Predicted versus measured production differences using summer air conditioning for lactating dairy cows.

Predicted summer production losses were made for Holstein cows in three widely separated locations in the United States where field investigations of air conditioning for dairy barns have been completed. These losses, based on a functional relationship developed from Missouri Climatic Laboratory data and local climatological data, were compared to the measured losses at each location. Correlation of losses for Holstein cows varied from excellent in northern Ohio to good in southern Louisiana and central Missouri. Variations due to hotteror cooler-than-normal summer seasons and level of production were reasonable at all locations. Establishment of the strong link between laboratory production data and commercial herd production enables projection of expected production losses for cows in any area having adequate climatological records. Predicted losses, in turn, allow economic evaluation of environmental control or modification alternatives. Investigations of environmental requirements for livestock have for many years been centered on establishing the adverse effects of climatic extremes on production and other physiologic functions. For dairy cattle, laboratory research has provided a functional relationship between production and suitable climatic parameters. Extension of the laboratory results to field situations has been slow to develop, primarily due to the large capital outlay for air-conditioning equipment and the continuing costs incurred to operate the equipment. Studies in Missouri, Louisiana, and Ohio have provided the only known applications of environmental control to herd situations (3, 5, 6). However, interest in applications to commercial herds has been exhibited by producers in such diverse tropical and subtropical areas as Puerto Rico, Guam, Republic of Ivory Coast, Hong Kong, and Mexico, in addition to southern and western areas of the United States. Recent developments in combining laboratoryestablished relationships with probabilities of occurrence of climatic parameters have permitted prediction of seasonal production losses for livestock. For lactating Holstein dairy cows, milk production decline has been related to the Temperature-Humidity Index (THI) 1 by Berry et al. (I) : MDee ~ 1.075 -1.736 N L + .02474(NL) ( T H 1 ) where MDec is absolute decline in milk production, kilograms/day-cow, N L is normal level of production, kilograms/day, T H I is daily mean Temperature-Humidity Index value. Hahn and MeQuigg (2) developed the necessary methodology whereby summer milk production losses for June 1 through September 30 could be predicted using this equation and empirical probabilities of occurrence of T H I for Columbia, Missouri. These predictions have now been completed for 16 stations in the United States (4) ; Figure 1 shows a map with isolines of summer season production losses for cows of 13.5 and 23.0 kg/day normal levels of production. To check the validity of the predictions, comparisons were first made between predicted and measured production losses for Holstein cows (16 cows per gToup; 18.1 to 22.7 kg/cow-d~y 1 Temperature-Humidity Index is a derived statistic computed from the relation, T H I = t ~ I .36 t~ + 41.2 where t~ is dry-bu]b temperature, C, and t~p is dew-point temperature, C. ~. DAIEY SOIENC ~, ]rOL. ~ , NO. 6 v. 52, no. 6 (June 1969) TECHNICAL NOTES 8 0 1 normal level of production) provided summer air conditioning from late June until early September in the Columbia, Missouri area (3). The measured net increase in production over cows confined to a dry-lot with an open-front shed during the switchback experiment was 0.58 kg/cow-day in 1964 when the average ambient temperature was 0.83 C above normal for the period and 0.44 kg/eow-day ii1 1965 when the average temperature was 1.72 C below normal. On a 122-day basis, summer production losses totaled 71 kg/cow in 1964 and 53 kg/eow in 1965. Predicted production losses, based on a normal season of the same length, were 60 kg/cow for 18.1 kg/cow-day NL and 89 kg/cow for 22.7 kg/eow-day NL. The reasonable correlation of predicted and measured production losses for cows in midMissouri gave incentive to compare observations made in the two other field investigations with predictions based on laboratory research, particularly since the field locations are widely separated. In the Louisiana study (5), three groups of Holstein cows (ten cows per group, with about 11-14 kg/cow NL) were respectively placed in an air-conditioned chamber, an open shed, and on permanent pasture with adequate tree shade from May through mid-August near Baton Rouge. Measured mean daily milk produetion was 1.1 kg/eow-day higher for the airconditioned group when compared to the pasture with shade group during the hotter-thannormal season (average ambient temperature was 0.83 C above normal for the period). Extending the results of the 112-day experiment to the standard 122-day season used in Figure 1 gives a smnmer production loss of 138 kg/eow; from Figure 1 for 13.5 kg/eow NL, the predicted losses are 127 kg/eow. Again, predicted values correlate reasonably well with measured production losses. The Ohio study (6) was conducted during the summers of 1962 through 1964 on the farms of four private cooperators, each having a different breed of cows. The air-conditioned cows were compared to cows exposed to the naturally varying environment for the duration of the experiments (about June i through September 30 each year). Results from the 1962 and 1963 test seasons were not presented in a form that permitted comparison with predicted values. Presentation of the 1964 results obtained during an approximately normal summer in Ohio did permit comparison; using Farm 3 in Wayne County (northern Ohio) with Holstein cows having a normal production level of about 20 kg/cow, the difference between air-conditioned and outside (control) cow groups was 22 kg/eow, over the standard 122day season. Predicted losses are in excellent agreement, being nearly 23 kg/cow for this level of production (Fig. 1). For Farms 1, 2, and 4, with Brown Swiss, Ayrshire, and Guernsey herds, respectively, the agreement between predicted and actual production for the period was not as good. The obvious possibility is that the functional relationship for Holstein cows may not be applicable to other breeds, but other factors may also have been involved. These correlations between observed and predicted milk production losses for Holstein cows o£ varying production levels, located in widely spaced locations, and subjected to differing management systems, have established a strong ]ink between production obtained under laboratory conditions and that obtained under commercial herd situations involving environmental control. This link permits projection of expected production losses for cows in any area having adequate climato]ogieal records, thereby allowing estimation of the economic feasibility of environmental control or