On estime généralement qu'il existe un taux de charge optimal des
parcours, que ce taux permet de préserver les parcours tout en portant les profits au
maximum, que l'évaluation de ce taux est une simple affaire de technique et que le taux
optimal s'applique à tous les types de production animale, depuis les systèmes purement
commerciaux jusqu'aux systèmes exclusivement axés sur la subsistance. En dépit de ces
hypothèses, les responsables de l'élaboration des politiques, les organismes de
développement et les spécialistes de l'élevage et des parcours ont été en désaccord
constant avec les pasteurs africains au sujet de la valeur de ce taux optimal. On attribue
d'ordinaire cette divergence d'opinions à l'ignorance des populations pastorales et l'on
recommande de recourir à des campagnes de vulgarisation ou à une réglementation
administrative. Selon les experts, les parcours africains sont surexploités, et il
importe que les habitants des zones sèches de l'Afrique apprennent rapidement - ou soient
forcés d'adopter - de nouvelles pratiques.
Cette série de trois articles traitent d'une autre explication possible. Ils font valoir
que certains observateurs techniques se sont fait une idée confuse de la nature exacte du
surpâturage, que les taux de charge actuels sont souvent appropriés aux objectifs de
gestion pastorale et que, dans de nombreux cas, ils s'avèrent en outre écologiquement
rationnels.
Fondé sur des données expérimentales tirées d'essais concernant les taux de charge
propres aux bovins de boucherie, le premier article de la série fait état des confusions
que peut susciter la notion de surpâturage. L'auteur juge peu vraisemblable l'hypothèse
d'un taux de charge optimal unique et estime qu'aux diverses pratiques et options
d'élevage doivent correspondre des objectifs de chargement différents.
Le deuxième article examine les objectifs de gestion types des pasteurs en Afrique
semi-aride ainsi que les ressources dont ils disposent. Les éleveurs traditionnels optent
souvent pour des produits différents de ceux que choisissent les grands éleveurs
commerciaux. Dans les systèmes d'élevage africains, la relation qui unit facteurs de
production et produits sur le plan technique exige généralement que l'on emploie des
taux de charge élevés pour maximiser les rendements, qui s'avèrent d'ordinaire égaux
ou supérieurs aux rendements par hectare obtenus par les grands éleveurs commerciaux
dans un milieu naturel comparable.
Le surpâturage pratiqué dans le cadre des systèmes d'élevage traditionnels peut se
révéler rentable, mais néanmoins écologiquement non viable à long terme. Le
troisième article aborde la question de la dégradation des parcours. Il présente une
technique d'évaluation des coûts environnementaux correspondant à différents taux de
charge, fondée sur l'appréciation de l'ampleur des pertes en sol ou des modifications
économiquement néfastes de la végétation des parcours.
Existe una densidad de pastoreo óptima para los pastizales; dicha
densidad permite conservarlos obteniendo beneficios máximos, su estimación es una
cuestión técnica y la densidad óptima es aplicable a todos los tipos de producción
pecuaria, desde los sistemas totalmente comerciales hasta los que son completamente de
subsistencia. A pesar de estas hipótesis, las autoridades, los organismos de desarrollo y
los científicos especializados en ganadería y en pastizales han demostrado su desacuerdo
con los pastores africanos sobre lo que es la densidad óptima. Esta diferencia de
opinión se suele atribuir a la ignorancia de los pastores y se recomiendan campañas de
extensión o nueva reglamentación administrativa. Los pastizales de Africa, según los
expertos, están sometidos a sobrepastoreo y los residentes en las zonas secas del
continente deberían aprender con rapidez algunos nuevos hábitos, o bien habría que
forzarlos a ello.
En esta serie de tres artículos se trata de encontrar una explicación alternativa. Se
argumenta que algunos de los observadores técnicos han partido de una idea confusa de lo
que es el sobrepastoreo, que las densidades actuales de pastoreo son a menudo apropiadas
para los objetivos de ordenación de los pastos y que esas densidades pueden ser
sostenibles desde el punto de vista ecológico.
Utilizando datos experimentales obtenidos en ensayos sobre la densidad de pastoreo de
vacunos de carne, en el artículo inicial de la serie se señalan las posibilidades de
confusión que entraña el concepto de sobrepastoreo. Se argumenta que no es probable que
haya una sola densidad de pastoreo óptima, siendo apropiadas distintas metas para
prácticas y objetivos de explotación diferentes.
En el segundo artículo se examinan los objetivos característicos de la ordenación y la
dotación de recursos de los pastores de las zonas semiáridas de Africa. Los pastores
tradicionales eligen con frecuencia unos resultados de la producción distintos de los que
tienen los productores comerciales. Las relaciones técnicas entre los insumos y los
productos suelen requerir una densidad de pastoreo mayor para obtener un rendimiento
máximo en los sistemas de explotación africanos, que normalmente igualan o superan la
producción por hectárea de los productores comerciales en condiciones naturales
comparables. La densidad elevada de pastoreo puede ser rentable para los pastores
tradicionales, pero resultar insostenible desde el punto de vista ecológico a largo
plazo.
En el tercer artículo de la serie se aborda la cuestión de la degradación de los
pastizales. Se presenta una técnica para la estimación de los costos ecológicos,
medidos en función de la tasa de pérdida de suelo o de cambios perjudiciales desde el
punto de vista económico en la vegetación de los pastizales, que pueden derivarse de
distintas densidades de pastoreo.
It is widely believed that there is an optimum stocking rate for
rangeland, that this rate will conserve range while maximizing profit, that estimating
this rate is a technical matter and that the optimum applies to all types of livestock
production, from fully commercial to totally subsistence systems. Despite these
assumptions, policy-makers, development agencies and livestock and range scientists have
persistently disagreed with African pastoralists over what the optimum rate is. This
difference of opinion is commonly attributed to pastoral ignorance, and extension
campaigns or administrative regulation are recommended. Africa's rangelands, the experts
tell us, are overstocked, and the residents of dry Africa should quickly learn, or be
forced to adopt, some new habits.
This series of three articles explores an alternative explanation. It argues that some
technical observers have had a muddled idea of what constitutes overstocking, that
existing stocking rates are often appropriate to pastoral management objectives and that
these rates may be ecologically sustainable.
Using experimental data from stocking rate trials for beef cattle, the first article in
the series exposes the potential for confusion embedded in the concept of overgrazing. The
article argues that there is unlikely to be a single optimum stocking density, as
different stocking targets are appropriate to different husbandry practices and
objectives.
The second article examines the characteristic management objectives and resource
endowments of herders in semi-arid Africa. Traditional herders often choose to produce
different outputs from those produced by commercial ranchers. Technical relationships
between inputs and outputs commonly require higher stocking rates to maximize yields in
African husbandry systems, which usually equal or exceed per hectare output from
commercial ranchers in comparable natural environments.
It may be profitable but environmentally unsustainable in the long term for traditional
herders to stock heavily. The third article in the series addresses the issue of rangeland
degradation. It presents a technique for estimating environmental costs, measured in terms
of rates of soil loss or economically disadvantageous changes in rangeland vegetation,
which result from different stocking densities.
A herd of Brahmin cows with their calves grazing in an improved
pasture at an experimental station in Panama
Un troupeau de vaches Brahmin et leurs veaux broutant dans un pâturage amélioré d'une
station expérimentale au Panama
Hato de vacas Brahmin y sus terneros pastando en pastizales mejorados en una estación
experimental en Panamá
Photo/Foto: United Nations
Well-fed Baoulé cattle on lush pastures in Burkina Faso
Bétail Baoulé bien nourri des riches pâturages du Burkina Faso
Ganado Baoulé bien alimentado en los frondosos pastizales de Burkina Faso
A series of three articles by:
R. Behnke may be contacted c/o FAO Representation, PO Box 24185,
Windhoek, Namibia.
N. Abel is with the National Rangelands Program, Division of Wildlife and Ecology, CSIRO,
PO Box 84, Lyneham, ACT 2602, Australia.
"Overpopulation may be defined rigorously as too many animals,
but the rigor ends there."
Graeme Caughley
Intensification involves additional investment to achieve additional
output. Overstocking involves excessive investment in livestock and a loss of rangeland
output. Intensification is good, overstocking is bad, and the difference should be plain.
But it is not.
Overstocking is not a single concept assessed by like-minded observers according to a
standard set of generally accepted criteria; rather it is a group of ideas clustered
loosely around the notions of too many animals and too little grass. Because different
versions of the overstocking concept possess a common name and exhibit a certain family
resemblance, they are often confused with each other and with intensification.
The clarification of this conceptual muddle is the focus of this paper, which reviews the
experimental literature on stocking rate trials for beef cattle. This paper models the
effect on output of "intensifying" a beef production system by investing in more
and more cattle while holding con-stant the supply of grazing land. The results of this
process of intensification are expressed as a production function for meat output at
various stocking densities. This production function identifies at least six optimal
stocking densities beyond which a rangeland might reasonably be judged to contain
"too many animals", i.e. six operationally distinct definitions of overstocking
which are described and discussed below.
The relationship between product output and cattle densities on a beef
ranch is expressed diagrammatically in Figure 1, which summarizes the results of numerous
grazing intensity experiments on a wide variety of pasture types (Conway, 1974; Jones and
Sandland, 1974; Malechek, 1984; Butterworth, 1985). In Figure 1, the vertical axis
measures output in terms of weight gain (kg of beef produced), either per animal or per
unit land area; the horizontal axis marks the stocking density measured in animals, or
livestock units (LU), per hectare.
With respect to weight gain by commercial beef breeds, individual animal performance
(dashed line) can be represented by a pair of straight lines. At very low stocking
densities (from 0 to MN [maximum nutrition]), weight gain per animal remains constant
because forage is so abundant that it constitutes no constraint and diminished amounts of
forage have no impact on animal performance. When forage does become a limiting factor at
densities above MN, weight gain per animal decreases as an inverse linear function of
stocking density.
Beef production in terms of weight gain per hectare (solid line) is a somewhat more
complicated affair. The shape of this curve is a function of the per caput output of
individual animals at different stocking densities, multiplied by the total number of
animals at those densities. The result is a parabola which intersects the horizontal axis
at two points. At densities between MN and K (ecological carrying capacity) in Figure 1,
the relationship between productivity per animal and stocking rate is expressed as:
gain per animal = a - bS (1)
where S is the stocking rate in animals per unit land area and a and b are constants for particular pastures or types of livestock. Productivity per unit area is therefore:
gain per unit area = aS - bS2 (2)
with a, b and S as in Equation 1.
These intersections identify the only stocking densities at which the
grazing system is potentially at natural or "unmanaged" equilibrium. At zero,
there are no animals and so the system is stable although unproductive. The other
intersection, to the far right along the horizontal axis at K, also marks an unproductive
but potentially stable state. At this density, sometimes termed ecological carrying
capacity, the animal population ceases to expand because it has grown so large that, on
average, it receives only a maintenance diet and animals die at the same rate they are
born and gain weight at the same rate they lose it.
At K, an animal population produces no physical output in terms of average weight gain per
animal or per unit land area. Most owners of domesticated livestock therefore find it
profitable to contain herd growth at a population size short of this ecological ceiling.
To arrest herd growth artificially and hold herd size short of K requires the constant
culling of animals at a rate which will offset the natural capacity of the herd to grow.
However, culling - the harvesting of a steady crop of beef from the system - is precisely
what the rancher wants to do.
Halfway between 0 and K - at the peak of the parabola - is the stocking density MY
(maximum [biological] yield). It is at this density that the tradeoff between individual
animal performance and total animal numbers is most advantageously poised to give the
highest potential rate of mass gain. Between 0 density and MY, adding more animals to the
grazing system increases total output, but at a diminishing rate as densities approach MY.
At densities greater than MY, the reverse process takes over: competition for feed is so
intense that the addition of more animals progressively undermines both individual and
total herd output, until output falls to 0 at K. Ranchers who want to maximize beef
offtake will, therefore, seek to operate at stocking densities in the vicinity of MY.
Figure 1 displays the biological output or physical yield of a grazing system. However,
commercial ranchers are intent on maximizing economic returns rather than biological
outputs. A technique for assessing the profitability of alternative stocking rates is
illustrated in Figure 2, which converts the physical outputs shown in Figure 1 into cash
equivalencies and then compares these returns with operating costs at different stocking
densities.
MN = Maximum nutrition
MY = Maximum yield
K = Ecological carrying capacity
1
Stocking rate and beef production
Taux de charge et production de viande de bœuf
Densidad de pastoreo y producción de carne de vacuno
MN = Maximum nutrition
MP = Maximum profit
MY = Maximum yield
MO = Open access equilibrium
K = Ecological carrying capacity
2
Economically and biologically optimal stocking densities
Taux de charge optimaux, tant du point de vue économique que biologique
Densidades óptimas de pastoreo desde los puntos de vista económico y biológico
Nomads and their cattle in Ethiopia
Nomades et leur bétail en Ethiopie
Población nómada y su ganado en Etiopía
In Figure 2 the value of output and costs of production (vertical axis)
are displayed relative to alternative stocking densities (arranged along the horizontal
axis as in Figure 1). Output is expressed per unit area and price per unit of output is
arbitrarily set at one ($1 per pound or £1 per kg, etc.) so that both physical yield and
total revenue can be represented conveniently by the same curve (Jarvis, 1991).
Costs in Figure 2 refer only to operating expenses which increase in proportion to herd
size. For simplicity, these "variable" costs - e.g. for veterinary supplies or
hired labour - are assumed here to be constant per beast and, therefore, to increase
linearly with the addition of each animal.
The value of the rancher's own management input, family labour and land are treated as
"fixed" costs, since they do not increase with increases in animal numbers nor,
at least in the short term, can they be avoided by stocking fewer animals. These fixed
expenses are not treated as costs but do receive "rent", defined as the
difference between the total variable costs and gross returns to the enterprise. For the
commercial rancher on private land, the economically optimal stocking density, MP (maximum
profit), is that density which maximizes rent, the differential between total revenue and
total vari-able costs. This point is reached at the level of production in which the last
additional unit of output adds the same amount of revenue as costs.
While more elaborate and precise techniques can be employed, the economically optimal
stocking density in Figure 2 can be roughly identified by visual inspection; it occurs at
the point of greatest vertical distance between the revenue and variable cost curves.
Profit per unit area is:
PA = P[aS - bS2] - cS - FC (3)
where Pa = profit per unit area; P = price per unit weight of beef; c = variable cost per animal; FC = fixed costs per unit area; and a, b and S are as defined for equations 1 and 2. For further discussion, see Booysen, Tainton and Foran (1975), Carew (1976), Hildreth and Riewe (1968), Workman (1986) and Wilson and MacLeod (1991).
Let us now review Figures 1 and 2, looking for the various stocking densities beyond which an observer - or producer - might be inclined to conclude that the system contains too many animals.
Scenario 1. The lowest of these values is density MN - the
density at which feed availability first becomes a constraint. Beyond this
"critical" stocking rate (Hart, 1980; Malechek, 1984) increases in density
entail a progressive decline in livestock nutritional levels, per caput animal
productivity and overall herd condition.
Because of these detrimental effects, MN has been widely employed as a baseline for
determining appropriate intensities of rangeland use. Routinely, empirical evidence to
establish the baseline in different grazing environments is provided by experimental
results from agricultural research station trials conducted at or close to nutritionally
optimal stocking densities. This research commonly documents a vast productivity gap
between animals on research stations versus those in adjacent or similar pastoral areas.
It is then concluded that the pastoral areas are overstocked, unproductive and poorly
managed (Behnke, 1985).
This type of "yield gap" analysis is premised, however, on a sleight of hand:
production must be expressed per head rather than per unit land area. As Figure 1
illustrates, stocking densities that sustain cattle at peak condition are unlikely to
match the aggregate output of more heavily stocked areas, despite record levels of
individual animal performance. In fact, few commercial ranchers could sustain the economic
losses incurred by employing such a low stocking rate. Ranchers producing very expensive
animals for the show ring or for their pedigree may be the only examples of commercial
enterprises that can afford to maintain stocking densities that maximize individual animal
performance. Since it would be unreasonable to transform Africa's open rangelands into a
pan-continental stud farm, this initial definition of optimal stocking density is
irrelevant, although it has contributed significantly to a vague and ill-defined notion of
"overstocked" African rangelands.
Scenario 2. A second possible target stocking density
occurs in the vicinity of point MP in Figure 2. MP is the most advantageous stocking
density for commercial ranchers who are trying to maximize their profits. The
self-interest of rangeland users will encourage the adoption of this stocking target
whenever rangelands are monopolized by one firm or producer who is in a position to
capture all the resource rents and profits generated by a restrained stocking policy.
The precise location of this commercial optimum is determined by a combination of
biological and economic factors, and may be effected by changing cost levels or output
prices, as will be discussed in the next paper in this series. Nonetheless, MP invariably
lies to the right of (at a higher density than) MN. So long as there are significant
variable costs in ranching operations, with few exceptions MP will also be positioned to
the left of (at a lower density than) MY, the next stocking threshold (Workman, 1986).
Scenario 3. Whereas density MN marked the point of maximum
per caput animal output, MY marks the density at which a herd owner can obtain maximum
aggregate output per unit area.
For the rancher pursuing commercial objectives on freehold land, MY marks no management
goal. On the other hand, the maintenance of densities near MY may, under certain
circumstances, be consistent with the objectives of subsistence-oriented African
pastoralists (Behnke, 1994). MY marks the stocking density that will maximize the combined
output of all herds using an area and, thereby, provision the largest human population
directly dependent on the livestock of that area. MY would therefore conform to the
political and strategic requirements of pastoral communities which were compelled to
defend their resource base by maintaining on it the largest sustainable human population.
Stocking densities that maximize aggregate output are, of course, significantly higher
than those that are appropriate for either specialized breeders or beef ranchers operating
in a commercial context and possessing secure title to their land.
Scenario 4. The next critical density, at MO (open access
equilibrium), maximizes the number of independent herding operations using an area.
Stocking densities in the vicinity of MO are often the unintended result of a situation in
which rangeland is unowned and herders are free to enter and use a pasture at their own
discretion. In this situation there is an incentive for new owners to add their private
animals to those already using an area, in an effort to capture for themselves part of the
unallocated economic rent available. The entrance of new herds and herd operators is
likely to continue until aggregate stocking densities approach MO in Figure 2, the point
at which total variable costs equal total revenue, removing any further incentive for new
operators to enter the area.
At MO, all potential resource rents will have been dissipated by excessive numbers of
livestock owners and livestock using the "open access" resource, and herd
operators will receive only an income sufficient to cover the costs of operation and
provide a minimum "opportunity cost" wage comparable to what they could expect
to earn if they abandoned pastoralism for some other occupation. MO therefore represents
the outer margin of viable economic operation on the rangeland in question. Stocking
densities beyond MO may be biologically possible but they are not economically sustainable
since, beyond MO, the costs of herd operation would exceed returns, rendering insolvent
anyone who persistently operated at these densities. For further discussion, see Gordon
(1954) on the relationship between fishing intensity and fish stocks, reapplied to
pastoral conditions in Jarvis (1991).
Open access/minimum wage equilibrium is not a desirable stocking target for any group of
producers except the very poor. For the poor, the use of unclaimed natural resources can
provide an escape from unacceptable working conditions elsewhere in the economy. Should a
sufficient number of poor people avail themselves of this option, free access to
underexploited natural resources could open up an economic frontier, create labour
scarcities in the wider economy and, for a time, increase minimum wages and standards of
living. While generally deplored by environmentalists, wage frontiers of this kind may
help to maintain economic equity and encourage the growth of democratic political
institutions.
Scenario 5. The fifth stocking ceiling, K, marks the limits of what is biologically feasible over the long term in a particular grazing system. K is what wildlife biologists are referring to when they talk about "ecological carrying capacity" - the level at which a herbivore population would naturally tend to stabilize, in the absence of predators and assuming a relatively constant forage supply from year to year. Of purely theoretical interest for the owners of domesticated stock, for some wildlife managers K may represent a positive stocking goal - a herbivore population undisturbed by human predation.
Scenario 6. The highest conceivable levels of overstocking
lie beyond K; hence they are not depicted in Figures 1 and 2. These levels of overstocking
- at what might be termed K+ - may be caused by an overabundance or sudden dearth of
vegetation and are, by definition, unsustainable.
Overabundant feed supplies can result in the sudden expansion of animal numbers when, for
example, new herbivore species are introduced into favourable habitats, temporarily
releasing normal controls on population growth. This is the typical herbivore eruption.
Animal populations overshoot available feed supplies because the herbivores consume the
forage "output" produced by plant growth and then proceed to eat the vegetative
"capital" represented by the plants themselves, thereby undermining the basis
for maintaining future plant growth. When the lagged effects of this "asset
stripping" are felt, the herbivore population may crash (Caughley, 1981).
The eruption of domestic livestock populations is illustrated by the introduction of
cattle, and the expansion and subsequent collapse of their numbers on the high plains of
the western United States in the late nineteenth century or, more recently, on the Mambila
Plateau of Nigeria. A population overshoot similar to that produced by an eruption can
also be caused by a precipitous drop in primary production for whatever reason. In Africa,
drought is the usual cause of these collapses in forage availability and associated
crashes in livestock numbers.
The biological asset stripping which underpins the herbivore eruption has its commercial
parallels. Assuming that a rangeland cannot maintain ranching incomes at levels comparable
to opportunities elsewhere in the economy, the accelerated "decapitalization" of
vegetative stocks at K+ animal densities is, at least in theory, a feasible commercial
proposition.
There are at least six optimal stocking densities which can be defined in terms of livestock production criteria and beyond which a grazing system might be said to be "overstocked". Confusion arises because different densities may be appropriate to different management and production systems or advocated by different sets of professional observers. The critical densities are as follows:
Maximum nutrition (MN), the highest stocking density consistent with maintaining optimum standards of animal nutrition and individual animal performance.
Maximum profit (MP), the density which optimizes operator profits
or "economic rent" per unit land
area, assuming rangeland is held in secure individual tenure.
Maximum (biological) yield (MY), the most advantageous stocking rate for pastoral communities that require the maintenance of high human population densities in order to defend land rights which are not legally secured.
The highest economically sustainable stocking rate, and the rate which, under conditions of open access equilibrium (MO), maintains the "maximum (number of independent herding) operations".
Ecological carrying capacity (K), the highest livestock populations that are biologically sustainable in a given setting.
A biological mining operation in which an unsustainably large livestock population temporarily maintains itself (K+) before slipping or crashing back to a more modest size.
There is little point in simply characterizing an area as "overstocked". Rangelands are over or understocked with reference to different - and potentially conflicting - sets of management objectives associated with alternative production systems and assessment criteria. The preceding discussion has identified six different sets of such criteria - animal nutrition, profits, yield, the number of herding operations and, finally, the total number of livestock which could be supported on a permanent or temporary basis - all of which might be maximized under different management regimes. The practical lesson to be drawn from this analysis is that:
"[Overpopulation] is not a single neat phenomenon but a set of them. Before management activities can be planned to cope appropriately with a case of overpopulation, we must know not so much why the area is overpopulated but rather in what sense it is overpopulated" (Caughley, 1981).
It follows that destocking can, on occasion, be both inappropriate and
expensive. At densities short of MY and MP, wherever these points may lie in a particular
grazing system, destocking entails reductions in rangeland output or pastoral revenue. In
sum, one range manager's "overstocking" may be another's
"intensification", and the distinction is only partly a technical one. Much also
hinges on whose management objectives one subscribes to.
We must first determine the kind of (alleged) overstocking we are dealing with before
weighing the evidence for its existence and the costs and feasibility of its control. As
the next paper in this series shows, this is a complex undertaking in semi-arid Africa,
where we are comparing systems as different as commercial ranching and
subsistence-oriented pastoralism.
The promotion of commercial livestock husbandry has long been seen as a
means of destocking African rangelands and increasing livestock output through increased
offtake. This paper argues that commercialization does exert a long-term downward pressure
on African stocking densities, which will make many policy-makers, administrators and
range scientists happy. However, the shift from subsistence to market-oriented forms of
range livestock husbandry also exerts downward pressure on total rangeland output and
undermines the capacity of rangelands to support human populations, a possibility that is
not likely to be warmly welcomed by displaced pastoralists. For husbandry systems
dependent on natural forage, commercialization is not a process of intensification, but
rather of factor substitution. In this process, capital investments and commercial inputs
displace labour and encourage both lower stocking rates and the production of a diminished
array and volume of output. For range husbandry systems, intensification, destocking and
commercialization may be contradictory rather than mutually compatible objectives.
Distinctive livestock breeds, species and output mixes, variable levels of market
involvement and different systems of land tenure ensure that the boundaries of what
constitutes "overstocking" are likely to be very different for commercial
ranching and African pastoralism. Nonetheless, despite policy concerns about pastoral
overstocking, there is little experimental data on mixed-product output levels from
different combinations of indigenous African breeds and species at alternative stocking
densities. It is, therefore, not possible to construct for multispecies, multiproduct
husbandry systems empirical production functions comparable to those in the first paper in
this series on beef ranching.
There exists, nonetheless, a body of experimental research on output from indigenous
African stock subjected to various levels of nutritional stress. It suggests that African
livestock respond very differently from European breeds of beef cattle to the nutritional
deprivation associated with increasing stocking density. Two sets of factors are
responsible for these distinctive responses: the physiology of indigenous African
livestock and the broad mix of products derived from pastoral and agropastoral herds. The
combined effect of these factors is, in general, to position comparable overstocking
thresholds at higher stocking rates in pastoral than in ranching systems.
Cattle kept by commercial ranchers have been selectively bred to respond to improved forage availability at low stocking rates on ranches. But in achieving positive responses to favourable conditions breeders have had to accept the reverse process as well - declines in productivity resulting from input withdrawal. In comparison, African cattle breeds are less sensitive to high stocking densities and low feed availability, and can survive, produce and reproduce under conditions that are inadequate by the standards of commercial breeds in temperate climates (Coppock, Swift and Ellis, 1986). The physiological mechanisms that sustain this resilience include:
Indigenous African cattle are smaller and lighter than improved breeds,
and can match neither the absolute level of output per animal nor the efficiency of the
rate of feed conversion into livestock product achieved by improved breeds (Richardson,
1994). What indigenous breeds do produce is the highest output per hectare or per kilogram
of metabolic or body mass. Because indigenous breeds have low dietary maintenance
requirements, output per hectare is maximized at higher stocking densities than with the
larger, improved breeds (Richardson, 1994; Tawonezvi et al., 1988). Indigenous
breeds are also better able, through the mechanisms described above, to survive drought.
While improved animals might be more productive in the favourable forage conditions
prevailing after the rains return, few of these animals would have been able to survive
the drought.
Although cattle have received the bulk of research attention, it would appear that African
small ruminants, especially camels, respond to nutritional deprivation in a manner
analogous to African bovines rather than commercial beef breeds.
The relative advantages of different stocking densities are also
influenced by the kinds of products managers are seeking. There are several reasons for
supposing that the density-dependent production functions for dairy produce, animal fibre,
fertilizer products and draught power - all important pastoral and agropastoral products -
are significantly different from the output curves for beef.
Western has calculated that pastoralists can obtain over 2.5 times more energy from
combined meat and milk offtake than from meat offtake alone (Western, 1982; Western and
Finch, 1986). This is because of the greater efficiency of conversion of both feed energy
(Blaxter, 1962; King, 1983) and nutrients - principally nitrogen (Spedding, 1971) - from
pasture into milk.
Milk, fertilizer, power and fibre also differ from meat, hides and carcass derivatives in
that the former are live animal products while the latter require animal slaughter.
Optimum animal densities for carcass production are those that generate the greatest
surplus for culling. As discussed in the first paper in this series, these densities occur
at the "explosive" stage of herd growth, at about K+ for beef cattle. On the
other hand, optimal densities for live animal production are those that sustain an
appropriate "standing crop" of animals rather than a rapid turnover in animal
numbers.
If production can continue during periods of weight loss, stocking rates that maximize
live animal outputs such as fibre, manure or milk will tend to be higher than those that
maximize meat output. This is illustrated in Figures 1a and 1b, which compare the wool and
meat outputs of several varieties of sheep on different kinds of pasture. Meat output
peaks and begins to decline at stocking densities well below those that would maximize
wool output (Donnelly, Morley and McKinney, 1983; Donnelly, McKinney and Morley, 1985).
1a
Changes in wool and meat production per ewe with stocking rate
Variations de la production de laine et de viande par brebis selon le taux de charge
Cambios en la producción de lana y de carne por oveja en función de la densidad del
pastoreo
1b
Changes in wool and meat production per hectare with stocking rate
Variations de la production de laine et de viande par hectare selon le taux de charge
Cambios en la producción de lana y de carne por hectárea en función de la densidad de
pastoreo
Similar patterns emerge with respect to milk and manure output compared with meat output at various densities. These relationships are modelled in Figure 2, which compares meat production from a cohort of steers with milk production from cows and manure output from both groups. The model predicts that cows will continue to produce milk at high densities when steers are losing weight. According to these calculations, even if humans drank only half of all milk production, they would still get more megajoules (MJ) from this source than they would from meat output, and the remaining milk would feed calves. As might be expected, the volume of manure output increases with density and is highest at the maximum density.
Note: Output over six months of milk from a cow herd compared with meat from a steer herd. Outputs are converted to energy contents. Herds are on similar land receiving a steady daily amount of rainfall totalling 300 mm. Grass grows at 2.2 kg/ha/mm of rainfall. Reasons for higher energy output from cows are higher efficiency of conversion of metabolic energy to net energy in lactation compared with mass gain; and higer intake rates per unit body mass of lactating cows compared with steers. Manure output from cows continues to increase with stocking density. Equations for production and maintenance are from Konandreas and Anderson (1982).
2
Stocking rate and production of milk or meat and manure
Relation entre le taux de charge et la production de lait, de viande et de fumier
Densidad de pastoreo y producción de leche o de carne y de estiércol
The combined production effects of indigenous breed characteristics and agropastoral output mixes are depicted in Figure 3, which presents a hypothetical revenue or combined physical product curve from ranch and pastoral stock at different densities. The salient differences between commercial and pastoral productivity at alternative stocking densities are as follows:
MP = Maximum profit
MY = Maximum yield
MO = Open access equilibrium
K = Ecological carrying capacity
3
Ranch and pastoral revenues and costs at various stocking rates
Recettes et dépenses propres aux systèmes de ranching et d'élevage pastoral selon le
taux de charge
Beneficios y costos de la ganadería comercial y pastoral con diversas densidades de
pastoreo
In sum, the shape of the pastoral output curve combined with a "flat" variable cost curve minimizes the differences between MP, MY, MO and K; it also positions these thresholds at very high stocking densities relative to commercial ranching. These factors help explain why African pastoral and agropastoral producers are inclined to push their stocking rates towards the ecological limit (Tapson, 1990; Jarvis, 1991). The next section examines the effects of these high stocking rates on pastoral output levels.
Table 1 summarizes evidence on the comparative productivity of
commercial ranching and African pastoralism. All the studies cited in Table 1 attempt to
capture in one unit of measure - be it protein, energy or cash values - the combined
utility of the diverse array of products generated by indigenous African herds. These
studies also express output on a per hectare basis, which makes possible a direct
comparison of rangeland productivity under various production systems.
The methodological problems involved in comparing fundamentally different systems are
immense, as are the problems of accurately assessing the combined value of products as
diverse as milk, meat, fibre, power and fertilizer. The safest interpretation of the
values in Table 1 is that they are reasonable, although rough, approximations of the
relative output of commercial ranching and pastoralism. The results are nonetheless
compelling. The case material comes from West, East and southern Africa, so the
geographical spread of the evidence is quite good. And across sub-Saharan Africa it would
appear that indigenous open range pastoral systems achieve, at the very least, output
parity with ranching systems in comparable natural environments; the indigenous systems
routinely exceed by a wide margin the yield from comparable commercial systems.
The evidence summarized in Table 1 supports the hypothesis illustrated in Figure 3: the
shift from indigenous African pastoralism to market-oriented ranching is not a process of
intensification. Table 1 suggests that output from the land falls in the transition to
commercial meat production, as does the intensity with which livestock capital is utilized
per hectare (i.e. the stocking rate).
1
Comparative productivity1 of commercial ranching and open-range pastoral
production
Productivité comparative des systèmes de ranching commercial et d'élevage pastoral en
parcours libre dans des conditions écologiques
Productividad comparativa de la ganadería comercial y la producción pastoral en
pastizales abiertos en condiciones ecológicas comparables
Country |
Pastoral vs ranch productivity (Ranching = 100%) |
Units of measure |
Mali |
· 80-1066% (relative to
United States) |
kg protein production/
ha/year |
Ethiopia (Borana) |
· 157% (relative to Kenya) |
MJ/ha/year of gross energy edible by humans |
Kenya (Maasai) |
· 185% (relative to East Africa) |
kg protein production/ha/year |
Botswana |
· 188% (relative to Botswana) |
kg protein production/ha/year |
Zimbabwe |
· 150% (relative to Zimbabwe) |
$Z/ha/year |
1 Under comparable ecological conditions.
Sources. Mali: Penning de Vries and Djiteye (1982); Ethiopia: Cossins (1985);
Kenya: Western (1982); Botswana: de Ridder and Wagenaar (1984); Zimbabwe: Barrett (1992).
The conversion of indigenous African livestock husbandry to
market-oriented meat production may depress stocking rates in the long term but this
conversion entails real costs in terms of declining total livestock output per unit land
area. Some classes of pastoral herd owners are likely to be able to bear these costs
profitably while others are not. In particular, increased market involvement and lower
stocking rates affect large and small African herd owners very differently.
Systems research conducted by the former International Livestock Centre for Africa (ILCA),
now incorporated in the International Livestock Research Institute (ILRI), among the
Maasai in Kenya and the Borana of Ethiopia has provided some of the best quantitative
evidence of these effects. Table 2 summarizes the results.
2
Comparative productivity of Maasai and Borana herds
Productivité comparative des troupeaux Masaï et Borana
Productividad comparada de hatos Maasai y Borana
Maasai |
Borana |
|||||
Small herd |
Medium herd |
Large herd |
Small herd |
Medium herd |
Large herd |
|
Harvested milk/cow/year (kg) |
235 |
175 |
50 |
251 |
238 |
219 |
Ratio of calf weight gain to harvested milk |
3.58 |
5.58 |
19.32 |
1.99 |
2.73 |
5.92 |
Output/head/year (US$) |
24 |
22 |
16 |
27 |
24 |
20 |
Hours of labour/TLU1 |
0.8 |
0.5 |
0.2 |
- |
- |
- |
1 Tropical livestock unit.
Sources: Bekure et al. (1991); de Leeuw (1995).
Among the Borana and Maasai, physical output - an indexed value combining
meat and milk offtake - was relatively constant irrespective of differences in herd sizes
(de Leeuw, 1995). This output was, however, captured in very different ways by large and
small herd operators and with very different financial consequences. As herd sizes
increased there was a steady drift away from labour-demanding dairy production towards
meat production. These trends are indicated in Table 2 by absolute declines in harvested
milk output, by the increasing importance of meat relative to milk and by declining labour
inputs per head as herds grow in size. Through the expenditure of household labour on
dairying, poorer households maximized their meagre income, added considerable value to
herd output and achieved higher economic returns per animal than those with larger herds.
These changes in management objectives were sustained by parallel changes in husbandry
practices and herd structure. Among the Maasai there was a tentative movement by large
operators towards what commercial ranchers would identify as a "growing out"
operation, to exploit the demand among neighbouring cultivators for plough oxen. This
shift in production orientation is evidenced by an increase in the retention of older male
animals in larger herds, a trend which was particularly advanced among Maasai who held
title to their land and purchased and finished rather than bred much of their male stock
(White and Meadows, 1981).
Comparable shifts in production orientation among the Borana were indicated by differences
in the management of the adult female herd component. Small Borana herd owners prolonged
the lactation period of their cows, thereby increasing milk production but delaying
reconception, which reduced calving rates and compromised meat production. Large
operators, on the other hand, accepted reduced milk yields and lactation periods in return
for improved calving rates and higher meat output. In essence, large Borana herds were
moving away from traditional dairying and towards a rudimentary cow-calf production system
(de Leeuw, 1995; Bekure et al., 1991; Coppock, 1994).
Although influenced by different regional and local conditions, evolutionary processes
broadly similar to those taking place among the Maasai and Borana have been documented for
rangeland areas throughout Africa (Kerven, 1992; Sikana and Kerven, 1991; Sikana, Kerven
and Behnke, 1993). This material substantiates a general shift to more extensive,
meat-oriented commercial ranching among larger African pastoral producers. At the same
time, the extensive nature of range-based meat production renders this form of pastoralism
unattractive to small herd owners who have abundant supplies of domestic labour but few
animals. While they may be heavily involved in animal sales to support essential grain
purchases, these smaller operators retain many "traditional" husbandry practices
which support combined meat and milk production, for both sale and home consumption.
The stocking rates associated with these different forms of emergent commercial production
are difficult to document, since both large and small herds mingle on the open range. The
interests of large producers can be judged, however, by their persistent attempts to limit
grazing pressure by restricting access to natural resources. Range enclosure and water
privatization movements dominated by large herd owners have been widely documented in
semi-arid Africa (Grandin, 1986; Ensminger, 1990; Peters, 1984; Hogg, nd; Behnke, 1985 and
1988; Schlee, 1991). In many cases these abridgements on customary tenure have been
bitterly contested within and between agropastoral communities.
The preceding analysis provides some insight into the divergent interests which motivate
these struggles over resource control. Destocking does not present the unqualified
benefits - more grass, more output and more profit - envisaged by its enthusiastic
proponents. It all depends on the animal densities prevailing when destocking begins and
ends and on what different man-agers want from the system, given the assets they control.
One herder's intensification is another's overstocking. The stocking rates which would
maximize profit (MP), are not identical to those that would maximize physical output (MY),
or the stocking rates which would suit the interests of poor people requiring free access
to rangeland resources (MO). Emergent commercial pastoralists and subsistence-oriented
herd operators maximize different kinds of livestock output, possess very different assets
and find different stocking rates appropriate to their purposes. With the
commercialization of Africa's pastoral economies, these divergent perspectives have come
to represent the conflicting interests of different classes of African livestock keepers.
We are now in a position to re-examine the longstanding debate on the
nature of pastoral price responsiveness and its effect on stocking densities. Some
observers have argued that pastoralists exhibit a negative or perverse response to price,
selling fewer animals as prices improve and thereby increasing stocking densities. Others
view improved livestock marketing systems and prices as mechanisms for increasing offtake
and reducing animal numbers. Empirical evidence which might decide this issue is
inconclusive and obscured by war, drought, erratic recording, multiple official and
unofficial marketing channels and the operation of very imperfect commodity and input
markets in semi-arid Africa. Yet the situation is also genuinely complex, as the following
discussion shows.
Economic optima (MP, MO) do not routinely coincide with biological optima (MN, MY, K) (see
Art. 1, Fig. 2), and the degree of disjunction depends on economic factors. As variable
costs (purchase and management costs per animal) increase, the stocking density that
maximizes net revenue will decrease, all else being equal. This relationship is conveyed
in Figure 4, where the revenue from a grazing system is held constant at two different
levels of variable costs, denoted costs and costs*. With increasing variable expenses
(from costs to costs*), there is a decline in both the stocking density, which produces
the highest net profit (from density MP to MP*), and the density at which open access
equilibrium occurs (from MO to MO*), (Jarvis, 1991; Wilson and Macleod, 1991; Mentis,
1977).
4
Impact of increase in variable costs on pastoral optimum
Incidence de l'accroissement des coûts variables sur l'optimum
Efectos del aumento de los costos variables sobre un sistema pastoral óptimo
Figure 4 suggests a number of policy-relevant conclusions. One obvious
observation is that the provision of subsidized or free inputs for pastoral producers -
feed or mineral supplements, water development or veterinary support - will reduce
variable costs and encourage high stocking rates. Programmes of subsidized input supply
are therefore inconsistent with a commitment to controlling livestock numbers.
Conversely, a grazing tax levied on each animal retained on the range would inflate
operating costs and depress optimal economic stocking densities. As Jarvis (1991) has
argued, however, taxation measures which were draconian enough to have a significant
impact on livestock numbers would be unworkable in practice. Destocking enforced through
taxation would drive poorer producers off the ranges even as production was rising, while
those who remained would be no better off than before, since all increases in productivity
would be passed on to the government. Annual adjustments in taxation levels to take
account of rainfall variability would also be politically difficult since they would
inflate tax levies to reduce herd sizes during periods of insufficient rainfall, when
pastoralists are already suffering hardship. (For a fuller discussion, see Jarvis, 1991.)
A livestock grazing tax may, therefore, be best viewed not as an effective instrument of
destocking but as a potential source of government revenue which is not inherently at
cross purposes with destocking initiatives.
Figure 4 depicts a situation in which operating costs increase while product prices remain
constant, i.e. the classic cost-price squeeze. Figure 5 illustrates the reverse process in
which variable costs remain stable, prices improve and herd owners expect prices to remain
high for some time.
The effect of these price changes on output levels will depend on the initial stocking
rate. When the initial stocking rate is below MY, MP moves to the right (to MP*), closing
the gap with MY and increasing both the stocking rate and the volume of product output. If
producers are operating at densities greater than MY, such as MO in Figure 5, price
increases will have the perverse effect of inflating stocking densities (to MO* in Fig.
5), while depressing levels of physical output as stocking rates approach K. In either
case, price increases uniformly encourage increases in stocking rate, at least in the
short term and among commercial producers (Jarvis, 1991).
5
Impact of increase in product price on pastoral optimum
Incidence de l'accroissement du prix des produits sur l'optimum pastoral
Efectos del aumento de precio de los productos sobre un sistema pastoral óptimo
Cattle surviving on maize stover in the Gambia
Bétail survivant grâce à une alimentation composée de tiges de maïs en
Ganado que sobrevive alimentándose de forraje de maíz en Gambia
The situation may be more complex on Africa's open ranges where producers tend to be involved in production both for the market and for home consumption. Intensification, Gass and Sumberg (1993) have observed, can proceed in two distinct ways:
"...through the progressive modification of existing production systems, or the establishment of entirely new systems. The former represents a positive movement along the production function or an outward movement of the production function itself; the latter, a move to a new production function."
Figures 4 and 5 may accurately depict the immediate response of
pastoralists to changing marketing opportunities - a readjustment of their position in
terms of their existing production function and husbandry system. But this does not
necessarily tell the whole story. Emergent commercial producers - larger Borana and Maasai
herd owners, for example - will be tempted not simply to adjust their subsistence-oriented
husbandry system to changing prices but, eventually, to shift to new production systems
tailored to commercial output. If the reasoning summarized in Figure 3 is accurate, this
shift will entail reductions in stocking rates which would dwarf the effects of short-term
adjustments to price fluctuations.
The effects of prices on pastoral stocking rates are therefore complex and, unless it is
carefully analysed, the evidence may appear contradictory. In the short term, and
especially among smaller producers forced into distress sales when conditions are
difficult, improved prices may signal a period of recapitalization, individual herd growth
and higher overall stocking rates. In the long term, and especially among larger producers
in a position to shift to commercial meat production, higher prices are likely to
encourage the adoption of new market-oriented husbandry systems and lower stocking
targets. Whether price changes increase or decrease stocking rates in a particular
situation will, therefore, depend on the working definition of "long" and
"short" term, adopted by researchers and policy-makers, the proportion of the
total herd held by large or small owners and the magnitude and permanence of the price
changes required to induce spontaneous commercialization.
Half a century of experience suggests that mandatory stock sales and stocking quotas offer no permanent remedy for high stocking rates in semi-arid Africa. Extension campaigns to educate herd owners on the virtues of low stocking rates are also ineffectual if high rates serve the interests of some producers. Effective policy measures to reduce pastoral stocking rates must therefore reverse the conditions that promote high densities. But the preceding analysis suggests that these important causative factors are not always susceptible to administrative manipulation, that the advantages of manipulation may not outweigh the costs and that manipulation may not yield predictable results. We are left with a few positive recommendations and a humbling awareness of what we do not yet know:
- species diversity of pastoral livestock holdings and an emphasis on
quick-breeding small ruminants in the recovery period following a drought;
- a strong female component in pastoral herd structures, which favours both dairy
production and rapid post-drought herd growth;
- a reliance on indigenous breeds and species that are drought-tolerant;
- spatial dispersal of livestock assets and risk spreading through stock loans and herd
splitting;
- herd mobility to exploit heterogeneous environ-ments and diminish the impact of
localized resource deficiencies;
- redistribution of livestock and the sharing of their produce.
The prevalence of these practices suggests that minimizing climatic risk
and maintaining minimally acceptable yields and herd numbers are important pastoral
objectives. Yield optimizing calculations are therefore but one part of a large number of
tactical considerations for herd owners. While they may accurately depict some factors and
goals which influence decision-making, the simple models presented here are probably
insufficient to predict pastoral behaviour in risky environments.
In sum, the theoretical simplicity depicted in this analysis gives way on close inspection
to considerable empirical complexity - in different kinds of natural environments, for
different classes of producers, under variable economic conditions and over different
time-scales. Careful field research may clarify the situation but the diverse ways in
which producers can respond to stimuli suggest that true, non-trivial generalizations may
be difficult to discover and that, for the foreseeable future, most recommendations
regarding destocking and overgrazing will be site- and situation-specific.
The authors are grateful to Ken Hodgkinson and Arthur Knight for their comments on the first draft of this paper, which was consequently much modified.
The mainstream view of rangeland degradation is that it is very widespread, serious and often caused by grazing (Williams, McCarthy and Pickup, 1995). A more sceptical school finds little African evidence in terms of declines in the productivity of either vegetation or livestock (Sandford, 1983; Behnke, Scoones and Kerven, 1993; Shackleton, 1993; Tapson, 1993). Despite this, the destocking of Africa's communal rangelands has been advocated since early this century for the good of both the pastoralists and the land (Beinart, 1984). The first and second of our papers in this series examined the question, "how many animals should there be for purposes of production?" We showed that the appropriate stocking rate for production cannot be set except in relation to the production strategy and the social and economic circumstances of the rangeland user - there is no single optimum density. The rate chosen to meet those objec-tives is selected by current generations for their purposes. The rate chosen is usually higher than range scientists would recommend. What if this causes degradation, thus reducing the productivity for future users? Should destocking be promoted in their interests? The dilemma is explored in this paper, using a framework for analysing resource use conflicts modified from Cullen (1990). Two of the three parts of the framework are described below. The third - structural elements - is covered at the end of this article.
Information elements. Disagreements over stocking rates and degradation may arise from differences in interpretation of the environment. These may result from choice of theory as well as from the difficulties of determining trends and time-scales and coping with the spatial variability and unpredictability of rangeland systems.
Psychological, cultural and value elements. Cultures are the sources of belief systems and values. The environment is perceived through a cultural filter. Most resource use conflicts can be interpreted as differences in values, which are expressed as the negotiating stances or interests of protagonists. Conflicts can sometimes be resolved through the negotiation of compromises between value sets; sometimes both protagonists can gain.
The purpose of this paper is to encourage range scientists, administrators, policy-makers and development agents to take a more critical approach to questions of stocking rate and sustainability. For too long, blinded by their own sources and interpretations of information and their own cultural filters, and driven by the structures they serve, public servants have been unable to see the views of other protagonists. We shall promote an approach which pays balanced attention to the information, psychological, cultural, value and structural elements in conflicts over stocking rate and range degradation.
The ostensible purpose of information on degradation and stocking rates is to clarify issues and promote better management or use of land. In practice, there are conflicting views on specific cases, principles and theories, so "facts" about degradation and sustainability confuse as often as they clarify (Behnke, Scoones and Kerven, 1993; Abel, 1993a). This can be because assessors who reach opposing conclusions are making their judgements in relation to different land uses. This is discussed in the section Psychological, cultural and value elements. Other reasons for confusion are discussed in the present section under the headings Choice of theory, Trends and time-scales, Spatial variability and Unpredictable systems.
Judging the decline in the performance of animals seems an obvious way
of assessing degradation (Abel and Blaikie, 1989; Fowler, 1981; Tapson, 1993). However, it
is usual for the reproductive rate to fall and mortality to increase as the stocking rate
rises because of competition for fodder. This may or may not indicate range degradation.
An estimation of the change in output per hectare would be more relevant. However, without
a linkage to changes in soils or vegetation, this approach offers no explanation of cause
and, consequently, no sense of trend, including reversibility. As secondary production is
the main purpose of pastoralism, its measurement should be one of the criteria of
degradation, but not the only one. Hence the use of approaches based on vegetation change.
Increase in shrubs is a criterion commonly used in the assessment of degradation because
of the reduction in grass growth which this (usually) causes. The balance of shrubs to
grass can become locked into a stable equilibrium through the inability of the grass to
obtain sufficient water and produce enough fuel to burn the shrubs, but generally the
balance is reversible (Walker and Noy-Meir, 1982). A reversible change should not be
called degradation, otherwise virtually all human economic activities are seen as its
agents and the term becomes useless (Abel and Blaikie, 1989). Moreover, shrubs can add
organic matter and nutrients to depleted soil, adding confusion to the label
"degradation".
When shrub encroachment occurs on ranches, cattle are deprived of grass during the growing
season, slowing growth, increasing costs, delaying benefits and reducing profits. On
ranches, an effectively irreversible increase in shrub cover can be classified usefully as
degradation. On communal lands, the main concern is output per unit of land (first and
second papers in this series). This is determined primarily by stocking density, which
itself depends on survival during dry seasons. Browse is often the only feed at such
times. Its quality is usually insufficient for production, but good enough for survival
(Abel et al., 1987). It is therefore likely that shrub encroachment is a useful
feature of a system managed for subsistence production. When cattle are run in parallel
with or replaced by goats and camels, the use of shrub cover as an indicator of
degradation makes no sense at all.
Probably the greatest source of confusion over range degradation is the succession theory,
which has been the mainstay of range management for most of this century. According to it,
a vegetation community at a site is believed to be able to exist in a number of stages,
each characterized by a particular composition of species. At the climax stage, the
community is in equilibrium with the climate. The equilibrium may be disturbed by fire,
herbivory, drought, cutting or some other factor, and thus regresses to an earlier stage.
Development towards the climax resumes if the disturbance is removed. The range manager
manipulates vegetation using fire and grazing intensity. Just enough grazing pressure or
burning should be applied for the range to regress to the stage best suited to the
domestic species in use. Pasture regressing too far from the climax, as revealed by a
replacement of climax species with plants of lower stages, can be rested.
The succession theory is the basis of the concept of rangeland carrying capacity,
according to which degradation occurs only if the carrying capacity is exceeded. The
carrying capacity is seen as a threshold, and managers must maintain stocking rates on the
safe side of it (however, see Fig. 3). It is being exceeded if species from too low a
successional stage begin to increase. This is unacceptable as an operational definition
because there is no way of deciding what an acceptable mix of species is. In practice,
range scientists give their opinion on this in an example of circular reasoning: the range
is degraded because species of too low a successional stage are established; they
"know" that these species are undesirable because they grow on degraded
rangeland. The theory offers no way of resolving a difference of opinion over whether or
not degradation has occurred.
The theory can predict changes in species composition resulting from grazing in conditions
where reliable rainfall permits equilibria to be established among animals, plant
communities and climate (Coppock, 1993). Over most of Africa's communal rangelands,
rainfall is highly variable and equilibria cannot be established. In these circumstances
rainfall, not stocking rates, is the chief determinant of plant species composition
(O'Connor, 1985). Other weaknesses in succession theory are discussed in Westoby, Walker
and Noy-Meir (1989), Behnke, Scoones and Kerven (1993) and Abel (1993b).
A typical scene of marginal grazing in West Africa
Une scène typique d'un patûrage marginal en Afrique de l'Ouest
Escena típica de pastoreo marginal en Africa occidental
The state and transition concept has been offered as a framework for managing non-equilibrial grazing systems (Westoby, Walker and Noy-Meir, 1989; Friedel, 1994). The range is seen as being in one of a set of quasi-stable states or in a phase of transition between states. The system does not stabilize in mid-transition. Changes between states are triggered by weather, fire or herbivory. The pastoralist is seen as a strategist and tactician who seizes opportunities for tipping the system towards a favourable state, such as burning shrubs when fuel loads are high or shrubs young. Adverse conditions are met with flexible responses, such as rapid destocking when that is appropriate (Hodgkinson, 1993). This is in tune with the opportunism of African pastoralists (Sandford, 1994). State and transition is an approach to decision-making rather than a theory with predictive power. It does not contain operational definitions of degradation, sustainability and productivity: the desirability of the state to be aimed for is left to the decision-maker. This may be a desirable feature, but the reason for it is that the approach is yet to have built under it theoretical foundations concerning soil and plant processes. Thus, the approach also lacks explanations of causal linkages among degradation, sustainability and productivity. We do not see the approach as a "stand-alone" theory, and its usefulness will depend on the skill with which it is linked to theories of ecological function.
The aim in assessing range degradation is to estimate the direction and
level of some undesirable change. The issue of "desirability" is discussed under
Psychological, cultural and value elements. If the change is reversible, it should not be
classified as degradation, as justified above. Some changes are clearly irreversible - the
extinction of species for example. Others may be thought of as reversible or permanent,
depending on the time-scale. Loss of perennial grass cover may be effectively permanent
for a rancher but reversible on a time-scale of centuries. Professional and occupational
time-scales vary widely. For example, soil loss may be effectively permanent to the range
scientist, but reversible to the geomorphologist. Confusion arises when protagonists think
they are arguing over the occurrence or otherwise of degradation when in fact they are
looking at different time horizons. The only clarification needed in these circumstances
is that time horizons and reasons for selecting them must be explicit and scaled to the
land use in question. These issues are illustrated in the following case-study.
Biot (1993) developed a method for estimating irreversible change over time. Changes in
primary productivity are calculated as a function of changes in soil depth, organic matter
and clay content. Structure and limitations of the primary production model are in Biot
(1990). Abel (1993b) linked the degradation model to a secondary production and financial
model. Sample outputs (from Abel, 1996) are shown below. Figures 1 and 2 show degradation
under the current stocking density in a communal rangeland in eastern Botswana. It is
compared with the rate of degradation under the "carrying capacity" recommended
by the Ministry of Agriculture (Field, 1977) and with no animals. Degradation occurs
without animals because soils on this landscape were formed under a different climate and
are not yet in equilibrium with the present one (Biot, 1988). Soil life is extended
greatly by destocking but on a time-scale that is much longer than is normally considered
by pastoralists, officials and development planners.
1
Decline in soil depth
Diminution de l'épaisseur du sol
Disminución de la profundidad del suelo
2
Decline in grass biomass production over time
Diminution de la biomasse herbacée en fonction du temps
Disminución de la producción de biomasa de gramíneas con el paso del tiempo
The influence of stocking rates on the length of the productive life of a soil is shown in Figure 3 (unlike Figs 1 and 2, it uses default parameters, so soil life is shorter). There is no evidence of a "carrying capacity", i.e. a threshold below which degradation does not occur. The effect of destocking on soil life is greater at the higher stocking rates and the massive destocking required to descend to the officially recommended "carrying capacity" would buy relatively few extra years.
3
Stocking rate and soil life
Taux de charge et durée de vie des sols (paramètres implicites)
Densidad de pastoreo y duración del suelo
Much of range ecology has been site-based until recently. Spatial
variability was regarded as an annoying feature which raised survey costs and hid
otherwise statistically significant differences between treatments in grazing and burning
experiments. Assessors thought they could measure degradational change at a site and scale
this up to estimate degradation at the scale of the catchment or vegetation type (Friedel,
1994). However, when soil erodes it does not usually wash quickly into the sea where it is
lost to future generations. For the Atlantic drainage of the United States, some 90
percent of the soil lost from the uplands during the last 200 years has been stored on
hill slopes and in valleys (Trimble, 1975 and 1983; Meade, 1982, quoted in Wasson, 1987).
Alarming soil erosion figures are commonly those for gross loss, and they do not show
deposition (Stocking, 1987). Moving soils from hillsides to valleys is not necessarily
detrimental from a human perspective.
Neglect of spatial variability has also meant ignoring the role of spatial pattern in
landscape function (Friedel, 1994). The landscape approach of Ludwig and Tongway (1995)
shows that the patchiness of range vegetation in Australia - lines and clumps separated by
bare areas - is necessary for landscape function. The patches salvage nutrients, mineral
particles and organic matter from runoff and wind. They also recycle nutrients in situ.
They are dependent on neighbouring bare areas for water and nutrients - without runon,
rainfall is insufficient to maintain the biomass of vegetation observed on the patches.
Grazing disrupts this system by reducing patchiness and thus the ability of the system to
retain, capture and recycle scarce resources. Response to rainfall and changes in
patchiness is measurable using satellite imagery, subject to the limitations of spatial
resolution (Knight, 1995). Degradation is defined in terms of a lowered response to
rainfall.
The consideration of sites as exporting, transmitting and receiving areas for mineral
material, organic matter, water and minerals should enable in principle the estimation of
net rates of degradation. The "erosion cell" concept attempts to achieve this
(Stafford-Smith and Pickup, 1993). An erosion cell is a land unit of variable size
containing a sediment production zone, a transfer zone and a sink. The cell moves upslope
as erosion progresses. The source zone is likely to be depleted in nutrients, organic
matter, silt, clay and seeds and is also likely to have shallower soils because of the
export of material. A decline in primary productivity, even in the absence of herbivores,
is likely. The transfer zone is by definition an unstable place where materials are in
transit. Soil depth and thus moisture storage capacity are likely to be greater than in
the source zone. The transfer zone is being enriched by material, including seed, from the
source zone and is therefore more productive than it. Mineral and organic matter
accumulate in the sink and enrich it, so soil depth and productivity will be greater.
Changes in the soil cause a disequilibrium, with opportunities for changes in species
composition being realized through the seed source from upslope. By integrating changes in
productivity of all the cells in a landscape, the net change, or degradation, could in
principle be estimated.
Pickup and Chewings (1986) have shown that the spatial patterns of erosion cells are
detectable on Landsat imagery. The attractiveness of this approach lies in its link to
fundamental landscape processes, its clear definition of degradation and its attempt to
estimate the net effect of degradation on the landscape. It has not so far been able to
predict trends in productivity but it should be possible to predict where the greatest
changes will occur.
The application of an inappropriate theory in the assessment of degradation is clearly a human error. However, the development of a more appropriate theory, such as landscape ecology, does not necessarily give us the power to predict, which was lacking in the succession theory. Rangeland systems tend to be intrinsically unpredictable, driven as they are by highly variable rainfalls on which there is no feedback from the rangeland, and comprising as they do multiple, interacting curvilinear and non-linear processes. In the case of soil loss and stocking densities, at least the following interacting relationships are curvilinear:
Intuition does not cope well with interacting curvilinear systems. Even with a single relationship, a rate of degradation can increase suddenly once a certain point of maximum curvature is passed. This can be magnified if thresholds of more than one relationship are crossed simultaneously. Curvilinear relationships are, however, amenable to modelling. Non-linear processes are more difficult. An example of a non-linear process is a stepwise increase in the erodibility of soil as the A horizon is removed to expose a dispersible B horizon. Another is the sudden change in ground cover characteristics (and therefore in the rate of soil loss) when a pasture is consumed by a locust swarm. Such changes are hard to model mathematically. More serious, because such systems can flip readily between different states, their behaviour may be very hard to predict and collecting extra data may not improve our ability to estimate trends, for the difficulty lies in the behaviour of the system rather than in the inadequacy of our models.
In the previous section we discussed how disagreement over information
can be a source of conflict. Another cause is human subjectivity. In part, this is the
result of our own biological limitations. We can receive only a tiny fraction of the
information in the electromagnetic spectrum and rely on instruments to collect information
about most of it. We live only a short time and tend to interpret changes only in relation
to our own brief life spans. Thus, cyclic or episodic changes which occur over a span of
time longer than the human lifetime may be seen as trends. Subjectivity is also explained
psychologically. According to Kelly (1991), data about our environment (sensed directly or
through the media) is interpreted in the mind according to a set of personal constructs
which both facilitate the processing and filter it selectively. Personal constructs differ
between cultures and subcultures. Aboriginal Australians do not have the linear,
progressive view of time held by European cultures. Geologists and evolutionary biologists
think in terms of millenia, ecologists in terms of centuries. Economists are constrained
by discount rates to time horizons of 20 years or so. Pastoralists may be driven by
survival to think only as far ahead as the next few seasons. Much of the fury in the
debate over rangeland degradation is due to the use of different time-scales by different
parties. The specification of a time-scale is a prerequisite for rational analysis.
Personal construct psychology holds that if one person is to communicate effectively with
another, it is not necessary to hold the same personal constructs but only to understand
the constructs of the other person. We may hold a different opinion from another party
about whether a particular change should or should not be allowed to happen, but
acceptance that our time-scales are different provides some basis for negotiation. Figures
1, 2 and 3 are an attempt to set out some of the parameters for such a debate.
The personal construct theory may offer an explanation of why the succession theory was
embraced so strongly and for so long despite contrary evidence: it may have fitted a
Western cultural belief that nature is both predictable and manageable. Neither is
necessarily true.
Our personal construct set causes us, while creating information, to interpret data in
certain ways. Thus, most disputes over information can be reinterpreted as conflicts over
values (Cullen, 1990). Obtaining more data will not resolve the conflict if values remain
in opposition. Science cannot create value-free information for arbitration in such
conflicts because it is not a unity but a set of disciplines, each with its own subculture
and value set (such as the different time-frames of ecologists and geologists). Nor can
science establish absolute truth or falsity, because "proof" can only be
relative, never
absolute (Chalmers, 1982). In disputes over range degradation and stocking rate, all
evidence, whether scientific or other, should be empirically validated, evaluated in
relation to the ideology of its constructors, compared with other sources of information -
local knowledge in particular - and judged according to its own merits.
A framework is needed for analysing conflicts of value over range degradation. A
human-centred framework is offered by environmental economics (Pearce and Warford, 1993).
The total economic value (TEV) of the environment, or a part of it, is:
TEV = DUV + IUV + OV + IV
This (crude) framework spans a range of values. Any subcultural group is
likely to hold a mix of the four values, with emphases varying between subcultures. Those
concerned with the short term and with material values (such as food for survival), and
who place humans outside nature, would emphasize the DUV at the expense of other values.
Groups, such as Australian Aborigines, who see humans as one part of nature, would
emphasize the IV. Those with a broad view, and those whose well-being is harmed by
degradation of life support systems, would focus on the IUV. Groups concerned about the
longer term would take the OV into account.
To illustrate how this framework might operate, the decline in primary productivity in
Figure 2 is converted to a gross margin in Figure 4 (Abel, 1993b). The gross margin is the
value of all livestock outputs, whether sold or consumed by the producer, valued at market
prices, minus variable costs of production. Figure 4 shows that, according to the model,
the gross margin at the current stocking rate does not fall below that at the recommended
rate for almost 600 years. However, benefits or costs not paid immediately are worth less
than their face value because of the cost of waiting to receive a benefit as well as the
benefit of delaying payment of a cost. The conventional way of estimating the benefit or
cost of delay is to discount at the rate at which capital invested now could earn
interest. A 6 percent real discount rate (i.e. corrected for inflation) is applied to the
gross margins to produce the curve in Figure 5. As discounting drives gross margins almost
to zero by year 60, the graph is truncated here.
4
Decline in gross margins
Réduction des marges brutes
Disminución de los beneficios brutos
5
Discounted gross margins
Marges brutes actualisées
Beneficios brutos descontados
The DUVs over time at the two stocking rates are the areas under each
discounted gross margin curve (capital costs are ignored here for convenience). The
current stocking rate is, from the stance of a short-term materialist, superior. Those
concerned with the benefits of future generations would consider the OV. They could not
estimate its value because the future use of the land is not known. We can, however,
estimate what it would cost to slow the decline in the OV by destocking: it is the area
between the two discounted gross margin curves in Figure 5 (for simplicity, IUVs are
ignored for now).
At this point, an interest group might object that using the range for pastoralism is
harming wildlife conservation. They may present their argument in terms of lost revenues
from tourism (DUV) or lost IV or both. Another group which lives downstream might complain
that the hydrological impacts of grazing are harming the ability of catchments to supply
their water - an IUV. The complexity of demands on the range in our simple example is
beginning to look realistic. Following is a method of managing this complexity.
an effectively permanent decline in the rate at which land produces forage for a given input of rainfall under a given system of management. "Effectively" means that natural processes will not rehabilitate the land within a time-scale relevant to humans and that capital or labour invested in rehabilitation are not justified.... This definition excludes reversible vegetation changes even if these lead to temporary declines in secondary productivity. It includes effectively irreversible changes in both soils and vegetation (modified from Abel and Blaikie, 1989).
If the purpose of the analysis is conflict resolution, then it might be set within the context of a negotiation procedure, wherein tradeoffs are made between generations and between interest groups. In such a case, no one pastoral group can be regarded as homogeneous: subgroups will have differing production strategies, values, access to resources, time horizons and differing responses to direct destocking and to policy measures designed to induce destocking (second article in this series).
Overstocking - could this have been avoided?
Surpâturage - aurait-on pu l'éviter?
Sobrepastoreo: ¿hubiera sido posible evitarlo?
At the beginning of this paper we said we wished to encourage range scientists, administrators, policy-makers and development agents to take a more critical approach to questions of stocking rates and sustainability. We have argued that the creation of information is strongly influenced by psychology, culture, values and ecological and economic circumstances. For example:
The influence of culture and values on the creation of information
applies as much to natural and social scientists as it does to officials and pastoralists.
Social scientists tend to support the needs and wants of pastoralists, who are often poor,
politically weak and vulnerable to governmental decrees on destocking. They might filter
environmental information and deny evidence of degradation. We suspect also that the
ideologies of range scientists and officials cause them to see degradation everywhere. We
believe that often what they are seeing is reversible change induced by increases in
stocking rates; sometimes they are seeing the redistribution of soils. Pastoralists may be
influenced by their values and priorities to ignore degradation that is already occurring.
Nevertheless, all these groups are specialists in some aspect of range use and have valid
viewpoints. By establishing negotiation procedures, and adopting something like the
analytical framework we have used, conflicts and tradeoffs could be identified and
bargaining initiated. However, to avoid ending on an unrealistic note, one more element in
the analysis of conflict must be considered.
Two elements of the framework for analysing resource use conflicts - information, and
subjective, cultural and value elements - were introduced at the beginning of this
article. The third type of element is structural. Examples are legal and other
institutions (land tenure for example), ethnic, class and caste structures, bureaucratic
and business organizations and, at the broadest scale, politico-economic structures.
Structures act as parameters in resource use conflicts, determining the limits of change.
They can also be a source of conflict, as when a bureaucratic organization with strong
values of timeliness, order and control attempts to change the activities of a pastoral
group, with an ideology of flexibility, opportunism and decentralized decision-making.
Bureaucracies have their own subcultures, aims, quests for power and political agendas,
and these may have to ameliorate before conflicting aims of pastoralists and officials can
be resolved. Structures may limit the scope for change in the short term but, in the
longer term, even the strongest are susceptible to social, economic and environmental
pressures.
Abel, N. 1993a. Reducing cattle numbers on southern African communal
range: Is it worth it? In R.H. Behnke, I. Scoones & C. Kerven, eds. Range
ecology at disequilibrium. New models of natural variability and pastoral adaptation in
African savannas, p. 173-195. London, ODI/IIED/Commonwealth Secretariat.
Abel, N. 1993b. What's in a number? The carrying capacity controversy on the
communal rangelands of southern Africa. Monograph No. 14, School of Development Studies,
University of East Anglia, Norwich, UK. 420 pp. (Ph.D. thesis)
Abel, N. 1996. Rangelands sustainability and ecological economics: an African case
study. Ecological Economics. (in press)
Abel, N. & Blaikie, P.M. 1989. Land degradation, stocking rates and
conservation policies in the communal rangelands of Botswana and Zimbabwe. Land
degradation and rehabilitation, Vol. 1, p. 101-123.
Abel, N., Flint, M., Hunter, N., Chandler, D. & Maka, G. 1987. Cattle
keeping, ecological change and communal management in Ngwaketse. Addis Ababa, ILCA.
200 pp.
Barrett, J.C. 1992. The economic role of cattle in communal farming systems in
Zimbabwe. Pastoral Development Network Paper 32c, London, ODI.
Behnke, R.H. 1985. Measuring the benefits of commercial versus subsistence
livestock production in Africa. Agric. Syst., 162: 109-135.
Behnke, R.H. 1988. Range enclosure in central Somalia. Pastoral Development
Network Paper No. 25b. London, ODI.
Behnke, R.H. 1994. Natural resource management in pastoral Africa. Dev. Policy
Rev., 12: 5-27.
Behnke, R.H., Scoones, I. & Kerven, C., eds. 1993. Range ecology at
disequilibrium. New models of natural variability and pastoral adaptation in African
savannas. London, ODI/IIED/Commonwealth Secretariat.
Beinart, W. 1984. Soil erosion, conservationism and ideas about development: a
southern African exploration, 1900-1960. J. South. Afr. Stud., 11(1): 52-83.
Bekure, S., de Leeuw, P.N., Grandin, B.E. & Neate, P.J.H. 1991. Maasai
herding: an analysis of the livestock production system of Maasai pastoralists in Eastern
Kajiado District, Kenya. Addis Ababa, ILCA.
Biot, Y. 1988. Forecasting productivity losses caused by sheet and rill erosion. A
case study from the communal areas of Botswana. University of East Anglia, Norwich, UK.
225 pp. (Ph.D. thesis)
Biot, Y. 1990. THEPROM - an erosion productivity model. In J. Boardman,
I.D.L. Foster & J.A. Dearing, eds. Soil erosion on agricultural land.
Chichester, UK, Wiley.
Biot, Y. 1993. How long can high stocking densities be maintained? In R.H.
Behnke, I. Scoones & C. Kerven, eds. Range ecology at disequilibrium. New models of
natural variability and pastoral adaptation in African savannas, p. 153-172. London,
ODI/IIED/Commonwealth Secretariat.
Blaxter, K.L. 1962. Energy metabolism of the ruminants. London, Hutchinson.
Booysen, P. de V., Tainton, N.M. & Foran, B.D. 1975. An economic solution to
the grazing management dilemma. Proc. Grassland Soc. South. Afr., 10: 77-83.
Butterworth, M.H. 1985. Beef cattle nutrition and tropical pastures. New
York, Longman.
Carew, G.W. 1976. Stocking rate as a factor determining profitability of beef
production. Rhodesian Agric. J., 73: 111-115.
Caughley, G. 1981. Overpopulation. In P.A. Jewell, S. Holt & Hart, eds. Problems
in management of locally abundant wild mammals. New York, Academic.
Chalmers, A.F. 1982. What is this thing called science? St Lucia, University
of Queensland Press.
Conway, A.G. 1974. A production function for grazing cattle, 3. An estimated
relationship between rate of liveweight gain and stocking rate for grazing steers. Irish
J. Agric. Econ. Rural Sociol., 5: 43-55.
Coppock, D.L. 1989. Bigger calves make better cows: fact or fantasy in variable
environments? ILCA Newsl., 8(4): 1-3.
Coppock, D.L. 1993. Vegetation and pastoral dynamics in the southern Ethiopian
rangelands: implications for theory and management. In R.H. Behnke, I. Scoones
& C. Kerven, eds. Range ecology at disequilibrium. New models of natural
variability and pastoral adaptation in African savannas, p. 42-61. London,
ODI/IIED/Commonwealth Secretariat.
Coppock, D.L., ed. 1994. The Borana plateau of southern Ethiopia: synthesis of
pastoral research, development and change, 1980-91. Addis Ababa, ILCA.
Coppock, D.L., Swift, D.M. & Ellis, J.E. 1986. Seasonal nutritional
characteristics of livestock diets in a nomadic pastoral ecosystem. J. Appl. Ecol.,
23: 585-595.
Cossins, W.J. 1985. The productivity of pastoral systems. ILCA Bull., 21:
10-15.
Cullen, P. 1990. Values and science in environmental management. Paper presented at
Symposium on Water Management in the Alligator Rivers Region. University of Canberra, ACT,
Australia, Water Research Centre.
de Leeuw, P.N. 1995. Maasai-Borana pastoralism. (in preparation)
de Ridder, N. & Wagenaar, K.T. 1984. A comparison between the productivity of
traditional livestock systems and ranching in eastern Botswana. ILCA Newsl., 3(3):
5-6.
Donnelly, J.R., McKinney, G.T. & Morley, F.H.W. 1985. The productivity of
breeding ewes grazing on lucerne or grass and clover pastures on the tablelands of
southern Australia (IV): Lamb growth. Aust. J. Agric. Res., 36: 469-481.
Donnelly, J.R., Morley, F.H.W. & McKinney, G.T. 1983. The productivity of
breeding ewes grazing on lucerne or grass and clover pastures on the tablelands of
southern Australia (II): Wool production and ewe weight. Aust. J. Agric. Res., 34:
537-548.
Ensminger, J. 1990. Co-opting the elders: the political economy of state
incorporation in Africa. Am. Anthropol., 92: 662-675.
Field, D.I. 1977. Potential carrying capacity of rangeland in Botswana.
Gaborone, Land Utilization Division, Department of Agricultural Field Services, Ministry
of Agriculture.
Finch, V.A. & King, J.M. 1979. Adaptation to undernutrition and water
deprivation in the African zebu: changes in energy requirements. In Proc. Research
Coordination Meeting on Water Requirements of Tropical Herbivores Based on Measurements
with Titrated Water, Nairobi. Vienna, IAEA.
Fowler, C.W. 1981. Density dependence as related to life history strategy. Ecology,
63(3): 602-610.
Friedel, M.H. 1994. How spatial and temporal scale affect the perception of change
in rangelands. Rangeland J., 16(1): 16-25.
Gass, G.M. & Sumberg, J.E. 1993. The intensification of livestock production
in Africa: experience and issues. School of Development Studies, University of East
Anglia, Norwich, UK. (mimeo)
Gordon, H.S. 1954. The economic theory of a common-property resource. J. Polit.
Econ., 62: 124-42.
Grandin, B. 1986. Land tenure, subdivision and residential change on a Maasai group
ranch. Dev. Anthropol. Net., 4: 9-13.
Hart, R.H. 1980. Determining a proper stocking rate for a grazing system. In
K.C. McDaniel & C. Allison, eds. Grazing management systems for southwest
rangelands. Range Improvement Task Force, New Mexico State University, Las Cruces,
USA.
Hildreth, R.J. & Riewe, M.E. 1968. Grazing production curves (II): Determining
the economic optimum stocking rate. Agron. J., 55: 370-372.
Hodgkinson, K.C. 1993. Tactical grazing can help maintain stability of semi-arid
wooded grasslands. Proc. XVII Int. Grasslands Congr., p. 75-76.
Hogg, R. nd. The politics of changing property rights among Isiolo Boran
pastoralists in northern Kenya. In P.T.W. Baxter & R. Hogg, eds. Property,
poverty and people: changing rights in property and problems of pastoral development.
Department of Anthropology, University of Manchester.
Jarvis, L.S. 1991. Overgrazing and range degradation: the need for and the scope of
government policy to control livestock numbers. East Afr. Econ. Rev., 7: 95-116.
Jones, R.J. & Sandland, R.L. 1974. The relation between animal gain and
stocking rate: derivation of the relation from the results of grazing trials. J. Agric.
Sci. (Cambridge), 83: 335-342.
Kelly, G.A. 1991. The psychology of personal constructs. Vol. I. A theory of
personality. London, Routledge. 405 pp.
Kerven, C. 1992. Customary commerce: A historical reassessment of pastoral
livestock marketing in Africa. London, ODI.
King, J.M. 1983. Livestock water needs in pastoral Africa in relation to climate
and forage. ILCA Research Report No. 7. Addis Ababa, ILCA.
Knight, A.W. 1995. REMA: A neutral model to reveal patterns and processes of cover
change in wooded rangelands. Remote Sensing Environ., 52: 1-14.
Konandreas, P.A. & Anderson, F.M. 1982. Cattle herd dynamics: an Integer
stochastic model for evaluating production alternatives. ILCA Research Report No. 2.
Addis Ababa, ILCA.
Lampkin, K. & Lampkin, G.H. 1960. Studies on the production of beef from zebu
cattle in East Africa (II): Milk production in suckled cows and its effect on calf growth.
J. Agric. Sci., 55: 233-239.
Ludwig, J.A. & Tongway, D.J. 1995. Spatial organisation of landscapes and its
function in semi-arid woodland, Australia. Landscape Ecol., 10(1): 51-63.
Malechek, J.C. 1984. Impacts of grazing intensity and specialized grazing systems
on livestock response. In National Research Council, National Academy of Sciences,
ed. Developing strategies for rangeland management. Boulder, Col., USA and London,
Westview.
Meade, R.H. 1982. Sources, sinks and storage of river sediment in the Atlantic
drainage of the United States. J. Geol., 90: 235-252.
Mentis, M.T. 1977. Stocking rates and carrying capacities for ungulates on African
rangelands. S. Afr. J. Wild. Res., 7: 89-98.
O'Connor, T.G. 1985. A synthesis of field experiments concerning the grass layer in
the savanna regions of southern Africa. A report for the Committee for Terrestrial
Ecosystems, National Program for Ecosystem Research. South African National Scientific
Programmes Report No. 114.
Payne, W.J.A. 1965. Specific problems of semi-arid environments. Qual. Plant.,
12(3): 269-294.
Pearce, D.W. & Warford, J.J. 1993. World without end. Economics, environment
and sustainable development. New York, World Bank/Oxford University Press.
Penning, de Vries, F.W.T. & Djiteye, M.A. 1982. The productivity of Sahelian
rangelands: a study of soils, vegetation and land use in the region. Wageningen, the
Netherlands, Puooc.
Peters, P.E. 1984. Struggles over water, struggles over meaning: cattle, water and
the state in Botswana. Africa, 54(3): 29-49, 127.
Pickup, G. & Chewings, V.H. 1986. Random field modelling of spatial variations
in erosion and deposition in flat alluvial landscapes in arid central Australia. Ecol.
Modelling, 33:269-296.
Richardson, F.D. 1994. Models for the selection of cow types for extensive meat
and milk production in developing areas (III): Cow types and biological efficiency of
production. Department of Applied Mathematics, University of Cape Town, South Africa.
(mimeo)
Sandford, S. 1983. Management of pastoral development in the Third World.
Chichester, UK, Wiley.
Sandford, S. 1994. Improving the efficiency of opportunism: new directions for
pastoral development. In I. Scoones, ed. Living with uncertainty. New directions
in pastoral development in Africa, p. 174-182. London, ITP.
Schlee, G. 1991. Traditional pastoralists: land use strategies. In Range
management handbook of Kenya. Volume II. 1. Marsabit District. Nairobi, Ministry of
Agriculture and Livestock Development.
Shackleton, C.M. 1993. Are the communal grazing lands in need of saving? South.
Afr. Dev., 10(1): 65-78.
Sikana, P.M. & Kerven, C.K. 1991. The impact of commercialisation on the
role of labour in African pastoral societies. Pastoral Development Network Paper 31c.
London, ODI.
Sikana, P.M., Kerven, C.K. & Behnke, R.H. 1993. From subsistence to
specialised commodity production: commercialisation and pastoral dairying in Africa.
Pastoral Development Network Paper 34d. London, ODI.
Spedding, C.R.W. 1971. Grassland ecology. Oxford, UK, Oxford University
Press.
Stafford-Smith, M. & Pickup, G. 1993. Out of Africa, looking in: understanding
vegetation change. In R.H. Behnke, I. Scoones & C. Kerven, eds. Range
ecology at disequilibrium. New models of natural variability and pastoral adaptation in
African savannas, p. 196-226. London, ODI/IIED/Commonwealth Secretariat.
Stocking, M.A. 1987. Measuring land degradation. In P.M. Blaikie & H.
Brookfield, eds. Land degradation and society, p. 196-226. London, Methuen.
Tapson, D.R. 1990. A socio-economic analysis of small-holder cattle producers in
Kwazulu. Vista University, Pretoria, South Africa. (Ph.D. thesis)
Tapson, D. 1993. Biological sustainability in pastoral systems: The KwaZulu case. In
R.H. Behnke, I. Scoones & C. Kerven, eds. Range ecology at disequilibrium. New
models of natural variability and pastoral adaptation in African savannas, p. 118-135.
London, ODI/IIED/Commonwealth Secretariat.
Tawonezvi, H.P.R., Ward, H.K., Trail, J.C.M. & Light, D. 1988. Evaluation of
beef breeds for rangeland weaner production in Zimbabwe (I): Productivity of purebred
cows. Anim. Prod., 47: 351-360.
Trimble, S.W. 1975. A volumetric estimate of man-induced soil erosion on the
southern Piedmont Plateau. In Prospective technology for predicting sediment yields and
sources, p. 142-152. ARS-S-40. Washington, DC, USDA.
Trimble, S.W. 1983. A sediment budget for Coon Creek Basin in the Driftless Area,
Wisconsin, 1853-1977. Am. J. Sci., 283: 454-474.
Walker, B.H. & Noy-Meir, I. 1982. Aspects of the stability and resilience of
savanna ecosystems. In B.J. Huntley & B.H. Walker, eds. Ecology of tropical
savannas, p. 577-590. Berlin, Germany, Springer.
Wasson, R. 1987. Detection and measurement of land degradation processes. In
A. Chisholm & R. Dumsday, eds. Land degradation. Problems and policies, p.
49-69. Cambridge, UK, Cambridge University Press.
Western, D. 1982. The environment and ecology of pastoralists in arid savannas. Dev.
Change, 13: 183-211.
Western, D. & Finch, V. 1986. Cattle and pastoralism: survival and production
in arid lands. Human Ecol., 14(1): 77-94.
Westoby, M., Walker, B. & Noy-Meir, I. 1989. Opportunistic management for the
rangelands not at equilibrium. J. Range Manage., 42(4): 266-274.
Wheeler, J.L. & Freer, M. 1986. Pasture and forage: the feed base for postoral
industries. In G. Alexander & O.B. Williams, eds. The pastoral industries of
Australia, p. 165-182. Sydney, NSW, Australia, Sydney University Press.
White, J.M. & Meadows, S.J. 1981. Evaluation of the Contribution of Group
and Individual Ranches in Kajiado District, Kenya, to Economic Development and Pastoral
Production Strategies. Nairobi, Ministry of Livestock Development.
Williams, M., McCarthy, M. & Pickup, G. 1995. Desertification, drought and
landcare: Australia's role in an international convention to combat desertification. Aust.
Geograph., 26(1): 23-32.
Wilson, A.D. & Macleod, N.D. 1991. Overgrazing: present or absent? J. Range
Manage., 44(5): 475-482.
Workman, J.P. 1986. Range economics. New York, Macmillan.