Comparison of dairy cow step activity under different milking schedules

Introduction

Consumers of animal products are increasingly taking animal welfare into consideration when making a purchase (European Commission 2007; Napolitano et al. 2008; Alonso et al. 2020). A widely accepted definition of animal welfare focuses on three aspects, namely, whether the animal is (1) functioning well, (2) feeling well, and (3) able to live a reasonably natural life (Fraser et al. 1997). Dairy farming has undergone intensification over time, which has contributed to an increase in average herd size (Clay et al. 2020). In pastoral systems, an increase in herd size has led to cows walking longer distances to the dairy, and an increase in time spent away from pasture (Beggs et al. 2019). Increasing time away from pasture, the amount of time spent standing on concrete, and the distance cows walk, may affect the expression of natural behaviours, and thus have an impact on animal welfare.

The milking frequency and milking interval can influence the time cows spend away from pasture. The milking frequency is defined as the number of milkings per day, and the milking interval is the time between milkings. Dairy farmers can adapt both of these to suit their needs.

Twice-a-day (TAD) milking frequency is common in pastoral dairy farming systems, and it usually results in greater milk production than a once-a-day (OAD) milking frequency (Edwards 2018). Milking more than TAD is rare in pastoral systems as the revenue from the additional milk does not cover the cost of the additional milkings (Culotta and Schmidt 1988). Milking three times in 2 days (3-in-2) is an intermediate option, between TAD and OAD, that gives some of the benefits of OAD with a lesser impact on milk production (Edwards et al. 2022). Changes in cow behaviour have been identified between OAD and TAD milking; however, there has been little research investigating the effect of milking 3-in-2 on cow behaviour.

There has been little research around the effect of milking frequency on activity (Hall et al. 2021). Most research has focused on the effect of milking frequency on behaviours such as eating and lying time rather than activity, as they are seen as more important to the cow (Munksgaard et al. 2005). Most research on total walking distance has focused on farm productivity metrics, such as decreased milk production and profit (Thomson and Barnes 1993) or lameness incidence (Crossley et al. 2020), rather than aspects of cow behaviour.

A recent study reported that on days where cows walked further, grazing time increased and rumination time decreased (Neave et al. 2021). The authors hypothesised that this was due to the increased energy expenditure of walking further. However, the maximum distance walked in the study was 4 km/cow.day under a TAD milking regime. This is a conservative distance for farms of over 200 cows. Another study investigated the relationship between cow lying time and their steps and found that for every additional 1000 steps/day, lying time decreased by 0.49 h (Beggs et al. 2018). As lying time is an important resting behaviour, it is likely prioritised by cows (Munksgaard et al. 2005), and they change other behaviours to preserve it. These observations suggest that there may be an effect of milking frequency, through walking distance, on behaviour.

The choice of milking frequency and interval affects the timing of milkings within a day. Grazing behaviour has a significant diurnal pattern, with major grazing bouts occurring in the early morning and late afternoon into early evening (Gibb et al. 1998), with the later grazing event being more intense. However, dairy cows milked TAD are typically milked in the early morning, and late afternoon, potentially affecting these natural cycles. In a previous study, researchers observed major grazing bouts after milking (Brumby 1955) as cows tried to rectify the disruption to their natural routine. This is likely further influenced by whether fresh forage is offered after milking. For example, where forage is offered later in the morning, the morning feeding bout can be delayed (DeVries and von Keyserlingk 2005). With respect to milking frequency, despite no difference in grazing time being measured between OAD and TAD, TAD cows grazed later into the evening because of the disruption of a second milking in the afternoon (Tucker et al. 2007). Rumination follows a grazing bout, often with a period of idling in between (Gregorini 2012), and these activities are important for efficient digestion and supporting high intakes (Beauchemin 2018). Factors that influence the timing of grazing, including the timing of milkings, will also likely affect the timing of rumination within the day.

Previous studies have noted a stage of lactation effect on behaviour (Bewley et al. 2010; Munksgaard et al. 2020); however, this may be confounded by the other factors that change with the stage of lactation, particularly in block calving pasture-based systems. These factors could include cow milk production, age, and body condition score (BCS), as well as the environmental temperature. Milking frequency affects udder fill but the existing research does not indicate that udder fill significantly affects lying time (Tucker et al. 2007; O’Driscoll et al. 2010, 2011). Research has shown that cows past their first lactation increase their lying time in later lactation (Munksgaard et al. 2020). This suggests that age in combination with stage of lactation affects lying behaviour. Cows with a lower BCS decreased their lying time and increased their grazing time in late lactation, when compared with higher-BCS cows (Matthews et al. 2012). This could potentially be caused by a drive to increase the lower-BCS cows’ heat production from rumination or because of hunger. Environmental temperatures that exceed a cow’s upper thermal neutral zone of 25°C (Avendaño-Reyes 2012) reduced grazing time, because, to dissipate heat, cows prefer to stand, seek shade, and decrease digestion as that produces large amounts of heat (Kadzere et al. 2002; Herbut et al. 2021). Consequently, it is important to separate the stage of lactation and any other factors that may influence changes in behaviour.

The objective of this study was to identify any differences in cow step activity, measured as steps per hour, among three milking frequencies and three milking schedules of 3-in-2 (milking three times in 2 days) milking, at two stages of lactation. Time spent eating was assessed to help explain differences in step activity within a day. These were investigated using data from five herds of 40 cows (n = 200) milked OAD, TAD, and three 3-in-2 intervals of 12–18–18 h, 10–19–19 h, and 8–20–20 h. The information from this study will allow farmers to assess whether the milking frequency and milking interval they use on farm are altering their cows’ behaviour.

Materials and methods

Results

CowManager behaviour data

Due to the small number of tags available per treatment group, raw means for each behaviour category are presented by treatment group and experimental period (Table 2).

Table 2.A comparison of early lactation and mid-lactation raw means by milking frequency and interval: once a day (OAD); twice a day (TAD); and three intervals of milking three times in 2 days (3-in-2); 8–20–20; 10–19–19 and 12–18–18 (n = 8 cows/treatment) for total minutes per day in the following behaviour categories: eating, ruminating, active, high active and not active, as defined by the CowManager tags, over the experimental period (n = 6 weeks).

Parameter	Early lactation					Mid lactation
	OAD	8–20–20	10–19–19	12–18–18	TAD	OAD	8–20–20	10–19–19	12–18–18	TAD
Eating	305	389	456	439	336	444	382	338	403	446
Ruminating	454	403	427	420	492	463	427	454	458	437
Active	164	168	131	114	196	134	243	211	213	160
High active	218	199	191	200	171	153	145	194	149	147
Not active	305	295	250	286	262	248	249	269	222	254

The eating time data were analysed to attempt to explain differences in the step activity of different trial groups (Table 1). In the early lactation trial period, cows’ eating time increased after milking for all the treatment groups (Fig. 1). After 10 am, when compared with the TAD treatment group, the OAD treatment group had an extended period of eating throughout the afternoon into the evening, with a less defined afternoon peak than for the TAD treatment group. The 3-in-2 8–20–20 and 12–18–18 intervals were also compared, as they had the largest difference in milking time and they had a significant difference in step activity (Table 1). The influence of milking times was evident. On Day 1, the 12–18–18 treatment group was milked at 17:00 hours and had a peak in eating time later in the afternoon, while the 8–20–20 treatment group was milked at 14:00 hours and had an extended period of afternoon eating. On Day 2, both treatment groups were milked in the mid-morning and their eating patterns were similar.

Fig. 1.

A graphical representation of early lactation eating time in minutes per hour per cow (n = 8) over a period of 24 h by miking frequency; once a day (OAD) vs twice a day (TAD) and milking three times in 2 days (3-in-2) intervals; 8–20–20 and 12–18–18, compared with milking times and divided into Day 1 (when cows are milked twice) and Day 2 (when cows are milked once) data for 3-in-2 treatments.

Discussion

The primary objective of this study was to identify the effect of milking frequency, including 3-in-2 milking intervals, on cows’ step activity, as indicated by the differences in mean steps per hour per cow. A secondary objective was to investigate diurnal differences in eating time to assist with the explanation of within-paddock differences in step activity.

An increase in milking frequency resulted in an increase in mean steps per hour in both the early and mid-lactation trial periods. As milking frequency dictated the number of times the cow walked between paddock and the dairy, with an average distance of between 1.4 and 1.5 km one way, the relationship between the number of milkings per day and average number of steps is not unexpected. However, when the data from the period that included travel to and from milking was removed, leaving step per hour values for the time spent in the paddock, a relationship remained between milking frequency and steps per hour.

Cows in the OAD treatment group had the lowest number of steps per hour during both the early and mid-lactation trial periods whilst in the paddock, despite having the greatest amount of time there due to their milking schedule. Differences in feed searching behaviour may be an explanation for this observation. Fig. 1 illustrates that the eating time of the OAD treatment group was more prolonged, with a slow rate of increase across the afternoon, when compared with the TAD treatment group. Perhaps the cows milked OAD felt a lesser need for rapid pasture intake as they were not removed for afternoon milking and were used to this routine, and other behaviours, such as grooming, ruminating or resting, were prioritised over the afternoon. It is also possible that due to less milk production from the OAD cows, their appetite was reduced and therefore grazing time was reduced, resulting in fewer steps. Previous research investigating the relationship between eating time and milking frequency (Tucker et al. 2007) or prolonged milking (Beggs et al. 2018) has not shown any differences in daily eating time, but diurnal variation and grazing intensity were not explored.

An alternative hypothesis could be that OAD cows were less active due to discomfort, as their udders were distended due to higher milk volumes caused by lower milking frequency. One study reported decreased lying time due to udder discomfort when a milking was omitted, showing that udder discomfort can change behaviour (O’Driscoll et al. 2011). In this trial, step activity was higher during the mid-lactation trial period, when cows produced less milk, than during the the early lactation trial period, and cows in the 3-in-2 treatment groups walked fewer steps on the day they were milked once. Only one study has investigated the correlation between activity and milking frequency. The study reported a relationship between decreased milking frequency from TAD to 3-in-2 and increased activity (Hall et al. 2021). More research is needed to understand which behaviours cows substitute for movement on days with only one milking, and the reasons for those differences.

When the step data associated with walking to and from milking were removed, the TAD trial group completed fewer average steps per hour than did the 3-in-2 trial groups, including on the day where 3-in-2 trial groups were milked twice. Greater activity in cows switched from TAD to 3-in-2 milking has been observed previously (Hall et al. 2021). This could have been due to an increase in feed searching behaviour, particularly in the evening, as a result of the larger area given per cow at each feed allocation. This hypothesis could be supported by the data from the early lactation trial period in our experiment; however, the pasture management changed for the mid-lactation trial period, where all treatments received the same-sized allocation each day, and the 8–20–20 herd remained more active. As the afternoon milking for the 8–20–20 herd occurred earlier than for the TAD herd (at 14:00 hours versus 16:30 hours), it is possible that the 8–20–20 herd had two grazing bouts, one occurring after milking and another in the early evening, resulting in more steps. Consequently, we speculate that the amount of walking TAD cows were required to do, due to their milking schedule, resulted in less step activity in the paddock.

A difference in steps per hour was also measured between 3-in-2 milking intervals, driven by differences on Day 1, when cows were milked twice. This suggests that the timing of milking influences step activity. The 8–20–20 trial group was milked at 06:00 hours and 14:00 hours on Day 1, and at 10:00 hours on Day 2, while the 12–18–18 trial group was milked at 05:00 hours and 17:00 hours on Day 1, and at 11:00 hours on Day 2. Cows have preferential grazing bouts early in the morning, and late afternoon into early evening (Gibb et al. 1998), with the afternoon bout being more intense. Milking times of the 12–18–18 trial group on Day 1 appear to have coincided with the cows’ preferred grazing time (Fig. 1). The 8–20–20 trial group, which was less disrupted by milking, clearly shows two separate peaks of eating over an extended period in the afternoon after milking, while the 12–18–18 herd had only one. The pattern of eating followed a similar trend when both trial groups were milked once on Day 2. Consequently, we hypothesise that the step-activity differences seen in the different 3-in-2 trial groups were driven by changes in grazing behaviour. Although there is little research on this topic, it is supported by one study that showed TAD cows grazing further into the evening than OAD cows (Tucker et al. 2007). The expression of natural behaviour is an important component of animal welfare, and thus the impact of milking timing on grazing behaviour is a consideration.

A key reason why farmers wish to reduce milking frequencies and extend milking intervals is to reduce work hours and early morning work starts, for their own welfare and so that they can attract and retain staff. However, it is important to also consider whether altering milking frequencies and intervals has a behavioural impact on cows. When looking at the impacts of milking frequency on a cow, most research has focused on farm metrics, such as milk production (Edwards 2018), or traditional welfare metrics, such as BCS (Roche et al. 2009). From our research, if we assume that the average length of the cow’s step is 1.5 m (Alsaaod et al. 2017), and the energy expenditure per kilometre walking over flat ground is 2 MJ (Nicol and Brookes 2007), then a difference between groups of 27 steps/h would provide a difference of energy expenditure of 2 MJ/cow.day. The largest significant difference we saw between the OAD and TAD herds was 115 steps/h at mid-lactation. If this is extrapolated to increased pasture demand with a large herd, this could result in a large energy deficit if step-activity differences are not considered when changing milking frequencies. Welfare also covers many facets, including normal behaviour expression. When farmers consider adjusting their milking times and frequency to suit their goals, they should ensure they account for potential changes in normal grazing behaviour and grazing allocation. Putting systems in place, such as limiting herd sizes to ensure that cows are not away from the paddock any longer than necessary and that efficient milking practices are used (Edwards 2013), will also help ensure that disruption to natural behaviour is minimised.

Conclusions

Milking frequency affected cow step activity, both when walking to the dairy was included and excluded from the data. Changes in diurnal eating patterns were also evident when comparing OAD and TAD trial groups. We believe that these differences were likely to be due to the amount of walking between the paddock and dairy, and changes in grazing behaviour or another explanation such as udder discomfort. However, because milking frequency also affects milk production, more research is needed to understand the relationship among milk production, milking frequency and behaviour. The timings of milkings also affected cow step activity, with significant differences seen amongst the 3-in-2 trial groups, driven by differences on the day cows were milked twice. There were differences in the eating pattern amongst the 3-in-2 trial groups, suggesting the timing of milkings, in addition to the frequency, affects grazing behaviour and, thus, step activity. More research is needed to determine the impact of these differences on animal welfare. Farmers should offset these changes in grazing behaviour due to milking frequency and timing by minimising the time cows spend out of the paddock, especially in the afternoon.

Data availability

The data that support this study cannot be publicly shared due to ethical or privacy reasons and may be shared upon reasonable request to the corresponding author if appropriate.

Conflicts of interest

We declare no conflicts of interest.

Declaration of funding

This study was part of the Flexible Milking Project, funded by the Sustainable Farming Fund (Project 405879), and the dairy farmers of New Zealand via DairyNZ Inc. (Contract TW2001).