ASAE Conference Proceeding
This is not a peer-reviewed article.
Thermal Environmental Effects on Feed Intake in Commercial Dairy Herds
D.M. Allen, J.G. Linn, and K.A. Janni
Pp. 205-212 in Fifth International Dairy Housing Proceedings of the 29-31 January 2003 Conference, (Fort Worth, Texas, USA), ed. K. A. Janni. ,Pub. date 29 January 2003 . ASAE Pub #701P0203
Most research on dry matter intake (DMI) of lactating dairy cattle has been conducted in the thermal neutral zone, between 5 and 20 degrees C, or under heat stress conditions. Generally studies have found that DMI decreases as the environmental temperature and humidity increases above the thermal neutral zone. Very little information is available on the change in lactating cow DMI in environmental conditions below the thermal neutral zone. The purpose of this study was to measure changes in DMI of lactating cows on three commercial dairy farms in Minnesota and Wisconsin over 12 continuous months. Data collected monthly from an average of 1485 lactating Holstein cows included group average milk production, days in milk (DIM), monthly body weights, DMI and feed refusals. Averaged temperatures and relative humidity levels in the naturally ventilated freestall housing were recorded hourly over the 12-month study period. No significant seasonal effects on DMI of lactating cows were observed on the three farms in this study. Small decreases in DMI during months when the maximum outside temperature exceeded 20° C were observed. No consistent effect on increasing DMI when maximum temperatures remained below the thermal neutral zone temperatures was observed. In the absence of extreme or prolonged humid conditions or temperatures outside the thermal neutral zone, the primary factors affecting DMI of lactating cows on these three farms were milk production and cow characteristics.KEYWORDS. Dairy, Dairy housing, Dry matter intake, Thermal neutral zone
The effect of environmental temperature and humidity on dry matter intake (DMI) and milk production of lactating dairy cows has only been studied to a limited extent (NRC, 1987; NRC, 2001). Most previous research has focused on the effects of heat stress or environmental conditions above the thermal neutral zone of 5 to 20° C. McGuire et al. (1989) measured DMI and milk yield of Holstein cows in a thermal comfort zone of 19 to 25° C and a thermal stress zone of 19 to 40° C. Cows in the thermal comfort zone ate 25% more DM (15.1 vs 11.1 kg/day) and produced about 3 kg/day more milk (19 vs 16.2 kg/day) than cows in the thermal comfort zone. Restricting DMI of cows in the thermal comfort zone to 75% of ad libitum DMI mimicked the effects of thermal stress.
West et al. (1999) studied the interaction of environmental conditions and four concentrations of neutral detergent fiber (NDF) in the diet on DMI of lactating Holstein cows. They found as the environmental temperature and humidity increased and the NDF concentration of the diet increased, DMI of cows decreased linearly. The decrease in DMI was greatest (22%) for cows fed the low NDF (30% of the DM) and least (14%) for cow fed the highest NDF diet (42% of DM) as the environment changed from less than 72 temperature humidity index (THI) to over 78 THI. Holter et al., (1997) reported a similar decrease (22%) in DMI of mid to late pregnant multiparous Holstein cows subjected to heat stress conditions.
Very limited information is available on the change in DMI when cows are in environmental conditions below the thermal neutral zone. In the 2001 Nutrient Requirements of Dairy Cattle (NRC, 2001) temperatures below the thermal neutral zone are suggested to alter nutrient metabolism and increase maintenance requirements.
Previous research on how environmental conditions affect DMI of lactating dairy cows has generally been short term studies and in research facilities. The objective of this research was to study changes in DMI of lactating dairy cows on commercial dairy farms in Minnesota and Wisconsin over 12 continuous months.
Materials and Methods
Daily DMI intake from 16 pens representing a monthly average of 1485 lactating Holstein cows was conducted on three commercial dairy farms: one in southern Minnesota (Farm B) and two in western Wisconsin (Farm A and Farm C). Farm descriptions follow. Data was collected for 12 consecutive months at each location beginning in November 1999 for Farms A and C and January 2000 for Farm B. Additional information is available in Allen (2001).
Data Collection - All Farms
Daily. Group average milk production, DIM, lactation number, and cows per pen were electronically captured daily on each farm using Dairy COMP305® (Valley Agricultural Software, Tulare, CA). All daily farm production data was downloaded monthly to obtain complete 365 day records for each pen on each farm except on Farm C; the hard drive on the computer storing production data was erased resulting in a loss of milk production data from April 25, 2000 to June 1, 2000. Bulk tank milk fat and milk true protein percent was determined daily on milk shipped from each farm by Land O’ Lakes (Arden Hills, Minnesota).
The average DM amount fed per head per day for each pen on all of the farms was recorded using EZ Feed® software (Valley Agricultural Software, Tulare, CA). Pen feed refusals were suppose to be weighed back daily on all farms, however, this did not occur. Actual days with recorded weigh back weights varied by farm; however, all farms recorded weigh back weights more than 50% of the time. An average weigh back percentage was calculated on a monthly basis for each farm based on their recorded weigh back weights. Weigh back was assumed to have the same DM concentration as the diet fed to the pen. Dry matter intake per day was calculated from the DM fed per pen minus the calculated refusal DM per pen. EZ Feed® software was broken February 18, 2000 to March 4, 2000 on Farm B so actual DMI was not determined during those days.
Indoor dry-bulb temperature and relative humidity data were collected in one barn at each site. Dry-bulb temperatures were measured using a type-T thermocouple junction in each barn with an assumed uncertainty of ± 1° C. Relative humidity was measured using an integrated circuit humidity sensor (IH-3602, Hy-Cal Sensing Products) with a manufacturers interchangeability of ± 5% up to 60% relative humidity and ± 8% at 90% relative humidity. The indoor temperature and relative humidity sensors were measured with a single thermocouple and one humidity sensor placed at one location in the barn in the middle of face-to-face freestalls at cow height, approximately 15 cm above the top rail of the stall. To protect the sensors they were placed inside an 8 cm diameter PVC pipe, open at the top and bottom for free airflow. The outside thermocouple and humidity sensors were located under either the north or east roof overhang in a short piece of PVC pipe to minimize solar radiation effects. Readings were taken every 5 minutes and averaged over an hour. Data at each barn were collected using a data logger (21X Micrologger, Campbell Scientific Inc.) and multiplexer (AM416, Campbell Scientific Inc.) connected to a cellular phone through a modem. The hourly stored data were downloaded onto a personal computer through a modem (Janni and Allen, 2001).
Monthly. Body weight was determined on 30 to 35% of the cows in each pen every month. No attempt was made to measure the same cows monthly. A random sample of cows within the pen were measured either in headlocks using a heart girth tape (Farm A and Farm C) or weighed with an electronic scale exiting the parlor (Farm B). All cows were weighed at a similar time of day and month by location.
Statistical analysis. Correlation statistics on DMI, temperature and humidity were preformed using SAS® (SAS® Inst., 1996). Data were compiled and averaged over 3-month periods for analysis. Farm was the experimental unit and only one temperature-humidity recording instrument was used per farm. Therefore, there were no experimental unit replications for standard analysis of variance statistical methodology.
Farm A - Description
Animals. Eight hundred lactating Holstein cows were cared for according to standard operating procedures on the farm. All cows were milked three times per day and bovine somatotropin was administered to cows according to farm protocols. All cows were grouped according to stage of lactation and reproductive status within parity. Only four early lactation pens (average 97.8 DIM) are represented in the data analysis, as these were the only pens with headlocks for body weight measurements. Parity of the 353 cows used for DMI data was 56% multiparous and 44% primiparous.
Housing. Cows were housed in two six-row, curtain-sided freestall barns with an open ridge running east and west. The milking parlor was attached and located south of the barn. Mattresses were used in freestalls and bedded with sawdust. Data was collected from only Barn 1 that contained 400 freestalls divided into five pens with headlocks along the feed bunk. Each quadrant of the barn contained two water troughs in the walkway. The summer cooling system consisted of sprinklers over the feed bunk and in the holding pen with 122 cm high-speed axial fans placed over the freestalls.
Feeding System and Diets. Cows were fed a total mixed ration (TMR) once daily using a horizontal Reel Auggie ® mixer manufactured by Knight Manufacturing (Brodhead, WI). The bunk management goal was to feed for an empty bunk at the same time every day and incorporate six scheduled feed push-up times per day. Average nutrient composition of diets fed during the study was 53 ± 3.7% DM, 18.5 ± 0.4% crude protein (CP), 30 ± 0.7% neutral detergent fiber (NDF) and 38.5 ± 1.5% nonfiber carbohydrates (NFC). Diets averaged 53% forage, 47% concentrate on a DM basis with the forage portion consisting of 25% corn silage, 23% alfalfa haylage and 5% alfalfa hay.
Farm B - Description
Animals. Six hundred Holstein cows were managed and cared for according to standard operating procedures implemented by the farm. Cows were milked three times daily and bovine somatotropin was administered to cows greater than 90 DIM according to farm protocols. Data was collected from four production groups: 1) Fresh cow group averaging <15 DIM, 2) First lactation cows, 3) Early lactation second and greater lactation cows 4) Mid-lactation to late lactation cows of mixed parity. Cows from Farm B averaged 144.3 DIM and produced 38.8 kg of milk/d.
Housing. Cows were housed in three curtain-sided freestall barns. Barn 1 was a three-row freestall barn with drive-by feeding along an open rail. The barn was oriented north and south and contained 100 sand bedded stalls. Barn 2 was oriented east and west and contained four rows of sand bedded freestalls without a bunk. Barn 3 was a three-row freestall barn with sand bedded stalls and oriented north and south. Cows in Barn 2 and Barn 3 were fed out of J-bunks in Barn 3. All cows had access to a minimum of three water troughs. The summer cooling system in all barns included 122 cm high-speed axial fans placed along the feed bunk, and in the holding pen. Sprinklers were placed along the feed bunk and in the holding pen.
Feeding System and Diets. All cows were fed a TMR twice daily using a Reel Auggie ® feed mixer manufactured by Knight Manufacturing (Brodhead, WI). Six push-ups were scheduled for barn 1 with no push-ups in barn 2 or 3 because of feeding in a J-bunk. Feeding strategy was to feed for less than 3% refusals or empty bunk at the initial AM feeding. Therefore, the order in which pens were fed could differ daily. Average nutrient composition of diets fed during the study was 52 ± 2.9% DM, 18.5 ± 0.3% CP, 29.4 ± 1.1% NDF and 37.4 ± 1.4% NFC. Diets contained an average of 28% alfalfa haylage, 22% corn silage, 1% alfalfa hay and 49% concentrate, DM basis.
Farm C - Description
Animals . Six hundred lactating Holstein cows were utilized in the study and all cows were managed and cared for according to farm protocols. Cows were milked three times daily and bovine somatropin was administered according to farm protocols. Data was collected from eight lactation pens on Farm C. Cows were then divided into groups based on parity and reproductive status. One of the eight pens consistently contained primiparous cows. Farm C was going through an expansion period during the last five months of the data collection period. Therefore, due to cow purchases, some pen characteristics did not remain constant from month to month. Cows averaged 143 DIM and produced 36.3 kg of milk/d.
Housing. Cows were housed in two, four-row freestall barns oriented east-west. Headlocks were located along the feed bunk in each barn. The milking parlor was south of both barns. Barn 1 contained 240 freestalls with mattresses bedded twice weekly using sawdust or oat hulls. Barn 2 was the northern most barn. A flush system was used in each barn. Two water troughs per pen were placed on either end of the freestalls. The summer cooling system in each barn consisted of sprinklers over the feed bunk and 122 cm high-speed axial fans in the holding pen and over the freestalls.
Feeding System and Diets. All cows were fed a TMR once daily using a Reel Auggie ® feed mixer manufactured by Knight Manufacturer (Brodhead, WI). The feed bunk was managed for a target of 3% or less feed refusals with the daily TMR fed at a similar time each day. Feed refusal amounts were weighed and recorded daily. Average nutrient composition of diets fed during the study was 49.8 ± 3.0% DM, 18.3 ± 0.5% CP, 29.7 ± 1.2% NDF, 38.5 ± 1.0% NFC Diet DM consisted of corn silage (27%), alfalfa haylage (21%) and alfalfa hay (4%) and concentrate (48%). of the diet DM.
Results and Discussion
The average number of cows, DMI, milk production, days in milk and lactation number for each of the three farms during the 12-month data collection period is in Table 1. The average DMI of lactating cows across the three farms was very similar during the collection period averaging 23 ± 1 kg/head/day. Farm A had the highest milk production per cow (39.9 kg/day), lowest DMI average (22.7 kg/day) and lowest days in milk (98 DIM) of the three farms. This is because only four early lactation pens were used for data collection on this farm compared to the other two farms where DMI of cows in all stages of lactation was collected. Farm A also had the lowest average lactation number and therefore, more first lactation animals were included in the data set as a percentage of total cows compared to Farms B and C.
Table 1. Average farm data from November 1999 to January 2001.
Average number of cows
DMI , kg/head/day
Days in milk
22.7 ± 3.91
39.9 ± 6.1
98 ± 28
23.3 ± 4.9
38.8 ± 9.1
145 ± 106
23.7 ± 3.6
36.4 ± 8.3
146 ± 91
1 Standard deviation
Outside minimum and maximum temperatures by farm are in Figure 1. All farms had similar average monthly minimum and maximum temperatures, as linear distance between the three farms was less than 120 km. Maximum temperature exceeded the upper thermal neutral zone (20° C) for cattle during the months of June, July, August and September. During the hottest week (July 7-14) the average indoor stall temperatures were 23.6, 25.1, and 24.5° C for farms A, B, and C, respectively. Hourly average stall temperatures ranged from 17 to 32 C in the three barns during the warmest week (Janni and Allen, 2001). This indicated that heat stress during the study was minimal as nighttime temperatures (minimum temperatures) always dropped below 20° C. In addition, THI values during the warmest week were less than 70 for 58, 28, and 41 hours for farms A, B, and C, respectively. THI values exceeded 78 for 19, 21 and 16 hours respectively (Janni and Allen, 2001). Measured outside temperatures at all three farms were near the historical average for July (Janni and Allen, 2001).
Cows were exposed to several months of cold stress (temperatures below 4° C) even though January was warmer than historical average (Janni and Allen, 2001). During the coldest week (January 21-28) the average indoor stall temperatures were -6.1, -9.2, and -6.0 C for farms A, B, and C, respectively. Hourly average stall temperatures ranged from -17 to 4 C during the coldest week (Janni and Allen, 2001).
Figure 1. Outside minimum and maximum temperatures for the three farms over the 12-month study period.
Average milk production of pens measured on the three farms over the 12-month period are in Figure 2. Farm A had a gradual rise in milk production (early lactation pen measurements only) over the 12-month period, whereas, both Farm B and C (all stages of lactation pens) had a gradual decline in milk production. All farms tended to slightly decrease in milk production starting in July and continuing through September. This decrease cannot be attributed to increasing DIM as farms either held constant or decreased DIM over these months. The continual decline in milk production for Farm C after September can be partially attributed to their expansion startup at this time. The moderate thermal stress that occurred during July through September (Figure 1) had a minimal impact on milk production.
Dry matter intakes of lactating cows over the 12 months are shown in Figure 3. After the first three months of the study, DMI remained relatively constant (23 ± 1 kg/head/day) across all farms. Decreases in DMI during June through September when maximum temperatures averaged greater than 20° C were less than expected and less than what others have reported during heat stress periods (Eastridge et al., 1998; West et al., 1999). The heat abatement measures (sprinklers and fans) on these farms along with minimum temperatures decreasing into the thermal neutral zone probably minimized the day time heat stress affects on cows.
Temperatures below the thermal neutral zone did not appear to increase DMI as typically observed in practice on farms. This probably reflects the warm, mild conditions of the 1999/2000 winter and the lack of an extended day and night time period (more than 1 week) below -15 C.
Figure 2. Average daily milk production of lactating dairy cows on the three commercial dairy farms over the 12-month study period.
Figure 3. Average daily dry matter intake of lactating dairy cows on the three commercial dairy farms over the 12-month study period.
Correlations between DMI and minimum and maximum humidity and outside temperature within each farm are shown in Table 2. Dry matter intakes were grouped into seasons (January – March, April – June, July – September and October – December) for analysis and interpretation. On Farm A, maximum humidity had a high negative correlation while both minimum and maximum temperature had a high positive correlation with DMI for three of the four seasons. Only spring (April – June) was humidity and temperature not correlated with DMI. On Farm B, maximum humidity was positively correlated with DMI from summer, fall and winter, but negatively correlated with DMI during the spring. Temperature was highly correlated with DMI for all seasons except summer (July through September) on Farm B. On Farm C, DMI was only highly correlated with humidity during January through June. Minimum and maximum temperature were only highly correlated and then negatively with DMI during colder months of January through March and then again from October through December. Over all farms, maximum humidity appears to be more consistently correlated with DMI, although both negatively and positively, then either minimum or maximum temperature. The generalization of high temperatures lowering DMI and cold temperatures increasing DMI were not supported by the data collected on these three farms from November 1999 through January 2001.
Table 2. Correlation between dry matter intake of lactating dairy cows on three commercial dairy farms with minimum and maximum humidity and outside minimum and maximum temperature over a 12-month period.
January - March
April – June
July – September
October - December
January - March
April – June
July – September
October - December
January - March
April – June
July – September
October - December
* P<.01, **P<.05
The outdoor temperatures during the study were above normal in January and near normal in July, creating little thermal stress. Without prolonged thermal stressful conditions, no consistently significant seasonal effects on DMI of lactating cows were observed on the three farms. Small decreases in DMI during months when the maximum outside temperature exceeded 20° C were observed. No consistent effect on increasing DMI when maximum temperatures remained below the thermal neutral zone temperatures was observed. In the absence of extreme or prolonged humid conditions or temperatures outside the thermal neutral zone, environment was a tertiary factor affecting DMI of lactating cows behind milk production and cow characteristics on these three farms.
Allen, D. M. 2001. Dry matter intake and neutral detergent fiber intake of dairy cattle. Ph.D. Thesis. University of Minnesota, St. Paul, MN.
Eastridge, M. L., H. F. Bucholtz, A. L. Slater and C. S. Hall. 1998. Nutrient requirements for dairy cattle of the national research council versus some commonly used software. J. Dairy Sci. 81:3049-3062.
Holter, J.B., J.W. West and M.L. McGillard. 1997. Predicting ad libitum dry matter intake and yield of Holstein Cows. J. Dairy Sci . 80:2188-2199.
Janni, K. A. and D. M. Allen. 2001. Thermal environmental conditions in curtain sided naturally ventilated dairy freestall barns. In Livestock Environment VI, Proc. Sixth Int. Livestock Env. Sym . 367-376. R. Stowell, R. Bucklin, and R. W. Bottcher, eds. ASAE, St. Joseph, MI.
McGuire, M.A., D.K. Beede, M.A. DeLorenzo, C.J. Wilcox, G.B. Huntington, C.K. Reynolds and R.J. Collier. 1989. Effects of thermal stress and level of feed intake on portal plasma flows and net fluxes of metabolites in lactating Holstein cows. J. Anim. Sci . 67:1050-1060.
National Research Council. 1989. Nutrient Requirements of Dairy Cattle. 6th rev. ed Natl. Acad. Sci., Washington DC.
National Research Council. 2001. Nutrient Requirements of Dairy Cattle. 7th rev. ed Natl. Acad. Sci., Washington DC.
West, J.W., G.M. Hill, J.M. Fernandez, P Mandebvu and B.G. Mullinix. 1999. Effects of dietary fiber on intake, milk yield, and digestion by lactating dairy cows during cool or hot, humid weather. J. Dairy Sci . 82:2455-2465.