Principles of restoring vegetation

1. Promote the use of native species and minimize reliance on exotic species
China has extensively used exotic species to rehabilitate degraded vegetation. Such efforts have sometimes met short-term success, but problems such as slowed growth have generally occurred later. Experience suggests that exotic species will either fail, due to a lack of adaptation to local conditions, or will thrive and cause damage to natural ecosystems (see Annex A.4).
Native species have adapted to the full range of local conditions over thousands of years. The utilization of resources by different species in an ecosystem is inter-linked; species sharing ecosystems have an intricate balance of resource use. Any exotic species which can survive in local conditions is likely to disrupt this balance, killing or out-competing native plants for space and nutrition. Once native plant populations have been depleted, many other species (such as birds, mammals, invertebrates and fungi) which rely on them for suitable food and habitat will also decline and even disappear. This situation will reduce the resistance of the ecosystem to outbreaks of pest organisms.
Widely planting exotic species will first cause the loss of China's unique biodiversity resources, followed by a decrease in the value of ecotourism and local biological products. In addition, the reduction of ecological functions, such as soil erosion control, water conservation and biodiversity conservation of local vegetation, will cause much greater economic loss when evaluated on a larger scale.

It should be noted that nativeness is not a political entity. China encompasses many different biomes, and most species in China do not occur naturally throughout the country. A native species is one that occurs naturally in a given biogeographical unit; not all species from China are native to a given locality.

Recommendations

• Avoid use of exotic species in any replanting project.
  
• Invest in the techniques required to cultivate native (local) species for restoring vegetation.

2. Aim toward the appropriate climax vegetation
Without human disturbance, the different climates, parent rocks, soils and humidity conditions lead to corresponding vegetation communities known as climax communities. Efforts to restore vegetation should take into account the natural climax vegetation of a given zone, which is best suited to exploit the conditions sustainably.
Zonal climax vegetation ecosystems distributed across China include forests, grasslands and deserts (Box 3). Natural vegetation allows the most effective penetration of water into soils, and soil and water conservation (see Annex A.1). Programmes to rehabilitate degraded landscapes should aim to recreate the original, natural vegetation formations (i.e. forests, shrublands, grasslands) as far as possible. Any change to the components of the system would alter, and very often impair, the functioning of the original system. For instance, "returning" corn fields on slopes with a gradient of 25-35O in Sichuan Province to plantations of gingko or tangerines does not substantially improve environmental services (i.e. water conservation). In contrast, the original, species-rich sub-tropical forests, with mixed native species, are vastly more effective in providing these environmental services, and it is these that should be restored. Similarly, in those upland areas of China in which the natural vegetation was grassland and shrubland, there are no scientific grounds for planting willow, poplar or other trees. In such a case, it is most appropriate to use a good mix of native grasses, herbs and shrubs to recreate what had been there before these lands were degraded.
In terms of moisture and air temperature conditions, forest climax ecosystems can be divided into temperate and tropical, coniferous and deciduous, while grassland ecosystems can be further divided into meadow grassland, typical grassland, desert grassland, high and cold grassland, shrubland and tundra that is distributed in alpine and subalpine zones.
Climax communities are characterised not only by their canopy vegetation but also the understorey layers. In many parts of China the natural forest canopy is dominated by conifers, but there is almost always a broadleaf or bamboo understorey. Efforts must be made to ensure that both the upper and lower storey species will be included in the new forests. Plans will need to be made for the collection of wild seeds and the establishment of the necessary seedling nurseries. Climax tree species may need to be raised under shade and protected from wind. Where vegetation is at an early successional stage, coloniser plants and pioneer trees will often be required (see below). Bamboo understorey is difficult to replace because of the 20-30 year intervals between mass flowering events106. However, experiments in Sichuan have shown good success using vegetative propagation by planting small clumps of transposed bamboo intermittently in the undergrowth.
In forests a deep leaf litter, moss cover, bamboo thickets or dense undergrowth should be achieved. Plantations that are too dark or which naturally inhibit undergrowth are generally poor at protecting soils. Exposure of tree roots is a clear indication of high levels of soil erosion and the need to change the forest or plantation structure.

Box 3. China's climax ecosystems 6
Geographic region	Zonal climax ecosystem
North	Cool-temperate coniferous forest
	Mid-temperate coniferous and deciduous broadleaf mixed forest Warm-temperate deciduous broadleaf forest North subtropical deciduous broadleaf forest, with evergreen components Middle subtropical evergreen broadleaf forest South subtropical monsoon evergreen broadleaf forest North tropical rain forest and monsoon rain forest Coral reef forest on South China Sea Islands
South and Southeast
East	Forests
	Meadow grasslands Typical grasslands Desert grasslands Alpine grasslands Desert scrub
Northwest

Among the temperate forests of China, primary ecosystems are generally rare or absent. In some cases, primary vegetation can be inferred by comparison with ecosystems elsewhere. In the Great Xinggan Mountains, the cool-temperate Xinggan larch and mid-temperate Korean pine forests are extensions of the East Siberian coniferous forests and the Far-east mixed Korean-pine/broadleaf forests of Russia respectively. Since they are quite similar in the composition of flora, it is not very difficult to estimate their primary types.
In the warm-temperate zone, human damage to the various deciduous forest types is extremely serious. Most of the so-called primary ecosystems at present are actually still maturing after past clearance. Only when such pre-climax forest occurs at similar latitude and conditions to primary deciduous broadleaf forest can we provisionally determine the primary forest type. Yang YuPo and Xiang Chenghua^[45] describe several different succession series for montane forests in Sichuan that would be useful to follow in reforestation efforts.
Climax formations in the tropics are even more complex and take a much longer time. Cheung^[46] estimates the normal succession time for re-establishing sub-tropical broadleaf forest to be about 100 years compared to 1000 years for tropical semi-evergreen forests. The result is that continuing disturbance will cause recovering semi-evergreen tropical forests close to the climatic ecotone between tropical and sub-tropical to flick across to the shorter channel and develop into sub-tropical types.
Those natural grasslands dominated by annual species reach an arrested climax very quickly. Component species have well dispersed seeds and can colonise suitable spaces annually. Development into more perennial or woody vegetation is arrested by fire, grazing pressure, seasonal flooding or physical and edaphic factors. These grasslands are highly dynamic and any shift in these characteristics will have significant changes on the form of the grassland community.
On the other hand, many of China's most valuable grasslands occur at high elevations on the Qinghai-Xizang plateau - and these tend to be dominated by perennial sedges (such as Kobresia). These grasslands are very slow to reach a climax state following disturbance. While very productive and nutritious, the shoot to root ratio is low and these grasslands are characterised by a thick layer of binding roots which act to stabilise the topsoil. It is inadvisable to plough these grasslands - they provide their greatest ecological services when left intact.
Ecosystems have sometimes been "constructed" to restore certain aspects of natural ecosystem function. Such "optimised" artificial ecosystems are sometimes referred to as ecological projects; they allow the testing of certain ecological theories. They depend on simulation of natural ecosystem structure using ecological principles. Unfortunately, the approach is extremely difficult, and the scientific basis generally inadequate. An optimised artificial ecosystem has different aims compared with economic agricultural management, regional administration and reforestation. To succeed, artificial ecosystems require the close harmony of the various components with each other and with the environment. This method has already been tried in the subtropical regions of China, but faces many difficulties. One is that it may take many years for problems to emerge, due to uncommon weather conditions, pest outbreaks etc., such that apparent "success" can be short-lived. Another is that it is very difficult to apply the experiences of establishing artificial ecosystems from one place to another.

Recommendations

• Restored vegetation should, when mature, be structurally and functionally equivalent to the native vegetation in a given area. This calls for restoring the approximate species composition of natural ecosystems.
  
• All storeys must be adequately represented in the restored vegetation. These may include shrub and bamboo layers, herb and moss layers, and leaf litter.
  
• Where the primary ecosystems are no longer present, the appropriate climax vegetation should be inferred from available information.
  
• Artificial ecosystems do not provide all the environmental benefits and insurance that native climax vegetation provides. While they may have experimental value, they should not be relied upon as a mainstay of vegetation restoration.

3. Cover denuded ground
The most highly degraded lands, in which vegetation is absent, are responsible for much of the soil erosion and hence flooding (see Annex A.1.3) and loss of water storage capacity (see Annex A.1.1), as well as shifting sands (see Annex A.1.7). These lands call for emergency measures to restore vegetation cover.
In most cases it is not possible to go directly from bare land to climax vegetation. First, the land will need to be stabilised and covered over by coloniser or pioneer vegetation. In tropical regions this may be a fast growing tree such as Macaranga. In temperate regions it is more likely to be a cover of grasses, herbs and shrubs, though pioneer trees such as pines and birch may become established rather early in the succession.
Both in desert oases and in semi-arid areas, networks of farmland protection forests help to control sandstorms, moderate the local climate and prevent salinization. Such forests enable persistently high agricultural yields in arid regions. Farmland protection forests comprising "narrow forest belts (3-10m) and small lattice (150-180mu)" are widely planted in northeastern China. They have succeeded in stabilising conditions for forest growth, whilst reducing land and water conflicts between farms and forest, which are characteristic of "broad forest belts (20-22m) with big lattice (750mu)". Lands adjacent to desert and wind-eroded land are most affected by shifting dunes and sandstorm damage. Forest belts in these areas have provided protection against such problems. Large-scale wind and sand control forests established on the northern edge of the Wulanbuhe Desert, on the verges of the Gansu Corridor oasis, in the Caidamu Basin and in Xinjiang Province have served to protect oases and control shifting dunes.
In other desert and semi-desert areas, it is very difficult to restore vegetation. But usually there are soil deposits below the sand, and a low water table. Such conditions are suitable for superxeric shrubs and herbs. The restoration of natural vegetation through the construction of closed grass belts is an important component of the measures to protect oases and improve the environmental conditions of sandy regions^[15].
Throughout the country, there is exposed soil at land scars resulting from landslides, gulley erosion and engineering works. These small areas with no vegetation cover at all create a disproportionate contribution to local soil loss and sedimentation of water systems. Special programmes should be put in place for the treatment of these scars.
In some cases engineering work has to be undertaken to stabilise the land surface. Local crews must be on hand to tackle such jobs. In all cases bare areas should be quickly seeded to establish an immediate vegetation cover.
Ideal seeds are available in local waste areas and local biologists should study the natural pattern of plant colonisation of bare ground so that treatment can follow natural succession. Seeds of local grass and bushes should be collected for such scar treatment. In some cases, even school children could be usefully engaged in such a programme of seed gathering.
Ideal plants on land scars are fast growing native weeds, especially if they can root in loose stony soils and bind the substrate. Turf forming grasses, creeping plants such as Rubus and coloniser bushes such as Sambuccus and Buddleia give excellent cover in most temperate parts of China. Generally speaking, the most feared and aggressive weeds faced by farmers in their fields will be among the most successful coloniser species for quickly getting a first vegetation cover on bare land.

Recommendations

• China should launch a new programme for the emergency treatment of bare land scars such as landslides, erosion gulleys or lands where vegetation has been cleared during construction works. Emergency treatments will involve both physical engineering as well as suitable application of vegetation cover.
  
• At the small scale, land scars should be immediately seeded with native colonising plants.
  
• At the larger scale, lattices of protective vegetation may be effective. The appropriate natural vegetation is most likely to achieve positive environmental effects.

4. Promote heterogeneity and follow natural succession pathways
Because of human disturbance, most land in China is now degraded from climax ecosystems to one of a variety of successional phases. Succession of plant communities is a long process. Under constant human disturbance, vegetation may remain in an early phase, or even degrade further. For example, forests converted to farmlands may subsequently degrade to bare land after erosion (see Annex A.1.4) and lose their fertility (see Annex A.1.5).
All natural ecosystems have a characteristic degree and scale of patchiness. This is partly caused by the pattern of small-scale disturbances, such as tree-falls, which contribute to variation in physical conditions. Where the canopy is opened, light-requiring saplings can grow. Most natural forests are age-staggered, so that there are always well-developed sapling and recruitment layers as well as mature trees. This allows sustainability of the forest system. This is usually not the case in monoculture plantations, which are also susceptible to soil panning, and the loss of the hydrological sponge effect.

Recommendations

• Monocultures should be avoided, and regenerating vegetation should have a level of patchiness appropriate to the climax ecosystem.
  
• Forests should be age-staggered so that there are well-developed sapling and recruitment layers.
  
• In economic plantations the soil under the trees should be protected against panning. The best way is to allow establishment of appropriate understorey and litter layers.

5. Restore ecological interactions
"Ecological forest" is a term often misused in China. An adherence to ecological processes, including succession, is an important part of true ecological forestry, which is distinguished by "the emphasis placed on natural patterns and processes: understanding them, working in harmony with them, and maintaining their integrity, even when it becomes financially difficult or inconvenient to do so"^[8].
The soil and litter of natural ecosystems are important in the balance and cycling of water, nutrients, gases and minerals (see Annex A.1). Restoring this balance is a fundamental aspect of rehabilitating degraded land. Increasing the sponge effect of forests (see Annex A.1.1) requires increasing the soil permeability under the forest.
Natural forests have a rich biota of soil organisms (earthworms, springtails, termites, ants, burrowing lizards and mammals, etc.) that aerate the soil layer and actively increase soil permeability and fertility. In addition, the combination of deep-rooting and shallow-rooting trees enhances penetration of water to the soil and underlying rocks. In economic plantations the ground under the trees sometimes suffers from low water penetration, due to compaction of the soil. Herbaceous vegetation can improve this. Farmers may be tempted to clear weeds under their trees, even though this will increase erosion, as they suspect competition for resources. However, negative effects of herbs on trees are slight and some kinds, particularly legumes, may even help to increase the supply of nutrients through nitrogen fixation.
Ecological forestry requires an understanding of the diverse and complex roles played by animals in forests, e.g. in seed dispersal and pollination, as excavators of tree cavities used by other organisms, and as agents that control the spread of pests. Knowledge of the needs of different species can lead to simple forest treatments that can accelerate natural processes of regeneration.
Dead wood and leaf litter are important in forest functioning and recovery9. Nutrient cycling and energy flow in forest depend heavily on dead wood and litter. Fallen logs generally have a high proportion of carbon, nitrogen and phosphorus, and mineral elements such as potassium, calcium, magnesium and sodium. Fungi in fallen logs can also fix nitrogen. Seeds of trees (e.g. Abies and Tsuga) and spores of ferns and mosses germinate on rotting fallen logs, which provide seedlings with nutrition and water. Logs also lack certain ground-dwelling pathogenic bacteria, that otherwise limit seedling growth. Many microorganisms, invertebrates, reptiles, amphibians, birds and mammals depend on such logs for shelter and food. These help control forest pests and further contribute to ecosystem recovery. Removal of fallen logs will impair soil fertility and ecological functioning of forest.
Trees and other plants cannot reproduce and disperse without the agents that pollinate them, or disperse their seeds. Wild animals should be protected so as to use their function as seed dispersal agents. Some seeds adhere to the fur or feathers of animals and are passively dispersed. Some animals eat fruits and disperse seeds in their dung. Others, such as birds that eat conifer seeds, spread seeds accidentally during food collection. Primates, pheasants and pigeons can disperse large fruits of climax trees in this way. In tropical zones, fruit bats can also disperse large seeds. Smaller birds such as bulbuls and thrushes are effective at dispersing small seeds of berries and can significantly accelerate the natural spread of woody vegetation. The most important seed dispersers in tropical and subtropical regions include fruit bats, macaques, gibbons and civets among the mammals, and barbets, hornbills, fruit pigeons, broadbills, corvids, muscicapids, bulbuls, white-eyes, laughingthrushes, babblers and flowerpeckers among the birds^[10].
Planting shady cover for shy species (e.g. pheasants) or fast growing food attractants such as Rubus, Sorbus etc. at distances of a few hundred metres from the forest edge will encourage birds to walk or fly across open areas to feed in such spots and deposit other seeds from the forest as they do so. Where young forests lack tree cavities, appropriately-designed boxes may provide temporary nesting and roosting sites for woodpeckers, owls, tits and other birds, as well as bats, which consume vast numbers of insects that might otherwise threaten tree survival.
It is important to design as much connectivity as possible into the forest cover to allow genetic exchanges between wildlife populations and allow species to colonise the new forest areas. The greater the number of species able to enrich the forest the more ecologically robust it will be and the more likely to sustain itself.

Recommendations

• Degraded soil requires remediation of its physical, chemical and biological properties to restore its hydrological and ecological functions. This is best achieved by the full range of rooting, shading and nutrient flow occurring in natural ecosystems.
  
• Any replanting project should use native species that provide the necessary fruits, edible leaves, shelter, breeding sites for local wildlife species.
  
• For each ecosystem, the agents of plant dispersal, pollination and mortality should be thoroughly understood. Where necessary, research should be undertaken on these processes in natural ecosystems.
  
• Decaying plant matter (logs, leaf litter) is often an important substrate for fungi and small animals which are important in the food chain. Damp wood and vegetation should not be removed from recovering ecosystems.

6. Prioritise protection of existing natural ecosystems
The most important step in restoring the natural vegetation of China is to safeguard those surviving ecosystems with the highest ecological integrity. Besides giving the best models for how natural ecosystems should function, they are essential sources of fauna, flora and micro-organisms for the recolonization of degraded areas. "The first rule of intelligent tinkering is to keep all the pieces"^[12].
Rich natural habitats not only serve ecological functions best but have added benefits in the form of providing valuable biodiversity products (see Annex A.5.1), eco-tourism opportunities (see Annex A.5.2) and provide natural controls against the spread of fire (see Annex A.2.2) and pests (see Annex A.4).
Intact natural ecosystems have been shown to also provide the best ecological services in relation to climate moderation (see Annex A.2) and provision of water flow for human uses (see Annex A.1.2), flood control (see Annex A.1.3) and hydropower generation (see Annex A.6).
To some extent protection is achieved by the existing nature reserves system, but there are many problems. Some ecosystems are inadequately covered by nature reserves (e.g. lowland coastal plains deciduous forest). Others, while protected on paper, are nevertheless subject to ongoing exploitation and degradation.
The protected areas system should be designed systematically to encompass all natural vegetation types in China. It is also essential that this system should be actively managed to ensure ecosystem integrity and functionality are maintained and restored. True nature reserves should not be places for hunting, fishing, collecting, logging or grazing. Undisturbed streams and rivers are also important to the functioning of the surrounding ecosystems.

Recommendations

• Ensure all ecosystems are adequately represented by nature reserves.
  
• Ensure that protected areas are managed to conserve biodiversity. Conservation involves prevention of damage to terrestrial and aquatic habitats, and prevention of harvesting of wild species. This calls for capacity building on a massive scale.

7. Extend natural ecosystems through closure of degraded lands
"Natural restoration" means restoration of a degraded ecosystem without artificial assistance, but solely by natural succession. Land closure is the typical method of natural restoration. Closing forests or grasslands to human activity enhances natural renewal. Examples include involve total exclusion, limiting grazing to a rotation system, or controlling grazing intensity according to the carrying capacity of the land ^[22].
Closing deforested land has the advantages of shortening the time needed to achieve forest cover, preserve rare species and increase the stability of forest stands. Under closed conditions, forest will regenerate after 8-10 years in southern China, and 10-15 years in northern China and southwestern alpine regions.
In subtropical regions, where natural conditions are advantageous, moderately degraded ecosystems can return to the zonal natural vegetation quite quickly when human disturbance stops.
Allowing forests to regenerate naturally also has the advantage of small inputs but large benefits. The cost of artificial afforestation per hectare is generally orders of magnitude higher than that of closing forests; yet in terms of water and soil conservation, microclimate control and improvement, biodiversity protection and maintenance of atmospheric balance, artificial forests are inferior when compared with closed forest.
The work of closing forests should involve the county (forestry department), township (forestry centre) and village levels working in unison. Planning should be made on the basis of comprehensive investigations.
In the restoration of vegetation and ecological conservation, it is important to set up more natural reserves and implement strict protection in key areas such as coastal zones (see Annex A.2.3), headwater regions (alpine lakes and streams), water sponge forests (core areas of forest) and source regions of drinking water (reservoirs) ^[14],^[15].
It is not always necessary to create physical fences to close land, but regulations need to be imposed to strictly ban logging, firewood collection, burning of vegetation and free-range grazing by domestic animals. Ungulates create trails that develop into erosion gullies, and loosen the topsoil so that it may be washed away in heavy rains. Ungulates also eat young trees and other plants, retarding regeneration. Such animals should be confined to pastures or paddocks. In some situations the controlled cutting of herbs and grass for fodder could be permitted in closed forests.
Getting people to accept and respect forest closure is extremely difficult. Local communities should be organised into forest watch brigades and be paid to ensure that the rules are respected. Payment should be indexed to success, so that those communities that do a good job of ecosystem protection are duly rewarded. A share of the immense benefits to the country as a whole should be returned to these communities.
Following mountain closure, vegetation cover, species diversity and above-ground biomass all increase obviously. The natural regeneration of trees is also very important in the restoration of forests. Scattered trees or seedlings are often seen in some shrub and brushwood ecosystems, and will mature gradually under protected conditions.

Areas of forest that are regenerating naturally typically have mixed species and a vertical structure consisting of canopy, shrub understorey and herb layers. They generally have plenty of site conditions for further recovery.
Grasslands also respond well to closure. After three years of closure, the above-ground biomass in degraded meadow can double^[20]. Biomass and plant species richness increase and the species composition develop to suit the habitat^[21].

Recommendations

• Natural regeneration is the most powerful and economical force in restoring natural ecosystems. Closure should be the primary method for restoring vegetation.
  
• The work of closing forests should involve the county (forestry department), township (forestry centre) and village levels working in unison. Planning should be made on the basis of comprehensive investigations.
  
• For closed areas, regulations need to be imposed to strictly ban logging, firewood collection, burning of vegetation and free-range grazing by domestic animals.
  
• Local communities should be organised into forest watch brigades and be paid to ensure that the rules for closed areas are respected.

8. Use appropriate measures to accelerate natural regeneration
While natural restoration of ecosystems is very effective, it may be necessary to take additional artificial measures for various reasons. First, natural restoration of seriously degraded ecosystems can take a very long time. Second, the natural development of a seriously degraded ecosystem often departs from the direction toward a typical natural ecosystem (deflected succession). Thirdly, in some special cases, such as wastelands affected by heavy metal pollution or changes of soil acidity and alkalinity, natural restoration is very difficult without artificial intervention. Fourthly, some of the component species of the succession sequence may no longer be present.
Having determined the climax vegetation for the location, and the stage in natural succession from coloniser to climax, it is important to determine the limiting steps in vegetation recovery. This may be the soil condition, recolonisation by plants or animals, or mutually-restrictive relations between species^[22].
In some conditions, it is most important to improve the physical and chemical qualities of soil. For example, organic matter or sulphur waste can reduce excessively-high soil pH, while lime can raise excessively low pH. Organic fertilizer can reduce the excessive heavy metal content of soil through adsorption; irrigation can remediate soil with a high surface-salt content.
Different kinds of ecosystem have different needs of nutrient elements. It is very difficult to restore degraded ecosystems in conditions where nutrients are lacking. Directly applying nutrient elements in fertilizers may cost too much and need to be done year after year; artificial fertilizers also fail to restore healthy soil. A most effective and practical approach is to grow leguminous plants, so as to improve nutrient availability through nitrogen fixation. An alternative is to apply organic fertilizer, which will help increase the activities of soil microorganisms, and gradually restore the phosphorus content of degraded soil.^[23]
Where the original plant species and their associated animal species have been lost, a very important artificial intervention in the restoration of degraded ecosystems is to provide seed sources. There are various ways to do this. Aerial seeding has often been used in subtropical regions. Dibbling certain local pioneer tree species is also very effective.
To restore degraded ecosystems to the respective climax ecosystem sometimes requires tree species of the climax community, but sometimes mid-succession tree species must be chosen first. Where vegetation is at an early successional stage, coloniser plants and pioneer trees may be required. This depends on local conditions. Pioneer species can generally be grown in the open, whereas species characteristic of later successional stages may require shade.
Species that would not recolonise or germinate without assistance, especially those which perform important roles, are prime targets for cultivation. To some extent, however, nurseries must be opportunistic and maintain collections of seedlings of whatever species become available in given years. Oaks for instance have heavy 'mast' years only periodically. Some species may be better collected as wild seedlings from the forest floor, rather than reared from seeds artificially. Seedlings should be reared in good soils that have been mixed with small amounts of the wild forest soils from which the seeds were collected, to ensure the correct mycorrhizae are available. The more component species of the original forests can be raised, the better. Complexity brings greater ecological robustness and health, reduces the risks of disease or pest problems, and reduces dependence on further human intervention in future management.
Measures taken in the practical work also vary according to different soil conditions. Sometimes seeds can be sowed directly, and sometimes the soil needs to be improved first. Animals should also be considered in the provision of seed sources in degraded ecosystems. A healthy ecosystem should comprise all kinds of native fauna. Since most animals can move, the introduction of some of them requires no artificial aids. However, many animals have limited dispersal ability, and may not be able to recolonise from intact forest ecosystems. For some of these, such as earthworms, active translocation of native species may have great benefits to the resumption of ecological functioning.
To promote the healthy growth, development and reproduction of some species, it may be necessary to restrain the growth and development of certain others. Where grass is densely established, it may prove difficult to get natural tree regeneration to occur. Whether or not this objective is explicit, restraint is often used in practice. The most typical example is regular cultivation in the restoration of forests. In the early stage of the natural restoration of degraded ecosystems, artificial fostering can boost the regeneration and restoration process.
In areas where temperate bamboos have formed thickets, it will be necessary to physically cut the thickets away in swathes or holes to allow interplanting of young trees. Considerable labour will be required to clean around the trees and prevent choking by bamboo until such time as the trees are taller than the surrounding bamboos.
Close planting of trees may control the development of herb layer. This may mean rearing trees for longer than usual before planting into the field. In tropical regions many fast-growing tree species can be planted by staking or planting saplings of 1-2 metres height into the soil at the beginning of the rainy season.
Some species require some opening of the herb layer to allow seedling survival. Other species need a preliminary shade layer before seedlings can survive.
Wind is a powerful force that is used in nature for reforestation of bare areas by pioneer plants and some climax trees (e.g. conifers, Acer, Populus etc.). In regions where such native wind-dispersed trees are lacking, recolonisation can be enhanced by planting these species along the windward margins of target areas. Some species of herbs and shrubs can distribute their seeds in wind by as much as one km. Most conifer species are only able to spread seeds for about 100 m. In the latter case the use of strip planting can achieve a total forest cover at a fraction of the cost of total planting.
It is wise to leave adequate parent trees in managed forests to act as seed sources to promote reforestation. Similarly it is a useful technique to initially establish tree cover at the heads of valleys and along ridges from where natural wind dispersal of seed and the direction of gravity will allow the natural seeding of lands below.
In temperate zones, closure of land must often be accompanied by active coppicing of trees and especially of budding shrubs. If all the budding stems are left intact, nutrients will be dispersed, which results in slow growth of main stems and hinders the development of forests. In addition, there may be intense competition among shrubs as well as between shrubs. In this situation, coppicing may accelerate forest regeneration. It is more effective to prune branches and retain 2-3 strong budding stems in each clump. This method has been successfully employed in the mountainous regions of Beijing.
Birch woods usually come from degraded cool-temperate coniferous forests. When there is a source of spruce seeds, seedlings and young trees of spruce often grow in birch woods; otherwise there are only birches. To restore various spruce woods, strips of birch can be clear-cut and spruces interposed. As the spruces grow, mixed forests of birch and spruce will first come into being. Then birches will be replaced gradually by spruces. Spruce woods usually cover small areas, and clear-felled areas are more likely to be invaded by larches. If there are seedlings or young trees of spruces in the larch woods, they may form mixed forest or return to spruce woods gradually.
Because of different densities of Korean pines, mixed coniferous-broadleaf forests in the mid-temperate zone can mature to various mixed broadleaf forests, composed of Manchurian ash, Manchurian walnut, purple linden, painted maple, etc. Such types of broadleaf forests retain some primary microhabitats. Without seed sources of Korean pine, it is difficult to restore mixed Korean pine-broadleaf forests. Here, planting conifers and protecting broadleaf is an appropriate intervention to accelerate succession. Strips can be cut in broadleaf forests, and Korean pines interposed; mixed Korean pine/broadleaf forests will gradually develop. This measure has been widely implemented in the northeastern forest regions, with some success.
Measures to accelerate natural grassland habitats must first distinguish among the important grassland types in China. One important distinction is between grasslands dominated by annual processes (with natural seeding occurring each year) and those which are characteristically perennial in nature. Another distinction is the fragility of the ecosystem.
Most Chinese grasslands are extremely fragile. Those in arid areas are comprised of grasses that are keenly adapted to harsh conditions, and these localities are also unlikely to possess sustainable inputs such as water to promote increased growth. While some remedial activities may serve to start natural regeneration, it is particularly important to engage in these activities with a clear idea that the goal is to achieve a climax community of native grasses and forbs.
The high alpine grasslands of the Qinghai-Xizang plateau present another problem. These grasslands grow in an environment where the annual temperature is below freezing; they are primarily temperature-limited. In addition, many of the alpine grasslands are old and comprised of thick sod layers of roots, binding the soil and prohibiting erosion. These grasslands are difficult to re-establish, and all efforts should be made to leave them intact. In many places these alpine grasslands have been ploughed under for agriculture. At these altitudes many of these agricultural attempts have failed, and now the land is being replanted with annual grass seeds in an attempt to improve productivity for livestock. Most of these efforts are also failing, as wind erosion has depleted the top-soil layer. It may take physically planting sets of perennial Kobresia mats to begin the restoration process to the natural climax state. Thus, all efforts should be made to not plough or physically degrade these grasslands.
In certain circumstances specific land treatment measures may be used to improve and restore grasslands. In those cases where the soil has become highly compacted, deep or surface plowing may enhance productivity. Applications of fertilizer or water may initially be an important measure to speed up recovery and increase productivity - but these activities are expensive and unsustainable on the scale of the large grasslands in China. In annual grasslands selective controlled burning can have positive effects on degraded grassland. The effect of fire on grassland varies with the plant species and can promote or suppress the growth and development of different populations and change the quality and value of grasslands^[24],^[25]. Autumn fires in Chinese leymus grasslands have more prominent effect on increasing the yield of grasses than spring fires.

Recommendations

• Degraded soil may need active remediation. Restoration of soil nutrient balance can be achieved through nitrogen-fixing plants and organic fertilizers. Soil erosion can be reduced by maintaining dense plant cover and by preventing the panning of soils.
  
• Species that would not recolonize or germinate without assistance, especially those which perform important roles, are prime targets for cultivation.
  
• Some management of dense vegetation may be necessary to allow smaller plants to grow. For example, in areas where temperate bamboos have formed thickets, it will be necessary to physically cut the thickets away in swathes or holes to allow interplanting of young trees.
  
• For multiple-stem shrub species, coppicing may help accelerate canopy development.
  
• Animals which function as pollinators or dispersers of key species should be protected. Where necessary, they should be reintroduced, following the IUCN Reintroduction Guidelines^[26].
  
• Further conversion of natural grasslands into farmland should be strictly prohibited.
  
• Most grasslands in China exist in harsh conditions which make re-vegetation particularly difficult - arid and alpine grasslands should be left intact as any efforts to restore them to their climax state will take a long time and be expensive.
  
• As most remedial methods for restoring grasslands require high input, which cannot be sustainable, efforts must be put into place to reduce the rate at which natural grasslands are being destroyed; agents responsible for the loss of grasslands must be identified and their actions reversed.

9. Zone according to major land-use objective
Not all land areas will allow priority to be given to ecosystem restoration. Most of China's land is under pressure for resources including food, timber, fodder and fuel. Proper zonation will be needed to take account of the priorities and conditions.

1) Ecological zones: In the areas most critical for adjusting water flow and improving water quality in the main rivers of China, and most important to biodiversity conservation, the Government must provide reasonable long-term support (at least 30 years) to relieve the pressure for local economic returns. At present, the funds and food provided through the programme to return farmland to forest or grassland are generally enough for local people to survive. The important thing remaining is for Government to guarantee long-term support. In areas with steep slopes and high erosion, natural restoration of vegetation should be a high priority, and some degree of active human help will often be required. Until now, no economic plantation has been proved to provide adequate ecological functions. Therefore, establishment of economic tree/grass plantations in these areas should not be allowed, and no economic returns should be expected for at least 20 to 30 years. Only a limited area of existing highly-productive agricultural land on the lower slopes should remain, and no new areas should be turned into agricultural land. Local people should be engaged in taking care of the development of natural vegetation, such as collecting seeds of local species instead of spending money on nurseries of economic or alien species. Spare labour in these areas should be shifted to small towns, industries, ecotourism and other sectors.

2) Economic zones: In areas that are far enough from rivers and intact natural ecosystems, economic plantations could occupy much of the area. However, patches of natural vegetation, which serve as barriers for controlling pests and fires, as traps for eroded soil, and as refuges for local biodiversity, should be restored. Establishment of connections (corridors) between existing patches must have a high priority. The economic species should mostly be local species and each species should not cover too large an area. Farmers should be given access to information and training regarding appropriate agroforestry techniques. Agroforestry generally involves interplanting taller economic plants (trees, shrubs, bamboo, etc.) with shorter annual crops, ensuring more efficient use of sunlight, soil nutrients and water. Even in the economic zone, incentives should be sought to encourage the growth of plants with high capacity to stabilise and enrich soil, to enhance soil water retention and to improve carbon sequestration from the atmosphere, while meeting market demands. In these areas, local people would need at least five years' support from the government, and then could rely on the land to produce sustainable revenue.

3) Buffer zones: In areas between the ecological zones and the economic zones, there should be a gradation in the proportion of natural vegetation, with a higher proportion closer to the ecological zone. The level of government support should be related to the ratio of ecological to economic function.
Vegetation in the more ecological zones should approximate the local climax vegetation, or mimic natural succession toward the climax. In the more economic zones, local communities could be encouraged to help set the land-use objectives. The appropriate farming system will depend on a thorough assessment of local resources, soil and climatic conditions, potentially available inputs, and the farmer's ability to take risks^[27]. Where resources and inputs are scarce and crop failure might carry high personal cost, diverse cropping systems dominated by long-lived plants are likely to be more appropriate. The use of several systems appropriate for specific areas and dependent on each farmer's constraints would result in a landscape mosaic of agroecosystems in various stages of succession.

Recommendations

• Land should be zoned for management according to the most important functions. Due to the huge economic, environmental and social value of upper catchment slopes, management for ecosystem restoration should have priority in these.
  
• The forest areas of China should be zoned into various categories, including at least:
  Protection Forests (two main types would be biodiversity and hydrological functions)
  Natural Restoration Forests (degraded scrub or forest allowed to regenerate naturally by forest closure)
  Sustainable Production Forests (forests specifically managed or planted for the sustained production of timber and other products)
  Artificial Restoration Forests (man-made forests including shelter belt forests and desert-checking 'green walls')
  
• Zoning should be done at a national level to designate certain forest regions of China by their primary objectives; but also at the local level to plan detailed management, based on precise local conditions of floristic zone, altitude, climate, aspect, slope, soil and current forest condition.
  
• Identification of areas to be developed for timber production should be based on land capability and local ecological function requirements and not driven by demand projections. The ecological restoration needs should not be compromised by production wishes.
  
• Economic forests should be interspersed with ecological forests, to provide insurance against fire and pest outbreaks, and to help conserve regional biodiversity. Timber production should be within mosaics of different land uses.

10. Ensure that forestry, agriculture, grazing and collecting are sustainable
China has a massive demand for timber. Some land must be devoted to timber production, both for economic reasons and to reduce the immense pressure on forests elsewhere in Asia.
Those areas that are not too steep or susceptible to erosion may be zoned for sustainable timber production. In such zones, the emphasis will be more on generating a sustainable supply of timber or other non-wood products (see Annex A.5.1), rather than restoring the original forest composition.
The Forestry Departments at the provincial level already have many proven systems for sustainable forestry. However, in an effort to be practical and efficient, many systems have relied on the use of monoculture stands and introduction of exotic species (termed economic forests). In the short term some of these approaches may be viable, provided the risks of invasion are minimal. A long-term perspective would call for these efforts to be revised according to the principles outlined in this position paper.
In planning sustainable economic forests and grasslands, efforts should be made to maximise benefits by offering multiple services and to minimise environmental threats of forest fire, diseases, pests and soil erosion. Well-designed plantations can serve as recreation and picnic areas, produce timber as well as provide habitat for wildlife and also give benefits of soil and water conservation.
Ecological forests, with high ecological integrity and biodiversity, should be interspersed within a matrix of such economic forests. These ecological forests can serve as buffers preventing outbreaks of pests, disease and fire. In addition, they also provide spaces for the growth, feeding and breeding of plants and animals, which will greatly improve the level of local biodiversity.
Some native trees can be restored naturally on clear-felled areas. Thus the priority is to preserve their seed sources. Some mother trees should always be retained on logged areas, so as to harness the force of natural regeneration to restore forests. In the clear-felling of mature forests, seedlings of various tree species should be well protected. Thus forests could regenerate quickly with or without intervention, and grow up in 20-30 years. This is very significant in promoting natural restoration of forests.

Clear-felling and selective logging may have very different impacts on biodiversity. In one study28, a broadleaf Korean pine forest plot had 34 bird species, of which 33 were forest specialists. Upon clear-felling all these forest birds were lost. After two years, it had only eight bird species, all characteristic of scrub and grassland. After 50 years of regeneration, the plot had 15 bird species, including only 12 forest birds - less than half the number in the original forests. By comparison, in another broadleaf Korean-pine plot 40% of trees were selectively removed. One year later the plot still had 24 forest birds. Impacts on plants are similar. In one study, in the second year after selective logging, 90% of the understorey species were the same as in original forest, with only changes in relative abundance. However, in clear-cut forest it took 30 to 40 years to restore the original understorey species composition.
One approach for restoring degraded ecosystems in large areas is to combine forestry and agriculture. Where it is unfeasible to restore primary ecosystems, such agro-ecosystems can stop the degradation of soils, improve environmental conditions and increase economic benefits. Agro-ecosystems should have low requirements for nitrogen and phosphorus fertilizers; a good ability to use available nutrients, light and water; protection from pests; and low risk^[27]. In agricultural systems, many of these beneficial roles are served by establishing a fallow period. A fallow period (whether seasonal or more long-lasting) restores soil fertility, by allowing nitrogen fixation^[29], recovery of soil invertebrates^[30], and mycorrhizal fungi^[31]. It also causes a decline in agricultural pests^[27].
Besides various kinds of poplar and acacia, fruit trees are often the main components of agro-forestry systems in China. Since the space between young fruit trees is quite large, herbaceous plants such as vegetables, medicinal materials and cereals can be intercropped. This results in a multi-layered floral community, which helps to conserve water and soil and increase economic incomes.

Different zones place different climatic limitations on agriculture^[32]. Extremely cold lands have no potential for agriculture. Lands that are arid or infertile can be used successfully, although the cost of compensating for environmental limitations increases exponentially with increasing abiotic stress. Grazing animals (which act as trophic buffers between people and environment) have proven successful in dry and infertile environments, provided that stocking rates are low enough to prevent overgrazing. On the Qinghai-Xizang plateau native ungulates and livestock have co-existed for over 2,000 years; clearly it is possible for economic animal husbandry to thrive in this environment. The solution to the situation today, with much of the plateau overgrazed and degraded, is to search for remedies that can balance the grazing intensity of livestock (and native ungulates) with the condition of the grassland. Because this pasture land is unlike any other major grazing ecosystem in the world (by being alpine and temperature limited), it is essential to develop local solutions rather than to attempt to import ideas. The key to past sustainable use of this environment was the mobility of livestock and the flexible use of winter and summer ranges. Most current attempts to manage this ecosystem adopt mechanisms to ensure the livestock are less mobile (fences, etc.), a process that appears to be unsustainable. The vast knowledge of animal husbandry developed over the past hundreds of years should be incorporated into management decisions effecting herd mobility.
Humid tropical lowlands are environments of low abiotic stress but overwhelming biotic intricacy. Here it pays to imitate natural ecosystems rather than struggle to impose simplicity on ecosystems that are inherently complex.

Recommendations

• Harvests in forestry and agriculture should not deplete the system's capacity for renewal. Economic disincentives to sustainable practise should be identified and removed.
  
• Routine monitoring should be in place in all systems, to ensure impacts of forestry, agriculture, grazing and collecting do not exceed prescribed limits.
  
• Those agroforestry systems with sustainable features of natural ecosystems should be favoured.
  
• Controlled grazing may be possible in regenerating ecosystems, but only by non-browsing animals. Browsers, such as goats, destroy shrubs and tree seedlings as well as herbs.
  
• Methods should be enacted to ensure that a flexible form of livestock herd mobility is used on the alpine grasslands, to minimize the effects of overgrazing and to utilize the grassland resources sustainably.

11. Protect vegetation against fire
Fire is of greatest danger in the relatively dry northern forests and the very hot southern forests, and less important in the humid central forests. Forest fires were formerly commonplace, causing large annual losses to the forestry sector and human life (see Annex A.2.2).
The tragic Greater Xinggan Mountains forest-fire in 1987 burned 11,000 km, and destroyed 3,700 km of forest beyond regeneration. This was the turning point of China's forest-fire control programme. Since then, the Central Party, State Council and local governments at all levels have attached great importance to forest-fire prevention work.
China's current forest-fire prevention strategy includes monitoring with a combination of ground, air and satellite remote sensing. NOAA series weather satellites are used in current monitoring of forest-fire, but monitoring precision still needs improvement.
According to natural and social conditions and level of fire risk, China has been divided into different fire risk divisions. Monitoring is based on the degree of fire risk. Recently, the government has formally promulgated Weather Degrees of National Forest-fire Risk. But a national forecast system of fire risk has still not been established. It is imperative to reinforce the basic research and set up a national system of fire risk warnings as soon as possible.
Since almost all fires in China are man-made, it is especially important to properly manage fire sources. Management of combustible fuel and controlled fires are among the active measures used to reduce the risk and impact of wild fire. In some kinds of forests, small controlled fires are valuable in reducing fuel build-up. Controlled fires have other functions, including control of the spread of some forest plant diseases and insect pests, and increasing the availability of minerals and nutrients such as nitrogen, phosphorus and oxygen, to trees.
Growing belts of fire-resistant forest is essential to prevent the spread of forest-fire and reduce fire damage. Besides preventing and closing off fire, this procedure has many other benefits and functions, including the increase in diversity of forest structure, prevention of soil erosion, increase in soil fertility, establishment of a buffer against the invasion of plant diseases and insect pests, enhancement of an intensive level of forest management, and generation of a higher standard of economic income. Construction of fireproof forest belts has already been included in the 21st Century Agenda of China.
The construction of forest-fire prevention facilities has developed rapidly since 1988. This includes improved forecasting facilities, watchtowers, air surveillance, communication and fire breaks, as well as personnel training and mechanization of fire fighting. Upgrading of fire protection equipment and enhancing the capabilities of the current equipment is needed.
Improved capacity of fire fighters through improved training and equipment will increase effectiveness and efficiency. Professionals and the public should join together in the battle against forest-fires. The general public should be mobilised through education on fire prevention.

Recommendations

• Establish a national forecast system of fire risk, and reinforce fire-risk research in susceptible areas.
  
• Where forest fires are prevalent special fire prevention and fire-fighting measures should be introduced, such as cutting firebreaks or planting green barriers of non-inflammable species.
  
• In high-risk forests, combustible fuel could be managed, by controlled fires or harvesting of fuelwood, to prevent the build-up of dangerous levels of dry inflammable vegetation in the undergrowth.
  
• Village fire-watch and fire-fighting brigades should be organised and trained to detect and extinguish fires at the earliest instance.
  
• Special care should be taken in selecting tree species for planting that are less fire-prone. Generally speaking conifers and Australian exotics such as Eucalyptus and Acacia are extremely inflammable.
  
• Climate monitoring combined with maps of fire-proneness of vegetation should be used to provide an early warning of fire likelihood. These should be combined with on the spot measurements of humidity, wind speed and dryness of vegetation.
  
• Strict regulations should be imposed on the burning of vegetation on adjacent farmlands.
  
• The width of firebreaks should be fixed according to combustibility of trees and local wind speed. Care should be taken to select tree species that are less fire-prone for planting along the fire protection belt. In seasons when fire is prone to happen, professional fire-fighting brigades should be organised to guard the key sectors.
  
• In seasonally dry climates fires on grassy slopes may kill off any young tree seedlings that have established themselves. In such cases, efforts must be made to shade the grasses by planting taller trees and saplings. This may mean rearing trees for longer than usual before planting into the field or use of tall cuttings of secondary forest species that can grow from stakes.

12. Protect regenerating ecosystems against pests, diseases and invasive species/genes
During vegetation restoration, plant diseases and animal pests occur frequently owing to the impaired ecosystem structure and function, including the absence of natural enemies and plant chemical defences (see Annex A.4). Effective control measures against biological disasters are therefore key.
Chemical, biological and physical methods are sometimes employed. Chemical pesticides are problematic, as they tend to pollute the soil and water, and threaten non-target life forms such as birds and mammals. Many of these are themselves natural enemies of the pests or perform other important ecological roles, such that their loss impedes vegetation restoration. Any chemicals used should be as specific as possible, non-persistent, and non-accumulative in the food chain. Persistent organic pollutants, including organochlorine compounds, should never be used. Biological methods try to control pests and diseases using natural enemies or biological pesticides. These are less likely to endanger soil or water quality, but these too may impact non-target life forms. Physical methods are unlikely to damage non-target organisms, but are difficult to employ on a large scale.
Integrated pest management is based on the characteristics of the diseases and pests, the existing control technology and the principles of environmental protection. For instance, in the event of a devastating pest outbreak, a chemical pesticide might first be applied to quickly control its spread and then a biological method used to achieve a more lasting effect. The process will use the minimum chemical dosage to achieve the desired result.
A major limitation on integrated pest management is knowledge of the ecological roles of different organisms in natural and regenerating ecosystems. Some "pest control" programmes are entirely misdirected, due to a lack of ecological understanding. Voles, pikas and marmots in alpine grasslands are sometimes poisoned as pests. However, these mammals are essential elements in healthy grasslands providing aeration and drainage, which increases the water-holding capacity of the land. The burrows are used for shelter by many breeding birds, while the mammals are vital in the food chain for many important predators33. Poisoning of rodents and pikas contributes to severe degradation of the grasslands, severe loss of biodiversity, the reduced water sponge of grasslands and desertification.
Invasive species are now recognised to be globally a bigger cause of local extinctions than habitat destruction. They have been shown to be a major threat to both natural and regenerating ecosystems, while genetically modified organisms contain unstable genes which can have unpredictable impacts on local biodiversity.
The species that are most likely to become invasives are fast-spreading hardy colonists. These are exactly the properties that managers look for in a vegetation species that can quickly cover bare ground. The temptation to use dangerous exotic species is great, and indeed China is already importing such risky species for revegetation purposes. Preventing invasions is more economical than control.

Recommendations:

• Restrictions should be placed and enforced on the import and movement of species known to be invasive.
  
• Risk evaluation procedures should be formulated for all exotic species. Planting of exotic species or of genetically modified organisms should not be allowed unless they are proven harmless to ecosystem integrity and biodiversity conservation.
  
• Release of any species to the wild should strictly adhere to IUCN Reintroduction guidelines^[26].
  
• A monitoring system should be established, to collect, research and disseminate information on invasive species and their impacts.
  
• Public awareness needs to be improved, and the effective prevention measures well publicised.
  
• Any pest control programmes should be based on an adequate understanding of the ecology of the species.
  
• Research should be conducted into the ecological roles of native fauna and flora. In particular, the ecology of possible pests and of potential natural enemies of pests should receive concerted study.
  
• Pest management should be integrated. Potentially invasive species should be controlled manually if possible, without recourse to chemical or biological measures. Measures such as inundation, fire or shade can be also applied within a limited range.
  
• Vegetation should be managed for heterogeneity and structural complexity appropriate to the climax ecosystem. This should reduce the risk of the spread of pests or invasive organisms, and encourage biological diversity.
  
• When an invasive species has spread widely and caused serious damage to the local economy, human health or natural ecosystems, alternative control measures should be considered. Chemical pesticides should be avoided if possible. Biological control using non-native species should not be implemented unless the control agent has been shown to be benign.
  
• Following local elimination of invasive species, native vegetation should be re-established as quickly as possible. These areas should be closely monitored to prevent re-invasion.

13. Monitor and research biodiversity and ecosystem integrity
There are various reasons to monitor integrity and diversity of natural and regenerating ecosystems. One is to ensure that restoration and management are progressing towards the desired objectives, such as environmental services, biodiversity conservation and economic viability. Monitoring is needed to prescribe appropriate management changes.
Where soil and water conservation are major objectives, monitoring regimes should obviously cover the effectiveness of vegetation in meeting them. This might involve measuring water runoff in wet and dry seasons, and sediment loading of the water. The effective rainfall coefficient measures the rainfall filtering into the middle and lower layers of the soil, and therefore its water absorption capacity.
Monitoring other aspects of ecological integrity is also important. Most functions of vegetation are largely dependent on biomass and biodiversity, which are themselves dependent on the maturity and complexity of the ecosystem (see Annex A). More complex natural ecosystems are more likely to modulate water flow, prevent erosion and resist pest outbreaks. They also harbour more species of economic value. China's various grasslands support 7,000 species of higher plants; many of these are important species for food, medicine and fodder.
There are no universally agreed indicators for monitoring biodiversity and ecological integrity. All-encompassing indices of "species diversity" are not necessarily meaningful, as they are influenced by many factors, and should be used only with caution^[34]. The choice of indicators will depend on the objectives of management, and on logistic constraints. Possible indicators include structural attributes of the ecosystem (vertical structure of the vegetation, foliage cover of different strata, above-ground biomass), dominant species composition in different storeys, and evenness of species abundance. Measuring the ratio of annual to perennial plants gives one measure of ecosystem succession; perennial plants play a major role in most mature terrestrial ecosystems.
Monitoring of ecological processes is another approach. Measures of nutrient and energy flow, decomposition rates, productivity etc. may give useful information on ecosystem function. In subtropical soil, organic carbon content typically has a positive relationship with above-ground biomass and water-holding capacity. Indirect indicators of ecological processes can be made by estimating the abundance or biomass of various ecological guilds (decomposers, grazers, browsers, seed-dispersers, predators etc.), in different ecological strata or habitats. Certain species are of disproportionate importance in the functioning of ecosystems, influencing the overall species composition and richness. Keystone species might be animals, plants or micro-organisms. Populations of such keystone species (such as pika on the Qingzang plateau grasslands) should be monitored by an appropriate index of abundance or density. Healthy populations of these keystone species are indicators of the health of the ecosystems.
It may also be desirable to monitor populations of particular species of conservation concern. Monitoring of endangered species is of great national importance, and required under the Convention on Biological Diversity. Species listed by IUCN^[35],^[86] as globally threatened are a priority for such monitoring. Wild relatives of economically important species are another priority. The vast majority of species (e.g. insects, fungi, microorganisms) are undescribed by science, and many species are likely to be endangered without our knowledge. Research on taxonomy and community ecology should be strengthened on these groups. At the same time, monitoring of the overall diversity of insect communities should be conducted, since diverse ecosystems are likely to support more restricted species.
Effective land use is limited by our poor understanding of the ecological functions and relationships of natural ecosystems. Governmental policies should encourage relevant studies on ecosystem restoration and develop adequate knowledge bases for vegetation restoration using native species.

Recommendations

• Develop a monitoring programme for all vegetated areas according to land-use objectives and human resources available. Include appropriate indicators of ecosystem integrity and biological diversity.
  
• For natural ecosystems, include measures of increasing or maintaining high species diversity and boosting relative abundance of endangered species.
  
• For areas where natural vegetation is regenerating monitor changes in dominant species and relative abundance of endangered species.
  
• For areas with economic objectives monitor ecosystem functionality and relative abundance of endangered species.
  
• Study and assess the impact of different human activities on the ecological indicators measured.
  
• Conduct further research on the influence of vegetation on water conservation, climate, soil conservation and erosion control.
  
• Promote research on ecological relationships, interactions, succession and limits on regeneration of local species and ecosystems.
  
• Research and monitor the impact of alien invasive species on local ecology, biodiversity and the economy.
  These research and monitoring needs can only be met by mobilizing expertise from scientific institutions and universities to work in close collaboration with mainline bureaus responsible for land management.

14. Manage for threatened species
In addition to the general measures aimed at preserving the ecological processes and guilds of natural ecosystems, specific management may be required for species of special concern. The need for such intervention will be evident from existing information and the results of monitoring.
Some threatened species have specific ecological requirements. For instance a tropical frugivore such as a gibbon (Hylobates sp.) requires a closed canopy so that it can travel through its home range, but also needs a variety of different fruiting trees to feed on. The phenology of those trees must be staggered, as in an extensive natural forest, so that the gibbon can find some fruits in all months of the year. This is only possible in a fairly complex, species rich, evergreen forest. Gibbons cannot live in monocultures or forests without sweet fruits such as figs. Similarly, carnivores cannot live in small isolated forests because there is not a large enough prey base to support a viable population. So the size of habitat blocks and the species mix are very important considerations in planting forests that can act as wildlife habitat.
The connectivity principle is very important. Corridors may need to be planted or preserved to allow gene flow between semi-isolated populations to give those populations viability and avoid inbreeding and demographic crashes that occur in small and isolated populations.
Species need to live in populations of viable size. Populations that are too small suffer from genetic inbreeding, demographic events (e.g. the only male of breeding age being sterile), chance extinctions from disease or other causes, edge effects and other risks. By maintaining habitat linkages or corridors, wildlife populations remain connected to larger panmictic (randomly interbreeding) breeding units and their survival chances are higher. The larger the connected unit the more different species can co-exist in equilibrium. Equilibrium is reached when the local rate of extinction of species is balanced by the local rate of colonisation by new species.

• It is important to design as much connectivity as possible into the forest cover. Corridors should be planted or preserved to allow genetic exchange between wildlife populations and allow species to colonise restored areas.
  
• Most threatened species will be adequately served by an ecosystem perspective. In selecting species for particular management, preference should be given to 'umbrella species' for which management would also serve the ecosystem as a whole.

15. Involve the community and promote public awareness
The need and scope for public participation
A national programme of reforestation and protection of natural vegetation cannot succeed without a high level of support and input from local communities. It is equally clear that the degree of protection or plantation effort and respect for vegetation by local communities will depend very much on the stake they have in the vegetation cover that results.
It has proved possible to encourage local farmers to plant trees on bare hillsides and even on their own fields so long as the farmers themselves retain ownership of such trees and future products. It is much less easy to encourage farmers to invest in planting of ecological forests or non-commercial vegetation, and it has proved difficult to maintain high levels of public involvement in fire prevention and emergency fire-fighting. Under the experimentation phase of the programme of returning steep farmlands to tree or grass cover, one of the main shortcomings has been the consistent tendency of farmers to opt for economic rather than ecological forests. This has thwarted Government plans that 70% of forested lands should be of the ecological type.

This problem raises questions about the way in which farmers can be rewarded for inputs into public (rather than private) projects, and also about the degree to which they should be involved in planning the details of vegetation restoration.
On the positive side, farmers and foresters have a great deal of local knowledge about the exact potential of the lands in question. They know which species will naturally re-colonise the area, which native species are locally available as seeds or seedlings, and which secondary forest species can be propagated by the planting, for example, of trimmed stakes, as used regularly for planting living fences. Such knowledge should certainly be tapped when planning the specific species that should be planted or protected in each site.
However, sole reliance on local people in this planning process will result in economic decisions being made to benefit only the immediate local community. These decisions may not be in the best interests of the larger number of downstream stakeholders, who depend mainly on the environmental functions of the planted areas. It is the role of Government to represent this wider interest in negotiating with local farmers. Even provincial governments may have different interests to the downstream users who are usually the principal beneficiaries of good water control.
The local community should be involved in all aspects of vegetation restoration: covering denuded ground; protecting natural and regenerating ecosystems from encroachment and utilisation; harvesting seeds and cultivating native species; coppicing, etc. to accelerate regeneration; guarding against fire; guarding against invasives, diseases etc.; and monitoring biodiversity and ecological processes.
Social forestry
Social forestry, also known as community or participatory forestry, is a model of forest management and organisation involving local farmers and multiple objectives: to directly benefit the farmers; to improve the rural ecological environment; and to facilitate integrated, harmonious and sustainable development in rural areas^[36]. It is contrasted against conventional forestry (sometimes misnamed "traditional" forestry), which has the single objective of timber extraction. Social forestry is in fact far closer to the traditional forestry practised by indigenous people, in China and elsewhere. Today, social forestry can draw on both the knowledge and perspectives of local communities, and on centralised and systematised scientific knowledge, to form an integrated, sustainable and adaptable approach to forest management.

As noted above, a key issue in social forestry is land tenure. Clarifying and defining tenure can motivate farmers to make informed decisions about forest management, taking into account the balance between benefits and responsibilities. Where farmers are being asked to change their practises, it is essential to provide support for them over a reasonable time-scale. The management priorities, and the support required, might take different forms in different regions, according to erosion risk and other factors.
Different models of social forestry have been attempted in different regions, and these provide many examples of successful and unsuccessful experiments. In general, the most promising models are those which have stood the test of time.

Employment opportunities in vegetation restoration
Some existing agricultural labour can be absorbed into new activities such as forest protection and planting and into restoring or creating terraces on steep mountainsides. Local farmers can be employed in maintaining tree nurseries, and collecting the requisite tree seeds from local forests. For these and other new activities, special training will be needed. Special recruitment priority could be given to farmers that have abandoned and reforested their fields.
Sustainable forestry offers a range of new economic opportunities for local communities, some of which call for short-term Government investment. Some farmers should also be paid to maintain valuable germ-plasm by continuing to cultivate local variants of crops and vegetables or grow wild relatives of domestic varieties. This is a vital function even if it does not have immediate commercial value, and thus will need subsidisation.

Ecotourism
Due to the numerous benefits provided by intact natural ecosystems, non-consumptive uses should be promoted where possible. One possibility is ecotourism (see Annex A.5.2). At the core of tourism is the human inclination for a change from everyday surroundings, and curiosity about the world. As urban people have increasing amounts of leisure time and money, there is a growing demand for tourist destinations. Tourism employed 6.5% of the global workforce in the early 1990s^[37]. Nature tourism, which satisfies fundamental human needs for peace, beauty and diversity, has grown steadily and today has become a multi-billion dollar industry worldwide.
Due to the resultant pressures on natural areas, the tourism industry has developed codes of ethics. Ecotourism is defined, by The International Ecotourism Society, as "responsible travel to natural areas that conserves the environment and improves the well-being of local people". Since it does not inherently involve the removal of products, it is sometimes considered the most benign economic use of natural habitats^[38].
Ecotourism is economically beneficial in four ways: it is a growth industry; the market comes to the resource; tourism helps diversify the economy; and it stimulates economic growth in rural areas^[39]. However, in China and many other countries, it is usually entrepreneurs far away from the destination who put up the investments for tourism facilities - transport, hotels, restaurants, etc. - and it is these outside investors who reap the lucrative profits available. Very little of the tourists' money finds its way into the local economy at the point of destination. It is also clear that tourism developments often fail to help conserve the natural environment, but rather damage and put additional pressures upon it.
Outdoor recreation and poorly-managed nature tourism can also have a wide range of negative environmental impacts^[40]. Besides large-scale associated developments, such as roads and buildings, these include impacts from large numbers of visitors in vehicles or on foot. Impacts may affect soils (compaction, altered hydrology, litter disturbance, loss of soil organisms, erosion and soil deposition), plants (loss of plant cover, spread of exotic flora along paths, collection, trampling damage) and animals (through interruption of tranquillity, habitat changes, and injury or death).
Better development of ecotourism could certainly be a very positive force in helping to justify and pay for the conservation and rehabilitation of natural vegetation, and the process is self-reinforcing as the better the environment becomes, the more attractive it will be for further ecotourism development. The revenues from ecotourism can meet many of the social needs locally, increasing the stake of local communities in the state of their natural environment^[41].
The general guidelines below include proposals for improved development of ecotourism. More detailed guidelines are given by Ceballos-Lascuráin^[37], including suggestions on marketing (establishing inventories of attractions, targeting visitors, evaluating appeal, and promotion), education and interpretation (topics and target groups, interpretive media), ecotourism training, and planning physical facilities (site plans, design techniques and materials, design for the disabled, refuse treatment, transport and circulation, road and path design, nature trails).

Recommendations

• The local community is key to the success of efforts to restore vegetation. They should be fully informed of the objectives, involved in the process, and included in the economic benefits of restoration. It is most important to ensure that if farmers are to be denied direct harvesting rights of the vegetation cover concerned, they are given adequate compensation and incentive for their involvement.
  
• Given the land-use objectives of regenerating vegetation, appropriate models should be developed based on successful examples of sustainable social forestry. Conflicts between objectives of residents and the government should be discussed openly, and resolved in advance where possible. Means to encourage dedication to the restoration process should be explored.
  
• Restoring and protecting vegetation will create various employment opportunities. Local people, especially those from ecological zones or who have abandoned their farmland, should have the priority to be hired first. Employment opportunities include work as forest guards, foresters, fire protection staff and nurserymen.
  
• In areas being returned to natural vegetation, support for farmers' education should be strengthened, to ensure the nine years of compulsory education.
  
• Ecotourism could be explored as one possible source of sustainable income from recovering vegetation. This should be subject to the guidelines in the position paper.
  
• Tourism developments should not cause any irreversible damage to the nature or ecology of the environment. Strict regulations controlling allowed activities, zoning and numerical limits will need to be applied.
  
• Ecotourism should promote positive environmental ethics, and foster "preferred" behaviour in its participants.
  
• Ecotourism should concentrate on intrinsic rather than extrinsic values. Facilities and services may facilitate the encounter with the intrinsic resource, but should never become attractions in their own right, or detract from the resource. Ecotourism should be oriented around the natural environment.
  
• Ecotourism should be developed in a way that brings genuine economic benefits to those counties and local communities whose protective activities have preserved, and are needed to continue conserving, the biotic, scenic and cultural sites and tourism destinations.
  
• The environment should attain a net benefit, contributing to its sustainability and ecological integrity. Systems should be put in place to establish baseline data for monitoring the ecological and cultural impacts of all ecotourism developments. Benefits can be measured socially, economically, scientifically, managerially, and politically.
  
• Physical structures developed for ecotourism - hotels, lodges, bridges, trails, watchpoints etc. - should be attractive, clean, in harmony with the surrounding scenery and in line with local cultural style. Energy efficiency should be a major consideration, and construction in local natural and preferably degradable materials is strongly preferred.

16. Plan holistically
National development planning
China can afford and should be willing to pay the apparently high costs of these reform programmes. It has been estimated that to apply the current programme of restoring all agricultural lands over 15% slope in China would cost RMB 4 billion per year (US$0.5 billion)^[44]. This would be a low price to pay for the halving of all soil erosion in a country whose GNP is close to RMB 4,000 billion (US$500 billion) and where environmental disasters relating to poor vegetation cover now cost many billions of US dollars per year.

Local development planning
Provincial and local governments have responsibility for assuring the short-term development of communities. Perhaps more importantly, they must ensure that long-term and broader perspectives are not compromised by short-term development. It is sometimes necessary to eat 'bitter medicine' to cure an environmental problem. It is easier for Government to 'prescribe' such medicine than to expect local communities to voluntarily propose such activities. Government must therefore define the goals of re-greening projects whilst allowing local people to refine the fine details.
One weakness of micro-planning within such activities as forest protection, or restoring vegetation on bare lands or steep farmlands, is the lack of integration with wider regional considerations. For instance, the degree to which farmland can be converted to natural vegetation is compromised by the need to provide continuing support for the farmers concerned. Where alternative employment can be offered by new industries in rural townships, it may be relatively easy to get farmers to completely abandon their lands after replanting. In areas with less potential for absorbing displaced farmers or loggers, local government is forced to adopt other schemes for allowing such persons to earn a living on forested lands, e.g. by allowing some levels of harvesting of economic products or levels of sustainable wood harvesting.
Where county governments have the necessary capacity, they should be allowed to develop integrated county-level plans for restoring hydrological function. In such plans they can integrate measures, with greater flexibility, in relation to specific local circumstances. Flexibility should be allowed in the setting of budgets, so that specific local considerations can be properly taken into account. In some steep, remote or dry areas instituting these procedures is certainly more difficult and it will be more expensive to restore vegetation cover. In some counties, restoring vegetation on bare land, or stabilising landslide zones, may be a higher priority than restoring vegetation on steep farm slopes. Currently, sites for restoring vegetation are determined by the feasibility of conversion with the available budget, rather than by priority of treatment of the site. Plans should be flexible enough for budgets to be diverted to the most important tasks in priority areas.
Inter-institutional boundaries should not hamper practices that cut across sectors, e.g. "agriculture" and "forestry". Such boundaries are a potential problem worldwide (see box 13).

Flexibility should also be allowed in applying national policies based on the local ecology. Conditions vary so much across China that in one area a 25% slope may be very dangerous to open up for agriculture, but in another situation, especially if deep terracing is possible, such slopes could be farmed with negligible erosion. It some such situations, investment in terraces may be more effective than conversion to economic tree plantations.
Besides legislation and positive investment, other fiscal mechanisms can be