Text%2f2009-06!22!03!00!34hdmpbg45mbmm31buvsyslm55_characteristics Of Desertification And Its Rehabilitation In China

Journal of Arid Environments (1997) 37: 419–432

Characteristics of desertification and its rehabilitation in China

Yong Zha* & Jay Gao†‡ *School of Geographical Sciences, Nanjing Normal University, Nanjing 210097, People’s Republic of China †Department of Geography, University of Auckland, Auckland Private Bag 92019, New Zealand (Received 20 March 1997, accepted 16 June 1997) The definition of desertification and its causes in the Chinese literature are reviewed and compared with those in international publications. Both Chinese researchers and their western counterparts have difficulty in reaching a generally accepted definition for desertification and an agreement upon the exact role played by human activities and environmental settings in desertification initiation and development. Tremendous efforts in China have gone into rehabilitating desertified land into productive uses with great contribution to existing knowledge in reclaiming desertified land. The early biological-oriented measures based solely on economic return have recently been replaced by a much more successful, multi-disciplinary approach of rehabilitation combined with preventive measures that follow sound ecological principles. ©1997 Academic Press Limited Keywords: desertification; causes of desertification; severity assessment; rehabilitation of desertified land; land reclamation; China

Introduction With a territory of 9·6 million km2, China is one of the most severely desertified countries in the world. Desertification is threatening the lives of close to 400 million people and has affected about 3·3 million km2 of land (Chen et al., 1996). It is thus very important to study desertification and rehabilitate desertified land into productive uses. Although sand transport and sand dune movements were studied in the 1960s (Zhu et al., 1964; Wu, 1965), these efforts were highly limited in their scope and quantity. Spurred by the United Nations Conference on Desertification (UNCOD) held in Nairobi, Kenya in 1977, immense research on desertification and its rehabilitation has been carried out with fruitful results. In this paper the characteristics of desertification in China are identified through a review of published papers. The literature cited, with a few exceptions, comes chiefly from journals and books recently published in Chinese. Wherever relevant, the issues under consideration are discussed ‡Corresponding author. 0140–1963/97/030419 + 14 $25.00/0/ae970290

© 1997 Academic Press Limited

420

Y. ZHA & J. GAO

in a wider scientific context through citing articles published internationally in English books and journals. According to UNCOD (1978), desertification refers to the ‘diminution or destruction of the biological potential of the land that can lead ultimately to desert-like conditions.’ Increasingly, it has been agreed that the term ‘desertification’ should be restricted to dryland environments only (Thomas, 1993). Therefore, land degradation in humid regions is beyond the scope of this paper. The various definitions of desertification are first presented, followed by consideration of its spatial extent and the magnitude of the desertification problem. The causes of desertification and its development are discussed next. Efforts to monitor desertification and to rehabilitate desertified land into productive uses are reviewed. Finally, the outcome of the rehabilitating efforts are summarized.

Definition of desertification Coined by the French botanist and ecologist Aubr´eville (1949) nearly half a century ago, the term ‘desertification’ has undergone numerous modifications in its meaning since then. More than one hundred definitions have appeared in the English literature so far (Glantz & Orlovsky, 1983). For instance, Rapp (1974) defined it as ‘the spread of desert-like conditions in arid or semi-arid areas due to man’s influence or to climatic change.’ However, no single definition is generally accepted (Dregne, 1983). Much confusion in the literature has occurred as a result of its unscrupulous use (Thomas & Middleton, 1994) in three aspects: (a) indiscrimination between the process of desertification and its state; (b) non-consensus regarding the geographic regions to which it applies; and (c) its exact causes. Recently, Rhodes (1991) and Thomas (1993) suggest that the concept of desertification be revised in light of renewed scientific advances that have enhanced our understanding of the problem. Namely, natural fluctuation in environment causing long-term detrimental impact must be distinguished from land degradation caused by human actions. The concept of ‘desertification’ was not introduced into the Chinese literature until after the UNCOD in 1977 (Chen et al., 1996). Prior to that, the term tudi shahua (land sandification) was in common use (Dong & Liu, 1993). It refers to the coarsening process of the land surface after fine sandy and nutrient particles are lost to aeolian erosion. Though close to desertification in meaning, it at most forms a stage in the development of desertification (Zhu et al., 1989). Another related term is called fengshahua (aeolian sandification). It refers to the process of forming desert-like landforms by sand outside arid and semi-arid zones (Zhu, 1986). However, Li (1988) argued that this process should be called strictly land degradation. Profoundly affected by its constantly changing international meaning, desertification has been dissimilarly defined by Chinese researchers. Zhu & Liu (1981) referred to it as ‘the process of environmental degradation in non-sandy areas where the fragile ecology is disturbed by excessive human activities’. It was defined by Yang (1987) as a series of climatic and geomorphologic processes in arid, semi-arid, and some semihumid sandy areas under the influence of various conditions at diverse time scales. According to Chen (1991), desertification is the contemporary process of land degradation that is caused mainly by sand in a fragile ecosystem and forms a desert-like landform. It is ‘the process of environmental change that is characterized by sandblasting and forms a desert-like landform in formerly non-sandy areas’ (Dong et al., 1988). Apparently, these definitions differ from one another widely in the process and time scale involved. Lack of agreement in defining desertification originates in part from its confusion with desertization because of inappropriate translation. Referring to desert encroachment in arid and semi-arid areas of non-desert landforms due to improper human

DESERTIFICATION AND REHABILITATION, CHINA

421

activities, desertization was translated as shamohua (desertification), whereas desertification was translated as huangmohua (barrenification) in Chinese. Zhou & Pu (1996) argued that the international definition was by no means perfect and had to be altered to suit desertification peculiarities in China. The term huangmohua should be used in its broadest sense to encompass desert creeping, land degradation in the forms of soil erosion, waterlogging and soil salinization to avoid confusion. Unlike the international ones, these Chinese definitions place a much greater emphasis on the material (sand) that is essential in desertification initiation than on climatic, especially precipitation, variables that are incorporated in the definition implicitly. All sandy deserts and lands are located in northern China that has an arid or semi-arid climate (Fig. 1). The proposed adoption of huangmohua will undoubtedly make the concept of desertification in Chinese closer to its international meaning. Severity of desertification Historically, many parts of China are susceptible to desertification. All of them are concentrated in the north-western, northern and north-eastern (‘Three North’) dryland (Fig. 2). Some of these historical events of desertification have been documented by various scholars. Zhu et al. (1986) cited notable instances of widespread desertification in the semi-arid steppe (A in Fig. 2) dating back to the Han Dynasty (202 BC–AD 220). Dong et al. (1988) found that the Mu Us Sandy Land (B in Fig. 2) has existed since the Quaternary, even though its size fluctuated over the years. It has been subject to the southward encroachment of a sandy desert since the

Figure 1. Distribution of sandy deserts (1–8) and lands (9–12) with respect to climatic zones in China. Sandy deserts and sandy lands are differentiated because the latter is formed out of human activities (Source: modified from Fullen & Mitchell, 1994).

422

Y. ZHA & J. GAO

Tang Dynasty (AD 618–906) (Guo et al., 1989). As a remarkable example of desertification in Chinese history, the Ordos Plateau (Fig. 2) contains 120,000 km2 of land that were desertified during the prehistoric period (Guo et al., 1989). At its southern fringe a belt of migratory sand about 60 km wide formed along the Great Wall during the last three centuries. The history of desertification in the Taklimakan Desert (1 in Fig. 1 and F in Fig. 2) can be dated back to 31,000 years ago (Wang & Dong, 1994). At present China still faces a serious problem of desertification. Approximately 13% of the territory comprises of deserts and desertified land (Qu, 1980). It is estimated that 3·3 million km2 have been affected by desertification, accounting for 34% of total land area (Chen et al., 1996). Desertified land in China totals 1·1 million km2 by the account of Zhu & Cui (1996), but 2·2 million km2 by the account of Zhou & Pu (1996). None of the authors provided accuracy for their estimates. The sheer scale of the desertification problem, combined with its complex causes, makes accurate estimates impossible (Fullen & Mitchell, 1994). The massively disparate figures reported are attributed to three reasons: (a) definition of desertification. The land affected by a specific type of degradation was included in one figure, but not in another. Guo et al. (1989) reported a total of 1·3 million km2 of desert and desertified land without specifying the quantity for desertified land alone; (b) types of desertified land. Some authors included desertified land in arid and semi-arid areas whereas others also counted the land degraded by erosion in humid and semi-humid areas. Zhu

Figure 2. Distribution of historical and contemporary desertification in China. Numbers represent sandy deserts/lands; for their names refer to Fig. 1 (Source: modified from Sheehy, 1992).

DESERTIFICATION AND REHABILITATION, CHINA

423

and Cui (1996) included desertified land by water erosion (0·37 million km2) and the areas affected by physical and chemical erosion (0·38 million km2) in their estimate. The figure given by Zhou & Pu (1996) included land affected by soil erosion (1·79 million km2) and salinization (0·065 million km2); and (c) degree of desertification. Areas already affected by desertification were included in one figure whereas areas vulnerable to desertification were also counted in another. For instance, Zhu et al. (1989) included 81,000 km2 of land susceptible to desertification in their estimate. According to Thomas (1993), hyperarid environments should not be considered desertified because they are desert-like in their natural state. Furthermore, vulnerability to desertification should be distinguished from desertification itself (Rhodes, 1991). The amount of desertified land estimated by different authors converges at around 33·4 million ha (Fig. 3) if the revised international definition of desertification by Rhodes (1991) is adopted. This trend of drastic reduction confirms that ‘previous assessments of desertification may have over-estimated the worldwide extent of the phenomenon’ (Thomas, 193), at least for China. Desertified areas are widely scattered in a few clusters in northern drylands (Fig. 3). The most prominent cluster is formed by 207 agropastoral counties in 13 ‘Three North’ provinces where 109·5 million ha of land have been desertified, accounting for 9·2% of the total area in China (Hou, 1985). In this zone alone, 26·9% of the affected land is severely desertified, 25·7% strongly under development, and 47·4% under development. Characterized by a landform of partially stabilized sand dunes covered with shrubs, these areas are usually located at the periphery of a desert, oasis, or the lower stretch of a rive (Figs 2 and 3) (Guo et al., 1982). All of them have a patchy and fragmented pattern of spatial distribution (Zhu & Cui, 1996).

Figure 3. Distribution of contemporary desertification in China separated into actual and potential categories (Source: modified from Zhu, 1992).

424

Y. ZHA & J. GAO

The problem of contemporary desertification has worsened in a number of areas. In North China 50,000 km2 of land were desertified in the second half of this century (Guo et al., 1989). In the agropastoral zone (Fig. 2) desertified land increased from 137,000 km2 in the late 1950s to 176,088 km2 in the 1970s, with another 158,000 km2 potentially vulnerable to desertification (Wang, 1990). In the 1980s desertified areas expanded at an annual rate of 2103 km2 (Shou et al., 1992). Desertification of steppe grazing land throughout North China has reached crisis proportions (Sheehy, 1992). The percentage of desertified land in the Korqin Steppe (A in Fig. 2) grew from 20 in the 1950s to 52 in the late 1970s (Zhu et al., 1984a). Sandy land in Uxin Qi (an administrative unit equivalent of county), Inner Mongolia (Fig. 1) increased from 4193 km2 in 1957 to 5685 km2 in 1977 (Lin et al., 1983). Mobile and semi-mobile sand dunes encroached upon the oasis in the Gurban Tunggut Desert (2 in Fig. 1), Xinjiang Autonomous Region, by 0·5–2·6 m annually during 1958–1986 (Anon, 1987). Affecting millions of people over a vast area, desertification has caused colossal environmental detriment and economic loss. Its direct destruction includes reduced soil fertility, degraded soil structure, and deteriorated vegetation quality (Zhu & Cui, 1996). Loss of soil nutrients by aeolian erosion totals 5600 tons, or the equivalent of fertilizers worth 17 billion yuan (Luo et al., 1994). It causes a direct economic loss estimated between US$2–3 billion (Anon, 1994; Chen et al., 1996). The indirect loss associated with desertification is 2–3 times more.

Causes of desertification Internationally, the causes of desertification have been identified as overcultivation, overgrazing, deforestation, and salinization (Goudie, 1990; Thomas & Middleton, 1994). These human-related factors have also been reported to cause desertification in China. Destructive human activities range from overcultivation, overgrazing of livestock, excessive gathering of fuelwood and plants for medicinal purposes, mining, to construction of transportation routes (Zhu et al., 1981; Sheehy, 1992). Under the same natural settings, plowing sandy land accelerates aeolian erosion by tens, even hundreds of times (Dong et al., 1987). Frequent ethnic wars, recurrent conversion and reversal of land use from crop to pasture triggered desertification 30 km south of the Great Wall in the 1670s (Fig. 2) (Bao et al., 1984). In addition, human errors in policymaking were responsible for rapid desertification in drought-prone sandy lands between the mid-1960s and mid-1970s (Zhu & Cui, 1996). Of the 3·4 million ha newly desertified land in the agropastoral north, 42·9% was caused by overcultivation, 31·1% by overgrazing, 22·2% by excessive collection of fuelwood, and the rest by mining and construction (Zhu et al., 1994b). Although not explicitly identified as a separate factor, overpopulation is the reason for most of the excessive human activities mentioned above. Zhu et al. (1984b) recommended that population be controlled to prevent desertification from worsening. Dong (1992) identified a close correlation between changes in desertification and populations of human beings and livestock. However, he questioned whether overpopulation caused desertification, a process arising from multiple elements. No quantitative relationship between population growth and desert expansion has been established yet. Unlike human-related factors, the environmental settings conductive to desertification in China are quite unique. Commonly identified ones include sandy, loose surfacial sediment deposits and the coincidence of droughts with the windy season

DESERTIFICATION AND REHABILITATION, CHINA

425

(Zhu & Cui, 1996). The former provides the materials to be transported, and the latter fuels the power to move them. No consensus has been reached regarding the exact role played by the two categories of factors. On the one hand, Dong (1992) argued that historical desertification was caused mainly by natural factors, especially climatic fluctuation. It was found that desertification in the Mu Us Sandy Land (9 in Fig. 1) was caused primarily by climatic fluctuation during the Ice Age, and secondly by modern tectonic activities and inappropriate human activities (Dong et al., 198). According to the Expedition Team of Academia Sinica (1978), desertification in ancient agricultural areas resulted from the worsened physical environment, especially climate change. On the other hand, many others are of the opinion that human factors are more important. The main reason for desertification in the Ordos Plateau (Fig. 2) lies not in climate change but in human activities (Hou, 1985). Human activities are largely responsible for desertification in arid and semi-arid China (Zhu, 1982). Desertification worsened in all areas heavily influenced by human activities in north-east Urumqi ¨ (Fig. 1), but remained little changed wherever human influence was small or nil (Liu & Jiang, 1996). After correlating the desertification rate in the Mu Us Sandy Land (9 in Fig. 1) with yearly precipitation, Luk (1983) found that droughts did not always lead to desert expansion, but excessive clearing of land for cultivation and grazing did. Sites of ruined ancient cities in the Mu Us Sandy Land convinced Lin et al. (1983) that desertification was not problematic in historic times. Instead, large-scale cultivation and grazing since the mid-seventeenth century triggered rapid desertification. The seeming contradiction of these findings can be resolved by a simultaneous examination of both categories of factors. Anthropogenic factors are intrinsically interacting with environmental settings in desertification initiation and evolution. Neither of them can function independently without the other, and thus they should be analysed simultaneously. Based on the results from principal component analysis, Dong (1992) found that human factors accounted for 60% of the variation in desertification, and natural ones 40%. Feng (1987) reported that only 10% of the desertified land in China was caused by natural factors such as droughts and aeolian erosion, and the remaining 90% by human activities. Results published in the English literature contribute little to elucidate the debate. Le Hou´erou (1992) thought that global warming could accelerate the process of desertization. Similarly, Wang & Dong (1994) found that global warming would cause desertification in the Taklimakan Desert (1 in Fig. 1) to continue, and the process would be accelerated by human impact. Rising temperatures and declining rainfall for a period of 30–50 years in Sudan may accelerate desertification there (Alvi, 1994). Because of the limited length of climatic records and thus the difficulty in establishing long-term prediction of climate, the question whether continual climate change gives rise to desert expansion cannot be answered with confidence (Anon, 1977). Indeed, it is difficult to separate human- and climatically-induced changes. Desertification research in the revisionist era requires an ongoing awareness of anthropogenic vs. climatic influences on dryland resources (Rhodes, 1991). More detailed studies on the extent of desertification and its long-term monitoring at regional and national levels are needed for the realistic assessment of roles played by desertification-triggering factors (Thomas, 1993). In the absence of convincing evidences from the western literature, the conflict of opinions is reconciled by taking into account the differential temporal and spatial scales of desertification initiation and development. While environmental conditions and physical factors created a fragile ecosystem and initiated the formation of deserts, human elements were principally responsible for their deterioration and expansion (Dong & Liu, 1993). Natural variables played a major role in historic desertification. Anthropogenic elements such as improper land management practices taking precedence over ecological principles are blamed for contemporary desertification. At

426

Y. ZHA & J. GAO

the regional level the physical elements are more important than the human ones whose importance becomes increasingly larger as the scale is progressively reduced to a local one.

Desertification rehabilitation The history of rehabilitating desertified land in China is essentially the history of converting it into productive uses in sandy areas (Chen et al., 1996). Most of the rehabilitating efforts concentrated on the agropastoral zone (Fig. 2) because 55 million people and 10 million ha of cropland and pasture have been affected by desertification (Guo et al., 1989). A prerequisite of a successful rehabilitation scheme is the identification of desertified areas and assessment of desertification severity.

Monitoring and assessment of desertification Historically, desertification information is obtained from ground surveys and expeditions. Such inefficient methods of information acquisition have gradually been replaced by the increasing use of small-scale aerial photographs (Zhu et al., 1984b; Shou et al., 1992). As a supplementary means, field trips are carried out occasionally to verify the results obtained from photointerpretation and to assess their accuracy (Zha, 1989; Guo, 1990; Liu & Jiang, 1996). Since the emergence of space-borne remote sensing, satellite images such as Landsat Multispectral Scanner (MSS) and TM have been utilized to delineate the extent of desertified areas (Luk, 1983; Guo, 1990; Liu & Jiang, 1996). At a spatial resolution of 30 m, TM data enable them to be mapped at an accuracy level comparable to that from aerial photographs (Liu & Jiang, 1996). Desertification severity levels mapped from TM images are consistently within 90% of those obtained from colour infrared aerial photographs (Guo, 1990). Despite these high accuracy levels, satellite images have not completely eliminated the need for aerial photographs. The results interpreted from historical aerial photographs guided the mapping of desertified land from MSS data (Luk, 1983). Two methods have been used to process the remotely sensed data, manual interpretation for aerial photographs and satellite images, and digital analysis for satellite data (Luk, 1983; Guo, 1990; Liu & Jiang, 1996). Desertification monitoring comprises identification of changes in desertified areas, which can be accomplished by overlaying time-sequential data such as remotely sensed images. Comparison of one 1977 MSS image with one 1989 TM image revealed the shrinkage of vegetative cover in a 2100 km2 area near Urumqi ¨ (Fig. 1), Xinjiang, and the southward shift of a sandy desert (Liu & Jiang, 1996). Since satellite images were not available prior to the early 1970s, historical aerial photographs were relied upon to determine the expansion of desertified land (Bao et al., 1984). In the absence of historical aerial photographs, ground survey results were used as a surrogate (Lin et al., 1983). Comparison of the two sets of results could indicate the general trend of desertification change, but not the location where the changes had occured. Overlay of multiple images or maps is ideally carried out in a Geographic Information System (GIS). So far GIS has found limited applications in identifying desert expansion. Zha & Gao (1997) overlaid two desertification distribution maps of Yulin County (Fig. 1), Shaanxi Province, to identify the areas desertified or converted to productive uses between 1960 and 1987. The acquired spatially-based information facilitated the identification of desertification causes at some sites inside the study area. Wang & Kang (1990) outlined a prototype microcomputer-based information system for desertification rehabilitation, dynamic monitoring, and trend forecasting at the

DESERTIFICATION AND REHABILITATION, CHINA

427

country level. However, no further progress has been reported on the planned research. Various criteria have been proposed to assess desertification severity. Bao et al. (1984) employed the depth of underlying sand, the degree of aeolian erosion, and the amount of vegetative cover and shifting sand dunes to map severity at four levels (latent, ongoing, severe and most severe). Percentage of mobile sand dunes was also used to classify desertification as severe, strongly affected and under development (Zhu et al., 1984b). Zhu et al. (1981) applied a combination of the amount of aeolian erosion and the change in surface landforms. An annual removal of 3 cm and deposition of > 5 cm of sand were considered severe, < 1 cm of erosion and deposition slight. However, vegetative cover proves to be a more popular criterion, especially if the results are obtained from remotely sensed images. A vegetative cover of < 5, 15, 30 and 50% is considered, respectively, extremely severe, severe, moderate and slight (Guo, 1990). Similarly, Liu & Jiang (1996) considered a vegetative cover of < 5% extremely severe and < 20% severe. However, a vegetative cover of > 80% represents no desertification hazard. Instead of using a single factor, Dong (1996) derived desertification severity levels from weighted averaging of 16 desertification contributing factors, including potential hazard, current status, desertification rate, human population and livestock size. All these studies were carried out at the regional or local level. No research has been reported on the assessment and classification of desertification severity at the national level. GIS has not been utilized to map desertification severity or to assess desertification hazard and its environmental impact.

Desertification rehabilitation In sandy drylands desertification occurs in two manners, direct encroachment of mobile sand dunes upon grazing land, and deposition of drifting sand over grasses, both under the action of wind (Lin et al., 1983). Therefore, rehabilitating desertification is essentially to slow down wind velocity through increased surface roughness (Zou et al., 1981). Construction of engineering works and planting of vegetation are applied to halting the advance of migratory sand dunes (Dong et al., 1987). Engineering works such as straw checkerboards can effectively reduce wind velocity and minimize the amount of sand transported, even though their optimal width is still debatable (Feng et al., 1994). Checkerboards at a height of 0·15–0·2 m above the ground increase the roughness of a sand surface by 400–600 times, and reduce wind velocity by 20–40% at a height of 0·5 m and by 10% at 2 m above the surface (Zou et al., 1981). The quantity of sand transported over a checkerboard is only 1% of that over a shifting sand dune (Zhu, 1992). Moreover, checkerboards increase soil organic content by 23-fold after a surface crust is formed. Soil crust with moss growing on it can resist aeolian erosion force within a speed as high as 25 m s–1. Nevertheless, engineering works alone cannot eliminate desertification hazards on roads (Chen, 1992). If combined with vegetation networks, their effectiveness is considerably improved (Zou et al., 1981). Planting of shrubs and trees brings more ecological and economic benefits from the control efforts, making it sustainable. The direct benefits are decreased wind velocities, increased soil temperature and organic matter inputs from biomass, improved soil moisture retention, and reduced soil erosion (Yang, 1990; Fullen & Mitchell, 1994). Species of dwarf shrubs suitable for entraining sand include Salix flavida, Hedysarum scopariu, and Caragana korshinkskii planted at an interval of 1–2 m. At a density of > 20%, shrubs can achieve the same effect as a 1 m by 1 m checkerboard (Yang, 1990). However, planted vegetation alone

428

Y. ZHA & J. GAO

in partially desertified areas cannot bring desertification under complete control, especially during its later stages of succession (Chen, 1981). Because of the significance of vegetative cover in reducing wind speed and generating economic benefits, prior to the 1970s biological means were relied upon exclusively to stabilize sand dunes while other factors were neglected (Shou et al., 1992). However, if the land is exploited beyond its carrying capacity, no single control measure can function effectively (Dong, 1992). It was realized later that desertification is a process resulting from multiple factors. Accordingly, a multi-disciplinary approach of rehabilitation and prevention was adopted. The measures were designed for rehabilitating the deteriorated ecosystem based on both ecological principles and economic return. The rehabilitation methods include regulating growth of cropland and livestock, reconverting infertile cropland at the margin of steppe grazing land to semi-natural ecosystems, and developing land use patterns that integrate combinations of grazing land, woodland and cropland (Sheehy, 1992). A ratio of 3:3:4 for land allocated to farming, pasture and forestry in semi-arid sandy areas can lead to an ecological balance (Zhu & Cui, 1996). This ratio varies with the severity of desertification. The more severely an area is desertified, the higher the proportions for woodland and pasture (Zhu et al., 1984b). With a combination of rehabilitation and preventive measures following basic ecological principles, desertification can be harnessed and desertified land be reclaimed for productive uses. Common land reclamation strategies include windbreaks, irrigation with silt-laden river water, and dune stabilization using straw checkerboards and planted xerophytes (Fullen & Mitchell, 1994). The disastrous consequence of the mistaken policy of stressing the paramount importance of grain yield in the 1970s has been corrected by reducing cropland if its cultivation is conducive to aeolian erosion and desertification (Lin et al., 1983). The effectiveness of these rehabilitating measures has not been comprehensively assessed. Zhu & Cui (1996) qualitatively outlined the successful measures in arid and semi-arid areas. Zha & Gao (1997) found planting grasses at the fringe of sandy land is the least effective as they are readily topped by the shifting sand dunes. Instead, scrubs and trees are more resistant to being buried by sand and thus more effective in halting desertification.

Outcome of desertification rehabilitation The achievements of tackling desertification are exemplified by the ‘Three North’ project in the agropastoral zone of North China. Approved by the State Council in 1978, this programme of constructing shelter-belt systems was launched in an area of 4·069 million km2 spreading across 13 provinces. Internationally renowned as China’s Green Great Wall, this multiple-stage project is expected to be completed by the year 2025 when forest coverage will reach 14·95%. In the first stage 7·9 million ha of protective forests were planted (Guo et al., 1989), bringing forest cover from 5·05% in 1978 to 7·09 in 1989 (Zhu, 1990). Of the 5736 km2 of mobile sand dunes in the region, 3068 km2 have been stabilized (Yang, 1990). If they function as anticipated, the farmland-protective forests will protect 8·5 million ha of cropland and increase grain yield by 5·5 million tons. The direct economic benefits from the project are estimated at 25·55 billion yuan, or 27 times the initial investment. Valued at 44·6 billion yuan, the indirect benefits come from increased grain yield, soil conservation, sand fixation and protection of pasture. However, as unveiled by observations of areas planted with trees, the actual achievements are much less spectacular than reported due to the low survival rate of the trees (Becker, 1985). At a smaller scale, the reversion of desertified land to productive uses has taken place in a number of areas. A shelter belt of 170 km long by 300–400 m wide has been

DESERTIFICATION AND REHABILITATION, CHINA

429

established in north-east Ulan Buh Desert (7 in Fig. 1) in Inner Mongolia, protecting tens of thousand hectares of rangeland and cropland in more than 150 counties (Qu, 1980). Thanks to decades of desertification control efforts, oasis area in the southeastern Tengger Desert (6 in Fig. 1) increased by 120 km2 between 1959–1990, with the reclaimed land used for cropping, orchards, forestry, and shelter belts (Zhu, 1992). Desertified land in northern Shaanxi Province (around 9 in Fig. 1) has been reclaimed for farmland and woodland along the river banks and in low-lying moist alluvial fans (Bao et al., 1984). Although the moderately and slightly desertified areas in Yulin Country (Fig. 1), Shaanxi Province are predicted to increase by 104,000 ha, the most severely affected area is forecast to decrease by 190,000 ha, forming an overall decreasing trend (Kang et al., 1995).

Summary and conclusions Enormous efforts have gone toward tackling desertification in China since the United Nations held its first conference on combating desertification 20 years ago. These efforts concentrated on defining desertification, determining its causes, assessing its spatial distribution and severity, and rehabilitating desertified land into productive uses. Some of the problems facing Chinese scholars are identical to those facing their western counterparts. These issues include how to define desertification properly and how to assess the exact role played by human-related and environmental factors in desertification. Due to the insufficient amount of data collected, it is difficult to disentangle the impact of anthropogenic desertification effectively from that of environmental desertification. Because of the pressure generated by an ever increasing population and dwindling arable land in China, it is of paramount importance to rehabilitate the land lost to desertification to productive uses. Consequently, restoration of desertified land makes up a huge portion of the scientific endeavour in desertification research. A disproportionate amount of emphasis is placed on desertification control whereas insufficient attention is given to prevention. Subsequently, desertified land is rehabilitated to productive uses on the one hand, but on the other, formerly stabilized sandy land is encroached upon by shifting sand dunes. In taming desertification by biological means, huge efforts are devoted to planting grasses and tree saplings. However, inadequate efforts are made to ensure their survival and the sustainability of the rehabilitation programme. These earlier problems have been remedied after the realization that desertification resulted from a variety of factors, both human activities and natural settings. The multi-disciplinary approach of rehabilitation and prevention based on ecological principles and economic return has achieved much more success in rehabilitating the deteriorated ecosystem than the biological means. To conclude, China is facing a serious desertification problem. Most of the affected areas are located in the arid and semi-arid north. They are caused by both environmental settings and inappropriate human activities including overcultivation, overgrazing, and excessive gathering of fuelwood and plant species for medicinal purposes. The natural settings are important to the initial formation of desert conditions in historical times, whereas anthropogenic factors are critical to contemporary desertification. After preventive measures following sound ecological principles were adopted, the desertification trend has been reversed at various geographic scales. Engineering measures alone are not so effective in halting the encroachment of sand dunes as biological measures that can bring more economic return from the rehabilitating efforts. The multi-disciplinary, ecologically-sound rehabilitating approach proves to be most effective in restoring desertified land to productive uses. Jay Gao would like to thank the University of Auckland for granting him research leave during which this research was undertaken.

430

Y. ZHA & J. GAO

References Alvi, S.H. (1994). Climatic changes, desertification and the Republic of Sudan. GeoJournal, 33: 393–399. Anon (1977). Desertification: an overview. In: Secretariate of the UNCOD (Ed.), Desertification: its causes and consequences, pp. 1–61. Oxford: Pergamon Press. 448 pp. Anon (1977). Desertification in Xinjiang Autonomous Region gets worse day by day. Geographical Knowledge, 11: 32 (Reprinted from Jingji Ribao, 6 July 1987). Anon (1994). Desertification control. In: State Planning Commission and State Science Commission (Ed.), China’s Agenda 21, pp. 142–151. Beijing: Chinese Environment Science Publishing House. 192 pp. Aubr´eville, A. (1949). Climats, For´ets, et D´esertification de l’Afrique Tropicale. Societe de Editions Geographiques. Paris: Martime et Coloniales. 255 pp. Bao, Y., Wen, Z., Dong, S., Zhang, J. & Yang, T. (1984). Inventorying desertification types and its prevention in northern Shaanxi Province. Journal of Desert Research, 4(2): 35–44. Becker, J. (1985). Sand trickles through China’s other great wall. New Science, October 24. Chen, W. (1981). The changes in soil properties after an effort in artificial sand-fixation and plantation in the Shapotou area. Journal of Desert Research, 1(1): 40–48. Chen, G. (1991). On desertification and aeolian sandy land. Research of Arid Areas, 8: 57–60. Chen, G. (1992). Effect of sand-drifting prevention on roads in the Tengger Desert. Journal of Desert Research, 12(3): 38–41. Chen, G., Dong, Z. & Yan, P. (1996). Desertification: international research topics and research strategies of China. Exploration of Nature, 15: 1–5. Dong, Y. (1992). Role of human factors in desertification. Journal of Desert Research, 12(1): 16–26. Dong, Y. (1996). Assessment of hazard degree and analysis of desertification trend in sandy China. Journal of Desert Research, 16(2): 127–131. Dong, Y. & Liu, Y. (1993). Review and prospect of the studies on desertification in China. Geographical Research, 12(2): 94–102. Dong, G., Li, C. & Jin, J. (1987). Some results of simulated experimentation on aeolian erosion using a wind tunnel. Chinese Science Bulletin, 32: 297–301. Dong, G., Shen, J., Jin, J. & Gao, S. (1988). The conceptualization of desertification and desertization. Arid Land Geography, 11: 58–61. Dregne, H.E. (1983). Desertification of Arid Lands. London: Harwood Academic Publishers. 242 pp. Expedition Team of Academia Sinica (1978), Soil Geography in Inner Mongolia Autonomous Region and West of Northeast China, pp. 12–28. Beijing: Science Publishing House. Feng, J. (1987). Willing to be the foundation upon which desert science is developed: a visit with Professor Zhu Zhenda of the Lanzhou IDRAS. Geographical Knowledge, 12: 20–21. Feng, L., Lu, J. & Di, Y. (1994). Review on the prevention of sand damage to railway lines in desert areas of China. Journal of Desert Research, 14(3): 47–53. Fullen, M.A. & Mitchell, D.J. (1994). Desertification and reclamation in north-central China. Ambio, 23: 131–135. Glantz, M.H. & Orlovsky, N. (1983). Desertification: a review of the concept. Desertification Control Bulletin, 9: 15–22. Goudie, A.S. (1990). Desert degradation. In: Goudie, A.S. (Ed.), Techniques for Desert Reclamation, pp. 1–33. Chichester: John Wiley & Sons. 271 pp. Guo, S. (1990). Application of remote sensing imagery in studying desertification in Yulin Country. In: Wang, C. (Ed.), Comprehensive Study on the Forested Areas in the Loess Plateau, pp. 188–194. Beijing: Science and Technology Publishing House. 402 pp. Guo, S.L., Qi, W.H. & Li, L.X. (1982). Use of the catastrophe theory model to study the process of desertification — an example of sandy land of northeast China. Acta Geographica Sinica, 37: 183–193. Guo, H., Wu, D. & Zhu, H. (1989). Land restoration in China. Journal of Applied Ecology, 26: 787–792. Hou, R. (1985). Ancient city ruins in the deserts of the Inner Mongolian Autonomous Region of China. Journal of Historical Geography, 11: 241–252. Kang, G., Yao, F., Wang, X. & Dong, Y. (1995). Application of gray prediction theory in desertification monitoring. Journal of Desert Research, 15: 353–356.

DESERTIFICATION AND REHABILITATION, CHINA

431

Le Hou´erou, H.N. (1992). Climatic change and desertization. Impact of Science on Society, 166: 183–201. Li, X. (1988). Explanation of barrenification, sandification, and desiccation and their uses in China. In: China Natural Resource Research Committee (Ed.), Study on Natural Resources in Arid and Semi-arid China, pp. 23–26. Beijing: Science Publishing House. 230 pp. Lin, Y., Chen, C.K. & Chen, C.T. (1983). A discussion on the direction of production development and way of halting desertification in the Mu Us sandy land, Ordos Plateau — based on its physical conditions. Acta Geographica Sinica, 38: 230–240. Liu, X. & Jiang, X. (1996). Extraction of desertification information from the optical and digital processing of multiple-date remotely sensed images. Arid Land Geography, 19(3): 1–6. Luk, S.H. (1983). Recent trends of desertification in the Maowusu Desert, China. Environmental Conservation, 10: 213–224. Luo, B., Zhou, S., Zhang, X. & Meng, X. (1994). World desertification status and the successful experience of desertification control in China. Journal of Desert Research, 14(2): 49–54. Qu, G. (1980). Deserts in China and their prevention and control. Mazingira, 4(2): 74–79. Rapp, A. (1974). A review of desertization in Africa — water, vegetation and man. Sweden: Secretariat for International Ecology, Report No. 1. 77 pp. Rhodes, S.L. (1991). Rethinking desertification: what do we know and what have we learned? World Development, 21: 1137–1143. Sheehy, D.P. (1992). A perspective on desertification of grazingland ecosystems in North China. Ambio, 21: 303–307. Shou, W., Li, N., Zhou, J., Zhang, X. & Zhao, Y. (1992). Discussion on the fundamental models of controlling desertified land in China. Journal of Desert Research, 12(3): 22–27. Thomas, D.S.G. (1993). Sandstorm in a teacup? Understanding desertification. The Geographical Journal, 159: 318–331. Thomas, D.S.G. & Middleton, N.J. (1994). Desertification: exploring the myth. Chichester: John Wiley & Sons. 194 pp. UNCOD (1978). United Nations Conference on Desertification, Round-up, Plan of Action and Resolutions. New York: United Nations. 43 pp. Wang, S. (1990). Ecological environment in north, northwest and northeast China. In: He, H. (Ed.), Study on Remote Sensing of Ecology and Environment, pp. 1–15. Beijing: Science Publishing House. 375 pp. Wang, Y. & Dong, G. (1994). Sand sea history of the Taklimakan for the past 30,000 years. Geografiska Annaler, 76A(3): 131–141. Wang, Y. & Kang, G. (1990). A microcomputer-based desertification environment information system. In: ESCAP/UNEP (Ed.), Remote Sensing for Desertification and Vegetation Mapping, pp. 95–98. Tehran: UNEP. 134 pp. Wu, Z. (1965). Observations on sand dune movement and preliminary study on the prevention of desertification hazard. Research on Desertification Control, 7: 7–14. Yang, G. (1987). Discussion on several issues related to desertification. Research on Arid Regions, 1987: 73–77. Yang, Z. (1990). Benefit analysis of the “Three North” protective forest in Yulin prefecture. In: Wang, C. (Ed.), Comprehensive Study of the Forested Areas in the Loess Plateau Using Remote Sensing, pp. 112–121. Beijing: Science and Technology Publishing House of China. 402 pp. Zha, Y. (1989). Research on Quantifying the Desertification Processes Using Remote Sensing, Unpublished Master’s thesis, Institute of Remote Sensing Applications, Academia Sinica, Beijing. 34 pp. Zha, Y. & Gao, J. (1997). Effectiveness of measures to combat desertification in Yulin Country, Shaanxi, Western China using remote sensing and GIS. Applied Geography (submitted). Zhou, J. & Pu, L. (1996). Rustic opinions on concept of desertification and its practical significance. Journal of Desert Research, 16: 191–195. Zhu, Z. (1982). Historical physical geography. In: Physical Geography of China (Ed.), pp. 249–260. Beijing: Science Publishing House. 262 pp. Zhu, Z. (1986). Problems of land desertification in humid and semi-humid zones. Journal of Desert Research, 6(4): 1–12. Zhu, J. (1990). Comprehensive evaluation of the ecological benefit of the entire “Three North” Protective Forest. In: Nu, X. & He, H. (Eds), Study on Remote Sensing of Ecology and Environment, pp. 16–25. Beijing: Science Publishing House. 375 pp.

432

Y. ZHA & J. GAO

Zhu, Z. (1992). China: desertification mapping and desert reclamation. In: UNEP (Ed.), World Atlas of Desertification, pp. 46–49. London: Edward Arnold. 69 pp. Zhu, Z. & Cui, S. (1996). Distribution patterns of desertified land and assessment of its control measures in China. China Environmental Science, 16: 328–334. Zhu, Z. & Liu, S. (1981). Desertification Process in North China and Its Regional Control. Beijing: Chinese Forestry Publishing House. 83 pp. Zhu, Z., Guo, H. & Wu, G. (1964). Study on sand dune movement near the oasis in southwest Taklimakan Desert. Acta Geographica Sinica, 30(1): 35–50. Zhu, Z., Liu, S. & Xiao, L. (1981). The characteristics of the environment vulnerable to desertification and the ways of its control in steppe zone. Journal of Desert Research, 1(1): 2–12. Zhu, Z., Liu, S. & Di, X. (1984a). Retrospect of the history of desert research and problems in the field in China. Journal of Desert Research, 4(2): 3–7. Zhu, Z., Liu, S. & Yang, Y. (1984b). The possibilities and realities of re-managing desertified land in the interlacing agropastoral areas in northern China. Scientia Geographica Sinica, 4(9): 197–206. Zhu, Z., Liu, S., Wu, Z. & Di, X.M. (1986). Deserts in China. Lanzhou: Institute of Desert Research, Academia Sinica. 132 pp. Zhu, Z., Liu, S. & Di, X. (1989). Desertification and its Control in China. Beijing: Science Publishing House. 126 pp. Zou, B., Cong, Z. & Liu, S. (1981). A preliminary observation on the basic characteristics of sand-carrying currents and the effects of adopted prevention and control measures in Shapotou. Journal of Desert Research, 1(1): 33–39.