Previous Page Table of Contents Next Page


Section 1 - Module 6: Range resource evaluation


Part A: Definitions and concepts
Part B: Purposes
Part C: Types of data
Part D: Methods of data collection
References


Part A: Definitions and concepts


Vegetation definitions and concepts
Management definitions and concepts


This module is concerned with livestock-range interactions. It is particularly applicable to those systems of livestock production in which communal grazing is practiced and is perhaps most relevant to the semi-arid and arid regions of sub-Saharan Africa, where pastoral and agropastoral systems predominate.

This module follows the same general outline adopted for the other modules in this Section, except that it begins with a definition of the concepts which form the basis for a proper evaluation of range conditions and for any study of livestock- range interactions.

Vegetation definitions and concepts

Vegetation structure. This term refers to the arrangement, spacing and size of plants within a given area. Particular structural features of plants which are often measured are height, volume, stem size, crown size and spacing.

Species (or floristic) composition. This term refers to the proportion of different plant species found in association within a given area (Tothill, 1978). The proportion can be estimated by using one or more of the following measures:

· dry-matter weight of individual species per unit area
· number of plants of each species per unit area (i.e. density)
· occurrence, i.e. the relative frequency of different species per unit area, and
· cover1 per unit area occupied by individual species.

1 Cover is defined as the proportion of the soil surface covered by vegetation. That might be near 0% in the desert or in an unplanted cultivated field and as high as 100% in dense grassland or forest. The kind of cover should always be defined, thus: basal cover (percentage of the soil surface occupied by the bases of plants), litter or mulch cover, rock cover, tree canopy cover or foliage cover (Heady and Heady, 1982).

The particular measure chosen will depend on the objectives of the study and on the type of vegetation studied.

Note: The composition of a plant community can change because of many factors, including grazing practices, burning, drought and temperature effects, pests, diseases and erosion. Depending on the nature of this compositional change, the productivity of an area (in terms of its capacity to support livestock) may change. A change in plant composition results because of the relative adaptability of different species to these influences (Stoddart et al, 1975; Butterworth, 1985). The process of change from one equilibrium compositional state to another is known as plant succession.

Plant biomass. Plant biomass is the total dry-weight equivalent of plant matter per unit area. Vegetation utilised for grazing and/or browsing may be only part of total biomass. A distinction is often made between above-ground biomass and standing crop. The former includes plant litter in the estimation of dry-matter production in an area (Clarke, 1986).

Significant intra- and inter-seasonal variations in plant biomass and standing crop can be expected for most African range production systems, where rainfall is usually seasonally distributed and often highly variable (Sandford, 1982). Under similar rainfall conditions, soil type will have an effect on plant biomass production (Abel et al, Vol. 1, 1987). Since plant biomass and standing crop are affected by species composition and density, composition should be related to productivity (in terms of quality and quantity) when assessing vegetation-livestock interactions. The palatability of grass and browse is also affected by species composition (Module 7).

Range condition. This is the state and health of the range, which can be assessed on the basis of an area's vegetation composition, plant vigour, ground cover and soil status (Pratt and Gwynne, 1977).

The concept of 'condition' implies that an optimal or desired vegetation cover (in terms of quantity and composition) exists for each particular land system. However, since it will often be uncertain what the desired or 'optimum' condition is (particularly in areas which have undergone misuse for a considerable period of time), and since optimum range condition will differ according to the manner in which the range is used (e.g. cattle, sheep, wildlife), the comparison used should be clearly stated as well as whether this is based on actual measurements or whether it is assumed.

Range trend. This indicates the direction of change in range condition over time. The detection of range trend in the early stages of change may be difficult because vegetation characteristics will often fluctuate widely as a result of seasonal variations in climate (Heady, 1981; Ng'Ethe, 1986).

Climatic variations affect the stability of the environment or range, while trends in range productivity reflect the sustainability (or resilience) of existing management practices and of the ecological system itself (see text below and Figure 1). The criteria most commonly used to assess range condition and trend are (Pratt and Gwynne, 1977):

· herbaceous cover - i.e. species composition, vigour, percentage of total cover and stand density

· shrub/tree cover - i.e. number of trees (shrubs)/unit area, species composition, height, vigour and age

· forage value - i.e. nutritive value of forage, seasonal variation, potential productivity and palatability, and

· soil stability and fertility - i.e. depth of soil, structure, rainfall infiltration and nutrient status.

Range stability. Stability is the degree to which range productivity remains constant despite normal fluctuations in environmental variables, such as climate (Conway, 1985). Figure 1 illustrates the concept by comparing a stable range production system with a relatively unstable one.

Range sustainability. The term 'sustainability refers to the ability of the range to maintain its long-term productivity when subject to particular environmental and management stresses or disturbances2 (Figure 1).

2 Conway (1985) defines stress as a regular (sometimes relatively small and predictable) change in environmental conditions. A disturbance, on the other hand, is an irregular, infrequent, relatively large and unpredictable change, such as drought.

Management definitions and concepts

Stocking rate. This is defined as the actual number, at a specific time, of tropical livestock units (TLUs) per hectare, where a tropical livestock unit is the equivalent of one bovine animal of 250 kg liveweight.

Standard livestock-unit conversion factors are often used in the estimation of area stocking rates. The factors used by ILCA are shown on a table on the next page.

A distinction is often made between gross and net stocking rate per hectare. The first concept does not differentiate between land used for grazing or other purposes (such as cropping), while the second relates animal numbers to land specifically allocated for grazing.

Figure 1. Stability and sustainability: Illustration of concepts.

A: Sustainable systems

B: Unsustainable system

Species

Conversion factor (head to TLU)

Camels

1.0

Cattle

0.7

Sheep

0.1

Goats

0.1

Horses

0.8

Mules

0.7

Donkeys

0.5

Also, the concept of en 'optimum stocking rate' has been debated in the literature (Sandford, 1982; Butterworth, 1985). The definition of 'optimum' will depend on the criteria of evaluation used or the grazing strategy adopted (see below under carrying capacity). When maximum productivity per head is the objective, a lower stocking rate will be used, but when the aim is to maximise productivity per hectare, higher stocking rates will be applied.

Under the latter circumstances, animal productivity per head will fall, but overall productivity per unit area will (within limits) normally rise (Stoddart et al, 1975; Abel et al, 1987) (Figure 2). This conclusion is particularly relevant to African livestock systems since traditional producers normally attempt to maximise returns per hectare, not per animal (Behnke, 1984; de Ridder and Wagenaar, 1986). The long-term environmental implications of these different approaches to optimising productivity will need to be carefully considered.

Grazing pressure. This is the number of animals per unit of available forage. Different animal species will have different preferences for different kinds of forage (e.g. grass or browse), and this will influence grazing pressure within a given area (Cooperrider and Bailey, 1981). The grazing habits of different animal species may be complementary or competitive: preferences may vary between seasons, thereby altering the degree of complementarily/competitiveness over time.

Figure 2. The relationship between stocking rate and liveweight gain per head and per hectare, Zimbabwe.

a Matopos, 1967

b. Marondera, 1968/69

Source: Butterworth (1985).

Grazing capacity. Grazing capacity is the maximum stocking rate which the rangeland can support without causing damage to vegetation or other range resources (such as soil and water). If the actual stocking rate exceeds this level for a sufficiently long period and leads to irreversible changes in range condition,3 the area is said to be overgrazed. An area may, however, be overstocked in the short term, without being overgrazed (i.e. the number of animals carried on the area may exceed its carrying capacity at that point in time).

3 Distinction should be made between irreversible and reversible changes in range condition. According to Abel et al (1987) loss of clay and silt particles from the rangeland is detrimental and irreversible. Loss of soil organic matter is, in theory at least, slowly reversible. Bush encroachment caused by overgrazing can be either.

Understocking may also result in a deterioration of grazing resources. This occurs when, in the absence of grazing, the vegetation composition on the range tends to change in favour of less palatable or lower-producing plant species (or plants become less palatable because of senescence). In these circumstances, selective grazing by fewer animals (or animal species) means that large areas remain ungrazed and become increasingly infested with less palatable species. Also, because grazing is confined to patches of the range, localised overgrazing and deterioration of range condition will occur.

The use of vegetation as a primary indicator of environmental degradation caused by overgrazing is based on the premise that vegetation is a reliable mirror of its ecological environment (Mouat and Johnson, 1981). The vegetation indicators which are usually considered cause for concern (Stoddart et al, 1975) are:

· low plant cover

· preponderance of plants of low palatability, and

· change from a species composition in which perennials predominate to one dominated by annuals, particularly fortes (i.e. herbaceous weeds).

Other indicators of environmental status are soil erosion, decreased soil water availability, and livestock condition or productivity which may, however, lag behind a deterioration in vegetation or soil conditions.

Carrying capacity. This is the maximum number of animals which can be carried on a given area of land, during a period of the year when the quantity and/or quality of forage is at its lowest (Butterworth, 1985), or in dry years.

A distinction between opportunistic and conservative grazing strategies needs to be made in this context. An opportunistic strategy is one in which the holder/owner varies the number of stock according to current forage availability - i.e. grazing pressure is held at a constant level. This strategy is said to be efficient if livestock numbers are varied at the appropriate time (Sandford; 1982, 1983a). An efficient opportunistic grazing strategy will not, normally, lead to environmental degradation.

A conservative strategy is one which maintains the population of grazing animals at a low level, in both good and bad years - i.e. grazing pressure is variable. Under this strategy, no attempt is made to exploit periods of forage abundance by increasing livestock numbers. It is essentially a risk-averting management practice which is less likely to cause environmental degradation.4

4 This statement needs qualification If, for instance, forage abundance were to occur over a 3-year period without a compensatory increase in livestock numbers, undesirable changes in vegetation composition could occur, resulting in lowered carrying capacities in the long term.

Neither of these strategies will be wholly attainable in practice, and both approaches to herd/flock management are likely to be evident in African livestock production systems.

The tendency to retain or increase livestock numbers, even under adverse environmental conditions, is characteristic of many sub-Saharan African production systems. Coupled with the practice of communal grazing, this tendency is often said to be causing widespread environmental deterioration, particularly in the semi-arid and arid areas of the region. However, the evidence for environmental decline on a wide scale is inconclusive (Horowitz, 1979; Breman et al, 1979/80; Sandford, 1983a). Relationships between livestock management practices and the environment have rarely been quantified and have often been misinterpreted or misunderstood. Technological interventions have thus often been inappropriate.

Part B: Purposes


Livestock- Range interactions
Scope for technological improvement


The main objectives in range resource evaluation, therefore, are to:

· quantify important relationships and interactions between livestock management practices and range resources (vegetation, soil and water),5 and

· assess the necessity and scope for improvement through technological intervention. Technologies may be aimed at improvements in range condition and/or livestock production.

5 However, the different types of cyclical change present in all ecosystems must be separated from the effects of management, to avoid confounding of the two and ascribing changes in range condition to the wrong causes (Heady, 1981).

Livestock- Range interactions

For any livestock production system, management practices will directly affect the productivity of the range.

For instance, exploitative management practices may alter species composition towards less palatable, lower-producing plant forms which are less able to support animal populations. Long-term overstocking will lead to overgrazing, soil loss and reduced water infiltration.6 Such trend relationships need to be understood since they will affect the stability and sustainability of rangelands. If used on a wide scale and over a long period of time, inefficient opportunistic (and conservative) management practices will have similar effects.

6 The effect of overgrazing on the environment will also depend on the existing level of soil fertility and, to some extent rainfall. Heavy stocking on high-fertility soils may, for instance, induce the growth of higher-producing plant species, while on low-fertility soils, unfavourable species may become predominant.

Range conditions also vary in the short term (i.e. within and between seasons), and this can affect performance levels and the management practices adopted. Seasonal variations in the quantity and quality of feed can, for instance, affect the movement of stock (e.g. herd splitting - see Module 10) as well as the amount of time an animal spends grazing/browsing (Dicko-Touré, 1980).7 These, in turn, can affect birth and growth rates, mortality rates, and milk and meat production.

7 The traditional requirements of different animal species requirements, priority rights for animals in different age/sex categories and cultural or security considerations will also affect stock movement and grazing browsing habits. Seasonal stock movement is, however, largely due to an attempt to match feed supplies with feed requirements i.e. to equilibrate grazing pressure. If the movement does not or cannot occur (e.g. because of encroachment of pastoral grazing areas by agropastoral communities), stocking rates may become too high and performance levels may decline as a result.

Proper diagnosis of short- and long-term relationships between livestock and the range is thus an essential first stage in the design of appropriate technologies. Some of the important livestock-range interactions are between:

· stocking rate per hectare and vegetation composition

· stocking rate per hectare and animal productivity per hectare and per head

· plant species composition and animal performance

· grazing pressure/stocking rate per hectare and the incidence of soil erosion (and the effect on water availability)

· livestock movements and seasonal range productivity, and

· seasonal range production and seasonal animal production.

The relationship between stocking rate per hectare and vegetation composition. Table 1 gives an example of this relationship under research-station conditions in Uganda.

The relationship between stocking rate per hectare and animal productivity per hectare and per head. In studies of this nature, variables such as birth rate, mortality rate, growth rate, meat and milk production (Module 5) could be related to grazing pressure or stocking rate in studies of this nature. Figure 2 provides examples of such a relationship for two locations in Zimbabwe.

The relationship between plant species composition and Animal performance. When examining differences in composition, the attention should be on the nutritive value of the feed consumed. This is because changes in vegetation composition are unlikely to affect performance unless they are associated with changes in the nutritive value (e.g. as a result of a higher/lower proportion of legumes in the pasture or browse).

Example:

Table 1. Botanical composition at different stocking rates, Serere, Uganda, 1970.



Botanical composition (%) at a stocking rate (animal/ha) of

2

3

5

Hyparrhenia rufa

71

68

26

Stylosanthes guianensis

19

22

20

Sporobolus pyramidalis

1

2

39

Other

9

8

15


Source: Butterworth (1985).

Scope for technological improvement

Knowledge about the factors which constrain livestock production provides a basis for technological design. Two approaches may be used to effect technological improvement, including the development of technologies which:

· WORK WITHIN EXISTING CONSTRAINTS TO PRODUCTION

In this case, an attempt is made to exploit flexibilities for short-term improvement within the system itself.

For instance, overgrazing and overstocking may be localised near water courses or water points. Range condition and animal productivity could be improved by a spatial re-distribution of grazing pressure to areas which are relatively less affected (e.g. through the provision of additional water points) (Sandford, 1983b). Similarly, grazing pressure could be re-distributed through time by adjusting stocking rates more rapidly to changes in seasonal forage availability (Abel et al, 1987) or by improving marketing facilities.

· ATTEMPT TO BREAK EXISTING CONSTRAINTS TO PRODUCTION

In this case, an attempt is made to achieve improvements of a long-term nature.

This strategy will be applicable particularly when:

· livestock production per hectare and per head is limited by range condition, or

· management practices limit the scope for improvement in range condition.

For instance, the scope for significant improvements in range condition arid livestock production per hectare will be very limited if, because of culturally entrenched management practices, the vegetation is seriously degraded. A change in community perspectives may first be necessary before improvements of a long-term nature can be affected.

An understanding of the regional distribution of livestock numbers between seasons and of the seasonal availability of forage is, therefore, an important aspect of range resource evaluation. The effect of individual animal species on the range environment should also be determined. An understanding of the economic goals of the target farmer/pastoral group is equally necessary, if the right approach to technological design is to be adopted.

For instance, the farmer/pastoralist may aim to maximise returns per hectare rather than per head, or he may prefer opportunistic stocking rate strategies to conservative ones. The approach adopted to technological improvement must reflect these goals to be meaningful.

A number of basic questions will, therefore, need to be asked when designing new technologies for improved rangeland and livestock productivity. They are:

· Will the new technology be compatible with the range environment? How will it affect range stability and sustainability?

· Will the new technology be compatible with existing management practices? If not, will a change in management be culturally acceptable or economically rational from the recipient's point of view? Which group(s) in society will benefit if the technology is adopted?

· Can animal productivity per head or per hectare be improved without making simultaneous improvements to range condition? If not, what changes will need to be made?

· Can range condition be improved without making simultaneous changes to livestock management practices? If not, which changes in management will need to be made, and what are the costs and benefits involved?

Such questions should always be asked to determine the right approach to data collection in range resource evaluation. The approach adopted will, in turn, depend on the time available, cost considerations and manpower resources, because tracking trends in range condition can be both time-consuming and costly.

Part C: Types of data


Objectives of data collection
Ground, aerial and remote sensing data


During the initial stages of livestock systems research, baseline data on the environment as well as on the historical, economic, ethnographic, infrastructural and political characteristics of the target area should be collected. Data on the environment should provide details on climate, topography, vegetation, soil and water resources and human agricultural activities (Module 1, Section 1). This information can often be obtained from secondary data sources (e.g. meteorological reports).

Objectives of data collection

To define the objective of data collection in range resource evaluation, it is useful to answer the following three questions (Clarke, 1986):

· How big an area needs to be covered to meet the overall objectives of the study?
· How often and over what time period should data collection be made?
· How detailed should information be?

There are essentially two types of study conducted in range resource evaluation (USDA Forest Service, 1981):

· resource inventory studies, and
· resource monitoring studies.

Resource inventory studies. The aim of these studies is to assess resource use and availability at one point in time. They may be part of an initial baseline data survey, or they may be discrete studies which are more detailed in nature.

Resource inventory studies provide information about livestock numbers and herd/flock structures within an area (Module 9), as well as about existing vegetation characteristics, soil loss, water resources and stocking rates. Because they rely on data from one point in time, inferences about the relationship between management practices and range condition need to be treated with caution. They may indicate potential problems in resource use and availability, but trends cannot be established. However, the determination of trend relationships is often the chief concern in range resource evaluation studies!8 (Pratt and Gwynne, 1977; Schmid-Haas, 1981).

8 Methods applicable to these kinds of studies are outlined in Part D of this module.

Resource monitoring studies. studies The aim of these studies is to track range resource trends through time and to relate these to livestock management practices. They tend to be costly, particularly in the arid/semi-arid regions of Africa where vast areas may need to be covered for several years in succession (Clarke, 1986).

They give information on vegetation trends, by monitoring successive changes in plant composition which should be distinguished from changes caused by intra - and inter-seasonal climatic influences. Successive changes are directional and relatively predictable (Heady, 1981). Meteorological data, if available over a sufficiently long period, may provide some indication of longer-term cyclical influences which affect the environment. Informal interviews with producers who have lived in an area over a sufficiently long time period may also be useful (Module 1, Section 1).

The level of detail in data collection will depend on the type of study conducted and cost, manpower and logistics considerations. It will often be more efficient to confine attention to a few key indicators of range condition rather than to attempt a broad- based study on a whole series of variables (Husch, 1981; Mouat and Johnson, 1981). This, of course, implies that the useful indicators of range condition are known by the researcher.

Ground, aerial and remote sensing data

Data used for range resource evaluation can be collected using ground, aerial and/or remote sensing surveys.

Ground surveys. Ground-sampling techniques are commonly used to Ground-truth' data obtained from aerial and satellite surveys, by providing finer detail on environmental characteristics and management practices (e.g. plant composition, grazing pressure, grazing capacity and carrying capacity). They are also used to monitor animal sex/age structure, smallstock numbers and ground-cover species composition, which cannot, at present, be accurately monitored by other means. Ground-survey techniques are relatively expensive and time-consuming.

Aerial surveys. There are two main kinds of aerial survey. Low-altitude, usually sample-based, aerial survey using both observer's direct eyesight and some photography, and higher-altitude, complete photographic coverage. In the following discussion we are referring to the low-altitude sample survey, unless the context suggests otherwise. Such surveys9 are suited to:

· estimate the numbers and densities of larger herbivores and wild animals. Where tree density is not high, seasonal stock movements can also be traced by using aerial counts or photography

· map major vegetation types in terms of physiognomy (i.e. woodland, scrubland, grassland) and basal or foliage cover

· map topography and soil types

· map of human settlement patterns and agricultural activities

· locate major seasonal water resources, and

· provide evidences of soil erosion and its causes (e.g. the distribution of erosion gullies, soil scalds and their relationship to human/livestock densities).

9 Although Clarke (1986) classified aerial surveys as the most cost-effective of the three methods, problems of aircraft availability and lack of skill in the use of the technique currently limit its application on a wide scale. In addition, much of the data obtained by aerial survey needs to be ground-truthed.

Aerial surveys conducted over a number of years can provide useful information about changes in range conditions, the distribution of human activities over time and livestock densities. General indications of the causes for and direction of change will often provide a useful starting point for more detailed range - livestock interaction studies.

Remote sensing. Satellite information has been used to map ecological resources, monitor environmental conditions and estimate changes in green biomass (Lamprey and de Leeuw, 1986). The satellite series most often used for ecological monitoring and mapping purposes are the Landsat, NOAA and SPOT-1.10

10 The satellites are equipped with sensors which record radiation reflected from the earth's surface, and the data are received in digital or image (photographic) form. Rangeland studies use satellite data based on the differential absorption of red and infra-red wavelengths by green vegetation (the so-called 'false colour' differences) to estimate changes in green cover. The data must be correlated with ground and/or aerial survey information for accurate interpretation.

The usefulness of satellite data for the estimation of rangeland grass cover is presently limited by a number of factors (Abel and Stocking, 1987):

· The techniques currently used depend on per cent cover of the green material present. This means that estimates must be confined to the growing season and that dry cover cannot be measured.

· It is uncertain whether the imagery can distinguish between the canopies of the woody vegetation layer and of the ground-layer vegetation.

· When ground cover is sparse, soil reflectance distorts the reflectance of vegetation (Robinove, 1981). Differences in soil type and topographic features create additional complications.

· Cloud cover distorts remotely sensed data, and atmospheric absorption of radiation can result in non-systematic biases in results.

Summary

Ground, aerial and remote sensing techniques complement each other (Figure 3). Each has its own particular limitations in terms of time, cost and the level of detail which can be provided, but, together, the three methods are a useful tool for a comprehensive assessment of range resources and regional land-use patterns (ILCA, 1983; de Leeuw and Milligan, 1984). They can also be used to identify specific areas requiring attention or as a basis for system description and diagnosis.

Figure 3. Complementary rangeland survey techniques.

Source: Thalen (1981).

Part D: Methods of data collection


Ground-survey methods for vegetation studies
Ground-survey methods for soil studies


This part of Module 6 outlines some of the common ground-survey methods used in the evaluation of livestock-range interactions. The general principles involved in sample selection are discussed in detail in Module 2 of Section 1.

It is not possible (or practical) to give a comprehensive coverage of remote sensing data collection methods in this manual. The literature on this topic is extensive and the user is referred to the reading list, and in particular to Clarke (1986), for some of the more important references currently available Aerial survey methods used to assess vegetation and soil resources are discussed in manuals dealing specifically with these topics. Ground and aerial survey techniques used to obtain information about livestock numbers are described in Module 9.

Ground-survey methods for vegetation studies

Vegetation studies concerned with livestock-range interactions will normally concentrate on:

· plant species composition to assess the effect of livestock grazing practices on range condition and trend, and

· plant biomass production. Used together with data on livestock species and numbers estimates, data on plant biomass production provide a basis for the assessment of:

· the quantity and quality of feed available, by season and by animal species, and

· grazing pressure, grazing capacity and carrying capacity, by season and by animal species.

· basal or foliage plant cover. Data on basal cover will be needed when vegetation composition by percentage cover is studied, while data on foliage cover will be collected when soil structure, soil erosion or moisture conservation are suspected.

The selection of the method(s) to collect such data will depend on:

· the relationships which need to be studied
· the key composition or biomass variables which are to be measured
· the level of detail, and
· the sampling techniques most appropriate for vegetation studies.

Methods for the assessment of vegetation composition

Vegetation composition can be determined by using plot (or quadrat), intercept and plotless methods.

PLOT (QUADRAT) METHODS. These involve establishing sample plots (quadrats) of vegetation whose size and shape will depend on the type of vegetation being sampled, the density of the vegetation encountered and the type of survey being undertaken.

For instance, range scientists often use a large sample of small quadrats (50 x 50 cm), because it is easier to absorb information quickly from small units. However, when the vegetation is sparse (as is the case in the arid/semi-arid rangelands), larger quadrats should be used to ensure adequate sample data (e.g. IPAL, 1983).

The general guidelines on the size and shape of quadrats applicable to three different types of vegetation are shown in Table 2 below.

Table 2. Size and shape of sample plots.

Type of vegetation

Quadrat shape

Quadrat dimensions

Ratio of quadrat sides (rectangular plots)

Herbaceous


Rectangular

0.5-1.0 m2

1:2

Circular

0.2-0.5 m2

n.a.

Shrub

Rectangular

50-100 m2

1:5-1:10

Trees

Rectangular

200 -1000 m2

1:5 -1:10

n.a. = not applicable.

Source: Clarke (1986).

'Nested' quadrats are usually used for mixtures of trees, shrubs and herbaceous vegetation. A nested unit might include a 1000-m2 quadrat for the estimation of tree composition, several 200-m2 shrub quadrats and further, randomly sited, herbaceous units enclosed within the larger area.

Vegetation composition can be estimated on the basis of frequency of occurrence, density, dry weight (biomass) and basal cover.

The frequency sampling method records the species present/absent in each quadrat. Large areas can be surveyed very quickly, especially if attention is focused on a few key indicator species (Tothill, 1978).11 The presence/absence method can be used to estimate the composition of tree, shrub and/or herbaceous species. Because it reflects the distribution of vegetation species within an area, it can also be very useful in the study of vegetation patterns. Pattern recognition is easier if the sample layout is regular rather than random.

11 Using a small number of indicator species within an area will usually give the information required, e.g. the relative importance of palatable versus unpalatable species or of annuals versus perennials. The key species are not necessarily the dominant species of the area (Werger, 1981).

The density method is suitable when individual species are clearly distinguishable, as in woodland, and it is, therefore, often used to estimate woody species composition.

Foliage cover is affected by seasonal effects; for this reason, the measure is not suitable for determining long-term trends in composition. This can be done using basal cover, especially that of perennials, which is not subject to seasonal influences.

Biomass or dry-matter weight measurements (for herbaceous material and foliage from low shrubs) will help establish vegetation relationships between:

· composition and plant biomass production, and
· biomass production and soil type or rainfall.

The conventional method involves cutting, sorting and weighing the different species in each quadrat by hand, which is time-consuming and requires considerable skill, particularly in multi-species grasslands.

Therefore, the dry-weight-rank method (t'Mannetje and Haydock, 1963; Jones and Hargreaves, 1979) is often used. This method involves the selection of the first, second and third heaviest species (on the basis of biomass) within each quadrat,12 each of which is then assigned a weighting teased on standard multipliers which have been shown to be applicable over a range of pasture types in Australia, the United States and Zimbabwe (Jones and Tothill, 1985). The multipliers are:

Rank 1 (heaviest):

0.70

Rank 2 (middle):

0.24

Rank 3 (lightest):

0.06

12 The three most important species usually account for about 90% of the overall biomass of a quadrat and also for about 90% of plant species, if composition is assessed on the basis of frequency or cover.

The values for each quadrat are then summed for each species and expressed as percentages of the total score (see example below). This approximates the percentage contribution by weight of each species, from which the overall composition of the sample area is derived.

Example: Determine the average plant species composition over four sample quadrats with plant species A, B and C, using the dry-weight-rank method.

Table 3. Estimated biomass (g) and rank (in brackets).

A = 60 (1)

A = 65 (1)

A = 10 (2)

A = 5 (2)

B = 5 (3)

B = 10 (3)

B = 40 (1)

B = 20 (1)

C = 15 (2)

C = 15 (2)

C = 6 (3)

C = 3 (3)

By applying the rank weights given above, the following overall percentage biomass is contributed by each species:

Table 4. Percentage of biomass contributed by each species.

Species

Sum of quadrat scores

Total rank score

Per cent composition

A

0.7+0.7+0.24+0.24

= 1.88

47.0

B

0.06+0 06+0.7+0.7

= 1.52

38.0

C:

0.24+0.24+0.06+0.06

= 0.60

15.0


4.00

100.0

Based on the overall percentage biomass of each species given in Table 4, the overall composition of the sample area would be:

Table 5. Overall composition for the sample area.

Species

Sum of quadrat

Total weight (g)

Per cent composition weights (g)

A:

60+65+10+5

= 140

55.1

B:

5+10+40+20

= 75

29.5

C:

15+15+6 +3

= 39

15.4


254

100.0

If there are only two to four species making up the pasture, the method could also be used to estimate proportions or the weight of the component species. Field estimates are recorded directly onto data sheets for computer processing or into small portable computers. The data are then processed, e.g. by using a computer programme called BOTANAL (Jones and Tothill, 1985).

The dry-weight-rank method assumes that there is variability between quadrats and that there are at least three species present in the majority of sample units. If the number of species is less, the problem can be overcome by increasing the size of each quadrat. Alternatively, direct estimates of the proportions of different species can be used.

INTERCEPT METHODS. These involve the sampling of plant species which contact a line transect. Transects are usually placed within areas which have relatively homogeneous vegetation and soil characteristics. Such areas may be identified by ground or aerial reconnaissance surveys or from maps. Three intercept methods are commonly used:

· the line-intercept method
· the step-count method, and
· the wheel-point or frame-point method.

The line-intercept method measures actual vegetation intercepts with a tape measure. These data can then be used to determine foliage or basal cover and frequency of occurrence.

The method tends to be tedious and time-consuming and has limited applicability in situations in which vegetation is highly variable. Cruder systems, such as the presence/absence sampling, often provide information on vegetation composition of comparable or even better accuracy, and take less time to perform.

The step-count method counts the number of different plant species contacted by the point of the foot when walking along a transect. If bare soil is contacted, this is also recorded to give an indication of cover.

The step-count method is a simple and rapid means of making a preliminary assessment of composition (based on frequency) and of basal or foliage cover, and is best suited to areas with low-growing herbaceous vegetation. It can therefore be used for rapid rangeland monitoring where data precision is not essential.

The wheel-point (Tothill, 1978) or the frame-point method (Clarke, 1986) is a more sophisticated modification of the step-count method. In both cases, points mounted on a frame are used to measure vegetation characteristics related to composition or cover. Meeuwig (1981) describes the use of this method for estimating shrub cover and biomass.

To be able to measure changes in the composition or cover by the intercept methods, the same transect lines must be used each time measurements are made. Transects are therefore often permanently marked and may be set in a grid pattern or along a basal marker. Grid methods of laying out sample areas have two advantages:

· the transects are easy to identify, and
· the grids help define vegetation patterns within the sample area.

PLOTLESS METHODS. The best known of these methods are the nearest-neighbour and the point-quarter methods (Clarke, 1986). Both involve the use of transects along which reference points are taken at random (e.g. by pacing according to a random number table) to measure composition in terms of density.

The nearest-neighbour method. With this method, the plant closest to the sample reference point and its neighbour are identified and the distance between each is measured. This procedure is repeated at each of the random reference points, and the overall data are then used to calculate composition on a density basis.

The point-quarter method uses a sampling frame (or compass). The frame, which has two arms outstretched at a right angle to one another (forming a cross), is placed at each selected reference point. The closest plant to the centre of each of the four quarters of the cross is identified and the distance is measured (Figure 4). The area occupied by each plant is calculated by squaring this distance. The overall average along the transect is then calculated to give mean vegetation density (see example below). Composition can be obtained by recording each species type along the transect line.

Figure 4. Schematic representation of the point-quarter method.

Example: Assume a transect of 100 m long on which 20 reference points have been identified. At each point, the distance to four plants is measured and the sum of all these distances is 1000 metres. For 80 plants (4 x 20), the average point to plant distance is 1000/80 or 12.5 m and the average area occupied by each is 125 x 12.5 m or 156.25 m. Since 1 ha is 10 000 m2, the density of plants per hectare is 10 000/156.25 = 64. Of the 80 plants identified, 30 are of species A, 20 of species B. 20 of species C and 10 of species D. Thus, the composition of species along the transect is:

Species

Composition (%)

A

37.5

B

25.0

C

25.0

D

12.5

Plotless methods are relatively quick and low-cost methods suitable for determining the composition of tree and shrub communities. They can be combined with plot methods by placing quadrats at each reference point to make more detailed measurements of plant cover, frequency and/or biomass.

Estimation of plant biomass and standing crop

Vegetation biomass and standing creep are measured to determine the dry weight of plant material available for grazing and/or browse during different seasons of the year and over the long term.

Plant biomass and standing crop can be directly measured by cutting and weighing of all individual samples, but since both procedures are costly and impractical in most situations, indirect or double-sampling methods have been developed where the procedure is:

· Locate sample quadrats or plots (along transects or at randomly determined sites) within the area to be surveyed.

· Select a number of these quadrats and then measure in them dry- matter weight (usually by destructive sampling, i.e. cutting and weighing of sample material) for the whole sample area or for the individual species present. The latter is most meaningful when only a few species are present (species).

· Use the dry-weight data to calibrate visual estimates of dry matter in other sample quadrats. This is done by means of linear regression relationships (Module 11) established from initial dry-weight measurements (e.g. IPAL, 1983, pp. 264-318). Repeated estimates from the same site can then be made within or between years.

The procedure makes it possible to estimate biomass or standing crop over large areas, at relatively low cost. Because the initial measurement of dry weight takes some time, details about other relevant vegetation characteristics (e.g. density, cover, frequency and/or composition) may also be obtained. Such data can be of value when the sample needs to be stratified, e.g. on the basis of productivity differences resulting from factors such as irregular grazing, burning, soil characteristics etc.

The measurement of plant biomass and standing crop for herbaceous plants and shrub and tree species is outlined below.

ESTIMATING THE PLANT BIOMASS AND STANDING CROP OF HERBACEOUS PLANTS. This is done by the scale or comparative-yield method which was developed in Australia (Haydock and Shaw, 1975) and involves the following steps:

· Examine the survey area to determine the likely variation in biomass/standing crop.

· Select visually four plots, two each close to the 'average low' and the 'average high' levels of forage yield and give the pair in each group values of 1 to 4, respectively. One plot of each group is cut and the forage is weighed green, the others are left uncut.

· Select a fifth plot which falls within the middle of these two extremes. The plots estimated to be slightly above and below this middle plot are then cut and their forage is weighed to decide whether the identified plot does indeed fall within the middle of scales 1 and 5. If so, the plot is assigned a scale of 3.

· Repeat the same procedure to determine scales 2 and 4, which lie half-way between scales 1 and 3, and 3 and 5, respectively.

· Use the five ranking standards to rank visually other plots in the sample area. One or several individuals (recorders) may be involved in the ranking.

Note: Periodic checks to ensure consistency are required, particularly if the process is carried out over several days. Photographs can also be used for reference purposes: ideally, each recorder should have a set of photographs to carry out the ranking exercise.

· Assign, at random, about 15 plots to each recorder for ranking on the basis of a scaling factor. The plots are then cut and the dry matter is weighed.

· Determine linear regression equations for each recorder relating estimated weight (or scale) to actual weight, and correct any recorder bias identified.

· Determine overall herbaceous biomass for the sample area. Depending on the accuracy of the initial weight:scale calibration, results under this system have been shown to be remarkably accurate. The initial process is time consuming but once completed, large areas can be surveyed in a relatively short time period.

Estimates of species composition can be made if the comparative-yield method is carried out alongside the dry-weight ranking method. The method can also be used to determine changes in production resulting from seasonal or long-term influences. Under range conditions, the same quadrat sites (or very similar and close to them) should be used during each sampling, but on each sampling occasion, new ranking standards may need to be established.

The scale or comparative-yield method was modified by ILCA field researchers in Kenya. ILCA's method relies on estimation of green biomass, green percentage, percentage cover and leaf canopy height in 0.5 m2 quadrats selected at random in the survey area. After visual estimates have been made, one plot in 10 is cut and the forage weighed. The sample is then dried to obtain the dry-matter percentage (DM %), and regression relationships are established between the estimated and actual green-matter weight. Finally, visual estimates are corrected by the equation derived (Module 11), which makes it possible to make DM extrapolations over large areas.

ESTIMATING THE BIOMASS OF STANDING CROP OF SHRUB AND TREE SPECIES. This is done whenever browse forms a significant component of the animal diet in an area. Methods used to estimate browse productivity are, however, comparatively undeveloped and detailed work has been confined to relatively few sites (Bille, 1980).

One of the problems has been that different standards of measurement have been applied to different situations. In some cases, only the edible or accessible portions of the vegetation have been measured while in others, total biomass or standing crop production was determined. The plant parts which are accessible or usable by livestock vary between seasons, species and environments. In addition, there is considerable variation in production within species, which makes it difficult to obtain reliable data without increasing sampling time and cost.

Highly sophisticated methods have been devised to measure browse production (e.g. by the use of radio-isotopes), but these are not broadly applicable on African rangelands and, therefore, are not discussed in this module. The general procedure in the more commonly used methods is:

· Identify the major browse species within the survey area, using ground and/or aerial surveys (for woody cover) local vegetation maps. This provides the basis for stratifying the area into major vegetation types.

· Within each vegetation type, determine such characteristics as height and density, using the survey methods described above.

· Estimate biomass production by means of direct harvesting of selected trees or shrubs. The extent of destructive sampling will depend on the objectives of the study. For instance, if we are interested in the accessible parts of the plant (e.g. leaves, new twigs, flowers and fruits), samples will be taken from those parts only.13

13 Bille (1980) briefly discussed the relative importance of individual plant components as browse biomass, for different tree species in West Africa.

The amount and type of vegetation actually consumed will normally depend on the browsing habits of different animal species. The IPAL (1983) study in northern Kenya determined biomass production on the basis of the following assumptions made with respect to how high the browsing animal can reach - i.e. up to 1 m for sheep; up to 2 m for goats; and up to 4 m for camels.

In addition, litter samples may be collected in trays placed under the canopies of similarly sized trees or shrubs which are protected from browsing. By taking into account plant litter (which is an important part of the diet of domestic stock, particularly during the drier months of the year), the yield potential of the sampled plant species will be estimated more accurately (IPAL, 1983, p. 287).

· Measure particular tree/shrub attributes (e.g size or height, crown diameter, stem diameter) and relate the measurements of these indicator (or surrogate) variables to actual, recorded biomass using regression analysis (Module 11). The indicator variables may also be used to predict biomass over the entire survey area.14

· Predict the browse-species biomass in the surveyed area on the basis of the density and composition estimates obtained during initial surveys.

14 For instance, crown diameter was found to be a good indicator of wood biomass for the woody species in northern Kenya. As a measure of forage biomass, however, it was generally unreliable (IPAL, 1983).

The use of indicators in predictive equations means that biomass production can be estimated over large areas at relatively low cost. The estimates of utilisable browse can then be related (together with data on herbage production) to the numbers of different animal species in an area to determine the overall grazing pressure. Trends in browse productivity can be monitored by repeating the process over time, taking care to exclude seasonal and cyclical influences.

Ground-survey methods for soil studies

Soil erosion is influenced by soil type, rainfall intensity, topography (slope and length of slope), vegetation cover and composition. Therefore, the methods used to determine vegetation cover and composition can also be used to indicate:

· the likelihood of soil erosion occurring in an area, and

· the scope for preventing soil erosion through changes in livestock management practices.

Assessments of this nature can result in false conclusions about the causes for and/or the extent of soil erosion in an area, but actual measurement of soil erosion in African rangelands is fraught with problems, as well. According to Abel and Stocking (1987), some of these problems are:

· Because of the vast areas involved, accurate assessment of soil loss is both cliff cult and costly to make, requiring between 10 and 20 years to obtain conclusive results.

· Lack of technically feasible and low-cost methods applicable to the rangelands.

· Erosion itself is difficult to measure because rates of soil loss are highly variable over space and time. Periodic observations (e.g. comparison of aerial photography over time) may, for instance, indicate average changes but will not register the events which determine the average.

· Precise measurements of soil loss taken in one area cannot be readily extrapolated. Extrapolations of soil loss from trial plots will, for instance, tend to overestimate actual losses over large areas.

· When soil loss from one area results in soil deposition in another area, the net effects of soil relocation need to be considered. Losses (in terms of productivity) from the eroded area may, for instance, be countered or even outweighed by the gains which occur in the deposition area, but this may prove to be extremely difficult to estimate in practice.

Of the field methods developed for the measurement of soil loss, the erosion pins method is the most common and the most likely to be applicable to African rangelands. It involves placing pins (or pegs) at randomly located sites throughout the sample area, which then represent fixed reference points for the measurement of changes in soil movement over time. A net drop in soil level is taken as an indication of soil loss from which estimates of rates of erosion are made.

The advantage of the method is that it is low cost and that the pins can be sited over large areas in a relatively short period of time. One of its disadvantages in the field is that the pins can be easily lost or destroyed, or they may be difficult to locate because of vegetation growth between survey periods.

On large areas, the movement of soil is complex and apparent losses may be compensated for by deposits elsewhere. Unless such deposits are accounted for, by determining net changes, the method will not measure actual erosion rates. Furthermore, measurements must continue over a sufficient time period before the results can be considered conclusive. This is because short-term changes will often be reversed due to changes in climate, management practices etc.

Instead of using pins, some researchers have used tree-root exposure and pedestal development to measure erosion rates. In some instances, soil loss measured on experimental plots has been used to predict erosion rates over larger areas, but such extrapolations are dubious, particularly when applied to rangelands.

Soil-loss equation models have been developed in the United States, which are based on regression relationships established from trial plot data. Most of these models rely on technical expertise and data precision not found in developing countries. A model has, however, been developed for the southern African rangelands which uses aerial-survey data. It is relatively low cost and adaptable to conditions elsewhere in Africa, provided that an adequate data base exists (Abel and Stocking, 1987).

References

Abel N and Stocking M. 1987. A rapid method for assessing rates of soil erosion from rangeland: An example from Botswana. Journal of Range Management 40(5):460-466.

Abel N O J. Flint M E 5, Hunter N D, Chandler D and Maka G. 1987. Cattle-keeping ecological change and communal management in Ngwaketse. Vol. 1. Summary; Vol. 2. Main report. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia; Integrated Farming Pilot Project, Ministry of Agriculture, Gaborone, Botswana and Overseas Development Group; School of Development Studies, University of East Anglia, Norwich, UK Vol. 1, 45 pp.; Vol. 2, 82 pp. + tables, figures, references.

Behnke R. Jr. 1984. Measuring the benefits of subsistence versus commercial livestock production in Africa. In: Flora CB (ed), Animals in the farming system. Proceedings of Kansas State University's 1983 Farming Systems Research Symposium, Manhattan, Kansas, 31 October - 2 November 1983. Farming Systems Research Paper Series 6. Kansas State University, Manhattan, Kansas, USA. pp. 564-592.

Bille J C. 1980. Measuring the primary palatable production of browse plants. In: Le Houerou H N (ed), Browse in Africa The current state of knowledge. Papers presented at the International Symposium on Browse in Africa, Addis Ababa, 8-12 April 1980. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 185-196.

Breman H. Ciss A M, Djitjeye M A, Elberse W Th. 1979/80. Pasture dynamics and forage availability in the Sahel. Israel Journal of Botany 28:32.

Butterworth M H. 1985. Beef cattle nutrition and tropical pastures. Longman Inc., New York, USA. 500 pp.

Clarke R (ed), 1986. The handbook of ecological monitoring. Clarendon Press, Oxford, UK 298 pp.

Conway G. 1985. Agricultural ecology and farming systems research. In: Remenyi J V (ed), Agricultural systems research for developing countries. Proceedings of an international workshop held at Hawkesbury Agricultural College, Richmond, New South Wales, Australia, 12-15 May 1985. ACIAR Proceedings 11. ACIAR (Australian Centre for International Agricultural Research), Canberra, Australia. pp. 43-59.

Cooperrider A Y and Bailey J A. 1981. Computer simulation of forage allocation between big game and domestic livestock. In: Arid land resource inventories: Developing cost-efficient methods. Proceedings of an international workshop held in La Paz, Mexico, 30 November- 6 December 1980. General Technical Report WO-28. USDA (United States Department of Agriculture), Forest Service, Washington DC, USA. pp. 547-554.

Dicko-Touré M. 1980. Measuring the secondary production of pasture: An applied example in the study of an extensive production system in Mali. In: Le Houerou H N (ed), Browse in Africa: The current state of knowledge. Papers presented at the International Symposium on Browse in Africa, Addis Ababa, Ethiopia, 8-12 April 1980. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 247-254.

Haydock K P and Shaw N H. 1975. The comparative yield method for estimating dry matter yield of pasture. Australian Journal of Experimental Agriculture and Animal Husbandry 15(76):663-670.

Heady H F. 1981. Arid land characteristics, resources and uses - an overview. Section 1. In: Arid land resource inventories: Developing cost-efficient methods. Proceedings of an international workshop held in La Paz, Mexico, 30 November - 6 December 1980. General Technical Report WO-28. USDA (United States Department of Agriculture), Forest Service, Washington DC, USA. pp. 5-7.

Heady H F and Heady E B. 1982. Range and wildlife management in the tropics. Longman Group Inc. New York, USA. 140 pp.

Horowitz M M. 1979. The sociology of pastoralism and African livestock projects. AID Program Evaluation Discussion Paper 6. USAID (United States Agency for International Development), Washington DC, USA. 102 pp.

Husch B. 1981. How to determine what you can afford to spend for inventories. In Arid land resource inventories: Developing cost-efficient methods. Proceedings of an international workshop held in La Paz, Mexico, 30 November - 6 December 1980. General Technical Report WO-28. USDA (United States Department of Agriculture), Forest Service, Washington DC, USA. pp. 98-102.

ILCA (International Livestock Centre for Africa). 1983. Pastoral systems research in sub-Saharan Africa Proceedings of the IDRC/ILCA workshop held at ILCA, Addis Ababa, Ethiopia, 21-24 March 1983. ILCA, Addis Ababa, Ethiopia. 480 pp.

IPAL (Integrated Project in Arid Lands). 1983. Proceedings of the IPAL Scientific Seminar, Nairobi, Kenya, 1-2 December 1981. IPAL, Nairobi, Kenya. 355 pp.

Jones R M and Hargreaves J N G. 1979. Improvements to the dry-weight rank method for measuring botanical composition. Grass and Forage Science 34:181-189.

Jones R M and Tothill JC. 1985. Botanal - a held and computing package for assessment of plant biomass and botanical composition. In: Tothill J C and Mott J (eds). Ecology and management of the world's savannahs. Australian Academy of Science, Canberra, Australia. pp. 318-320.

Lamprey R H and de Leeuw P N. 1986. The use of NOAA AVHRR imagery for rangeland monitoring. In: Hansen R M, Woie B kM end Child R D (eds). Range development and research in Kenya Proceedings of a conference held at the Agricultural Resources Centre, Egerton College, Njoro, Kenya, 1-5 April 1986. Winrock International Institute for Agricultural Development, Morrilton, Arkansas, USA. pp. 441-453.

de Leeuw P N and Milligan K 1984. A review of integrated surveys for resource inventory end monitoring of pastoral production systems in sub-Saharan Africa ILCA Bulletin 20: 18-22. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia.

Meeuwig R O. 1981. Point sampling for shrub biomass. In Arid land resource inventories: Developing cost-efficient methods. Proceedings of an international workshop held in La Paz, Mexico 30 November - 6 December 1980. General Technical Report WO-28. USDA (United States Department of Agriculture), Forest Service, Washington DC, USA. pp. 323-326.

Mouat D A and Johnson R R. 1981. Vegetation inventories and interpretation of environmental variables for resource management: Grand Canyon National Park, a case study. In: Arid land resource inventories: Developing cost-efficient methods. Proceedings of an international workshop held in La Paz, Mexico, 30 November - 6 December 1980. General Technical Report WO-28. USDA (United States Department of Agriculture), Forest Service, Washington DC, USA. pp. 300-307.

Ng'Ethe J C. 1986. Implications of range condition and trend in resource management. In: Hansen R M, Woie B M and Child R D (eds). Range development and research in Kenya. Proceedings of a conference held at the Agricultural Resources Centre, Egerton College, Njoro, Kenya, 1-5 April 1986. Winrock International Institute for Agricultural Development, Morrilton, Arkansas, USA. pp. 461-466.

Pratt D J and Gwynne M D (eds). 1977. Range management and ecology in East Africa. Hodder and Stoughton, London, UK 310 pp.

de Ridder N and Wagenaar K T. 1986. Energy and protein balances in traditional livestock systems and ranching in eastern Botswana Agricultural Systems 20(1): 1-16.

Robinove C J. 1981. Efficient arid land monitoring using Landsat images. In: Arid land resource inventories: Developing cost-efficient methods. Proceedings of an international workshop held in La Paz, Mexico, 30 November - 6 December 1980. General Technical Report WO-28. USDA (United States Department of Agriculture), Forest Service, Washington DC, USA. pp. 256-259.

Sandford S. 1982. Pastoral strategies and desertification: Opportunism and conservatism in dry lands. In: Spooner B and Mann H S (eds). Desertification and development: Dryland ecology in social perspective. Academic Press, London, UK pp. 61-80.

Sandford S. 1983a. Management of pastoral development in the Third World. Overseas Development Institute, London, UK and John Wiley and Sons, Chichester, UK 316 pp.

Sandford S. 1983b. Organisation and management of water supplies in tropical Africa. ILCA Research Report 8. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. 45 pp.

Schmid-Haas P. 1981. Monitoring change with combined sampling on aerial photographs and on the ground: In: Arid land resource inventories: Developing cost-efficient methods. Proceedings of an international workshop held in La Paz, Mexico, 30 November - 6 December 1980. General Technical Report WO-28. USDA (United States Department of Agriculture), Forest Service, Washington DC, USA. pp. 383-388.

Stoddart L A, Smith A D and Box T W. 1975. Range management. Third edition. McGraw-Hill Book Co., New York, USA. 532 pp.

Thalen D C P. 1981. An approach to evaluating techniques for vegetation surveys on rangelands. In: Low-level aerial survey techniques. ILCA Monograph 4. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 85-97.

t'Mannetje L (ed), 1978. Measurement of grassland vegetation and animal production. Bulletin 52, Commonwealth Bureau of Pastures and Field Crops, Hurley, Berkshire, England. Commonwealth Agricultural Bureaux, Farnham Royal, Buckinghamshire, England. 260 pp.

t'Mannetje L and Haydock K 1963. The dry-weight rank method for botanical analysis of pasture. Journal of British Grasslands Society 18.

Tothill J C. 1978. Measuring botanical composition of grasslands. In: t'Mannetje L (ed) Measurement of grassland vegetation and animal production. Bulletin 52, Commonwealth Bureau of Pastures and Field Crops, Hurley, Berkshire, England. Commonwealth Agricultural Bureaux, Farnheim Royal, Bucks, England. pp. 22-62.

USDA (United States Department of Agriculture) Forest Service. 1981. Arid land resource inventories: Developing cost-efficient methods. Proceedings of an international workshop held in La Paz, Mexico, 30 November - 6 December 1980. General Technical Report WO-28. USDA Forest Service, Washington DC, USA. 620 pp.

Werger M J A. 1981. Phytosociological inventory as a basis for resource evaluation in arid lands. In: Arid land resource inventories: Developing cost-efficient methods. Proceedings of an international workshop held in La Paz, Mexico, 30 November - 6 December 1980. General Technical Report WO-28. USDA (United States Department of Agriculture), Forest Service, Washington DC, USA. pp. 338-341.


Previous Page Top of Page Next Page