Livestock Research for Rural Development 20 (2) 2008 Guide for preparation of papers LRRD News

Citation of this paper

Dairy heifer rearing under increasing intensification of smallholder dairy systems in the Kenya highlands

B O Bebe

Department of Animal Sciences, Egerton University,
Box 536 Egerton-Njoro, Kenya
obebeb@yahoo.com

Abstract

The objective of this study was to assess dairy heifer rearing practices in relation to varying levels of intensification in the central highlands of Kenya. Data were obtained from a random sample of 987 farm households in a cross-sectional survey.

 

Grazing management, reflecting varying levels of intensification, had a significant influence on rearing practices for replacement animals.  For calves, bucket feeding was more common (p<0.05) and tick control less common (p<0.05) in semi-zero and zero grazing than in free grazing farms.  Semi-zero and zero grazing farms had lower (p≤0.017) median proportions of heifers to cows (PHC) (0.00 to 0.18) compared to free grazing farms (0.31).  Compared to free grazing, semi-zero and zero grazing farms kept PHC in the farm lower by weaning replacement females one month earlier (p≤0.017) and selling them when (8 mo) 4 months younger.  Lower PHC was associated with hectares of maize planted per unit cattle (p<0.0003), and cash spent on purchase of fodder per unit cattle (p<0.0818), but higher PHC was associated with the more available land per unit cattle (p<0.0248).  Keeping PHC low is likely a rational choice by smallholder farmers to increase farm productivity because of scarcity of land to plant fodder, and scarce capital to buy supplementary feeds.  Policy reforms supportive to vertically integrated dairy development can encourage farms with comparative advantage to produce replacement heifers for farms unable to produce their own replacements.

Key words: dairy intensification, farm productivity, heifer proportion to cows, smallholders


Introduction

Lack of dairy replacement heifers is one of the major limitations to the development of smallholder dairy production in developing countries (De Jong 1996).  In Kenya, in the period soon after independence, smallholder farmers sourced dairy replacement stock from public and private large-scale farms (Conelly 1998).  However, rapid expansion of smallholder dairy farms occurred as large-scale farms rapidly declined (Stotz 1979; Conelly 1998).  An earlier study of herd dynamics in smallholder farms in the central highlands of Kenya indicated that smallholders mainly rear their own dairy replacements or purchase them from other smallholdings in the Rift valley area (Bebe et al 2003).  Sourcing of dairy replacements from large-scale farms was very minimal.  However, half of the smallholder dairy farms did not keep any heifer, with majority keeping mostly cows.  As a result, the proportions of heifers to cows (PHC) were insufficient for replacement of the cows that left the herd.  The PHC in these farms is likely to be related to farm feed resources and the need to increase farm productivity where farmers lack access to grazing pastures and have to adopt more intensive dairy production systems. 

 

It can be expected that the trends towards more intensive dairy production will increase the constraints to rearing of replacement heifers in smallholder crop-dairy systems.  Thus a better understanding of heifer rearing practices in relation to farm feed resources and productivity as smallholders intensify their production systems will be needed in order to identify technologies and policies that can support productive and sustainable smallholder dairy production.  The possible relationship between PHC and own produced feeds, purchased feeds, and milk yields has not been quantified in earlier studies that focused on dairy heifer rearing on smallholder farms (Gitau et al 1994a and b; De Jong 1996, Lanyasunya et al 1999).  The objective of this study was to assess dairy heifer rearing practices in relation to varying levels of intensification in smallholder farms in the central highlands of Kenya.

 

Materials and methods 

Data collection 

A stratified random sample cross-sectional survey of 1755 households in the Kenya highlands was conducted between June 1996 and April 1998 to obtain data on livestock management practices as part of characterisation studies of dairy systems supplying the Greater Nairobi milk shed.  Stratification was by agro-ecological potential (for cropping and dairying) and milk market access as main determinants of dairy system development (MoA 1998; Staal et al 1998).  There were 987 households keeping cattle in the total survey sample.  A detailed description of the study sites, survey methodology and herd management are presented in an earlier part of this study (Bebe et al 2003). 

 

Cattle households were interviewed to obtain information on management practices concerning female replacement animals (from birth until age at first calving) on their farms.  Information was collected on methods of milk feeding and tick control, use of anthelminthics, and the ages at weaning, at selling and at first calving (AFC).  Information on AFC was collected on cows present in the herd.  The breeds of those cows were identified.  Data collected on feeding management included system of grazing, total farm size, land size allocated to maize and fodder growing and purchase of fodder and concentrates in the past one year. 

 

The total land owned by a household, land allocated to maize and land allocated to napier were chosen to reflect the availability of own produced feed resources.  Purchased fodder and concentrates reflected the level of dependency on external feed resources.  Hectares of land for own produced feeds and cash spent on purchased feeds were expressed per Tropical Livestock Unit (TLU=1 for bull, 0.7 for cow, 0.5 for heifer and young bull, 0.2 for calves).  These variables (own produced and purchased feeds) were considered to be measures of level of intensification of the dairy farm system.

Statistical analyses

Cross tabulations was used to determine the difference between categories of grazing systems in the methods of feeding and tick control and use of anthelminthics. Event-time regression was used to quantify the effect of grazing system and breed types on the AFC.  The starting point of the measurement period was set to the lowest reported AFC (20 m).  All cows that had a record of AFC were considered to have a failure event (AFC), hence this record was considered uncensored.  The AFC data were analysed by using a semi-parametric Cox proportional hazard model because it does not require specification of a distribution for the baseline hazard function, defined as the probability of occurrence of first calving event at some observed time ti conditional upon survival to that time.  The model was represented as:

 

λ(t;x) = λ0 (t) e(x’β)

where:

λ(t;x) =  hazard of event for a cow at time t with covariates x

λ0 (t)=  baseline hazard function describing the hazard of event for an hypothetical situation when all covariate values are set to zero

e(x’β) =  term specific to individuals with covariates x 

 

The covariates included the fixed effects of grazing system (free, semi-zero and zero-grazing) and breed, categorised as Bos indicus comprising East African Zebu, Boran and Sahiwal (ZB); small mature size Bos taurus breeds comprising Guernsey and Jersey (GJ); and large mature size Bos taurus breeds comprising Friesian (FR) and Ayrshire (AY). Hazard ratios (eβ) from this model were obtained with their 90% confidence interval. 

 

The medians for the heifer rearing variables (age at weaning, age sold, and PHC), owned produced feeds, purchased feeds, and milk yields were estimated for the categories of grazing systems.  These variables were non-normally distributed, thus significant differences between pairs of grazing systems were investigated by the Mann-Whitney rank sum test (two tailed).  Multiple comparison tests were performed with significance level adjusted using the Bonferroni test (Thomas et al 1985).

 

Logistic regression (Proc GENMOD; SAS Institute Inc.1996) was used to explain the multivariate relationships between PHC and feed variables.  Prior to logistic regression analysis, co linearity between the variables was investigated by means of the bivariate correlation coefficients (PROC CORR; SAS Institute Inc.).  None of the variables had a correlation coefficient larger than 0.50.  Thus all the potential explanatory variables were offered to multivariate logistic regression model.

 

Results and discussion 

Milk feeding and tick control 

In the free-grazing farms, majority of the farmers reared at least a heifer and regularly used acaricide for tick control, but fewer farmers practiced bucket feeding and deworming compared (p<0.05) to the practice in semi-zero- or zero-grazing farms (Table 1). Compared to suckling, bucket feeding is has some disadvantages.  It is labour intensive, use of calf pens is necessary (Das et al 1999), and has been associated with lower weight gains in smallholder herds (Gitau et al 1994a).


Table 1. Percentage of farmers practicing heifer rearing, bucket feeding of calf and regularly use acaricide and anthelminthics for replacements in smallholder farms in the Kenya highlands

Grazing system

Number
 of farms

Percentage of farmers

Rearing
≥1 heifer

use bucket feeding

regularly use acaricide

regularly use anthelminthics

Free

227

62

38

92

15

Semi-zero

326

54

61

86

14

Zero

434

48

68

67

28

Total

987

53

60

84

21


Acaricide application was mainly through hand spraying. The practice of tick control in semi-zero or zero grazing farms is less frequent than in free-grazing farms because the exposure to tick infestation is lower, thus farmers perceive it to be less important (Gitau et al 1997; Gitau et al 1999; Siamba et al 1999). In a study of serum antibody prevalence of most important tick borne disease parasites (Theileria parva, Theileria mutans, Babesia bigemena, Anaplasma marginale) in smallholder farms of the central Kenya highlands, Gitau et al (1997) demonstrated that higher prevalence of serum antibodies was significantly associated with calves in free grazing than those in semi-zero or zero grazing farms.  In rating of animal health problems by these farmers, farmers practicing semi-zero or zero grazing rated tick borne diseases lower than those practicing free grazing (Bebe et al 2003). 

Weaning and selling of female replacements 

Table 2 shows the means and medians for the age at weaning and at selling of female replacements. The median age at weaning varied from 3 to 4 months.  Calves in farms practicing zero grazing were weaned one month earlier (p≤0.017) than those in farms practicing free grazing.  Early weaning in zero grazing is likely due to households’ needs to have more milk for home consumption and for sale to earn cash income as reported in a study in Bahati division within the central Kenya highlands (Lanyasunya et al 1999). That study showed that the amount of milk fed to calves from birth to 3 months of age was lower in zero grazing (193 kg/calf) and in semi-zero grazing (230 kg/calf) than in free grazing (321 kg/calf) farms.


Table 2.  Median age at weaning and at selling for replacement females in smallholder dairy farms in the Kenya highlands

Variable

Grazing system

Number of farms

Mean

Standard deviation

Median

Age weaned, months

Free

221

5.5

2.7

4.0c

Semi-zero

270

4.6

2.1

4.0d

Zero

395

3.8

1.4

3.0c,d

Age sold, months

Free

47

10.8

5.2

12.0

Semi-zero

35

7.7

5.3

8.0

Zero

103

9.6

6.2

8.0

Significant difference (Mann-Whitney rank-sum test p≤0.017) between: cfree and zero grazing; dsemi-zero and zero grazing


Because inadequate feed quantity and quality is a pervasive constraint in smallholder zero-grazing farms (Omore et al 1996; Reynolds et al 1996), the practice of early weaning requires that farmers offer supplementary feeds if they are achieve adequate growth of calves.  Without adequate supplementary feeding, early weaning will result in poor growth leading to delayed age at first calving and subsequent low lactation yields (Gitau et al 1994a; de Jong 1996; Lanyasunya et al 1999).  In addition, high mortalities can be expected (Gitau et al 1994b) thus leaving less room for selection of replacement heifers and production of surplus heifers for sale.  Regular use of supplementary feeding, however, would depend on the availability of financial resources, because cost of commercial feeds is high and may be out of reach for most of the smallholder farmers.  Support to farmer cooperative movement is one approach to help relax the constraints to supplementary feeding through provision of feeds on credit arrangements.

 

Though not significant, replacement females sold from farms practicing zero grazing were younger by 4 months than those sold from farms practicing free grazing (Table 2).  Replacement females weaned early and then sold soon after weaning in zero grazing farms is likely a managerial strategy adopted by farmers to reduce spending scarce feed resources on raising the yet unproductive animals (heifers).  Early weaning and sale soon after weaning of heifers can partly explain the lower (p≤0.017) median PHC in farms practicing zero grazing (0.00) and semi-zero grazing (0.18) than in those practicing free grazing (0.31) presented in Table 3.  In an earlier study of the herd dynamics in these smallholder farms, it was estimated that 21 to 25% of the cows leave the herd annually (Bebe et al 2003).  Thus the replacement needs of these herds are about 25% in annually.  Farms practicing free grazing, with a median of 0.31 PHC, would be able to meet their replacement needs, but those practicing semi-zero or zero grazing, with a median of 0.00 to 0.18 PHC, would have to purchase heifers to meet their herd replacement needs. 

Own produced feeds, purchased feeds, milk yields 

Indicators of quantities of own produced (ha∙TLU-1 of total land, maize planted, and napier planted) and purchased (Ksh∙TLU-1 of fodder and concentrates) feeds were estimated across categories of grazing systems (Table 3).  The medians indicate that own produced feeds, except for napier, were higher in farms practicing free grazing whereas purchased feeds were higher in farms practicing zero grazing, reflecting greater dependency on purchased feeds in farms practicing with increasing intensification of dairying.  Greater dependency on purchased feeds and planting of napier grass fodder in intensification of smallholder dairying has also been observed with smallholder dairy systems in India (Patil and Udo 1997) and in the Hindu Kush-Himayas in Asia (Tulachan 1999). The median milk yield per cow per day varied from 3 to 5 litres and was highest in farms practicing zero grazing and lowest in farms practicing free grazing (Table 3).  This reflects higher productivity of cows in farms practicing zero or semi-zero grazing compared to those practicing free grazing system, achieved with use of planted fodder and purchased feeds.


Table 3.  The proportion of heifers to cows (PHC), feed resources, and milk yields by grazing systems in smallholder farms in central highlands of Kenya

Variable

Grazing system

Number of farms

Mean

Standard deviation

Median

Proportion of

Free

227

0.40

0.53

0.31b,c

heifers to cows

Semi –zero

254

0.36

0.46

0.18b,d

 (PHC)

Zero

363

0.29

0.37

0.00c,d

Land available

Free

227

2.26

3.95

0.99c

 (ha∙TLU-1)a

Semi –zero

285

1.63

2.71

0.90d

 

Zero

425

1.09

3.08

0.59c,d

Maize planted

Free

190

0.25

0.38

0.13c

 (ha∙TLU-1)

Semi –zero

265

0.27

0.32

0.15d

 

Zero

417

0.18

0.27

0.08c,d

Napier planted

Free

187

0.08

0.15

0.00c

 (ha∙TLU-1)

Semi –zero

261

0.09

0.23

0.00d

 

Zero

406

0.13

0.19

0.06c,d

Fodder purchased

Free

190

396

1346

0.00c

 (Ksh∙TLU-1)

Semi –zero

265

499

1677

0.00d

 

Zero

417

1047

2389

83c,d

Concentrates

Free

190

2332

5009

0.00c

Purchased

Semi –zero

265

1419

4069

0.00d

(Ksh∙TLU-1)

Zero

417

3135

4352

1085c,d

Milk yields

Free

227

3.84

3.17

3.00b,c

 (L∙cow-1∙d-1)

Semi –zero

245

4.80

2.93

4.55b,d

 

Zero

323

5.57

3.56

5.00c,d

a TLU=1 for bull; 0.7 for cow; 0.5 for heifer and young bull; 0.2 for calves
Significant difference (Mann-Whitney rank-sum test p≤0.017) between: bfree and semi-zero grazing; cfree and zero grazing; dsemi-zero and zero grazing


 The PHC in relation to own produced and purchased feeds, and milk yields

 

Table 4 shows the logistic regression estimates for the associations between PHC and the selected measures (own produced and purchased feeds) of the level of intensification in the dairy farm.  For the analysis, complete data was available for 706 farms out of the 987 farms with cattle.  Lower PHC was significantly associated with zero grazing (p<0.0107) compared to free grazing, hectares of maize planted per unit of cattle (p<0.0003), and cash spent on purchase of fodder per unit of cattle (p<0.0818). The available land per unit of cattle had positive and significant associations with PHC (p<0.0248).  Hectares of napier grass planted and cash spent on purchase of concentrates per unit of cattle were negatively but not significantly associated with PHC.


Table 4.  Logistic regression estimates for the associations between the proportion of heifers to cows (PHC) and measures of the level of intensification in smallholder farms (n=706) in the central Kenya highlands

Independent variables

Estimate

Standard Error

P -value

Intercept

-0.834

0.109

0.0001

Gazing system

 

 

 

Free

Reference

 

 

Semi-zero

-0.040

0.111

0.7210

Zero

-0.303

0.118

0.0107

Land available, ha∙TLU-1

0.052

0.023

0.0248

Maize planted, ha∙TLU-1

-1.557

0.432

0.0003

Napier planted, ha∙TLU-1

-0.345

0.366

0.3460

Fodder purchased, Ksh TLU-1

-0.0001

0.0001

0.0818

Concentrates purchased, Ksh∙TLU-1

-∞

0.3859


The lower PHC in farms that practiced zero grazing compared to those that practiced free grazing implies that farmers who adopt zero grazing opt to keep herds with low proportion of heifers relative to cows.  This is likely due to greater dependency on planted napier grass fodder and use of higher levels of purchased feeds (Table 3), which would make heifer rearing too costly in terms of feeding.  The negative associations between PHC and purchased fodder, and napier grass and purchased concentrates (though not significant), indicates that the increased use of these feeds tend to be associated with high proportion of cows relative to heifers in smallholder dairy farms.  Keeping the proportion of cows higher relative to heifers is expected to contribute to increasing farm productivity because cows are the most productive animal class.  This is likely a rational choice considering that farm sizes are small, a median of 1.2 ha per household (Bebe et al 2003) in which food crops also must be grown.  As a result, the area allocated for growing napier is generally inadequate for the total herd in the farm.  In the central highlands of Kenya, smallholders allocate an average of 0.2 ha to napier growing (Staal et al 1998), which is half less the recommended area of 0.40 ha for a cow and heifer in these farming systems (Valk 1990). 

 

An increase by one ha of land per unit of cattle was estimated to result in an additional 0.05 PHC in the farm (Table 4).  Land available (ha∙TLU-1) reflects the availability of grazing pastures.  The availability of grazing pastures reduces the dependency on purchased feeds and planted fodder, hence cheaper for rearing of replacement stock.  The negative association between hectares of maize planted and PHC result from reduction in the available grazing pastures because maize is planted mainly as food for humans, but become available for cattle feeding as by products (stover and leave thinning. 

Effect of grazing systems and breed on the age at first calving 

Table 5 shows the effect of type of breed and grazing system on the probability of first calving.


Table 5.  Estimates of hazard ratios for age at first calving for cows stratified by breed and grazing system in smallholder farms in the Kenya highlands

Factor

Number of records

Regression coefficient

Hazard ratio (HR)

95% confidence interval of HR

Breed

 

 

 

 

   Friesian

498

1.09 ± 0.13

2.97

2.40, 3.66

   Ayrshire

287

1.22 ± 0.14

3.40

2.71, 4.25

   Guernsey/Jersey

150

1.11 ± 0.22

3.03

2.11, 4.36

   Bos indicus

348

ref.

ref.

 

Grazing system

 

 

 

 

   Free

394

ref.

ref.

 

   Semi-zero

460

0.65 ± 0.13

1.91

1.53, 2.38

   Zero

429

0.82 ± 0.17

2.27

1.071, 3.02

Model overall score 219.95, df=11, P=0.00)


The hazard ratios show that the probability of first calving was significantly higher in Bos taurus dairy breeds (AY, GJ and FR) than in Bos indicus cattle (ZB), and higher in zero- and semi-zero-grazing than in free-grazing system.  The AY, GJ and FR breeds, respectively, were 3.41, 3.03 and 2.97 times more likely to attain first calve earlier than were the ZB.  Heifers raised in zero- and semi-zero-grazing were 2.21 and 1.97 times, respectively, more likely to calve earlier than were heifers raised in free-grazing systems (Table 5).  Farmers practising zero- and semi-zero-grazing systems achieved earlier age at first calving likely by planting more napier fodder and purchasing more feeds, both concentrate and fodder (Table 3).

 

Conclusions 

 
Acknowledgements

 

The first author was supported by a research grant from The Netherlands Foundation for the Advancement of Tropical Research-WOTRO.  The authors acknowledge the support of the Smallholder Dairy (R&D) Project (SDP) of the Kenya Ministry of Agriculture and Rural Development, the Kenya Agricultural Research Institute (KARI) and the International Livestock Research Institute (ILRI) for this study.  SDP was funded by the UK Department for International Development (DFID) for the benefit of developing countries.

 


References

 

Bebe B O, Udo H M J, Rowlands G J and Thorpe W 2003 Smallholder dairy systems in the Kenya highlands: cattle population dynamics under increasing intensification. Livestock Production Science 82: 211-221

 

Conelly W T 1998 Colonial Era Livestock Development Policy: introduction of improved dairy cattle in high potential farming areas of Kenya. World Development 26: 1733-1748

 

Das S M, Wiktorsson H and Forsberg M 1999 Effects of calf management and level of supplementation on milk yield and calf growth of Zebu and crossbred cattle in the semi-arid tropics. Livestock Production Science 59: 67-75

 

De Jong R 1996 Dairy stock development and milk production with smallholders. PhD Thesis, Department of Animal Production Systems, Wageningen Agricultural University, The Netherlands 303 pp

 

Gitau G K, McDermott J J, Adams J E, Lissemore K D and Waltner-Toews D 1994a Factors influencing calf growth and daily weight gain on smallholder dairy farms in Kiambu district, Kenya. Preventive Veterinary Medicine 21: 179-190

 

Gitau G K, McDermott J J, Waltner-Toews D, Lissemore K D, Osumo J M and Muruiki D 1994b Factors influencing calf morbidity and mortality in smallholder farms in Kiambu District of Kenya. Preventive Veterinary Medicine 21: 167-177

 

Gitau G K, Perry B D, Katende J M, McDermott J J, Morzaria S P and Young A S 1997 The prevalence of serum antibodies to tick-borne infections in cattle in smallholder dairy farms in Murang’a district, Kenya: a cross-sectional study. Preventive Veterinary Medicine 30: 95-107

 

Gitau G K, Perry B D and McDermott J J 1999 The incidence, calf morbidity and mortality due to Theileria parva infections in smallholder dairy farms in Murang’a district, Kenya. Preventive Veterinary Medicine: 39, 65-79

 

Lanyasunya T P, Wekesa F W, de Jong R, Udo H, Mukisira E A and Ole Sinkeet N S 1999 Effects of a calf rearing package introduced to smallholder dairy farms in Bahati division, Nakuru district, Kenya.  pp. 450- 457 In: Proc. 6th Biennial KARI Scientific Conference held on 9-13 November 1998, Nairobi, Kenya

 

MoA 1998 Kenya’s Dairy Development Policy: Towards the development of a sustainable dairy industry. Ministry of Agriculture, Nairobi, Kenya. 49 pp

 

Omore A O, McDermott J J and Gitau G K 1996 Factors influencing production on smallholder dairy farms in Central Kenya, pp 370-379 In: Proceedings of the. 5th Scientific Conference of the Kenya Agricultural Research Institute, October 14-16, 1996, Nairobi, Kenya

 

Patil B R and Udo H M J 1997 The impact of crossbred cows in mixed farming systems in Gujarat, India: milk production and feeding practices, Asian-Austrialian Journal of Animal Sciences 10: 253-259

 

Reynolds L, Metz T and Kiptarus J 1996 Smallholder dairy production in Kenya, World  Animal Review  87 (2) 66-72 http://www.fao.org/docrep/W2650T/w2650t07.htm

 

SAS Institute Inc 1996 Release 6.12  SAS Institute Inc Cary NC, USA

 

Siamba D N, Lokwaleput  I, Onyango  T A, Nampaso  J and Nyakira B S  1999  Ticks and tick-borne diseases: Epidemiology, control practices and prospects for integrated ticks and tick-borne disease management in Bahati, Nakuru. pp. 34-46 In: Proceedings 6th Biennial KARI Scientific Conference held on 9-13 November 1998 Nairobi, Kenya

 

Staal S B, Chege L, Kenyanjui M, Kimari A, Lukuyu B, Njubi D, Owango M, Tanner J Thorpe W and Wambugu M  1998 Characterisation of dairy systems supplying the Nairobi milk market: A pilot survey in Kiambu district for the identification of target groups of producers. Project report of the MoA/KARI/ILRI Smallholder Dairy (R and D) Project. Nairobi, Kenya 85 pp

 

Stotz D 1979 smallholder dairy development in past, present and future in Kenya. PhD Thesis, Hohenheim University, Germany

 

Thomas D C, Siemiatycki J, Dewar R, Robins J, Goldberg M and Armstrong B G 1985 The problem of multiple interference in studies designed to generate hypotheses, American Journal of Epidemiology 122: 1080-1095

 

Tulachan  P M  1999 Livestock production and management strategies in the mixed farming areas of the Hindu Kush-Himalayas, Asia, pp 123-134 In: Tulachan P M, Mohamed-Saleem, M A, Maki-Hokkonen J, Partap T (Editors). Contribution of livestock to mountain livelihoods, International Centre for Integrated Mountain Development, Kathmandu, Nepal

 

Valk Y S 1990 Review reports of the Dairy Evaluation and Advice Forms (DEAF) surveys during 1989. National Dairy Development Project (NDDP/M41/200) Ministry of Livestock Development, Nairobi, Kenya, 69 pp



Received 7 June 2007; Accepted 10 November 2007; Published 1 February 2008

Go to top