2.1. The Nutrient Cycle and the Chemical Fertility of the Soil
2.2. The Structure and Physical Protection of the Soil
2.3. Soil Organic Matter and the Carbon Cycle
2.4. The Plant Water Relationship
2.5. Other Hydrological Processes
There are two, perhaps three, basic cycles in which the main plant nutrients circulate. The first consists of the uptake of nutrients from the soil, their incorporation into plant tissue, and their direct return to the soil. This natural cycle dominates in almost all natural vegetation, and is one of the biophysical bases of shifting cultivation. In agriculture there is an additional loop, in that nutrients may be removed in harvested material, but partly returned in human or animal wastes. In far too many parts of the less-developed world this nutrient cycle is no longer efficient, because the duration of the fallow period has become too short. Finally, all vegetation has an external open "cycle", where nutrients escape to the atmosphere or to rivers or groundwater, and are indirectly and partly replaced from the atmosphere or soil mineral breakdown. This external "cycle" increases greatly during soil degradation. It is also a main component of soil acidification, which is due to loss of cations from the soil profile (Uexkull and Mutert, (1995)).
Where the social and economic conditions are such that nutrients in harvested produce are not returned to the field because it is sold to cities or abroad, the harvested nutrients become part of the external cycle also. The removal in the harvest from intensive agriculture can be very large, up to over 200 kg/ha of N or K. It is then necessary to replace these losses with fertilizer. Such heavy cycling is difficult to control accurately, which is the main reason for the losses in this type of agriculture. Considerable advances have been made in modelling movement of N in plants and soils over the past 20 years, so that control can now be improved (Addiscott et al., 1991). A key requirement is that nutrients, particularly nitrogen, shall be supplied as far as possible at the time when they are required by the growth of the plant, so that they are taken up relatively rapidly. This is particularly important where heavy rainfall and light-textured soils lead rapidly to leaching. However, the unknown local variations in growth, yield and naturally present soil nitrogen mean that precise control is still difficult. The decision on the correct amount of fertiliser for a particular piece of land under a particular crop is still very open to error, even in highly developed agriculture.
The very low fertility of soils under shifting cultivation found in much developing-country agriculture mean that yields and nutrient offtakes are correspondingly low, and are often decreasing (Table 1 and 2). As the fallow period shortens, the internal nutrient cycle becomes still less effective. It is imperative to import nutrients as manure, residues or fertilisers in such cases. There are many ways in which the nutrient cycle can be more tightly closed, such as by recycling all plant, animal and human residues, applying hedgerow or agroforestry prunings, and the use of legume crops, but this does not make the soil productive if it starts from a severely depleted state. Even the best-managed cropping will have some losses through the external cycle, which have to be replaced. Nitrogen is a special case because it can be fixed from the atmosphere, though this is often not sufficient to meet crop needs. Cropping of a soil which is continually losing nutrients, particularly phosphorus and potassium, is ultimately non-sustainable (Smaling, 1993). This is the basic argument for the use of fertilisers. The challenge is to use these for intensification so skilfully that the pollution problems seen in temperate (and some tropical) agriculture are avoided, and that their use becomes sustainable. It is even more challenging to do so in economical and affordable ways for resource-poor farmers with small farms, especially in Africa. Nutritional problems are often severe in shifting cultivation systems as these become more intensive. Table 1 shows several classifications, all based essentially on the ratio of cropping years to the total cycle length, emphasizing the importance of this ratio for the productivity of the system. The way in which productivity decreases as the system moves from Phase 1 to more intensive states depends greatly upon the soil type and its resilience. Very few soils will stand continuous removal of nutrients in continuous cultivation. Table 2 contains some excessively low yield values, that are presumably largely due to nutrient depletion. Nutrients contained in organic residues are very valuable, and should be used as extensively as possible, both to enhance nutrient cycling, and because they have slow-release characteristics that may in some circumstances make them more efficient than fertilisers for the same total amount of nutrient. Their prime drawback is that the quantity is limited.
It is argued above that higher production will eventually require the use of fertilisers, which are sometimes considered suspect on grounds of sustainability. Potassium and phosphorus are produced by extractive industries, and their supply is therefore not indefinitely sustainable. In fact, the potential reserves of potassium are about 1000 times greater than the annual production, and phosphate rock resources are about 450 times greater (Louis, 1993). Also, nutrients from fertilizers may escape from the local nutrient cycle, and cause pollution of water ways and groundwater, with eutrophication and health hazards. This problem can be largely controlled by careful agronomy, and this must become a larger research topic in less-developed countries in the future. (Nye, 1992)
This consideration of nutrient cycles and balances suggests that greater efficiency in uptake by plants is of only partial utility, because nutrient must at some stage be resupplied if it is to be taken off in crops. However, more efficient uptake does allow the soil in a field to be operated at a lower nutrient concentration than otherwise, and this will tend to minimise losses. Low soil concentrations of nitrogen also encourage resupply of nitrogen by associative or symbiotic fixation, which can supply all or part of the requirement. Associative fixation had an uncertain record at one time, but it appears that some grasses can fix important amounts. The fungal mycorrhizal root associations can improve the uptake ability for phosphorus, and thereby make more of the soil phosphorus available at a given plant growth rate (Sieverding, 1991).
Table 1: Proposed Divisions of Shifting Cultivation Systems
|
Shifting cultivation land |
Recurrent cultivation land |
Permanent and semi-permanent land |
Allan, 1965 |
|
|
Long. medium. short |
||||
|
L > 10* |
L = 7 - 10 L = 5 - 7 L = 3 - 5 |
L < 2 or = 2 - 3 |
|
|
|
Phase I |
Phase II |
Phase III |
Phase IV |
Greenland, 1974 |
|
Simple shifting cultivation |
Recurrent cultivation May be complex |
Recurrent cultivation with continuously cultivated plots |
Continuous cultivation |
|
|
Extensive shifting cultivation |
Intensive shifting cultivation |
Semi-permanent cultivation |
Permanent farming |
Ruthenberg, 1976 |
|
Forest fallow |
Bush fallow |
Short fallow |
Annual/multicropping |
Boserup, 1981 |
* C = number of cropping years/cycle. F = number of fallow years/cycle, L = (C + F)/C.
**R = 100C/(C + F).
Table 2: Yield per Unit Area (kg/ha) of Some Major Staples in Some African Countries
(selected data from FAO Production Yearbooks)
|
Country |
Crop |
1961/65 |
1969/71 |
1979/81 |
1989/91 |
1993 |
|
Côte d'Ivoire |
Rice |
890 |
1168 |
1171 |
1174 |
1334 |
|
Maize |
680 |
773 |
700 |
713 |
831 |
|
|
Cassava |
2500 |
3300 |
5266 |
5680 |
4521 |
|
|
Sorghum |
500 |
507 |
538 |
563 |
600 |
|
|
Ghana |
Maize |
550 |
982 |
1078 |
1300 |
1509 |
|
Cassava |
8500 |
7419 |
8647 |
|
7226 |
|
|
Kenya |
Wheat |
1678 |
1678 |
2011 |
1747 |
1579 |
|
Maize |
1100 |
1489 |
1360 |
1300 |
1249 |
|
|
R&T |
|
7722 |
7993 |
|
8077 |
|
|
Niger |
Millet |
|
422 |
435 |
383 |
357 |
|
Sorghum |
480 |
445 |
432 |
280 |
153 |
|
|
Nigeria |
Maize |
920 |
869 |
1350 |
1300 |
|
|
Sorghum |
850 |
652 |
634 |
1093 |
967 |
|
|
R&T |
|
9585 |
9728 |
9930 |
10193 |
|
|
Cassava |
6200 |
10592 |
|
11150 |
10500 |
|
|
Sudan |
Millet |
|
567 |
397 |
166 |
|
|
Sorghum |
900 |
808 |
725 |
534 |
509 |
|
|
Tanzania
|
Maize |
1180 |
813 |
1306 |
1340 |
1404 |
|
Sorghum |
870 |
503 |
763 |
970 |
1102 |
|
|
R&T |
|
4902 |
9491 |
8280 |
8223 |
|
|
Cassava |
4100 |
4854 |
12071 |
10830 |
10400 |
|
|
Uganda |
Cereals |
|
1069 |
1551 |
|
1528 |
|
R&T |
|
4449 |
5802 |
|
6594 |
|
|
Zaire |
R&T |
12100 |
6795 |
6901 |
7562 |
7906 |
The worst understood part of soil science and plant nutrition today is almost certainly the soil microbial population and its functions in relation to higher plants (Lynch, 1990). Very few of the microbial species are identified, the systematics are fragmentary, and the functions are known accurately only for a limited number of specialist organisms. Many of the soil processes that they carry out are essential for plant nutrition, soil structural stability, and several of the most important biogeochemical cycles. Considerable strides have been made in the past decade, but the introduction of new molecular biology techniques should allow accurate identification of both species and function in the future.
Soil structure is still not well understood at a fundamental level, and much of the science is largely empirical. It is still not possible to describe 3-dimensional structure or porosity in fundamental terms, and a great deal of reliance has to be placed on inferences from the moisture characteristic and the hydraulic conductivity. The strength and impedance of soils are closely allied to structure and texture, but cannot be predicted in a mechanistic way. The quantity and behaviour of the particular clays in the soils (low-activity kaolinites or active smectites) together with the amount of soil organic matter are critical. In particular, it is difficult to predict the degree of resilience of a soil following structural damage. This rather weak theoretical basis means that much of the work in these subjects must be empirical, applied and adaptive, though absolutely essential.
It is extremely important to preserve structure, otherwise capping, loss of infiltration, wind and water erosion, and root impedance easily occur. The general field methods of preserving structure and porosity are well known from long experience; maintain soil organic matter levels, maintain vigorous plant growth, prevent heavy loads, prevent large raindrop impacts at terminal velocity, and cultivate with care at the right time. Even so, the immediate strength of the soil structure, and the speed with which it regains structure after damage (resilience) are not properly understood (Greenland and Szabolcs, 1994). Even soils with similar textures may show great differences in the persistence of structural damage. Research must therefore aim at better practical and site-specific methods of preventing damage, or remedying it in suitable cases, and a great deal of progress has been made in this way. Social aspects are important. For example, even when the value of mulching is recognised for structure protection, there are often alternative uses for the vegetable material which have priority.
Wind erosion can be devastating, and is a critical component of the loosely defined process called desertification. Once soil movement starts it helps to loosen more soil, and dust storms rapidly develop. The essential control measure is to maintain a sufficiently dense vegetation cover, so that soil does not start to move, which is why intensive grazing is so dangerous, especially under variable rainfall regimes and socioeconomically driven high herd and flock numbers.
Water erosion can occur as sheet, rill, gully or river-bank erosion, depending upon the topography, soil type, rainfall intensity and erosivity, infiltration capacity and length of run. Some erosion losses always occur, even under natural conditions, and less than 2 t/ha/y are usually regarded as acceptable, depending upon the assessed rate of formation of fresh soil at the bottom of the soil profile. However, in serious cases the loss may be over 100 t/ha/y. The provision of planted or natural strips of permanent vegetation along field contours and river banks is useful in lessening river silt load, though the latter may not affect the actual erosive process on the field itself. A variety of erosion models exist, but the whole process is extremely time and site-specific, and a more precise ability to predict the effects is needed. The pathway of the eroded material further down the catchment is often irregular and difficult to follow, so protective measures may be needed closer to the place where off-site damage is occurring while erosion continues.
Cultivation techniques are therefore very important. Mixed- and inter-cropping has the advantage that the soil is very rarely left completely bare (Greenland and Lal, 1977). Zero tillage has obvious attractions, assuming that a layer of residues is left on the soil surface, and was shown to be beneficial in West Africa and elsewhere (Greenland, private communication). However, there are reports from East Africa that cultivation produced better yields than zero tillage, while the soil structure was maintained. It seems likely that these differing reports are a typical case of the site-specificity of soils work, and a consequence of less erosive rainfall and different soil types in East compared to West Africa. The question is whether understanding of the processes is sufficient to predict where zero tillage will or will not be superior to other techniques, and Lal (1983) has identified certain characteristics that favoured no-till systems. The clear definition of soil types and their properties in the cultivated state is an essential pre-requisite for this type of work.
Soil organic matter (SOM) occupies a crucial position with regard to soil fertility and its ability to grow crops. The microbial breakdown (mineralization) of SOM provides nitrogen, phosphorus and sulphur as plant nutrients - in the absence of fertilizer or organic wastes, this is the only source. SOM can also increase the ability of the soil to hold other nutrients such as potassium. The structure of soil is also dependent upon SOM, which cements and stabilizes soil aggregates and thereby makes the structure more porous and more stable. The increased porosity is valuable in holding water in the soil. The presence of adequate SOM is associated with the level and activity of the soil biological population, which is essential for soil health. The ability to maintain an adequate level of SOM is therefore critical for soil as a medium for plant growth (Woomer and Swift, 1994). Models are now available that can predict the future SOM level from the environmental conditions, the inputs of vegetable material and the soil type, but these still require improvement and testing.
In the last decade the behaviour of this reservoir of carbon in the SOM has become of heightened importance because it partly determines the carbon dioxide level in the atmosphere. World soils contain roughly 1500 Gt, with about 750 Gt in the atmosphere and about 550 Gt in the land biota (almost wholly vegetation), so the possibility of carbon storage in standing vegetation or in the soil is of great interest. However, the most immediate question is how much carbon is lost from deforested and newly cultivated soils, currently thought to be about 2 Gt per year (IPCC, 1990). However, the most recent work suggests that tropical forests are in total net absorbers of about 0.9 Gt carbon per year (Grace et al., 1995). Much work on this subject is still needed.
No plants can grow without losing water, because of the need to take in carbon dioxide for growth through their stomata, though some plants have various mechanisms for water sparing or drought resistance (Smith and Griffiths, 1993). The weight of water transpired for unit weight of dry matter formed (the transpiration ratio) varies widely both with the plant species and the climate, but is usually of the order of 200-1000. This ratio is likely to be decreased by the steadily increasing level of carbon dioxide in the atmosphere, but the detailed effects are still not clear.
The total amount of water that can be transpired by a crop during its growth cycle is determined by the climate, the radiation interception, the canopy structure, and the internal physiology. If that amount of water is not available, as rain, irrigation or water stored in the profile, the yield is lowered. In rain-fed agriculture, the amount of available water stored in the profile is critical, and depends upon previous rainfall, the water-holding capacity of the soil and the partition between infiltration and run-off (Passioura, 1988). The extent and ramification of the root system determines how much of this water in the soil profile can be utilised, and at what rate. The water not used in evapotranspiration or in run-off is then available for percolation and groundwater recharge - the second point of partition.
Water is also lost from the soil surface during plant growth, at a rate dependent upon the surface soil water content, and the rate at which radiant or adventive energy reaches the soil surface. Where the crop canopy is sparse, a large part of the total water may be lost in this way, and dryland crops often transpire less than half the rainfall (Le Houerou, 1984). The use of mulch may improve this, but a thick mulch can retain rainfall, and allow it to be evaporated without reaching the plant roots. The frequently poor utilisation of water by crops in dryland environments offers a number of possibilities for improvements. These include concentrating rainfall into a fraction of the land area to increase depth of percolation (a form of water harvesting), or use of intercropping to give a more complete or a longer-lasting canopy. However, there has to be a compromise between using the rainfall for immediate cropping, and allowing it to recharge the groundwater or maintain the river flow.
The total water use by a stand of plants on a given soil with given rainfall varies with the structure of the canopy, which determines the "surface roughness," and thereby the interaction between canopy and atmosphere. It also depends upon root penetration, so that the profile can be dried out to different depths. This is the cause of the general finding that woodlands transpire considerably more water than do short crops or grasslands. The replacement of one type of land use by another can therefore sharply alter the amount of water that can be left in the profile, or percolate down and recharge the groundwater or maintain river flow.
All these processes have been built into a number of models, based upon the single-dimension Soil-Vegetation-Atmosphere-Transport (SVAT) models and the Penman-Monteith equation, which has been extended to deal with incomplete canopies. At larger scales these are coupled to above- and below-ground water transport models (see below), incorporating variations in soil type, vegetation structure and topography if necessary (Wallace, 1995a). These are often large and complex, but are becoming steadily more accurate. In the limit such models are being built into the Global Circulation Models used for climate prediction. The theoretical underpinning is therefore strong, with the weakest part probably in the simulation of the distribution and function of root systems in relation to soil properties.
One of the most difficult situations to research and to understand is that with controlled competition between two or more species, as in intercropping or agroforestry (Wallace, 1995b). The mechanism of competition may be for light, nutrients or water, of which the last is likely to be important in all systems in which there is a net water deficit. The measurement of how and to what effect the different species and plant individuals divide the limited available water supply is technically complex, but obtaining a net benefit from having two or more species present depends upon this partition process. Most research in more-developed countries has focused on the monoculture situation, even in work on forests. Advances are being made at ICRAF in this subject, but much more work will be needed.
Climate is defined by the annual means, the distribution of the variables according to season, and the between-year variability. All are critical for rainfall, because this determines the amount of precipitation, its intensity and the frequency of extreme events. These affect infiltration, runoff and erosion, horizontal and lateral transport and storage in soil profile, drainage and groundwater recharge, and associated questions of water quality. Groundwater is a most important component, and its lateral flow towards the rivers helps to maintain their volume. Lateral flow of groundwater may cause transfer from one watershed to another, but it is unlikely to be a major factor. The underlying theory of water transport in soil is strong, but the accurate measurement of the actual soil parameters in the field situation can be difficult.
The basic understanding of the movement of fertilisers, pollutants and other solutes is good, but work in the field is difficult because of the heterogeneity of soil structure. Strongly structured soils often have a bimodal porosity, which means that during rapid percolation, when large pores are waterfilled, the rate of penetration of solutes can be much faster than expected. When rivers are contaminated by such chemicals, the water quality will fairly soon return to normal if the sources can be controlled. Contamination of groundwater is potentially far more serious, in that groundwater may take decades or centuries to be replaced.
Watershed research is heavily dependent upon the application of computers and mathematical modelling, which are now used in data storage, real-time control systems for river management, simulation models for pollution of rivers, and large models of subsurface flow, groundwater behaviour and total watershed hydrology (Maidment, 1992). The SHE (Système Hydrologique Européenne) is an example of such large and data-hungry models for predicting water movement over considerable areas.
The use of modelling in controlling large irrigation schemes is one aspect of this type of work, which is made more difficult because high salt and sodium contents can change the hydraulic properties of a soil markedly. Other uses are in predicting the water relationships in watersheds under different forms of land management (Gregersen et al., 1987). There is a large hydrological modelling programme for the whole Mississippi basin operated by a part of the World Climate Research Programme. There are also now programmes that aim to connect hydrological models to other spatially arranged models, such as the modelling system for hydrology, farm economics, land use, ecology and water quality that has been constructed for two river basins in the UK (O'Callaghan, 1995).
The subject is advancing rapidly, partly due to the increasing power of computers. Whilst most of the underlying theory is strong and dependable, the problems are in the sheer spatial and temporal complexity of natural systems, and the consequent possibility of error. Careful validation of models is therefore necessary. Little such work has been done thus far in the CGIAR System, with the major cases being the small-watershed management models developed at ICRISAT.
