Tải bản đầy đủ - 0trang
Chapter 1. Desertification and It's Relation to Climate Variability and Change
HILLEL AND ROSENZWEIG
Ecosystems in semiarid and arid regions around the world appear to be undergoing various processes of degradation commonly described as desertiﬁcation.
According to UNEP (1992), all regions in which the ratio of total annual precipitation to potential evapotranspiration (P/ET) ranges from 0.05 to 0.65 should be
considered vulnerable to desertiﬁcation. Such regions constitute some 40% of the
global terrestrial area, which totals about 130 million km2 (13 billion ha). Dregne
(1983) calculated that the arid, semiarid, and dry subhumid regions of the world
occupy 12.1, 17.1, and 9.9% of the world’s total land area. Relatively dry areas
cover much of northern Africa, southwestern Africa, southwestern Asia, central
Asia, northwestern India and Pakistan, southwestern United States and Mexico,
western South America, and much of Australia (Fig. 1, see color insert).
Arid and semiarid regions cover over a fourth of the world’s land area, and
are home to nearly one-sixth of the world’s population (WRI, 2000). The total
population of the world has doubled in the last four decades, resulting in the
current total of about 6 billion. As of 1998, some 80% of humanity resided in the
so-called developing countries, which contain only 58% of the total land area and
54% of the total cropped area. Moreover, many of the developing countries are
located in semiarid regions that are most vulnerable to degradation.
According to a report published by the World Resources Institute (WRI, 1998),
the total area of land under cropping has increased by some 25% since 1950. In
the same period, the world’s population has more than doubled, so the area of
cropland per capita has been reduced by nearly a half.
At present, the annual growth rate of cropland (0.2%) is only one-seventh the
growth in population (Lal, 1997), so the decline in arable land per capita is continuing. That decline is most severe in the developing countries, which are expected
to increase their populations most rapidly and will therefore be most in need of
increased food production. In sub-Saharan Africa, for instance, the per capita area
of arable land, which was 1.6 ha in 1990, is projected to fall to 0.63 ha by 2025
(Scherr, 1999). The lands still available for the expansion of farming are, in large
part, marginal lands of relatively low productivity and high vulnerability.
Desertiﬁcation is an emotive term, conjuring up the specter of a tide of sand
swallowing fertile farmland and pastures. The United Nations Environmental Programme (UNEP) sponsored projects in the early 1980s to plant trees along the
edge of the Sahara, with the aim of warding off the invading sands. While there are
places where the edge of the desert can be seen encroaching on fertile land, the more
pressing problem is the deterioration of the land due to human abuse in regions
well outside the desert. The latter problem emanates not only from the desert but
also from the centers of population; not only from the spread of the sand dunes
but also from the spread of people and their mismanagement of the land (Hillel,
1992). Therefore, protecting the front line may do nothing to halt the degradation
DESERTIFICATION: CLIMATE VARIABILITY AND CHANGE
behind it. The true challenge is not so much to stop the desert at the edge of a
semiarid region as to protect the entire region from internal abuse of its vegetation
and of its soil and water resources.
A vicious cycle is already operating in many areas: as the land degrades, it is
worked ever more intensively so its degradation accelerates; and as the returns
from “old” land diminish, “new” land is brought under cultivation or grazed by
encroachment onto marginal or submarginal areas. But attempts to encapsulate
these complex problems in the catchall term “desertiﬁcation” may have obscured
its true character and confused the search for its amelioration.
In this paper, we review the concepts, deﬁnitions, and processes pertinent to
desertiﬁcation, and offer an alternative, more inclusive term, namely, “semi-arid
ecosystem degradation.” We use the long-term drought in the Sahelian region of
Africa as a case study for analyzing the complex set of climatic, biophysical,
and social factors that interweave to create the process of semiarid ecosystem
degradation, and we evaluate current monitoring techniques, including remote
sensing. We next consider the potentialities and hazards of irrigation development
as a possible means to improve agricultural production in semiarid regions. We then
ask the question, “How might global climate change affect the Sahelian region of
Africa?” and analyze a set of recent projections derived from global climate change
scenarios, in light of the region’s vulnerabilities. Finally, we offer our views on
prospects for sustaining semiarid ecosystems and agroecosystems in the future.
II. CONCEPTS AND DEFINITIONS
Desertiﬁcation is a single word used to cover a wide variety of effects involving the actual and potential biological productivity of ecosystems in semiarid and
arid regions. The term desertiﬁcation (or desertization) was apparently coined by
the French ecologist LeHouerou (1977) to characterize what was perceived to
be a northward advance of the Sahara in Tunisia and Algeria. It gained currency
following the severe drought that afﬂicted the Sud region of Africa in the early
1970s, and again in the 1980s, during which the Sahara was reported to be advancing southward into the Sahelian zone as well. For example, Lamprey (1975)
estimated that during the period from 1958 to 1975, while mean annual rainfall
diminished by nearly 50%, the boundary between the Sahara and the Sahel had
shifted southward by nearly 100 km.
As deﬁned in recent dictionaries, desertiﬁcation is the process by which an area
becomes (or is made to become) desert-like. The word “desert” itself is derived
from the Latin desertus, being the past participle of deserere, meaning to desert, to
abandon. The clear implication is that a desert is an area too barren and desolate to
support human life. An area that was not originally desert may come to resemble
a desert if it loses so much of its formerly usable resources that it can no longer
HILLEL AND ROSENZWEIG
provide adequate subsistence to humans. This is a very qualitative deﬁnition, since
not all deserts are the same. An area’s resemblance to a desert does not make it
a permanent desert if it can recover from its damaged state, and, in any case, the
modes of human subsistence and levels of consumption differ greatly from place
The United Nations Conference on Desertiﬁcation (UNCOD) was held in
Nairobi in 1977. It was convened in response to the severe drought that had befallen
the Sahel from the late 1960s through most of the 1970s. Its report deﬁned desertiﬁcation as “the diminution or destruction of the biological potential of land that
can lead ultimately to desert-like conditions . . . under the combined pressure of
adverse and ﬂuctuating climate and excessive exploitation.” That statement leaves
open several questions, such as the deﬁnition of the land’s “biological potential,”
the type and degree of damage to the land that can be considered “destruction,”
and the exact meaning of “desert-like” conditions.
Mainguet (1994) characterized desertiﬁcation as the “ultimate step of land
degradation to irreversible sterile land.” This deﬁnition ignores the complex set
of processes that progress gradually (and, for a time, reversibly) at different rates.
Rather, it conﬁnes the term to the ﬁnal condition that is the extreme culmination of
those various processes. An alternative approach would be to deﬁne the processes
themselves and characterize the degree of degradation at which their separate or
combined effects may be considered to have become irreversible.
In recent years, the very term desertiﬁcation has been called into question as
being too vague, and the processes it purports to describe too ill-deﬁned. Some
critics have even suggested abandoning the term, in favor of what they consider
to be a more precisely deﬁnable term, namely, “land degradation” (e.g., Dregne,
1994). However, desertiﬁcation has already entered into such common usage that it
can no longer be recalled or ignored (Glantz and Orlovsky, 1983). It must therefore
be clariﬁed and qualiﬁed so that its usage may be less ambiguous.
The United Nations has since modiﬁed its deﬁnition of desertiﬁcation as
follows: “Land degradation in arid, semiarid, and dry subhumid areas resulting
from various factors, including climate variations and human activities” (Warren,
1996). That deﬁnition still does not either clarify the relative importance of the
two potential causes or imply the possibility that they may be interactive. It merely
shifts the issue to the deﬁnition of “land degradation.” Does the latter pertain
to the soil, and, if so, to just what qualities or attributes of the soil (physical,
chemical, and/or biological)? Does it also pertain to the vegetation present on
the land, and, if so, to what attributes of the vegetation (biomass, photosynthesis,
respiration, transpiration, growth rate, ground coverage, species diversity, etc.)?
And what of the animal life associated with the land?
“Land degradation” itself is a vague term, since the land may be degraded with
respect to one function and not necessarily with respect to another. For example, a tract of land may continue to function hydrologically—to regulate inﬁltration, runoff generation, and groundwater recharge—even if its vegetative cover is
DESERTIFICATION: CLIMATE VARIABILITY AND CHANGE
changed artiﬁcially from a natural species-diverse community to a monoculture,
and its other ecological functions may be interrupted.
Rather than “land degradation,” we prefer the term “semiarid ecosystem degradation.” A semiarid ecosystem encompasses the diverse biotic community living
in this given domain. Included in this community is the host of plants, animals,
and microorganisms that share the habitat and that interact with one another
through such modes as competition or symbiosis, predation, and parasitism. It
also includes the complex physical and chemical factors that condition the lives
of those organisms and are in turn inﬂuenced by them. A semiarid ecosystem may
be a more or less natural one, relatively undisturbed by humans, or it may be an
artiﬁcially managed one, such as an agroecosystem.
Each ecosystem performs a multiplicity of ecological functions. Included among
these are photosynthesis, absorption of atmospheric carbon and its incorporation
into biomass and the soil, emission of oxygen, regulation of temperature and the
water cycle, as well as the decomposition of waste products and their transmutation
into nutrients for the perpetuation of diverse interdependent forms of life. Integrated
ecosystems may thus play a vital role in controlling global warming and in absorbing and neutralizing pollutants that might otherwise accumulate to toxic levels.
An agroecosystem is a portion of the landscape that is managed for the economic
purpose of agricultural production. The transformation of a natural ecosystem into
an agroecosystem is not necessarily destructive, if the latter is indeed managed sustainably and if it coexists harmoniously alongside natural ecosystems that continue
to maintain biodiversity and to perform vital ecological functions.
In too many cases, however, the requirements of sustainability fail, especially
where agricultural systems expand progressively at the expense of the remaining
more or less natural ecosystems. The appropriation of ever-greater sections of the
remaining native habitats, impelled by the increase of population as well as by
the degradation of farmed or grazed lands due to overcultivation or overgrazing,
decimates those habitats and imperils their ecological functions.
In the initial stages of degradation, the deteriorating productivity of an agroecosystem can be masked by increasing the inputs of fertilizers, pesticides, water, and
tillage. Sooner or later, however, if such destructive effects as organic matter loss,
erosion, leaching of nutrients and salination continue, the degradation is likely to
reach a point at which its effects are difﬁcult to overcome either ecologically or
Key processes related to desertiﬁcation include drought, primary production and
carrying capacity, soil degradation, and water resources. The role of social factors
is also important.
HILLEL AND ROSENZWEIG
A typical feature of arid regions is that the mode (the most probable) amount
of annual rainfall is generally less than the mean; i.e., there tend to be more years
with a below-average rainfall than years in which the rainfall is above average,
simply because a few unusually rainy years can skew the statistical average well
above realistic expectations for rainfall in most years. More than 90% of the total
variation in annual rainfall can generally be encompassed within a range between
one-half and twice the mean.
The variability in biologically effective rainfall is yet more pronounced, as
years with less rain are usually characterized by greater evaporative demand, so
the moisture deﬁcit is greater than that indicated by the reduction of rainfall alone.
Timing and distribution of rainfall also play crucial roles. Below-average rainfall,
if well distributed, may produce adequate crop yields, whereas average or even
above-average rainfall may fail to produce adequate yields if the rain occurs as
just a few large storms with long dry periods between them.
In semiarid agricultural regions, “drought,” like desertiﬁcation, is a broad, somewhat subjective term that designates years in which cultivation becomes an unproductive activity, crops fail, and the productivity of pastures is signiﬁcantly
diminished. Drought is a constant menace, a fact of life with which rural dwellers
in arid regions must cope if they are to survive. The occurrence of drought is a
certainty, sooner or later; only its timing, duration, and severity are ever in doubt.
It is during a drought that ecosystem degradation in the form of devegetation and
soil erosion occurs at an accelerated pace.
Any management system that ignores the certainty of drought and fails to provide
for it ahead of time is doomed to fail in the long run. That provision may take the
form of grain or feed storage (as in the Biblical story of Joseph in Egypt), or
of pasture tracts kept in reserve for grazing when the regular pasture is played
out, or of the controlled migration of people and animals to other regions able to
accommodate them for the period of the drought.
There has been a prolonged period of drought in the Sahelian region of Africa
since the early 1970s (Fig. 2). Various hypotheses involving both natural and
Figure 2 Rainfall ﬂuctuations 1901–1998, expressed as a regionally averaged standard deviation
(departure from the long-term mean divided by the standard deviation) for the Sahel. (Source: IPCC
WG II, 2001).
DESERTIFICATION: CLIMATE VARIABILITY AND CHANGE
anthropogenic factors have been advanced to explain the persistence of this
1. Atmospheric Dust
One hypothesis is that the recent droughts are due to a cooling of the land
masses of the Northern Hemisphere by about 0.3◦ C between 1945 and the early
1970s, owing to an increase in atmospheric dust from drylands, as well as from
air pollution and volcanic eruptions. The cooling may have changed the patterns
of air mass movement (Tegen et al., 1996). Evidence in support of this hypothesis
seems to be contradicted by the heavy rains that occurred in the Sahel during the
1950s when the Northern Hemisphere cooled, and by the severe Sahel drought
that occurred during the early 1980s when the Northern Hemisphere experienced
2. Ocean–Atmosphere Dynamics
Another hypothesis links drought in the Sahel to changes in ocean–atmosphere
dynamics, speciﬁcally changes in sea–surface temperatures (SSTs) in the world’s
oceans. Such changes might tend to reduce the northward penetration of the
Intertropical Convergence Zone (ITCZ)—the great band of equatorial clouds
whose shifting pattern brings monsoonal rain to the humid tropics as well as
to the Sahel (Nicholson, 1986). Many studies have linked interannual variation of
SSTs and seasonal precipitation variability in the region (e.g., Druyan, 1987; 1989;
Folland et al., 1986; Lough, 1986; Rowell et al., 1995). Droughts in the Sahel tend
to be coincident with positive SST anomalies in Southern Hemisphere oceans and
the Indian Ocean, and negative SSTs in the Northern Hemisphere oceans, especially the subtropical North Atlantic Ocean. Abundant rain in the Sahel is often,
but not always, linked with SSTs of the opposite sign in the Atlantic and other
oceans (Lamb and Peppler, 1991, 1992). The interhemispheric SST gradient in the
Atlantic Ocean appears to be a key mechanism for precipitation in the Sahelian
latitudes (Fontaine and Janicot, 1996; Ward, 1998).
Warmer than normal SSTs in the tropical Paciﬁc related to the El Ni˜no/Southern
Oscillation (ENSO) phenomenon have similarly been linked with droughts in
Australasia, India, South America, and Southern Africa, though these droughts
typically do not persist for more than one or two seasons. The Sahelian region
of Africa, on the other hand, has had many dry years that are not correlated
with Paciﬁc SSTs, so the persistence of the Sahelian drought sets it apart from
droughts in other parts of the world. There does appear to be some ENSO-driven
teleconnection to drought in West Africa (e.g., Fontaine and Janicot, 1996), but
Janicot et al. (1996) show that the strength of the correlation of Sahel rainfall with
the Southern Oscillation Index (SOI) is quite variable. Hunt (2000) proposes a
mechanism by which tropical Paciﬁc SSTs inﬂuence Sahel rainfall by modulating
HILLEL AND ROSENZWEIG
the North Atlantic Oscillation (NAO) via the Paciﬁc–North America oscillation. Druyan and Hall (1996) suggest that extreme Paciﬁc Ocean SST anomalies inﬂuence climate variability in the Sahel through wave disturbances of the
tropical easterly jet, with associated effects on convergence, humidity, and precipitation. These and other ocean–atmosphere relationships are being used to forecast
seasonal rainfall in the region (Nnaji, 2001; Ward, 1998).
3. Land–Surface Change
Still another hypothesis is that droughts can be caused or worsened by largescale changes in the land surface of Africa, and speciﬁcally by the deforestation and
overall denudation of the land (Charney, 1975; Sud and Molod, 1988). A process
may thus have started whereby the drought can become self-reinforcing. According
to the theory of “biophysical feedback,” losses of vegetative cover resulting from
the drought as well as from overcultivation, overgrazing, and deforestation, along
with the consequent increase of the dust content of the air, combine to enhance
the area’s reﬂectivity to incoming sunlight. That reﬂectivity, called “albedo,” may
rise from about 25% for a well-vegetated area to perhaps 35% or more for bare,
bright, sandy soil. As a larger proportion of the incoming sunlight is reﬂected
skyward rather than absorbed, the surface becomes cooler, and so the air in contact
with the surface has less tendency to rise and condense its moisture so as to yield
An additional effect of denudation is to decrease interception of rainfall by
vegetation and inﬁltration, while increasing surface runoff, thereby reducing the
amount of soil moisture available for evapotranspiration. Crops and grasses, which
have shallower roots than trees and in any case transpire less than the natural mixed
vegetation of the savanna, transpire even less when deprived of moisture during a
drought. The meteorological consequences of such changes have been explored in
modeling studies (Xue and Shukla, 1993). The hypothesis is that such changes may
have some effect on regional precipitation, since in many continental areas rainfall
is derived in signiﬁcant part from water evaporated regionally. It proposes that the
biophysical and physical processes interact, as lower rainfall leads in turn to more
overgrazing, less regrowth of biomass, and further reduction in reevaporated rain
owing to the decline in soil moisture. Thus, the feedback hypothesis offers its own
explanation as to why the drought in the Sahel has tended to persist for so long.
There is still no conclusive evidence, however, that even large-scale changes in
land surface conditions do actually affect regional-scale climate (Nicholson et al.,
1998; Nicholson, 2000).
Key components in semiarid ecosystem degradation processes are increased
surface albedo (the reﬂectance of solar radiation) and increased generation of
dust, both of which are consequences of the exposure of bare, dry ground following
removal of the original vegetative cover.
DESERTIFICATION: CLIMATE VARIABILITY AND CHANGE
The albedo of a bare soil depends on the organic matter content and the mineral
composition of the topsoil. It also depends on the moisture content of the soil
surface. A moist soil is generally less reﬂective (i.e., “darker”) than a dry soil
(Hillel, 1998). Thus, Nicholson et al. (1998) found that near the southern edge
of the African Sahel (at a latitude of 15 degrees north), where the rainfall was
450 mm, the albedo was about 30%. However, near the northern boundary of
the Sahel, where the mean annual rainfall was only 200 mm, the surface albedo
was about 43%. Albedo is also affected, to some degree, by the smoothness or
roughness of the surface. Above all, however, it is affected by the vegetative cover
and its above-ground residues.
A widely cited hypothesis, promulgated by Charney (1975), Charney et al.
(1975), and Otterman (1974, 1977, 1981), suggested a feedback mechanism between land use and climate change. Speciﬁcally, they raised the possibility that an
increase in albedo resulting from anthropogenic denudation of the land can in turn
cause a diminution of rainfall. The mechanistic reasoning underlying this hypothesis is that an increase in surface reﬂectivity implies a reduction in the absorption of
solar energy, which entails a reduction in soil surface temperature and a consequent
reduction in sensible heating of the atmospheric layer in contact with the soil.
Proponents of the Charney hypothesis speculated that because a more highly
reﬂective surface should tend to be cooler, it should enhance the subsidence of
warm dry air and hence exacerbate the area’s aridity. This, in turn, reduces the
upward convective rise of warm air that normally results in condensation of vapor
and the formation of clouds. If the rise in albedo occurs over a large enough area,
it might thus reduce the regionally generated rainfall. Hence, so the reasoning
goes, surface denudation—which is the common effect of humans attempting to
survive with their livestock during a drought—is a self-reinforcing process that
exacerbates the very drought that initially induced it. Lare and Nicholson (1994)
imply that if desertiﬁcation (i.e., denudation) is extreme, it could indeed evoke the
sort of feedback originally postulated by Charney.
A striking example of the albedo difference between grazed and ungrazed land
can be seen along the border between the western Negev of Israel and northeastern
Sinai of Egypt. The two contiguous areas of this arid region had been grazed
to the same degree until 1948, after which the newly established State of Israel
restricted grazing on its own side of the border. Consequently, the area within
Israel developed a relatively dense vegetative cover that appears much darker on
aerial and satellite photographs than the neighboring area on the Egyptian side.
According to Otterman (1977, 1981), the protected area of the Negev had an albedo
of 12% in the visible light and 24% in the infrared range, whereas the corresponding
values on the overgrazed Sinai side were as high as 40 and 53%.
Recent studies have shown, however, that the darkening is due not only to the
shrubs and grasses growing in the area but also to a biological crust (consisting of
algae, fungi, and cyanobacteria) that developed on the surface of the sandy soil.
HILLEL AND ROSENZWEIG
A vegetated area, though it appears darker in aerial photographs, may not be
warmer than a bare area, as long as the plants are actively transpiring. The process
of transpiration involves the absorption of latent heat and therefore tends to cool
the foliage. During the dry season, however, many of the indigenous plants curtail
transpiration so that they, along with the area as a whole, may indeed become
warmer than it would be if it were bare of vegetation. Otterman and Tucker (1985)
reported radiometric ground temperatures (evidently made in the summer season)
of about 40◦ C in Sinai and about 45◦ C in the Negev. More recently, Otterman et al.
(2001) reported that measurements made by NOAA satellites have consistently
shown the Negev to be warmer than Sinai by about 4.5◦ C during the generally dry
period of May to October. In contrast, Balling (1988) and Bryant et al. (1990) found
that the surface temperatures on the darker (more densely vegetated) U.S. side of
the Mexican border were 2 to 4◦ cooler than on the overgrazed and lighter-colored
Mexican side. The latter measurements may well have been made during a period
when the vegetation was actively transpiring, and hence produced a cooling effect
despite its lower albedo.
The persistent presence of dust in the atmosphere itself has an effect on an area’s
radiation balance (Fouquart et al., 1987). It tends to scatter and reﬂect a fraction of
the solar (shortwave) radiation, while absorbing longwave radiation emitted from
the Earth. In some cases, a turbid atmosphere may actually warm the air near the
ground, while in other cases it may do the opposite, depending on such variables
as its density as well as its reﬂective or absorptive properties.
Recent studies on the potential effects of aerosols on rainfall have advanced another feedback hypothesis. Denudation of an area’s vegetation is usually associated
with biomass burning, which releases smoke into the air. In addition, denudation
also results in deﬂation of the soil surface by wind erosion, which in turn creates a
“dust bowl” effect. Rosenfeld and Farbstein (1992), Rosenfeld (1999, 2000) and
Rosenfeld et al. (2001) have presented evidence that concentrations of such aerosols in the troposphere can suppress rainfall signiﬁcantly.
The postulated mechanism is that moisture condensed on the dust particles forms
small droplets that do no coalesce sufﬁciently to generate rainfall. The detrimental
impact of dust on rainfall is less than that caused by smoke from biomass burning,
but the abundance of desert dust in the atmosphere renders it important. The
reduction of rainfall affected by desert dust can cause drier soil, which raises still
more dust, thus creating a feedback loop to further reduce rainfall.
B. PRIMARY PRODUCTION AND CARRYING CAPACITY
The biological productivity of any ecosystem is due to its primary producers
(known as autotrophs), which are the green plants growing in it. They alone are
DESERTIFICATION: CLIMATE VARIABILITY AND CHANGE
able to create living matter from inorganic raw materials. They do so by combining
atmospheric carbon dioxide with soil-derived water, thus converting radiant energy
from the Sun into chemical energy in the process of photosynthesis. Green plants
also respire, which is the reverse of photosynthesis, and in so doing they utilize
part of the energy to power their own growth. The net primary production then
becomes available for the myriad of heterotrophs, which subsist by consuming
(directly or indirectly) the products of photosynthesis. A stable ecosystem is one
in which production and consumption, synthesis and decomposition, are in balance
over an extended period of time.
When humans enter into an ecosystem and appropriate some of its products
for themselves, they normally do so in competition with, and at the expense of,
other potential consumers. Historically, in the hunter-gatherer phase of subsistence,
humans merely selected the most readily obtainable and useful (or desirable) plant
and animal products, leaving the remainder more or less intact. As their population
increased, humans began to manage the ecosystem so as to promote the production
of the goods they desired, and to suppress the species that competed for those
products. At a still later stage, humans tended to take over sections of the ecosystem
entirely, aiming to eradicate all species that did not serve them directly, and to
plant (and harvest) only the plants and animals they chose to domesticate. In the
process, the ecosystem’s biodiversity and natural productivity were profoundly
affected (Hillel, 1992).
As long as the tracts dominated by humans consist of small enclaves within a
large and continuous ecological domain, the ecosystem as a whole is not seriously
affected. However, as population grows progressively and human management
becomes both more extensive and more intensive, the ecological integrity of entire
regions is threatened. Especially affected are areas within the semiarid and arid
regions, which, because of the paucity of water and the fragility of the soil (typically
deﬁcient in organic matter, structurally unstable, and highly erodible) are most
vulnerable and least resilient.
The term “carrying capacity” has been used to characterize an area’s productivity
in terms of the number of people or grazing animals it can support per unit area
on a sustainable basis (Cohen, 1995). However, the productive yield obtainable
from an area—and hence the number of people deriving their livelihood from it,
at whatever standard of life—depends on how the area is being used. Under the
hunter-gatherer mode of subsistence, an area may be able to carry only, say, 1 person
per square kilometer, whereas under shifting cultivation it may carry 10, and under
intensive agriculture perhaps 100. The more intensive forms of utilization also
involve inputs of capital, energy, and materials, such as fertilizers and pesticides,
that are brought in from the outside to enhance an area’s productivity. As the usable
productivity is affected by the availability of water (i.e., by seasonal rainfall),
it varies from year to year and from decade to decade, and a long-term average
HILLEL AND ROSENZWEIG
(as well as variability) is difﬁcult to determine, especially given the prospect of
climate change. It is therefore doubtful that any given regions can be assigned an
intrinsic and objectively quantiﬁable “carrying capacity.”
Human pressure on the meager resources of arid ecosystems arises primarily
because of increasing population and the trend toward sedentarization of formerly
nomadic people. What typically follows includes the cutting down of wooded
plants for fuel, overcultivation, and overgrazing by livestock (especially in the
immediate peripheries of water supply centers such as wells, cisterns, or surfacewater impoundments). The denuded and pulverized soil surface then falls prey to
erosion by wind (during the dry season) and by water (during the rainy season).
Wind erosion blows away the fertile topsoil and greatly increases the content of dust
in the atmosphere. Water erosion also scours away the topsoil and often cuts into
the soil to produce deep gullies. During fallow periods, rainfall may also leach
away soluble nutrients. The net result can be an overall reduction in biological
Over a long period of time (say, centuries), and in the absence of human intervention, even a severely eroded soil can recover. However, on the time scale of years
to a few decades, especially if humans continue to overgraze and/or overcultivate
the land, soil erosion may be, in effect, irreversible. One problem is to measure the
productivity of an area and its gradual change from year to year or from decade
to decade. Quite another problem is to assess the recoverability (or resiliency) of
an area following a partial loss of productivity, and the rate of potential recovery,
i.e., the time pattern of gradual restoration of productivity and the period needed
for its completion (Dregne, 1994).
Desertiﬁcation from anthropogenic and climatic factors in Senegal caused a fall
in standing wood biomass of 26 kg C ha− 1 y−1 in the period 1956–1993, releasing
carbon at the rate of 60 kg C cap− 1 y− 1 (Gonzalez, 1997). The signiﬁcance of these
quantities in the global balance may be small, but perhaps important nonetheless
(Bouwman, 1992; Lal, 2001).
C. SOIL DEGRADATION
An important criterion of soil degradation (itself a major component of land and
ecosystem degradation) is the loss of soil organic matter. Compared to soils in more
humid regions, those in arid regions tend to be inherently poor in organic matter
content, owing to the relatively sparse natural vegetative cover and to the rapid
rate of decomposition. The organic matter present is, however, vitally important
to soil productivity. Plant residues over the surface protect the soil from the direct
erosive impact of raindrops and from deﬂation by wind and help to conserve soil
moisture by minimizing evaporation. Plant and animal residues that are partially
decomposed and that are naturally incorporated into the topsoil help to stabilize its