per cow. The strategies (1 to 7) were characterized by the criteria used to ... come over feed cost per cow for group- ... FCM (.4 x (milk yield) + 15 x (fat yield)).
Evaluation of Criteria Used to Group Lactating Cows Using a Dairy Production Model C. 8. WILLIAMS Agricultural Research Service, USDA Roman L Hruska US Meat Animal Research Center Clay Center, NE 68933 P. A. OLTENACU Department of Animal Science Cornell University Ithaca, NY 14853
ABSTRACT
INTRODUCTION
A dairy herd production simulation model was used to evaluate the effectiveness of Seven grouping strategies in terms of annual mmme over feed cost per cow. The strategies (1 to 7) were characterized by the criteria used to rank individual lactating cows for placement into feeding groups: 1, required nutrients (energy and protein) per kilogram of DMI, 2, required nutrients (energy and protein) per kilogram of NDF intake capacity; 3, DIM;4, test day milk; 5, test day FCM; 6, dairy merit (kilograms of FCM per kilogram of BW.75); and 7, dairy merit weighted by DIM. In terns of annual income over feed cost per cow, strategies 1 and 2 were found to be the most effective and strategy 4 the least effective, with average potential production levels of 8000,9000, and 10,000 kg of milk per 305d lactation, using two or three feeding groups. With the milk-feed price relationships used, mean annual income over feed cost per cow for grouping strategies 1 and 2 with three feeding groups was $21, $33, and $40 geater than with two feeding groups for potential production levels of 8000,9000, and 10,OOO kg of milk per 305-d lactation, respectively. (Key words: grouping criteria, computer model, dairy)
Feeding of TMR to groups of lactating cows is practiced widely, especially in large, high producing dairy herds (6, 7, 13). Several studies have evaluated the economics and milk production response with TMR (3. 8) or have discussed the concept, advantages, and disadvantages of TMR fed to different feeding groups of dairy cattle (6, 13). Evidence suggests higher income over feed costs (IOFC) with multiple feeding groups (12, 14). The extent to which IOFC is maximized will depend primarily on decisions concerning number of feeding groups and on criteria used to rank cows for placement into feeding groups (14). Several strategies have been used for placing lactating cows into feeding groups, but only McGilliard et al. (9) have reponed comparisons of some of these strategies. McGilliard et al. (9) developed a statistical clustering methodology (cluster method) that groups lactating cows simultaneously by megacalories of NEL required per kilogram of predicted DMI and CP required as a percentage of predicted DMI. One disadvantage of using predicted DMI is that, if ration characteristics such as crude fiber or NDF that influence fill are used in the prediction equation, DMI will decrease when cows are switched to a lower feeding group. This may result in higher protein and energy requirements per kilogram of predicted DMI of the lower quality ration. One way of avoiding this potential problem is to express the cow’s requirements on a per unit basis of a variable not associated with the ration. We propose that, for the purpose of placing cows into feeding groups, requirements can be expressed on a per kilogram of NDF intake capacity (NDFIC) basis because the
Abbreviation key: NDFIC = NDF intake capacity, IOFC = income over feed costs.
Received November 19. 1990. Accepted July 24, 1991. 1992 J Dairy Sci 75:155-160
155
156
WILLlAMS AND OLTENACU
(10) to predict DMI of lactating dairy cows. The model calculates daily potential and actual mi& production, total energy requirements, and DMI for each animal in the herd. The potential milk production was calculated with an incomplete gamma function, using parameters estimated by Congleton and Everett (4). Total energy requirements are obtained in the model by summing the separate requirements for maintenance, potential milk production, pregnancy, growth, and body reserve tissue repletion. The equations used to calculate these energy requirements are documented MATERIALS AND METHODS (16). The actual milk production achieved in Using different criteria to rank individual the model depends on nutrient intake, which, cows for placement into feeding groups, seven in turn, is partially dependent on nutrient regrouping strategies were defined and evaluat- quirements. ed. The following criteria were used to rank For strategies 1 and 2, the energy requireindividual cows in a herd. Strategy 1 required ments as calculated with the model were used nutrient concentration expressed as megacalo- to calculate the required energy concentration; ries of NEL required per kilogram of predicted the CP requirements for maintenance and proDMI and CP required as a percentage of pre- duction were calculated using the same equadicted DMI. Strategy 2 required nutrient con- tions as McGilliard et al. (9). The DMI in centration expressed as megacalories of NEL strategy 1 was predicted with the formula of required per kilogram of estimated NDmC and Brown et al. (2). and IL'WFIC in strategy 2 was CP required as a percentage of estimated estimated as a percentage of BW (16). Values NDHC. Strategy 3 used DIM. Strategy 4 used for NDFIC at specific stages within the calving test day milk yield. Strategy 5 used test day interval are given in Table 1. FCM (.4 x (milk yield) + 15 x (fat yield)). The model was run for 10 yr with a herd Strategy 6 used dairy merit (100(FCM)/ size of 100 cows to generate three base herds, (BW75)). Strategy 7 used dairy merit weighted representing three levels of potential 3 0 5 4 lactation milk production (8000, 9000, and by DIM (100(FCM)/(BW.75)) x Each grouping strategy was evaluated con- 10,000 kg of milk per cow). Each grouping sidering either two or three feeding groups for criterion and number of feeding groups stratlactating cows and one feeding group for dry egy was simulated for a 5-yr period, for all cows. The data used to evaluate these grouping production levels, using the appropriate base strategies were generated from a dairy herd herd. The formulations of the complete rations production computer model (16). This model used in these simulations are in Table 2, and is an extension of that described by Mertens the composition and prices of the feeds used in animal's NDFIC places an upper physical limit on DMI for a particular ration (10). Several other criteria are used in the field to rank lactating cows for placement into feeding groups. These criteria are based on BW, stage of lactation, reproductive status, or milk production (5). The objective of this study was to evaluate and compare various grouping strategies with respect to milk production and economic consequences. Computer modeling and simulation were used to meet t h i s objective.
TABLE 1. Neutral detergent fiber intake capacity means' for first lactation and older cows at different stages of the calving interval. Stage Category
calving
100 d Postpartum
First lactation cows Older cows
.85 .80
120 1.20
300 d Postpartum 1.00
.95
Next CalVina
.80 .80
'Means are expressed as a percentage of BW. and values at intermediate stages were calculated by linear interpolation. For example, the NDF intake capacity of a 500-kg first lactation cow 200 d postparhm is ((1.2 + ((1.01.2)/(300 - 100)) X (200 - 100))/100) X 500 = 5.5 kg of NDF. Journal of Dairy Science VoI. 75, No. 1, 1592
157
EVALUATION DAIRY GROWING cR[TERIA
TABLE 2. Composition' of complete rations fed to cows with different potential milk production levels within two feeding and three feeding p u p systems.
+
Ration and composition
High Forage DM? % mL* CP, %
NDF, 96 MediUm Forage DM? % m L Mcavkg CP, % NDF, %
mee feeding groups
T w o feeding gmups
sod
9Ooo
1o,o00
so00
9Ooo
1o.ooo
62 1.665 18.26 33.94
57 1.683 19.21 32.05
50 1.707 20.54 29.40
55 1.689 19.59 31.29
50 1.707 20.54 29.4
45 1.726 21.48 27.5 1
... ...
... ... ...
. . .
...
... ... ... ...
70 1.617 17.44 36.33
65 1.636 18.34 34.48
60 1.656 19.24 32.64
80 1.578 15.65 40.02
78 1.586 16.01 39.28
70 1.617 17.44 36.33
85 1.558 14.75 41.86
80 1.578 15.65 40.02
75 1.597 16.55 38.17
...
LOW
Forage, DM4 % NEL, McaUkg CP, 96
NDF, %
'AU values expressed on a DM basis. 2Potential production level (average 305-d lactation production of 8o00, 9O00, and 10,000 kg of milk). 370% Corn silage, 30% hay crop silage. '60% corn silage, 40% hay crop silage.
these ration formulations are in Table 3. Lead factors were not used in formulating any of the rations. Test days were scheduled to occur on d 1 of each month, and, on each test day, individual cows were evaluated and ranked according to the grouping criteria being implemented then. For grouping strategies 1 and 2, the cluster method developed by McGilliard et al. (9) was used to rank lactating cows simultaneously by their required energy and CP concentrations. For strategies 3 through 7, the lactating cows were ranked in descending order on the basis of the grouping criteria appropriate for each strategy. The feeding groups were of fixed and equal sizes, and, after the lactating cows were ranked, they were assigned to the appropriate feeding group according to their rank.
Days dry and days open were assumed to be normally distributed. The mean (62.0, 107.0), standard deviations (10.0, 17.0), minimum (38.0, 75.0), and maximum (103.0, 185.0) values reported by Williams (15) were used to generate randomly days dry and days open, respectively. The only differences between the base herds used in 5-yr production runs were the individual cows' potential, corresponding to different herd production levels. Therefore, little changes in nodeed costs between the production runs for different levels of 3056 lactation production (SOOO, 9000, or 10,OOO kg of milk per cow) and different number of feeding groups (two or three) were expected. Consequently, IOFC was used to compare the different grouping strategies
TABLE 3. Estimated nuhient composition and cost per weight DM of the feeds used in ration formulations. Feed
Bromegrass hay Corn silage Haycrop silage Concentrate mix
NEL
(Mcavkg DM)
1.20 1.60 1.35 1.89
CP
NDF (5)
68 51 42 10.5
9.0 8.1 18.0 30.0
cost
( S k DM) a37
.OS29
.lo22 .19
Journal of Dairy Science Vol. 75, No. 1, 1992
158
WILLIAMS AND OLTENACU
TABLE 4.Average annual milk production per cow (lactating and dry) and average annual income over feed cost per cow (lactating and dry) for each grouping strategy by potential production level with two and three feeding groups. Production level, kg milk/305 d per cow
Grouping strategy Two groups 1. Required nutrient concentration' 2. Required nutrient concenh?ldon2 3. DIM 4. Test day milk 5. Test day FCM 6. Dairy merit @M) 7. DM Weighted by DIM Three groups 1. Required nutrient concentration' 2. ~ q u i r e dnutrient concentration' 3. DIM 4. Test day milk 5. Test day FCM 6. Dairy merit @hf) 7. DM Weiehted by DIM
9Ooo
10,Ooo
6459 6463 6431 6276 6384 6428 6430
7219 7234 7168 7002 7129 7161 7181
8408 8438 8353 8181 8306 8357 8402
6629 6638 6549 6385 6498 6566 6595
7483 7501 7383 7232 7315 7414 7444
8625 8661 8501 8300 8427 8582 8572
8 W
9ooo
10,ooo
1144 1144 1137 1100 1126 1136 1137
1301 1304 1289 1249 1279 1287 1292
1531 1538 1519 1477 1507 1519 1529
1164 1166 1145 1107 1133 1149 1155
1333 1338 1310 1275 1295 1317 1324
1570 1579 1542 1494 1524 1560 1558
EO00
-(kg rmlidyr per cow) - -($/yr per cow) -
'Required energy and C P divided by kilograms of predicted DMI. 2Required energy and CP divided by kilograms of estimated NDF intake capacity
8O00,9O00, and 10,OOO kg, feed costs per cow averaged over strategies 1 and 2 were $20.67, $40.15, and $20.61 greater, respectively, with three fesding groups than with two. However, because of higher levels of milk production per cow with three feeding groups, IOFC per cow averaged over strategies 1 and 2 were $21, $33 and $40 greater, for potential production levels of 305d lactation production of SOOO, 9000, and l0,OOO kg, respectively. The lower level of milk production obtained with two feeding groups is probably due to the drop in milk RESULTS AND DISCUSSION production that occurs when cows are switched Average annual milk production and IOFC from a high to a much lower energy diet (1, per cow (lactating and dry) for each grouping 11). Smith (14) reported that cows in a one strategy are given in Table 4 for three potential feeding group system produced 120 kg more production levels, considering two and three FCM per cow than cows in a two feeding feeding groups systems. Milk production and group system. With three feeding groups, the IOFC per cow were highest with strategies 1 dietary change that occurs usually is not as and 2 and lowest with strategy 4 for all poten- large as with two feeding groups, and the tial production levels within two and three negative effect on milk production is feeding groups. Milk production and IOFC per minimized. Consequently, the impact of three cow also were higher with three feeding feeding groups on annual IOFC per cow is groups than with two for all treatment combi- greater as production level of cows increase. nations. These results indicate that strategies 1 There were small differences in annual and 2 were superior to the others in placing IOFC per cow between strategies 1 and 2. This cows into feeding groups so that their nutri- suggests that nutrient requirements expressed tional requirements were more closely met. on a per kilogram of NDFIC basis is just as For 305d potential lactation production of effective as nutrient requirements expressed on within potential production levels and between two and three feeding group systems. During the last year of each 5-yr production run,data on total feed cost, total milk produced, and number of cow-days (lactating and nonlactating) were collected. Milk was valued at $12.50145.4 kg, and the annual IOFC per cow was calculated as 365 multiplied by the difference between total income and total feed cost divided by the total cow-days in the herd.
Journal of Dairy Science Vol. 75. No. 1, 1992
C V A L U A l l U N UAIKI
a per kilogram of DMI basis in ranking lactating cows for placement into feeding groups. Grouping strategy 4 (based on test day milk production as the criterion for ranking cows) had the lowest IOFC per cow for all treatment combinations. This is not surprising, because milk production is more related to absolute nutrient requirements and is less capable of explaining differences in required nutrient concentration between cows (9). The addition of fat test information to the milk production (grouping strategy 5 ) resulted in higher annual IOFC per cow compared with strategy 4. Adding BW information to the fat test and milk production information (grouping strategy 6) increased the IOFC per cow over strategy 5. Grouping strategy 3 (based on DIM as the grouping criterion) was just as effective as strategy 6 (dairy merit) in terms of IOFC per cow for the three potential production levels with two feeding groups. The effectiveness of DIM as a grouping criterion is because it places all fresh cows and firstcalf heifers into the high feeding group, and it also would tend to keep the firstcalf heifers in the high feeding group for a longer time. With three feeding groups, dairy merit was superior to DIM for the three potential production levels. These results probably are because, if DIM is used to rank cows with three feeding groups, fint calfheifers would be kept in the high feeding group for a shorter time than with two feeding groups. This switch to a lower feeding group earlier in the lactation would result in lower milk production and a reduction in the repletion of body reserves. Strategy 7, in which grouping was based on combined DIM and dairy merit information, resulted in IOFC per cow that were higher than grouping based on only dairy merit for all treatment combinations except for three feeding groups with milk potential of 10,OOO kg. The deviations in IOFC per cow of strate gies 3, 4. 5, 6, and 7 from the mean IOFC per cow of strategies 1 and 2 are shown in Figure 1 for two and three feeding groups. These deviations show that, if three feeding groups are used, more can be gained by using either strategy 1 or 2, especially for high potential production levels. The poor performance of strategy 4 (test day milk) is because no other information is used along with test day milk. However, a common management strategy
20
2 Feeding graups 8000
-
0
z
-20
H
9
3 Feeding groups
1oow
9000
8000
10000
9000
P 0
-40
w
0
"1
-M)
Y
Daysinm~k Tes;FCM
0
Daw
~esidaymllk ,
m , Ihqme:;
,
men1 weighted by days in m i l l
100
8000
9000
10000
8000
9000
10000
POTENTIAL PRODUCTION LEVEL (kg mllk1305-d)
Figure 1. Difference between the mean income over f e d cost (IOFC)per cow of grouping strategies 1 and 2 and the IOFC per cow for the other grouping strategies.
used with grouping systems based on test day fresh milk is to force cows less than 2 mo into the high feeding group, retain fistcalf heifers and cows in poor body condition in that group for a longer time, and force cows late in lactation into the low feeding group. These modifications to the use of test day milk alone may explain why test day milk still remains a popular method used by dairy producers to group lactating dairy cows. CONCLUSIONS
Using annual IOFC per cow to evaluate the effectiveness with which the seven grouping strategies placed lactating cows into feeding groups, groupings based on required nutrient concentration in the ration (strategies 1 and 2) were superior for all treatment combinations. Ranking cows by test day milk production was the least effective of the grouping strategies, but the addition of fat test and BW information to the test day milk resulted in higher IOFC per cow. These four strategies (1, 4, 5 , and 6) also were ranked the same by McGilliard et al. (9), who used actual production records from 80 Holstein herds to study how effective the strategies were in grouping cows homogeneously by nutritional requirements. Our results suggest that three feeding groups should be used for potential production levels of 9000 Journal of Dairy ScienCC Vol. 75. No. 1, 1992
160
WILLIAMS AND OLTENACU
and 10,000 kg of milk. For low potential production levels (8000 kg of milk or less), the use of only two feeding groups is better &ecause the additional nonfeed costs associated with three feeding groups may not be offset by the small increase in annual IOFC per cow. Other management considerations may have to be used to justify a larger number of feeding groups. The small differences in IOFC per cow between strategies 1 and 2 suggest that either NDFIC or DMI may be used to calculate required nutrient (protein and energy) concentrations. But using NDFIC may be slightly more advantageous because it can be easily estimated; also, problems associated with predicting DMI are avoided. ACKNOWLEDGMENT'S
This research is a component of the NC-119 Regional Project and was in part supported by USDA and by the Agricultural Experimental Station at Comell University, Regional Hatch Project Number 476. REFERENCES
1Akinyele, I. 0..and S. L. Spahr. 1975. Stage of lactation as a criterion for switching cows from one complete feed to another during early lactation J. Dairy Sci. 58:917. 2 Brown, C. A., P. T. Chandler, and J. B. Holter. 1977. Development of predictive equations for milk yield and dry matter intake in lactating cows. J. Dairy Sci. 60:1739. 3 Cassel, E. K.,W. G. Merrill, T. V. Bui, R A. Milligan, and R W.Guest. 1982. Evaluation of systems for feeding supplemental concentrates to group fed cows.
J o d of Dairy Science Vol. 75, No. 1, 1992
Comell Univ. Anim. Sci. Mimeo. Number 59, Ithaca,
NY. 4 Congleton, W. R, and R. W. Everett. 1980. Application of the incomplete gamma function to predict cumulative milk production. J. Dairy Sci. 63:109. 5 Coppock, C. E. 1977. Feedii methods and grouping systems. J. Dairy Sci. 601327. 6Coppodr. C. E.,D. L. Bath, and B. Harris, Jr. 1981. From feeding to feeding systems. J. Dairy Sci. 64: 1230. 7 Evason, R. A., N. A. Jorgeoson, J. W. Crowley, E. L. Jensen, and G.P. Barrington. 1976. Input-output of dairy cows fed a complete ration of a constant or variable forage-tc-grain ratio. J. Dairy Sci. 591776. SHoglund, C. R. 1973. Dairy facility investments and labor economics. J. Dairy Sci. 56488. 9Mdjilliard, M. L.,J. U Swisher, and R. E. James. 1983. Grouping lactating cows by nutritional requirements for feeding. J. Dairy Sci. 66:1084. 10 Mertens, D. R 1987. Predicting intake and digestibility using mathematical models of ruminant function. J. Anim Sci. 64:1548. 11 Moseley, J. E., C. E. Coppock, and G.B. Lake. 1976. Abrupt changes in the forage-concentrate ratios o€ complete feeds fed ad libitum to dairy cows. J. Dairy Sci. 59:1471. 12Polan, C. E.,and T. H. Friend. 1975. A comparison of herd feeding: all togetber or by production groups. Proc. 1975 College Dairy Fecd Cod. Board, Blacksburg, VA. 13Raka, A. H. 1969. Complete rations for dauy cattle. J. Dauy Sci. 522370. 14Smith, N. E. 1976. Maximizing income over feed costs: evaluation of production response relationships. J. Dairy Sci. 59:1193. 15Williams, C. B. 1985. Correlation analysis of dairy practices and management factors on New York dairy farms, 1982. Comell Univ. Agric. Ikon. Res. Rep. 85-3, Ithaca. NY. 16Witliams. C. B., P. A. OItenacu, and C. J. Sniffen. 1989. Application of neutral detergent fiber in modeling feed intake, lactation response and body weight changes in dairy cattle. J. Dairy Sci. 72:652.
