doubled with increased space. Furthermore, cold and wet weather increased shelter use.
1. Introduction During winter, cattle housed outside are exposed to cold, rainy and windy conditions and might, therefore, benefit from protection by natural vegetation or artificial shelters. In general, beef cattle breeds such as Angus Aberdeen, in good body condition and health, can tolerate low ambient temperatures without being in risk of cold stress (Webster, 1970). This cold resistance depends on the combined effect of the individuals’ own heat production and its insulation by fat tissue and fur coat. However, wind lowers the insulating effect of the fur coat and increases heat loss by transduction, which can be exacerbated by rain as heat loss is increased from wet skin (Schütz et al., 2010). Previous studies have shown that outdoor-wintered beef cattle increase their use of protected areas during times with precipitation, lower temperatures and increased wind speed (Graunke et al., 2011; Van laer et al., 2015). Thus, cattle kept in areas with limited natural protection against wind and rain might benefit from a well-designed shelter to mitigate the risk of cold stress. Indeed, lying is a highly prioritised need in cattle, and limited access to proper lying areas can have a negative effect on animal welfare (Ekesbo, 2011; Munksgaard et al., 2005). Artificial shelters
⁎
should, therefore, be large enough to provide such lying areas for the entire group simultaneously. According to Danish recommendations, a shelter should provide a minimum space of 4 m2 per adult individual (> 500 kg) (Table 1; SEGES, 2016). However, these recommendations have not been investigated scientifically, and it remains unknown whether the recommended space is sufficient for all animals in a group to lie down simultaneously. The objective of this experiment was to evaluate the effect of space allowance on shelter use. The prediction was that increased space allowance per individual would increase simultaneous use by all animals in a group. An additional objective was to investigate the effect of temperature, wind speed and precipitation on the use of artificial shelters in an area without access to protection by natural vegetation. 2. Materials and methods 2.1. Subjects The data collection took place in a private Angus herd in Jutland, Denmark. A total of 35 animals were included. Of these, 9 were heifers,
Corresponding author at: Blichers Allé 20, DK-8830 Tjele, Denmark. E-mail address: [email protected] (J.W. Christensen).
https://doi.org/10.1016/j.applanim.2018.04.007 Received 14 December 2017; Received in revised form 12 March 2018; Accepted 8 April 2018 0168-1591/ © 2018 Elsevier B.V. All rights reserved.
Please cite this article as: Fogsgaard, K.K., Applied Animal Behaviour Science (2018), https://doi.org/10.1016/j.applanim.2018.04.007
Applied Animal Behaviour Science xxx (xxxx) xxx–xxx
K.K. Fogsgaard, J.W. Christensen
Table 1 Recommendations for cattle on minimum space allowance per individual in shelters for outdoor winter housing of cattle (SEGES, 2016). Body weight, kg 2
Minimum shelter area (m /individual)
< 60
60–100
100–150
150–200
200–300
300–400
400–500
> 500
1.2
1.4
1.7
2.0
2.5
3.0
3.5
4.0
Table 2 Overview of animals in each group in each of the three periods. Group
1 2 3 4 5 6 7 8 9
Table 3 Distribution of the nine experimental groups (1–9) in a 3 × 3 crossover design. The groups were randomly selected into each category. Treatments were based on 100%, 150% and 200% of the recommended space availability in shelters (See Table 1).
Animals Period 1
Period 2
Period 3
3 cows, 1 bull, 1 calf 5 heifers 1 cow, 4 heifers, 1 calf 3 cows 3 cows 3 cows 3 cows 4 cows 3 cows
3 cows, 1 bull, 1 calf 5 heifers 1 cow, 4 heifers, 1 calf 3 cows 3 cows, 1 calf 4 cows 3 cows, 1 calf 3 cows 3 cows
3 cows, one bull, 1 calf 5 heifers 1 cow, 4 heifers, 1 calf 3 cows, 1 calf 3 cows, 1 calf 3 cows, 1 calf 3 cows, 2 calves 4 cows 3 cows, 2 calves
Treatment
Period 1 Period 2 Period 3
T100%
T150%
T200%
2, 4, 6 3, 7, 8 1, 5, 9
1, 5, 9 2, 4, 6 3, 7, 8
3, 7, 8 1, 5, 9 2, 4, 6
subjecting all nine groups to all three treatments (Table 3). The experimental period (January 2017–March 2017) consisted of three treatment periods, each consisting of an adaption period (approx. 10 days) followed by a 16 days registration period. During the adaption period, the group had access to the space available in the following treatment to habituate them to the new space allowance and minimize any potential order effects. For all groups, the available space per group was calculated as the sum of m2/individual based on Table 1. For example, in T100%, cows had 4 m2/individual, heifers had 3 m2/individual and calves 1.2 m2/individual (see Table 1). The number of individuals in the groups varied between the treatment periods as a few groups required regrouping and due to calving (Table 2). These changes occurred outside the registration periods and the space available in the shelter was always adjusted to fit the space requirements based on group composition (Table 1). The only exception was when a cow calved within a treatment period. In this case, the space availability was unadjusted to avoid disturbance of the cow and calf. Space availability in the shelters was adjusted using movable, galvanized fences, which allowed shielding part of the shelter. The depth of the shelter was constantly 5 m, while the width of the opening varied between treatments, but was never less than 2.5 m.
2 were calves, 1 was bull and 23 were cows. Furthermore, an additional seven calves were born during the experimental period. The animals were divided into nine groups (Table 2). Due to practical circumstances at the private farm, the bull had to be included into one of the groups. Each group had access to a fenced paddock with a squared shelter (5 × 10 m) with an open long side. The shelter was constructed of metal with metal roof and open, triangular gables in both ends (Fig. 1a). The shelter was bedded with barley straw. Hay silage was provided as feed in a trough within the paddock, and water was available ad libitum. The nine paddocks were located on two fields separated by a track with three paddocks on the east field and six on the west field (Fig. 1b). The paddocks were separated by an electric fence, and all access to vegetation was barred, so the only possible shelter was provided by the artificial shelters. The animals were separated into groups and had access to the shelters 14 days before the experimental period. The experiment was conducted during the winter 2016/2017 and complied with EU and Danish Ministry of Justice legislation concerning animal experimentation.
2.2. Experimental design
2.3. Recordings
Experimental treatments were based on Danish recommendations (SEGES, 2016) on shelter space for beef cattle (Table 1). Three experimental treatments were tested: 1) 100% of the recommended m2/ individual (T100%), 2) 150% of the recommended m2/individual (T150%) and 3) 200% of the recommended m2/individual (T200%). The study was designed as a 3 × 3 crossover design, randomly
The use of the shelters was recorded by infrared trail cameras (1 camera/shelter, Black IR Trail Camera, ScoutGuard, USA) taking a picture every 15 min. From these pictures, the use of the shelters was estimated by recording of the number of animals inside the shelter and their position (lying/standing) on each picture. Since some pictures were lost due to camera/flash failure (see results), the following
Fig. 1. a) Shelter design. b) Overview of the nine paddocks with shelters (black rectangles), placed with the opening towards the east/southeast. 2
Applied Animal Behaviour Science xxx (xxxx) xxx–xxx
K.K. Fogsgaard, J.W. Christensen
Fig. 2. Shelter use measured as the percentage of pictures (mean ± se) where all the animals in a group are (a) inside the shelter or (b) lying inside the shelter. The horizontal axis refers to three treatments (100%, 150% and 200% of the recommended space available per individual). Bars with different letters differ at P < 0.05.
Results are presented as LSmeans ± standard error (SE) and considered significant when P-values were ≤ 0.05 and as a tendency when 0.05 < P ≤ 0.1. Throughout, P-values were based on the Satterthwaite approximation for the denominator degrees of freedom.
variables were calculated as percentages of the total number of usable pictures per group for each 24-h period: 1) Percentage of pictures where at least one animal is inside the shelter (%one_animal), 2) percentage of pictures where all individuals in a group are inside the shelter (% all_animals), 3) percentage of pictures where at least one of the animals in the shelter is lying down (%one_lying) and 4) percentage of pictures where all animals in a group are lying at the same time in the shelter (%all_lying). Temperature loggers (iButton DS1923; Maxim Integrated, San Jose, CA, USA) were placed under the ridge in all shelters, obtaining humidity (%) and temperature (°C) every hour throughout the experimental period. Furthermore, daily meteorological data consisting of average wind speed (m/s/day) and precipitation (mm/day) were collected from a nearby meteorological station (Karup Airport; www. wunderground.com) located 6 km from the experimental location. As an estimate of the additional chilling effect of wind, a chill factor index (CFI) was calculated based on the equitation of Environment Canada (Webster et al., 2008):
3. Results The final data set consisted of 39,095 pictures (out of 41,472 possible) obtained during the three treatment periods in all nine shelters. For two groups (groups 2 and 3), data were removed for two subsequent days in period 1 because the animals broke through the fence, causing a temporary disturbance. Furthermore, black pictures from periods with camera or flash failure were removed from the final data. A total of 2,377 pictures were lost due to these circumstances. The shelters were used during a large part of the total observation period (e.g. at least one individual was inside on an average of 53.7 ± 0.9% of all pictures obtained across groups, treatments and periods, and all individuals in the group were inside on an average of 14.6 ± 0.7% of all pictures). The cattle predominantly used the shelters during the dark hours (e.g.% one_animal: Light hours (09:00 h-16:00 h): 17.6% vs. Dark hours (22:00 h-05:00 h): 92.6%). When the available space per individual was the recommended space (T100%), the shelters were used less compared to when 150% and 200% of the recommended space was available (T150% and T200) for the variables indicating how often all animals in a group were present or lying in the shelter at the same time (Fig. 2). A similar tendency was found for the variables where at least one animal was inside (%one_animal: T100%: 50.1 ± 3.6, T150%: 59.3 ± 3.2 and T200%: 52.1 ± 3.5; F2,50.6 = 2.78, P = 0.07) or lying inside the shelter (% one_lying: T100%: 44.1 ± 3.6, T150%: 53.4 ± 3.1 and T200%:
CFI = 13.12 + 0.62 * T − 11.37 * V.16 + 0.40 * T * V.16 where T = temperature (°C) and V = average wind speed (converted into km/h). 2.4. Statistical analyses The data were analysed in SAS Enterprise Guide 7.1 (SAS Institute, Inc., Cary, NC). Data were considered normally distributed based on graphical inspection of the residuals. The effect of the three experimental treatments and weather conditions were examined in a mixed model (PROC MIXED in SAS Enterprise Guide), taking the effect of repeated measures on group (1–9) within period and the random effect of period (1, 2, 3) into account. A separate model was created for each dependent variable (%one_animal, %all_animals, %one_lying and % all_lying) with treatment (T100%, T150%, T200%), chill factor index, precipitation and relative humidity included as explanatory variables. Measurements on different days obtained in the same group within period were assumed correlated, leading to the modulation of covariance structure. The choice of covariance structure (compound symmetry, autoregressive or unstructured) was determined based on the best fit, assessed by the Akaike’s information criterion (lowest values preferred; 9.1, SAS Institute Inc., Cary, NC). An autoregressive model (AR(1)) was chosen for all models. This choice is also supported by the experimental set-up as data is obtained over a period of time. In case of significant differences between treatments, a pairwise comparison of treatments (t-test) was conducted using the LSmeans statement in SAS.
Table 4 Summary of weather conditions (per 24 h) measured during the entire experimental period (Jan–March 2017). Temperature and humidity were obtained from data loggers placed inside the shelters. Additional data were obtained from a nearby weather station (Karup Airport, located 6 km from the experimental site).
Temperature (°C) Humidity (%) Precipitation (mm/24 h) Wind speed (m/s) Chill factor index
3
Mean ± se
Minimum
Maximum
4.9 ± 0.2 91.4 ± 0.3 1.6 ± 0.1 15.4 ± 0.3 −6.9 ± 0.2
−2.4 75.0 0 2 −14.5
13.9 99.9 7.9 34 6.3
Applied Animal Behaviour Science xxx (xxxx) xxx–xxx
K.K. Fogsgaard, J.W. Christensen
lying position or preferred distance to nearest neighbour. The most common lying position on both pasture and in stalls are on sternum with the body tipped to one side and the head positioned either stretched out on the floor or bent along the side. On pasture, cows were also observed lying on the side with head and legs stretched from the body, a position rarely observed indoor with more limited space (Kilgour et al., 2012; Krohn and Munksgaard, 1993). Previous studies examining the effect of space per individual have primarily focused on the total space available per individual in a pen or box and not lying area per se as in the present study. Hickey et al. (2002) found no differences in lying time and weight gain in beef cattle housed in out-wintering pats with wind shelters with 3, 6 or 18 m2 per individual. However, across treatments, a minimum dry lying area of 2.2 m2 was maintained. The lack of variation in dry lying area between the treatments, despite increased space availability, might explain the lack of difference in lying time across treatments. In indoor housing, Simmental heifers (mean weight 468 kg) were found to have reduced daily weight gain and lying time when the space availability per individual in a box was 1.5 m2 compared to individuals kept in similar housing with increased space per individual (> 2 m2) (Fisher et al., 1997). In dairy cattle, increased space per individual was shown to result in increased lying time and less aggressive interactions between herd mates (Schütz et al., 2015). Furthermore, a limited number of stalls (150% overstocking) changed the 24-h lying pattern, resulting in increased lying during daytime (Winckler et al., 2015). Schütz et al. (2015) concluded that dry cows (approx. 500 kg) need at least 4–6 m2 per individual to meet their needs and avoid negative effects on lying behaviour as well as reduce the degree of aggressive behaviour. This estimation of 4–6 m2 per individual includes feeding area and alley. Therefore, direct comparisons to the present study on shelter space were not possible. The animals in our study were more frequently observed lying down when the available lying area increased from 4m2 to 6m2 per adult individual, which indicates that for beef cattle under the present circumstances 4 m2 may not be sufficient. Extensive breeds of cattle such as Angus Aberdeen are considered relatively resistant towards low temperatures. However, when exposed to rain and wind, the skin temperature is reduced, increasing the risk of discomfort or perhaps cold stress (Schütz et al., 2010). This was supported by our findings of increased use of the shelters with decreased chill factor index and increased precipitation. Similar results were previously reported for crossbreed Angus cattle where the animals sought shelter behind windbreaks, in vegetation or stayed closer together in groups with increased wind speed and decreased chill factor index (Graunke et al., 2011). Furthermore, the cattle were found to use protected areas in windy and cold periods, although these areas had decreased foraging opportunity compared to other areas, suggesting that these areas were used as protection against the weather (Graunke et al., 2011). Higher wind velocities partially destroy the insulation effect of the hair coat and result in increased convective heat loss (Ames and Insley, 1975). In North European countries, high wind speed during winter is often confounded with colder temperatures, exacerbating the risk of cold stress. However, the present study was not designed to investigate the importance of access to shelter in relation to animal welfare. More research is needed to address the importance of access to protection against wind and rain in extensive breeds of beef cattle. In the present study, the artificial shelters were the only option for protection against wind and rain. A Belgian study involving different breeds of cattle examined the use of artificial shelters and natural vegetation in areas where both were present during winter. Overall, they found a high preference for natural vegetation and very little use of artificial shelters, except in an area with very limited and low natural vegetation where an artificial shelter was used (Van laer et al., 2015). Further research is needed to explore preferences for natural and artificial shelter in cattle and the effect of availability of other resources, such as food and water.
Table 5 Results from statistical models (mixed models) examining the effect of the continuous variables, precipitation and chill factor index, on shelter use (% pictures). Estimated intercept %one_animal Precipitation (mm/24 h) Chill factor index %all_animals Precipitation (mm/24 h) Chill factor index %one_lying Precipitation (mm/24 h) Chill factor index %all_lying Precipitation (mm/24 h) Chill factor index
Estimate
FNumDF,Den-
P-value
DF
48.2 ± 3.5 1.7 ± 0.3 −1.0 ± 0.3
23.61,332 14.31,164
< 0.001 < 0.001
0.7 ± 0.2 −0.4 ± 0.2
7.41,352 4.31,382
0.006 0.03
1.3 ± 0.3 −1.1 ± 0.3
15.31,323 18.51,218
< 0.001 < 0.001
0.3 ± 0.2 −0.3 ± 0.2
1.91,353 3.91,374
0.1 0.05
14.6 ± 3.1
44.3 ± 3.6
12.1 ± 2.4
48.2 ± 3.3; F2,54.5 = 42.96, P = 0.06). An overview of the weather conditions across all periods is presented in Table 4. The relative humidity was not a significant factor (P > 0.2) and was therefore removed from the final statistical models. Increased precipitation and a lower chill factor index increased the use of shelters across all four variables (Table 5). 4. Discussion This study aimed to examine the effect of shelter space availability on shelter use by beef cattle. The results suggested that when the available space follows the national recommendations for cattle (4 m2 for an adult), the shelters were used less compared to treatments where either 150% or 200% of the recommended space per individual was available. This was especially evident when examining how often an entire group was present inside their shelter, and how often all animals in a group were lying down simultaneously. This was observed on approx. 6% of the pictures when 100% of the recommended space was available, while the whole group was observed lying down together on more than 12% of the pictures when 150% or 200% of the recommended space was available. These results indicate that although there appears to be physical space for the animals to lie down also at the 100% treatment, the animals may choose not to lie down when space is limited. Furthermore, higher-ranking individuals may prevent lowerranking group members from lying/standing inside the shelters. In order to further investigate interactions between low and high-ranking animals, individual identification of the animals is necessary. In the current study, individual marking of the animals was performed prior to the experiment, but unfortunately the used marking paint did not withstand the wet weather and the identification therefore failed. Further studies are required to explore individual shelter use in relation to social rank. In addition, this study was based on a limited number of animals and groups, and not all treatment sequences were included. A 10-days habituation period was included between treatments but the extent to which previous experience affects shelter use in following periods (order effects) remains to be investigated. For cattle kept in indoor housing facilities, it was shown that increased competition for access to free stalls resulted in decreased lying time (Fregonesi et al., 2007). Furthermore, limited space per individual might impair optimal lying and getting up/down movements for the animals. The scientific knowledge on preferred lying space for beef cattle is sparse. From algometric calculation, a 500-kg individual was estimated to take up 1.54 m2 when lying on sternum (Baxter, 1992). This is the estimated space which the body of the animal takes up. However, this does not necessarily reflect the preferred size of lying areas for the individual, as the calculations do not include preferred 4
Applied Animal Behaviour Science xxx (xxxx) xxx–xxx
K.K. Fogsgaard, J.W. Christensen
5. Conclusion
Fisher, A.D., Crowe, M.A., O'Kiely, P., Enright, W.J., 1997. Growth, behaviour, adrenal and immune responses of finishing beef heifers housed on slatted floors at 1.5, 2.0, 2.5 or 3.0 m2 space allowance. Livest. Prod. Sci. 51, 245–254. Fregonesi, J.A., Tucker, C.B., Weary, D.M., 2007. Overstocking reduces lying time in dairy cows. J. Dairy Sci. 90, 3349–3354. Graunke, K.L., Schuster, T., Lidfors, L.M., 2011. Influence of weather on the behaviour of outdoor-wintered beef cattle in Scandinavia. Livest. Sci. 136, 247–255. Hickey, M.C., French, P., Grant, J., 2002. Out-wintering Pads for Finishing Beef Cattle: Animal Production and Welfare, Animal Science. Cambridge University Press, pp. 447–458. Kilgour, R.J., Uetake, K., Ishiwata, T., Melville, G.J., 2012. The behaviour of beef cattle at pasture. Appl. Anim. Behav. Sci. 138, 12–17. Krohn, C.C., Munksgaard, L., 1993. Behaviour of dairy cows kept in extensive (loose housing/pasture) or intensive (tie stall) environments II. Lying and lying-down behaviour. Appl. Anim. Behav. Sci. 37, 1–16. Munksgaard, L., Jensen, M.B., Pedersen, L.J., Hansen, S.W., Matthews, L., 2005. Quantifying behavioural priorities-effects of time constraints on behaviour of dairy cows, Bos taurus. Appl. Anim. Behav. Sci. 92, 3–14. SEGES, 2016. In: Ingvortsen, B.K., C, V., Spleth, P. (Eds.), Fact sheet: Keeping of outwintering animals – Cattle, horses and sheep (In Dansih: Faktaark: Hold af udegående dyr i vinterperioden – kvæg, heste og får). SEGES P/S, Denmark. https://www. landbrugsinfo.dk/oekologi/kvaeg/oksekoedsproduktion/sider/4054_faktaark_ udegaaende_dyr_vinter.pdf. (Accessed October 2017). Schütz, K.E., Clark, K.V., Cox, N.R., Matthews, L.R., Tucker, C.B., 2010. Responses to short-term exposure to simulated rain and wind by dairy cattle: time budgets, shelter use, body temperature and feed intake. Anim. Welfare 19, 375–383. Schütz, K.E., Huddart, F.J., Sutherland, M.A., Stewart, M., Cox, N.R., 2015. Effects of space allowance on the behavior and physiology of cattle temporarily managed on rubber mats. J. Dairy Sci. 98, 6226–6235. Van laer, E., Ampe, B., Moons, C., Sonck, B., Tuyttens, F.A.M., 2015. Wintertime use of natural versus artificial shelter by cattle in nature reserves in temperate areas. Appl. Anim. Behav. Sci. 163, 39–49. Webster, A.J.F., 1970. Direct effects of cold weather on the energetic efficiency of beef production in different regions of Canada. Can. J. Anim. Sci. 50, 563–573. Webster, J.R., Stewart, M., Rogers, A.R., Verkerk, G.A., 2008. Assessment of welfare from physiological and behavioural responses of New Zealand and dairy cows exposed to cold and wet conditions. Anim. Welfare 17, 19–26. Winckler, C., Tucker, C.B., Weary, D.M., 2015. Effects of under- and overstocking freestalls on dairy cattle behaviour. Appl. Anim. Behav. Sci. 170, 14–19.
Increasing space availability from 4 m2 to 6 or 8 m2 per adult individual resulted in higher use of the shelters as simultaneous use by a whole group was found to double when the available space was increased. Shelter use was also found to be affected by weather conditions as colder and harsher weather with lower chill factor index resulted in an increased use of the shelters. In the present study, the cattle did not have access to natural vegetation, and further research is needed to clarify whether cattle have preferences for specific types of shelter and the overall need for shelters in relation to animal welfare. Conflicts of interest None. Acknowledgements The authors would like to thank the owner and staff on the participating farm as well as research technicians, Anton Steen Jensen and Henrik Krogh Andersen, Aarhus University. This study was funded by the Danish Authorities and Aarhus University. References Ames, D.R., Insley, L.W., 1975. Wind-chill effect for cattle and sheep. J. Anim. Sci. 40, 161–165. Baxter, M.R., 1992. Overstocking reduces lying time in dairy cows. In: Phillips, C.P.D. (Ed.), Farm Animals and the Environment. CAB International, Wallingford, UK, pp. 67–81. Ekesbo, I., 2011. Farm Animal Behaviour: Characteristics for Assessment of Health and Welfare. CAB International, Oxfordshire, UK.
5
Contents lists available at ScienceDirect
Applied Animal Behaviour Science journal homepage: www.elsevier.com/locate/applanim
Influence of space availability and weather conditions on shelter use by beef cattle during winter Katrine K. Fogsgaard, Janne W. Christensen
⁎
Department of Animal Science, Aarhus University, Blichers Allé 20, DK-8830 Tjele, Denmark
A R T I C LE I N FO
A B S T R A C T
Keywords: Animal welfare Behaviour Cold stress Out-wintering Shelter Weather
The objective of this experiment was to evaluate shelter use by beef cattle in relation to space allowance per individual and weather conditions. Nine groups of Angus cattle (n = 35 in total, 3–6/group) were kept on paddocks with a squared shelter (5 × 10 m) with an open long side. In a 3 × 3 crossover design, three experimental treatments were tested based on national recommendations: 1) 100% of the recommended m2/individual (4 m2 per adult), 2) 150% (6 m2 per adult) and 3) 200% (8 m2 per adult). The shelter area was fenced off according to treatment and the number and size of animals in each group. Shelter use was estimated from pictures taken every 15 min with infrared trail cameras placed in all shelters. When the available space per individual was 100% of the recommended space, the shelters were used less compared to when 150% and 200% of the recommended space was available (e.g. percentage of pictures where all animals were inside the shelter (%all_animals): P < 0.001). There was a significant effect of weather conditions on shelter use (e.g. %all_animals; chill factor index: P = 0.03, and precipitation: P = 0.006), i.e. the shelters were used more with decreasing chill factor index and with increased precipitation. In conclusion, beef cattle increased their use of the shelters when the space allowance per individual increased with 50% or 100% compared to the current, national recommendations; e.g. simultaneous use by a whole group doubled with increased space. Furthermore, cold and wet weather increased shelter use.
1. Introduction During winter, cattle housed outside are exposed to cold, rainy and windy conditions and might, therefore, benefit from protection by natural vegetation or artificial shelters. In general, beef cattle breeds such as Angus Aberdeen, in good body condition and health, can tolerate low ambient temperatures without being in risk of cold stress (Webster, 1970). This cold resistance depends on the combined effect of the individuals’ own heat production and its insulation by fat tissue and fur coat. However, wind lowers the insulating effect of the fur coat and increases heat loss by transduction, which can be exacerbated by rain as heat loss is increased from wet skin (Schütz et al., 2010). Previous studies have shown that outdoor-wintered beef cattle increase their use of protected areas during times with precipitation, lower temperatures and increased wind speed (Graunke et al., 2011; Van laer et al., 2015). Thus, cattle kept in areas with limited natural protection against wind and rain might benefit from a well-designed shelter to mitigate the risk of cold stress. Indeed, lying is a highly prioritised need in cattle, and limited access to proper lying areas can have a negative effect on animal welfare (Ekesbo, 2011; Munksgaard et al., 2005). Artificial shelters
⁎
should, therefore, be large enough to provide such lying areas for the entire group simultaneously. According to Danish recommendations, a shelter should provide a minimum space of 4 m2 per adult individual (> 500 kg) (Table 1; SEGES, 2016). However, these recommendations have not been investigated scientifically, and it remains unknown whether the recommended space is sufficient for all animals in a group to lie down simultaneously. The objective of this experiment was to evaluate the effect of space allowance on shelter use. The prediction was that increased space allowance per individual would increase simultaneous use by all animals in a group. An additional objective was to investigate the effect of temperature, wind speed and precipitation on the use of artificial shelters in an area without access to protection by natural vegetation. 2. Materials and methods 2.1. Subjects The data collection took place in a private Angus herd in Jutland, Denmark. A total of 35 animals were included. Of these, 9 were heifers,
Corresponding author at: Blichers Allé 20, DK-8830 Tjele, Denmark. E-mail address: [email protected] (J.W. Christensen).
https://doi.org/10.1016/j.applanim.2018.04.007 Received 14 December 2017; Received in revised form 12 March 2018; Accepted 8 April 2018 0168-1591/ © 2018 Elsevier B.V. All rights reserved.
Please cite this article as: Fogsgaard, K.K., Applied Animal Behaviour Science (2018), https://doi.org/10.1016/j.applanim.2018.04.007
Applied Animal Behaviour Science xxx (xxxx) xxx–xxx
K.K. Fogsgaard, J.W. Christensen
Table 1 Recommendations for cattle on minimum space allowance per individual in shelters for outdoor winter housing of cattle (SEGES, 2016). Body weight, kg 2
Minimum shelter area (m /individual)
< 60
60–100
100–150
150–200
200–300
300–400
400–500
> 500
1.2
1.4
1.7
2.0
2.5
3.0
3.5
4.0
Table 2 Overview of animals in each group in each of the three periods. Group
1 2 3 4 5 6 7 8 9
Table 3 Distribution of the nine experimental groups (1–9) in a 3 × 3 crossover design. The groups were randomly selected into each category. Treatments were based on 100%, 150% and 200% of the recommended space availability in shelters (See Table 1).
Animals Period 1
Period 2
Period 3
3 cows, 1 bull, 1 calf 5 heifers 1 cow, 4 heifers, 1 calf 3 cows 3 cows 3 cows 3 cows 4 cows 3 cows
3 cows, 1 bull, 1 calf 5 heifers 1 cow, 4 heifers, 1 calf 3 cows 3 cows, 1 calf 4 cows 3 cows, 1 calf 3 cows 3 cows
3 cows, one bull, 1 calf 5 heifers 1 cow, 4 heifers, 1 calf 3 cows, 1 calf 3 cows, 1 calf 3 cows, 1 calf 3 cows, 2 calves 4 cows 3 cows, 2 calves
Treatment
Period 1 Period 2 Period 3
T100%
T150%
T200%
2, 4, 6 3, 7, 8 1, 5, 9
1, 5, 9 2, 4, 6 3, 7, 8
3, 7, 8 1, 5, 9 2, 4, 6
subjecting all nine groups to all three treatments (Table 3). The experimental period (January 2017–March 2017) consisted of three treatment periods, each consisting of an adaption period (approx. 10 days) followed by a 16 days registration period. During the adaption period, the group had access to the space available in the following treatment to habituate them to the new space allowance and minimize any potential order effects. For all groups, the available space per group was calculated as the sum of m2/individual based on Table 1. For example, in T100%, cows had 4 m2/individual, heifers had 3 m2/individual and calves 1.2 m2/individual (see Table 1). The number of individuals in the groups varied between the treatment periods as a few groups required regrouping and due to calving (Table 2). These changes occurred outside the registration periods and the space available in the shelter was always adjusted to fit the space requirements based on group composition (Table 1). The only exception was when a cow calved within a treatment period. In this case, the space availability was unadjusted to avoid disturbance of the cow and calf. Space availability in the shelters was adjusted using movable, galvanized fences, which allowed shielding part of the shelter. The depth of the shelter was constantly 5 m, while the width of the opening varied between treatments, but was never less than 2.5 m.
2 were calves, 1 was bull and 23 were cows. Furthermore, an additional seven calves were born during the experimental period. The animals were divided into nine groups (Table 2). Due to practical circumstances at the private farm, the bull had to be included into one of the groups. Each group had access to a fenced paddock with a squared shelter (5 × 10 m) with an open long side. The shelter was constructed of metal with metal roof and open, triangular gables in both ends (Fig. 1a). The shelter was bedded with barley straw. Hay silage was provided as feed in a trough within the paddock, and water was available ad libitum. The nine paddocks were located on two fields separated by a track with three paddocks on the east field and six on the west field (Fig. 1b). The paddocks were separated by an electric fence, and all access to vegetation was barred, so the only possible shelter was provided by the artificial shelters. The animals were separated into groups and had access to the shelters 14 days before the experimental period. The experiment was conducted during the winter 2016/2017 and complied with EU and Danish Ministry of Justice legislation concerning animal experimentation.
2.2. Experimental design
2.3. Recordings
Experimental treatments were based on Danish recommendations (SEGES, 2016) on shelter space for beef cattle (Table 1). Three experimental treatments were tested: 1) 100% of the recommended m2/ individual (T100%), 2) 150% of the recommended m2/individual (T150%) and 3) 200% of the recommended m2/individual (T200%). The study was designed as a 3 × 3 crossover design, randomly
The use of the shelters was recorded by infrared trail cameras (1 camera/shelter, Black IR Trail Camera, ScoutGuard, USA) taking a picture every 15 min. From these pictures, the use of the shelters was estimated by recording of the number of animals inside the shelter and their position (lying/standing) on each picture. Since some pictures were lost due to camera/flash failure (see results), the following
Fig. 1. a) Shelter design. b) Overview of the nine paddocks with shelters (black rectangles), placed with the opening towards the east/southeast. 2
Applied Animal Behaviour Science xxx (xxxx) xxx–xxx
K.K. Fogsgaard, J.W. Christensen
Fig. 2. Shelter use measured as the percentage of pictures (mean ± se) where all the animals in a group are (a) inside the shelter or (b) lying inside the shelter. The horizontal axis refers to three treatments (100%, 150% and 200% of the recommended space available per individual). Bars with different letters differ at P < 0.05.
Results are presented as LSmeans ± standard error (SE) and considered significant when P-values were ≤ 0.05 and as a tendency when 0.05 < P ≤ 0.1. Throughout, P-values were based on the Satterthwaite approximation for the denominator degrees of freedom.
variables were calculated as percentages of the total number of usable pictures per group for each 24-h period: 1) Percentage of pictures where at least one animal is inside the shelter (%one_animal), 2) percentage of pictures where all individuals in a group are inside the shelter (% all_animals), 3) percentage of pictures where at least one of the animals in the shelter is lying down (%one_lying) and 4) percentage of pictures where all animals in a group are lying at the same time in the shelter (%all_lying). Temperature loggers (iButton DS1923; Maxim Integrated, San Jose, CA, USA) were placed under the ridge in all shelters, obtaining humidity (%) and temperature (°C) every hour throughout the experimental period. Furthermore, daily meteorological data consisting of average wind speed (m/s/day) and precipitation (mm/day) were collected from a nearby meteorological station (Karup Airport; www. wunderground.com) located 6 km from the experimental location. As an estimate of the additional chilling effect of wind, a chill factor index (CFI) was calculated based on the equitation of Environment Canada (Webster et al., 2008):
3. Results The final data set consisted of 39,095 pictures (out of 41,472 possible) obtained during the three treatment periods in all nine shelters. For two groups (groups 2 and 3), data were removed for two subsequent days in period 1 because the animals broke through the fence, causing a temporary disturbance. Furthermore, black pictures from periods with camera or flash failure were removed from the final data. A total of 2,377 pictures were lost due to these circumstances. The shelters were used during a large part of the total observation period (e.g. at least one individual was inside on an average of 53.7 ± 0.9% of all pictures obtained across groups, treatments and periods, and all individuals in the group were inside on an average of 14.6 ± 0.7% of all pictures). The cattle predominantly used the shelters during the dark hours (e.g.% one_animal: Light hours (09:00 h-16:00 h): 17.6% vs. Dark hours (22:00 h-05:00 h): 92.6%). When the available space per individual was the recommended space (T100%), the shelters were used less compared to when 150% and 200% of the recommended space was available (T150% and T200) for the variables indicating how often all animals in a group were present or lying in the shelter at the same time (Fig. 2). A similar tendency was found for the variables where at least one animal was inside (%one_animal: T100%: 50.1 ± 3.6, T150%: 59.3 ± 3.2 and T200%: 52.1 ± 3.5; F2,50.6 = 2.78, P = 0.07) or lying inside the shelter (% one_lying: T100%: 44.1 ± 3.6, T150%: 53.4 ± 3.1 and T200%:
CFI = 13.12 + 0.62 * T − 11.37 * V.16 + 0.40 * T * V.16 where T = temperature (°C) and V = average wind speed (converted into km/h). 2.4. Statistical analyses The data were analysed in SAS Enterprise Guide 7.1 (SAS Institute, Inc., Cary, NC). Data were considered normally distributed based on graphical inspection of the residuals. The effect of the three experimental treatments and weather conditions were examined in a mixed model (PROC MIXED in SAS Enterprise Guide), taking the effect of repeated measures on group (1–9) within period and the random effect of period (1, 2, 3) into account. A separate model was created for each dependent variable (%one_animal, %all_animals, %one_lying and % all_lying) with treatment (T100%, T150%, T200%), chill factor index, precipitation and relative humidity included as explanatory variables. Measurements on different days obtained in the same group within period were assumed correlated, leading to the modulation of covariance structure. The choice of covariance structure (compound symmetry, autoregressive or unstructured) was determined based on the best fit, assessed by the Akaike’s information criterion (lowest values preferred; 9.1, SAS Institute Inc., Cary, NC). An autoregressive model (AR(1)) was chosen for all models. This choice is also supported by the experimental set-up as data is obtained over a period of time. In case of significant differences between treatments, a pairwise comparison of treatments (t-test) was conducted using the LSmeans statement in SAS.
Table 4 Summary of weather conditions (per 24 h) measured during the entire experimental period (Jan–March 2017). Temperature and humidity were obtained from data loggers placed inside the shelters. Additional data were obtained from a nearby weather station (Karup Airport, located 6 km from the experimental site).
Temperature (°C) Humidity (%) Precipitation (mm/24 h) Wind speed (m/s) Chill factor index
3
Mean ± se
Minimum
Maximum
4.9 ± 0.2 91.4 ± 0.3 1.6 ± 0.1 15.4 ± 0.3 −6.9 ± 0.2
−2.4 75.0 0 2 −14.5
13.9 99.9 7.9 34 6.3
Applied Animal Behaviour Science xxx (xxxx) xxx–xxx
K.K. Fogsgaard, J.W. Christensen
lying position or preferred distance to nearest neighbour. The most common lying position on both pasture and in stalls are on sternum with the body tipped to one side and the head positioned either stretched out on the floor or bent along the side. On pasture, cows were also observed lying on the side with head and legs stretched from the body, a position rarely observed indoor with more limited space (Kilgour et al., 2012; Krohn and Munksgaard, 1993). Previous studies examining the effect of space per individual have primarily focused on the total space available per individual in a pen or box and not lying area per se as in the present study. Hickey et al. (2002) found no differences in lying time and weight gain in beef cattle housed in out-wintering pats with wind shelters with 3, 6 or 18 m2 per individual. However, across treatments, a minimum dry lying area of 2.2 m2 was maintained. The lack of variation in dry lying area between the treatments, despite increased space availability, might explain the lack of difference in lying time across treatments. In indoor housing, Simmental heifers (mean weight 468 kg) were found to have reduced daily weight gain and lying time when the space availability per individual in a box was 1.5 m2 compared to individuals kept in similar housing with increased space per individual (> 2 m2) (Fisher et al., 1997). In dairy cattle, increased space per individual was shown to result in increased lying time and less aggressive interactions between herd mates (Schütz et al., 2015). Furthermore, a limited number of stalls (150% overstocking) changed the 24-h lying pattern, resulting in increased lying during daytime (Winckler et al., 2015). Schütz et al. (2015) concluded that dry cows (approx. 500 kg) need at least 4–6 m2 per individual to meet their needs and avoid negative effects on lying behaviour as well as reduce the degree of aggressive behaviour. This estimation of 4–6 m2 per individual includes feeding area and alley. Therefore, direct comparisons to the present study on shelter space were not possible. The animals in our study were more frequently observed lying down when the available lying area increased from 4m2 to 6m2 per adult individual, which indicates that for beef cattle under the present circumstances 4 m2 may not be sufficient. Extensive breeds of cattle such as Angus Aberdeen are considered relatively resistant towards low temperatures. However, when exposed to rain and wind, the skin temperature is reduced, increasing the risk of discomfort or perhaps cold stress (Schütz et al., 2010). This was supported by our findings of increased use of the shelters with decreased chill factor index and increased precipitation. Similar results were previously reported for crossbreed Angus cattle where the animals sought shelter behind windbreaks, in vegetation or stayed closer together in groups with increased wind speed and decreased chill factor index (Graunke et al., 2011). Furthermore, the cattle were found to use protected areas in windy and cold periods, although these areas had decreased foraging opportunity compared to other areas, suggesting that these areas were used as protection against the weather (Graunke et al., 2011). Higher wind velocities partially destroy the insulation effect of the hair coat and result in increased convective heat loss (Ames and Insley, 1975). In North European countries, high wind speed during winter is often confounded with colder temperatures, exacerbating the risk of cold stress. However, the present study was not designed to investigate the importance of access to shelter in relation to animal welfare. More research is needed to address the importance of access to protection against wind and rain in extensive breeds of beef cattle. In the present study, the artificial shelters were the only option for protection against wind and rain. A Belgian study involving different breeds of cattle examined the use of artificial shelters and natural vegetation in areas where both were present during winter. Overall, they found a high preference for natural vegetation and very little use of artificial shelters, except in an area with very limited and low natural vegetation where an artificial shelter was used (Van laer et al., 2015). Further research is needed to explore preferences for natural and artificial shelter in cattle and the effect of availability of other resources, such as food and water.
Table 5 Results from statistical models (mixed models) examining the effect of the continuous variables, precipitation and chill factor index, on shelter use (% pictures). Estimated intercept %one_animal Precipitation (mm/24 h) Chill factor index %all_animals Precipitation (mm/24 h) Chill factor index %one_lying Precipitation (mm/24 h) Chill factor index %all_lying Precipitation (mm/24 h) Chill factor index
Estimate
FNumDF,Den-
P-value
DF
48.2 ± 3.5 1.7 ± 0.3 −1.0 ± 0.3
23.61,332 14.31,164
< 0.001 < 0.001
0.7 ± 0.2 −0.4 ± 0.2
7.41,352 4.31,382
0.006 0.03
1.3 ± 0.3 −1.1 ± 0.3
15.31,323 18.51,218
< 0.001 < 0.001
0.3 ± 0.2 −0.3 ± 0.2
1.91,353 3.91,374
0.1 0.05
14.6 ± 3.1
44.3 ± 3.6
12.1 ± 2.4
48.2 ± 3.3; F2,54.5 = 42.96, P = 0.06). An overview of the weather conditions across all periods is presented in Table 4. The relative humidity was not a significant factor (P > 0.2) and was therefore removed from the final statistical models. Increased precipitation and a lower chill factor index increased the use of shelters across all four variables (Table 5). 4. Discussion This study aimed to examine the effect of shelter space availability on shelter use by beef cattle. The results suggested that when the available space follows the national recommendations for cattle (4 m2 for an adult), the shelters were used less compared to treatments where either 150% or 200% of the recommended space per individual was available. This was especially evident when examining how often an entire group was present inside their shelter, and how often all animals in a group were lying down simultaneously. This was observed on approx. 6% of the pictures when 100% of the recommended space was available, while the whole group was observed lying down together on more than 12% of the pictures when 150% or 200% of the recommended space was available. These results indicate that although there appears to be physical space for the animals to lie down also at the 100% treatment, the animals may choose not to lie down when space is limited. Furthermore, higher-ranking individuals may prevent lowerranking group members from lying/standing inside the shelters. In order to further investigate interactions between low and high-ranking animals, individual identification of the animals is necessary. In the current study, individual marking of the animals was performed prior to the experiment, but unfortunately the used marking paint did not withstand the wet weather and the identification therefore failed. Further studies are required to explore individual shelter use in relation to social rank. In addition, this study was based on a limited number of animals and groups, and not all treatment sequences were included. A 10-days habituation period was included between treatments but the extent to which previous experience affects shelter use in following periods (order effects) remains to be investigated. For cattle kept in indoor housing facilities, it was shown that increased competition for access to free stalls resulted in decreased lying time (Fregonesi et al., 2007). Furthermore, limited space per individual might impair optimal lying and getting up/down movements for the animals. The scientific knowledge on preferred lying space for beef cattle is sparse. From algometric calculation, a 500-kg individual was estimated to take up 1.54 m2 when lying on sternum (Baxter, 1992). This is the estimated space which the body of the animal takes up. However, this does not necessarily reflect the preferred size of lying areas for the individual, as the calculations do not include preferred 4
Applied Animal Behaviour Science xxx (xxxx) xxx–xxx
K.K. Fogsgaard, J.W. Christensen
5. Conclusion
Fisher, A.D., Crowe, M.A., O'Kiely, P., Enright, W.J., 1997. Growth, behaviour, adrenal and immune responses of finishing beef heifers housed on slatted floors at 1.5, 2.0, 2.5 or 3.0 m2 space allowance. Livest. Prod. Sci. 51, 245–254. Fregonesi, J.A., Tucker, C.B., Weary, D.M., 2007. Overstocking reduces lying time in dairy cows. J. Dairy Sci. 90, 3349–3354. Graunke, K.L., Schuster, T., Lidfors, L.M., 2011. Influence of weather on the behaviour of outdoor-wintered beef cattle in Scandinavia. Livest. Sci. 136, 247–255. Hickey, M.C., French, P., Grant, J., 2002. Out-wintering Pads for Finishing Beef Cattle: Animal Production and Welfare, Animal Science. Cambridge University Press, pp. 447–458. Kilgour, R.J., Uetake, K., Ishiwata, T., Melville, G.J., 2012. The behaviour of beef cattle at pasture. Appl. Anim. Behav. Sci. 138, 12–17. Krohn, C.C., Munksgaard, L., 1993. Behaviour of dairy cows kept in extensive (loose housing/pasture) or intensive (tie stall) environments II. Lying and lying-down behaviour. Appl. Anim. Behav. Sci. 37, 1–16. Munksgaard, L., Jensen, M.B., Pedersen, L.J., Hansen, S.W., Matthews, L., 2005. Quantifying behavioural priorities-effects of time constraints on behaviour of dairy cows, Bos taurus. Appl. Anim. Behav. Sci. 92, 3–14. SEGES, 2016. In: Ingvortsen, B.K., C, V., Spleth, P. (Eds.), Fact sheet: Keeping of outwintering animals – Cattle, horses and sheep (In Dansih: Faktaark: Hold af udegående dyr i vinterperioden – kvæg, heste og får). SEGES P/S, Denmark. https://www. landbrugsinfo.dk/oekologi/kvaeg/oksekoedsproduktion/sider/4054_faktaark_ udegaaende_dyr_vinter.pdf. (Accessed October 2017). Schütz, K.E., Clark, K.V., Cox, N.R., Matthews, L.R., Tucker, C.B., 2010. Responses to short-term exposure to simulated rain and wind by dairy cattle: time budgets, shelter use, body temperature and feed intake. Anim. Welfare 19, 375–383. Schütz, K.E., Huddart, F.J., Sutherland, M.A., Stewart, M., Cox, N.R., 2015. Effects of space allowance on the behavior and physiology of cattle temporarily managed on rubber mats. J. Dairy Sci. 98, 6226–6235. Van laer, E., Ampe, B., Moons, C., Sonck, B., Tuyttens, F.A.M., 2015. Wintertime use of natural versus artificial shelter by cattle in nature reserves in temperate areas. Appl. Anim. Behav. Sci. 163, 39–49. Webster, A.J.F., 1970. Direct effects of cold weather on the energetic efficiency of beef production in different regions of Canada. Can. J. Anim. Sci. 50, 563–573. Webster, J.R., Stewart, M., Rogers, A.R., Verkerk, G.A., 2008. Assessment of welfare from physiological and behavioural responses of New Zealand and dairy cows exposed to cold and wet conditions. Anim. Welfare 17, 19–26. Winckler, C., Tucker, C.B., Weary, D.M., 2015. Effects of under- and overstocking freestalls on dairy cattle behaviour. Appl. Anim. Behav. Sci. 170, 14–19.
Increasing space availability from 4 m2 to 6 or 8 m2 per adult individual resulted in higher use of the shelters as simultaneous use by a whole group was found to double when the available space was increased. Shelter use was also found to be affected by weather conditions as colder and harsher weather with lower chill factor index resulted in an increased use of the shelters. In the present study, the cattle did not have access to natural vegetation, and further research is needed to clarify whether cattle have preferences for specific types of shelter and the overall need for shelters in relation to animal welfare. Conflicts of interest None. Acknowledgements The authors would like to thank the owner and staff on the participating farm as well as research technicians, Anton Steen Jensen and Henrik Krogh Andersen, Aarhus University. This study was funded by the Danish Authorities and Aarhus University. References Ames, D.R., Insley, L.W., 1975. Wind-chill effect for cattle and sheep. J. Anim. Sci. 40, 161–165. Baxter, M.R., 1992. Overstocking reduces lying time in dairy cows. In: Phillips, C.P.D. (Ed.), Farm Animals and the Environment. CAB International, Wallingford, UK, pp. 67–81. Ekesbo, I., 2011. Farm Animal Behaviour: Characteristics for Assessment of Health and Welfare. CAB International, Oxfordshire, UK.
5