Introduction

The primary environmental factors that produce heat stress in the dairy cow are temperature and humidity (Hahn et al. 2003; West 1995; Armstrong 1994). The effect of hot humid weather represents high costs to dairy farmers. There is not only a decrease in milk yields during hot weather (Valtorta et al. 1997; Valtorta and Gallardo 2004; West 2003)—milk also has a lower solids content, and cows show impaired reproductive performance and greater susceptibility to ill-health (West 1995). Two important aspects related to hot weather management for lactating dairy cows are: (1) management of the cow environment, and (2) feeding management.

Effects of weather conditions on feed intake and milk production are mostly mediated by changes in core body temperature (NRC 1981; West 1999), which could be affected by management practices.

Under grazing conditions, there is no doubt about the efficacy of shade in alleviating heat stress during the summer months (Davison et al. 1988; Valtorta et al. 1996, 1997).

In hot and humid climates, such as those in Georgia (West 1999) and the Argentine central milk supply area (Valtorta and Gallardo 2004), additional benefits from the combination of sprinklers and fans may be achieved.

During hot weather, energy from nutrient intake is critical because of the decline in feed intake that occurs. Energy content is one the most limiting factors in dairy diets, especially during high production and heat stress. The diet must be more energy rich to provide sufficient energy to maintain milk yield (West 1995). One way of increasing energy in the diet is to increase concentrates and to decrease forage, thus reducing the forage:concentrate ratio.

Since both environmental and nutritional factors could affect dairy cow performance, a trial was designed to evaluate the combined effects of diet and cooling with sprinklers and fans in the holding pen before milking in a grazing system. The responses in terms of rectal temperature (RT), respiration rate (RR) and milk production and composition were analysed.

Materials and methods

Location and period

The trial was carried out at the Experimental Dairy Unit, Rafaela Agricultural Experimental Station (INTA), 31°11′ S, from 12 January to 3 March 2003.

Animals and treatments

Fifty-eight Holstein cows in mid lactation were randomly assigned to four treatments, consisting of a combination of two diets: control (CD) and balanced (BD) with two levels of environmental control before milking: sprinklers and fans (SF) and no sprinklers or fans (NSF). Overall average body weight was 560.19±53.3 kg/cow. Average days in milk and milk production were 132.09±15.58 days and 26.2±3.9 l/cow per day, respectively. The average number of lactations was 2.53±2.03.

Management and feeding

All animals grazed an alfalfa pasture, in strips assigned to each group daily.

From 1000 hours until afternoon milking all cows stayed in a pen adjacent to the milking parlour where natural tree shade was available. There, water was offered ad libitum to all cows. Animals in the BD treatment also received alfalfa hay.

During milking, at 0500 and 1700 hours, concentrate mixes were offered in a constant amount to the cows of each group. Half of the total mix was fed at each milking.

In order to obtain different forage:concentrate ratios in both groups (nearly 80:20 in CD and 70:30 in BD; Table 1) a restricted strip grazing area was offered to cows in the BD treatment (on average 70 m2/cow per day) vs. CD (on average 100 m2/cow per day). The CD diet was formulated according to regional feeding practices, while the BD diet was calculated to obtain a better protein and energy equilibrium. In both cases the CPM-Dairy (1998) was used to determine the formulation. Concentrate mix for CD cows consisted of 30% sorghum grain, 30% corn grain, 20% wheat grain and 20% wheat bran. The BD mix, on the other hand, contained 36% corn germ, 26% wheat grain, 26% corn grain and 12% soybean seed. The composition of CD and BD is shown in Table 1.

Table 1 Composition of diets fed to cows for the control diet (CD) and balanced diet (BD)
Full size table

Animals in SF received a combination of sprinklers and fans, for 20 min before the morning milking and 30 min before the afternoon milking, in the holding pen. Fans were 1 HP, 1.2 m in diameter and were placed with a 30° tilt. Air speed was about 3 m/s. The sprinklers presented a 30 l/h flow rate. Since the cooling period was 30 min, sprinklers and fans were working continuously, because no differences in the effects of this system vs. the sprinkling–ventilation cycles were observed during cooling periods <1 h (Davison et al. 1996).

Measurements

Milk production was recorded daily, during both morning and afternoon milking, with an automatic Afimilk system (Afikim, Israel). Milk composition was determined twice a week on individual morning and afternoon samples, which were analysed for milk fat, protein, lactose, urea and non-fat solids. All values were obtained with a Milkoscan autoanalyser (Foss Electric, Hillerød, Denmark).

RT and RR were measured once a week for all animals before they entered the holding pen and immediately after they left the milking parlour, in the afternoon. RT was recorded by means of clinical veterinary thermometers, and RR by flank movements.

Body condition score (BCS) was determined at the beginning and at the end of the trial period using the 1–5 points scale described by Edmonson et al. (1989).

Individual dry matter (DM) intake of concentrate was measured once a week, by weighing the amounts offered and rejected.

On pasture, forage availability was determined whenever a phenological stage change was observed, by hand-plucking (Meijs et al. 1982). Six 0.25-m2 samples were obtained per strip.

Individual pasture consumption was recorded monthly. In order to estimate it, pairs of animals were allocated to individual pens. Forage availability in these pens was determined the day before the consumption measurement. The day after, the remaining pasture was measured. Consumption was estimated by the difference between offered and refused pasture (Gallardo et al. 2005).

Feed quality of all feed components (pasture and concentrate) was analysed every 2 weeks. DM, crude protein (CP), neutral detergent fibre (NDF), acid detergent fibre, lignin, ether extract and ashes were determined (Association of Official Analytical Chemists 1990).

Meteorological data

Air temperature and relative humidity data were obtained from a meteorological station located about 500 m from the experimental dairy farm. Average daily temperature humidity index (THI) was calculated after Armstrong (1994).

Experimental design

Productive and physiological data were analysed by GLM (SAS 1989), using the following two-factor model:

$$Yijk = \mu + Di + Rj + DRij + Covk + Eijk $$

where μ=overall mean, Di=diet effect, Rj=cooling effect, DRij=interaction between Di and Rj, Covk=covariate effect, and Eijk=random error.

Individual milk production 2 weeks prior to the beginning of the trial was the covariate for the productive data.

The analysis of DM intake also included an effect of month, in a similar model.

Differences were considered to be statistically significant when P<0.10, since the trial was performed under grazing conditions (Kolver and Muller 1998).

Results

Table 2 shows estimated pasture daily allowance for animals receiving CD and BD. There were differences in allowance between months within the trial, therefore, the data for January and February are presented. In February, the allowance was about 13% higher for CD, and about 18% higher for BD.

Table 2 Characteristics of the grazed alfalfa, and pasture allowance during the experiment, according to month, for CD and BD. DM Dry matter; for other abbreviations, see Table 1
Full size table

The chemical composition of pasture offered to all cows is shown in Table 3. No large differences were observed between months. However, there was a trend for the February pasture to have a somewhat better quality.

Table 3 Chemical composition of pasture fed to all cows in January and February.CP Crude protein, NDF neutral detergent fibre, ADF acid detergent fibre, NEL net energy of lactation; for other abbreviations, see Tables 1 and 2
Full size table

Table 4 presents the chemical composition of the concentrate mix fed to each diet treatment (CD and BD). The high protein of BD concentrate mix was due to the inclusion of corn germ.

Table 4 Chemical composition of concentrate mix fed to CD and BD cows. For abbreviations, see Tables 1, 2 and 3
Full size table

The high concentration of CP in the total diet, nearly 19% in both treatments, was due to the inclusion of pure fresh alfalfa pasture as a high proportion of the diet. In many cases fresh alfalfa CP was >20%, the maximum being close to 25%.

As stated previously, the CD was formulated following the common feeding practices in the area, i.e. high forage diets (Table 1). The final forage:concentrate ratios were 78:22 for CD and 65:35 for BD.

Pasture and total DM intake also differed between months within the trial (Table 5). Pasture intake was 16% higher for CD and 40% for BD in February, as compared to January. Total diet intake was affected by both month and diet. The interaction between month and diet was significant (P<0.07). The highest intake was for the BD in February, while the remaining interactions were not significant.

Table 5 Effects of CD and BD diets in combination without (NSF) and with cooling with sprinklers and fans (SF) on pasture and total dry matter intake (kg/cow per day) during January and February. For other abbreviations, see Table 1
Full size table

On the basis of the quality of ration components and DM intake, the overall energy concentration of the total diet was 1.48 Mcal net energy of lactation NEL/kg DM for CD and 1.60 Mcal of NEL/kg DM for BD, respectively, as calculated according to NRC (2001). The metabolizable protein balances were +90 g/cow/day for CD and +105 g/cow/day for BD.

Table 6 shows overall mean RT and RR values before the afternoon cooling period, for all treatments. Responses to sprinkling and fans are also shown.

Table 6 Effects of CD and BD in combination with NSF and SF on rectal temperature (RT) and respiration rate (RR) before afternoon milking (mean±SD), and mean variation (δ) of RT and RR (after–before afternoon milking). For other abbreviations, see Tables 1 and 5
Full size table

Average air temperature, relative humidity and THI during the whole experimental period were 25.4±3.9°C, 71.5±11.2% and 74.4±5.4, respectively, as computed from mean daily data. However, meteorological conditions were quite different for each week. Table 7 shows maximum temperature and mean THI recorded the day when physiological parameters were measured. The values for the day before are also presented. Since the work was performed under field grazing conditions, there is a logical time variation. Extreme temperatures were 12.6 and 37.7°C.

Table 7 Maximum temperature and mean temperature humidity index (THI) recorded during the day before (Day−1) and the day when RT and RR were measured (Day 0). For other abbreviations, see Table 5
Full size table

Since there were differences between weeks, the time effects on RT responses are presented in Fig. 1.

Fig. 1
figure 1

Week effects on variation of rectal temperature (RT), values after and before cooling. Treatments are combinations of control diet (CD) and balanced diet (BD) without (NSF) or with cooling with sprinklers and fans (SF) for 30 min in the holding pen before afternoon milking

Full size image

Milk production and composition are shown in Table 8, in which change in BCS is also presented.

Table 8 Effects of CD and BD diets in combination with NSF or SF on cows’ performance. MUN Milk urea nitrogen, BCS body condition score; for other abbreviations, see Tables 1 and 5
Full size table

There was a significant interaction between diet and cooling. Therefore, given the different physiological responses to SF among weeks (Fig. 1), milk fat concentrations per week are shown in Fig. 2, for all treatments.

Fig. 2
figure 2

Weekly effects on milk fat content for animals on CD and BD in combination with NSF and SF for 20–30 min in the holding pen before milking. For abbreviations, see Fig. 1

Full size image

Discussion

Under grazing conditions, pasture availability and intake are highly environment dependant. It is well recognized that they vary significantly between seasons. However, there are also variations in the same season (Tables 2 and 5).

The differences in DM intake observed between diets could be due to more comfortable environmental conditions in February and also to the higher pasture allowance and quality (Tables 2, 3). During January the pasture intake measurement was carried out on a hot day (THI >74). During the February determination, weather conditions were milder (THI <72). It is well recognized that feed intake decreases under heat stress (West 2003). It is also well demonstrated that a higher pasture allowance leads to higher levels of consumption (Hodgson 1990), as reflected in the results obtained in February, as compared to January.

The objective of balancing the diets to improve productivity was achieved by controlling pasture allowance and varying the forage:concentrate ratio, thus obtaining a better dietary energy, protein and fibre equilibrium (NRC 1981; West 1995, 2003). In terms of metabolizable protein balance, the supply of amino acids was higher than the requirements according to CPM Dairy. In the CD the lysine balance was +12.2 g/cow per day and methionine balance 0.1 g/cow per day, the respective values for BD being 23.1 and 5.7 g/cow per day. The excess was mainly due to the inclusion of alfalfa pasture as a high proportion of both diets, especially in CD. It is very difficult to balance the nutrient concentration in the diet, when alfalfa pasture represents a high proportion of the total ration (Castillo et al. 1998).

With respect to supplying supplements in grazing systems, Valtorta et al. (1996) showed that the inclusion of corn grain in the diet of cows with access to artificial shade lead to an improvement in their productivity.

RT and RR decreased (P<0.05) in response to evaporative cooling with sprinklers and fans before milking (Table 6). These results agree with several trials conducted with heat-stressed cows (Armstrong 1994; Chen et al. 1993; Tarazón-Herrera et al. 1999). The proposed management only included cooling before milking, since this is compatible with farm practice. Between morning and afternoon milking, cows only received shade, which does not achieve complete comfort. Therefore, animals in all treatments had RT >39°C before afternoon milking.

Environmental conditions differed during the trial (Table 7). However, most of the time THI was above the threshold value of 72 (Armstrong 1994). The responses in RT and RR were generally higher during the weeks where the cows showed higher stress, as assessed by the mean THI recorded when physiological parameters were measured (Fig. 1). These observations suggest that the use of a sprinkler and fan system prior to milking under grazing conditions could be effective when animals are cooled for only 50 min/day (Valtorta and Gallardo 2004). This response would be more evident when balanced diets were offered.

Milk production and milk protein concentration (Table 8) were higher in the BD treatments (P<0.005 and P<0.01, respectively). The increase in milk yield could probably be due to the higher density of the diet, 1.60 versus 1.48 Mcal NEL/kg DM, which would provide sufficient energy to increase milk yield under heat stress (NRC 1981; West 1995). Similar responses were observed by Drackley et al. (2003) when cows in mid-lactation received a diet containing 1.60 Mcal NEL/kg DM as compared to the controls fed a 1.52 Mcal NEL/kg DM diet, during summer. There were no treatment effects on BCS (Table 8). These results would indicate that both factors acted in such a way that energy was more efficiently diverted to milk production.

No significant diet effects were observed on MUN. However, there was a effect of cooling on this parameter (Table 8). It is probable, that cooling the cows lowered the extra energy demand to eliminate body heat, thus leaving more energy for milk production (NRC 2001). Also, balancing the diet by decreasing the forage:concentrate ratio could lead to the availability of more energy for microbial protein synthesis, which would in turn result in an increase in milk protein content. It is possible that there was higher utilization of ammonia in the rumen, also considering the higher protein intake in BD (Bargo et al. 2003). On the other hand, there could be a lower utilization of amino acids as an energy source in SF.

There was an interaction effect between diet and cooling on fat content (Table 8). Cooling with sprinklers and fans seemed to have a stronger effect on milk fat content during weeks with higher thermal stress (Fig. 2), though this was not statistically evaluated. A trial carried out during a heat wave event in the area where this study was performed showed a decrease in milk components in response to the lack of a cooler recovery period (Valtorta et al. 1997). Thus, the heat-stress relief in this study during the hotter weeks could have determined the response in terms of fat content. Cooling effects seem to be diluted during cooler periods. This differential response could have determined the interaction between both experimental factors.

These results show that under grazing conditions the effects on milk production and composition are enhanced when the diet is specially formulated for hot periods, by controlling pasture allowance to decrease the forage:concentrate ratio. The alleviation of heat stress before milking also affects animal performance and, in some cases, there is an interaction between diet and cooling. Although the cooling periods were not very long, they were long enough to have some beneficial effects on the physiological status of the cows, as determined by RT and RR.