Introduction

The current dairy cattle production is mostly characterised by a demand for high production per cow in the environment that is rapidly changing to a less adequate one in terms of cattle breeding. Climate change is imminent according to the forecasts (IPCC 2007) and will have great impact on animal production worldwide. Battisti and Naylor (2009) stated that by year 2050, most of the world will experience median temperatures in the summer that are warmer than the warmest temperatures on record. Accordingly, Reiczigel et al. (2009), in Hungary, determined the increase of heat stress days/year (temperature-humidity index, THI > 68) from 5 to 17 in a period of 30 years. Also, Gauly et al. (2013) stated that in the global warming scenarios, the heat stress of high-producing dairy cows will be an increasing concern of milk producers in Europe, while Segnalini et al. (2013) emphasise the necessity of appropriate adaptation strategies development in order to minimise the negative effects of warming in farm animals in the Mediterranean basin. Dunn et al. (2014) determined that in the future, the number of days exceeding the THI threshold value in southern parts of the UK could increase from an average 1–2 per year to over 20 per year by 2100. Regarding Europe, GIRA–Consultancy and Research Prospective and Strategie (2012), in the analysis of Regional movements in EU Milk Production, forecasts the movement from regions with intensive farming towards regions around the Atlantic with less intensive farming and more land suitable for pasture (meaning lower production costs). Hansen (2013) stated that production increase makes cows more susceptible to heat stress; therefore, even if the climate change does not occur, heat stress will become an important problem. The modern dairy cow, with its high levels of productivity, begins losing the ability to regulate its body temperature at air temperatures as low as 25–29 °C. Studies of Bohmanova (2006) and Collier et al. (2006) showed that the high-producing cows are much more susceptible to heat stress than low-producing. Kadzere et al. (2002) suggested that the thermoregulation physiology of dairy cattle has been changed due to intensive genetic selection for milk production. The cows that produce more milk have larger frames and larger gastrointestinal tracts which allow the digestion of more feed. This results in more metabolic heat and reduces cow’s ability to maintain normal temperature at heat stress conditions. Kadzere et al. (2002) concluded that the increase of milk yield, feed intake and metabolic heat shifts the thermoneutrality to lower temperatures. Furthermore, Berman (2005) stated that the increase in the daily milk yield from 35 to 45 kg/day leads to a higher sensitivity to thermal stress and reduces the threshold temperature for intermediate heat stress by 5 °C. Heat stress in dairy cows reduces dry matter intake, milk production (West et al. 1999; Casa and Ravelo 2003) and reproductive performances (Bohmanova et al. 2007; Ravagnolo et al. 2000). Considering that heat stress affects feeding behaviour of dairy cattle, Collier et al. (2006) presume that predispose of dairy cows to metabolism disorders (acidosis/ketosis) could also be affected. Conversely, Sanker et al. (2013) found that heat stress did not influence the incidence of metabolism treatments. Also, heat stress is associated with changes in milk composition, somatic cell counts (SCC) and mastitis frequencies (Bouraoui et al. 2002; Collier et al. 2012; Correa-Calderon et al. 2004; Gantner et al. 2011; Ravagnolo et al. 2000; St-Pierre et al. 2003; West 2003; Hammami et al. 2013; Smith et al. 2013). Furthermore, the heat stress condition induces significant loss of profit, for example in the USA between $897 million and $1500 million per year (St-Pierre et al. 2003). The heat stress could be measured variously, but the most common measure in dairy cattle is the temperature-humidity index (THI). THI presents the combination of ambient temperature and relative humidity and is a useful and easy way to assess the risk of heat stress (Kibler 1964). Du Preez et al. (1990a, b) determined that milk production and feed intake is affected by heat stress when THI values are higher than 72. Bouraoui et al. (2002) put the threshold on 69, while Bernabucci et al. (2010) as well as Collier et al. (2012) on 68. Vitali et al. (2009) suggested that the risk of cow’s death starts to increase when THI reaches 80. The significant decrease of daily milk traits (yield and contents) was also determined in Croatian environmental conditions with the highest decline during the summer period in Eastern and Mediterranean Croatia (Gantner et al. 2011). The question is how do we effectively decrease the impact of heat stress on Croatian dairy farms? There are many methods to decrease the impact of heat stress for example shading, cooling and nutrition (Valtorta et al. 1997; Kadzere et al. 2002; West 2003). Also, selection for heat stress resistance could be an effective, long-term method (Bohmanova 2006). The antagonistic relationship between cow’s production and heat tolerance was determined by Ravagnolo et al. (2000). This relationship implies the deterioration effect of the selection on productivity on cow’s resistance to heat stress. The unfavourable genetic relationship between THI and productive and reproductive traits was also found in other studies (Ravagnolo and Misztal 2002a, b; Freitas et al. 2006; Aguilar et al. 2009). On the other hand, the high-producing Holsteins in Israel show that selection on production could be successful in terms of heat stress (Aharoni et al. 1999). Most of the research concerning heat stress was conducted on Holsteins, only few studies compare milk production of Jersey and Holstein breeds in terms of heat stress (Harris et al. 1960; Collier et al. 1981; Smith et al. 2013). Our earlier research (Gantner et al. 2017) aiming at the determination of THI threshold value for daily milk traits (yield, fat and protein content) of dairy cattle (Holsteins and Simmentals) indicated higher resistance to heat stress in Simmentals than Holsteins.

The objectives of this research were to determine the variation in daily milk yield (DMY) and somatic cell count (SCC) as well as the differences in resistance to heat stress due to production level and breed of dairy cattle in Croatia.

Materials and methods

Data

Individual test-day records of Holstein and Simmental dairy cows collected during the regular milk recording performed by alternative milk recording method (AT4/BT4) in the period from January 2005 to December 2012 in Croatia were used for the statistical analysis. Monthly, at each recording, milk yields were measured during the evening or morning milkings. Additionally, at each recording, ambient temperature and relative humidity were recorded. Daily temperature-humidity index (THI) was calculated using the equation by Kibler (1964):

$$ \mathrm{THI}=1.8\times \mathrm{Ta}\hbox{--} \left(1\hbox{--} \mathrm{RH}\right)\times \left(\mathrm{Ta}\hbox{--} 14.3\right)+32 $$

where Ta is the average temperature in degrees Celsius and RH is the relative humidity as a fraction of the unit. Records with lactation stage in (<6 and >500 days), age at first calving in (<21 and >36 months), missing or parity >6/7 (Holstein/Simmental), and missing or nonsense Ta and RH value were deleted from the dataset. Regarding the production level, cows were divided into two classes (high (≥20/15 kg) and low (<20/15 kg) with limit put on 20 kg of milk/day for Holsteins and 15 kg of milk/day for Simmentals. Regarding the parity, cows were divided into three groups: first, second and third+. Also, only cows with minimum three test days per parity were taken into analysis. Data, provided by the Croatian Agricultural Agency, after logical control consisted of 1,070,554 test-day records from 70,135 Holsteins reared on 5679 farms and 1,300,683 test-day records from 86,013 Simmentals reared on 8827 farms in Croatia.

Statistical analysis

The variation in daily milk yield (DMY) and somatic cell count (SCC) due to heat stress was determined by least square analyses of variance for each given THI value (from 68 to 78) with regard to the production level (high, low) and breed (Holstein, Simmental) separately for each parity class (first, second, third+) using the PROC MIXED procedure in SAS (SAS User’s Guide 2000). Used statistical mixed model included following effects: days in milk as lactation curve (i = 6 to 500 day); calving season as fixed class effect (j = 1/2005 to 12/2012); age at calving class as fixed class effect (k = 21 to 36 month)*only for first parity; region as fixed class effect (l = Croatian counties); and THI class as fixed class effect (m = 0 (normal condition—values under the given threshold) or 1 (heat stress condition—values equal and above the given threshold)).

The significance of the differences between the THI classes was tested by Scheffe’s method of multiple comparisons.

Results

Mean monthly values of daily ambient temperature, relative humidity and THI measured during milk recording of dairy cows in period of 2005–2012 are presented in Figs. 1 (Holsteins) and 2 (Simmentals). Mean monthly ambient temperature during milk recording of Holsteins significantly exceeded 20 °C in June, July and August, while in September, it was a bit lower. Mean monthly THI reached the value of 70 during July. The analysis of the mean monthly values of microclimate parameters showed similar conditions during the milk recording of both breed, Holstein and Simmental. The maximum values of ambient temperature, relative humidity and THI measured during milk recording of dairy cows in period from June to September in accordance to recording year are presented in Tables 1 (Holstein) and 2 (Simmental). The extremely high temperature and THI values (>35 °C; >90) indicate that both breeds were in heat stress conditions during the summer period, with slightly less unfavourable conditions measured in September.

Fig. 1
figure 1

Mean daily ambient temperature, relative humidity and THI measured during milk recording of Holstein cows

Full size image
Fig. 2
figure 2

Mean daily ambient temperature, relative humidity and THI measured during milk recording of Simmental cows

Full size image
Table 1 Maximum values of ambient temperature, relative humidity and THI measured during milk recording of Holstein cows in period from June to September in accordance to recording year
Full size table
Table 2 Maximum values of ambient temperature, relative humidity and THI measured during milk recording of Simmental cows in period from June to September in accordance to recording year
Full size table

The differences in the daily milk yield production of high and low-producing first parity Holsteins and Simmentals due to heat stress conditions are presented in the Table 3. Statistically highly significant decrease of daily milk production in amount of 0.158 (THI = 68) to 0.335 kg/day (THI = 72) was observed in high-producing first parity Holsteins. Lower drop in daily milk production was observed in low-producing first parity Holsteins with first significant drop in amount of 0.059 kg/day at the THI = 73. The high-producing first parity Simmentals experienced statistically highly significant decrease of daily milk production at all tested THI values with the highest drop when THI = 69. Although high-producing Simmentals exposed to the heat stress decreased their milk production, the drop was almost two times lower than in high-producing Holsteins. The analysis of the daily milk yield in low-producing first parity Simmentals showed small but significant increase of daily production when THI in (68, 77) and small, insignificant increase when THI = 78.

Table 3 Least square means (LSM) of daily milk yield (kg) regarding the given threshold value in accordance to the production level (high, low) and breed (Holstein, Simmental) for first parity cows
Full size table

Least square means of daily milk yield (kg) regarding the given threshold value in accordance to the production level (high, low) and breed (Holstein, Simmental) for second parity cows are shown in the Table 4. A similar decreasing trend of daily milk yield due to heat stress as in first parity high-producing cows was also determined in second parity cows. In Holsteins, the drop was in interval from 0.424 to 0.768 kg/day and significantly higher than in first parity cows. A much lower drop in (0.064, 0.234) was observed in second parity Simmentals. Low-producing cows, of both breeds, experience statistically significant increase of daily milk yield at every tested THI value (68–78).

Table 4 Least square means (LSM) of daily milk yield (kg) regarding the given threshold value in accordance to the production level (high, low) and breed (Holstein, Simmental) for second parity cows
Full size table

The variations in daily milk yield due to heat stress in the third+ parity cows regarding the production level and breed are presented in the Table 5. The decreasing (in high-producing) and increasing (in low-producing) trends of daily milk production were similar as in second parity cows. In high-producing Holsteins, the drop varied in (0.422, 0.763), while in Simmentals, the drop was smaller and varied in (0.090, 0.235). The increase in daily milk yield in low-producing Holsteins (0.023–0.130 kg/day) was a bit higher than in Simmentals (0.066–0.094 kg/day).

Table 5 Least square means (LSM) of daily milk yield (kg) regarding the given threshold value in accordance to the production level (high, low) and breed (Holstein, Simmental) for third+ parity cows
Full size table

The variation in somatic cell count (SCC) due to heat stress in high and low-producing first parity cows is shown in the Table 6. Statistically highly significant increase of SCC was determined in high-producing first parity Holsteins at all THI classes (68–78). A significant increase of SCC due to heat stress was also determined in low-producing first parity Holsteins, but the increase was lower than in high-producing cows. When Simmental first parity cows were analysed, a low decrease in SCC was determined until THI = 77; while in THI ≥ 78, an insignificant increase was determined. In low-producing first parity Simmentals, statistically significant decrease of SCC was determined at all tested THI classes (68–78).

Table 6 Least square means (LSM) of somatic cell count (in 000) regarding the given threshold value in accordance to the production level (high, low) and breed (Holstein, Simmental) for first parity cows
Full size table

The response to heat stress with regard to the breed and production level of second parity cows is presented in the Table 7. A similar response as in high-producing first parity Holsteins, that is significant increase of SCC, was determined in second parity cows. However, low-producing second parity Holsteins in heat stress condition tend to decrease the SCC. In high-producing second parity Simmentals, the variation in SCC due to heat stress was insignificant while low-producing cows tend to cause statistically significant decrease of SCC.

Table 7 Least square means (LSM) of somatic cell count (in 000) regarding the given threshold value in accordance to the production level (high, low) and breed (Holstein, Simmental) for second parity cows
Full size table

The results for third+ parity cows are shown in the Table 8. The increase of SCC was determined only in high-producing Holsteins, while in high-producing Simmentals, the variations were mainly insignificant. In low-producing group, a significant drop was observed in Holsteins when THI in (68, 69, 70, 71), while in Simmentals, a statistically highly significant drop was determined at all tested THI values (68–78).

Table 8 Least square means (LSM) of somatic cell count (in 000) regarding the given threshold value in accordance to the production level (high, low) and breed (Holstein, Simmental) for third+ parity cows
Full size table

Discussion

The analysis of the microclimatic conditions during the milk recording of dairy cattle in Croatia indicates that mean daily temperature significantly exceeded 20 °C during the summer period with slightly less hot conditions in early autumn. The determined maximum values of daily ambient temperature (>35 °C) in combination with high maximum values of relative humidity (>90) resulted in incidence of summer days with THI > 90. Johnson (1987) estimated the optimal temperature zone for lactating dairy cows in (−0.5, 20 °C), while Berman et al. (1985) put the upper critical air temperature for dairy cows at 25–26 °C. The dependence of temperature and humidity was shown in the study of Bianca (1965) who reported that at a temperature of 29 °C and 40% relative humidity, the milk yield of Holstein, Jersey and Brown Swiss cows was at 97, 93 and 98% of normal production, but when relative humidity increased to 90%, yields dropped to 69, 75 and 83% of normal production. Regarding the THI values, Vitali et al. (2009) suggested that the risk of cow’s death starts to increase when THI = 80. These findings indicate that both dairy cattle breeds that were analysed were exposed to heat stress conditions during the summer and early autumn.

The analysis of the daily milk yield variations indicates that the response to heat stress highly depends on daily production level, cow’s breed and parity. As for instance, high-producing cows of both breeds experienced decrease in daily milk yield in terms of heat stress in all parities (first, second and third+) with the highest decrease in the second parity. Furthermore, the Holsteins experienced higher drop in daily milk yield than Simmentals. The first significant drop of daily milk yield, in the high-producing cows, was observed at THI = 68 which could be taken as the threshold value for genetic evaluation. On the other hand, low-producing cows, with exception of first parity Holsteins, increased their daily milk production in the environment above the given THI.

The negative effect of heat stress on daily milk yield of dairy cattle was quite well researched. For instance, Casa and Ravelo (2003) concluded that heat stress in the warmer months in Argentina induces production decrease, in relation to the normal production level of 22 l/cow/day, in amount of 6% (9%) depending of the region. Bernabucci et al. (2010) and Collier et al. (2012) determined a decrease when THI = 68, while Bouraoui et al. (2002) in the Mediterranean climate observed a decrease in milk production of dairy cows when THI ≥ 69. Bouraoui et al. (2002) determined a decrease of milk yield by 0.41 kg per cow per day for each point increase in the value of THI above 69. Du Preez et al. (1990a,b) observed that dairy cows in South African conditions are affected by heat stress in case when THI values are higher than 72. The same authors determined that the amount of milk yield for Holstein cows decreased about 10 to 40% during the summer period in comparison to the winter period (Du Preez et al. 1990b). A significant decrease of daily milk yield when THI ≥ 72 was also noticed in Croatia (Gantner et al. 2011). Research conducted in the USA determined different threshold values regarding the region, for example 72 in Georgia and 74 in Arizona (Bohmanova et al. 2007). The difference between determined threshold values could be due to better adapted cows, farm management or special housing characteristics. For instance, Lambertz et al. (2014) determined that a drop of fat-corrected milk (FCM) in Holstein cows highly depends on housing systems with the highest drop observed in the classic indoor system. They concluded that heat stress resulted in decreasing milk yield, fat and protein contents, and increasing somatic cell score (SCS).

The differences in response to heat stress due to the breed were analysed in several studies. For example, a higher drop of milk production in Holstein cows comparable to the other breeds (Jersey and Brown Swiss) was determined by Bianca (1965). Also, Collier et al. (1981) suggested that Jersey cows may be more heat tolerant than the Holsteins with respect to milk yield production. Smith et al. (2013) determined the increase of milk production of Jersey cows during heat stress. Also, in the same study, the authors determined a decrease of milk production of Holstein cows during heat stress. They concluded that Jersey cows used in their study could be more heat tolerant than Holstein cows.

Similarly, like daily milk yield, the differences in somatic cell count (SCC) due to heat stress highly depend on daily production level, breed and parity. The highest increase in the daily SCC was determined in first parity high-producing Holsteins, while statistically significant but lower increase was determined in multiparous cows. The first significant increase of SCC, in the high-producing Holsteins, was observed at THI = 68 which could be taken as threshold value for genetic evaluation. However, Simmentals, high and low producing, tended to decrease SCC with higher decrease in multiparous and low-producing cows.

The milk SCC presents the most important indicator of the udder health. The results indicate that in heat stress condition, Holsteins may experience higher prevalence of subclinical and clinical mastitis. The negative effects of heat stress with increasing SCC from spring to summer were also determined by Bouraoui et al. (2002). Lambertz et al. (2014) observed that increasing THI values were associated with increasing SCS in Holsteins, in four different housing systems. Smith et al. (2013) concluded, based on the analysis of heat stress effects in Holsteins and Jersey breed in terms of milk and component yields and somatic cell score, that Jersey cows appeared to be more heat tolerant than Holstein cows.

The results of this research comply with the findings of Kadzere et al. (2002), Bohmanova (2006), Collier et al. (2006) and Hansen (2013) which imply that the high-producing cows are much more susceptible to heat stress than low-producing. Also, the results of this research indicate that Simmental breed could be more resistant to heat stress than Holstein. Taking into consideration that the climate conditions are certainly going to deteriorate in Croatia, which means higher summer temperatures with more frequent prevalence of heat stress days, an adequate development strategy for dairy farming is necessary in order to ensure effective and profitable dairy farms. Further research should determine whether Simmentals are genetically more appropriate for the challenges that face the future milk production in this region. Furthermore, could an adequate production level be achieved with Simmentals by maintaining of the heat resistance?