Chapter VII: State of the Art and Tools - Silviculture and Silvo-Pastoralism

1. Definition and evolution of the system
2. Silvicultural operations
3. Harvesting
4. Silvo-pastoralism

This chapter evaluates achievements over the past 10 to 15 years in both these areas. The first part of the chapter reviews the various forestry operations and forms of harvesting, except for pollarding which is dealt with in Chapter VII paragraph 4.2. With regard to silvo-pastoralism, which is a recent science undergoing all-out development (because foresters, shepherds and livestock farmers only knew about relations of force and exclusion until quite recently), only the main principles of exploitation and mention of a few improvement schemes will be dealt with.

1. Definition and evolution of the system

The silvicultural system is a set of rules applied to a forest stand in order to ensure its renewal.

The three basic systems are: the coppice system; the coppice-with-standards system; and the high forest system. The coppice system is made up of stump and/or root sprouts, originating from “rejuvenation” cuttings, which constitute a vegetative reproduction system. The ‘high forest’ is a stand made up of trees directly grown from seed on site. The ‘coppice-with-standards’ is a mixed system designed to perpetuate stands with trees of which some have originated from seeds and others are derived from vegetative regeneration. The ‘coppice-with-standards’ system involves the removal at a fixed, short-rotation period, of all coppice trees except for some carefully selected standards which constitute a distinct storey of the high forest, and the partial felling of the ‘reserved trees’ (of seedling origin) selected at each rotation. The mixed stand resulting from this comprises a coppice under a high forest of multiple age classes. The high forest ‘selection method’ is a stand of mixed, scattered individuals and/or groups of trees of multiple ages and dimensions.

Other systems can also be envisaged, such as ‘coppice selection’ method, which is a specific treatment in which only the selected individuals of merchantable dimensions are removed at each felling operation, which produces uneven-aged sprout stands. One speaks of the ‘simple coppice’ system when a stand is clear-felled in order to obtain an even-aged rejuvenated stand, made up of stumps and/or root sprouts. Conversely, the ‘coppice of two rotations’ system is used when certain shoots are set aside during a logging operation in order to take part in the next revolution, in order to produce wood of larger dimensions (Métro, 1975).

In the dry deciduous forests, the coppice may be overshadowed by a high forest. In open woodlands and a fortiori in tree steppes and savannas, this is not the case. This is why some authors utilize such expressions as ‘coppice selection’, coupe par furetage (in French) or selective felling which, while clearly expressing the idea of crop selection, do not take into consideration the method of reproduction, which is not exclusively vegetative. Mazoyer (1992) prefers the French expression prélèvement par furetage-jardinage, which may be approximately translated as ‘selective extraction’.

There is no one forestry term which is perfectly suitable to define all the different multiple-use forms of managing tree savanna formations or open woodlands with both sexual and asexual regeneration, because it involves managing a forest area for the benefit of so many different beneficiaries (livestock, farmers, foresters, bee-keepers, hunters, environmental and territorial managers).

From the strict forestry viewpoint, the development of the silvicultural system has come about gradually. Initially, forest managers very often proposed the simple coppice system, and more rarely a pseudo ‘coppice-with-standards’ system in regions with higher rainfall, and more recently they have proposed the ‘coppice selection’ system, particularly in the open woodlands and savannas.

a) The simple coppice system and the coppice-with-standards system

The simple coppice system has been in use for a long time in tropical dry woody formations whose main extracted products is firewood. In the West-African Sahelo-Sudanian formation, the simple coppice rotation was often established at 8 to 15 years and even 20 years.

For the management of the miombo formations under the coppice-with-standards method, Fanshawe (in CSA/CCTA, 1959), recommended to remove the coppice material every 40 years and to harvest the standards (reserves) every 60 to 100 years, depending on the quality of the produce. In the open woodlands in the Democratic Republic of the Congo (miombos), “attempts were made to generate well-formed coppices by cutting trees at ground level. During clear-felling operations, successive early prescribed fires are then ignited until the coppice vegetation is dense enough to prevent grass development. A subsequent thinning out enhances the production of the best stems. Growth in diameter declines after the twentieth year.” (Malaisse, 1979)

Subsequent management operations have kept to these principles: the undergrowth is generally made up of less valuable species which are mainly used for fuelwood (Combretaceae, Acacia, etc.). In the upper storeys are the timber species (Bombax costatum, Pterocarpus erinaceus, Terminalia macroptera, Anogeissus leiocarpus) which are harvested following rotations of several dozen years (Catinot, 1985; Arbonnier and Faye, 1988).

In Côte d’Ivoire the Badénou Forest is currently being managed for timber. Tree savannas and savanna woodlands are subjected to thinning, while shrub savannas are regenerated by clear-cutting in order to convert them into high forests (Part Four, Case Study 2).

b) The coppice selection system and selective cutting

Since the mid-1980s, within the framework of integrated forest management, fruit, fodder or medicinal species, which the traditional systems protected, have been taken into consideration. Nouvellet (1993) has emphasized an idea that has become increasingly more popular in the past 5 to 10 years in Burkina Faso, namely that the clear-felling of coppices is not suitable, as it creates a number of regeneration problems for certain useful trees (such as the Butyrospermum paradoxum, whose seed production is only abundant after about 20 years). The impact of repeated felling on regeneration by stump or root sprouts is, moreover, very little known.

Little by little, in West Africa, forest managers have gradually moved away from the simple coppice system and from the coppice-with-standards regime, with clear-cutting of the coppice and felling under long rotations of standards (reserves), to a selective coppice-with-standards system, that maintains some reserves (Catinot, 1994), or to either selective cutting or the coppice selection system (Kaboré and Ouedraogo, 1995) which are known as the ‘Say method’ (Peltier et al., 1994). The latter are based on the following principles:

* a group of species characterized by one dominant type of use (fuelwood, timber, fruit trees, fodder crops, etc.) is established;

* for each group of species, a particular management method is chosen depending upon the minimum merchantable diameters, more or less long periods of rotation, and protection and conservation systems;

* generally speaking, the dying-back or dead individuals are systematically harvested;

* the fragile zones in which deforestation can lead to degradation or pollution (low-lying backwaters, slopes, etc.) are protected; and

* supporting complementary silvicultural activities.

To be noted that these new approaches make it possible to maintain a minimum cover in the compartments being harvested, thereby reducing erosion and the decline in fertility. Moreover the fact that certain trees are now being considered for non-wood forest products means that the local people are involved in managing the woodlands to suit their own interests.

Mazoyer (1992) has nevertheless pointed out that this type of harvesting consisting of deadwood gathering, small wood and large wood selection and pruning branches off park trees was common practice in Niger and was wholly appropriate there so long as wood extraction did not exceed production. He therefore considered that “this form of harvesting has already had its day in all the underharvested areas” around Niamey. However, in the overharvested zones this selective coppice system could and should be maintained.

Table 16 compares management proposals by Nouvellet (1992), Catinot (1994) and Peltier et al. (1994) in the Sudano-Sahelian zones.

Table 16: Comparing three management proposals under 600-800 mm of annual rainfall

Silvicultural system	Nouvellet (1992)	Authors Catinot (1994)	Peltier et al. (1994)
Coppice	- simple coppice - 8<DME<15 cm - short rotation (7_2 years) for fuelwood, fodder and other uses. - Medium rotation (14_2 years) for timber and ‘utility’ wood	- Coppice selection system - Rotation 7-8 years and selection of clump sprouts for ‘utility wood’ under 14-16 years rotations	- coppice selection system or exploitation felling: DME 6 cm at the base for Guiera senegalensis and Combretum nigricans and C. glutinosum.
High forest	- long revolution, in excess of 21 years, for timber.	- revolution linked to a 50 years rotation (timber).	- DME 35 cm for exploitation felling of Pterocarpus lucens, etc. - felling of dying-back trees.
Fruit species	- total protection with selective thinning by removal of dying-back trees.	- protection - no logging - collecting rights during fruiting period	- protection until death of trees
Fodder species	High-bole trees protected Controlled pruning	-	- trimming out tolerated above 2 m high

DME = minimum merchantable diameter (from the French Diamètre Minimum d’Exploitabilité)

2. Silvicultural operations

2.1 Thinning and release cutting
2.2 Pruning
2.3 Plantations
2.4 Direct seeding
2.5 Fire-breaks
2.6 Total protection versus early burning

2.1 Thinning and release cutting

Thinning can be carried out both on standing trees and on large clump sprouts, in order to increase production. The term ‘release cutting’ will only be used for specific types of thinning, relating for example to young sprouts (or young saplings) in order to improve individual growth.

* Thinning

The principle of thinning out is to encourage the development of certain trees considered to be valuable, usually for economic purposes, by eliminating neighbouring ones. This technique can only be used in closed stands. One must also be sure that the investment in terms of time and manpower guarantees a reasonable economic return in terms of the quality and quantity of the volumes harvested. Lessons learned from plantation thinning operations lead to think that under advanced thinning, the trees retained react weakly, while early thinning produces the best responses.

While the techniques and effects of thinning are comparatively well-known in the case of plantations, their application to natural stands has not been studied in depth. Indeed, the density of the trees in dry tropical formations (except for certain dry deciduous forests) does not allow thinning operations. These are very exceptionally and rarely justified in economic terms.

In a natural forest, thinning is always to the detriment of very common species as well as species for which there is a low demand (wood, fruit, leaves, roots, etc.). In the Tiogo and Laba (Burkina Faso) reserved forests, the not widely used species Entada africana, Annona senegalensis and Gardenia ternifolia, are felled before fruiting and the branches are then laid out on the bare ground to serve as seed traps and therefore encourage the sexual regeneration of other woody and herbaceous plants (Nouvellet et al., 1995).

It could also prove to be useful to thin out species which are not sought after locally during the least favourable season for their vegetative multiplication or at a height, to be established, which would inhibit shooting. Conversely, in order to encourage root suckers or sprouts on multiple-use species, thinning operations could be planned during the most favourable period.

Jackson et al. (1983) considered that managing the tree cover through intermediate felling or regeneration cutting is not realistic in most of the Sahelian countries. Resorting to selective thinning, in order to encourage specific species, demands a great deal of know-how and at the present state of knowledge this is not possible (particularly with regard to the proportion for natural vegetative multiplication and the portion for sowing).

At Morondava (Madagascar), results have shown that selective thinning can double the growth in diameter of heliophilous species, particularly in the lower storeys or the intermediate storeys for at least five years (Deleporte, 1995).

* Release cutting

Coppice shoot selection applied to indigenous species (namely when natural mortality and browsing are no longer a threat) has rarely been investigated. The gains resulting from an artificial release cutting (in terms of stump volume yields or growth rates allowing shorter differed grazing periods) have virtually not been investigated.

Observations made in Mali over a period of one year, have shown that selected sprouts from timber species (Daniellia oliveri, Isoberlinia doka, Pterocarpus erinaceus), display 50 percent higher growth rates (Irkonanan, 1994). Thinning by release cutting seem to be very useful for improving a stand.

In Botswana, natural stands, virtually all mono-species, of Colophospermum mopane cover the region of Boteti in the central province. In a non-repetitive experimental design, three compartments were subjected to variable intensity release cuttings, to be compared with a control compartment (Table 17). All three release cutting intensities tested improved tree growth. Furthermore, the stumps of harvested trees produced many sprouts in the course of the year (Coe, 1992).

Table 17: Release cutting in four 10 × 10 m compartments in Botswana

Items		Compartment 1	Compartment 2	Compartment 3	Compartment 4
Original density		7900	8300	8600	8000
Density following release cutting (shoots/ha)		2800	3400	8600	1500
Percentage of initial plant material left standing		35.4%	40.9%	100%	18.7%
Diameter in cm at 10 cm above ground level
	- before release cutting	2.71	2.36	2.89	2.73
	- after release cutting	4.42	3.26	-	3.99
	- 11 months later	4.65	3.61	3.09	4.35
Height in metres
	- before release cutting	1.09	1.02	1.31	1.19
	- after release cutting	1.82	1.54	-	1.71
	- 11 months later	1.85	1.60	1.37	1.73
Basal area in cm2/ha
	- before release cutting	616.16	454.46	731.29	563.73
	- after release cutting	472.43	315.10	-	205.53
	- 11 months later	525.46	381.27	756.85	241.80

Average stem diameters, heights and basal areas of Colophospermum mopane, before, immediately after, and 11 months after variable intensity release cutting operations.

Many unanswered questions remain with regard to thinning and release cutting: the vitality and longevity of the stumps after several thinning operations, the yield expected after the release cutting of sprouts, the effect of felling on root suckers’ production by species. Furthermore, the effects of thinning or release cutting on grass production have not yet been studied.

2.2 Pruning

The different possibilities for cutting trees (some of which are extreme) vary according to the purpose of the primary harvest (Figure 17). Cutting-back makes it possible to use the simple coppice system, while topping is used to harvest branches at a given height. Pruning consists in cutting-off side branches level with the trunk in order to improve the shape of the bole and the quality of wood. In the dry tropical savanna zone, pruning, as generally conducted, does not aim at improving bole wood, eliminating crotches or improving bole straightness. It is a practice used to supply fodder and/or firewood and to reduce crop shade. Pollarding (trimming-out) consists in cutting-off tips of branches or young adventitious shoots on the main stem of a tree (see Chapter VII,4.2-e).

Topping, which in these regions very often involves the virtual removal of all the branches is also a common practice with a number of different objectives:

- fodder production (at the end of the dry season, this non-advisable technique gives the tree the shape of a half-open umbrella);

- fruit production (by controlling tree height to facilitate fruit gathering); and

- sprouts’ positioning (high enough to stay out of reach of browse or fire damage).

In northern Burkina Faso, productivity under Faidherbia albida cover is increased by undertaking periodic pruning operations. In Kenya some tribes top their trees in order to collect fuelwood and increase construction timber production (Shepherd, 1992).

In Latin America, the system of trimming-out the branches (for fuelwood) is very widespread and forms an integral part of the natural stand management system used by the small farmers (Maldonado Aguirre, no date; Parra Hake, 1985). In Mexico, Prosopis spp. have virtually all their branches pruned between ages five and seven. The trees are then left to rest for another four to five years in order to produce sturdy sprouts which are much prized for making charcoal. Lastly, they are cut flush with the ground when their diameter at 30 cm above ground reaches 15 cm.

To the forester, the purpose of pruning is to improve the quality of timber or to increase the production of fuelwood by selective branch cutting (uncontrolled cutting of branches destroys the wood potential). For the small farmer who is in most cases, not concerned about the impact on the wood quality of the bole, it is above all a means of rapidly obtaining sticks. This variety of uses may constitute yet another difficulty when trying to conduct research into the effects of pruning. This is a vacuum which should be filled in the near future by, for example, taking greater interest in the possibility of pruning fruit and fodder trees, and by changing the legislation because the practice is often prohibited by law in most cases.

2.3 Plantations

Many management plans contain a chapter devoted to reforesting cleared areas. Before discussing the silvicultural techniques, it should be recalled that the first condition for success is to ensure that the species is suited to the environment.

Seedling production in nurseries is now under full control in the case of a great many species. The use of polythene bags and more recently of vermiculite clods has become widespread and has developed at the expense of more costly containers. The use of bare-rooted seedlings or the planting of ‘stumps’ (barbatelles in French) (ground grown seedlings, the main stem of which is cut at 2.5 cm from the collar, sometimes at 30-40 cm, while the main root is kept some 20-25 cm in length) are costly techniques that demand particular care from nurserymen. The use of striplings (seedlings with an intact stock and a maximum of roots, planted leafless) is not widely used in dry zones. In Senegal, it has occasionally been used for some species (Khaya senegalensis, Azadirachta indica) which do not withstand the ‘stump’ practice. In West Africa, the collar of certain young plants is buried in order to increase the number of roots.

It is important to differentiate between reforesting on cleared agricultural land from plantations in forest openings. The latter are often gently sloping depressions, where nothing should be planted at all because they become flooded in the rainy season. Improvement planting using the strip method have proven to be costly in terms of upkeep (Casamance) and have not been able to be carried through with sufficient care.

In Mali, the plantations in the forest ‘openings’ seem to provide better results than using the strip method (Cuny, 1993), with a better survival rate and better growth (Table 18). These results however, must be used with care.

Table 18: Average heights and growth rates by species according to plantation technique (forest ride or patch) - Farako Site, 1988 plantation

Species	Initial number of plants		Height in cm				Height increase (cm) 10.1990-11.1991
			October 1990		November 1991
	Strip	Opening	Strip	Opening	Strip	Opening	Strip	Opening
Afzelia sp.	90	107	40.2	68.1	45.3	91.5	5.1	23.4
Daniellia sp.	70		14.2	-	Failure	-	-	-
Khaya sp.	80	97	90.8	124.4	124.3	177	33.5	52.9
Parkia sp.	95	80	16.1	27.1	18	36.4	1.9	9.3
Pterocarpus	75	59	10.8	21.6	16.6	25.2	5.8	3.6

Opening = forest clearing

Source: Cuny et al., 1992

These results clearly show that the success of a plantation of indigenous species is often quite unpredictable. A brief study (into the survival, growth rates, pedological and topographical features of the sites, etc.) could be undertaken in the arboreta, tree collections and experimental designs in different countries or sub-regions with a similar ecology.

The drawbacks with improvement plantings in dry tropical zones are numerous: unpredictable outcome closely related to rainfall regularity, need for protection under long deferred grazing /browsing periods, initial medium-to-low growth rates, indispensable beating up and intermediate treatment and tending operations, high costs etc. As much as possible, management via natural regeneration (seedlings, sprouts, root suckers, etc.) should be preferred to planting in stand establishment and/or regeneration.

2.4 Direct seeding

Many direct seeding trials have been carried out in the dry tropical zones over the last 50 years. They have shown that their success depends on the following:

- the need for quality seed in physiological and genetic terms;

- proper pre-treatment using simple and economical methods;

- seeding must take place at the beginning of the rainy seasons;

- the seed must be carefully sown (the hard topsoil must be scarified) and techniques simplified;

- grass growth must be checked or if possible species must be selected with young tap roots, because weed control requires considerable manpower which is involved for agricultural tasks during that period);

- low anthropogenic disturbance and a favourable human environment (agriculture evolving towards sedentary farming systems);

- a well-advised selection of species capable of fast initial growth, to rapidly free the crown from browsing by livestock and wildlife; and

- species (forestry, fodder, fruit) must be carefully chosen in terms of the requirements of the farmers, herders and other users.

On account of its low cost and comparative simplicity, direct seeding is therefore a simple and effective silvicultural technique, easily accessible to small farmers, when rainfall is regular. However, fast-growing or non-palatable tree species must be selected in order to shorten the protection period required. Relying upon natural vegetative reproduction is also a way of reducing the protection (through deferred browsing) duration.

2.5 Fire-breaks

The purpose of fire-breaks is to create discontinuity in the forest stand in order to reduce fire intensity and effectively control it at specific points. They must be established perpendicular to the prevailing winds. If badly designed, they can cause erosion. However, they are generally ineffective unless they are set up on limited areas with very clearly defined objectives, linked to intensive actions. It is impossible to comprehensively protect large areas.

The choice of a fire-break type (Figure 13), depends upon soil conditions, investment means available and people’s aspirations. The fire-breaks’ establishment and maintenance period coincides with that of intense agricultural field labour requirements, during which farmers are rarely at hand. In the absence of mechanical facilities, this constitutes a limiting factor in dry tropical regions.

There are four types of fire-breaks (Arbonnier and Faye, 1988):

- The bare fire-break, which is completely isolated, sometimes effective to protect small areas if it is on a hard-pan lateritic soil and if it is at least 30 m in breadth;

- Fire-breaks under natural vegetation (submitted to annual prescribed burning): They supplement existing 5 m-wide fire-breaks and extend to some 30 to 40 m in breadth, amidst the vegetation subjected to controlled burning; the advantage of this type of fire-break is that it does not require frequent maintenance;

- Fire-breaks under cultivation: standing on the best soils, they bear early maturing crops, the residues of which are removed after harvest (a rapid decline in the productivity of crops is generally noted in this type of fire-break);

- Wooded fire-breaks: the purpose here is to set up a permanent and closed crown cover to prevent grass from growing. The species that are suitable for this in West Africa include Anacardium occidentale, Azadirachta indica, Khaya senegalensis, Ziziphus mauritania (often the foliage is too open), Z. mucronata or other species with dense foliage (Gmelina arborea, Mangifera indica). It is recommended that species with small leaves be chosen because the layer of leaves that fall to the ground is less flammable than a cover made up of large leaves.

1. Bare fire-break

2. Fire-break under natural vegetation

2.1. natural vegetation under prescribed burning

2.2. fire-break under cultivation

2.3. wooded fire-break

However, fire-break establishment alone does not solve the problem. Regular maintenance and effective surveillance must be carried out. Fire-breaks are not isolated elements, sufficient to combat bush-fires. They must form part of a basic infrastructure and can be used to demarcate forest boundaries, delimit transhumance corridors and in some cases facilitate forest access (bare fire-breaks). No fire-break is totally reliable. The establishment of a fire-break depends upon the composition of the open or closed woody layer. In the steppe zones it will be economical. It is in such formations that fires are less frequent, particularly in the Sahel. Conversely it is particularly difficult to establish fire-breaks in the Sudanian zone where fires are frequent and intense. Their establishment and maintenance are extremely costly and cannot be taken in charge by village communities. For a 2 000 ha forest, some 100 to 200 ha of fire-breaks should be established, which constitutes a gigantic amount of work. To ease the problem, the following two solutions may be envisaged on fertile land:

- Fire-breaks under cultivation, which require full approval by all villagers (compensating them for the remoteness and limited yields of the fields, or providing non-financial incentives, since priority is given to agricultural work in the fields during that same period), together with continual monitoring.

- Wooded fire-breaks. Planted with non-palatable woody species playing the role of transhumance corridors, to be managed by the herders.

In the Badénou Forest of Côte d’Ivoire, farmers and herdsmen are in principle involved in managing fire-breaks (Part Four, Case Study 2), as is the case in the Bouaké region where earlier such involvement produced positive results.

2.6 Total protection versus early burning

Many different people burn vegetation - beekeepers, herdsmen, hunters, tourists, children, etc. The most frequently incriminated economic agents are the shepherds of the Sudanian domain who use fire to enhance aftergrowth of hardy grasses and to facilitate their herds’ movements. It is these fires which cause the most serious havoc in forests, because they are ignited at the height of the dry season when the biomass is more flammable. These are not just running fires but violent conflagrations which reach tree crowns. Very often fire-breaks constitute helpless defences against fires of such intensities and magnitude.

It is in the more humid parts of the dry tropical zones (for example with a Sudano-Guinean climate) that fires are most frequent. It is here that the joint involvement of herders, farmers, foresters and other economic agents should be encouraged. The Sahelian zone range lands are better protected.

Since many years, much attention has been devoted to the study of prescribed early brush-wood burning and its impacts on forest stands. The experimental design set up at Kokondékro in Côte d’Ivoire is the oldest. The main lessons drawn from it are:

Aubréville (1953) agrees with Bégué that the climax of the Bouaké region is the semi-deciduous closed forest, and that fire is the determining factor in the regression of these forest formations. In order to support his convictions, in 1936 he set up a network of three experimental designs in northern Côte d’Ivoire to study the impact of fire on woody vegetation. The one at Kokondékro, near Bouaké, is the only one that remains. It is now the oldest experiment that is still continually being monitored on bush fires in Africa.

Kokondékro is in a transitional zone between Sudano-Guinean and Guinean Forest climates, according to Aubréville’s definition (1949). The dry season runs from November to March. The rainy season generally has two rainfall peaks (June and September) but sometimes the short dry season, July through August, does not materialize. Mean annual precipitation (1974-1990) has been slightly under 1 100 mm, with large inter-year fluctuations.

The experimental design set up in the Kokondékro forest reserve is surrounded by natural forest formations (the part upstream of each compartment) and man-made plantations (of Senna = Cassia), Gmelina arborea and Tectona Grandis, dating back to the early 1940s. It is composed of three rectangular 2-ha compartments (100 × 200 m each) without repetition, separated by 10 m wide fire-breaks.

To each compartment corresponds a given treatment which has been preserved and maintained since 1937.

- compartment X: total protection against fire;

- compartment Y: early prescribed burning, on 15 December at the beginning of the dry season; and

- compartment Z: late prescribed burning at the end of the dry season, beginning in the first half of March.

Compartment X (integral protection) has become in 1994, after 65 years of fallow, of which 58 years were under total protection, a semi-deciduous closed forest. The canopy is practically continuous and is invaded by climbers. The understorey, which was impenetrable even 20 years ago has been largely cleared as a result of the rise and closing of the canopy. On this compartment, 117 species with a circumference of more than 2 cm have been identified, totalling about 6,900 individuals per hectare. The basal area is 28.1 m²/ha.

Compartment Y (early prescribed burning) comprises two distinct plant formations. The highest and most fertile part is covered by closed forest which appears much less mature than the totally protected compartment. A tree savanna covers the poor soils on the lower part. In this ‘early prescribed burning’ compartment, measuring individuals with a circumference over 2 cm, made it possible to identify 79 species for an average density of 2 200 ha and a basal area of 15.6 m²/ha. In comparison with the ‘integral protection’ compartment, the stand’s specific diversity has been reduced by 33 percent, while the total number of individuals (except Phyllantus nummularifolia) has fallen by 67 percent, and the basal area by 44 percent.

Compartment Z (late prescribed burning) which is burnt every year at the end of the dry season, is covered in the lower part, over two-thirds of its total area, by a shrub savanna, while the more fertile upper part is covered by a tree savanna. After 58 years of burning, there are no more than 20 species, with 214 individuals per ha, and 3.05 m²/ha of basal area. A considerable difference exists between this compartment and the one under the ‘integral protection’ treatment. In comparison with the latter, this one has lost 83 percent of its specific diversity, 97 percent of the tree population, and 89 percent of the basal area.

The initial conclusions to be drawn from this test in the pre-forest Sudano-Guinean zone are obvious: the different ‘prescribed burning’ treatments applied to a fairly homogenous tree savanna has created, after 58 years, widely differing plant formations (Figures 14 and 15).

In the absence of fire on two soil types of differing fertility, a semi-deciduous closed rain forest is developing. The late fires rapidly destroy the pole stands and the stump sprouts are unable to develop. The stumps gradually die. The adult trees of the fire-resistant species tolerate these fires and continue to develop, but there is no guarantee of regeneration. After about 40 years, the large individuals die without being replaced by saplings. The cover becomes increasingly weakened and the fires increasingly more violent. The mortality rate of the old trees rises. They only manage to survive on the more fertile soils. It is likely that they will rapidly disappear, giving way to grassy savanna.

With early prescribed burning, soil fertility is of primary importance. On fertile soils, the cover closes here and there, probably on old termite hill sites (these are not covered with grass, and represent isolated areas which are partially protected from the effects of bush fires). Fire is no longer a regular occurrence; the cover keeps on closing. Patches of closed forest appear and grow until crowns eventually meet. The number of species in semi-deciduous closed forest increases. However, these are at risk of disappearing if a running fire takes place in the forest during an abnormally dry year. On poor soils, the developments over the first few years are similar to what occurs in the ‘late prescribed burning’ compartment. The young individuals are killed by the fire. However, some saplings of savanna species escape (perhaps in very wet years, when fires are not intense). This creates a sort of balance which keeps the woody stand stable and there is very little wood production.

The recolonization of a fallow site by a forest formation can therefore only be envisaged under total protection, on all kinds of soils or, if early bush-fire is prescribed on the more fertile soils. Late brushwood fire is, wherever possible, to be avoided everywhere, just as early prescribed burning is to be avoided on the poorest soils. The forester who sets out to convert a degraded stand into a closed forest, that is productive and sustainable, must take account of this, particularly if he intends to use fire as a management tool (Louppe et al., 1995).

* Are there any conclusions to be drawn?

In the Sahelian zone and elsewhere on low fertility soils, fires ignited during early prescribed burning operations or resulting from wild bush fires, must be avoided.

In the Sudanian zone, early prescribed burning will still be used for a long time in conjunction with an optimum livestock carrying capacity. Total protection can only be justified for small areas of closed forest stands which are hardly ever affected by running fires. In open woodlands, early prescribed annual burnings induce higher fodder yields (Table 20). As regards wood production, it is negatively affected by regularly resorting to early prescribed burning. The loss in wood production is nevertheless less important than in the case of wild bush-fires (following a more or less long protection period). Continual use of early prescribed burning leads inevitably to some reduction of biological diversity through elimination of all species with fine bark that are sensitive to fires, whatever the sort.

Figure 14: Compartments at Kokondékro - Individuals (Côte d’Ivoire)

Figure 15: Compartments at Kokondékro - Basal area (Côte d’Ivoire)

Source: Louppe et al., 1995

3. Harvesting

The stump height or level above ground at which trees must be harvested has been studied very little. In the case of species that develop sprouts, all authors agree that a comparatively young tree, cut at ground level, will produce resistant sprouts. In the case of older trees, most felling instructions recommend to cut at ground level so as to enhance proventitious buds and, for some species, root suckers (Bellefontaine, 1995-b) A higher cut will mainly result in the production of adventitious sprouts which are less robust, contributing furthermore to weakening the stump.

Some loggers prefer to cut at 50 cm or even 1 m high. A height of between 20 and 40 cm is frequently recommended. If the cut is flush to the ground, the proventitious buds develop their own root system. The existence of more or less autonomous sprout-root sectors within the same stump has been revealed (for Eucalyptus camaldulensis) in Morocco (Riedacker, 1973).

In the case of silvo-pastoral management, another issue arises, namely that of young sprouts’ browsing by livestock. This has been raised with regard to Pterocarpus erinaceus which is difficult to regenerate without protection, given its slow growth and the high palatability of its young sprout shoots. Tests involving variable cutting heights (ground level, 0.5, 1.0 and 1.5 m above ground level) have been undertaken for this species in Mali (Anderson, 1994). Initial observations have shown that Pterocarpus erinaceus sprouts well at the various cutting heights and that sprout growth at 1.5 m above ground level is still reasonable. Cutting at chest height is therefore possible in order to place the sprout shoots out of the reach of certain animals. These tests must be pursued because one may wonder whether wood production, after a cut of this kind, is sufficiently attractive to justify the loss of 1.5 m of stem. Furthermore one assumes that the vitality and number of sprouts varies with the age of the stump.

Optimum cutting period: regarding trials in terms of optimum cutting periods, most of those documented were not designed according to a rigorous statistical design.

- Burkina Faso. A non-repetitive test to determine the most favourable period for inducing stump sprout regeneration was carried out on three sites from 1983 to 1984 (Kaboré and Renés, 1987). The cutting periods were set out as follows: beginning and end of the rainy season (May and September/October, respectively), dry cold season (December) and dry hot season (March). Even though it is difficult to draw objective conclusions, the best period for cutting would seem to be the start of the rainy season. However, authors emphasize the difficulty of showing the impact that the cutting period may have on the stump’s sprouting ability, given the heterogeneous nature and the general condition of the test compartments, the variability in soil quality and the irregularity of rainfall. Nevertheless the following provisional proposals were made:

* in the south Sudanian domain, it was recommended to cut between October and March, when the dense tall grass cover dries out gradually, reducing its competitiveness towards sprouts that should reach a reasonable height with the return of rain;

* in the north Sudanian domain, characterized by a cover of annual grasses, cutting can run from October to May;

* in the sub-Sahelian domain, the cutting period could run from April to June, to avoid a strongly marked dry season.

- In Mali, seven economically valuable species have been studied (Anderson, 1994): Isoberlinia doka, Daniellia oliveri, Anogeissus leiocarpus, Pterocarpus erinaceus, Combretum fragrans, Burkea africana and Khaya senegalensis. Generally speaking all of these seem to produce vigorous sprouts whatever the felling date. I. doka and D. oliveri having a constant growth rate throughout the year, the cutting period should not matter with regard to regeneration. Conversely P. erinaceus and A. leiocarpus grow at a faster rate after felling at the beginning of the dry season. Cutting down at that period of the year would be the most profitable.

- In Niger, non-repetitive regeneration experiments in 1974 have revealed that cutting operations between the end of the dry season and at the beginning of the rainy one, are most favourable to vigorous sprouting (Bouzou, 1984).

In addition to these biological aspects, account should also be taken of the human factors such as available manpower, the wishes of the loggers, the traditions and transport possibilities. Other considerations also regularly emerge following people’s involvement in managing, monitoring and deriving revenue from the forest: the loggers are first of all small farmers whose agricultural calendar coincides with the wood harvesting period recommended by foresters, which gives rise to conflicting situations in terms of labour availability.

Rotation is the period between two successive cuts of the same nature on the same compartment (Dubourdieu, 1989). In the absence of accurate experimental results, the simple coppice rotation period was estimated at around 15 to 20 years.

In Senegal, beyond eight years after harvesting, it was found that the basal area of the shoots suitable for charcoal remained comparatively stable. Arbonnier and Faye (1988) therefore envisaged an eight-year harvesting cycle instead of the 20 years previously indicated.

On the basis of experiments carried out at Gonsé in Burkina Faso, a rotation system of between eight and ten years is envisaged (to be specified subsequently).

At Tientiergou in Niger, six years is sufficient between two successive logging operations. The coppice selection system only harvests trees whose stems have already exceeded the minimum harvesting diameter. Short intervals between successive cuttings are possible without harm for the stand. However, under too short rotation periods, extraction only concerns the smallest categories of wood, i.e, those with dimensions just above minimum commercial diameter. This justifies the adoption of an intermediate rotation period of six years to be able to harvest other diameter categories of wood. There are also other reasons for this choice (Peltier et al., 1994-a):

- Management facilities can also make it possible for the number of compartments on a given area to be reduced when it is decided to set up a compartment layout to be managed on an annual basis. This reduces the length of the rotation; and

- In terms of rotations, decision-making should be based on local consumption habits. It is pointless to decide on long rotations, knowing that the wood categories harvested would exceed in diameter, that requested by markets (6-8 cm in diameter).

The rotation and revolution periods applied to high forests vary widely:

- The merchantable diameter generally in use in West Africa, lies between 50 and 60 cm. Depending upon species, rotation periods may range between 60 and 70 years for Afzelia africana, 50 years for Bombax costatum and 30 to 40 years for Khaya senegalensis.

- On the west coast of Madagascar, in the dry deciduous forest of Morondava, the initial 25-year rotation period was raised to 50 years and then to 100 years. For Commiphora spp., represented by two species accounting for 70 percent of the total merchantable volume, the time needed to reach the legal commercial diameter of 38 cm is put at 400 years (Part Four, Case Study 3).

- In Niger it is considered necessary to set up merchantable diameters sufficiently high, in the order of 35 cm (basal diameter) for Pterocarpus lucens (Peltier et al., 1994-b);

- In Burkina Faso, rotation periods in excess of 21 years have been proposed.

- In northern Côte d’Ivoire, at Badénou, the tree savannas and savanna woodlands are thinned, while the shrub savanna is regenerated by clear-cutting at the beginning of the management programme in order for them to be converted into high forest. The rotation is 20 years. Managed according to the high forest system, the best individual trees will be enhanced through thinning operations until they reach commercial dimensions set at 60 cm diameter on an empirical basis for a revolution of 80 years (Part Four, Case Study 2).

To date, it appears that no accurate studies have been undertaken with regard to felling techniques. It is a fact that shearing and removing bark from a stump, as a result of poor workmanship, using unsuitable or blunt tools, affect negatively the survival and growth capacity of sprouts as compared with stumps properly treated. Furthermore, the use of an axe, rather than a saw, leads to substantial timber loss.

It is important to lay down a minimum set of felling rules in order to prevent harvesting operations that are detrimental to the forest particularly the dry deciduous forest.

In addition to the height and periods of cutting, with which compliance is essential, the notion of directional felling should also be taught to the loggers. Indeed, the largest trees are generally felled without paying attention to the surrounding stand. The damage thus caused can be considerable and can hamper future regeneration which is vital in regions where trees have a particular importance (in steppes or savannas). This technique should also be encouraged in the forests which are rich in construction timber where the damage has direct financial repercussions (open woodlands with Isoberlinia doka, or the closed forests in Madagascar, etc.).

In the dry tropical forests on the southern edge of area concerned, a sound precaution to be taken could be to remove climbers from the large trees before felling.

To be also noted that in the Paraguayan Chaco, only a few species are felled with an axe for timber extraction: Schinopsis balansae (parquet, stakes for fences, tannins), Aspidosperma quebracho-blanco (parquet and charcoal), Bulnesia sarmientoi (parquet, stakes for fences, oil), and Calycophyllum multiflorum (cabinet-making). This logging operation relates to less than 10 trees per ha. These forests are mostly used as range lands, particularly in Argentina and Bolivia, where they are cleared, and 90 percent of the woody material is burnt (Gerber, 1995).

In small-wood hauling operations, the use of animal traction is still fairly unusual, instead, firewood and wood for rural construction and fences (‘utility wood’) are usually carried by man.

In the open savanna zones, large-diameter wood is either square-sawn on the spot, or chopped for firewood. In the zones with greater resources (open woodlands or dry deciduous forests) timber removal is a problem.

At Morondava, in Madagascar, interesting results have been attained in hauling operation trials based on animal traction using zebus (Wyss, 1990; Rakotonirina, 1991):

- while semi-mechanized logging requires substantial investments, hauling based on animal traction is much more economical, and is within the small farmers’ capacity;

- the ecological impact on the forest is less severe;

- comparatively few infrastructure schemes are necessary;

- material and financial returns are high; and

- it is a highly appropriate and locally suited technique, as it is based on a system known to the villagers who are well acquainted with the utilization of zebus and carts for other purposes.

The logging track system is often very inadequate: one or two logging tracks, more or less permanently negotiable, cross the forests. The small farmers/loggers mainly use footpaths to get to the logging sites. The wood is usually yarded/dragged to the landing site using these footpaths.

The feeder road system should be optimized in order to:

- minimize damage to the stand;

- facilitate access to logging sites (if the compartments to be harvested are not easily accessible, the loggers will harvest another compartment which is more accessible);

- return from the logging site with the felled timber (the ‘skipper strips’ are temporary tracks which only last throughout the period of the wood extraction);

- monitor forest access; and

- optimize fire prevention.

- the peripheral strip road which is operational during one year at least;

- the transversal ‘mechanized trails’, separated by 500 m to be used by the truck loggers; and

- perpendicular to the transversal mechanized trails and at some 500 m from one another, are the feeder roads used to facilitate wood extraction.

According to Wyss (1990), experimentation in terms of forest logging track system has been undertaken in Madagascar in order to study animal traction in logging operations (Figure 16):

- A system of skidding (yarding) and skipper strips was developed, in which zebus positioned within the skipper roads, by means of a rope and a pulley, drag the logs lying in 1 m wide skidding (yarding) strips. This has proven to be a very arduous and poorly performing system.

- A system of main skipper strips, communicating with distinct single 2.5 m wide subsidiary skipper strips, developed to reach each log harvested.

- A non-systematic network of skipper strips reaching every felled tree, with the possibility of making use of one lane (strip) in zigzag, to reach several felled logs. This adaptable infrastructure system is best suited for hauling, using animal traction, because obstacles can be skirted and the strips kept short if the paths joining several felled trees are carefully chosen. This latter type of network is the most suitable for haulage by animal traction.

The logging track system is therefore important as it reduces regeneration damage and conditions the access to the forest (for marketing forest products, for effective fire control and ease of forest grazing and monitoring). A good network provides good market access for forest products and fosters the rational development of the area. Where possible, this should be studied in advance, and as best as possible incorporated into the fire-break system.

Figure 16: Haulage by animal traction infrastructure

A. Systematic infrastructure with skidding track

B. Systematic infrastructure with skidding lanes opened by each lelled tree

C. Infrastructure with lanes opened according to fellings

Source: Wyss, 1990

4. Silvo-pastoralism

4.1 Definition of carrying capacity
4.2 Husbandry principles
4.3 Land management
4.4 Improvements

Half the world’s livestock is found in the dry zones of the planet. The management of forest tree formations in range land country and on agro-pastoral lands must therefore take account of this fact. Many forestry projects have failed because the pastoral component has not been properly integrated.

In the case of dry tropical tree formations, silvo-pastoralism is defined as the utilization by transhumant and agro-pastoral herds, of natural environments that may be occasionally, but not permanently farmed or cultivated. In West Africa, by ‘natural wild lands’ is meant the Sahelian steppes, the Sudanian savannas, the river flood plain zones (the interior delta of the Niger in Mali for example), the fallow systems, of variable lasting periods. After the harvest, the livestock feeds on crop residues.

There are a number of fairly complex rules governing the organization and husbandry of natural and agricultural resources.

4.1 Definition of carrying capacity

The notion of livestock carrying capacity cannot be accurately defined unless there is perfect control over the grazed lands, the number of animals and fodder production. Range lands are usually open and heterogeneous areas which the animals utilize irregularly and imperfectly, depending upon their movements. The use of the notion of carrying capacity nevertheless has the advantage of giving some idea of the magnitude of the number of heads of livestock that can be fed, on the basis of a reasonable estimate of future production.

The stocking rate of a range area is defined according to the number of livestock heads, or even better, according to the number of TLU it can feed, without compromising future production. Knowing that the TLU’s daily consumption of dry matter is set at 6.25 kg per day, it is easy, once the biomass of the range in terms of dry matter per hectare is known, to determine the number of grazing days per hectare for the anticipated period of utilization.

However the biomass measured at the end of the rainy season declines as the dry season progresses. Boudet (1991) estimates that in the Sahelian zone only one-third of the measured biomass can be consumed because of trampling, wind and termites. In the Sudanian zone, one-half of the edible part will be eaten by the livestock. This consideration must therefore be taken into account when calculating stocking rate.

If the stocking rate of a range can be evaluated over one biological year, it could also be evaluated taking into consideration the seasonal constraints and therefore be evaluated by season (dry season, rainy season). Equally, the stocking rate of flood-plain ranges (pastures) during subsidence of water, or that of the post-fire regrowth, are expressed in terms of the six dry season months.

Account must be taken of the availability of crop residues (millet shoots, for example) which represents a substantial quantity of the livestock diet.

The stocking rate of a range may be expressed either in terms of TLU (for example 0.2 head/ha) or in terms of number of hectares needed to keep one TLU (5 ha/head).

The production and carrying capacity for various bio-geographical zones of West Africa are set out in Table 19.

In some miombos, Setaria thermitaria and Brachiaria brizantha grasses and various legumes (Indigofera, Crotolaria, Eriosema and Vigna spp.) are the most important fodder plants. Given the fact that they are rarely dominant and considering the low global palatability of miombo pastures, means that the stocking rate of these range lands is at best equivalent to 15 ha/LTU (Malaisse, 1979).

In Ethiopia, the middle Awash Valley is a sub-desertic environment between the Afar Desert and the sharp slopes of the high East African plateaux. Here are found wide open savannas of the Awash region, open woodlands close to the Arba River, grass savannas and savanna woodlands in the valleys. The amount of fodder available varies from 700 to 3 000 kg of dry matter/ha, depending upon the site. The potential carrying capacity is between 4 and 12 ha per head of cattle weighing 300 kg (Corra, 1992).

In North-east Brazil, extensive livestock farming has been encouraged in the Sertão zone in the caatingas. Campello (1995) has explained that “the management of the vegetation is totally linked to the zone’s anticipated final land use (...). Deforestation is viewed as a way of raising land prices. Furthermore, the lack of value set upon fuelwood, coupled with a weak taxation system applied by the forestry authorities, is causing people to lose interest in forest management. Management for pastoral purposes is the most common use of the caatinga and also one of the most empirical. Under normal conditions, the carrying capacity of the area is 12 ha/head of cattle (area necessary to keep a large animal for one year) but there are many cases of overgrazing (...). In order to enhance pasture development, woody vegetation is eliminated, with the exception of some fodder species such as Ziziphus joazeiro.”

In Paraguay, in the Chaco humedo (lying between the Paraguay and the Pilcomayo Rivers) the carrying capacity is 5 ha per head of livestock (Gerber, 1995).

Table 19: Ranges’ production and carrying capacity in the bio-geographical zones of West Africa

Climate	Biogeographic zone	Annual rainfall (mm)	Rainy season (days)	Vegetation type and landscape	Production (kg/ha)	ha/LTU	Dominant tree species	Grass species	Animal	Land use
Desert	South Sahara	100	15	Contracted landscape	-	-	Acacia tortilis Acacia ehrehbergiana Leptadenia pyrotechnica	Stipagrostis pungens	Camels Goats Sheep	Pastoralism/no rain-fed cultivation
Hyper-arid	Saharo-Sahelian	200	30	Diffuse steppes Contracted landscape	400	17	Acacia. Tortilis Acacia ehrehbergiana Leptadenia pyrotechnica	Panicum turgidum	Camels Goats Sheep	Nomadic and transhumant Pastoralism , no rain-fed agriculture
Arid	Sahelian sensu stricto	400	75	Mimosae Savanna with annual grasses	1 000	7	Acacia Senegal Acacia tortilis Balanites aegyptiaca Boscia senegalensis	Aristida mutabilis Aristida funiculata S. gracilis (annuals)	Camels Goats Sheep Zebus	Transhumant pastoralism with millet cultivation
Semi-arid	Sudano-Sahelian	600	120	Combretaceae Savanna with annual grasses	1 500	6	Ombretum spp. C. micranthum Sclerocarya birrhea Acacia seyal	Cenchrus. Biflorus Eragostis trimula Loudetia togoensis	Zebus Sheep Goats	Pastoralism with millet cultivation
Sub-humid	Sahelo-Sudanian	900	180	Combretaceae savanna with perennials	800 - 2 500	3-9	Khaya senegalensis Parkia biglobosa Butyrospermum paradoxum	Andropogon gayanus Diheteropogon hagerupii P. pedicellatum	Camels Zebus Sheep Goats	Sedentary crop cultivation and pastoralism (millet, sorghum, groundnuts)
Sub-humid	Sudanian sensu stricto	1 200	210	Wooded savanna with perennial	3 000	7	Anogeissu leiocarpus Combretum nigricans	Andropogon gayanus A. pseudapricus Pennisetum subanustum	Zebus and Bulls 1	Cotton, cassava, groundnuts, mango cultivation
Humid	Guinean	1 200	300	grasses	6 000	1	Anona senegalensis C. febrifuga Bridlea diplandra	A. macrophylla Hyparrhenia rufa H. diplandra	Zebus and Bulls \

¹ = Trypano-resistant / Source: Grouzis et al., 1989

4.2 Husbandry principles

a) Range lands

Natural steppe and savanna range lands have been traditionally managed by herdsmen practising transhumance to varying degrees. While using these range lands, they try to strike a balance between grazing grass in its optimum stage, seeking water (ponds, rivers, wells), minimizing the inconveniences and risks from insects (flies, horseflies, tse-tse flies) and moving around avoiding conflict with the farmers.

In West Africa during the rainy season the herds use the pastures to the north of the rain-fed agriculture boundary. They drink at the temporary ponds and/or wells and are generally spread out widely. By this means they use range lands which would be difficult to access at the peak of the dry season, particularly in the northernmost zones. This is also the period of salt cures. Moreover, insects are fewer than in the Sudanian zone, and there are few risks of conflict with farmers. At the beginning of the dry season, after the crops are harvested, the herds can move southwards to find greener and hence more nutritious grass. There they graze on natural vegetation, fallow-lands, post-crop residues, river flood-plains, and on post-fire regrowths. The herds then move northwards again as the first rains arrive, when the peasant farmers begin planting.

Such use of natural resources is the result of a long experience which has led the herdsmen to make the very most of the resources while minimizing health risks and social conflicts.

During the course of these movements, the herds can cover considerable distances, up to several hundred kilometres. In tropical Africa these transhumant movements are well-known and have been described in the CTA/CIRAD-EMVT Atlas Elevage et potentialités pastorales sahéliennes (1985-1990). This atlas covers six Sahelian countries (Chad, Niger, Burkina Faso, Senegal, Mauritania). Two other atlases, by the same publishers cover Northern Cameroon (1992) and the Sudan (1993), describing the movements and the use of natural resources in these two countries.

The transhumant herdsman therefore stake out huge geographical areas in which, season after season, they find a fairly balanced diet for their livestock. The very well-known example is that of Cheikou Ahmadou’s ‘dina’ in the nineteenth century which ruled the use of river flood-plains in the interior Niger delta during the dry season (Gallais, 1967). In Mali there is also a pastoral code for animal traction that has been drawn up (Gallais and Boudet, 1980).

In East Africa miombo forests have a number of peculiar features, of which one is particularly outstanding: many trees in these forests produce new leaves two months before the rainy season arrives (Lawton, 1980), and the same is done by some rare species in the Sudanian area. There is no documented explanation of this apparent physiological dysfunction, which is widely exploited by the herdsmen.

In the Sahelian zone, apart from a few attempts to farm land around urban areas, transhumant herdsmen are the sole users of the natural environment. In the Sudanian zone, on the other hand, they use non-farmed land areas. Over the past two decades, the drought in the north and the increased agricultural population in the south have complicated this arrangement. Transhumance is now longer, the herdsmen are adopting a sedentary life-style, and the natural environment is increasingly being cleared. Furthermore, by using animal traction, the farmers have created what is fast-becoming a large pool of animal husbandry. This is creating conflicts between herdsmen and farmers.

b) The agro-pastoral lands

In the Sudanian zone, these lands were originally used for agriculture. The main agricultural activity is based on food crops (sorghum, millet, groundnut and sometimes rice) and industrial crops (groundnut and cotton). Around the village three more or less concentric circles spread outwards: the first are the almost permanent crop fields, then the fallow lands of variable duration, interspersed by temporary agricultural fields, and finally the natural vegetation areas. This whole complex is certainly of value to the pastoralists thanks to the crop residues (straw, leaves), the annual grasses in cornfields and the reinstallation of Andropogon gayanus in the fallow fields.

Gradually, on these lands cropping by animal traction has given rise to livestock husbandry. Now the farmers have small herds. These are often entrusted to a paid shepherd, and are kept far away from the village during the cropping period. After the harvest, the livestock graze on the crop residues and the cornfield annual grasses. It is in this period of the food calendar (November to February) that they compete with the pastoral herds in movement.

At the present time the agro-pastoral lands have enough resources to feed the village livestock provided that account is taken of the nearby natural grazing lands, and the possibility of having fairly limited herd movement. Furthermore, their geographical situation means that they can grow fodder crops and be supplied with agro-industrial by-products, particularly from the cotton and groundnut industries.

c) Overgrazing, resting and grazing exclusion

Apart from changes in the climate, about which man can do nothing, overgrazing and bush fires are two phenomena which are always raised when dealing with natural range lands.

Overgrazing may be defined as action by livestock which modifies the potential of a range land. The first manifestation of overgrazing is the modification of the floristic composition. The sought-after palatable disappear giving way to non-palatable species which are not sought after which have been given the chance to multiply. This disappearance of sought-after species may be due to the depletion of the root system, as will be seen later. But in physiognomic terms, this development is not particularly visible. The other visible manifestation of overgrazing is better known, because it brings erosion, and sometimes to a spectacular degree. The gradual disappearance of the grass cover, and even its total disappearance, and trampling encourage water erosion. This is particularly acute in hilly areas, such as in the Adamawa in Cameroon (Hurault, 1974).

In areas with low rainfall and in other marginal range lands it is sometimes difficult to determine and attribute the share of the blame to the climate or to the impact of the livestock. After the 1973 drought, Gaston (1981) showed that in the Chadian Sahel, climate was mostly responsible.

It should also be noted that the whole notion of overgrazing must be used with great caution. This is the case with the deep boreholes in Ferlo (Senegal). In the dry season, because of the presence of large herds (sometimes over 5 000 head of livestock if the neighbouring boreholes are out of order) the ground around the boreholes looks very degraded (and remote sensing makes this more evident). However, the situation is quite different in the rainy season. This area, which has received a large amount of nitrogenous and grain input through animal faeces, produces herbaceous vegetation of high quality. This ambiguity has been very clearly emphasized by Valenza (1984). For the real problem is how to ensure the sound operation of all the boreholes, use all this biomass which is practically destroyed by trampling at the beginning of the dry season, and implement measures in order to preserve and restore the tree and shrub component, which is often damaged and becomes impossible to regenerate.

More recently, Carrière (1995), studying the impact of pastoral systems on the environment has concluded that the combination of a variety of different factors “increase in farmed areas, reduction in the grazing areas, increased concentration of impacts around agricultural poles, transfer of livestock ownership, change in feeding systems, reduced access to resources by herdsmen” is the cause of “the gradual decline of natural pastoral resources over a quarter of a century, the pressure of man and livestock on the natural environments has doubled as has the number of users”. Since no analysis has been made of this development, one is presently faced with degraded ecosystems and no strategy to rehabilitate them and conserve what is still conservable. Quite rightly, Carrière concludes that “it is on the scale of the main eco-climatic region that one faces the twin problem of livestock production and environmental protection”, while recognizing the key role of the NGOs in solving these problems “at the level of the village and the small region.”

Resting is a comparatively simple way of countering the effects of overgrazing for what is generally only a short period of time. In extreme cases, resting is needed for several years. It may also be accompanied by some light regeneration operations, as has been initiated in Burkina Faso (Toutain and Piot, 1980), using procedures which will be examined in greater detail below (Chapter VII, 4.4-b).

Short resting may be one element of grassland management. In the Sahelian zone, during the rainy season, it permits optimum development of annual grasses, a good dissemination of seeds and may help to regenerate woody plants. In the Sudanian zone, it guarantees greater grass production which, once burnt, makes it possible to properly control brush encroachment. However, it should always be borne in mind that in the case of range lands, only one land usage right exists, and resting may be introduced within a changing and unstable context. This requires lengthy negotiations and the achievement of a broad consensus. The fallow period must be established in advance.

The two notions of resting and deferred grazing are not the same. But in practice resting coincides most of the time with deferred grazing of an area.

Enforcing grazing exclusion within newly exploited forest compartments over a long period of years seemed to be popular until quite recently. This operation was prescribed in every forest management plan: all harvested compartments were subjected to total protection (at least in theory) against livestock and fires. The period of protection varied between 18 months (the more modern development plans of the kind implemented at Nazinon in Burkina Faso) and five years. The main reasons justifying this decision were the following:

- to protect all highly palatable natural regeneration from browsing; and
- to protect against soil compaction following exposure and after clear-cutting, to both sunshine trampling by livestock.

According to Mazoyer (1992) imposing strict grazing exclusion for several years should be avoided as far as possible because it is by no means always necessary in dry zones. The decision must be suitable for each particular case, and impose the fewest constraints possible.

Other works carried out in Niger under the Energy II project, consider grazing exclusion to be less beneficial than believed. At Tientiergou there was no significant effect of enclosures, and the livestock had no impact on the growth of woody plants: “these grazing exclusions are not really necessary except to guarantee sufficient grass regrowth, after harvesting the woody plants and provide grazing resources after two or three years under deferred grazing. In this case, protecting shoots from browsing is of secondary importance.” (Peltier et al., 1994-a.) At the present time Peltier and colleagues suggest that for the Tientiergou forest a very short period of deferred grazing should be imposed during the rainy season following felling.

Considering the difficulty of enforcing grazing exclusion without creating discontent or conflict with the pastoral stock-breeders, there seems to be a tendency now to gradually reduce or even do away altogether with this approach which is too top-down and authoritarian. Setting up a short deferred grazing period of a few months is preferable if accompanied by efficient grassland management (reduction of stocking rate to be agreed upon jointly with livestock owners, improving pastures, setting up rotational grazing, developing water points at strategic sites, etc.). It is also essential to make the local people more responsible in order to limit the amount of livestock per village or the length of time taken for the transhumant herdsmen to pass through. This seems presently to be the only way to effectively integrate a standardized and acceptable form of silvo-pastoralism, avoiding overgrazing, thus ensuring resources’ sustainability. To achieve this, it means getting to know better the various categories of livestock-breeders and involving them, taking their needs into consideration.

In the case of forests under timber production management, the above observations need to be moderated, knowing that the highly valuable species (Khaya, Pterocarpus, Prosopis, Afzelia, etc.) are often very palatable. In order to prevent them from acquiring a shrub-like development, it is indispensable to protect them from browsing. In this context, grazing exclusion at Badénou (Côte d’Ivoire), has been set up for 3-year periods (Part Four, Case Study 2).

d) Bush fires viewed from the point of view of the pastoralists

When discussing bush fires two questions arise: their causes and their usefulness. CIRAD-EMVT has addressed the question of bush fires in the paper Fiches techniques d’élevage tropical (Ministy of Cooperation and Development, 1990), and what follows is largely taken from this paper.

Apart from rare cases in which they are due to lightning, the main cause is man. Fire is deliberate (fires ignited by hunters and herdsmen, or of malevolent origin) or accidental (travellers’ fires not properly extinguished, agricultural clearing fires which get out of control, etc.).

In the Sudano-Guinean landscape of Africa, fire is one way of maintaining tree savannas or savanna woodlands, given the trend towards the development of closed deciduous forest. In fact, tree savannas are maintained thanks to early brushwood burning ignited in November. Indeed, burning partially dried grass cover is less damaging to woody plants that are beginning their rest period, except in the case of young seedlings. Conversely, late brushwood fires which take place in March, when the grass cover is completely dry, reach as high as the branches of where young leaves begin to develop. These trees are then forced to put out new shoots and by so doing, become depleted (in terms of food and energy sources).

The paper mentioned above concludes “in this climatic zone it is man who maintains the savanna vegetation by practising annual burning. Savannas and open woodlands, whose floras are virtually identical, need fire in order to subsist”.

The impact of fire on the chemical contents of the soil is small (Box 19). Virtually all the mineral elements return to the soil after burning, except for nitrogen. However, “it should be remembered that the organic matter supplied to the soil comes essentially from the underground system, which in humid regions, has a greater biomass than the aerial complex. The rate of root renewal in tropical soils is extremely fast”.

Considering the aspects of natural resource management through use of early brushwood burnings and the problem of livestock production, the author of this report concludes as far as the Sudano-Guinean area is concerned that “fire should not be viewed as a factor of transformation but as a factor of savanna conservation. It is necessary in order to maintain the floristic variety of the savanna, and particularly the grass cover which is indispensable to livestock”.

An experimental design established since 1993 in Burkina Faso, compares the grass cover biomass to the woody biomass (a split-plot statistical design was used in the Tiogo and Laba forests) in compartments, some of which are subjected to early brushwood burning, and others protected from fires (Table 20). The grass layer is dominated here by the Poaceae family (over 80 percent of the total population), thus essentially by annual species such as Andropogon pseudapricus, A. Fastigiatus, Loudetia togoensis, Pennisetum pedicellatum, Diheteropogon hagerupii, Rottboellia exaltata, Microchloa indica and Hackelochloa granularis. These species are typical of shallow and poor soils. On fairly deep soils hardy Poaceae such as Andropogon gayanus, A. ascinodis, Diheteropogon amplectens are more frequently found. The grasses make up the bulk of domestic and wild ruminants’ feed. In 1994, after two years of protection against fire and following abundant rainfall, the biomass was found to have declined by 54 percent in the case of annual grasses and 25 percent in the case of hardy species. The annual production of the grass biomass is strictly linked to rainfall. These results need to be confirmed, particularly because 1994 was a year with a favourable rainfall. For the future management of these forests, the presence or absence of fire will have a direct influence on the choices taken by forest managers: the absence of fire encourages woody plants while early fires favour grasses (Nouvellet et al., 1995).

Table 20: The development of the biomass of annual and hardy grasses (in the forests of Tiogo and Laba in Burkina Faso) with and without early brushwood burning

Rainfall per year		Annual grasses (t/ha)			Hardy grasses (t/ha)
Year	Annual rainfall (mm)	A-Early burning	B-No burning	A-B (%)	C-Early burning	C-D D-No burning	(%)
1992	894	2.54	2.49	+2	4.05	4.20	-4
1993	748	1.96	1.22	+38	3.12	2.49	+20
1994	1 130	3.30	1.51	+54	4.79	3.60	+25

Source: Nouvellet et al., 1995.

The sparse woody layer of the Sahelian steppes frequently make it necessary to set up fire-breaks. But in addition to their high maintenance cost, they are not totally effective. It is preferable to work together with the local stock-breeders and to make them aware of the need to limit deliberate burning. It is very useful, additionally, to train them so that they can immediately set about putting out fires which break out near their camps.

Box 19: Biomass, mineral content and mass of the herbaceous stratum and of ashes before and after a bush fire in Loudetia simplex grass savanna
		C		N		P		K		Ca		Na
	Biomass g/m2	Content (%)	Minimum mass g/m2	Content (%)	Minimum mass g/m2	Content (%)	Minimum mass g/m2	Content (%)	Minimum mass g/m2	Content (%)	Minimum mass g/m2	Content (%)	Minimum mass g/m2
1972
Grass cover	606.7	4 115.6	249	0.191	1.157	0.0642	0.390	0.226	1.369	0.301	1.83	0.0077	0.047
Ash	57.4		9.0	0.268	0.154	0.371	0.213	0.985	0.565	2.33	1.34	0.0304	0.017
Restitution (%)			3.61		13.31		54.6		41.3		73.2		36.2
1973
Grass cover	379 494	40.1	152	0.277	0.051	0.0561	0.213	0.276	1.045	0.3	1.14	0.013	0.049
Ash		10.1	5.0	0.142	0.070	0.513	0.253	1.510	0.746			0.035	0.017
Restitution (%)			3.29		6.66		118.8		71.4				34.7
Average restitution			3.49		10.14		77.3		54.3		73.2		35.4
In a Loudetia simplex grass savanna at Lamto, Côte d’Ivoire, the contents of certain elements were measured in the vegetation immediately before the fire, and then in the ash after the bush fire. The contents together with the mineral masses per surface unit are given in the table above. They make it possible to calculate the restitution rate of each element (the percentage of the element recovered in the ash). Considering the average results gathered over a two year observation period, one may say that 54 percent of the potassium, 73 percent of the calcium and 77 percent of the phosphorous were handed back (restitution) by the ashes, whereas, 90 percent of the nitrogen content was lost. Source: Fiches techniques et élevage: les feux de brousse (1990). Ministy of Cooperation and Development - CIRAD-EMVT.

e) Pollarding and lopping

Pollarding consists in pruning branches level with the main stem, trimming-out end parts of branches or apical shoots of the crown (Figure 17). It is very widely used by the dry tropical zone pastoralists in order to increase fodder availability at the end of the dry season. The woody vegetation provides about 50 percent of the proteins in the lean period between the dry and the rainy season. This is therefore a vital part of the diet of the herds.

A study carried out on the Peuhl livestock husbandry system in Burkina Faso, shows that it has very little harmful effect on the environment, except on the trees which are pollarded (De Boer and Kessler, 1994).

Pollarding and cutting back of Pterocarpus lucens seem able to considerably prolong the period in which the leaves can be collected in the second half of the dry season (Le Houérou, 1980). Pollarding is more productive in the transitional season (October to January in the Sahel) than in the hot dry season (March-April) while cutting back would produce the opposite effect (Sepp, 1986).

Different types of pollarding and leaf stripping (extraction of leaves and twigs) operations have been carried out in Mali (Cissé, in Le Houérou, 1980) on three species (Combretum aculeatum, Cadaba farinosa and Feretia apodanthera): total leaf stripping every 15 and 50 days respectively, partial leaf stripping the total removal of the leaves every 15 days, the total removal of leaves every 30 days, partial leaf removal and a control treatment. The results have shown that there is a direct influence on leaf production of both the periodicity and the period of lopping:

- the leaf biomass is larger after the rainy season;

- a partial pruning of the branches and stripping of the leaves is more productive than a total extraction;

- when comparing the two leaf stripping operations at 15 and 30 days, production is higher in the case of the 30-day leaf stripping;

- leaf stripping can have both a depressive effect on leaf production (Cabada and Combretum) and a stimulating effect (Feretia); and

- the protein content is inversely proportional to the frequency of pruning.

If it is not carried out on a yearly basis, lopping does not induce negative repercussions on diameter increment of Prosopis cineraria in India (in Jensen, 1995).

Tests carried out on Acacia seyal in Burkina Faso (Toutain and Piot, 1980) have shown that pruning all branches of trees managed using the pollard system (Box 20) is very harmful to this particular species. The least harmful was the lopping of one-third of the crown. However, the authors stated that it was not possible to recommend any one particular type of lopping, on the basis of their results which would guarantee the sustainability of the individual trees. Total lopping causes a high mortality rate (60 percent).

Excessive stumping (pollarding), trimming-out and leaf stripping may affect the stand and bring about a change in species’ composition which is detrimental to fodder species: Cissé (in Le Houérou, 1980) mentions the example of the replacement of Combretum micranthum and Acacia ataxacantha with Perocarpus lucens, Acacia seyal and Combretum aculeatum.

Box 20: Pollarded oak stands The pollarded oak stands in the Basque country may be described as sparse woodlands with 20-50 stems of ‘pedunculate’ oak per hectare. The boles reach up to 3 m and bear strongly developed crowns with large branches. The stand’s clearings are covered with fern heath. This very particular type of forest cover still found over entire series of communal forests, has been fashioned by decades of original treatment. This came as a result of the Basques’ relentless willingness to reconcile unrestricted grazing with productions that are essential to their pastoral way of life: production of fodder to sustain their livestock, production of wood for domestic use and collecting forest litter. To this avail, the trees have been topped every 10-15 years either completely, or partially, a few metres above ground level. This ‘topping’ of the trees has the advantage of encouraging the development of sprouts above the ‘browsing line’, and enables the crown to grow abundantly. From the beginning of the nineteenth century until 1930, the forest management systems defined series which they called ‘pollard series’ where the harvesting regulations laid down the number of pollards which should be cut each year and the potential volume to be harvested, together with the cropping rules to be observed (maintaining a number of sap-drawers). However, the development of new harvesting techniques coupled with the non-renewal of trees ‘depleted’ by the frequency of the cuttings, have gradually led to the abandonment of this forest management method which some - not without malice - have named ‘coppice without standards’. Source: Aureau, 1989

In Zambia, the shepherds use fire as the dry season proceeds: burning and supplementary feed rations are necessary. The type of shifting cultivation (called chitimene) used in miombo forests can reconcile some of the interests of the herdsmen, foresters and farmers. The trees are cut back during the dry season to between 1 and 2 m in height leaving several branch stumps (topping / pollarding result in faster regeneration than cutting level with the ground). The leaves are browsed by the livestock. The branches are then piled up on the ground and burnt just before the beginning of the rainy season. This shifting cultivation system extends for three to five years before giving way to a fallow period based on Brachystegia-Julbernardia, which gain the upper hand again after about 20 years (Celander, 1983).

Few surveys and studies have been conducted into the response capacity of different species to pollarding, lopping and stumping. Those that have show that:

- if overdone, these practices threaten the survival of species;

- the trees are more subject to termite attack;

- in the event of a fire, the whole tree is affected, especially umbrella trees whose partially sectioned branches hang to the ground, protecting grass from grazing (hence constituting a stock of flammable straw);

- the effect of pollarding varies according to the species and the season; and

- pollarding extends the fruit gathering period into the second half of the dry season for some species.

- in Burkina Faso, in the Nazinon Forest, it is strictly forbidden to pollard the trees;

- Nouvellet (1992) has proposed controlled pruning in order to prevent overextraction, leading to early mortality;

- in Niger, in the Tientiergou Forest, the management plans permit pruning at more than 2 m above ground level for species which are able to produce aerial fodder (Peltier et al., 1995); and

- in Mali, the contracts for the village management of the fodder trees, particularly Pterocarpus erinaceus (threatened) in the strict reserved forest of Monts Mandingues (for the urban rearing of sheep) is probably the most interesting and instructive experiment so far in this area.

- specifying rights and guaranteeing them;

- respecting the carrying capacity;

- transferring certain duties, including investment in fodder species, and self-supervision of the logging in the zone;

- respecting traditional organization; and

- simplicity.

It is noted that lopping is generally forbidden in most West African countries, which creates a serious problem for farmers who wish to be able to use the trees in the parklands without restrictions. Tests on the technique, timing and the intensity of lopping, heading, pruning, and pollarding are not very numerous, which sheds doubts on the possibility of regularly renewing the tree parks which, in a few years’ time, will certainly constitute the main source of wood in the over-populated Sahelian zones. Legislation on trees scattered in the croplands should be brought up to date again.

In order to improve the interpretation of experiments of this kind, the synthesis of observations and results should be prefaced by an accurate definition of all the terminology used and of the harvesting procedures/modalities (season, periodicity, duration of the trial, the estimated age of the trees, the statistical method used, etc.). Figure 17 gives a better understanding of the meaning of the expressions and terminology used.

Research is still needed in order to better understand the consequences of this type of harvesting system and to propose extraction levels and procedures which do not weaken or kill the trees.

Figure 17: Pruning trees

Source: von Carlowitz, 1991

Figure 18: Potential production of mulberry leaves according to the shape and spacing of the trees

4.3 Land management

a) Rules of use: principles, surveillance structures

In theory, range land management could be organized according to the principles which existed before the drought, if there is a good understanding between farmers and herdsmen in the Sudanian zone, and a natural food chain is used.

In reality, to quote Boudet (1990) this is a question of a “legal and cultural imbroglio”, in the sense that since the use of range lands is merely a right of use, it is not considered as land development operation. Consequently agricultural development is taking place at the expense of the range lands, which are now at all events shrinking in the north because of the drought.

This being so, it is indispensable to lay down rules of use governing the exploitation of the range lands by transhumant herdsmen:

- the livestock feed regime and climatic risks require the herdsmen to move;

- this mobility has stood the test since centuries facing the vagaries of the climate that have seriously affected these people; and

- the geographical entities into which the transhumant stock-breeders developed must be recognized by the government following the achievement of a wide-ranging consensus by all the parties involved.

In technical terms, and with reference to Boudet (1990), the following principles could be laid down for the range lands:

- the Sahel could be moderately used during the rainy season;

- the Sahelian range lands could be left to rest periodically;

- the dry season stocking rate could be adapted according to the fodder stocks;

- standing fodder reserves could be set up for the hot dry season;

- an agreement could be reached on the use of the agro-pastoral lands after the food-stock harvests;

- intermediate areas could be identified on which the transhumant pastoralists could have rights of use; and

- fire management, water point protection, the supervision of the water points in the flood zones, after the mid-dry season drought.

The range lands and the agro-pastoral lands face a serious space problem in view of the climatic and demographic changes. Two fundamental parameters must be managed:

- use of range lands (appropriation, rights and rules of usage) which fall in the area of social sciences and administrative organization; and

- use of the soil and vegetation resources, which are directly related to agronomy and ecology of plant environment.

b) Continuous grazing: an insidious form of overgrazing

The work carried out by César (1992) in northern Côte d’Ivoire shows that biomass production depends on climatic factors and is also affected by water shortages. The edaphic conditions and tree cover also affect biomass production.

Studying grazing under various intensities, the author proposed a number of equations predicting optimum pasture use, under which the fodder potential would be maintained. Pastures of highly productive and palatable species can be maintained when exposed to high stocking densities in a limited period of time. César mentions “it therefore seems that it is preferable to use these savannas under high stocking rates for limited periods of time, so as to provide the most regular feed possible, while leaving in return, sufficient resting periods of about 30 days each, for the pasture to recover. This form of management would seem to be preferable to the usual system which is almost continuous grazing.” (Figure 19)

Continuous grazing, which often leads to overgrazing, reduces the production of regrowth. The grasses are unable to reconstitute their underground reserves, with the result that the root system is depleted by losing mass. This reduces production the following year. However, by introducing a rest period, since the soil is not depleted, the grazing potential is restored.

Continuous grazing is an insidious form of overgrazing which does not translate into visible deterioration of the ground surface. “The root biomass is reduced, it no longer provides organic matter to the soil and the humus layer becomes thinner and disappears. The savanna fodder grasses die because of depletion and all that remains is sparse annual psammophilous grass species that are able to withstand the new edaphic conditions. This development can be completed within five to ten years” (César, 1992).

Studies of this phenomenon show that these resources, old fallows and savannas which seem to be fairly hardy and have sufficient biomass need to be managed with great care. Range management in savanna areas must in fact operate to maintain the pasture potentialities, while avoiding bush encroachment.

Figure 19: Evolution of the floristic composition resulting from different treatments on the Abokouamékro Ranch in Côte d’Ivoire.

1) On the natural savanna at Abokouamékro, the favourable species (Andropogenea) are dominant (balance 1)

2) Continual grazing causes a regression in the favourable species, benefiting Loudetia arundinacea (balance 2). Allowing the savanna to rest, does not alter this new balance.

3) Intensive reaping reverses the direction of development, with a regression of Loudetia and an extension of the Andropogenea (balance 3).

In this connection César (1992) points out that “the development of woody layers was the natural evolution of every stand in the Sudano-Guinean domain and this trend increases with grazing”. The solution he proposes to reduce woody plants is deferred grazing (to provide combustible grass matter) followed by early bush prescribed burning.

Brushwood encroachment has also been known for a long time in the Central African Republic. As long ago as 1965, Bille drew attention to the problems created by Harungana madascariensis. At the present time, because of grazing, which encourages the dispersion of the seeds of certain species, invasion by Chromolaena odorata (Laos grass), which is a recently introduced weed, is increasing (Audru, 1988).

c) Thorough multidisciplinary management

In conclusion, when managing the pastures within the dry tropical woody formations, a balance must be sought in the plant landscape by curbing the natural tendency towards brushwood encroachment, while encouraging the sustainable existence of fodder grasses.

This balance, which is fairly easy to state, in reality requires from the technical viewpoint, excellent fire control and adequate skills on the part of the pastoralists, aided by technicians, to enable them to carry out a diagnosis of the grazing lands in order to exploit the grass at the optimum time while maintaining high soil fertility. This fertility is vital. It guarantees good fodder production, and is a condition of success if the long-term fallow land is to be used as arable land again.

In what might be called sociological terms, the management of grazing lands in dry tropical forests necessarily requires certain rules of use be generally accepted and adopted. These include the adoption of:

- geographical pastoral units;

- complementary approaches and objectives between pastoralists, agro-pastoralists and foresters; and

- respect for regular rest periods, calendar of early bush-fire prescribed burnings and stocking rates agreed upon.

Monitoring, both in terms of early warning and the medium-term development, is an indispensable scientific tool for implementing any renewable natural resource management system on a sustainable basis. It must be the result of the participation of agricultural, forestry, veterinary, water, meteorological and physical planning services.

These techniques and proposals at the level of the farmer can be implemented immediately. But the improvement of the livestock husbandry also requires such techniques as haymaking and pasture regeneration.

4.4 Improvements

We shall only deal here with two types of improvement requiring inputs, which while being fairly modest nevertheless constitute a major curb on their extension: haymaking and pasture regeneration.

a) Hay

Hay is defined as dry fodder. It comes from the grass which is cut when green and dried in the sun, and then stored for several months before being given to the animals. If at the moment of reaping the grass is already dry standing, it is not hay but straw. This distinction is very important. In the first case, the hay has conserved most of its nutritious qualities, but in the second they have virtually disappeared entirely.

In order to obtain high quality hay, the green grass must be reaped when it is mature. Reaping may be manual (by sickle or scythe) or mechanized (animal traction or motorized). The manual reaping is tiring and requires a great deal of manpower. A family may store reserves for a few head of cattle in the case of a farm, but this method is impossible for intensive livestock-breeders. The mechanized reaping makes it possible to reap much larger areas. A yoke of oxen can pull a mechanical mower.

Once reaped, preferably early in the day, the grass must be dried in the sun and turned over several times. It normally takes two to three days for drying to be complete, which implies storage in small haystacks for the night, and in the event of rain. While the grass keeps its qualities when drying is rapid, the rain leaches all the mineral elements from the hay, particularly if it falls when the latter is almost dry. It becomes then, very similar to straw. This comparatively simple haymaking technique requires some organization on the part of the small farmer. It requires him to estimate the amount of hay to be reaped, depending upon the amount of people available and taking account of the risk of rain.

After harvesting, the hay must be stored, possibly in the shape of compressed haystacks in order to protect it from the sun and also to reduce its bulkiness. It is quite a dangerous makeshift arrangement to store the straw on the roofs of the houses. Other fodder reserves are made by the small farmers and herdsmen. They store cereal stubble, groundnut haulm, Faidherbia albida husks, and brush straw which is sometimes sold in the towns, green Echinochloa stagnina (bourgou) which is sold green on the markets, etc.

With such a high demand for fodder, this tendency to store has become widespread in recent years. The demand can only keep growing and it is possible that haymaking will develop at least in certain specific areas.

b) Pasture improvement

This activity which demands a great deal of manpower often at a time when it is not available, is becoming evermore costly. It can therefore only be undertaken on limited areas, which will facilitate guarding, because once they are restored, grazing lands are often used by herds coming from elsewhere.

Work on this subject-matter has been carried out in Africa, mainly in the Sahel: Burkina Faso (Toutain and Piot, 1980), Chad (Guervilly and Bouba, 1992; Toutain, 1993) and Senegal (Diatta and Mandret, 1991; Roberge, 1994).

The most simple improvement technique consists in soil preparation. As far as possible, the soil is turned over or ploughed along the contour lines. The lines are about 10 m apart (Figure 20). This light preparation is sufficient to retain water and the seeds of annual species. After the rains, the lines of grass become clearly visible, and are effective for several years, in helping restore the soil, until its surface cover is totally rehabilitated.

This presupposes that for two to three years the area is under full grazing exclusion. This technique can be improved by sowing fodder species along the lines. At the beginning, the normal practice is to use seeds of natural species collected during the previous dry season. The results are encouraging but the village community that undertakes this work needs to exert perfect control and supervision over its land.

Figure 20: Schematic profile of a sub-soiled and crescent-ridged compartment

Source: Toutain (1993)

These techniques, which have shown their effectiveness on highly degraded lands, are not in themselves a solution to the problems of range land deterioration. But if they are managed properly with a general sense of responsibility on the part of the rural communities for the management of their own lands, they provide an excellent demonstration of the potential of natural vegetation, which is particularly well-suited to the extreme conditions of marginal lands.