Budgeting and allocation of feed resources

P.N. de Leeuw, B.H. Dzowela and R. Nyambaka
International Livestock Centre for Africa
P.O. Box 46847, Nairobi, Kenya

Introduction
Materials and supplies
Results
Discussion
References

Introduction

In most of the semi-arid Eastern Kenya, small-scale farmers cultivate maize, beans, cowpeas and pigeonpea as well as keeping cattle, goats and sheep. The majority of farms are between 2 and 15 ha in size and most land is privately owned. Cropped land increases with farm size from 1.5 ha to 5 ha, the remainder of the land being left for grazing (Rukandema, 1984; Tessema et al, 1985). Annual rainfall is between 600 and 900 mm per year falling in two seasons with a growing period of 50-80 days each (Jaetzold and Schmidt, 1983; Stewart and Kashasha, 1984). Cattle are kept for milk, traction and cash sales while small stock are sold for cash. Cattle are the most important, comprising 75-85% of the total livestock mass. Stocking rates are high and decrease from 1.5 TLU/ha (Tropical Livestock Units of 250 kg) for a 5/ha farm to 1.0 TLU/ha for a 10 ha farm. Forage from rangeland and crop residues are the principal feed resources. Except in good rainfall seasons demand for livestock feed are likely to exceed supplies resulting in overstocking, land degradation and low productivity (de Leeuw, 1988).

On-farm research has shown that the fodder supply in smallholder farms can be increased by 1) establishment of grass-legume leys, 2) planting of grass and legume forages and 3) better utilisation of crop residues. Household welfare and cash income was improved by channeling most of the planted forage and high quality fodder to one or two crossbred cows for milk production (Tessema et al, 1988). Livestock productivity can be further enhanced by more efficient utilisation of feed resources through proper budgeting and allocation of feed resources to the various livestock enterprises maintained on farms.

There is, therefore, a need to develop efficient feed budgeting and allocative procedures that can assist in manipulating feed resources to the best advantage of the farmer and in line with his priorities for his various livestock enterprises. The objectives of such procedures are 1) to access the quantity and quality of all feed components, and 2) to determine the production goals of each enterprise and allocate feed in relation to their requirements within the framework of overall feed supplies.

Materials and supplies

To develop procedures for feed budgeting and allocation, the following assumptions were made:

1. The existence of an average 'model' farm with a fixed land area and a stable land-use pattern resulting in known areas under crops and rangeland.

2. A fixed monthly growth rate for pastures and planted forages and a fixed yield of crop residues at harvest together with quality parameters for each component.

These assumptions are based on actual data and the conditions under which smallholder farmers operate (de Leeuw, in press). The projected 'model' farm has 9 ha of land, of which 3 ha was cropped, one ha was allocated to planted forages and legumes, and 5 ha was used as rangeland (Table 1). All farmed land had maize and one third was sole-cropped and the remainder was intercropped with beans, cowpeas and pigeonpea. Estimated seasonal yields are derived from regular surveys data measured during the 1987/8 period in a number of sample farms. In total the 9 ha farm produced 32 t DM of feed or 3.7 t DM/ha/annum over two growing seasons (Table 1).

Table 1: Land, Crops and Feed Resources of the 'model'

Type	ha	Yield t DM/ha		Total yield (t DM/ha)
		1st Season	2nd Season
Rangeland	5.0	1.7	1.5	16.0	(49%)
Maize Residues	3.0	1.7	1.7	10.2	(31%)
Beans and Cowpea	1.0	0.6	0.5	1.1	(3%)
Pigeonpea	1.0	-	0.6	0.6	(2%)
Planted grass	0.5	3.0	2.6	2.8	(9%)
Planted legume	0.5	1.9	1.7	1.8	(6%)
Total for farm	11.0	16.6	15.9	32.5	(100%)

Source: de Leeuw (in press).

Quality parameters were restricted to crude protein content and digestibility; these were estimated monthly for rangeland and planted forages. Since crop residues were produced at fixed times after harvest, only two quality classes were relevant: either when fed or grazed immediately after harvest or when fed later leading to a small reduction in quality (Tables 2 and 3).

Table 2: Crude protein (CP) content and digestibility (DIG) of rangeland, planted grass and legume forage (in % DM ^-a)

	Rangeland		Grass		Legume
Period	CP	DIG	CP	DIG	CP	DIG
Early season	12	65	12	65	20	65
Mid season	8	55	10	60	17	60
Late Season	6	50	8	55	15	55
Dry season	5	45	6	50	15	50

^a/means calculated from estimated monthly values (de Leeuw, unpublished)

Table 3: Crude protein (CP) content and digestibility (DIG) of crop residues (in % DM).

Harvest time	Early		Late
Crops	CP	DIG	CP	DIG
Maize	6	50	4	45
Cowpea/beans	12	60	10	55
Pigeonpea	12	60	10	55

Source: de Leeuw (in press)

Four main livestock enterprises were kept on the 'model' farm. Crossbred cattle for milk production, oxen for traction, a few zebu cattle and 15 small-stock (goats and sheep). The total livestock mass was 2440 kg or 9.8 TLU, which converts to a stocking rate of 1.1 TLU/ha (Table 4). Their daily feed intake has been taken as uniform across enterprises at about 3% of their liveweight and was kept constant during the year. Based on the total stock weight of all enterprises, daily and annual feed requirements were estimated at 73 kg and 26.3 t DM respectively (Table 5). Thus, the 'model' farm produced sufficient feed for its stock. Given the distribution of body mass across the livestock groups, the crossbreeds required 40%, the oxen 29% and the remaining 31% was needed for the zebu and small-stock (Table 5).

Table 4: Livestock holdings and daily feed requirements.

Class/type	Number	Weight (kg)		Daily Intake
		Unit	Total	Kg DM
CB Cows	2	350	700	21
CB heifer	1	200	200	6
CB calf	1	70	70	2
All crossbreds	4		970	29
Work oxen	2	350	700	21
Zebu cow	1	250	250	8
Zebu heifer	1	150	150	4
Zebu calf	1	70	70	2
All zebus	3		470	14
Goats	10	20	200	6
Sheep	5	20	100	3
All smallstock	15		300	9
All stock	24		2440	73

Source: de Leeuw (in press)

Table 5: Monthly and annual feed requirements of different stock classes.

	Requirements, t DM
Class/type	Month	Year	% of total
Crossbreds	0.87	10.5	40
Work oxen	0.63	7.6	29
Zebus	0.42	5.0	19
Smallstock	0.27	3.2	12
Total	2.19	26.3	100

Source: Calculated from Table 4.

The budget started on 1 October in one year and ended on 30 September in the following year. For each feed component, monthly changes (in t DM/ha) were determined and cumulative totals were entered on the spread sheet. For instance, DM yield of rangeland on 1st October was 1 t DM/ha as a left-over of the previous year, which increased in monthly steps to a maximum of 2.7 t (i.e. a 1.7 t increment in line with data in Table 1) at the end of the growing season in February, declined by 15% during the short dry season due to seed and leaf fall as well as termite attack and insect damage. During the second season, yield increased by 1.5 t/ha falling progressively by 5% per dry month up to the end of the 'model' war. Cumulative yields for forage grasses and legumes were assessed in a similar manner.

Crop residues are easier to assess due to their fixed harvest times and outputs. They are harvested twice during January-February and May-June; as farmers did not harvest all crops at once, availability was spread over several months. For each monthly entry, protein content and digestibility coefficients were added (Table 3).

The allocation of feed sources to each livestock enterprise was expressed as a percentage of their total monthly feed consumption (Table 6). The four groups were reduced to three by combining small-stock with the zebu cattle herd. As there was no shortage of total feed, the allocation was based on quality with a priority ranking from crossbred, through work-oxen to the zebu/small-stock group.

Table 6: Diet composition (%) of three livestock enterprises over three-monthly periods.

Period	Oct-Dec	Jan-Mar	Apr-Jun	Jul-Sep	Annual Mean
Feed Source	Crossbreds
Rangeland	54	30	43	10	34
Maize	13	29	30	43	29
Grain legumes	7	7	3	10	7
Forage grass	23	17	17	20	19
Forage legumes	3	17	7	17	11
	- - -	- - -	- - -	- - -	- - -
	100	100	100	100	100
	Work oxen
Rangeland	83	50	56	27	54
Maize	10	37	34	53	33
Grain legumes	-	7	4	7	5
Forage grass	4	3	3	7	4
Forage Legume	3	3	3	6	4
	- - -	- - -	- - -	- - -	- - -
	100	100	100	100	100
	Zebus and smallstock
Rangeland	89	50	53	30	55
Maize	11	43	40	53	37
Grain legumes	-	7	7	17	8
Forage grass	-	-	-	-	-
Forage legumes	-	-	-	-	-
	- - -	- - -	- - -	- - -	- - -
	100	100	100	100	100

Source: Calculated from assumed monthly diet composition.

The monthly consumption of each feed component was calculated and subtracted from the cumulative total. It was assumed that harvesting did not affect subsequent growth; hence for each month the balance consisted of the quantity of feed at the end of the previous month and growth minus consumption and losses during the current month. A number of test runs were done to adjust diet composition to available feed components to avoid negative values. Thus, this part of the procedures was done manually and no computer programme has as yet been developed. For the final run, the crude protein and digestibility contents of each of the three diets were determined on a monthly basis.

Results

In principle, the monthly feed allocation was based on the availability and the quality of each feed source with adaptation to priority rating for each livestock group. However, since forage from rangeland and maize stover constituted 80% of the total feed, all groups had to rely on these two sources as the main components of their diet. Maize residues were available only during two distinct periods and as large-scale storage was thought unfeasible, most was allocated soon after harvest.

Diet composition for the three livestock enterprises are summarized over four 3-monthly periods (Table 6). Crossbred cattle relied more on rangeland in the wet than in the dry seasons and they, together with the oxen, were the main beneficiaries of planted forages, which constituted 30% and 8% of their respective annual diet. When the quality was high, these forages were cut and carried and fed after grazing to promote greater daily intake, whereas during the dry season they also assisted to improve diet quality. The residues fed during the first three months were stored feeds from the previous year (Table 9).

Work oxen were given access to some planted forage and legume residues in July-August to improve their body condition for ploughing in the dry season and again in October-November when tillage was required after the first rains. No planted feed was supplied to the zebus and small-stock and they relied entirely on natural pastures and crop residues (Table 7).

Table 7: Annual allocation of feed resources for crossbreds, work oxens and other stock.

Source	Crossbred	Oxen	Others	Total use
	% of total available
Rangelands	22	26	28	76
Maize residues	30	24	30	84
Legume residues	41	21	38	100
Forage grass	71	11	-	82
Forage legume	64	17	-	81

Source: derived from Tables 4 and 6.

Table 8: Average crude protein content (CP%) and digestibility coefficient (D%) in diets of three livestock enterprises over 3-monthly periods.

Enterprises	Oct-Dec	Jan-Mar	Apr-Jun	Jul-Sep	Annual Mean
Crossbreds	(CP%) 9.6	7.3	8.2	8.0	8.3
	(D%) 58	50	54	51	53
Work oxen	(CP%) 9.2	6.0	7.3	6.0	7.1
	(D%) 58	48	53	48	52
Zebus and smallstock	(CP%) 9.2	5.3	7.0	6.0	6.9
	(D%) 58	47	52	48	51

Source: Derived from Table 2, 3, 4 and 6.

Table 9: Annual feed utilisation and balance (t DM).

	Start 1/10	Livestock use	lost	end 30/9
Rangelands	5.0	12.3	4.1	4.6
Maize residues	0.9	8.6	2.4	0.1
Legume residues	0.2	1.7	-	-
Forage grass	0.5	2.3	0.5	0.5
Forage legume	0.2	1.4	0.4	0.2
Total	6.8	26.3	7.4	5.4

Source: Derived from Table 1, 5 and 6.

In terms of protein content, the crossbreds received the best diet, followed by the work oxen. Differences in diet protein between groups were pronounced, because protein content is easy to manipulate as several sources of high-protein feed were available, be it in limited quantity (Table 7). Differences in digestibility were much less pronounced and therefore providing crossbreds with a higher digestible diet was difficult to accomplish. The overall value in feeds given to crossbreds was only 2% higher than that in the diets of zebu cattle and small-stock.

On an annual basis about 26 t of feed was used by the three livestock groups or about 80% of the total annual production of 32 t (Table 1); another 23% was assumed 'lost' to other consumers (termites, wildlife, etc.) and decomposition (Table 9). Overall, utilisation was slightly higher than production with the result that less feed was carried over at the end of the 'model' year than at its start. However, for rangeland and planted forages sufficient herbage was left standing to insure continued growth in the following year.

Discussion

The feed supply of the model farm demonstrated that intensive livestock production was possible in normal years provided one hectare was set aside for planted forages and crop residues were fully utilised. The timing of the allocation was crucial and geared to specific animal requirements. In addition, sufficient herbage was left -for the next year to avoid overgrazing and to secure sustained feed output. The assumed stocking rate was realistic and 1.1 TLU/ha corresponded to the actual rates recorded in the sample farms and in other parts of the semi-arid zone (de Leeuw, 1988; Jaetzold and Schmidt, 1983; Tessema et al, 1985).

Further fine-tuning of the 'model' is required particularly with respect to the quality parameters of the different feed resources. Since monthly inputs were used, it is possible to adjust feed allocation more closely to the animal requirements by taking into account calving dates, stage of lactation of crossbred, zebu and small-stock, or if required, including feed for fattening stock for sale. Iterative procedures can be included in the programme so that feed resources are allocated automatically according to set priority ratings of quantity and quality as was done for feed allocation in smallholder farms in Bangladesh (Hermans, 1986).

The feed allocation programme is flexible and can be adjusted for farm size, cropping pattern, herd size and composition and production goal of farmers as well as changes in yields and quality of each feed components. It can also become a tool to test the effects of feed resources utilisation on farm output when it is linked with holistic economic farm models.

References

Hermans, C. 1986. Adaptation of the Kahn model for a mixed farming system in southeastern Asia. In: N. de Ridder, H. van Keulen, N.G. Seligman and P.J.H. Neate (eds). Modelling of extensive livestock production systems, 189230. ILCA, Addis Ababa, Ethiopia.

Jaetzold, R. and Schmidt, H. 1983. Farm management handbook of Kenya: Vol II, East Kenya, Ministry of Agriculture Nairobi Kenya. de Leeuw, P.N. 1988. Livestock and fodder resources of smallholder farms in semi-arid eastern Kenya. Draft Report, ILCA, Nairobi, Kenya.

Rukandema, M. 1984. Farming systems of semi-arid eastern Kenya: a comparison. E. Afr. agric. For. J. 44: 422-435.

Stewart, J.I. and Kashasha, D.A.R. 1984. Rainfall criteria to enable response farming through crop-based climate analysis. E. Afr. Agric. For. J. 44:58-79.

Tessema, S., Emojong, E.E. Wandera, F.P. and Nderito, M. 1985. Features of traditional farming systems as they affect livestock production. A case study of 18 small-scale farms in the dryland areas of the Eastern Province of Kenya. Dryland Farming Research and Development Project, Document No. 6, Machakos, Kenya.

Tessema, S., Emojong, E.E., Maluti, M. and de Leeuw, P.N. 1988. A strategy of livestock research adapted to semiarid small-scale mixed farming systems; Katumani experience. In: Dzowela, B.H. (ed). Proc. of 3rd PANESA workshop, p. 378-390, ILCA, Addis Ababa.