Chapter VI: State of the Art and Tools - Forest Resources Evaluation

1. Inventories
2. Estimating forest resources
3. Growth and productivity
4. Some data on productivity
5. Allowable cut and compartment layouts
6. The main technical approaches
7. Measurement of fodder biomass and density of cover
8. Evaluation and monitoring of resources: the new tools

This chapter summarizes the main methods for observing and measuring the arboreal, woody and herbaceous components of the forest ecosystems in dry tropical zones. At the end of this chapter, the methods for evaluating and monitoring the resources are described.

1. Inventories

1.1 Regional or global inventories
1.2 National inventories
1.3 Local inventories

Inventories should provide quantitative and qualitative evaluations of the resources and their distribution with the necessary degree of accuracy according to the particular objectives to be attained. Because of the wide variety of different objectives, inventories exist for large geographic regions, and even whole continents, countries, administrative sub-regions and forests. These have been established for macroeconomic, bio-geographic, physical planning purposes, as aids for local management purposes. Each one of these has specific, if not unique, methods of observation and evaluation techniques. This is due to the diversity of the situations, but becomes more controversial when this diversity is due to a lack of minimum standardization or communication. It is therefore an area which demands a major effort to standardize practices and gather information. We would add that this could be encouraged by improving certain monitoring and observation techniques (using radar satellite imagery).

A distinction is normally drawn between three major types of inventories: regional or global, national and local. For the forest manager the latter two are the most relevant. National inventories are associated with physical planning, and local inventories with forest management.

1.1 Regional or global inventories

These are based on macroeconomic or bio-geographic approaches and only provide quantitative or qualitative global evaluations relating to the fauna, flora and sometimes the elements examined for a mesological approach. These mainly refer to forests in the strict sense of the term, ignoring small rural wood-lots, agroforestry cover and in particular woody plants scattered around fields constituting a substantial biomass. The latest global and regional inventory was the one drawn up by FAO (1993) which excluded the densely cultivated zones from the dry regions.

1.2 National inventories

This is the most common tool used for physical planning and national forest management. It is a national planning tool. It is clearly more detailed than the regional inventories just mentioned, because it describes, quantifies and identifies the site of the resources more precisely, but in particular it relates the forest to other agricultural, urban and industrial components in the territory. It should therefore take account of the ‘off-forest’ woody resources, but this is not always done.

Emphasis on the importance of the off-forest resources is based on an estimate by Jensen (1995) regarding the volume of wood taken from the fields, namely, that at the national level, the growing trees on fallow land and crops account for 30 percent of the wood resources in Burkina Faso and 19 percent in the Gambia (Box 10 and Table 6).

For the same reasons, it should be noted that the total output (namely wood, fodder and farm production) of a forest-park or shrub-fallow land may exceed the total output of a forest. In some parks, the trees are better protected against fire, and the density of the shoots is not so great which gives them greater individual growth; in a forest stand, competition from the grass cover from the time of the first rains reduces water availability and hence limits growth. Strictly from the wood production point of view, the disappearance of the forests to the benefit of the parks carrying trees for multiple uses is not always disastrous. In future, in the dry tropical zones, cultivated lands will certainly be increasingly used for wood and fodder production, and can be incorporated into the national inventory.

National inventories have been drawn up in many countries, particularly in West Africa: Burkina Faso (Table 6), the Gambia (Box 10), and Senegal (Box 11). Mali decided on an inventory of the southern part of the country, and this was completed in 1991 (Table 7). Niger has a regional inventory dating back to 1986, and the Planning and Utilization of Soil and Forest (Planification et Utilization des Sols et Forêts) project worked from 1984-1989 on the inventory of wood tree sources in the non-desert part of the country. This inventory was updated (Landsat TM [thematic mapper]) by the Energy II project (Energie II).

For the developing countries, the methodological difficulty is transferring the results of the national inventories in order to provide data that can be used for local development and planning. This is an avenue of research to be undertaken in future, because there is so much pressure that these wood stands are rapidly being changed by the local people. It is therefore vitally important to begin sustainable sub-management of all the forest stands in every country, as Niger is doing (SEED-CTFT, 1991), using all available inventories.

Box 10: National forest inventory in the Gambia The Gambia has a national inventory since 1980 (Forester, 1983). The mapping of stand types was implemented through interpretation of a complete set of panoramic, panchromatic and infra-red photographs at scale 1/25 000 dated October 1980. The background map is a mosaic of panchromatic photographs enlarged to scale 1/10 000. The volume is measured over bark, starting from a 5 cm above ground felling point. It concerns all trees whose diameter at 1.30 m is superior to 10 cm. Trees under 10 cm diameter are numbered and accounted for as regeneration. The inventory covered almost the entire territory, including trees on cropland. Some 590 inventory plots out of 2 035 were established on non-forested village land. Several volume tables were used for timber as well as for fuelwood species. For trees under 40 cm diameter, the volume tables for timber species correspond to those established for fuelwood species. For large dimensions (over 60 cm diameter), timber volume tables are 7.5 percent superior to fuelwood volume tables. Growth increments have not been measured. They were estimated using sophisticated mathematical applications based on hypothesis which were not verified. Distribution of surface areas, volumes, wood productivity and regeneration under various natural vegetation formations in the Gambia (DFS/GTZ, 1983).
Type vegetation formations	Area (1000 ha)	Volume		Increments		Regeneration
		1000 m3	(m3/ha)	(1000 m3/year)	(m3/ha/year)	(number/ha)
Mangrove	14 980	2 748	183.5	72	4.8	609
Woodland and woodland savanna	89 446	5 217	58.3	123	1.4	860
Shrub savanna	346 981	5 370	15.5	156	0.4	1 075
Total forest formations	451 407	13 335	29.5	351	0.8	1 016
Fallow and cropland	394 423	3 213	8.1	56	0.14	847
Total natural formations inventoried	845 830	16 548	-	407	-	937
Source: Jensen, 1995

Table 6: National forest inventory in Burkina Faso - Distribution of areas, volumes and wood plant productivity by different type of plant stand (FAO)

Vegetation formation types	Area (1000 ha)	Volume in (1000 m3)	Volume (m3/ha)	Increments
				(1000 m3/year)	(m3 /ha/year)
Woodland and woodland savanna	4 848	193 803	40.0	-	-
Shrub savanna and ‘tiger’ thickets	10 572	155 545	14.7	-	-
Total and average for forest formations	15 420	349 348	22.7	8 595	0.56
Fallow and cropland	8 770	152 851	17.4	1 684	0.19
Total inventoried and average	24 190	502 199	20.8	10 279	0.42

Box 11: National forest inventory in Senegal Senegal has a national inventory that was drawn up in 1978. The cartography of the different types of forest stands was based on Landsat MSS (multi-spectral scanner) image interpretation on a scale of 1:500 000 in 1978, whose pixels represented an area 57 × 79 m (0.45 ha). The cartography only covered forest stands. The volume inventory was never made but the average volumes of different types of stands has been estimated on the basis of data from different inventories and measurements made by the Forestry Service (Service Forestier) and the National Forestry Research Centre (Centre National de Recherches Forestières). It indicates the volume of greenery on bole overbark to the 7 cm cut of all trees with a diameter above 10 cm. The range begins at 0.25 m3/ha for the shrub steppe to 120 m3/ha for the closed forests. In order to express the total volume (trunk + branches up to 4 cm in diameter) the bole volume must be multiplied by 1.4. To estimate productivity, the figures used range from 0.02 m3/ha/year for the shrub steppes to 3 m3/ha/year for the closed forests. These estimates are multiplied by the area of the stand types taken from the cartography.
Forest areas, bole volume and fuelwood productivity of different types of stands in the 1978 Senegal National Forest Inventory
Wood stands	Area 1978 1000 ha	Volume		Increments
		1000 m3	m3/ha/year	1000 m3	m3/ha/year
Mangrove	169	5 936	35.1	-	-
Closed forest	13	4 500	104.5	26	2.0
Galery forest	45	1 560	104.5	135	3.0
Open woodland	349	20 952	60	699	2
Savanna woodland	2 500	73 557	29.4	3 702	1.5
Tree savanna	4 852	28 209	5.8	3 229	0.7
Shrub savanna	469	432	0.9	64	0.1
Tree steppe	46	455	9.9	23	0.5
Tree steppe and grassland	3 159	3 248	1	367	0.1
Tree and shrub steppe	2 160	326	0.2	52	0.02
Palm groves	40	-	-	-	-
Coastal thicket	43	-	-	-	-
TOTAL	13 845	139 175	-	8 297	-
The overall growing fuelwood is 194.845 million m3 (namely 139.175 million m3 × 1.4) which is comparable to the estimated productivity of 8.297 million m3/year. NB: The rural woodlands are not included in this estimate. Source: CTFT/SCET, 1981

Table 7: Potential of main stands in Mali

Vegetation formations	Rainfall (mm)	Gross volume (m3/ha)	Dominant height (m)	Dominant species
Open woodland with Isoberlina doka	935-1 085	39-61	13-15	Isoberlina doka
	1 295-1 355	34-90	11-14.5	Isoberlina doka
Other open woodlands	1 000-1 310	50-58	11.1-11.8	Daniellia oliveri Bombax costatum, etc.
Shrub savanna (sand dominant)	500-700	4.5-10.1	4.9-7.9	Guiera senegalensis
Hard pan plain-land stands	450	5.5-6.0	5.5-9.4	Pterocarpus lucens Combretum spp.
Shrub savanna (heavy soils)	450-575	2.7-8.7	4.3-8.3	A. seyal, Balanites aegyptiaca Anogeissus leiocarpus,
South Sudanian shrub savanna	980	15.4-18.9	6.4-7.3	Terminalia macroptera etc.
Tree savanna (hydromorphic soils)	630-700	7.7-16.2	7.4-10.3	A. seyal, A. leiocarpus, etc.
Tree savanna (non-hydromorphic soils)	560-700	10.5-18.4	8.3-11.5	Combretum spp., etc.
South Sudanian savanna	935-1 085	25.0-57.6	8.8-11.7	Pterocarpus erinaceus Combretum spp., etc.
Sudano-Guinean tree savanna	-	22.0-45.7	8.7-11.3	D. oliveri, P. erinaceus Combretum spp., etc.
A few South Sudanian and Guineo-Sudano ‘orchards’	975-1 310	21.3-27.4	8.3-10.5	Vitellaria paradoxa P. biglobosa, S. birrhea

NB: The gross volumes and dominant heights are the extreme per type of stand. For each type of stand several sites were inventoried and their averages recorded.

Source: Nasi and Sabatier, 1988a

1.3 Local inventories

a) General features

Local inventories relate to forests for which management actions are intended. In order to define the objectives and the management procedures correctly, it is essential for a prior survey of the human and physical environments interacting with the forest to be carried out, and in particular to specify the socio-economic and environmental roles of these environments. This will partly facilitate the definition of the variables (to be measured and/or benchmarked). They also make it possible to spell out more clearly such more general and global aspirations as the notion of sustainability. This situation must then be coupled with a geographical description and demarcation of the forest and the zones which are only slightly forested, agriculture and rural.

Before drawing up the inventory, the forest may also be divided up into to several complementary partitions. In particular, zoning can be carried out in terms of usage, and in terms of the major vegetation stands. Photo-interpretation and remote sensing are valuable tools for the latter, the first being based upon surveys or field observation. Lastly, the forest manager must also have a stratification so that the inventory can be adapted to each layer. After this, the techniques used will be taken from the traditional methods that are always used when carrying out stratified surveys.

Whether a priori stratification is used, which is the method advised, or whether this is done a posteriori, as is often necessary because of time and cash constraints, in most cases the survey is carried out systematically in the areas identified, set out regularly along equidistant rows. A preliminary sample with a very low rate should be envisaged in order to characterize the heterogeneity of the resources, namely the value of the variation coefficients of the variables studied, in order to be able to estimate the minimum number of experimental plots necessary to achieve a specific level of precision.

Box 12: Sample size Let us take a population P of N individuals and a variable X which for individual i has the value Xi. The average 0 of N values Xi may be estimated as the average of the Xi measured on a sample of size n. Using the traditional hypotheses: sample without remainder, random and Gauss distribution of X (the normal law), one can construct a confidence interval of 0 at level 95 percent which is written as follows: Noting the precision over 0 and CV the coefficient of variation of X (CV is estimated by ), one can show that to obtain a given precision p0 a sample of size n is needed such that; , etsi N est grand, NB: N is the number of potential plots in the area inventoried; n is the number of plots measured. X may be a volume, a land area, a height... attached to each plot.

However, this pre-sampling is difficult to implement and increases the costs. One can also calculate the variation coefficient as the inventory proceeds or refer to the values obtained in similar situations. Table 8 shows the main variation coefficients taken from the literature.

The size of the experimental plots being surveyed generally varies between 0.1 and 0.5 ha. It would appear that an area of 0.125 ha (25 × 50 m for example) is a good compromise between the desire to reduce the percentage of error and to economize on time and money.

On the basis of a study carried out in the Koumpentoum Forest in Senegal, Arbonnier (1990) estimated that a 1 percent survey rate, representing the ratio between the total area sampled and the total area of the forest, is acceptable for making a partition of the forest into homogeneous compartments each of 500 ha for a silvo-pastoral management scheme (Figure 6). He also showed that accuracy has nothing to do with the shape or the size of the experimental plots (Box 13).

Table 8: Values of several coefficients of variation (inventoried) in the literature

Forest-Country	Vegetation	Rainfall (mm)	Experimental plots	CV %	References
Woro-Mali	Shrub and tree savanna	600-700	0.1 ha	CV(N) = 68.9 CV(Vol) = 79.8	Parkan Benembarek Meijer 1988
Yabo-Burkina Faso	Riparian formation Savanna woodland Shrub savanna	600	0.25 ha	CV(Vol) = 79.8 CV(Vol) = 40-70 CV(Vol) = 50	Compare Laban, 1983
Dabo-Senegal	Open woodland and Tree savanna	800-1 000	0.5 ha	CV(N) = 31.5 CV(Vol) = 34.9	Fall, Gueye 1993 - a and b
Faira-Niger	Striped and leopard bush	600	0.1 ha
	2 < diameter < 8 cm			CV(Vol) = 35-60 CV(N) = 30-40	Mengin-Lecreulx
	8 < diameter < 20 cm			CV(Vol) = 65-85 CV(N) = 60-75	Chabanaud, 1986
	Diameter > 20 cm			CV(N) = 100
Faya-Mali mountains Mandingues Dioforongo	Poor to rich formation	750-1 000	0.2 ha	CV(G) = 37	Sabatier, Nasi, 1988
Morondava-Madagascar	Dry closed forest	500-1 000	Sample strip 40 × 1 000
	Species harvested Diameter > 37 cm			CV(Vol) = 28 CV(N) = 26
	Species harvested 30 £ Diameter < 37cm			CV(N) = 33
	Not harvested species Diameter ³			CV(N) = 49

CV = Coefficient of variation
Vol = Volume/ha
N = Individuals/ha
G = Basal area/ha

Stem: diameters measured at collar height < 20 cm
Foot: diameters measured at collar level > 20 cm

Source: Arbonnier, 1990

The parameters to be measured should translate to objectives set:

- forestry aspects: the number of shoots, diameters, height (total, bole, etc.), volumes (total, construction timber, fuelwood, over or under bark, for each diameter class, etc.), land areas, number of young plants, number of dead trees;

- fodder aspects: types of grassy vegetation, recovery of woody and herbaceous plants, accessibility of aerial fodder (height);

- the phenology of the woody vegetation in order to identify the optimum period for harvesting fruit, pruning, collecting other forest products; and

- estimating the bare soil zones in order to reconstitute the cover.

b) The need for inventories

Some doubt the usefulness of inventories because of the extra costs (in terms of time and money) involved and the poor matching between the volumes calculated and those actually exploited. This applies particularly to open woodlands and savanna, for which the heterogeneity and the twisted shapes of the trees and bushes make it difficult to calculate volumes.

For some writers, the inventory is irrelevant to stands that do not reach 50 to 60 m³/ha of growing stock. They then recommend using the basal areas, which make it possible to roughly calculate the volumes. But as Kaboré (1989) has pointed out, “there would then not be any inventories as such when one realizes that very few Sahelian forest stands meet this condition”.

The purpose of the inventory is to provide the forest manager with objective data on the resources in both quantitative and qualitative terms, but above all to give him a better understanding of the forest, to describe all the various aspects of it (soil, grass and tree cover, biodiversity, etc.) and to produce a balanced picture of it in terms of the areas that are valuable and those which are poor. This information is necessary in order to reconcile forest resources, needs and sustainable management.

2. Estimating forest resources

a) Estimating volumes

Since wood is mainly used as fuel, estimating volumes in dry zones differs from what happens in humid zones where the main objective is timber production. Estimates relate to the whole of the woody biomass including the bole and the branches of each tree, except the shoots that are too small in diameter and protected species or those not considered suitable for the use as fuelwood. The traditional methods for measuring volumes of standing stock are not really suitable for these environments. Due to their important specific heterogeneity, crown wood volumes are difficult appreciate both in terms of their evaluation and their modelling. It is often necessary to resort to destructive methods in order to work out volume tables and estimates that are as close as possible to the total timber volume.

The results of the inventories which are often published without specifying the methodology used, must be analysed with caution. The measures that have been used should be studied and the volume tables used must be carefully scrutinized. In particular, it is necessary to know the following:

- whether all the species have been taken into account or if some have been excluded because they do not match the objective set. Some species, for example, are suspected of giving off toxic fumes, or do not burn well (they have a low calorific value, and produce too much ash) and they are not taken into account in a fuelwood inventory;

- the minimum dimensions used for the species, so that the total fuelwood volume is not underestimated. Small twigs abandoned on the ground vary enormously in terms of the distances for transporting it. Some volume tables indicate the diameter at felling point, but this is sometimes not mentioned (the smallest cut to be taken into account for fuelwood consumption would be established at a diameter of 3 cm in view of the importance of small-sized shoots in supplying domestic energy in rural areas);

- the conversion factors used between weight (tonnes) and volume (cubic metres and steres); and

- whether the volumes have been calculated with or without the bark. This is particularly important for fuelwood (case of Guiera senegalensis and Combretaceae). The volume of bark is important for such species as Terminalia spp. and much less important for species of the Combretum genus.

For fuelwood solid volumes are measured with the bark, except for dead standing stock. However, timber volumes and volumes of ‘wood for rural construction’ are easily overestimated. Bark can account for up to 25 percent of the total volume. It is therefore important to carefully assess the thickness and the volume of the bark when estimating volumes. Figures relating to the percentage of bark in total volume, depending on diameter and age classes are hard to come by.

Van Laar (1981) studied the volume with and without bark in the Gambia for two species (Table 9) in terms of diameter. In Mali one study has given a few details about eight species (Table 10). Pterocarpus erinaceus has the highest percentage of bark, with 17.6 percent (Anderson et al., 1991).

Table 9: Percentage of bark in the Gambia in terms of diameter for two species of construction timber

Diameter in cm	Bark in Percentage
	Daniellia oliveri	Pterocarpus erinaceus
10	18	25
20	12	14
30	9	10
40	8	9
50	7	8
60	7	7

Source: Van Laar, 1981: in Jensen, 1995

Species	Percentage bark	Number of measurement
Pterocarpus erinaceus	17.6	154
Khaya senegalensis	15.6	58
Combretum fragrans	14.7	29
Bombax costatum	14.1	14
Daniallia oliveri	14	112
Burkea africana	12.7	47
Isoberlina doka	10.9	87
Anogeissus leiocarpus	8.1	119

Source: Anderson et al. (1991)

c) Volume tables

Volume tables are divided into two categories:

- Stand volumes tables. The total volume carried by a particular area is expressed in terms of the features of the associated stand: the number of stems by diameter class, average height, the height of the tallest trees, the number of trees less than 5 m high, etc. These tables are sometimes used (even though they are often criticized by users) for dwarfed stands or those with many small stems which are plentiful in the dry zones; and

- Individual tables. These are used to calculate the volume of a tree of a particular species. In practice, this type of table is applied to species with comparatively large merchantable diameters, often destined for ‘wood for rural construction’ and timber production. When no ‘serviceable’ stand volume tables exist, the individual tables are also applied to trees with a small diameter.

The minimum diameter (minimum diameter above which, trees are inventoried ‘minimum inventory diameter’) is of crucial importance when estimating resources, particularly in thickets, or on steppes and shrub savanna, where women responsible for collecting fuelwood often cut woody plants of less than 10 cm diameters. Several management plans implemented in the 1980s were based on inventories with an initial ‘minimum inventory diameter’ of 20 cm. Diameters of 20 or even 10 cm can lead to a substantial underestimate of the fuelwood resources in some regions. The national inventory of Mali, in the Doukoloma Forest (rainfall of 750 mm/year) showed that 19 percent of the total volume of this forest was made up of trees with a basal diameter of less than 7 cm. In Niger, in the Faïra Forest which is mainly leopard bush, the volume produced by stems of 2 to 4 cm in diameter (measured at 25 cm above ground level) accounted for 68 percent of the volume of growing stock (Mengin-Lecreulx and Chabannaud, 1986).

A minimum diameter of 7-10 cm could be envisaged but only for inventories in regions where the fuelwood demand is not too great (however the number of shoots of a smaller diameter must be counted in order to estimate the regrowth potential). Conversely, in highly degraded woody formations, such as one finds in the Sahelian area, the lack of information available as a result of only considering minimum diameters of 7 cm means that it is necessary to carry out a comprehensive inventory of all growing stock.

d) Stacking (piling) wood and measurements’ conversion (weight, volume)

The coefficients for stacking wood in steres vary widely from one place to another. In theory, the stacking coefficient should not exceed 0.785 (the volume obtained when all the stacked wood is of regular shape and circular in diameter). In the literature the coefficients vary according to category, species and environment, from 0.2 for large gnarled balks to 0.8 in the exceptional case of quartered logs. For each forest managed it is therefore advisable to calculate them.

Some authors, in the absence of accurate figures have simplified the volume/weight conversion by estimating that a cubic metre of green wood weighs 0.5 tonnes of dry wood with 12 percent moisture, which could lead to skewed estimates. In the same way, the stere/volume conversion has for a long time been 0.5 m³ for a stere.

Depending on viewpoint and trade, the terms for evaluating the resource vary. Some speak in terms of ‘steres’, others of ‘cubic metres’, some in weight or energy equivalent. Economists prefer to speak of quantities of fuelwood used or consumed in tonnes rather than in steres or cubic metres. Tonnage measurement presents one major drawback - transport costs depend very largely on estimating the volume taken up by the cut timber.

Several studies have made it possible to establish correspondences between all these units.

Arbonnier and Faye (1988) offer the following figures for the Koumpentoum Forest (Senegal):

- for very small-wood (2 cm<diameter<4 cm): 1 stere = 0.26 m³; and
- for small-wood (4 cm<diameter<8cm): 1 stere = 0.31 m³.

- Small-wood: 3-6 cm in diameter
- Medium-wood: 7-12 cm in diameter
- Large-wood: ³ 13 cm in diameter

Diameter at thin end		3-6 cm			7-12 cm			³ 13cm
		N	m3/stere	Stere/m3	N	m3/stere	Stere/m3	N	m3/stere	Stere/m3
Situation	North of isohyet 900 mm	52	0.25	3.9	40	0.39	2.6	34	0.51	2.0
	South of isohyet 900 mm	57	0.28	3.5	58	0.47	2.2	27	0.57	1.7

N = number of measurements

3. Growth and productivity

Evaluating the annual productivity of forest stands is very valuable. It makes it possible to establish the allowable cut and the potential stand renewal. It may be said that while the inventory and the associated information are indispensable for the spatial management of a forest, evaluating its productivity is necessary for its sustainable management.

It is important to note that for the forest manager, the productivity of a stand must be considered in relation with the logging and management methods used (clear felling, coppice selection, etc.).

The methods for measuring global production capacities that are being addressed refer to clear felling, namely the simple coppice management system. There are three possible approaches:

- measuring production at a given age of the stand;
- setting up permanent experimental plots; and
- studying growth rings.

This is the most commonly used method. It is rather inaccurate, because even when the year of the previous clear felling is known, which is rarely the case, the local people have generally continued to extract wood either by coppice selection or by selective cutting. In the case of clear-cut felling followed by the total protection of the compartment, production P is simply calculated as follows:

P = V/A

Where:	V =	the actual volume (in the case of clear-cut felling) or the estimated volume (by volume table) in cubic metres, steres or kilograms, and
	A =	the age of the forest, or the time since the previous clear felling (in years).

Unfortunately many writers give an arbitrary age to a particular stand, which leads to miscalculated productivity.

b) Setting up permanent experimental plots

These are perfectly defined plots, identified on the ground (using concrete boundary marks, corner ditches, paint marks on the outermost trees, individual numbering on the bark repeated each year for the full-grown trees, detailed mapping, etc.) to be the subject of a periodic measurement and observation programme (growth, regeneration, mortality). These regular intervals will be determined by the objectives and the budgetary funding available, the time for drawing up the inventory and the growth rate according to the climate. This might be every three or five years, for example. These plots could be the ones that have previously been harvested in order to establish the volume tables. This is the most accurate method, but it requires regular monitoring and is costly. It is advisable to work on the basis of a network of permanent plots set up in different environments.

At the present time in West Africa it is only in Burkina Faso (the Laba, Tiogo, Gonsé, Bissaga, Yabo, Sa and Toumoussèni Forests) and in Niger (the Tientiergou Forest where 35 monitoring plots of 0.1 ha each have been installed) and in the Gambia (in three forests under the German-Gambian project) that these types of experimental plots have been very recently installed.

This method of calculation is very demanding and costly, but the advantage is that it can be adapted to different types of exploitation. Indeed, if any method other than clear felling is used, the other approach has to be used on the permanent plots and the monitoring arrangements must then be adapted.

c) Study of growth rings

Even today many people interested in trees and forests believe that there are no annual growth rings in the wood of species growing in the tropical regions. In the wake of the work by Coster (1921-28) which showed the existence of annual growth rings in the trees in Java, several researchers have tried to demonstrate the existence of annual growth rings in tropical woods, but few of them have been interested in the trees growing in dry zones, except for Fahn (1958; 1959; 1981) on the Tamarix and Acacia, and Mariaux (1967; 1975; 1979) on a number of Sudano-Guinean species, studied using the annual marking method (Box 14).

Annual growth marks do exist in all species. But for the most, they are either not at all visible or not easily visible to the naked eye. If the formation of the wood continues at a regular pace through the alternation of a dry season and a rainy season, the visibility of the growth rings is not due to the rest period (the dry season in the tropics or winter in temperate climates) but to a particularity of the wood. There may be a difference in the look of the initial wood in one year compared with the final wood in the previous year, or there may be a characteristic line between these two zones. In fact the visible expression of the annual growth rings depends on the anatomical structure of the wood of the botanical genus in question and not on the features or the length of the vegetative rest period.

As far as we know at present, very few species in the dry tropical zones of Africa do not have visible annual growth rings: Commiphora africana and Khaya senegalensis are the two that are known. But it must be admitted that the detection of growth rings is not always easy in some species such as Boscia angustifolia, Daniallia oliveri, Lannea acida, Sclerocarya birrea and Anogeissus leiocarpus or with slow-growing trees, particularly the Parinari curatellifolia and Butyrospermum paradoxum.

In the American continent very few studies have been carried out on trees of the tropical dry zones. In the Gran Sabana region in Venezuela, Worbes (1989) has shown the limits of the growth rings marked by a thin line of parenchyma, highlighted by variations in the spacing of the parenchyma lines during growth in the cases of Eschweiler sp., Ficus sp., and Pouteria elegans (= Neoxythece elegans). He makes no mention of any anatomically distinct growth rings in Tapira guianensis.

The growth rate of the trees in dry tropical zones analysed in terms of growth rings is very variable depending on the species, but also on the site and the individual tree. An attempt to link annual rainfall and the quantity of wood formed by Acacia tortilis subsp. raddiana (Mariaux, 1975) produced no results because while on the one hand the amount of wood formed annually is difficult to estimate when the width of the ring varies continuously around the circumference, the amount of useful rainfall used by the tree is even more difficult to define. This is particularly true since the adult trees have sufficiently deep roots to reach the water table under favourable conditions and have quite important reserves so they are not affected by certain changes in the climate. Moreover the damage caused to trees (pruning/lopping or fire) certainly affects the annual formation of the wood more than variations in rainfall do.

Considering the breadth of the annual growth rings or the relationship between the diameter and the age of certain trees, it is hardly possible to quantify the growth rate of all these species. It is certainly true that the heavier species, Acacia spp., Combretum spp. or Terminalia laxiflora have narrower growth rings than the lighter woods as a whole, such as Faidherbia albida, Ficus dicranostyla or Parkia biglobosa, for example. Nevertheless the heavy wood trees can grow quickly (with growth rings 4 mm and over) in the case of certain specimens of Afzelia africana, Isoberlina doka, Prosopis africana or Butyrospermum paradoxum.

In fact the most practical technical contribution to the study of growth rings would be the possibility to date a stand which had been clear felled, for example, whose historical background is unknown. The fact of being able to establish a rotation cycle with fairly considerable certainty is essential for productivity calculations.

Box 14: The annual marking method What is the best way to identify a number of wood layers, each corresponding to one vegetative year and how to identify one or more features of the wood plane by tracing out the boundaries of these annual layers? One of the simplest methods is by annual marking. Each year, when the main dry season arrives in which the trees for the most lose their leaves and put their cambium at rest, a cut is made in the bark as far as the cambium which must be destroyed or damaged. This cut must be 4-5 cm in length and be as narrow as possible, only 2-5 mm at the most. A cut is made each year into the tree, always to the right or always to the left, about 1 cm from the previous one and always at the same height, generally, for the sake of convenience, at shoulder height. This slight wounding of the bark and the cambium will cause scarring which will be subsequently quite visible in the wood, and lead to the necrosis of the subjacent tissues. In many species, this necrosis creates rapid local heart-wood formation, and a small coloured stain indicates the place where the sapwood has been injured, being always light in colour. The concentric layer of wood between the two scar marks of the two cuts made at one-year intervals is the layer of the wood formed by the tree during the course of that year. Very often the scarring and the cut made during the vegetative rest period is on a line or on a fine strip of wood which can easily be seen, and this indicates the limit of annual growth. These operations last a long time. It is preferable to make four marks each year on each tree, and one year after the last marking. The period between commencing the experiment with the first cutting and observing the wood after felling is therefore four years, but four annual growth rings can then be studied. It is possible to fell one tree earlier, two or three years after the first marking, but one or two growth rings alone will be available for interpretation. The growth rings are examined on a portion of the bole making a cross-section through all the points where the scarring is visible on the bark, which remains there for a long time. In order to characterize the annual growth rings properly, and their boundaries in a particular species and to discover the various aspects of any false growth rings and their possible frequency, a number of growth rings must be observed in several trees. Four annual cuts made into five trees of the same species which makes it possible to study the characteristics of 20 annual growth layers seems to be an adequate minimum to define the look and the regularity of the growth rings in the particular species. It is advisable, as far as possible, to choose trees which are 20-40 cm in diameter, and which seem to be healthy. Furthermore, since the experiment takes a long time and the life of a tree is subject to many unpredictable events, it is recommended to cut into six or seven trees in order to be sure of being able to harvest five. There are two other suggestions for ensuring that the experiment is successful: - when cutting the bark use a solid knife with a retractable blade and a small screwdriver to extract the sliver of bark. It is necessary to scratch the bottom of the cut with the screwdriver in order to be sure that the cambium is damaged, because the inner bark may have a number of layers of leafy liber deep down in some species; - the bole round section of which a portion cuts through all the old slits should be completely dry before being emery-papered. The better the preparation of this section bearing the scarring, the more easily identifiable will be the annual growth rings. It is recommended to finish the wood with a 320 or 400 emery paper. Source: P. Detienne, CIRAD-Forêt (according to Mariaux, 1967)

4. Some data on productivity

Little data is available on productivity. At the beginning of the century, foresters estimated the potential rather roughly. In Senegal, “the felling on the edges of the railway line between Kaolack and the Mali border, in the gonakiés (Acacia nilotica) forests (in the Podor Department) produced 50-80 steres per ha. The stand closed very quickly and 20 years later was able to produce 40-65 steres.” (Giffard, 1974)

Clément (1982) proposed for West Africa, productivity estimates in graph form and in terms of rainfall and forest status (Figure 7). The following equation gives the relationship between the potential productivity (expressed in m³/ha/year) of an unprotected stand, and annual rainfall (P in metres) at rotation age 20 (Table 12):

Productivity = i₀ = 0.05129 + 1.08171 × P²

Clément also defines the maximum productivity values for savannas either protected against bush fires or under management:

i_max = 1.25 i₀;

the minimum productivity values in the degraded zones:

i_min = 0.25 P².

Clément knew perfectly well that some of the very rare data available in 1982 were sometimes very approximate (for example, the age of the stand was estimated arbitrarily when there were no even-aged compartments).

A provisional summary drawn up by Goudet (1985-a) on West Africa shows the magnitudes of forest cover production natural stands not subjected to degradation either by overextraction of wood products, or overgrazing and/or late annual bush fires. Estimates of the maximum production of a stand without any particular protection made by Catinot in 1985 have also been addressed (Table 12).

Figure 7: Productivity estimates in terms of rainfall and state of the stand

Source: Clément, 1982

Since this study, a few production figures have been published. In most cases the new results obtained in different countries (Sefa and Koumpentoum in Senegal; Yabo, Bissiga, Sa and Toumousséni in Burkina Faso, etc.) do not contradict Clément’s proposals, which can still be used as benchmarks, provided they are referred to with caution and a critical mind. Only the Gonsé site in Burkina Faso (Nouvellet, 1993) has actually displayed a level of production two or three times equivalent to those forecast; however, the measurements refers to the total biomass (excluding leaves and roots) whereas, current evaluations generally refer to wood dimensions greater than 10 cm in circumference. This 2.95 m³/ha/year (mean annual increment) includes therefore wood of minor dimensions (circumference less than 10 cm) which accounts for 1.24 m³. In order to be able to compare the Gonsé’s production with those based solely on wood dimensions greater than 10 cm in circumference, the production should be put at 1.71 m³/ha/year (rather than 2.95 m³/ha/year) as indicated in Figure 7.

Elsewhere, the results did not reach up to expectations. In the Koumpentoum Forest (Senegal) for example, maximum production was put at 0.3 m³/ha/year, but in this case one also has to add the customary extraction for which no figures are available (Arbonnier and Faye, 1988).

The production of three natural forests in Burkina Faso has been estimated using figures derived from monitoring the coppice cuttings implemented in 1983 and in 1985 (Kaboré and Ouedraogo, 1995):

- Bissinga Forest: 0.80 m³/ha/year
- Sa Forest: 1.04 m³/ha/year
- Yabo Forest: 0.61 m³/ha/year

Rainfall in mm	Clement 1982	Goudet 1985-b	Catinot 1985
300
		0.1 à 0.5
400
			0.25 à 0.75
500	0.32
600	0.44
		0.5 à 1	0.50 à 1.75
700	0.58
800	0.74
			1.25 à 2.50
900	0.93
		1 à 1.5
1000	1.13
			1.75 à 2.75
1100	1.36
1200	1.61

NB: The 1982 data were underestimated according to some writers. This is both an advantage, in that it encourages caution, but it is also a drawback because it has led certain economists and technicians to infer that the natural stand is uneconomical.

Table 13: Annual average growth rates calculated in the Gambia over four years (1990-1993) for four species

Species	Bama Kuno forest			Katilenga forest
	N	ID	IH	N	ID	IH
Khaya senegalensis	7	1.2	0.6	26	1.4	1.4
Pterocarpus erinaceus	34	0.5	0.7	48	0.4	0.9
Daniellia oliveri	19	0.3	1.2	31	0.5	0.9
Cordyla pinnata	6	0.4	0.6	6	0.1	0.8

N = number of trees measured
ID = mean annual increment in diameter (cm/year)
IH = mean annual increment in height (m/year)

It is difficult to compare production results from East and South Africa, with those of West Africa. Inventories differ in terms of accuracy, situation (local inventory work in West Africa and regional/national in East Africa) and in terms of their presentation. However it would seem that Clément’s model is only partially supported by East Africa production data (Wormald, 1984):

- in Somalia, with an average rainfall of 600 mm productivity has been put between 0.15 and 0.7 m³/ha/year for bush vegetation and 2-2.5 m³/ha/year for savanna woodlands;

- in Kenya, productivity ranges from 0.5 m³/ha/year in the arid zones to 3 m³/ha/year in the more humid parts in the arid zones;

- in Tanzania, where 60 percent of the country is arid or semi-arid, productivity ranges from 0.3 m³/ha/year in the thicket formations to 1.5 m³/ha/year in dry miombos;

- in Zambia, with rainfalls ranging between 600 and 1 000 mm, management trials in the most productive zones have shown that over 20 years the mean annual increment is 2 m³/ha/year after an initial severe thinning; in the drier zones the increment rate is lower: 1 m³/ha/year; and

- in Zimbabwe, measurements made in the mopane (Colophospermum mopane) savannas gave a figure of 0.34 m³/ha/year, while for the Brachystegia spiciformis coppice, the figure is 1.69 m³/ha/year.

Kigomo (1995) mentions the following highly variable production figures for the miombos, without any further information on the age, stand type or harvesting method:

- 3.50 m³/ha/year in Malawi (Masamba, 1987);

- 1.00 m³/ha/year in Kenya (Blackett, 1994), as in Tanzania (Lundgren, 1975);

- 1.30 to 2.00 m³/ha/year for experimental plots in Zambia (Endear, 1967);

- 1.93 m³/ha/year of annual fuelwood production in the dry miombo regrowth aged 6-34 years in Zambia (Chidumayo, 1988);

- the same author notes that the exploitation of successive regrowths does not seem to affect productivity. The annual growth of the woody biomass in the regrowth is estimated at 2-3 t/ha, of which 1.1 t is for wood stacked for charcoal production. In the experimental coppice plots aged 3-29 years, the average annual growth rate was 1.97 t/ha/year and 1.68 t/ha/year on compartments aged 48-49 years;

- in Zambia, in the Copperbelt region, Chidumayo (1990) found that the average total woody biomass was 92.5 t/ha of which 3.6 percent in leaves and 96.4 percent in wood;

- in Tanzania the Tropical Forestry Action Plan estimated the volume of commercial timber from the open woodlands at 40 m³/ha (TFPA, 1989); and

- Boaler (1966) mentions a basal area in Tanzania equivalent to 14 m³/ha for the miombos.

In Central and South America, the information seems to be much less complete. In north-east Brazil, Campello (1995) cites the data from EMBRAPA (Empresa Brasiliera de Pesquisa Agropecuaria): the caatinga and the cerrado display very variable growth rates ranging from 1.2 to 8 steres/ha/year. During work in the Serido region (State of Rio Grande do Norte) the UNDP/FAO/IBAMA/FRA/87/007 project showed the following growth rates:

- 1.2 steres/ha/year for the open caatinga;
- 1.8 steres/ha/year for the tree caatinga; and
- 5.0 steres/ha/year for the dense caatinga.

5. Allowable cut and compartment layouts

The allowable cut is the yield that may be collected each year, or at regular intervals, on a particular forest area for a given period of time. In the dry zones, this allowable cut is above all related to the extraction of wood which should, in theory, be close to the productive capacity of the stand; this is done in order to ensure that harvesting is compatible with the renewal of the forest’s resources, which is one of the principles of sustainable development. However, in order to restore a degraded stand, the forest manager may also settle for an allowable cut which is inferior to the productive capacity of the stand. From a purely speculative point of view, the reverse could also apply, such as in the case of a forest stand conversion. The allowable cut is more than a quantified value, and is above all a management tool because it establishes an extraction procedure and intensity.

Generally speaking, the method which is used to mobilize the allowable yield is decisive. It is the very essence of the methodologies used for forest management and lies at the heart of the debate on the sustainability of one or other practice.

Depending upon the allowable cut, the forest is generally divided into areas of variable size called cutting units, each of which is able to contain a potential in the order of the allowable cut and which must be harvested in a clearly defined period of time (for example over one year). This applies to managements directed exclusively towards wood (timber, fuel) production.

The period between two successive cuts relating to the same area, or (cutting unit), is called rotation. In principle, since cuts are all of the same kind, the potential extracted should be of the same magnitude in order not to compromise the sustainability of the system.

This division of the forest area leads to the notion of compartment layout, which is the sum total of all the compartments meeting the harvesting objectives: cutting units, reserves, range lands, etc. Another name for this is zoning.

6. The main technical approaches

Firstly, one can separate all the approaches which require setting up a compartment layout from the approaches which do not use zoning. As far as fuelwood harvesting is concerned, the coppice selection system is considered the most viable management method in regions where there is little anthropogenic disturbance.

Approaches to establishing a layout traditionally contradict the definition of cutting units by content and capacity, with an indication of the content (Box 15). Since all these approaches are equivalent in a homogeneous environment, it would appear to be reasonable in the initial phase, to tend towards such a situation by dividing the forest area in more or less homogenous units or strata (forest stratification). The strata correspond to forest types which offer fairly important potential, but subject to different kinds of uses. This notion of use leads, for example, to classify into two different strata two identical areas in terms of forest cover, of which one is to be used regularly for livestock rearing, and the other not. It is on these strata that the compartment layout will be established.

In the dry zones, and in relation to fuelwood, it is difficult to obtain accurate and reliable volume estimates. It would therefore be fairly pointless to try to adopt a management scheme in which cutting units are defined according to volumes (of fuelwood) to be harvested. This could be otherwise when considering timber extraction from dry deciduous forests. The stratification initially carried out, makes it possible to propose a preliminary layout that brings us closer to a management scheme based on cutting units defined by area, with an indication of volumes to be extracted. It is enough, to adapt within each stratum, the area of the compartments to the ‘value’ of the stratum (Figure 8).

Figure 8: Area of the compartments to the ‘value’ of the stratum

The size of the compartments depends on the richness of the layer.

The succession of the cuttings may ignore the strata.

Since the layout is set up by incorporating all the strata, the second stage consists in defining the order and magnitude of the harvesting campaigns.

Many alternatives may be proposed here. In particular, account may be taken of the distance of the forest. In this case, two possibilities may be explored.

* With regard to areas which are distant from the villages, extraction is geared to transport facilities. This can be taken into account when setting up a layout.

* The compartment, that is to say the plot of land which will be harvested during the year, may be fragmented with one part close to the village (which will be exploited during periods when manpower is not easily available) and one part some distance away, which is used during other periods of the year (Figure 9-b). One might even envisage different ways of exploiting these two components of the compartment, noting that this alternative makes it possible to propose compartments spread on different strata.

a) The further away from the village the larger is the size of the compartment

b) The compartment is split into two parts, one near the village, and the other distant from it

Figure 10: Principle of progressive harvesting

The first year of the second rotation, Area 1 will be cut again. Since the demand will have by then increased, the reserve will have to be harvested beginning with Area 1-bis (Parkan et al., 1988).

While managing the Woro and Dialakoro Forests (Mali), the FAO implemented project, given the lack of available criteria, did not define the allowable cut either by volumes or by surface area to be harvested. It is only later that an annual allowable cut as implemented by farmers has been recognized. As a result, it has been possible to establish a pragmatic compartment layout based on field factual reality. The wood extraction is carried out progressively, while bearing in mind a number of silvicultural rules (Figure 10).

The diagram above illustrates this simple (but not simplistic) principle of progressive harvesting. It is very flexible: If the demand increases, an additional plot or area (taken from a reserved zone, set aside for this very purpose) can be added to the amount harvested for the same year. Conversely, if the villagers do not harvest the annual area that they themselves have defined (as in the case of Niger, Part Four, Case Study 4) one can remove one of the non-harvested plots. While this principle is easy to apply, it is still difficult to manage, because the supervision and sustainability issues still have to be faced.

Box 15: Management based on allowable cut by surface area (by identical surfaces) Forest management is a necessity for reasons which have to do with sound resource management, improvement and conservation. On this latter point, management based on allowable cut by surface area is not a satisfactory answer. Other possibilities exist. Management based on allowable cut by surface area provides no guarantee of the sustainability of the forest in all its diversity and productivity. It is obvious that the first thing to do when managing a forest, is to define/materialize its legal limits, map it and develop a compartment layout. This makes it possible to locate the various activities to be implemented within the forest. However, it is not enough to locate the intervention, one must also define them as accurately as possible according to the state of the art and current knowledge of the stand. An inventory of the resource potential is indispensable in order to evaluate the annual allowable cut and extract less volume than is produced annually by the forest. Establishing cutting units based on allowable cut by surface area, namely without guarantee regarding the volumes to be extracted and without defining the volumes to be harvested, cannot be regarded as sustainable management and provides no assurance of perenniality either to the forest stand or to its biological diversity. Merely establishing cutting units is not adequate management. This approach could perhaps be a first stage in the management process, but this must necessarily be followed by more elaborate actions. Knowing existing volumes and increments and acceptable extraction rates all form part of the elements that are necessary to guarantee better management than management based merely on cutting units defined by surface area. Research carried out in natural dry forests over the past 20 years or so have produced know-how about the dynamics of natural forest, their reaction to different forms of harvesting and thinning on their regeneration. This knowledge forms the basis for making silvicultural proposals. Resource inventory techniques have been used for many years, and are continually being improved. Tools therefore exist for realistic management without being simplistic, which guarantee the sustainability of the forest stands in technical terms. Sustainable forest management presupposes that harvesting is kept within limits that are not only linked to commercial considerations, which has always been the situation in the past, but also to silvicultural considerations. Cutting units based on surface area leave very little room for silvicultural considerations (perhaps a minimum diameter for harvest ability?) and impose no other constraint than the land area to be harvested. Source: CIRAD-Forêt, 1995

7. Measurement of fodder biomass and density of cover

On natural range lands, available fodder for domestic livestock comprises herbaceous vegetation, but also woody vegetation which is either accessible or made accessible to the animals.

In Africa, in the Sahelian dry zone, the grass cover is essentially made up of fine, annual grasses (Aristida spp., Chloris spp., Schoenefeldia gracilis, Cenchrus spp., etc.) which germinate when the first rains fall in June and July, and dry up at the end of September. This grass cover, largely dependent on rainfall and either be continuous or discontinuous, reaches a height of 50-60 cm, and, depending upon the rainy season, a density of cover ranging between 10 and 80 percent. Before the drought, which has affected tropical Africa since the 1970s, the hardy graminacea (Panicum turgidum, Cymbopogon schoenanthus, Cyperus jeminicus, etc.) were not rare, and their density of cover reached up to 20 percent. In the Sudanian zone, annual grasses (Hyparrhenia bagirmica, Diheteropogon hagerupii) or hardy grasses (Andropogon gayanus) are taller. They germinate and sprout when the first rain falls from April to May in a landscape which is devastated by bush fires. Recovery is more substantial, varying 40-80 percent while heights range between 1.5-2 m. However, it should not be forgotten that grasses are heliophilous and that according to the Yangambi classification, the open woodland is the only physical forest formation where the graminiceae family can develop.

Some woody species are sought after by ruminants, particularly at the end of the dry season when the grass cover is sparse: low branches, sprouting shoots, pods and husks, contribute substantially to animal feed. Young individuals are systematically browsed. Sometimes, shepherds chop off a few low branches for their livestock. In the case of tree fodder, biomass estimation which is a complex matter, is not based on the density of cover of trees.

a) Measuring the herbaceous biomass

First of all it should be made clear that in the field of pastoralism, priority is given to the herbaceous biomass that appeals to animals. The evaluation of available fodder, which is expressed in terms of kilogram of dry matter per hectare (kg dry matter/ha) only refers to the edible fraction. Edibility, in the case of animals, is a completely relative notion, not only as between different animal species but also on the degree of difficulty that the livestock have in finding food. One example suffices to illustrate this: in the Sahel, before the drought, Calotropis procera was completely ignored by the livestock, but since the drought, this is no longer the case.

The measurement of biomass is made by sampling and cutting: it has been termed the ‘destructive method’. The herb is cut with shears or a sickle within a square of 1 m. Levang and Grouzis (1980) proposed that this should be repeated at random 30 times in the Sahel, after which there is little improvement in accuracy.

It is interesting to know the total herbaceous biomass and the edible herbaceous biomass. After harvesting, the total weight is measured, and then what is considered to be edible (after observations, surveys, and according to the literature) is sorted. After homogenization, part of the mixture is taken and weighed, and then examined in the laboratory to evaluate in the vacuum dryer the dry matter content. With a simple calculation using the production per square metre expressed in terms of dry matter and area of reference, one obtains the available dry matter per hectare in kilograms.

But this figure is not sufficient because it is also necessary to know the ‘fodder value’ (nutritional value of fodder) by analysing it chemically. However the number of chemical tests, which is limited by the cost must be sufficient to be representative for the whole region. The tests can be carried out on a mixture of grasses or on the dominant species.

There are also ‘non-destructive’ methods for evaluating the biomass based on using field radiometers. It is not easy to use these in the natural environment because of the variety of flora. The cost of calibration is compounded by the cost of maintaining the equipment.

b) Measuring the biomass of fodder trees

This refers to the part that is accessible to animals, which varies with the size of the domestic animal species. Dromedaries make the most out of tree fodder as compared to goats. Furthermore, not all ruminants have the same food regime. It is therefore a complex matter to measure this biomass. The method is fairly well-established, but it is very tedious to apply.

Delacharlerie (1994), in a bibliographical summary, has listed the “methods for studying tree/shrub fodder availability, application to calculating carrying capacities”. In short, this method consists of characterizing a tree stand and choosing a sample plot (a compartment containing some ten trees). The biomass is evaluated in terms of the species that are known to be attractive to the livestock, and, species by species, distinguishing the leaves, fruits, etc., according to the growing seasons.

At the present time, because of the destructive and tedious character of direct measurement, the tendency is towards allometric relations. However the formulae that have been established (Box 16) must be adjusted locally by the standard branches method.

Only part of this total biomass which is evaluated in the field is actually consumed by the livestock. That is the accessible part using height and penetrability as evaluation criteria. Ickowicz (1995) brought all these together under the term ‘useful volume’. For this author, tree fodder is characterized by the volume of its accessible part (penetrability) namely the outer part of the crown below 1.5 m. It is therefore this volume which must be taken into account (Figure 11).

It is not easy to measure the edible woody biomass. The methods described here have been drawn up for the Sahel and the dry zones in America and have produced relations which must be validated for each situation. They make it possible to assess the magnitude of the fodder contribution from the woody formations studied (Tounkara, 1991).

Box 16: Allometric relations between the maximum leaf biomass and the measurement of wood species
SPECIES	ALLOMETRIC RELATIONS	SOURCE
Acacia laeta	BM = 142 D + 216.6	Piot et al., 1980
Acacia tortilis	BM = 52.5 D - 44.64	Piot et al., 1980
	BM = 0.5 C2.35	Cissé, 1991
Acacia senegal	LnBM = 1.40 lnC +0.46	Poupon, 1980
	BM = 14.05 C1.40	Cissé, 1991
Boscia senegalensis	LnBM = 0.47lnS + 0.77lnN + 0.9 lnH - 4.85	Cissé and Sacko, 1987
	BM = 2.34 C1.88	Cissé, 1991
Combretum aculeatum	BM = 60.57 H - 17.66	Piot et al., 1980
	BM = 1.55 C2.33	Cissé, 1991
Guiera senegalensis	BM = 3.09 C1.89	Cissé, 1991
Pterocarpus lucens	BM = 0.95 C2.07	Cissé, 1991
Ziziphus mauritiana	BM = 1.38 C1.91	Cissé, 1991
Ln = Naperian Logarithm BM = Biomass at maximum flowering (grams of dry matter) D = Basal diameter of trunk (cm) C = Circumference of trunk 40 cm above ground (cm) H = Height of crown (cm) S = Crown cover (projection of crown to ground) in cm2 N = Number of clumps Source: Ickowicz 1995, in: Delacharlerie 1994

Figure 11: Total, available and accessible biomass

c) The fodder value

In order to express the contribution of natural vegetation as livestock feed, the criteria to be taken into consideration are the production of edible matter (measured in the field and expressed in kilograms of dry matter per hectare), the energy value and the nitrogen value (Box 17). In addition to these quantitative criteria there is also one quality criterion: palatability, which is not always easy to establish. This requires a period of field observation throughout the year. But at all events, when there is severe drought, most of the species which may be considered as not being desirable are eventually eaten.

Box 17: Food value of fodder In order to express the food value of a type of fodder, the following notions are used: Energy value: the Fodder Unit or more recently the Milk Fodder Unit (in units of kg/dry matter), and Meat Fodder Unit. This is mainly inferred from digestibility: the Digestibility of Organic Matter (in percentage form). The nitrogen value: the Total Nitrogenous Matter, the Digestible Nitrogenous Matter, the Proteins Digestible in the Small Intestine (in grams per kilogram of dry matter). It is also necessary to know the dry matter (the dry matter content of a moist product is given as a percentage of green matter) and the Voluntarily Ingested Dry Matter (in grams per kilogram of metabolic weight).

The value of the palatable/edible fodder produced is measured in the field and the energy value is provided by laboratory analysis (Table 14). The energy value, which is obtained on the basis of digestibility, either measured or estimated, is expressed in terms of Meat Fodder Unit and Milk Fodder Unit.

Table 14: Composition of certain ruminant feed-crops

DESIGNATION			TIME	PERIOD OF VEGETATION	DM P 100	CHEMICAL COMPOSITION (gDM.kg-1)					ENERGY (KG/DM)			N. VALUE (gDM.kg-1)
						OM	TNM	GC	Ca	P	FU	MFU	MtFU	DNM	PDIN	PDIE
A. NATURAL AND CULTIVATE GRAMINACEAE GRASS (EXAMPLES)
Annual Sahelian/Sudanian grasses
	Aristida mutabilis
	- Vegetative stage		RS7		23	909	133	311	5.3	1.9	0.67	0.78	0.70	93
	- flowering stage		RS8		32	920	72	356	2.9	1.8	0.37	0.55	0.45	32
	- fruiting stage		RS9		49	926	59	365	3.4	1.4	0.33	0.53	0.42	19
	- straw		DS10-12		94	934	32	417	3.4	1.0	0.24	0.46	0.35	<0
		straw	DS3-5		95	942	6	438	2.3	0.3	0.21	0.44	0.33	<0
Hardy Sudanian/Guinean grasses
	Andropogon gayanus
		- vegetative stage	RS6-7		28	918	111	311	3.7	1.9	0.56	0.69	0.61	71
		- flowering stage	RS8		27	912	86	318	4.6	1.6	0.45	0.61	0.51	46
		- post-cut re-growth	DS2	15 days	38	905	107	295	5.0	2.3	0.55	0.68	0.60	67
		- post-cut re-growth	DS2-3	25 days	39	922	96	304	4.4	2.0	0.50	0.65	0.56	56
		- post-fire re-growth	DS1	20 days	37	915	84	284	5.3	1.5	0.47	0.63	0.53	44
		- post-fire re-growth	DS2	60 days	37	899	63	328	4.3	1.4	0.36	0.54	0.44	23
		- post-fire re-growth	DS2-3	100 days	44	936	57	344	5.5	1.4	0.35	0.54	0.44	17
B. NATURAL HERBACEOUS DICTYLEDONAE (EXAMPLES)
Annual Sahelian/Sudanian legumes
	Zornia glochidiata
		- vegetative stage	RS7		15	918	200	222	6.9	1.8	0.57	0.72	0.63	146		23
		- flowering stage	RS9		32	934	160	277	9.7	1.4	0.46	0.64	0.53	104		26
		- fruiting stage	RS9		33	929	134	297	13.2	2.2	0.38	0.58	0.47	85
		- straw	DS11		94	947	82	367	5.4	0.9	0.23	0.47	0.36	40		43
C. WOODY FODDER PLANTS (EXAMPLES)
Green leaves
	Acacia linarioïdes (Australian sp.)		DS	= 30	941	125	312	7.2	1.1				8			15
	Leucaena leucocephala		DS		903	181	231	2.1	1.8	0.07	0.37	0.25	128			20
	Piliostigma reticulatum		DS	= 30	909	91	297	18.4	1.2	0.65	0.77	0.69	9			6

Source: IEMVT ISRA

Figure 12a: Value of the oligo-elements (PHOSPHOROUS)

Figure 12b: Value of the oligo-elements (LEATHER)

Figure 12c: Value of the oligo-elements (ZINC)

Source: Guérin et al, 1994

For an African zebu the average daily consumable fodder is put at 6.25 kg of dry matter on average, with sharp seasonal variations, depending upon the type and availability of the fodder. The benchmark livestock unit is the Tropical Livestock Unit (TLU), namely an adult zebu weighing 250 kg on the hoof. When using the results of a survey, the livestock population is converted into this unit which facilitates the carrying calculations (cattle: 0.75 TLU; sheep and goats: 0.15 TLU; camelidae and horses: 1 TLU; donkeys: 0.5 TLU).

Table 15: Tropical Livestock Unit production in terms of the value per kilogram of dry matter (assuming an average ingestion of 2.5 kg dry matter per 100 kg weight on the hoof)

Dry matter contribution Level of Production	MFU	DNM (g)	DNM. FU
Maintenance/day (per LTU)	0.45	25	55
Weight gain/day-1
100 g .....................................	0.50	29	60
300 g ....................................	0.60	37	60
500 g ....................................	0.70	45	65
700 g .....................................	0.80	52	65
Milk production/day (1.j-1)
11 .......................................	0.50	34	70
31 .......................................	0.60	53	90
41 .......................................	0.70	63	90
61 .......................................	0.80	82	100

Source: Boudet, 1991. For the legend, see under Table 14.

8. Evaluation and monitoring of resources: the new tools

Producing cartographic documents should form part of every forest management operation. These tools make it possible to visualize the information that the forest manager requires such as forest inventory, stand features, potential and dynamics of the formation, the ecological conditions, etc. All of which information is indispensable and once in map form helps identify specific relations and makes it possible to gain a better understanding of the situation.

The traditional mapping procedures, very largely manual, no longer meet today’s requirements, which make it necessary to be able to continually update the information available on phenomena which are rapidly evolving and involve a wide area under study. The techniques used have made considerable progress over the past two decades, and they are increasingly developing today. In this area, the new tools that have been developed relate in particular to data acquisition, processing and output.

a) Remote sensing

Non-satellite remote sensing using aerial photographs (panchromatic and infrared emulsion) is an extremely effective tool for work on a very large scale, given the quality of the information supplied, but it is incompatible, being constrained by their date of issue (the data of aerial missions are often outdated), the necessity for repetitiveness and the large areas under study.

Since 1972, earth observation satellites are making it possible to automatically acquire vast quantities of data, particularly on areas that are difficult to access (Box 18).

The scale, the frequency of repetition of the viewing, the multi-spectral features, their numeric form, the stereoscopy for certain instruments (SPOT since 1986) are features that are particularly useful and are constantly being improved.

The satellite sensors record part of the electromagnetic radiation emitted, reflected or scattered by the object under observation. The sensors used to study wood formations in dry tropical zones are of three types, depending upon their spatial resolution.

* High and medium resolution sensors

Box 18: Remote sensing in Mali A new mapping of wood resources and land occupancy in Mali by remote sensing: the Inventory of Wood Resources (Projet d’Inventaire des Ressources Ligneuses) project. This project, which was implemented between 1985 and 1991, comprised two phases: - the study of test zones representative of the main ecological environments in Mali and the production of maps on a scale of 1:50 000 based on aerial photographs and forest inventories; - the mapping of the part of the country south of the 17th parallel, using SPOT satellite images and the results from the test zones. The study has resulted simultaneously, in building up a large database on the plant formations in Mali, and in producing a land occupancy cartography. The maps were published on a scale of 1:200 000, with two themes: plant formations and agricultural occupancy. A general summary map containing all the data was also published at scale of 1:1 000 000. This project, which brought together land inventories, satellite data and maps, is an indispensable source of information today for evaluating and managing Mali’s natural resources.

The amount of information available in image form is ten times the amount produced by Landsat MSS (ground resolution of about 80 m). Future programmes will have even finer resolution (10 m in multi-spectral and 5 m in panchromatic for SPOT 5 in the year 2000).

* Low resolution sensors

Furthermore, the recurrence of the data (two images per day) and the width of the recording band (2 700 km) are major advantages for global level monitoring of certain phenomena such as fires.

Interpreting and validating low resolution images generally requires associating high resolution images during the sampling procedures. This approach will be facilitated even more with SPOT 4, scheduled to be used in 1997, which can simultaneously acquire high and low resolution views using the VEGETATION and SPOT.HRV.IR instruments.

* Using vegetation indices

The study of seasonal variations or annual variations in this index reflects the changes in the photosynthetic activity of the vegetation. Periods of drought are therefore marked by a substantial fall in the NDVI.

* Radar images

NB: There is no doubt that satellite data constitute a very valuable working tool, which has led to changes in the approaches to environmental studies. However the data form only one part of the information required. It is also essential to be able to get down to ground level (the ‘field truth’ is fundamentally important) or better still begin with a field problem in order to better understand what one is trying to understand.

b) From digital cartography to Geographic Information Systems

Mapping methods, thanks to the development of computer science, have undergone far-reaching changes over the past few years. The maps made in the traditional manner are no longer suitable to meet present needs. They often require much time for processing, are difficult to file and disseminate and their updating is always a painstaking operation, which explains why they remain static.

Initially the new techniques related to the drawing: digitizing the outlines, automatic shaping, instantaneous rescaling, etc. All these operations have become more flexible and are faster to implement, making updating and cartographic restitution easier. The data which is automatically acquired is stored and used to build up a cartographic database which can be used at any time.

The range of different software used for all these operations form part of the so-called CAM (Computer-Aided Mapping) systems. However these tools do not make it possible to manage the data associated with the drawings, and these must therefore be processed using an independent software known as Data Base Management System. The GIS comprise all these functions and it manages within one and the same system, the cartographic data and the corresponding information.

The two major differences between a GIS and the Computer-Aided Mapping software are the fact that the GIS is designed around a Data Base Management System and the fact that it can carry out spatial analysis operations.

To a certain extent the GIS bring together all the advances made over the past few years. They are therefore capable of carrying out four sets of different tasks: collecting, managing, analysing and releasing cartographic information.

Because of their potential, the GIS are a major tool for managing natural resources. The potential for storing useful information on the region under study (satellite data constitute a large part of this), the capacity to be able to cross-link them to highlight specific phenomena, such as the degradation of forest areas, and explain the process taking place, based on all the factors of the physical and human environment, justify the use of these systems provided, naturally, that one is aware of their limitations.

Many projects use them both in the area of research and development, but there are still methodological and strategic problems with their use. In addition to the quality of the data stored and processed, the results obtained have to be analysed very cautiously, depending upon the methods used. In the same way, the abundance of data and the ease of ‘superimposing maps’ should not create the mistaken impression that there exists an amalgam of relations, because this is not the case.

Lastly, the installation of such systems is extremely costly in terms of hardware, but also in terms of data capture, recovery and updating, as well as manpower training.

c) Geographic positioning. An operational system: Global Positioning System

One of the merits of the GIS is that it makes it possible to cross-link information. It is therefore necessary to use the systems very carefully and particularly recall that the geographic positioning of the data demands great accuracy.

At the present time, objects on the ground (crossroads, trees, bridges, etc.) can be captured much more easily using a Global Positioning System (GPS), which are also known as ‘Navstar’ (Navigation Satellite Timing and Ranging) systems.

A GPS system comprises three sectors:

- space sector currently made up of a constellation of 24 satellites in orbits at about 20 000 km;

- control sector with five American stations; and

- user sector in the shape of a set of sensors which decode the GPS signals (entirely under the control of the American Defence Department).

- improving the accuracy in verifying the ground position of satellite data to be interpreted;

- initiating the materialization of forest strips as a basis for establishing a forest inventory sampling system;

- delimiting and subsequently mapping very accurately the areas to be put under management; and

- locating benchmarks: it should be added that in the dry tropical zones it is often difficult to find noticeable benchmarks, and in such cases the GPS is an indispensable system to be able to rapidly position the field observations without using complicated methods which require professional expertise.