Chapter 4
Key bio-physical constraints and opportunities

Biophysical factors of production

The factors of the bio-physical environment facing an individual farm household, such as climate, soil type, topography, pests and diseases, will impose biological and physical limitations on plant growth and therefore directly influence the component enterprises within an individual farming system. Agriculture is ultimately based on the exploitation of plants, whether they be crops, trees or grasses. The growth of a particular plant species will depend on the combination of bio-physical conditions at a specific locality. Different crops have different environmental requirements for optimum growth, likewise fuelwood and multi-purpose tree species, pasture grasses and legumes will respond in different ways to variations in the bio-physical conditions, notably differences in climate, soil properties and hydrology. Equally environmental influences affect livestock enterprises either acting directly on the animals eg. temperature and humidity, or indirectly through their effects on the growth of pastures and browse species (Douglas 1992, Young 1984).

The limitations imposed by the combination of bio-physical conditions in a specific locality, in conjunction with the level of management applied, determine the level of production that can be achieved for a specific agricultural enterprise. Comparison of the natural environmental conditions in an area with those required for optimum production of farmers' existing crop, tree and livestock enterprises will determine their ecological suitability and whether the constraints to sustaining and increasing production are in part due to natural causes.

Recognition of variations in land potential

The terms high potential and low potential tend to be used generally to imply the relative value of land for arable purposes. However land potential is only really meaningful in relation to a specific use, e.g. land with potential for maize production does not necessarily have potential for paddy rice, land with high potential for pastoralism may have a low potential for rainfed crop production (Douglas 1994).

An area with high potential is one where the land qualities match the biophysical requirements of the agricultural enterprise for optimum production so that few, if any, of the constraints to increasing production will be due to natural causes. An area with low potential is one where one or more of the factors of the natural environment impose sufficiently severe constraints (e.g. low rainfall, adverse soil properties, steep slopes etc) as to limit, or prevent, production on a sustainable basis.

Within the Asia Pacific Region the term marginal land denotes areas with low potential for the sustainable production of staple food crops, e.g. upland areas with steep slopes and shallow soils, lowland areas with waterlogged soils, or semi-arid and arid areas with soil moisture constraints. In much of the region, population pressure coupled with inequitable access to high-potential land is forcing more rural households into such marginal areas. The use of inappropriate management practices in land with low potential for crop production is leading to considerable land degradation.

Whereas yields can be increased in marginal areas by moisture conservation and use of drought tolerant cultivars, sustainable yields will always be lower than for high potential areas. As a result the potential population supporting capacity (FAO 1982) of marginal areas will be lower than that of high potential areas. Improving agricultural productivity within high potential areas will enable far higher population levels to be supported per unit area than improving productivity within low potential areas. Likewise preventing degradation within high potential areas will give higher returns to resources invested in conservation in terms of "saved yield" (Stocking 1984). Investment in land reform and agricultural development programmes in high potential areas, enabling them to support larger numbers of farm households, could be one way to reduce pressure on adjacent marginal areas.

The above paragraph should not be seen as a justification for directing investment solely to high potential areas. The reality in the Asia Pacific Region is that a significant proportion of the rural population of the region live in many of the low potential, or marginal, areas (some 263 million people, according to Hazell and Garrett [1996]). From a social equity point of view there is a need to invest in such areas so as to ensure that their inhabitants have the means to meet their subsistence needs, either by improved on-farm food production or from on-farm and off/non farm activities. There is also a growing belief that the possibilities for sustainable agricultural intensification in some marginal areas are much greater than originally thought. The key is to improve the productivity of the natural resources and people there with the right investments, technologies, policies and institutions (Ibid.).

Concentrating agricultural development in the high potential areas may ensure national food security. However, it does not solve the problem of how individual households in low potential areas gain access to the food they need. Distribution of food through famine relief or food for work programmes is not a viable long-term solution (Douglas 1994). The United Nations calculates that relief operations cost about US$70 per person in external resources. Investment in less favoured areas could save part of that money by reducing the need for such programmes (Hazell & Garrett 1996).

Soil as a resource

Soil is basic to life. It is the primary means of food production, directly supporting the livelihood of most rural people, and indirectly everyone, in the Asia Pacific region. It is an essential component of terrestrial ecosystems, sustaining their primary producers (all living vegetation) and decomposers (microorganisms, herbivores, carnivores) while providing major sinks for heat energy, nutrients, water and gases (Wild 1993).

The soil resources of landscapes vary widely in their suitability for use. Each soil type has its limitations, and each agro-ecological zone its climatic factors restricting crop growing seasons (FAO 1978). For example in the humid tropics, plant stresses are mostly the result of nutrient deficiencies exacerbated by leaching and surface removal making the soil acidic and nutrient-poor. In contrast, plant-available water capacity is the major handicap in some areas of Asia and the Pacific dependent on monsoon rainfall which occurs only part of the year. Indeed soil degradation may manifest itself in many forms. What often masquerades as drought is nothing more than the reduced ability of soils to retain sufficient water for plant growth between normal gaps in rainstorms (Stocking 1995).

Crop, tree and pasture species have certain specific site preferences for optimal production. Some are very particular and site-specific, others are less demanding and more versatile or adaptable in this regard. For each particular site, the integrated effect of climatic and soil conditions determines the production from a given plant, cultivated or in the case of natural woodland and pastures exploited under a specified management system. The effect of climate, through temperature and moisture supply, is overriding in determining biomass and plant production. The effects of soils and land management are modifiers. In general, the modifying effects of soils and land management become greater, the more the climatic conditions of a given area deviate from the optimum for the growth of the crops, tree or pasture species. As climatic conditions become more adverse, soil and management conditions become more important in determining production.

The basic soil requirements of crops can be characterised in relation to their internal and external properties, are show below (FAO 1978):

Internal requirements

• soil temperature regime
• soil moisture regime
• soil aeration regime
• natural soil fertility regime
• effective soil depth
• soil texture and stoniness
• absence of soil salinity and/or other toxic substances
• other specific properties, e.g. soil tilth

External requirements

• slope and other landform characteristics
• occurrence of flooding (incidence, regularity and depth)
• soil accessibility and trafficability under certain management systems.

Regional climatic constraints

There is considerable variation in the macro-climatic conditions in different parts of the Asia Pacific Region, ranging across the spectrum from arid to humid. Annual rainfall varies from over 10,000 mm in parts of the Central Highlands of Papua New Guinea to virtually zero in the Gobi and Australian deserts. Likewise across the region there is considerable variation in recorded temperatures. During the winter months in Mongolia the temperature commonly falls to below zero whereas in the summer months in the arid regions of Pakistan and Australia daytime temperatures can rise to over 50^oC.

Within the region can be found considerable variability in the length of growing season. Rainfed agriculture is restricted in many countries to that period of the year coinciding with the monsoon season(s). However in some of the more humid parts of the region rainfall occurs throughout the year giving an effective 12 months growing season. In addition in the northern and high altitude parts of the region the length of growing season will be curtailed by the occurrence of low temperatures (below 5^oC).

Tropical cyclones and typhoons are a feature of much of the region and result in heavy downpours with the risk of high runoff and flooding. The worst effects of the strong winds, tidal surges and heavy rainfall are mostly felt in coastal and island areas; the influence of some cyclones may extend into the interior of the Asian continent. For instance in early November 1995 the weather system associated with an unseasonal cyclone in the Bay of Bengal resulted in heavy snowfall in the mountains of Nepal, subsequent avalanches resulted in the deaths of several farmers and trekking tourists.

Islands in the Pacific and the Philippine archipelago are especially vulnerable to typhoons and cyclones. The smallest islands cannot deflect typhoons and cyclones, and are not large enough to moderate general climatic circulation patterns, making them vulnerable to drought and other climatic events, which can destroy complete ecosystems. Certain island ecosystems are resilient to such events e.g. the "typhoon forests" and indigenous farming systems of the Batanes Islands in the north of the Philippines (ASOCON 1990).

Where island ecosystems are used for human purposes, such long-term ecological resilience is commonly inadequate, and the natural disaster of a severe typhoon/cyclone event can be highly damaging to the local agricultural sector in the short term. The productive agricultural resource base is eroded, and natural regeneration (of soil and vegetation) is not fast enough to restore essential ecosystem processes (Bass and Dalal-Clayton 1995). The unusually destructive Cyclone Ofa that struck Western Samoa in February 1990 is reported to have so affected local food supplies that it took eight months before 80% of its agricultural capacity had recovered (Clarke 1994).

Several features of the climatic variability within the Asia Pacific region lead to high natural hazards of degradation. Much of the region's rainfall is associated with storm events and hence can be expected to fall at erosive intensities. In those parts of the region with distinct wet and dry seasons, a severe dry season will lead to the death of annual vegetation and much bare soil exposed to the first rainstorms. In the arid and semi-arid parts of the region, rainfall is not only low but highly variable, leading to recurrent drought with consequences for wind erosion and loss of vegetative cover.

Many countries of the region have mountain ranges which typically exhibit a wide variety of micro-climates as both temperature and rainfall can vary significantly depending on altitude and aspect. There is a marked decrease in mean temperature with increasing altitude. The highest mountain ranges in the region (the Hindu Kush/Himalayas) may progress from tropical climatic conditions in their footslopes to arctic conditions at their peaks. Above certain altitudes the occurrence of regular frosts will limit crop production. The problems of cold may be exacerbated by strong winds. Rainfall usually increases with altitude particularly on the side of a mountain range facing the prevailing rain bearing winds. On the leeward side rainfall may drop off markedly. Within a mountain ecosystem there may be localised and severe rain shadow effects. Within and across mountain ranges the climate may vary from very humid to desert conditions.

While climatic variability in the region is a constraint to SARM it is also regarded as an opportunity. The range of macro and micro climatic conditions means that a great variety of annual crops, as well as perennial tree crops, can be grown within individual countries and across the region as a whole. This contributes to the diversity of agricultural production within the Asia Pacific region.

Regional soil constraints

Variability in the nature and properties of the soils contributes to the diversity of agricultural production within the region. However many of these soils have properties which limit or preclude their use for sustainable agricultural production. The process of grouping land areas according to their soil constraints for SARM is particularly complex because in many cases individual tracts of land will exhibit a combination of soil and agro-climatic constraints (Dent 1990). It should also be noted that what may be a severe constraint for one agricultural enterprise may be less severe or a requirement for another. For instance waterlogging within the rooting zone will severely reduce the yield of many annual and perennial crops, but is essential for high yields of paddy rice.

Any attempt to assess the areal extent of soil constraints within the region has to be general in nature given the biophysical diversity and the variability in the availability and reliability of the necessary data. One such assessment has been made using data from the FAO/UNESCO Soil Map of the World (mapping scale of 1:5 million) to identify the major natural constraints to agricultural production (Dent 1990). Twelve soil constraint categories were recognised (see box 6) and figures for their proportional extent by country and for the region as a whole (see table 3) were arrived at, utilising a process of elimination based mainly on a somewhat arbitrary choice of the most limiting constraints.

Because of the limited data on which they are based the figures in table 3 should be regarded as first approximations rather than definitive assessments of the situation at either the country or regional level. However they do show that, with few exceptions, only a limited proportion of the land area of each country within the region is constraint-free. Where there are natural constraints, sustainable agricultural production may still be possible, providing the limitations can be overcome, with the aid of appropriate land management practices. Inevitably sustainable use of such areas involves higher investment costs than in constraint free areas, as well as good land management skills. However where individual constraints, or combinations of constraints, are particularly severe the costs and means of developing such areas for agricultural purposes may be beyond what is possible with today's knowledge and technologies.

The implications of the above for SARM are that the first priority should be to prevent misuse and loss of those lands that are constraint-free. Other areas cannot be ignored as the demand, from a growing population, for food means that agricultural production will have to be take place in areas where it is technically and financially possible to overcome the natural constraints. However the costs of such production can be expected to be higher, and returns lower, than in constraint-free areas.

Water resource constraints

Water like land, is becoming a scarce resource within the Asia Pacific region. Much of the extra food produced in Asia between 1960 and 1980 was grown in irrigated land. Unfortunately the rapid expansion of the irrigated areas in that period is not continuing, and the demands on existing supplies for non-agricultural purposes (e.g. urban and rural drinking water, industrial and mining enterprises)

increasingly compete with irrigation. It is becoming more difficult to identify new opportunities to develop water storage facilities. The construction of large dams is now generally recognised as uneconomic when based on a realistic estimate of the costs, which include the need for erosion control in the catchment area and drainage facilities in the command area (Greenland et al 1994). Also many dam projects in the Asia region are faced with growing opposition from local communities and the wider public due to their potential environmental, social and cultural impact.

Rather than large-scale dam construction, developing water-harvesting methods and microcatchment storage systems appear to merit more attention. However for crop production the most important problem is to improve water-use efficiency. This requires efficiency in the supply of water to the field, and efficiency in the use of water by the crop to increase economic yield. Many irrigation schemes also function at very low efficiencies (Greenland et al 1994). The expected life of many irrigation systems has also had to be drastically reduced because of soil erosion problems in the catchment area of the reservoir, causing excessively rapid siltation (box 7). In addition to improving the efficiency of the irrigation system there is scope for better crop selection. Different annual and perennial crops (or cultivars) will have varying water requirements for optimum growth. Thus crops with high water demand could be substituted for less demanding crops in order to match crop water requirements with water availability at critical times of the year.

^{Source Narayana & Ram Babu 1983}

Soil productivity

In the past too much emphasis was placed on assessing soil degradation on the basis of the weight of soil lost (expressed in tonnes of soil lost per ha, or millions of tonnes of sediment carried by rivers). The real issue is not the amount of soil lost or the area of land degraded, but the effect of this loss on the productivity of the land (Hudson 1992). Within the Asia Pacific region innumerable experiments have sought to quantify erosion, but only a handful have measured the loss of plant nutrients, and even fewer have attempted to correlate the nutrient loss with productivity (Stocking 1988).

Table 3. Extent of problem soils within Asia and the Pacific

Source: Total land area taken from FAO RAPA 1996 Selected Indicators of Food and Agriculture Development in Asia Pacific Region 1985-95 RAP Publication 1996/32

Definition

In the context of SARM, productivity can be defined as the productive potential of the soil system that allows accumulation of energy in the form of vegetation (crops, pastures, trees and shrubs) of value to farmers (after Stocking 1984, Stocking and Peake 1985). Soil productivity is a function of many factors including individual soil parameters, climate, vegetation, slope and management. It is a central element to any discussion on sustainable soil use because productivity implies the potential for future agricultural production.

Crop yield a proxy indicator

Like soil fertility, soil productivity is a real property of the soil, but is incapable of direct physical measurement (Stocking 1984). Crop yield is therefore commonly taken as a useful proxy indicator of soil productivity because of its measurability, its relevance to farmers and planners, and the possibility to quantify it in monetary terms (Stocking and Peake 1985).

A superficial look at the figures in many countries suggests that mean crop yields have increased over the past years. However this should not be taken to mean that soil productivity has not declined. Improved crop husbandry practices (use of higher yielding varieties, chemical fertiliser, pesticides etc) may have masked a decline in the soil's productive potential. The real question is, what level of agricultural production could have been achieved if improved crop husbandry practices had resulted in significantly increased yields, instead of compensating for reduced soil productivity (Norman and Douglas 1993)? It is clear that had soil degradation not taken place average crop yields would be higher, or could have been achieved at lower cost.

Very few long-term yield experiments have been conducted within the region to assess long term soil productivity. One of particular interest is a 33-year fertiliser experiment at Ranchi, Bihar. This found that despite changes to improved varieties, wheat yields declined substantially over the period with N, NP and NPK fertilisation, whereas they rose with farmyard manure (Goswami and Rattan 1992).

Factors underlying soil productivity decline

No single parameter consistently explains the loss of yield potential following soil degradation. The most important factors would appear to be (after Stocking and Peake 1985):

• Adverse changes in the chemical and biological status of the soil, e.g. depletion of nutrients, loss of organic matter, change in pH (acidification), salinisation and other toxicities.
• Reduction in the water available to plants due to: reduced soil depth as erosion brings limiting horizons (those that provide a lower limit to rooting depth) progressively nearer to the surface; and reduced water capacity of the remaining soil, as the coarser sands that remain following selective removal by erosion of the organic matter and fines have little ability to retain water.
• Decline in structural stability and increase in bulk density e.g. crusting, compaction and decrease in porosity will influence seedling emergence and root development.

Within the humid tropics the first of these factors, with the exclusion of salinisation, is probably the most important contributor to yield loss. Decline in available water capacity, and locally salinisation, will be most significant in arid, semi-arid and strongly seasonal climatic regimes. When physical degradation occurs it can have a significant impact in any climatic zone.

Erosion-induced loss in soil productivity

It has long been accepted that the productivity, or yield potential of soils is reduced by erosion (see Stocking 1984, Stocking and Peake 1985). However, erosion research has focused mainly on rates of soil loss, the detailed processes, and the variables which might be used to estimate rates. A good understanding of the processes of soil degradation is therefore now available, and rates of erosion may be predicted in many environments with reasonable accuracy (Stocking and Sanders 1992).

Research is still largely focused on the causes and description of erosion, with far less attention given to the consequences. Despite this there is an emerging consensus that:

• Erosion rate is a poor indicator of impact as measured by crop yield.
• Erosion can have a large impact even when rates of erosion are low (applies particularly to the tropics).
• Assessments of soil erosion need to be quantified in order to generate data which can allow an economic value to be calculated (Stocking and Sanders 1992).

The work undertaken so far shows that there is no simple equation that can be used to calculate that a soil loss of `x' mm (or tonnes/ha) will result in a `y' kg/ha reduction in crop yield. However comparison of the results of different trials does suggest that the consequences of erosion-induced loss in soil productivity are far more severe in the tropics than in temperate regions. The impact of a unit loss of tropical soil on yield can be at least 20 times greater than its temperate equivalent (Stocking and Peake 1985). It also needs to be remembered that it is the `quality' of the soil that remains, rather than the volume and properties of the soil that has already been lost, that will determine the future productive potential of the land (ABLH in press).

Restoring the productivity of a degraded soil is usually costly and will require considerable time and effort on the part of the farmer. It may also not be fully successful with yields remaining below those of adjacent uneroded sites (Stocking 1984). The implication is that it is better to focus resources on the prevention of soil degradation, thereby preserving the productive potential of a soil, rather than on rehabilitating degraded areas.

More to maintaining productivity than erosion control

Whereas there is a clear link between soil erosion and yield decline there is more to the maintenance of soil productivity than just the installation of runoff control measures. In the past such sayings as "Soil conservation must be done before yields can rise," and "Soil conservation raises yields," have been used to justify the construction of conservation banks in farmers' fields. If bank construction is all that is recommended in the name of sustaining soil productivity, then farmers are being deceived (Shaxson 1992). Even where land is "protected" by earth banks, productivity will continue to decline due to mismanagement of the interbank areas, resulting in adverse changes in the chemical, biological and physical properties of the soil (e.g. nutrient loss, decline in organic matter, crusting, compaction etc).

Soil resilience and sensitivity

Resilience and sensitivity are critical to the sustainability of soil productivity (see box 8). Central to the concept of resilience and sensitivity in SARM is the soil architecture and its regeneration after damage (Shaxson 1996). The spaces between the particles and structural units regulate movement and water retention as well as fluxes of oxygen and carbon dioxide in the root zone. They also affect root growth and function, and house the mass and species diversity of soil inhabiting micro-, meso- and macro-organisms. The reformation of relatively stable soil architecture after it has been damaged - by collapse, compaction, interstitial sealing, pulverisation - is achieved primarily through the activity of organisms acting on organic materials produced in situ and/or brought in from elsewhere (Shaxson 1996).

Role of vegetation in sustainable soil use

Combating vegetation degradation usually figures prominently in any programme concerned with the sustainable use of agricultural soils. This is because vegetation, whether natural (forests, woodlands and grasslands) or planted (crops, pastures, trees and shrubs), has the potential to contribute directly to the maintenance and improvement of soil productivity. The role played by vegetation can be considered under the following headings: protection; conservation and increase of soil organic matter; as a nutrient source; and improved moisture status (after Ingram 1990).

Protection

The ground cover provided by vegetation can prevent splash erosion by protecting the soil surface from the impact of erosive rains. The cover may be provided by the leaves and other parts growing above the surface (the canopy) or the dead materials deposited on the soil surface below the plants (litter). In a natural system the litter may be composed of leaves, stems, twigs, branches, seeds and fruits, whereas in cropping and agroforestry systems it may consist of deliberately applied mulch and/or crop residues.

Research on crop and rangeland productivity has found that because of the curvilinear relationship between erosion and cover, provided that mean cover exceeds 40%, erosion is low (less than 10% of that on a bare plot) under tropical

Source of definitions and matrix table Stocking 1995

Source: Wiersum 1984 cited in Qwist-Hoffman 1994

conditions (Elwell 1980, Zobisch 1992).¹⁶ Such a degree of protective cover is commonly achieved by traditional shifting cultivation systems in the humid tropics where some trees may be left standing while the trunks and branches of those

felled are left on the soil surface with the crops planted in between. It is also achieved by many traditional mixed cropping systems practised in the sub-humid to semi-arid tropics (Stocking 1985). The figure of 40% critical cover indicates that rangelands do not require a continuous sward of grasses, pasture legumes and browse species (something that may be difficult to achieve, particularly in arid and semi-arid environments) to protect the soil from erosion. It is also possible with improved crop husbandry practices to quickly provide this amount of ground cover with the leaves of well grown crops.

Many environmentalists believe that erosion can be stopped by planting trees. Regrettably it is not as simple as that. It all depends on the way the trees are planted and managed, as benefits in soil and water protection do not accrue automatically by having trees on the land (Hamilton 1986). On a recent trip to You Xi county Fujian Province PR China it was observed that the forestry department had burned all the vegetation prior to planting tree seedlings. The end result was a series of very steep and bare slopes at serious risk of erosion in the first 2-3 years of the forest plantation, when surface ground cover would be inadequate to protect against raindrop impact.

The forestry practice of screefing (scraping bare the soil surface around planted seedlings) is undertaken to reduce weed competition during establishment. When trees are planted in straight lines on steep slopes (a common practice) screefing can produce cleared strips of bare soil up and down slope leading to excessive runoff and erosion during periods of heavy rainfall (personal observation in northern Thailand and other parts of Asia). It is also clear that the litter below the trees rather than the tree canopy itself provides the bulk of the protection against erosion (see box 9). If the litter is removed for mulch, fodder, fuel etc then the conservation benefits from planting trees are seriously reduced.

It is also questionable that trees are more efficient at protection than annual crops which can cover soil far quicker (Ingram 1990). When mulched or managed with low tillage, annual crops give the same results for soil loss as do secondary forests (Lal 1977). Trees can take more than two years to close canopy in humid tropics during which time the ground is bare, as litter is still insufficient to cover it. This contrasts with annual crops which can provide adequate cover within 30 -45 days and pastures within 2-6 months (Sanchez 1987). Well-managed rotational cropping or well-managed pasture may be preferable alternatives to poorly managed forestland use (Shaxson 1992a).

Plants can also protect against rill and gully erosion by reducing runoff velocity and increasing infiltration. Litter, crop residues and a continuous grass sward will provide a sufficiently rough surface to reduce the velocity of moderate levels of runoff. Hedgerows of closely spaced tree and shrub species and grass strips across the slope can provide partly permeable barriers that reduce the velocity of higher levels of runoff, while encouraging infiltration and trapping soil on the uphill side. Widely spaced trees and shrubs as in an orchard or woodlot will have little direct effect on runoff as their trunks are too far apart to have any barrier effect. It is their litter and any herbaceous undergrowth that offer protection against runoff.

Trees and shrubs can protect soil against wind erosion by retarding the movement of soil particles. Hence the well-established use of windbreaks and shelterbelts to combat soil degradation in semi-arid regions (FAO 1976b). Tree and grass roots are also valuable in anchoring unconsolidated soils (e.g. sand dune stabilisation). The landslides that follow logging of steep hillsides (AIADP 1990) indicate that forest cover also protects against mass movement in geologically unstable landscapes.

Conservation and increase of soil organic matter

Soil organic matter is critical for SARM. Plants contribute directly to soil organic matter from two sources: below ground (roots) and above ground (leaves, stems, twigs, branches, flowers, seeds and fruit). In the case of perennial crops and plants, roots make a constant contribution to soil organic matter through sloughing off, rapid decay and exudation. These below ground processes are significant for maintaining soil organic matter levels and are part of the "hidden" benefits of perennial cropping and agroforestry systems, where tree root biomass is typically 20-30% of a tree's total biomass. The percentage will vary depending on environment, for instance in rain forests it can be as low as 15%, in moist savanna it may be 35-40%, and can rise well above 50% in semi arid vegetation (Young 1989).

For annual crops, in addition to the sloughing off and exudation during the growth period, most of the roots will be left in the soil to decay following harvest. This root-derived soil organic matter is inadequate to maintain high amounts but will ultimately stabilise soil humus at a low-level equilibrium. In the tropics this may be at 30-40% of the level under natural vegetation (Young 1976). Thus if farmers take no measures other than those necessary to prevent physical erosion, crop root exudation and decay will maintain soil organic matter in a low but steady state.

To sustain the productive potential of soils used for agriculture, soil organic matter levels should be maintained at a level of at least 50-75% of that under natural vegetation (Ibid.). Given that these levels cannot be attained from the root contribution alone, additional inputs of organic material will be required from above ground sources.

The actual quantities of such plant residues that need to be added to the soil, to maintain adequate soil organic matter levels, are estimated at 8,000 kg of dry matter/ha/yr in the humid tropics, 4,000 kg DM/ha/yr in the subhumid tropics and 2,000 kg DM/ha/yr in the semi-arid zones (Young 1989). In natural ecosystems this is no problem as the net annual primary production of above ground biomass is more than adequate (see box 10). However the amount of organic material available may be below what is required when the land is used for agricultural purposes, particularly annual crops. Not only may the total annual biomass production be reduced but much of it will be removed in the form of harvested products.

Under traditional shifting cultivation systems the deficit in available organic material during the cropping period is compensated by the ultimate surplus accumulated during the long bush fallow period. Given that long bush fallow systems are, due to land shortage, no longer an option for much of the tropics, there is a need for alternative means of supplying the necessary plant materials required to sustain soil organic matter levels.

One option is for shorter fallows in which the natural bush fallow is "enriched" with the introduction of faster growing tree species and herbaceous legumes. As this still involves leaving land idle and `unproductive' this is not an option where farm family holdings are small in size and alternative land is unavailable.

Ensuring that all crop residues are returned can make a significant contribution to sustaining soil productivity. Returning the residues from a maize or sorghum crop could well restore half the organic matter lost during one year of cultivation (Young 1976). However the common practice of burning crop residues during land preparation seriously reduces the quantity of organic matter returned to the soil let alone the quantities of nutrients.

Another option is to grow specific plants as a source of "green manure." In the case of agroforestry systems this involves taking the prunings from nitrogen fixing trees or shrubs and either applying them as a mulch or digging them into the topsoil. An alternative is to grow a herbaceous crop, usually a legume, specifically for the purpose of hoeing or ploughing it into the soil. This would have the effect of very short "enriched" fallow with the crop typically occupying the land for no more than 12 months.

The planting of a grass ley as part of a crop/livestock production system has economic value with proven capabilities of improving the properties of agricultural soils (Young 1976). Organic matter levels are raised by means of root exudation and the incorporation of the grass at the end of the ley. Grass roots also have a marked and beneficial effect on soil structure. The inclusion of a pasture legume with the grass seed, while not only improving the quality of the ley for livestock production, will also improve the soil's nitrogen status.

Plants as a nutrient source

Within the soil-plant nutrient cycle there are a number of ways in which plants can supply the soil with nutrients. Knowledge of these can lead to improved management systems that use the natural processes to maintain and enhance soil productivity for agricultural purposes.

There is a limit to the capacity of plant residues to supply nutrients, as what is not there in the first place cannot be recycled. The original source for phosphorus, potassium, and the secondary and micronutrients is rock weathering. If the soil parent material is low in these elements then, however closed the soil-plant system may be, it cannot become richer without external inputs. Nitrogen on the other hand originates from atmospheric fixation and can be increased in situ by biological means.

Plant litter, either naturally occurring or resulting from a farm practice such as mulching or leaving crop residues on the soil surface, will contribute to the replenishment of soil nutrients. As a nutrient source the humus resulting from litter breakdown has the following favourable characteristics (Young 1987):

• Supply is balanced across the range of primary, secondary and micronutrients.
• Nutrients in the form of organic molecules are protected from leaching.
• There is a steady release of nutrients in an available form through mineralisation.

Different plant residues (i.e. from different parts of the plant as well as from different plants) will decay at different rates and vary in their chemical components. From a biological soil management perspective, there are differences in the "quality" of different plant residues (Swift et al 1979). Litter of high quality (high in nutrients, low in lignin and polyphenols) decays and releases nutrients rapidly; that of low quality (low in nutrients, high lignin and/or polyphenols) decays slowly. Woody residues (stems, branches, twigs and coarse roots) are of low quality, but so are some herbaceous products including straw.

The significance of litter quality for agriculture is that it opens up the possibilities of using residues from different plants for varying purposes. High quality residues, because they decay rapidly, could be used to provide a short-term release of nutrients, their application timed to meet peaks in crop requirements. Low quality residues when applied as a mulch will remain as a protective cover for much longer while giving extended release of nutrients, protected against leaching until mineralised. This has been recognised in the context of agroforestry where leaves of different trees and shrubs vary widely in their quality and rates of decomposition. For instance the leaves of Leucaena leucocephala decay within a few weeks, those of Cassia simea, at an intermediate rate, whilst Gmeligna arborea, Acacia mangium and many Eucalyptus species are relatively slow decaying (Young 1989). Knowledge of differences in litter quality offers the scope for using combinations of plant materials to provide for both a rapid release of nutrients and a slower regular release over a longer period.

There are four management alternatives for using litter to supply nutrients: placement on the surface, burial in the soil, composting, or use as fodder with the nutrients returned via the manure. Buried litter decomposes faster than surface litter (Wilson et al 1986) but surface placement is desirable for erosion control. Burial, composting or use as fodder and/or livestock bedding may be more desirable for cereal crop residues, which are high in lignin, than for the generally high litter quality of tree leaves.

The decay of dead roots, the below ground equivalent of litter, is also a source of plant nutrients. Little data is available on the nutrient content of root residues or on rates of addition of residues to the soil, but given the amount of root biomass in proportion to the total plant the contribution from this source can not be ignored (Ingram 1990).

Biological nitrogen fixation takes place in the soil through non-symbiotic and symbiotic means. Non-symbiotic fixation is that carried out by free-living soil organisms. It can be of substantial importance relative to the modest requirements of natural ecosystems, but is small in relation to the greater demands of agricultural systems. Symbiotic fixation occurs through the association of plant roots with nitrogen-fixing bacteria. Many legumes are associated with Rhizobium, while a few non-leguminous species are associated with Frankia (Young 1989).

Nitrogen fixation by herbaceous legumes has long been a recognised agricultural practice (either as a productive crop, e.g. pulses, groundnuts), a green manure crop (e.g. Stylosanthes spp, Centrosema pubescens, including grass-legume leys), or a cover crop in perennial plantations (eg. Pueraria phaseoloides). In improved cropping systems it is usual to recommend that legumes be grown in rotation with non-legume crops. In many traditional cropping systems in Asia and the Pacific cereals and rootcrops may be intercropped with one or more legume crops such as beans, pigeon peas and groundnuts. Such traditional systems, along with intercropping of annual crops with nitrogen-fixing multi-purpose trees and shrubs, have in recent years become the focus of much research attention as a means of introducing organic nitrogen to the soil, for the benefit of the non nitrogen-fixing crops.

With regard to crop production in the tropics, the conventional wisdom is that high yields, particularly of cereal crops, depend on inputs of commercial nitrogen fertilisers, low soil nitrogen levels being a major factor in low yields under low input systems. On the basis of work undertaken by ICRAF and others it appears possible to identify trees and shrubs with a nitrogen-fixing capability (when grown in agroforestry systems) of 50-100 kg N/ha/yr (Young 1989). This opens up the possibilities for developing low external input farming systems using on-farm biological means to raise soil nitrogen levels thereby increasing crop production.

In addition to making available additional supplies of nutrients, there is scope for improving the efficiency of nutrient cycling in agricultural systems by improving the uptake capacity of plants. Under low nutrient conditions in nature, most plants are infected by one or other type of root-inhabiting mycorrhiza. The very fine and extensive mycelial network put out by these fungi improves the efficiency of nutrient and water uptake by greatly increasing the plants effective absorbing surface within the soil (Swift and Sanchez 1984). Like nitrogen, low phosphorus levels limit the agricultural productivity of tropical soils. It appears that mycorrhiza not only improves the efficiency with which plants take up available phosphorus, but with other organisms they can make rock phosphate soluble and transfer it to the host plant (Kugler 1986).

Research into mycorrhiza suggests that most of the important crop plants in both commercial and subsistence agriculture in the tropics have the capacity to form associations with appropriate mycorrhizal fungi. In low nutrient status soils crops may derive very considerable benefits from such associations (Swift and Sanchez 1984). Mycorrhizal infection is favoured by minimum tillage and low inputs of fertiliser and pesticides and so lends itself most readily to agriculture under low-input constraints. As a plant's ability to fix nitrogen may be improved by inoculation with the appropriate Rhizobium, in the future it may become possible to improve the plants uptake of phosphorus and other soil nutrients by inoculation with the appropriate mycorrhiza.

The established feeder root system of trees and shrubs is believed to exploit a greater volume and depth of soil for soil nutrients than those of annual or pasture crops (Swift and Sanchez 1984, Ingram 1990). In particular tree roots are believed to be able to capture nutrients freshly released by weathering in the deeper soil layers (below 2m) and to transfer them to the above ground parts of the plant. Such nutrients can then be made available to annual crops by utilising the natural tree litter and prunings (leaves and fine stems) as a mulch.

Agroforestry combinations appear to have considerable potential for enhancing the sustainability of agricultural soils in terms of nutrient status. The identification and use of multi-purpose trees and shrubs that can carry both nitrogen-fixing bacteria and mycorrhizal fungi infections therefore offer opportunities for the development of sustainable low external input farming systems.

Improved moisture status

Plant canopy, litter and mulch can reduce moisture loss from the soil surface by shading it from the direct rays of the sun. They can also act as a barrier (windbreak) intercepting and reducing the speed of winds that would otherwise reduce soil moisture by means of evaporation and evapotranspiration. The resulting improved moisture conditions facilitate the decomposition of organic materials. The improved surface structure and reduced runoff under trees and grasses also favour infiltration.

Hydrological conditions

The hydrological conditions prevailing at a particular site can have a marked influence on land use; this is critical to SARM. Of major significance for plant growth is the degree of surface waterlogging. This can be taken as the basis for a simple threefold classification into areas which are wet permanently, seasonally or not at all. With annual crops, whether or not waterlogged conditions extend into the rooting zone during the growing season will be critical. For paddy rice surface waterlogging is necessary whereas crops like maize and tobacco can suffer yield losses if the rooting zone is waterlogged for no more than 24 hours. Equally important is depth to the groundwater table. Perennial crops may be dependent on getting their roots down to the water table to survive a prolonged dry season, whereas a rising water table due to poor irrigation practices can result not only in a risk of waterlogging but also an increase in salinity within the rooting zone.

Small-scale farm households depend on water supplies not only for domestic use, but for livestock and irrigation. There may be more than one source and they may be used for different purposes, e.g. boreholes and wells used for domestic use, with livestock watered from rivers or dams, and shallow wells in valley floor sites used for dry season vegetable production. The location of water sources (surface or groundwater), water quality and seasonal availability can determine the suitability of an area for particular land use enterprises.

Many of the upland and mountain landscapes within the Asia Pacific region have high energy river systems with considerable ability to scour soil and transport coarse as well as fine sediment. Steep slopes lead to high rates of surface runoff and high velocity channel flows. Runoff is often seasonal in occurrence associated with the monsoons and in the case of the high mountain ranges of Asia, Australia and New Zealand the melting of winter snows.

Pests and diseases

The presence or absence of pests and diseases is an important characteristic of the biophysical environment for resource poor small-scale farmers. Vanuatu is reported as one the world's few truly cattle disease-free countries (Keith-Reid 1997). As a result it produces high-grade beef free of the drugs used in other countries to ward off disease. It thus reduces production costs and gives it a premium on price in the Japanese market.

The incidence of particular pests and diseases and the technical/ financial difficulties in controlling them could be a major constraint with regard to the productive potential of particular agricultural enterprises that would otherwise be ecologically suitable. The sudden occurrence of a pest or disease in an area where it has not previously been recorded can be devastating. The traditional farming system in Western Samoa was based on Colocasia taro as the main staple food crop. The outbreak of taro leaf blight in 1993, virtually wiping the crop out, forced on farmers significant changes in their cropping systems. Whereas previously Colocasia taro was the dominant crop grown by most farmers, within 12 months of the outbreak farmers had replaced the crop with significantly increased plantings of bananas, yams Alocasia, Xanthosoma taro, kava and cassava (Rogers and Losefa 1996). The outbreak has had severe economic and social consequences at both the national and household levels. Taro had been a major export crop for the country as well as being culturally important for Samoans.

Many land factors influence the relative incidence of pests and diseases, the most common being climate and soil. Among climatic factors, high humidity is particularly likely to cause increased incidence or effects of plant diseases. Soil-borne pests may be affected by texture; for example, nematode damage to sugarcane is most serious on sandy soils. The facts of biological distribution in themselves constitute a land characteristic, even if unrelated to climate, soil or other factors. For instance in 1995 the FAO Investment Centre formulation mission for the Indonesia Second National Estate Crop Protection Project found during its field visits that it was still possible to grow cocoa as a cash crop on some islands whereas the presence of cocoa pod borer on other islands had such an adverse effect on yield as to rule this out. During the same mission farmers on Bali whose citrus orchards had been infected with the citrus greening disease had no option but to abandon citrus production and look for alternative cash crops.

Pesticide-induced pest resistance and resurgence

The conventional approach of scientists and development planners to the problem of pests and diseases has been to promote the application of pesticides. The aim has been to prevent pest damage, yet ironically they can cause outbreaks themselves. Pesticides can be inefficient for several reasons (Pretty 1995):

• They can cause resurgence by killing off the natural enemies that control pests.
• They can produce new pests by killing off the natural enemies of species which hitherto were not pests.
• They can induce resistance in pests to pesticides.

Resistance can develop in a pest population if some individuals possess genes which give them a behavioural, biochemical or physiological resistance mechanism to one or more pesticides. These individuals survive applications of the pesticide, passing their genes to their offspring so that with repeated applications the whole population becomes resistant. High and frequent applications of pesticides exert the greatest selection pressure on populations (Pretty 1995). Resistance has now developed in all insecticide groups; at least 480 species of insect, mite or tick have been recorded as resistant to one or more compounds (Georghiou 1986). Unfortunately natural enemies appear to evolve resistance to pesticides more slowly than herbivores, mainly because of the smaller size of the natural populations relative to pests and their different evolutionary history (Risch 1987).

Resistance has also developed in weeds and pathogens. Before 1970 few weeds were resistant to herbicides but now at least 113 withstand one or more products. Likewise some 150 fungi and bacteria are also resistant (WRI 1994). An agroforestry trial in Western Samoa found, when monitoring weedy species under alley-cropping systems, a shift away from grasses to broad-leaved species (which were less competitive and easier to control), whereas the excessive use of the herbicide Gramoxone (paraquat) to control weeds over the past decade had led to increasing dominance by rhizomatous grassy species (Rogers et al 1993).

There are numerous reports of disease and insect outbreaks from within the Asia Pacific region (Khush 1990, Kenmore 1991, Winarto 1994). Brown planthopper (Nilaparvata lugens) outbreaks have at various times destroyed hundreds of thousands of ha of rice in countries from India in the west to the Solomon Islands in the east (Pretty 1995). In Indonesia the first problems started in 1974. Losses jumped in 1975 after the government started subsidizing pesticides and in 1997 over 1 million tonnes of rice were lost, enough to feed some 2.5 million people (Kenmore 1991). In 1979 750,000 ha were infested, followed by lower, but not insignificant, levels of infestation of between 20-150,000 ha per year during the 1980s. During this period the pest was only really checked with the release of new rice varieties containing genes that confer resistance, though even some of these have been attacked by new biotypes of the pest (Khush 1990).

Studies in the Philippines and Indonesia showed that outbreaks occurred after increases in insecticide use (Kenmore et al 1984, Litsinger 1989, Winarto 1993). Brown planthopper is kept under complete biological control in intensified rice fields untreated by insecticides. Even with over 1000 reproducing adults per square metre the natural enemies exert such massive mortality that rice yields are unaffected (Pretty 1995). As described in a report (Kenmore 1991):

Other countries in Southeast Asia still suffer significant losses to brown planthopper. In Central Thailand some 250,000 ha were infested in 1990, the worst year on record (Pretty 1995). In addition to the problems of pest resistance and resurgence there are other economic, environmental and related social costs associated with the use of pesticides (see box 11). Hence a need to take into consideration the full costs and benefits of pesticides before advocating their use for sustainable agricultural production.

Geographic isolation

Many small-scale farmers live in remote, often ecologically fragile areas, with few roads and little in the way of physical infrastructure such as trading centres, schools and health facilities. Such people are generally politically marginalised as well as economically disadvantaged, their access to government development funds is limited and as a result have few alternatives to exploiting their local natural resources on a short-term basis for their immediate survival needs. Historically, due to the nature of the terrain, highland and mountain communities in particular have been comparatively isolated (both from each other and the lowlands) requiring them to be largely self-reliant.

Many such isolated societies have evolved land use practices adapted to their local ecological conditions which enabled them in the past to meet their needs on a sustainable basis. These are now beginning to break down as population pressure increases due to the provision of better health care facilities.

Likewise increased exposure to the market economy encourages greater exploitation of natural resources (e.g. timber and charcoal) to generate cash for the purchase of consumer goods.

Improving road communications within small-scale farming areas can have mixed consequences. On the positive side overcoming geographic isolation through road construction may:

• Stimulate an increase in production by providing farm households with better access to farm inputs and markets (fertiliser can come in, and produce go out by truck rather than being hand carried); and
• Open up opportunities for the growing of alternative higher value crops, some of which may be better from a soil conservation point of view e.g. perennial tree crops.

However road construction may escalate land degradation problems if:

• New market opportunities encourage short-term non-sustainable cash crop production, e.g. ginger growing on steep slopes.
• New roads attract new settlers, not only leading to increased pressure on available resources, but also land degradation should the settlers follow the traditional land use practices of the area they have come from and it turns out that these are unsuited to the specific ecological conditions of the areas they have moved into, e.g. lowland farmers moving into marginal upland areas.
• Poor road construction results in the loss of protective vegetation, unstable roadside banks and concentrated runoff thus directly contributing to increased soil erosion and downstream sedimentation problems.

Geographic isolation is a particular problem for many island states, especially in the Pacific (Commonwealth Secretariat 1985). Not only are they a long distance from the nearest continent (e.g. Western Samoa is over 4,320 km from Australia), but also from each other. Many Pacific states are made up of widely scattered component islands. In consequence, although the capital islands are separated from each other by an average of upwards of 1,120 km., in several cases the outlying islands of two states may be less than 400 nautical miles apart. This gives rise to problems of the delimitation of their respective 200-mile exclusive economic zones.

Remoteness is reinforced by the difficulty of establishing viable transport and communication links. While new technology is easing the communications problem, efforts to improve transport links are impeded by the increasing use of container vessels and wide-bodied aircraft, which require large volumes of traffic. Whereas the main island in each Pacific Islands country may have reasonable links with the outside world in terms of an international airport and a port capable of handling large container vessels, the other islands frequently have to transship their imports and exports using smaller ships and poorer handling facilities, greatly adding to their costs and increasing the disadvantages of the isolated location. The shipping costs for bulk produce from Vanua Levu to Viti Levu (the main island) were reported at 1� times what it cost from Viti Levu to Australia.

Similar problems occur in some of the Asian archipelago states such as Indonesia and the Philippines where the economies of scale, important for reducing transport costs, work against the smaller and more remote island communities when it comes to the import of external agricultural inputs and the export of surplus produce. For instance there is anecdotal evidence that it is often cheaper for the supermarkets in Manila to purchase imported apples from the USA than to buy tropical fruits from the islands in the south of the Philippines.

In terms of total effect and the number of states affected, remoteness is a significant problem mainly for small island states. It encourages persistence of a subsistence economy. In contrast, modernisation and outward orientation have come early to the island states that are more favourably located on sea routes or nearer to other states and continents (Commonwealth Secretariat 1985).

This isolation can at times be a bonus as many Pacific Island countries are free of most pest and diseases that affect crops and livestock in Asia. However, maintaining this state requires strict enforcement of quarantine regulations. Should a new pest or disease reach one of these countries its effect can be devastating due to the lack of natural resistance amongst the local plants and animals.

¹⁶ Although the cited references refer to research work undertaken in Africa it is believed that the same relationship between mean ground cover and erosion applies in the Asia Pacific region.