Effect of environmental factors on lactation curves of milk production traits in Holstein-Friesian cows reared under North African conditions

M Bouallegue, B Haddad, M S Aschi and M Ben Hamouda*

Institut National Agronomique de Tunisie,
43 Av. Ch. Nicole, 1082 Tunis Mahragène, Tunisie.
cfpaa2006@yahoo.fr
* Institution de La Recherche et de l’Enseignement Supérieur Agricoles, (IRESA),
Ministère de l’Agriculture, Tunisie

Abstract

The Wilmink function was used to fit lactation curves of Holstein- Friesian cows in Tunisia using the Levenberg-Marquardt’s iterative method in the nonlinear regression procedure in SAS. The 305 days yields for all production traits were estimated with the Wilmink function. The influence of the non genetic factors on the 305-day milk, fat and protein yields in relation with the classical curve’s traits was investigated and a procedure for editing environmental factors in groups (Calving age-parity, herd-calving year and calving season) was derived. Test-day records of 259,776 observations of 5,649 Tunisian Holstein-Friesian cows were used in the analysis.

The results showed that the WIL function was a good fit for the lactation curves in different grouped effects and explained 98 % of variation in daily yield of milk, 96 % for fat and 97 to 98 % for protein yields. The average yields for milk, fat and protein shoed many fluctuations in different environment effect groups. Calving age-parity, herd-calving year and calving season was highly significant (P <0.01) for the 305-day milk, fat and  protein yields and for all curve traits, except for the persistency of fat yield which was obviously not affected by the herd-calving year factor (P>0.05). The highest significance of environmental effects on lactation curve parameters for milk traits allows having different shape lactation curves. The Shapes of the WIL lactation curve suggest that cow with a higher initial level of production attain faster the peak and thereafter decline at a slower rate than those with a lower initial level of production. Cows have lower peak milk but greater persistency in first than later calving age-parity groups. In all environmental factors Fat trend to be more persistent than milk and a little difference was observed between fat and protein. The 305 days yield of milk, fat and protein were positive and highly correlated with yield at the beginning of lactation (r=0.74, 0.50 and 0.57 respectively for milk, fat and protein yields) and peak yield (r=0.98).

Keywords: test day, lactation curve, environnemental effects

Introduction

The direct modelling of individual test-day (TD) records was proposed as a replacement for the traditional use of estimated accumulated 305 day yields. Some advantages of TD models include the ability to account for environmental effects of each TD, the ability to model the trajectory of the lactation for individual genotypes or groups of animals and a more precise adjustment for temporary environmental effects on TD (Jensen 2001).

Knowledge of the lactation curve of lactating dairy animals is required for feeding, breeding and economic management of a dairy herd. Lactation curves are especially important when making decisions. Dairy cows with a flat lactation curve are considered to have more persistent lactations and produce milk with a lower cost (Tekerli et al 2000).

Milk yield of individual cows is influenced by many factors (Van Tassell et al 1995), including genetic ability, parity, DIM (Days in Milk), age, physiological state, and management. In the lactation curve studies, calving age is generally used as a variation factor than the age of the animal; the parity was reported to be a significant source of variation and the lactation curves are usually developed using the average TD data on the milk yield within parity groups (1, 2 and ≥ 3) (Van Tassell et al 1995).

The combined exponential and linear model proposed by Wilmink (1987) is one among several models reported in the literature to describe the lactation curve of dairy cows. The Wilmink (WIL) model was developed in Holland and was the original function employed to describe the lactation curves in the official program of the genetic evaluation in Canada (Schaeffer et al 2000). In this form, it is considered the best three parameters of lactation curve (Olori et al 1999) and it is used in the random regression analysis. Moreover, WIL model is also used in some studies detecting the shape of the lactation curve (Macciotta et al 2005 and 2006, Roshanfekr et al 2010). This model is specifically conceived to model the lactation curve and WIL parameters can also be easily related to the characteristics of the lactation curve shape (Macciotta et al 2005).

Objectives

The main objective of this study was to analyze some environmental factors affecting lactation curve’s traits and the 305-day yields, for Holstein-Friesian cows raised under the Tunisian conditions by fitting Wilmink model to test day records.

Materials and Methods

Data

Data were 259,776 TD records of milk trait yields recorded on 5,649 Holstein-Friesian cows were collected by the National Centre of Genetic Improvement of Tunisia (CNAG: Sidi Thabet, Tunis) on 188 herds during the years 1994-2002. Lactation records with less, than 10 consecutive test-days were not included in the analysis. Likewise, test-day yields recoded up to 5 days following calving, biologically unacceptable fat or protein yields and a very low milk yield (< 3 Kg) were omitted. The lactation period was limited to 5-305 day after calving. Table 1 summarized the average milk yield (MY), fat yield (FY) and protein yield (PY) and their standard deviations for six calving age-parity groups.
 

Lactation curve Model

The WIL model (Wilmink 1987) was used to fit individual lactation curves for milk, fat and protein was used:
                                               
Where: e is the Neper number, a, b and c are the parameters that determinate the curve’s shape.

Curves were fitted using the Levenberg-Marquardt’s iterative method via the NLIN procedure of SAS [SAS 2001]. The parameter K is connected to the time of peak lactation and usually assumes a fixed value, derived from a preliminary analysis made on average production The peak yield (Peak) and the time to which one observes this maximum (DIMP) was calculated as:

Persistency for milk traits was calculated as the yield at one test expressed as a percentage of yield at an earlier test. The main criteria used to evaluate the goodness of fit of WIL model were, the coefficient of determination (1- the ratio of the residual sum of squares to the total sum of squares) and the residual (RES), defined as the absolute values of the difference between the predicted and real yields. The 305-days yields () were estimated by the WIL model to the following formula:
.

Factors affecting lactation curves

To evaluate the significant effect of age at calving, parity, month of calving, year of calving and herd on the lactation curve parameters and the 305-days yields, data used in this work (Table 1) were grouped to six groups (G1 to G6) of calving age connected with the parity of cow at TD  control. Three groups (HCY1 to HCY3) of herd-calving year were formed according to the level of production in herds on calving years. Season groups were defined as for 3-month intervals: winter (December-February), spring (March-May), summer (June-August), and fall (September-November).

The effects of calving age-parity, herd-calving year, calving season and first test-day date on milk, fat and protein yields and lactation curve parameters were analyzed using a general linear model:

Results and Discussion

Goodness of fit for WIL model

In a preliminary analysis, parameter K was estimated at 0.65 for MY and 0.10 for FY and PY. The value found for milk is agrees with that used by Silvestre et al (2006) and close to that (k=061) found by Olori et al (1999), whereas Wilmink (1987) proposed a value of 0.50 for Dutch- Friesians cattle. The WIL was a good fit for average lactation curves across cows (Table 2) in different grouped effects. The coefficient of determination from non-linear regression is 98 % for MY, 96 % for FY and ranged from 97 to 98 % for PY. The WIL also predicted better the yields; RES ranged from 1.70-2 Kg, 0.08-0.10 and 0.05-0.07 Kg, respectively for MY, FY and PY. Standard deviations of residual were uniformly smaller for protein than fat yield, similar results were noted by Schutz et al (1990). Thus, PY predicted yield with lower residual than FY in all environmental factors.

Factors affecting the shape of lactation curve for milk, fat and protein yields

The ANOVA mean squares of model effects are in Table 3. The coefficient of determination (R²) ranged from 0.05 for b to 0.25 for DIMP (days in milk at peak) for milk yield, these results are similar to previous studies (Tekerli et al 2000, Atashi et al 2009). For fat and protein yields, the R², ranged respectively from 0.01 (persistency) to 0.57 (DIMP) and from 0.01 (b) to 0.64 (DIMP). All grouped effects were highly significant (P<0.01), but only the herd-calving year was not significant (P>0.05) for fat persistency. The high significance of calving age parity is explained mainly by the effect of milk secretory tissue that requires more time for peak activity in primiparous cows than in multiparous cows Rao and Sundaresan (1979). The high level of significance of the herd-calving year effect can be explained by the difference in management between herds and also the diverse feeding level which changes according to the annual climate frequently has observed under the Tunisian conditions. The effect of calving season on lactation curve traits is explained by temperature and rain variations that affect fodder production, especially in summer under the North African climate when feeding resources are limited and heat stress effect is important. The highest significance of all grouped environmental effects on lactation curve parameters a, b and c for milk traits allows having different shapes of the curves. In the North of Africa conditions similar results were indicated in the literature but were limited only for milk yield. Boujenane and Hilal (2012) studied the genetic and non genetic effects on lactation curve traits determined by the incomplete gamma function for Holstein-Friesian cows in Morocco and reported that lactation curve traits (a, b, c, DIM, peak, persistency and Y₃₀₅) were affected by herd, parity, age at calving, season of calving and year of calving. Rekik et al (2003) fitted lactation curves of Holstein-Friesian cows in Tunisia in four types of herds and reported also some similar finding for milk yield.

The least squares means for the lactation curve traits are presented in Table 4 for MY and in Table 5 for FY and PY. In this investigation, the parameter a (represents a yield at the beginning of lactation), peak and 305-days yields were the lowest in the first calving age-parity group (G1) than other groups, but cows are more persistent for all yields traits. The highest persistency may be caused by the relatively lower peak yield resulting in a flatter lactation curve. The 305-day yields (Tables 4 and 5) show that the first group cows (G1) produce the lowest yields than the other groups, for example a difference of 635 Kg MY and 27 Kg FY and PY was noted between G1 and G3. Indeed, 97 % of the cows classed in G1 (Table 1) are in the first lactation (primiparous) and their calving age ranged from 16-36 months. Stanton et al (1992) observed the same effect in dairy cows and suggested that this pattern could be due to the fact that the body and mammary gland of young animals are still developing during the first part of lactation. Total yields increase with the calving age-parity groups and the yields peaked in the third group (G3) whose animals are mainly in the third lactation (77 %) and their calving age ranged from 50 to 62 months. Then, yields decline from the fourth lactation (G4), this decline in yields according to the advance of calving age is explained by the physiological activities of all body systems start to decrease and the secretary tissues of mammary gland is partially degenerated leading to gradual decrease in milk production with advancing age, in agreement with Keowen et al (1986).

For the herd-calving year groups, cows calved during 1993, 1995, 2000 and 2001 recorded more production of MY, FY and PY during their lactations. For calving season, the lowest level of production occurred for cows were calved in summer, but these cows are more persistent. While, the highest level happened for cows were calved in winter and fall for MY. Similar reports were obtained by (Tekerli et al 2000 and Atashi et al 2009). And there are a little variation between FY and PY in spring, fall and winter. However, summer months depress total milk, fat and protein yields (Keown et al 1986).

For DIMP, there is a tendency to decrease as calving age-parity groups increase with a little difference starting from the fourth group for MY and the third group for FY and PY. Peak yield occurs between 6^th at the 8^th week of lactation in all groups and traits. The great variation (3 to 6 days) was observed between the first and other groups according to the calving age-parity and between summer and the other calving season. Moreover, there is little variation in DIMP by herd-calving year effect. This was in accordance with the finding of Keown et al (1986).

For all production traits and all environmental effects, fat trend to be more persistent than milk and little difference was observed between fat and protein. Keown et al (1986) reported similar persistencies of fat and protein yields, protein tending to be more persistent than fat. Schutz et al (1990) indicate that protein was more persistent than fat yield for Holstein cows. Cows in high herd-calving year production groups have the highest peak production for all milk traits and more persistent for milk yield, but little difference in persistency with the three groups for fat and protein yields.

Lactation curves fitted for milk, fat and protein yields according to environmental factors are in Figures 1 to 6. Lactation curves for all yield traits and for all environmental factors were depressed at approximately 230 days in milk. This is explained by an increase in the influence of pregnancy depression at 7 to 8 months in gestation. The shape of the lactation curve differed between fat and protein yields in each environmental group. Figure 1 shows that the flattest curve was detected in the first calving age-parity group (G1) that started at a low level (18.4 kg), has a very slow increase (0.18 Kg/day) to the peak, which was not very high (23 kg) and was reached late in lactation (54 DIM), then the milk yield decreased slowly until the end of lactation. These trends in the shape result that cows’ group was more persistent. The most typical lactation curves were observed in the G3 and G4 cows’ group, with the majority of the animals are in third lactation (77%, for G3) and in the fourth lactation (70%, for G4). These cows started the lactation with a high level (23-24 kg) with a moderate rate of increase up to the peak production that was reached between 45-46 DIM at a high level (28 kg), they followed by a moderate rate of decrease of daily production. The shape of the lactation curves for cows in the fifth group were different to G3 and G4 only at the initial phase of lactation but from the peak the same trends were observed. This difference results in the fact that these animals produce 123 kg of milk less than those which have the maximum of Y₃₀₅ (G3).

The lactation curves for FY and PY according to the calving age-parity effect are presented In Figure 2, cows in the first group have a flatter curve for MY (Figure 1) and PY (Figure 2-b) than for FY (Figure 2-a). Keown et al (1986) found differences in lactation curves for fat and protein yields. Primiparous cows (G1) reach their peak faster for FY than for PY but present almost a similar trend in the decrease of yields. The highest curves for FY and PY (Figures 2-a, and 2-b) were obtained with cows which calving age are between 50 to 74 months classed in G3 and G4 groups. Although, those of G3 produce more than G4 in the second part of lactation and the curves of G3 and G4 were closer for FY than for PY. The highest level of production at the beginning of lactation was recorded in the third and the fourth parity of animals (G3 and G4) for all milk traits. While, the highest level of production at the end of lactation were obtained with cows in the first and in second lactation (G1 and G3).

The patterns of lactation curves in different herds classified by calving year are presented in Figures 3 and 4. The shape of the lactation curve for MY, was different between the three herd groups. Cows in HCY3 and HCY2 had the highest rate of increase to peak, but those in HCY3 produced the highest peak (Figure 3). Thus, in HCY1 the daily milk yield increased and decreased more rapidly with the lowest peak yield. However, cows in HCY3 presented a more typical lactation curve than HCY 1 and 2 so, this group produced the highest level of total milk yield (for example 6124 kg for HCY 3 and 5554 Kg for HCY 1).

The cows of the second and the third Herd-calving year groups present exactly the same shape of the lactation curve for FY, whereas they are different for PY. In all groups, FY increases rapidly after partition than PY, but with higher level for FY at DIMP. Curves for PY were very flattered than FY and closer to MY in general shape aspect.

The lactation curves for the different calving seasons are presented in Figure 5 and 6. Results show that the calving season affected the shape of fat and protein curves more than milk. The biggest fluctuation of MY between seasons was observed around the peak and until about 200 DIM. Whereas, season effect influenced FY and PY curves in the majority parts of lactation. For MY cows calved in fall, winter and summer started their lactation with the same level of production (Figure5) and this aspect was probably attributed to the body condition of the animal at calving and feeding level of the cow before and around partition than of the seasonality effect. While, those calving in summer increase fast to DIMP and recorded a lower peak yield. The higher peak yield with slower rate of decline was obtained from cows calving in the fall and winter. And curves for fall season were highly persistent until the end of lactation.

According to Figures 6-a, and 6-b fall and winter calving seasons were more favourable for fat and protein production. Indeed, initial yield and peak for fat and protein yields were greater for fall and winter. Cows freshening in the summer and spring have the lowest peak production for FY and PY but as lactation length increases; they augment their production and become more persistent. On the opposite, fat and protein production in winter and fall were higher and more persistent until about 200 DIM but after that production decreased fast and become inferior to other seasons. Indeed, cows calving in the spring and summer finish their lactation during cooler months, whereas those calving in winter and fall complete their lactations during warmer months and decline of more sharply (Keown et al 1986). This aspect explains the change of production levels between lactation stage and season effects.

Figure 6: lactation curves for calving season effect for FY (a) and PY (b)

Phenotypic correlations between the lactation curve traits are shown in Table 6. The highest correlation found was between a and peak yield (r=0.98), and the lowest correlation was between a and c for all milk production traits. The negative correlation between a and b and between b and peak suggest that cow with a higher initial level of production reached their peak faster and thereafter decline at a slower rate than those with a lower initial level of production. This aspect is observed with the third and fourth calving age-parity groups, in winter and fall for fat and protein yields and for second and third herd-calving year groups for all traits. The positive correlations between b and c indicate that cows with a higher rate of increase until the peak and also have a quicker decline after peak. The initial yield for MY, FY and PY (parameter a) was positively higher correlated with Y₃₀₅ and peak, but negatively associated with DIMP suggest that cows with bigger initial yields tend to have higher and peak yields and recorded the biggest total yields (Y₃₀₅). This suggestion is very clear with the result obtained by the third calving age-parity group, in winter and HCY3 group (Table 4 and 5). The correlation between c and persistency is higher and positive for MY and PY than for FY indicate that cows with a lower rate of decline have higher persistency. Finally, the results show no relation of persistency (P>0.05) with a total fat yield in 305 DIM. The correlation between peak yield and lactation yield (r=0.84, 0.57 and 0.65 respectively for MY, FY and PY) was higher than that observed between persistency and lactation yield .Thus peak yield seems to be more important in determining the total lactation yield than persistency.

Conclusion

• The direct modelling of individual test-day (TD) yield was effective in generating lactation curve estimates for milk, fat and protein yields. The curves developed from the Wilmink model as influenced by the various factors associate in groups in the present study, may serve as useful to help dairy producer. Different shapes of lactation curves detected suggest that the nature of the curves could provide a basis for planning and adjustment in the management of herds, particularly with regards to culling and assessment of the nutritional and health status of animals. Indeed, the interaction herd-calving year and calving months grouped in season effect of lactation make it possible to take into account differences related to the system in control of breeding and in particular to the food. Calving age-parity lactation curves offer the possibility to evaluate the animal performance with parity related to calving age.

  
  

• Initial yield, peak and 305-days yield were lower in the first calving age-parity group but their cows were more persistent for all yields traits. For all environmental factors, fat tend to be more persistent than milk and a little difference was observed between fat and protein. For calving season, the lowest level of production occurred for cow calved in summer, and the highest levels of production were found for cows calved in winter and fall for milk, in spring and fall for fat yield and with the same level of protein yield in winter, spring and fall. The shape of the lactation curve differed between fat and protein yields in each environmental group.

Acknowledgements

The National Centre of Genetic Improvement of Tunisia (CNAG: Sidi Thabet, Tunis) is greatly acknowledged for providing the data used in this study.

References

Atashi H, Moradi Sharbabak M and Moradi Shahrbabak H 2009 Environmental factors affecting the shape components of the lactation curves in Holstein dairy cattle of Iran. Livestock Research for Rural Development 21(5): http://www.lrrd.org/lrrd21/5/atas21060.htm 

Boujenane I and Hilal B 2012 Genetic and non genetic effects for lactation curve traits in Holstein-Friesian cows. Archiv Tierzucht 55 (5): 450-457: http://arch-anim-breed.fbn-dummerstorf.de/pdf/2012/at12p450.pdf 

Jensen J 2001 Genetic evaluation of dairy cattle using test-day models. Journal of Dairy Science 84, 2803-2812.: http://www.aseanbiotechnology.info/Abstract/21025694.pdf 

Keown J F, Everett R W, Empet N B and Wadell L H 1986 Lactation curves. Journal of Dairy Science 69: 769-781: http://www.journalofdairyscience.org/ 

Macciotta N P P, Vicario and Cappio-Borlino A 2005 Detection of different shapes of lactation curve for milk yield in dairy cattle by empirical mathematical models. Journal of Dairy Science 88: 1178-1191: http://download.journals.elsevierhealth.com/pdfs/journals/0022-0302/PIIS0022030205727843.pdf 

Macciotta N P P, Dimauro C, Catillo G, Coletta A and Cappio-Borlino A 2006 Factors affecting individual lactation curve shape in Italian river buffaloes. Livestock Production Science 104: 33-37. 

Rao M K and Sundaresan 1997 Influence of environmental and heredity on the shape of lactations curves in sahiwal cows. Journal Agricultural science Cambridge 92: 393-401. 

Olori V E, Brotherstone S, Hill W G and McGuirk 1999 Fit of standard models of the lactation curve to weekly records of milk production of cows in a single herd. Livestock Production Science 58: 55-63. 

Rekik B, Ben Gara A, Ben Hamouda M and Hammami H 2003 Fitting lactation curves of dairy cattle in different types of herds in Tunisia. Livestock Production Science 83: 309-315. 

Roshanfekr H, Mamouei M, Yarinejad F and Mohammadi K 2010 Phenotypic study of lactation curve in Iranian Holstein. Journal of animal and veterinary Advances 9(4): 721-725: http://docsdrive.com/pdfs/medwelljournals/javaa/2010/721-725.pdf 

Schutz M M, Hansen L B and Steuernagel G R 1990 Variation of milk, fat, protein and somatic cells for dairy cattle. Journal of Dairy Science 73: 484-493: http://download.journals.elsevierhealth.com/pdfs/journals/0022-0302/PIIS0022030290786961.pdf 

Stanton T L, Jones L R, Everett R W and Kachman S D 1992 Estimating milk, fat and protein lactation curves with a test day model. Journal of Dairy Science 75: 1691-1700: http://download.journals.elsevierhealth.com/pdfs/journals/0022-0302/PIIS0022030292779260.pdf 

Schaeffer L R, Jamrozik J, Kistemaker G J and Van Doormaal B J 2000 Experience with a test-day model. Journal of Dairy Science 83: 1135-1144: http://www.journalofdairyscience.org/ 

Silvestre A M, Petim-Batista F, and Colaço J 2006 the accuracy of seven mathematical functions in modelling Dairy cattle lactation curves Based on test-day records from varying sample schemes. Journal of Dairy Science 89: 1813-1821. 

SAS 2001 SAS User’s Guides Version 9.1 for Windows, SAS Institute Inc., Cary, NC. 

Tekerli M, Akinci Z, Dogan I and Akcan A 2000 Factors affecting the shape of lactation curves of Holstein cows from the Balikesir province of Turkey. Journal of Dairy Science: 83 1381-1386: http://jds.fass.org/cgi/reprint/83/6/1381.pdf 

Van Tassell C P, Jones L R and Eicker S W 1995 Production evaluation techniques based on lactation curves. Journal of Dairy Science 78: 457-465:  http://aipl.arsusda.gov/publish/jds/1995/78_457.pdf 

Wilmink J B M 1987 Adjustment of test day milk fat and protein yield for age, season and stage of lactation. Livestock Production Science 16: 335-348.

Received 31 December 2012; Accepted 28 April 2013; Published 1 May 2013

Go to top