Contents - Previous - Next

Plant nutrition management in farming systems: Interaction with efficient water use at watershed level


In most farming systems, the land is organized for the purpose of agricultural production in three main areas: the cropped area; the grazed area; and non-arable areas which include forests, rangelands and agriculturally-unproductive land. The ecological conditions prevailing in these three areas, the socio-economic conditions and the available know-how and genetic resources largely determine, through complex interactions, the geographical extent of these areas in a region. The historical patrimony is always very important and, in general, it is difficult strictly to connect the present development of the areas with the current conditions of farming. Their management by farmers, herders and foresters/hunters has an important impact on the availability of plant nutrients and water for the cropping systems implemented in the cropped area. Many traditional farming systems have developed differentiated use of these categories of land in which the transfer of water, biomasses and plant nutrients is organized.

Under conditions of low population densities, slash and burn practices are used in order to make available, through the burning of natural biomasses, a large part of the plant nutrients stored in the natural vegetation during long-term fallows. The productivity of these plant nutrients is low: run-off disperses a large quantity of them and leaching evacuates many of the nutrients in excess of what the crops can take up under the generally-extensive management of cropping systems. However, the infiltration of the soil usually remains quite high and the transfer of water within the landscape is limited where the proportion of cropped area is also limited.

Replacement of forests by perennial crops, industrial crops or traditional food crops (bananas, oil palm trees and other fruit trees) is less detrimental to the water balance and creates less plant nutrient losses in the watershed than slash and burn practices, when the population density increases. However, the capacity for replenishment of the plant nutrient reserves of the soil disappears through perennial cropping, because the period of storage through long-term fallows no longer exists. Cover crops may assist in providing some nitrogen if sufficient light is provided to the crop. Intercropping of annual crops between perennial crops at an early stage of the plantation is a valuable method for reducing the leaching of plant nutrients derived from the destruction of the natural vegetation. It also increases the efficiency of labour, sharing in particular the cost of weeding between the perennial and the annual crop, especially when pruning of the former is done at the same time (e.g. coffee, nuts). However, more water and more nutrients are consumed, which is detrimental to the perennial crop if there is a shortage of one of these two inputs (as with coffee plantations in Western Cameroon).

A. Angé, Chief, Plant Nutrition Management Service, FAO, Rome

With a further increase in population density, perennial use of large areas for crop production takes place, in rotation with fallow periods of varying duration. Farmers then have to face the problems of both water management and plant nutrition management for sustainable use of the land. In very rainy areas, the evacuation of surplus of water is the main problem, while in semi-arid tropical and temperate areas the shortage of water is a risk, as sometimes is the excess of water. In humid areas, the leaching and run-off of nutrients and the low efficiency of nutrients created by waterlogging are factors currently limiting crop yields. In other areas, insufficient water supply creates stress for the crop and lowers the efficiency of available plant nutrients, while leaching and run-off may also occur.

The competition for resources in the same area, by various social groups, is often a factor of destabilizing the proper management of land, water and nutrients. Thus, the design of improved general management often depends on negotiations between these groups resulting in trade-offs for access to the resources. In most cases, successful integration of various land use patterns, optimizing the management of water and nutrients, is the result of a long historical process, governed by powerful local hierarchies and well-established rules. These successful examples are adapted to a set of market conditions, production techniques and social attitudes, and there are many examples of disruption of ancient equilibriums, superseded by the evolution of population densities, social rules and political conditions, and by the modification of economic conditions. The development of rice paddies in Asia and of irrigated terraces in the Middle East and the Andes are examples.

The export of crop harvests mines the plant nutrient reserves in the soil if they are not replaced by natural supply from rain and irrigation water, sedimentation, dust, weathering of soil minerals and big-nitrogen fixation. Farmers have therefore thought up various methods for sustaining the availability of plant nutrients for their crops. There are two traditional ways: fallows and the transfer of nutrients. Fallows provide the accumulation of the natural plant nutrient supply over a certain period of time in order to make such a stock available for the following period of cropping. It must be noted that fallow periods also recreate better water management conditions through the alleviation of the crusting of the soil surface, the improvement of soil porosity by insects and deep rooting by pluriannual species, and replenishment of the soil organic matter content. Fallows imply the management of an area with the rotation of the cropped area in space and time. The extent and spatial organization of fallows have an important impact on the circulation of water in the watershed, by their influence on infiltration and run-off. Plant nutrients are also transferred from non-cropped areas to cropped areas: manures resulting from the grazing of forests, pastures and fallows; harvesting of natural straws, litter and leaves from forests and woodlands; and the transportation of topsoils.

Such forms of transfer have, in the long term, a great impact on water circulation in the watershed, if they exceed the natural renewal of these resources. Thus, the resulting denudation of hills in China and Vietnam and the overgrazing of large areas in the Andes, the Sahel and soudanian areas of Africa have dramatically increased run-off and finally degraded the cropped land, also threatening fertile alluvial soils and irrigated areas through flooding and abundant coarse sedimentation. The transfer of plant nutrients extends the management of those nutrients from the local cropped area to the whole territory concerned by the transfer. Plant nutrient management is then no longer a matter of the nutrient availability in the fields but is a problem of plant nutrient harvesting in farming systems and in a collective area.

In this respect, the connections between livestock systems and cropping systems are interesting. The organization of cropped and grazed areas is largely dependent on the social organization for the management of the livestock system. With very limited livestock, the organization in open fields will be prominent. When livestock systems are more important, hedges control the circulation of animals and also divide the rural space into management units, if the availability of the labour force is not sufficient to keep the herds away from the crops, or if the social organization is not adapted to the control of large herds by few herders, as in many semi-arid countries. When the control of animals is properly organized, open fields may co-exist with various levels of development of the hedges. However, the landscape is organized to preserve the access of animals to the essential resources: fodder in the rainy season in the upper parts of the landscape, fodder during the dry season in the lowest part of the landscape, access to drinkable water, access to crop residues after harvest. In most cases, access of farmers to the manures from these herds is well-regulated. The corresponding management of the landscape has important implications for the water circulation in the catchment. Maintaining the upper and lower parts of the watershed under natural vegetation protects it against erosion, but herders frequently set the vegetation on fi re, expecting new growth, and destroying the natural habitat of pest flies. This reduces the protecting effect of the vegetation and causes extensive losses of plant nutrients.

Intensification of livestock systems, through the collection of crop residues and the production of fodder, further modifies the management of the watershed and the balance of plant nutrients in the landscape. The denudation of the soils from intensive harvesting of crop residues facilitates runoff from rains coming before new planting. In contrast, the development of artificial pastures is an efficient method for controlling run-off. In most cases, the two practices concentrate the plant nutrients in the limited area where manures are spread. However, the development of leguminous fodder crops has an important role in restoring the nitrogen balance in the watershed.

Many traditional farming areas, especially in Africa and Latin America, are organized in concentric territories. The first of these, close to the houses, receives most of the domestic wastes and manures, and most of the available labour. This is the high intensification ring where, in general, food security is the primary issue for the farmers. The productivity of water for crops is very high, influenced by the ready availability of plant nutrients. In the second ring, peripheral to the first, distribution of agricultural inputs may occur and the land is used intensively. The bulk of agricultural products is generated there, but the land may be over-used. The productivity of water for crops is lower than in the first ring, especially if plant nutrients are not sufficiently available. Most crop residues are exported from these areas. The external ring, at some distance from the farms, is more extensively used and may produce quite good yields if not located on more marginal lands. However, the productivity of water is generally low because the supply of labour is usually insufficient. Thus, the distribution of labour in the watershed has important implications for the productivity of water, when combined with the regime for the supply and export of nutrients.


The relationship between the availability of water and the crop yield in a particular set of ecological conditions depends heavily on the management of the cropping system. In Senegal, on the pearl millet crop, the yield is linearly-correlated to the index of satisfaction of the crop for water supply (the ratio of real evapotranspiration to maximum evapotranspiration). This correlation depends on the level of intensification. On farmers' fields, where crop density, weeding and plant nutrition practices are of poor quality, the impact of water availability is less than in the research station. Therefore, the relationship between water availability, plant nutrition and crop yield must be related to ecological conditions and cropping system management.

The total availability of water has a direct impact on crop yield. It is well known that crops are particularly sensitive to the availability of water at special growing stages, such as at flowering time for cereals. However, the impact of water shortages and droughts is frequently exaggerated. Plant nutrient depletion in soils reduces their productivity, and aggravates the impact of water shortage on yields.

The severity of plant nutrient depletion can be roughly evaluated through apparent plant nutrient balance sheets using all identified plant nutrient supply sources and considering only the export of plant nutrients by harvested products and residues. This underestimates the export side because losses from runoff, leaching, volatization or immobilization are not considered.

In a case study from Central Senegal (Rabot, 1984), with an average annual rainfall of 700 mm, the influence of runoff and immobilization of plant nutrients is negligible, but leaching of N is important. The trial started in 1962 with a crop succession of groundout, sorghum, groundnut and fallow, transformed later into a rotation of groundnut, maize, cotton and sorghum, and ended in 1980. All crops of the rotation are cultivated each year in parallel plots, in order to give a clear picture of the influence of the cropping season on the yields. The trial compared a series without fertilizer supply and not tilled, in which all crop residues were exported, and a series with recommended applications of fertilizers and tillage in which the cereal straw and cotton residues were burnt and the groundnut residues exported. Plant nutrient exports and N2 fixation were evaluated throughout one rotation in each series.

In the series without fertilizer supply, the average yield of sorghum grain during 16 years was 0.75 t/ha, with large fluctuations. The 5-year average trend from 1965 to 1980 showed a decrease following an S-curve, from about 1.1 t/ha to about 0.45 t/ha. This reveals the development of a major limiting factor. Comparison between the annual rainfall and the yields cannot by itself explain the main trends of the yields. The moving 5-year average value of the annual rainfall shows a decrease from 880 to 680 mm, but the annual variation is not synchronized with the variation of the yields and is too small to explain the variation in yields, in spite of five rather severe droughts in 1968, 1972, 1973, 1977 and 1980.

In the series with fertilizer supply and tillage, the average sorghum yield over 16 years was 2.57 t/ha. Yields fluctuated around the average value +50 percent and -40 percent. The 5-year moving average of yields fluctuated between more than + 30 percent and less than -13 percent of the average yield, but there was little if any downward trend and the variation was not synchronized with the variation of the moving average for the annual rainfall. The effect of droughts on the yield is clear but irregular, and cannot explain the whole variation. The occurrence of droughts since 1972 did not modify the variability of yields around the average value as was the case without fertilizers. Therefore, the droughts alone cannot explain the major decrease of the yields observed on the plots receiving no fertilizers.

The apparent plant nutrient balance sheets assist in the interpretation of the data. On the plots receiving fertilizers, phosphorus and potash were supplied far in excess and cannot be limiting. The nitrogen balance was negative between 1971 and 1976. Therefore, nitrogen depletion could explain the low yields obtained during this period. When plant nutrients were correctly applied in this ecological condition for sorghum, the droughts which occurred between 1965 and 1980 had a limited impact (between 15 percent and 25 percent of the very high yields obtained continuously during this long period).

On plots receiving no fertilizers, plant nutrient depletion in the soil was severe, even though the rate of plant nutrient depletion is less when the yields are severely limited by the very low fertility of the soil. Even apart from leaching and other losses, net outflow from the capital of plant nutrients during the 16 years of the experiment exceeded 130 kg/ha of P2O5, 220 kg/ha of N and 220 kg/ha of K2O. This trend in plant nutrient depletion explains the decrease of the yield. Thus, the drought has a worse effect on yields when plant nutrient depletion is considerable, (which is not the case in the plots receiving fertilizers), thus increasing the losses of production mostly created by plant nutrient depletion. The drought makes the system less productive, but the vulnerability of the system is caused by plant nutrient depletion.

Attempts have been made to correlate crop yields and the position of the field in the watershed. As explained above, water availability, soil fertility and cropping conditions influence crop yields through multiple interactions. Therefore, particular attention to the experimental conditions is required, if the impact of water availability and soil conditions on yields is to be evaluated. Thus, a multifocal experiment was developed in a watershed in Central Senegal (1982-1983) in order to evaluate the distribution of the effect of fertilizers, manure and supply of potash and lime, in addition to the recommended fertilizer doses on millet in farmers' fields. Only the fields belonging to the supervisors of the farms, in regular millet-groundnut rotation, were selected. Those plots are prepared by ox-ploughing, sown mechanically and properly weeded. There is no supply of mineral fertilizers. A very detailed soil map was elaborated (1:10 000 scale) and the plots were grouped according to soil and landscape units.

The crop response to fertilizers was clearly correlated with the position of the field in the landscape. The control yield and the response to fertilizers were very poor in shallow soils on the upper part of the slope (where there are plough pans), on which the availability of water has been the limiting factor. On the middle part of the slope, where waterlogging was marked, the control yield was quite good, but the response to fertilizers was negligible. On the lowest but well-drained part of the landscape, the transfer of water from upper parts creates extremely favourable conditions for the crop. The transfer of nutrients over the years by the water flow is responsible for very high yields of the control, but the effect of plant nutrient supply from manure and fertilizers is highly significant as well. Plant nutrient availability and effect are influenced by water circulation in the landscape.

It can be concluded that the management of plant nutrients depends heavily on farming system management and on water availability. Factors determining the impact of plant nutrition management operate at different scales: the plot, the farm, the slope section, and the village territory.


Rabot, C. 1984. Vingt années de successions de cultures dans la moitié sud du Sénégal, impacts écologiques. DEA, écologie tropicale, université des sciences et techniques du Languedoc, Montpellier. 38 p. + annexes.

Contents - Previous - Next