This is the old United Nations University website. Visit the new site at http://unu.edu


Contents - Previous - Next


6. Contributed papers

Physical Aspects of Desertification

The first three papers in this session dealt with the climatic factor in desertification, and the fourth with a regional example of the impact of sand encroachment on oasis settlements.

A paper by A.T. Grove of Cambridge University, England, on Lakes as Indicators of Changing Climates in the Arid Zone proceeded, from a summary made with Alayne Street in 1976 of world evidence, to produce global maps of lakelevel fluctuations over the last 30,000 years. Evidence is still lacking from large areas of South America and Asia and is sparse for Africa south of the equator.

Uncertainty remains about the climatic implications of former high lake stands, although estimates of former precipitation can be made by making assumptions about former evaporation and run-off coefficients. The limits of fixed dunes, beyond those of present-day mobile sand, also provide useful indicators of past rainfall in more arid periods. The evidence is that mean annual rainfall in the last 20,000 years has fluctuated through a range comparable with the absolute interannual variation within the past century. Climatic fluctuations on this long time-scale have a bearing on desertification through the ages of fossil reservoirs of groundwater.

Judging from lake-level data, the last 1,000 years appear to have been generally dry compared with the late Quaternary record. Within the last century, and even the last few decades, lake-level fluctuations in closed basins have been marked in the arid zone.

Grove drew attention to the potential of LANDSAT imagery for monitoring lake areas as a guide to fluctuations in level, and the need to relate these to current meteorological data with a view to improving the palaeoclimatic interpretation of old lake shorelines.

In a paper on Above-Normal Rainfalls and Drought in Southern Africa, Peter Tyson of the University of Witwatersrand, South Africa, noted the economic significance of recurrent drought in the dry lands and its importance as a potential factor in desertification. He noted that rainfalls in these areas are usually erratic in spatial distribution and in time. Large inter-annual differences occur, with some tendency to quasi-biennial (two to three year-) rhythms, Overriding these are longer periods of persistent drought or abundant rainfall that do not occur with any firmly established frequency. Over large areas of southern Africa, however, the evidence for a quasi-twenty year recurrence is compelling, particularly in the summer rainfall regions of South Africa, Swaziland, Lesotho, Botswana and Namibia.

By submitting regionally averaged data for 62 stations in northeast South Africa to a five-term binomial filter and then to Fourier analysis, it has been possible to define the extended wet and dry periods and extend them back in time. By taking averages over the suggested wet and dry periods it is also possible to assess the extent to which the quasi-twenty year oscillation has persisted in time and manifested itself in space. Despite the scarcity of early records, the evidence suggests that the oscillation is prevalent in the data from the beginning of climatic recording in 1841. Alternating wet and dry spells continued up to 1905, with the period 1888 to 1896 being the wettest complete spell on record. The regular rhythm of change faltered just after the turn of the century, when the decade 1906 to 1915 was dry like that preceding it. However, the rhythm was immediately re-established and has continued uninterrupted from 1906 to the present. Since the turn of the century, the most consistently dry spell has been that of 1944 to 1953, when the regionally-average rainfall for every year was below normal, and the driest spell on average was that of 1963 to 1972. The most persistently wet spell has been the present one, with five consecutive years with above-average rainfall. In general, the dry spells have been more persistently dry than the wet spells have been wet, and in addition they have had a great aerial extent and homogeneity. With each successive wet spell since the turn of the century, the aerial extent of the excess-rain areas has increased. Departures for the latest incomplete wet spell are much higher than those for the 1888 to 1896 wet period.

The models suggest that the current run of wet years should continue with diminishing totals until 1982, when it is likely that a dry spell will succeed until about 1992. Tyson stressed that the model applies to regions and not to individual stations, that it does not indicate rainfall amplitudes, and that, on the average, two or three anomalous years occur in each hemicycle. The model is appropriate for strategic regional planning rather than for season forecast, but it remains a useful approach in the absence of a reliable predictive physicomathematical model for the general circulation of the atmosphere.

David Sharon of the Hebrew University of Jerusalem, Israel, reported on Studies of the Spatial Structure of Rainfall in Southern Israel, based on the identification of areas receiving rainfall within a generalized single event, using correlation analysis of daily rainfalls at adjacent stations. This is a useful tool where radar coverage is lacking, as so often in desert regions.

Regional variations in the size of rainfall areas have been identified through the correlation functions. Thus, rainfall areas are found to decrease rapidly southwards towards the fringe of the Negev Desert, but no such variation occurs in the extent of rainfall fields in the eastward transition to the fringe of the Syrian Desert. This indicates a structural difference between the rain-bearing systems in the south and east, as well as different atmospheric desertification mechanisms in the respective areas.

Studies based on a mesonet of rain recorders in northern Negev, combined with synoptic information and radar observations, show that typical winter storms in this area tend to break up into smaller subsystems that move southwards, away from the core of the main system that generally lies mostly north of the area considered. The storms then move eastwards or northeastwards into Jordan without change in dimension over large distances. This explains the difference in the size of rainfall areas mentioned above. The study also identified the occurrence in the Negev of unorganized local storms, mainly in the autumn and spring, that yield a very patchy rainfall. The high frequency of localized showers in the Negev at all seasons has been independently inferred also from the high frequency of convective rainfall intensities.

The results obtained for the northern Negev furnish a basis for sound interpretation of correlation functions in this part of the Middle East. This may make it possible to extend the analysis over the much wider region including southern Jordan, the entire Negev, Sinai and probably also parts of Egypt.

The spottiness of rainfall and the high rainfall intensities associated with it, typical of arid lands in general, have important implications for the hydrology and ecology of these regions. However, specific regional or seasonal rainfall regimes may differ in terms of the parameters mentioned above. Sharon recommended that comparative studies of these characteristics of desert rainfalls should be made in other parts of the world where such information is available.

In a paper on Problems of Protection of the Oases of Southern Morocco against Sand Encroachment, Monique Mainguet reported on a study carried out in the oases of the wadis Draa and Ziz at the request of FAO. The problem was approached in three ways: by mapping global wind dynamics for the area; by mapping areas of erosion, transport and deposition of sand; and by identifying for each of 20 oases the direction of the most serious threat of sand invasion, in order to assist in the design of shelter belts. Although the sand probably originates from tributary wadis in the north, satellite imagery showed that the wadis being threatened occur along an arc of predominant northeasterly sand drift under the influence of southwest winds, where these give place to strong winds from the northeast and that development and enlargement of the oases in the past have increased the threat of interception of moving sand along their western boundaries. Comparison of aerial photographs over a period of 14 years showed a significant increase in sand accumulation and the development of new barkhans, involving some displacement eastwards of the cultivated area.

Dr. Mainguet agreed that a major factor in the development of the problem was a growing lack of water for irrigation, partly due to the lowering of the watertable through the use of water pumps and partly as a result of a succession of dry Years in the last decade or so. This had led to some abandonment of irrigated land, with a resulting increase in sand encroachment.

Social and Regional Aspects of Desertification

A theme linking these papers, two on the Great Plains of the United States, two on Africa and one on the arid zone of northwestern India, was their focus on the importance of human adjustment to the dryland environment, and on the ways in which desertification appeared to have arisen from human intervention in the environmental balance.

Albert J. Larson, of the Chicago Circle Campus of the University of Illinois, traced the ebb and flow of settlement across the Great Plains of the USA as revealed by plotting the centres of gravity of population for Nebraska in successive decades for the period 1870 to 1970. Reversals of the general westward movement occurred as droughts strengthened the perception of "The Great American Desert," or in times of economic depression. The author claimed that this was a means of illustrating the effects of push-and-pull factors, whether physical, economic, social or political, both at the regional and local levels. A similar pattern of advance and retreat from the Plains was seen to characterize the other Plains States.

M.M. Khogali of the University of Khartoum, in a paper on Pastoral Nomadism and Desertification in the Republic of Sudan, noted that desertification was apparent in all ecological zones of the Sudan. In the north it consists mainly of the invasion of irrigated lands by moving sand, a centuriesold problem. It is in the semi-desert zone that widespread sand drifting gives the impression of a broad southward desert encroachment, claimed locally to be as much as 5 km per year. In the savanna zone, desertification takes the form of more localized degradation of pastures, particularly around watering points, and diminished crop yields. Both pastoral and agricultural land use are associated with desertification. Despite the traditional rotational grazing, increase in human and livestock populations has led to serious overstocking. Too-early stocking of the ephemeral pastures of the desert margins has prevented seedsetting and diminished the grazing resource, leading to increased sand movement. The dry-season dammering pastures have also deteriorated through selective grazing, with loss of valuable species, both in the semi-desert and short savanna zones. However, pastoral nomadism is not the only cause of desertification in the Sudan. Desertification in dryfarming areas has resulted from cleantillage, harvesting of ground nuts and a shortening of cropfallow rotations. There has also been a decrease in tree cover in the same areas, due to the demand for fuelwood and charcoal. Problems are likely to intensify with the inevitable sedentarization of nomads and general population increase. Control over livestock numbers and grazing will be necessary to maintain a rational rotational grazing, perennial crops should be encouraged in farming systems, and local population pressure should be eased through assisted outmigration.

J. Dresch, of the University of Paris VII, spoke on the impact of the introduction of modern technology in comparing desertification on the north and south borders of the Sahara. It is areas on the margins of the arid regions, rather than the arid and hyperarid zones, that are vulnerable to desertification, with the exception of the oases. Desertification is more advanced where there is the longest record of human occupation, where population numbers are greatest, where land is exploited at the same time by pastoralists and cultivators, and where towns have grown. However, it is the introduction of modern methods of production that has proved particularly disastrous and accounted for a rapid increase in the rate of desertification in the last few decades.

The northern and southern borders of the Sahara show very different records. In the south, particularly in the Sahel, desertification has occurred mainly since the colonial era, and within traditional land-use systems. Major factors have been the increase in human and livestock populations; social changes such as the settlement of nomads and their increasing engagement in agriculture; extension of cultivation, particularly of cash crops, with a shortening of fallow in rotations; the provision of perennial watering points in the absence of effective control of livestock numbers, leading to localized overgrazing; and an increasing demand for fuelwood from growing settlements. North of the Sahara, desertification and impoverishment of the vegetation have a long history; traditional land-use systems were already modified under the colonial regimes and degradation is in general more advanced. Major features here have been the extension of mechanized agriculture, often into climatically unstable areas, consequent accelerated erosion and the breakdown of mutual arrangements with nomadic or transhumant pastoral systems, and adverse environmental impacts of the growing towns. In both areas, the conservational measures proposed by the Nairobi Desertification Conference will succeed only if they are understood, accepted and practiced by all those populations affected.

S.P. Malhotra and H.S. Mann of the Central Arid Zone Research Institute at Jodhpur, India, in a paper on The Human Factor and Desertification in the Arid Zone of Rajasthan, described the impact of technology on the traditional social fabric of this region. The nomadic and sedentary populations previously formed an integrated unit providing for a balanced use of the desert ecosystems. This was achieved through functional specialization and inter-relation ships within society that enabled it to spread, both spatially and socially, the risks arising from drought hazard. The development of communications and the supply of relatively cheap manufactured goods led to some disruption of traditional exchange relationships between nomads and sedentary residents and between craftsmen and agriculturists. Improvements in medical and health services have resulted in a rapid growth of population and increasing pressures on the land. Veterinary programmes have allowed an accompanying growth in livestock numbers. These changes have resulted in desertification in the form of falling crop yields, reduction in vegetation cover, including tree cover, overgrazing, particularly by goats, and an increasing demand for fuelwood, all leading to accelerated wind erosion and sand drift. The authors stressed the need for a more integrated approach in the introduction of new technology, and an overall appreciation of the socio-cultural mechanisms affected, with a view to establishing the necessary social and cultural infrastructure for stable, development.

Erwin Kedar of William Paterson College, New Jersey, presented a paper on Desertification Processes on the Eastern Colorado Plains, a study originating from his experience of flying through a dust storm there in February 1977. He noted that the front zone of the storm was not continuous but consisted of scattered wind whirls, each of which was lifting dust from a particular field and raising it to a general ceiling dust cloud at about 2,400 m. He was subsequently able to locate the dust sources on the ground. They were fields of unprotected friable, light-textured soils of the Nunn-Bresser-Ascalon-Truckton associations, which are particularly favoured for pastures and winter wheat. The wind whirls, in the from of mini-tornadoes, occur within lee-wind storms along the Colorado Front Range, where they are triggered by local topographic and microclimatic conditions.

In these areas a rainfall of 500 mm is marginal for cropping, and in drought years the growth of pastures or crops is not vigorous enough to protect the soil. Vulnerability is enhanced by failure to leave crop residues on the land, as in the practice of stubble mulching. There is a long record of dust storms associated with drought in this area, particularly in the "Big Drought" of 1926 to 1931, when dustbowl conditions were widespread. Kedar estimated that the annual soil losses under conditions such as he observed could be as high as 180 tonnes per hectare, or more than five times the permissible rate of soil loss. He recommended more positive action by the US Soil Conservation Service, or local extension services, to identify farmers in these vulnerable areas, to carry out systematic investigations and to offer them advice, or threaten sanctions for their failure to meet their responsibilities as custodians of the soil.

Monitoring Desertification

Two papers discussed the establishment of systems for monitoring desertification based on the application of LANDSAT imagery, with its advantages of multispectral sensing and repetitive coverage.

Dietrich Bannert, of the Federal Institute of Geosciences and Natural Resources (BGR) of the Federal Republic of Germany in Hanover, described a system of monitoring using LANDSAT digital data, developed from two projects in the Sahelian Zone undertaken on behalf of the West German Ministry of Economic Co-operation. Among the indicators of desertification listed by the United Nations Conference on Desertification (UNCOD), the following can be determined from earth-resources satellite data:

- distribution of vegetation and soil cover;
- movement of sand dunes;
- extent of agricultural areas;
- extent of surface-water resources;
- distribution of bare ground.

The progress of desertification can be determined by the distributions and the relationships to each other of these indicators, if continuous observations are made over a period of time and recorded on thematic maps. Bannert proposed a system consisting of:

- a data-collection system, using LANDSAT;
- regional receiving centres, or in the interim the use of transportable receivers;
- data-processing facilities, preferably through contract with regional data centres;
- systems of automated rapid reproduction of multicoloured thematic maps at low cost, as developed by BGR.

In his paper on LANDSAT Applied to a Desertification Monitoring System, Erwin Kedar stressed the need for a multidisciplinary approach to the problem, and the consequent importance of synoptic monitoring. He proposed the establishment of a ring of test sites surrounding the Sahara, with, in each area, a test site in an inhabited area and a comparable control site which was virtually uninhabited. Ground truth would be based on 100 control cells in each test site, in each of which measurements would be made of a range of possible desertification indicators. From the indicators, factors would be established by factorial analysis and a synoptic index of arid environmental quality would be obtained, based on the factor scores of the different desertification indicators. Synoptic maps would be produced on this basis. The LANDSAT imagery would be analysed against the ground data to see which spectral band or combination of bands should be used to predict environmental quality. Computer simulation studies would be carried out, based on the relationships between indicators, in order to explore possible processes whereby past and present desertification had occurred, and the verified model would then be translated into a statistical model of desertification risk which could be used for synoptic forecasts of future desertification. Repetitive LANDSAT imagery would then be used to produce successive synoptic maps showing continuous change in environmental quality and desertification risk at each test site. This would assist the provision of economic assistance and the formulation of measures to combat desertification.


Contents - Previous - Next