The effect of seasonal thermal stress on milk production and milk compositions of Korean Holstein and Jersey cows

Objective In this study we investigated the effect of seasonal thermal stress on milk production and milk compositions between Holstein and Jersey dairy cows under the temperate-climate in Korea. Methods A total of 9 Holstein lactating dairy cows (2.0±0.11 parity) which had a daily milk yield of 29.77±0.45 kg, and days in milk of 111.2±10.29 were selected similarly at the beginning of the experiments in each season. Also, a total of 9 Jersey lactating dairy cows (1.7±0.12 parity) which had a daily milk yield of 20.01±0.43 kg, and days in milk of 114.0± 9.74 were selected similarly at the beginning of the experiments. Results Results showed that the average ambient temperature (°C) and temperature-humidity index (THI) were higher in summer, and were lower in winter (p<0.05). The average relative humidity (RH, %) was higher in autumn than that of other seasons (p<0.05). Milk production was significantly decreased (Holstein 29.02 kg/d and Jersey 19.75 kg/d) in autumn than in other seasons (Holstein 30.14 kg/d and Jersey 20.96 kg/d). However, the milk production was negatively correlated in Holstein cows, and positively correlated in Jersey cows with THI values increased from 16 to 80. In addition, milk yield was increased by 15% in Holstein cows and decreased by 11% in Jersey cows with the THI values increased from 16 to 20. The fat and protein content percentage was significantly higher in Jersey milk than in Holstein milk, furthermore the fat and protein content yield was higher in Jersey cow milk than that of Holstein cow’s milk with all THIs. Conclusion From the study results, we concluded that Jersey cows might be less adaptable to low temperature of the winter, and this would have a negative impact on dairy farmer income since Korea’s milk price estimation system places a higher value on milk yield than on milk compositions or sanitary grades.


INTRODUCTION
Climate change is transforming the earth ecosystems, and such change is producing a series of damaging abnormal weather phenomena, with the summer prolonged and the temper ature relatively increasing in the world [1]. Early research demonstrated that the increasing trend of mean/max/min surface temperature is evident during the summer season in Korea [2].
Heat stress (HS) is one of the most significant challenges to milk production facing dairy farmers during the summer. Climatic conditions in the Korean peninsula are such that the summer season is relatively hot, and generally accompanied by high relative hu midity (RH). Thus HS is chronic in nature, there is often little relief from the heat during the evening hours, and intense bursts of combined temperature and humidity further de press performance. Lactating dairy cows produce a large quantity of metabolic heat and accumulate additional heat from radiant energy. Production and accumulation of heat, coupled with compromised cooling capability because of environmental conditions, causes heat load in the cow to increase to the point that body temperature rises, intake declines and ultimately the cow's milk productivity declines [3].
In addition, as breeders select for higher milk production, they can mistakably also select for cows with less heat toler ance. Jersey breed is the second most important dairy breed in the world, the Jersey breed is smaller than Holstein breed, and their feed consumption is less than that of Holstein cows [4]. Besides, the faecal output, carbon gases emission, and space need for the breeding of Jersey cows are lower than that of Holstein cows [4,5]. Therefore, Jersey breeding is relatively beneficial to dairy farmers. Early research also demonstrated that as measured by milk yield Jersey cows might have more heat tolerance than Holstein cows [6]. For all these reasons, the dairy farmers and industries have a growing interest in raising Jersey breed. However, the relationship between breed and responses to seasonal HS has not been studied in Korea. Therefore, the objective of the current experiment was to evaluate the effects of seasonal changes on milk production and milk composition of Jersey breed compared with Hol stein breed in Korea.

Animals, management and experimental design
The study was conducted at the Department of Animal Re sources Development, National Institute of Animal Science (NIAS), Cheonan, Republic of Korea (36°55′53.2″ North, 127°06′22.1″ East, 23 m altitude). All dairy cows were main tained as stated in standard guidelines, and the experimental protocol involved in this research was approved by the In stitutional Animal Care and Use Committee (IACUC) at National Institute of Animal Science, Jeonju, Republic of Korea. A total of 9 Holstein lactating dairy cows (2.0±0.11 parity) which had a daily milk yield of 29.77±0.45 kg, and days in milk of 111.2±10.29 were selected similarly at the beginning of each season. Also, a total of 9 Jersey lactating dairy cows (1.7±0.12 parity) which had a daily milk yield of 20.01±0.43 kg, and days in milk of 114.0±9.74 were select ed similarly at the beginning of the season. The experiment was carried out from January 2016 to July 2017. The cows were raised under the same management and environmen tal conditions and housed in an open loose barn. The loose open barn was designed with the overshot roof of a ridge exhaust, fans to move and exchange the air in summer, and winchcurtain to block the cold wind in winter. The total mixed ration (TMR) was offered once a day at 09:00 h. The composition of the TMR remained the same throughout the year and included corn silage 28.1%, mixed hay 14.6%, alfalfa hay 8.3%, concentrate mixture 39.1%, soybean meal 4.4%, corn grain 4.4%, and mineral and energetic compo nents 1.0%. The TMR contained, on average, 47.3% dry matter (DM), 16.6% crude protein (CP), 4.96% ether ex tract (EE), 38.7% neutral detergent fiber (NDF), and 19.5% acid detergent fiber (ADF) on a DM basis. The percentages of DM, CP, and crude fiber in oat silage were 28.5%, 6.7%, and 35.1% and 30, 7.11%, and 34% for the spring and sum mer periods, respectively. The energy content of feed rations of all cows was 1.7 Mcal net energy of lactation/kg DM and total digestible nutrients 68.7%. TMR samples were taken monthly and stored at -20°C until analysis. Feeding was allowed throughout the 24hour period. Drinking water was made available at all times.

Measurement, sampling, and laboratory analysis
The ambient temperature (Ta, °C) and RH (%) was moni tored with a thermohygrometer (Testo 174H, Testo Inc., West Chester, PA, USA) with an accuracy of ±0.5°C, and ±3% RH. The thermohygrometer was set to record every day per 30 minutes and placed in the feeding area at a height of 2 meters above the ground. The temperature and humidi ty values were used to calculate several THI values; THI was calculated for each 30 min temperature and humidity mea surement according to the formula: THI = (0.8×°C)+(RH %×[°C-14.4])+46.4 [7]. The mean daily THI forms the mean of all available 30 minutes THI's per day. The maximum THI values for everyday were calculated according to the same formula of Fox and Tylutki [8], with the maximal tempera ture and the minimum RH/d, and the minimum THI values for everyday were calculated with the minimum tempera ture and the maximum RH/d. TMR samples were analyzed for DM, CP, and EE by AOAC (1995) procedure, and NDF and ADF by Van Soest et al [9]. The cows were milked two times a day (06:00 and 17:00 h) and milk yield of the individual cows was recorded at each milking on all experimental periods. Milk samples were collected from each cow twice daily at morning (06:00 h) and afternoon (17:00 h) biweekly, and was analyzed for pro tein, fat, and somatic cells using with a LactoScop (MK2, Delta Instruments, Drachten, the Netherlands).
Milk yield and milk protein and fat percentage were used to calculate the fat protein corrected milk (FPCM) yield [10]. The yields of milk fat and protein were calculated from milk yield and the contents of milk fat and protein, respectively.

Statistical analysis
The data analysis was performed with the statistical package SAS Enterprise Guide 7.1 (SAS Institute Inc., Cary, NC, USA). Seasons were defined as follows: spring (March to May), summer (June to August), autumn (September to November), and winter (December to February). The THI measured during the experimental period was divided into 13 classes based on the daily mean as follows: 16 to 20, 21 to 25, 26 to 30, 31 to 35, 36 to 40, 41 to 45, 46 to 50, 51 to 55, 56 to 60, 61 to 65, 66 to 70, 71 to 75, and 76 to 80.
The effects of the THIs and seasons on the milk produc tion and milk compositions of each breed (Holstein and Jersey cows) were analyzed by using the oneway analysis of vari ance (ANOVA). Tuckey's test was used to separate the means when significance was indicated. The effects of a breed and THI, or breed and season on the milk production and milk composition were analyzed by using the mixed procedure with a restricted maximum likelihood model. The breed, THI or season, and their interactions were included as fixed effects. For comparisons between breeds, THIs or seasons, and breed×THIs or seasons, the pvalues after Bonferroni post hoc test adjustment are presented. The values are pre sented as leastsquares means and standard errors of the means unless otherwise stated. Differences were considered significant if a probability (p) of <0.05 was found, and trends are discussed for variables with p≤0.10.

RESULTS
Meteorological data including the mean minimum and max imum Ta, RH, and THI measured during milk recording of each breed dairy cows in period from 2016 to 2017 are sum marized in Table 1. The recorded average daily Ta and THI were higher in summer (24.02°C±0.05°C, 72.55±0.09), and was lower in winter (-1.62°C±0.09°C, 35.11±0.12) than that of all other seasons (p<0.05). The results showed that the av erage RH was significantly highest in autumn (72.51%) and was significantly lowest in spring season (59.75%) (p<0.05). The average daily THI ≥72 was observed to be 0.96% in the spring, 57.78% in summer, and 2.11% in the autumn, which indicates that cows were exposed to the critical point of HS. In contrast, cows experienced the cold stress (THI ≤38) by 0.48% in the spring, 2.82% in the autumn, and 68.86% in the winter season.
The seasonal thermal stress effects on milk yield and FPCM yield trend between the Holstein and Jersey cows are repre sented in Table 2 and Figure 1. In Holstein dairy cows, milk production was decreased from the spring (30.94 kg/d) to autumn (29.02 kg/d), and then increased to winter season (31.12 kg/d) (p<0.05). However, in Jersey cows, milk pro duction was lower in autumn (19.75 kg/d) and winter (19.90 kg/d) season than that of spring (21.22 kg/d) and summer season (20.96 kg/d) (p<0.05).
Results in Figure 2 show that the milk production of Hol stein was highest (35.52 kg/d) at the 16 to 20 among the all THI ranges, then started to decline and remained constant (29.74 to 31.10 kg/d) when the THI values increased from 31 to 80 (p<0.05). On the contrary, the milk yield of Jersey cows was lowest (18.62 kg/d) at the 16 to 20 among the all THI ranges, and then the milk yield increased (20.63 to 21.22 kg/d) with a THI of 51 to 80 (p<0.05).
The present study has shown the relationship between the milk yield and THI values in Holstein and Jersey cows ( Table   Table 1    ue of 66 to 70. The seasonal thermal stress effects on milk composition between the Holstein and Jersey cows are represented in Ta ble 4. The study results showed that the milk fat percentage was higher in the winter (Holstein 4.19%, Jersey 5.62%) among the seasons in all breed's milk (p<0.05). In Jersey cows, the milk fat percentage was lower (4.93%, p<0.05) in summer than in the other season. However, the milk protein percent age of Holstein cows gradually increased from the spring (3.21%) to the winter (3.44%) (p<0.05). However, the milk protein percentage of Jersey cows was increased more in au tumn (3.88%) and winter (4.00%) than in spring (3.75%) and in summer (3.52%) (p<0.05). In Holstein cows, somatic cell count (SCC) was significantly increased in the summer (160.46×10 3 /mL). But, in Jersey cow, SCC was significantly higher in the autumn (275.07×10 3 /mL).
The effects of THI on the yield of milk fat and protein of Holstein and Jersey cows are presented in

DISCUSSION
The average daily Ta (°C) and THI were significantly higher in summer and lower in winter when measured data from each seasonal period were analyzed (p<0.05). However, the average RH (%) was highest in autumn (p<0.05). The THI represents the combination of Ta and RH, and used as a com prehensive indicator to assess the heat and cold stress degree of dairy cows. Some studies have explored the thermoneu tral zone which is between 5°C and 25°C [11], and is within the range of -0.5°C to 20.0°C Ta and 60% to 80% RH [12].  When THI exceeds 72, the cows begin to experience HS, while when THI is less than 38, the cows begin to experi ence cold stress [13]. Among our study results, the average daily Ta exceeded 20°C on 8.65% of the time in the spring, 92.00% in the summer, and 24.65% in the autumn season. Also, average daily Ta ≤5 was recorded on 10.10% in the spring, 11.97% in the autumn, and 96.41% in the winter season. The average RH of >80% were higher in summer (23.21%) and autumn (23.94%), compared with the spring (7.21%) and winter (13.17%) season. The study results con firmed that the cows may suffer from both the heat and cold stress in spring and autumn season, particularly cows ex perienced the stress with higher RH (>80%) in autumn season than in spring season. The average milk yield was reduced by 6.2% and 6.9% in Holstein and Jersey cows respectively as the average THI values changed from the spring to the autumn season and this lowered level may be due to the energy expenditure. Our study results are supported by Johnson et al [14] who re ported early that the greater heat production can explain the increasing rate of decline in milk yield for cows. Also, the same authors explained that the permanent drop in the cur rent lactation is proportional to the length of the HS. This study is consistent with earlier findings of Bohmanova et al [15] that the decline in milk production due to HS occurs with higher temperature in a semiarid climate (≥30°C, 25% RH) than in a humid climate (≥23°C, 75% RH).
Berman [16] reported that the higher milk yield is asso ciated with HS, which by itself would shift lower critical temperature (LCT) to lower ambient temperatures. The LCT decreases in -4°C [17] to values of -30°C during peak lactation [18], -37°C to -16°C with 30 kg/d of milk yield [19], -40°C with 36 kg/d of milk yield [20], or even -45°C [21]. Milk yield could be affected between Holstein and Jersey cows in the range of LCT. Hence, the average milk yield in winter season was increased by 7.2% in Holstein cows compared to the autumn season; with the average THI and Ta being 35.29°C and -1.49°C, respectively. But, no sig nificant difference was observed in Jersey cow's milk yield between the autumn and winter season.
In particular, the study results showed the different trends in the milk production of Holstein and Jersey cows, when the THI values increased from 16 to 80. A negative regres sion slope line was observed in Holstein cow' s milk production as THI increased from 16 to 80. Milk yield of Holstein and Jersey cows was reduced slightly (1%) at THI range of 76 to 80 as compared with THI range of 66 to 70. However, milk yields were increased by 15% in Holstein cows at the THI values of 16 to 20, but milk yield was dropped by 11% in Jer sey cows. Also, the milk yield dropped by 4% to 5% in Jersey cows, when THI values were between 21 and 50, and a loss of 11% when the THI values reached 16 to 20. Our study re sults confirmed that the LCT and upper critical temperature affected the milk yield in Holstein and Jersey cows, particu larly, Jersey cows experienced more cold stress than Holstein cows during each season in Korea.
The milk fat and protein content increased in Holstein cows when the THI value was 21 to 30 compared to Jersey cows (31 to 35 for milk fat and 26 to 30 for milk protein), and fat and protein content declined in Jersey cows more than in Holstein dairy cows, when the THI value was 76 to 80 and 46 to 50 respectively. Rodriquez et al [22] reported that HS environments have been associated with depres sions in milk fat percentage. Bouraoui et al [23] suggested that the depressed fat percentage could be attributed to the decrease in forage intake of the diet, and TMR could alleviate milk fat depression associated with HS by maintaining the intended forage to concentrate intake and, ensuring adequate fiber for proper rumen function. Many studies reported that a decreased milk protein is associated with increased maxi mum daily temperature [10,13]. The reduction in milk protein THI, temperature-humidity index; SEM, standard error of the mean. a-f Denotes comparison made within the same column (comparison between THI and Holstein and Jersey cows, respectively) (p < 0.05).
x,y Denotes comparisons made with same row (comparison between the Holstein and Jersey cows in the same THI) (p < 0.05).