3.1 Tillage practices for seedbed-rootbed preparation
3.2 Fallows
3.3 Fertilizer and lime requirements
3.4 Irrigation of ferralsols
In order to obtain good tilth several operations, including plowing, disking, rolling and harrowing may be necessary. They are all discussed in this chapter under the common name of tillage practices. Their most current objectives are to create optimum conditions for germination and for the development of a root system.
It is normally accepted that maximum yields are obtained from soil materials in which the aggregates that are close to the seeds are 3-12 mm. in size. Larger lumps may be present but should not be in contact with the seeds. WILSON and WINKELBLECH (1969) recognize two aspects in soil tillage, the seed zone and the root zone. In order to obtain uniform stands, the seeds should be surrounded by uniformly sized, closely packed crumbs particularly during the early stages of germination. The root zone, which operates once the root system is developed, requires less initial pulverization because it is exposed during much more time to the action of weather on the breakdown of soil clods,
Tillage operations as a rule should therefore aim at producing the highest proportion of suitable crumb (3-12 mm.) firmly packed at the depth where the seeds are placed. According to what has been said before, it is not required to nave the same degree of granulation achieved over the entire soil surface, but only at the vicinity of the seed. A rougher surface between the zones may increase the rainfall acceptance of the field, reduce erosion hazards, and prevent soil crusting.
Suitable tilth is obtained by correcting adverse soil properties. There may be other purposes besides structure for working the soil. The most important objectives are listed below without suggesting any rank of priorities or making them mutually exclusive:
i. change the pore size distribution in case it were inadequate for root penetration;ii. increase the water intake capacity of the soil in order to take maximum advantage of rainfall, reduce runoff and control erosion;
iii. improve soil aeration and thus stimulate the decomposition of organic matter, and enhance nitrification;
iv. mix fertilizers, green manures or other amendments with the soil, in order to incorporate them at suitable depth in the soil profile;
v. control weeds by turning them under at a depth which restricts their growth.
a. There is usually no possibility to carry out intensive plowing on virgin ferralsols which are recently cleared from the rainforest. Too many thick roots and stumps hamper the normal operations of tillage implements. It is moreover doubtful whether intensive plowing of forest ferralsols is needed, let alone whether it is desirable. Soil pores of suitable size in freshly cleared land are presumably present in sufficient amount.Unfortunately there are only few data on the minimum pore size requirements for roots. RUSSELL (1971) reports that small seeded, crops will only send their roots into pores that are larger than 100 microns (some grasses) or 200 microns (wheat). Virgin ferralsols in the tropical rainforest areas will certainly have enough pores of this size, and probably enough channels of this diameter, and therefore no deep plowing would be needed, in order to increase their porosity, when such plants are cultivated directly after clearing.
The bulk density of the plow layer, which is related to total porosity, has more often been referred to with relation to the effects of tillage operations. TROUSE and BAVER (1962) found that root elongation in low humic latosols (ferralsols) was reduced at a bulk density approaching 1.35 g./cm3. LE BUANEC (1970) reports that in savanna the sols ferralitiques of Ivory Coast the root development of cotton was stronger in surface layers having a total porosity of 55% (bulk density of 1.15), than in horizons having only 40% (bulk density of 1.50).
Penetration of the root tip in pores and channels is followed by expansion. Not all plants have similar requirements however. Trees and shrubs for example have roots which expand appreciable with age. Soil strength is one of the main factors which reduce the possibilities of root development in soils of critical bulk density. It is also known that soil materials are more rigid when dry and that mechanical impedance starts to restrict root development in soils of high bulk density (>1.55), before the water stress begins to retard plant growth. The main purpose of tillage operations in surface layers of high bulk density is to create large pores, channels and void planes in the rooting zone, resulting in a lower bulk density, and achieve a better distribution of air and water in the vicinity of the growing root tips.
LE BUANEC (1970) has found that the sols ferralitiques at Bouaké (25% clay) do not give satisfactory yields of igname, unless the soil has been loosened by deep plowing. Root crops need undoubtedly more plowong than small grains or grasses. BAVER et al. (1972) state that in low humic latosols (ferralsols) of Hawaii the rooting depth of sugarcane is generally restricted to the depth of tillage, and that deep plowing on these soils is therefore essential for optimum cane growth. In this case the soils had been cultivated over long periods of years, and the aeration effects of the proceeding natural vegetation had probably disappeared.
Most investigators concur that tillage of ferralsols results in higher yields (JURION and HENRY, 1969, page 130). It is not always clear however whether the benefits are due to improved pore size distribution and root penetration or whether other factors are involved.
It has become a matter of routine to insist on proper soil moisture content for plowing, but no accurate measurements on the optimum humidity have been found in the literature on ferralsols. The moisture content should be low enough so that the soil cohesion within the aggregates is sufficiently strong as to resist compaction and puddling. On the other hand, soil moisture should permit the breakdown of clods into pieces of suitable size. Deterioration of structure is more likely to occur in light textured ferralsols than in clayey ones, in the case that plowing had been done in too wet conditions. The high ironoxide content in the finer soils are usually a favourable factor for structure stability. Most problems have been encountered in coarse topsoils; for example surprisingly severe hardening of sandy red loams (72% sand) was observed in East Africa as a result of plowing under wet conditions. This hardening effect extended to a depth of 2-3 inches. (PEREIRA et al. 1958).
b. The second objective of tillage operations is to promote the water intake capacity of the soil. Oxic horizons and topsoils do not easily form surface crusts which are the main cause of limiting water acceptance; low fine silt contents and high percentages of free iron oxides do not seem to be conducive to crust formation, except in very fine sandy topsoils. The light textured surface horizons either become hard and massive when dry, or fall apart into single grained structures. In the case of hardening, plowing in too dry conditions will result in a cloddy surface condition, which may be adequate for rainfall intake, but not necessarily for germination, and require many post plowing treatments, or strip preparation of seedbeds.
There are no soils which can resist the high rainfall intensity of tropical regions, and adsorb water fast enough at any time as to avoid runoff; cultivation on the contour should always be the rule. BERTONI (1966) measured the effects of contour plowing and sowing versus downslope operations in ferralsols of São Paulo (Brazil), on 6% slopes under 1300 annual rainfall, and found that soil losses were reduced from 21,4 tons/ha/year to 4.1 tons. Water losses by runoff decreased from 64 mm /year to 36 mm /year. In this experiment it was shown that contour sowing is even more important for soil and water conservation than contour plowing. Conservation of the topsoil is particularly important in ferralsols, because the fertility is often exclusively present in the surface horizons.
In areas that are under the climatic stresses of erratic rainfalls, it is imperative to increase the water acceptance by the soil surface. A system of tied ridges, which consists of furrows tied at intervals to form basins, has been tested on lateritic soils in Kenya (PEREIRA, HOSEGOOD, and DAGG, 1967). It is essentially a system to make sure that maximum benefit shall be obtained from precipitation. In some years it increased the yield by 40% relative to those from the fields which had been cultivated on the flat (DAGG and MACARTNEY, 1968).
Ridging of ferralsols which are not exposed to erosion is not without presenting drought hazards however, especially in climates with strongly contrasting seasons (WALTON, 1962). It has been found that in the beginning of the growing period the moisture available in the topsoil is less under ridged than under land on the flat. The difference is attributed to the larger surface area exposed by ridging (ibid.) and to more important losses by evaporation. When plants have been sown on the ridge, they may suffer from drought at the initial stages of their development before their roots have reached deeper layers where the moisture content does not closely follow temporary dry spells. For early planting flat cultivation on land with less than 1% slope will most frequently produce the highest yields. The danger of the water strain may be lessened by planting on the side of the ridge, or by ridging after the establishment of the crop (WALTON, op. cit.).
The choice of the most suitable tillage practice will be governed by the slope gradient and its shape, and the characteristics of the rainfall. Tillage practices which build ridges along the contour, without closing the furrows at regular intervals, do not remove the risks of accelerated erosion caused by large masses of water which collect in the long furrows during heavy storms. Therefore PEREIRA, HOSEGOOD and DAGG (ibid.) emphasize the need of ridging and tying in one and the same operation.
Most experiments on ferralsols have shown the remarkable stability of slopes under intensive cultivation along the contour; there is usually only a slow downhill creep causing accumulation of soil materials and the formation of benches above terraces. If full conservation precautions are taken, the stability of the soils is adequate for arable cultivation on slopes up to 10%. The erosion hazards do not come as much from the plowed fields themselves as from running water collected by road drainage systems, housing, etc. which cannot be kept under control during intensive rainfall.
c. Soil tillage practices improve aeration; if they are applied after clearing, and coincide with a marked change in ecological conditions towards warmer temperature regimes, they will favour the nitrification of the organic matter, and increase nitrogen losses due to leaching. The rate of decomposition of organic matter .can for example be measured by CO2 release from the plowlayer. MEYER et al. (1959) report for ferralsols at Yangambi (Zaire) a period of about twenty days immediately after the first tillage operations, during which CO2 and K are released at a rapid rate. It is therefore important to accomplish the tillage operations as shortly as possible before the time of planting in order that the crops benefit from the available nitrogen; by this practice the losses due to leaching of nutrients are reduced to a minimum. It may be questioned whether such a flash release should be stimulated by aeration in the presence of a considerable labile organic nitrogen pool and under conditions of high temperatures. MEYER (1959) noted that turning the organic matter layer under a subsoil material during plowing, reduced the mineral N content of the soil markedly (from 60 ppm to 8 ppm, 8 days after tillage, in a soil with 0.196 % total N). It should however be avoided that seeds are located in materials which fix phosphorus at higher intensity than the top-soil.
The time of plowing with respect to clearing and sowing also depends on the kinds of tillage operations. Pre-plowing cultivation can help to reduce the time of exposure to direct sunlight. It avoids the formation of large clods or continuous soil slices held together by roots or by compaction, and shortens the weathering period needed after plowing for clod breakdown; it permits the operations for building tilth to be carried out immediately following one another. According to BOSHOFF and HILL (1969) it also reduces the cost of seedbed preparation by some 25% over the conventional technique where plowing is done first.
In deciding on the type of tillage to be carried out, the climate is important. In clayey ferralsols with high nitrifiable organic matter in the hot tropics, minimum tillage or burying of the humus horizon may be recommendable.
In cool climates, in humic ferralsols, or soils under grassland, where C/N ratios are higher, more intensive plowing may be justified, depending on specific purposes, and other management practices that are accomplished simultaneously; the same holds true when incorporating plant residues of high C/N ratios, depending on whether mineral fertilizers are used at the same time, and whether any immediate release of N is expected or if, on the contrary, biologic N fixation is one of the main objectives,
BOUCHARD and RAKOTOARIMANANA (1970) recommend plowing of the soil at the end of the cropping cycle as a means of increasing the structural stability in arable land which has suffered deterioration during clean weeding. They attribute the effect to aeration and drying during the rainless season.
d. The fourth purpose of the preparation of a rootbed is related to the mixing of amendments into the soil, especially in the case of nutrients that are slowly soluble, and do not migrate into deeper layers. Lime is more efficient in soils which lack calcium when it is incorporated at great depth, thus enabling roots to grow into horizons which received calcium carbonate. The crops can then benefit from a greater soil volume and water supply. Results of experiments conducted in ferralsols were discussed in chapter 2.3.4 (i), page 74.
e. Ploughing reduces the need for weeking, and permits to save labour and to keep more strictly to an agricultural timetable. JURION and HENRY (op. cit. page 130) report to have obtained a reduction by ± 60 mandays per hectare by plowing fields where the invasion by Imperata cylindrica was to be controlled. With the advent of pesticides, the economics of such practices are to be reinvestigated.
3.2.1 Restoration of chemical fertility
3.2.2 Restoration of physical properties
3.2.3 Soil and water conservation by fallows
Fallows are convenient management practices for restoring the productivity of the soil in countries where the price of new land and the cost of clearing the vegetation are cheaper than the overcharge due for the conservation of soil fertility in permanent cropland.
Under other circumstances the occupation of arable land by the fallow makes the rest period economically unsound, especially when it has been proved that the productivity can be increased either by fertilizers, irrigation, or pest control. In the case that modern technology cannot be applied, fallows are often the only alternative which is left after that continuous cropping has depressed the yields below acceptable standards.
The purposes which can be achieved by fallows are listed below:
i. replenish the surface horizon with fresh organic matter, in order to increase its nutrient supplying power, particularly of nitrogen, and improve the cation exchange properties. In addition, transform nutrients into more available compounds.Not all fallows are equally efficient in restoring the soil productivity. The time needed to rebuild an agricultural soil may vary from one season to several decades. There are even huge areas which never recovered from deforestation and where the original vegetation was never able to reoccupy the land.ii. develop a root system which draws Ca, Mg, K and other nutrients from lower layers, and concentrates them either in the vegetation, or in the topsoil.
iii. improve the structure of the soil, both by the development of a root system and by the addition of fresh organic matter.
iv. refill the soil with available moisture in order to use it during the growing season.
i. Fallows in rainforest regions
ii. Fallows in savanna regions
iii. Fallows in cool tropical regions
Among the different types of fallows in the hot humid tropics, the secondary forest is the most capable of bringing the soil productivity potential close to its original level. The trees operate by deep rooting and pick up cationic nutrients from the subsoil. Unlike savannas, they restore the readily decomposable organic matter pool without severe losses of N by annual burning. The forest may be very difficult to establish however and its development may be slow.
The efficiency of a forest fallow depends essentially on its ability to create in a short time a vegetation which protects the soil against high temperatures and erosion. The sooner the ecological conditions of a forest are reinstated, the faster the increase in organic matter of the soil. It is obvious that the control of soil losses due to erosion contributes to a better conservation of the gains.
The early stages of forest fallows are the most efficient in rebuilding the soil organic matter. LAUDELOUT (1960) considers that a duration of minimum ten years, and maximum fifteen is adequate for reaching nearly the original organic matter level of the rainforest. At later stages, the older regrowths immobilize the nutrients almost exclusively in the woody parts. In order to obtain a closed plant canopy right from the start of the fallow period, it is usually recommended not to end the cropping cycle with a clean weeded plant. Cassava or bananas, which tolerate mixing with pioneer forest species, are for example suitable transitions from crop to fallow. At the same time, plant residues that are rich in starch enhance N fixation and may accelerate the restoration process.
The importance of the type of crop which comes last in the relation and the ecologic conditions of the surroundings has been stressed by JURION and HENRY (1969). They contend that a weeded crop such as groundnuts at the end of the cropping cycle on medium textured orthic ferralsols in the center of a forest region, does not retard the recolonization by trees, provided that forest strips border the fields, that isolated trees act as seed bearers and perches for birds, and that stumps left in the fields give a quick start to the regrowth by sending up shoots. They obtained rates of spontaneous recolonization which made it unnecessary to try to plant forest species for the only sake of accelerating the establishment of a suitable fallow.
Such favourable circumstances may not prevail at the edges of the rainforest regions, and competition between savanna and tree species may become very harsh. Winds and fires are effective allies of savanna communities for taking advantage of ill-disposed soil properties.
Among the most difficult ferralsols for recolonization by trees are the acric ferralsols, followed by the dystrophic groups which are dominated in the effective base exchange capacity by aluminum. If the oxic horizons are thick (i.e. more than two meter), and when the profiles are freely drained, a short dry period, aided by fire, may favour the establishment of grasses and allow them to impede or to retard drastically the return of the rainforest.
Undeep ferralsols, in which the oxic horizon is underlain by weathering rock which may supply nutrients to penetrating roots, and contain more available water, are among the most suitable for the restoration of surface layers by forest fallows. They are approached in this capacity by the deep xanthic and orthic ferralsols, and only surpassed by the rhodic suborder. The eutrophic phases are usually the most favourable.
There are other soil properties which may retard or completely hold up the reclamation process of surface layers by forest fallows. Erosion causes most damage on convex slopes; sandy topsoils make ferralsols particularly sensitive to drought; strong declivities reduce water intake. These are conditions which should imperatively exclude agricultural uses, which involve clean-weeded crops, in acric, orthic and xanthic dystrophic ferralsols from areas which have no other resources than fallows for the rehabilitation of the arable topsoil, and where a dry season, however short, aided by fire may hamper the establishment of the forest pioneer plants. Even under better circumstances should it be recommended to apply management practices which stimulate the prompt establishment of a secondary forest; i.e. by leaving forest vegetation strips between fields, protection against intensive fires, wind breaks, and erosion control.
The importance of a rapid re-establishment of the vegetation for the efficiency of the restoration process has also been stressed by GREENLAND and NYE (1959). They estimated that a crop to fallow ratio of about 1:3 could maintain the humus level in forest soils at 75% of the equilibrium level.
The length of the fallow period in forest regions will depend on the quality of the soil, and the status which the fertility has reached at the moment the field was abandoned. In xanthic ferralsols which were dystrophic and medium textured, JURION and HENRY (1969) have described systems which allow for three to four years of cultivation followed by twelve to fourteen years of fallow; the crop to fallow ratio in this case was 3:12. How often ouch a sequence can be repeated could not be determined experimentally. The intermixed crops included bananas, cassava, rice and corn.
In areas which mainly consisted of eutrophic orthic ferralsols the best observed crop to fallow ratios were 6:15, 5:12 and 5:15. Some more demanding crops as cotton could be included in the rotation. Eutrophic ferralsols in the rainforest belt that are protected against fire allow about 30% of the arable land to be occupied by crops; dystrophic ferralsols in the same area would only support approximately 20% cropland. This means that farm holdings which are planned for forest fallow rotations, or shifting cultivation, should have five times as much arable land than the area which is actually cultivated, when the soils are predominantly dystrophic. In the case of eutrophic ferralsols the factor would be 3 or 4. The importance of land prices and soil properties in estimating the feasibility of management systems based on long fallows is obvious. Not only direct agrotechnical considerations come into play, but also the costs of building a suitable socio-economic infrastructure. The latter may be prohibitive in the case that the area of the rural community has to be extended fivefold particularly in the case the project has to be integrated into a modern market economy the objectives of which reach beyond traditional subsistance levels.
In order to cope with these problems, efforts have been made to shorten the fallow periods, or to increase the crop/fallow ratio. In the forest zone grasses have been used to replace the tree species. JURION and HENRY (1969) conclude that on the whole Pennisetum purpureum is not capable of maintaining soil fertility, except on eutrophic Rhodic Ferralsols. Other plants, such as Chloris gayana, Desmodium intortum, Canavalia ensiformis, and Stylosanthes gracilis, did not offer any improvement compared to natural fallows. They advocate the use of fertilizers in order to correct the deficiencies caused by continuous cropping, claiming that the effect of chemical amendments is relatively greater than that of organic dressings (JURION and HENRY, op. cit.).
Natural savannas on deep ferralsols only slowly succeed in creating suitable environments for restoring the soil productivity. Compared to forest regrowth, they poorly protect the soils against erosion, particularly when long dry seasons and fires reduce the plant cover to a minimum. The phosphorus and N levels under savanna are usually less than in the rainforest areas. The production of organic matter by grasses is strongly dependent on seasonal rain distribution; it may be high in perhumid climates, but low in regions with a dry season. AHN (1970) estimates that in West Africa the amounts of plant material added annually to the soil would in practice not exceed 2,5 Tons/ha which compares very unfavourably with the 15 to 20 Tons/ha produced by an established forest fallow.
The inefficiency of savanna fallow on ferralsols is often demonstrated by the need to scrape surface soil into individual mounds on which crops are grown, or to concentrate on spots where heaps of fallow vegetation have been burned.
There have been many attempts to improve the efficiency of savanna. Most include the control of fires. JURION and HENRY (1969) report that protecting spontaneous three years old grass fallows from fire on eutrophic orthic ferralsols increases the yields of seed cotton by about 170 kg/ha and of corn by 350 kg/ha dry grain in a subsequent two years rotation ending with groundnuts which did not show any response. Other investigations aim at replacing the natural grasses by herbaceous plants selected for their ability to regenerate the soil, control weeds, and possibly produce forage.
JURION and HENRY (op. cit.) mention Sotaria sphacelata and Brachiaria ruziziensis among the grasses which are adequate in controlling Imperata cylindrica. The legume Stylosanthes gracilis has a high feed value, and is capable of dominating the same weed. The experiments which lead to these conclusions were carried out on eutrophic ferralsols, or soils with a favourable base status. It is not certain, whether similar results could be obtained on dystrophic or allic soils, without the aid of chemical fertilizers. Weed pollution is indeed particularly severe in desaturated soils which are either poor in calcium or rich in exchangeable aluminum, or both.
In the hot tropics, under primitive management, which does not include the use of fertilizers, without effective erosion control, and without limitation or fires, etc. there has been no experience where the spontaneous regrowth of savanna on dystrophic orthic, xanthic or acric deep ferralsols improves the chemical fertility of the topsoil. There is not much evidence either that improved pastures would do any better, as they are usually invaded by weeds and require prohibitive inputs of labour or capital in order to keep the grassland clean.
The same limitations exist for grassland which is sown directly after clearing rainforest, or after a first crop, as it is traditionally practised in South America. Under extensive management systems, the intended pastures can neither resist the pressure of spreading poor savanna species in places where drought and fire impede the return of a forest, nor withheld the establishment of a secondary forest under wetter conditions. The action of the tree fallow is thus substantially retarded. The alternate use of grassland as fallows on dystrophic or acric ferralsols is only justified when it is complemented by fertilizers which take care of the mineral nutrient supply to the plants. The role of the fallow in the latter case is then essentially related to physical problems occurring in the soils after long periods of cultivation, and to the maintenance of organic matter levels.
Grazing of grass leys which are included in a crop-fallow rotation in savanna areas may have beneficial affects upon the subsequent arable crop. The action is particularly noticeable just at the opening of the arable cycle. STOBBS (1969) assumes that most benefits result from the greater quantity of nitrogen which accumulates in the grazed land. There may be some transference and concentration of fertility by moving animals from surrounding permanent grassland into the fallow area. Nutrients may be transformed into more available compounds(defoliation is thought to stimulate plant growth, and increase the efficiency of bringing up cations from lower layers into the surface horizons. Possible reduction of weed growth under heavy grazing is also mentioned as a possible reason for increased yields. The experiments reported by STOBBS had a 3/3 crop fallow year ratio, and production figures correspond to a 20% increase for night grazing, and a 10% increase for day grazing. The description of the experiment carried out by STOBBS does not give detailed information on soils, and it is not known whether the results are applicable to ferralsols with low content of bases. To be efficient there must be some mineral reserve in deeper layers, or suitable nutritional conditions in the topsoil, or transference of fertility from the outside.
There have been no basic investigations on the effects of natural fallows in humic ferralsols of the cool tropical climates. They have usually high organic matter contents with wide C/N ratios and are poorly saturated with bases. The dark humus rich horizons are thick enough, and the only possible benefit from fallows would be to supply the topsoil with bases or with readily decomposable organic matter.
These purposes are not easily achieved in soils with thick oxic horizons. JURION and HENRY (1969) estimate that approximately ten to twenty years of natural grass fallow would be needed in order to restore the soil after a one or two years crop rotation. The crop/fallow ratio would thus be 1/10, the lowest mentioned in primitive agricultural systems.
Most of the attempts of increasing the efficiency of fallow in the cool tropical savanna of Zaire aimed at controlling the invasion of couch grass (Digitaria Vestita), by introducing plant species which would hasten the restoration process. No promising results were obtained neither with Cassia didymebotria nor Setaria sphacelata (JURION and HENRY, op. cit. page 142), whether grazed or not. Setaria sphacelata moreover has the disadvantage that it cannot be ploughed into the soils with the implements which are available to local farmers.
Cropping necessarily leads to the deterioration of soil structure. MOREL and QUANTIN (1964) found in virgin savanna soil instability indexes of 0.4. Sandy topsoils degraded after two years, medium textured soils after four years, reaching indexes of about 1.5. Longer cropping periods lead to indexes of 2 or higher.
Fallows of running grasses, or with superficial rooting were able to restore the index to 1-1.3; crest grasses with deep rooting habits could achieve better remits and bring the structural index back to 0.8-0.4. Cover crops, such as Stylosanthes and Pueraria were rather poor in improving soil structure, except for Gajanus indicus which accomplished a better task in the amelioration of physical conditions.
If the structural instability has not reached extreme values, such as 2, it is suggested that 3 to 4 years of a natural grass fallow will be sufficient to correct structural deterioration.
The present experience thus seems to indicate that natural fallow communities including deeply rooted erect grasses are the most suitable for restoring soil structure. They improve the ability of the soil to accept rainfall and to transmit water, primarily by increasing the volume of freely draining very large pores and channels, (20 cm water tension, PEREIRA et al. 1954). In ferralsols they do not affect the distribution of the finer pores, for example those which are filled at field capacity. The influence of grasses on structure is mainly one of improving on soil aeration; according to PEREIRA et al. (1954) they do not confer continuing advantages, and the soil returns to its unfavourable state after the first year of cultivation. This is probably due to the fact that the better physical conditions are essentially the result of an increase in the amount of the large pores which are necessarily the most fragile.
Natural grass fallows however are very demanding on soil moisture and a dry season may deplete a 3 meter deep profile of all available water. PEREIRA et al. report that soils kept bare during the same time still contained 230 mm of available water. Such severe water deficits have deleterious effects on the following crops, especially if the rainfall distribution at the beginning of the growing season is erratic.
Grass species differ markedly in their ability to protecting the soil against erosion. The bunch grass, like Panicum Maximum are less suitable to reduce run off and soil losses than stoloniferous or sod types. SMITH and ABRUNA (1955) indicate that Melinis minutiflora, once it is established, provides excellent soil protection. Generally the losses during seedbed preparation and the early seedling stages of the grasses are higher than the total for several years after grass establishment (op. cit.).
The most difficult task for fallows is to restore the available water content of the profile. Only bare fallows may achieve such purposes in tropical regions. The difficulties to control soil erosion on a bare surface render this type of fallowing a dangerous technique however. PEREIRA et al. (1958) showed that volunteer covers may remove all available water in regions with a single rainy season where precipitations amount to approximately 500 mm, and where open water evaporation from a 120 cm diameter sunken pan averages 2108 mm per annum.
Natural fallows which are composed of indigenous plant species have usually deep rooting systems which exhaustively extract water from the entire profile. Introduced grasses with shallow rooting habits, when properly sown may suppress volunteer regrowth and afford some protection against erosion, without depleting the available water in deeper horizons.
3.3.1 Nitrogen fertilization (BARTHOLOMEW, 1972)
3.3.2 Phosphorus fertilization
3.3.3 Potassium fertilization
3.3.4 Lime requirements
i. Crop use requirements (NM)
ii. Natural supply of nitrogen (NS)
iii. Use efficiency of fertilizers
The quantity of fertilizer nitrogen (NF) which should be added to a soil depends on:
i. the amount of NM required by the crop in order to achieve a possible maximum yield. The maximum crop production is usually determined by limiting factors other than nutrients, such as environment, diseases, genetic plant characters, etc. (NM)Taking into account the first three factors, the following equation expresses the foregoing statements:ii. the nitrogen which is supplied from natural sources and absorbed by the crop during the growing season (NS).
iii. the efficiency of the soil-plant system in using the added fertilizers (f = fraction of fertilizer N added which is adsorbed by the plants).
iv. the cost/benefit relationship between the expenditures for the added fertilizers and the profits related to the increase in production.
A brief review of the present knowledge regarding the adequate techniques for evaluating NM, NS and f is given below. Current management practices act on all three components.
The average crop use requirements for corn, rice and wheat have been estimated by BARTHOLOMEW (1972), using experimental data obtained both in tropical and temperate regions. They are illustrated in fig. 16. The slope of the curves was defined by the following regressions:
DN = 30 +1.3 I for cornwhereDN = 58.1 + 0.005 I for wheat
DN = kg N needed to obtain a one ton increase in yieldI = the check yield in tons/ha from which the increase was measured
Fig. 16 - Nitrogen requirement a of cereal crops (BARTHOLOMEW, 1972)
The equations fit curves which indicate the trend toward greater nitrogen requirements per unit yield increment at high production levels. For rice it was estimated that 43 kg N were necessary to produce a one ton increment in brown rice.
Crop use requirement curves for other crops have not been computed; it should be noted that the equations consider the nitrogen needs of the entire plant, and not Just the export of the harvested product. It is also worth mentioning that in BARTHOLOMEW's approach it is assumed that the N fertilizer response is essentially rectilinear up to a maximum yield level beyond which the response to fertilizers is either zero or negative. This maximum obtainable yield defines a ceiling of response. Its position may be modified by acting on other nutrient levels, or it is determined by climatic conditions, water supply, etc.
It is very difficult, if not impossible, to predict the amount of N which will be supplied to the crop by the soil environment during a forthcoming season. The former history of the land, the weather conditions and the management practices cause considerable modifications in the rate of nitrogen release by the soil. They cannot be taken into full account by whatever laboratory methods.
Incubation procedures most closely approach the potential rate of mineralization of the organic matter in the field. They are time consuming however. Less reliable methods which chemically determine extracted fractions of inorganic N are more expeditive; for example the measurement of the amount of ammonium and nitrate which has accumulated in the soil during the dry season, or during a rest period, have been found to relate to actual mineralization rates of soil organic matter, including the crop residues. LATHWELL et al. (1972) found that the total N extracted by a 0.01 M CaCl2 and K2SO4 boiling solution highly correlated with the amounts of N mineralized during varying periods of incubation.
It is an absolute requirement that the laboratory diagnostic methods be calibrated against crop response measured in the field; since this experimental work has only a limited domain of applicability, the soil testing procedures themselves only give satisfactory results when they are restricted to well specified soil and cropping conditions.
The amounts of nitrogen supplied by the soil-plant system may also be evaluated directly from yields that are obtained in field check plots which do not receive fertilizer N. Traditional production levels may be a basis for assessing the natural NS supply, provided no other nutrients are limiting.
BARTHOLOMEW (1972) proposes such a system. For example the amounts of nitrogen used by corn which produces two tons would be approximately 50 kg per hectare (see figure 16). In the absence of fertilizers, this quantity would be a reasonable estimate of the quantity of natural nitrogen which is absorbed by the crop during the growing season. Tables which are established locally by observation on given crops and sites could become satisfactory guides for assessing the importance of the natural supply processes. They could also be used as standards for calibrating soil test procedures when a great variety of field conditions occur. More detailed discussions of the method can be found in BARTHOLOMEW's publication (1972).
It can be estimated from figure 16 that corn uses 140 kg more N per hectare to produce 8 tons than to produce the traditional amount of two tons per hectare. This supplement should be given as fertilizers to the soil. Only a part of it (f) is taken up by the crop however.
The importance of nitrogen losses by leaching in ferralsols has been discussed already.
The efficiency of nitrogen fertilizers can be increased by adapting the time of application to the growth pattern of the crop and to the periods of nitrogen supply by the soil. After a dry season there is usually an increased nitrogen release by the organic matter at the moment the rains start. No nitrogen fertilization is needed at that time. Later on split applications will therefore usually result in a better utilization of the fertilizer. The kinds of nitrogen carriers are also important, and slow release fertilizers may give good results. Placement should be such that the N reaches the active root zone when it is most needed by the plant. The moisture regime largely determines the time and the placement techniques. Band applications are only necessary when it is expected to have favourable interactions with phosphorus.
Since nitrogen uptake by the plant occurs mainly by mass flow through the transpiration stream, and because under tropical conditions the soil water movement in ferralsols is essentially rain dependent, split applications and slow release fertilizers should normally achieve the highest efficiency. AHN (1970) reports that split applications of nitrates to annual crops at one month and at two months after planting were more efficient than single applications either at planting or at two months after planting.
Nitrogen fertilization should not be in excess of phosphorus availability. The N:P ratios which have been proposed by DABIN (1967) are given in the following chapter.
Phosphorus moves to the roots mainly by diffusion through the soil water. It is the concentration of P in the soil solution which defines the rate of movement of phosphates to the root, by establishing a concentration gradient. This concentration is therefore called the intensity factor in phosphorus nutrition.
The solid phase must be able to provide sufficient phosphorus to the soil solution in order to avoid its depletion by the uptake of P by the crops. The quantity of P which is available for replenishing the soil solution is called the capacity factor.
There are several chemical reactions and adsorption processes which govern the equilibria between P in solution and the active P in the solid phase. Adsorption isotherms have been proposed to describe quantitatively these phenomena. The curves which have been obtained by FOX (1973) have been reproduced in figure 15, page 80.
In order to maintain a concentration of 0.2 ppm, which would be adequate for most crops, the Orthic Ferralsol (fig. 15) should have 380 ppm of sorbed P. This corresponds in a 20 cm thick surface layer of 1.3 bulk density to 988 kg per hectare of sorbed P, or 2261 kg P2O5.
How much of this is present in the soil, and to what extent it can provide phosphorus to the soil solution which is in contact with the roots, is a problem which has locally been solved by quick chemical extraction methods coupled to field experiments. There are no soil test procedures which have general applicability however.
CATE and NELSON (1965) mention tentative critical levels of P extracted from soils by different methods. The amounts of P below which the probability of response to fertilization is high, were:
|
6 ppm |
: Bray N° 1 method (0.1N HCl + 0.03N
NH4F) |
|
30 ppm |
: Bray N° 2 method (0.025N HCl + 0.03N
NH4F) |
|
18 ppm |
: North Carolina method (0.05N HCl + 0.025N
H2SO4) |
|
10 ppm |
: 0.1N HCl |
|
22 ppm |
: 0.7N HCl |
|
1.2 ppm |
: H2O extract. |
The management of P nutrition in ferralsols is closely related to organic matter content and to the nitrogen supply. BOYER (1970) mentions that for soils with a pH higher than 5.5, the total N to total P ratio should be between 9.1 and 4.6 DABIN (1967) indicates that a ratio of r = total N/available P of more than 22.9 would correspond to a shortage of nitrogen; a ratio of less than 45.7 would produce phosphorus deficiencies and consequently a poor utilization of nitrogen.
It is seldom possible to satisfy in one operation the phosphorus sorption capacity of the arable layer in order to achieve an adequate P concentration in the soil solution. The investment is financially too high to be borne completely in one growing season. Therefore management tends either to concentrate the fertilizer by placing it close to the seeds or the roots, or to granulate it as to reduce the contact area with the soil, or to block by other chemicals the fixation capacity of the soil.
Lime used in such amounts as to neutralize the exchangeable aluminium also increases the effectiveness of P fertilizers; it may also accelerate the decomposition of organic matter, and in this way contribute to the phosphorus supply from soil sources; where aluminium is dominant in the exchange complex, as in the younger members of the ferralsols, phosphorus fertilizers which contain both silica and calcium are usually most efficient. Basic slag and Rhenania phosphates give usually very satisfactory results in such soils.
Single superphosphates are most commonly used in typical ferralsols for immediate crop response. Less soluble forms are more convenient for soil fertilization than for crop fertilization.
Nutrient deficiency symptoms become noticeable in most crops when the amount of exchangeable potassium is less than 0.10 meq per hundred. gram soil. This critical level constitutes an absolute minimum which is valid for many soils, including the ferralsols. Only a few plants as cassava may be productive at lower contents (0,06 meq K per 100 g, ROCHE et al. 1959).
Sandy topsoil ferralsols are the most deficient in potassium. BOYE (1962) only found 0.02 meq K in oil palm plantations, where the trees responded to applications of 1 kg KCl per tree, raising the yields from 250 kg to 2000 kg oil per hectare.
Many crops present potassium deficiencies at higher percentages however. This may be due to a lack of sufficient soil volume for the roots to explore; soil chemists have rather stressed the importance of the balance of potassium with other nutrients, particularly magnesium and calcium.
Although only small quantities of exchangeable potassium are necessary to maintain suitable concentrations in the soil solution of ferralsols, it is usually recommended to have more than two percent of the sum of the exchangeable bases as potassium.
Comparable requirements for robusta coffee on ferralsols have been formulated by FORESTIER (1964), who parallels the cation exchange capacity with the clay plus silt content. He recommends critical exchangeable potassium levels of 0.12 meq K per 100 g soil, when there is 20% clay plus silt, but 0.50 meq K for ferralsols containing 65% particles smaller than 20 microns.
BOYER (1972) in his review of potassium in tropical soils states that a magnesium to potassium ratio of 3:1 seems to be favourable to the majority of crops. Ferralsols planted with oil palm should have Mg:K ratios greater than two, and Ca/K ratios of more than five (JULIA, 1962)(desequilibria resulting in excessive potassium uptake occur frequently when the exchangeable K content exceeds 1 meq per 100 g soil (FRANKART and CROEGAERT, 1959).
Fallow vegetation forms an appreciable reservoir of potassium. LAUDELOUT (1961) found that a thirty years secondary rainforest releases upon burning approximately 130 kg K per hectare; in the Yangambi ferralsols the exchangeable K was raised from 0.067 to 0.325 meq K per 100 g soil after clearing operations which included fire.
Leaching of potassium fertilizers increases in the following order: potassium metaphosphate, potassium chloride, and potassium nitrate (AHMAD and DAVIS, 1970).
Losses of potassium can be reduced by adjusting the time of application to the needs of the plants, for example by split applications. ROOSE et al. (1970) nevertheless estimate leaching losses to 50%. Well developed rooting systems are the best barriers to potassium lixiviation.
DE FREITAS, PRATT and VETTORI (1968) compared different rapid laboratory methods for the determination of the lime requirement of soil by calibrating the tests against the amount of CaCO3 needed to raise the pH of a soil sample during incubation. The time to reach the equilibrium pH was approximately six weeks.
The results of their experimental work on samples of ferralsols are given in tables 16 and 17.
It can be seen that the Ca(OAc)2 method (VETTORI, 1948), would bring the pH between 6.1 and 6.5; the WOODRUFF (1948) procedure would achieve the same results; adding calcium as to reach a base saturation of 86% (CATANI and GALLO, 1955) most frequently raises the pH to values between 5.6 and 6.0. The KCl method or the neutralization of exchangeable aluminium has no consistent influence on pH.
The authors stress the importance of the pH with relation to phosphorus availability which would be optimum near 6.0, Molybdenum uptake increases at high pH, and organic matter decomposition would be the fastest at pH 6.0. Aluminum is neutralized completely at pH 5.5, and there would be no toxic effects of manganese at pH close to 6.0.
The tables also indicate that the lowest amounts of calcium are those which only aim at neutralizing the exchangeable aluminium. Under many circumstances the minimum amount of lime is the only economically feasible for many farmers, and according to present experience it permits to obtain satisfactory yields with most crops.
The time of reaction to reach an equilibrium pH in the field left under natural vegetation takes at least six months in dystrophic heavy textured Latosol Roxo of Brazil (MUZILLI et al., 1969). The residual effect of lime, especially when it is only added in order to reduce the aluminium activity, is not well known. It would probably not exceed one year, and repeated applications are probably necessary in soils having more than the modal amounts of exchangeable aluminium.
Liming requirements do not depend only on inherent soil properties. The acidifying effects of fertilizers must also be neutralized. It is for example usually recommended to use one ton of lime per ton of ammonium-sulphate. Equivalent acidities of most common fertilizers, which indicate the amount of calcium carbonate required to neutralize the hydrogen ions released by 100 kg of fertilizers have been published elsewhere. Data from AHN (1970) are given in table 18.
The importance of incorporating lime into deeper layers has bean discussed in chapter 2.3.4 (i), page 74; deleterious effects on the structure of ferralsols having a net positive charge has been mentioned in chapter 2.2.1, page 49. Specific adsorption of calcium and blocking of exchange sites and a subsequent decrease of the CEC have been studied in Hawaii (UEHARA, SWINDALE and JONES, 1972).
Table 16 LIME REQUIREMENT (meq per 100 g) OF FERRALSOLS ACCORDING TO VARIOUS LABORATORY AND INCUBATION METHODS (DE FREITAS et al. 1968)
|
soil No.
|
pH after incubation |
(OAo)2 |
Woodruff
|
Base saturation
|
KCl
|
||
|
5.5 |
6.0 |
6.5 |
|||||
|
1 |
1.6 |
2.1 |
2.7 |
2.4 |
2.2 |
2.0 |
2.7 |
|
2 |
1.9 |
2.8 |
3.6 |
3.0 |
3.2 |
2.4 |
2.8 |
|
7 |
2.3 |
4.8 |
7.2 |
6.2 |
5.0 |
3.6 |
0.0 |
|
8 |
0.0 |
1.0 |
2.0 |
3.2 |
2.8 |
1.1 |
0.0 |
|
9 |
3.7 |
5.8 |
8.0 |
8.2 |
7.2 |
7.0 |
2.8 |
|
10 |
1.2 |
3.0 |
4.1 |
4.4 |
4.4 |
3.8 |
2.6 |
|
16 |
3.7 |
5.4 |
8.0 |
6.4 |
5.6 |
5.6 |
2.8 |
|
18 |
2.4 |
4.2 |
6.0 |
5.8 |
5.2 |
5.0 |
2.9 |
|
19 |
3.2 |
5.1 |
6.6 |
5.2 |
4.8 |
4.4 |
1.7 |
|
20 |
4.8 |
7.5 |
11.0 |
8.6 |
7.0 |
7.2 |
2.8 |
|
pH |
Ca(OAc)2 |
Woodruff |
Base saturation |
KCl |
|
6.6-7.0 |
1 |
1 |
1 |
0 |
|
6.1-6.5 |
8 |
5 |
3 |
1 |
|
5.6-6.0 |
1 |
4 |
7 |
3 |
|
5.1-5.5 |
0 |
0 |
0 |
4 |
|
4.5-5.0 |
0 |
0 |
0 |
2 |
|
Fertilizer and Formula |
Equivalent Acidity |
Equivalent Basicity |
|
NITROGEN |
|
|
|
Sodium nitrate (NaNO3) |
|
29 |
|
Ammonium sulphate
(NH4)2SO4 |
110 |
|
|
Ammonium nitrate (NH4NO3) |
60 |
|
|
Calcium nitrate (Ca(NO3)2) |
|
21 |
|
Urea (CO(NH2)2) |
80 |
|
|
Calcium cyanamide (CaCN2) |
|
63 |
|
Ammonium chloride (NH4Cl) |
120 |
|
|
PHOSPHATE |
|
|
|
Di-calcium phosphate (Ca(HPO4) |
|
25 |
|
Rock phosphate |
|
variable |
|
Basic slag |
|
variable |
|
POTASSIUM |
|
|
|
Potassium nitrate (KNO3) |
|
23 |
1/ The equivalent acidity is the number of kg of calcium carbonate required to neutralize 100 kg of the fertilizer; the equivalent basicity shows the neutralizing capacity, expressed as kg of CaCO3 of 100 kg of the fertilizer.
The topographic location of most ferralsols seldom allows irrigation; they frequently occur on elevated plateaus with deep water tables where the cost of bringing water are prohibitive.
There are a few examples however where well drained red ferralsols have been used for irrigation agriculture. CIAT (Centro Internacional de Agricultura Tropical, 1972) has reported its experience on acid highly leached ferralsols of the Llanos Orientales of Colombia, which were cropped with flooded rice. Their findings are briefly summarized below,
Irrigated rice production faced the problem of the appearance of a physiological disease called "anaranjamiento" or orange leaf disease. The symptoms are described as follows, (CIAT, 1972):
"Flooded rice grown in the llanos normally appears rather healthy and green during the first month of growth. During the second month, however, the plants become stunted, have insufficient tillering and the leaves start turning yellow to orange. Typical "anaranjamiento" begins with yellowing at the tip of the lower leaves, progressing down the leaf, especially along the margins, and moving up the plant to the higher leaves. The lower leaves eventually dry up and die ...»"It is postulated that the disease is due to the toxicity of a product formed by the reduction of the well drained ferralsol which is rich in free iron oxides."The roots of affected plants are generally short with few rootlets and are covered with a red iron oxide deposit. Sometimes the root tips are slightly enlarged and dark red. Most roots seem inactive."
The effects of flooding on Eh and the Fe concentration were measured. With irrigation the Eh decreased from 545 mV to a constant value of about 90 mV in 10-15 weeks. It was assumed that the free iron content buffered the soil solution at relative high Eh, and no potentials low enough to reduce sulphates were ever reached.
After several weeks the concentration of Fe in solution rises to a maximum of 300-350 ppm, but drops to a constant level of 150-200 ppm because of the influence of increasing pH during flooding. The Bin concentration was never higher than 2 ppm. No direct Fe or Mn toxicities seemed to be involved.
The conclusions of the CIAT investigations regarding the origin of the disease and the soil management problems related to irrigation of well drained iron rich ferralsols included the following points:
a. "Orange leaf disease is not a direct Fe toxicity since it may occur at relatively low Fe concentrations in soil solution, resulting in relatively low Fe levels in the plant. However it seems to be caused by damage of the root system by a reduction product, most likely Fe. The deposition of Fe oxide on the outside of the root not only limits root growth but also prevents the uptake of nutrients, especially P. In a soil already low in plant nutrients this limited uptake ability of the plants leads to an imbalance between supply and demand. In a large plant with a large demand, a small increase in Fe concentration in the soil solution and subsequent coating of the roots leads to a shortage of nutrients. The plant compensates for that by translocation of nutrients from the lower to the higher leaves resulting in oranging and early senescence of the lower leaves.For that reason, healthy plants, grown under conditions of high P and N or low Fe, are always first and most severely attacked by "anaranjamiento", once the Fe concentration in solution starts to build up. Similar conditions of restricted root growth in plants grown in too small a pot, plants grown on compacted soil, or plants from which the roots have been cut, will lead to the same "anaranjamiento" symptoms in other than the llanos soils. Since "anaranjamiento" is a root problem, and primarily a result of extreme P deficiency, foliar applications of P eliminate the symptoms."
b. "In soils with a rapid build-up of soluble Fe, the plant remains stunted from the beginning and there is no need to balance the top growth with the limited nutrient supply. In this case no typical "anaranjamiento" develops, but the plant may suffer from direct Fe toxicity."
c. "A slow reduction results in good initial plant growth, but the subsequent late occurrence of the Fe peak results in severe "anaranjamiento" and a considerable reduction of grain yields."
d. "The severity of "anaranjamiento" can be reduced by a combination of water management and fertilization practices designed to maintain a low level of Fe and a constant supply of soluble nutrients in the soil solution."
e. "A build-up of soluble Fe can be prevented by intermittent or rotational irrigation. A low Fe concentration during flowering, obtained by mid-season drainage, is advantageous for grain formation. Constant flooding with internal drainage maintains a low Fe level, but the loss of nutrients in the drainage water makes it counterproductive. Preflooding for three weeks has the advantage of passing the Fe peak before seeding or in the early stages of growth, but has the disadvantage that P applied at seeding is more rapidly fixed. For that reason preflooding for more than three weeks is not beneficial." (CIAT, 1972).
