EXISTING CASSAVA PRODUCTION/ENVIRONMENT KNOWLEDGE BASE

Technologies and crop/soil management systems

Crop/soil management to prevent nutrient depletion

Slash-and-burn

Over the ages farmers have developed ways to produce food while maintaining soil productivity. Some of these "sustainable" production systems were based on slash-and-burn agriculture, in which plots in the forest were cut and burned, crops were produced for 2-3 years until fertility had declined, after which new forest plots were burned and the cycle was repeated. Old plots gradually returned to forest and could be cut again after 20-30 years when the original fertility had been restored. This system worked well when the population density was low and land was amply available, and is still used in the less densely populated areas of Africa, Latin America and Asia.

Crop rotation and fallowing

In areas with a higher population density farmers try to maintain fertility by crop rotations and short-term fallows, sometimes in combination with organic or chemical fertilizers. When population pressure increases fallow periods may decrease to 1-2 years, which is not sufficient for regeneration of the native fertility, leading to a downward spiral of ever decreasing fertility and productivity.

On very acid and poor hillsides in Cauca Department of Colombia, cassava is one of the main crops, usually produced without fertilizer inputs but using bush fallow to maintain fertility. But even after 10-15 years of fallow the native vegetation consists mainly of small trees and underbrush. To study the effectivenes of fallowing on soil fertility maintenance and productivity, simple fertilizer trials were established on seven farms with different periods of fallow before cassava planting. Figure 13 shows that without fertilizer application cassava yields varied from 2 to 10 t/ha, independent of the length of fallow period which ranged from 1-2 to more than 15 years (CIAT, 1988). Yields were maintained at those low levels for the second and third cassava crops. However, with application of P alone, and especially with a combination of N, P and K fertilizers, yields could be doubled or tripled, irrespective of fallowing period. Simple economic analyses indicate that fertilizer application produced larger economic benefits than relying on bush fallow for fertility maintenance. Moreover, with fertilizer application, farmers could continuously cultivate the flatter areas and those closer to their homes, using the steeper slopes for planting fruit trees or coffee, or leaving these under permanent forest. This would also save a considerable amount of labor normally used for fallow clearing and land preparation.

1st. Planting

2nd. Planting

3rd. Planting

Figure 13. Effect of length of fallow period on the yields of three consecutive cassava crops grown with various fertilizer treatments in the area of Mondomo, Cauca, Colombia.
Source: CIAT, 1988.

In areas where soil and climatic conditions permit the rotation of cassava with other crops, this may be another alternative to prevent excessive depletion of nutrients (especially K) resulting from continuous cassava cultivation. In a long-term rotation experiment, conducted in Khon Kaen, Thailand, cassava was rotated in alternate years with peanut followed immediately by pigeon pea, crop residues of which were incorporated into the soil. In the 15^th year of cropping this resulted in a 78% yield increase over continuous cassava monocropping without fertilizers, and a 37% yield increase when fertilizers and soil amendmends (lime, rock phosphate and municipal compost) were applied. However, neither the fertilizers nor the crop rotation could completely restore the level of productivity of the original soil (Tongglum et al., 1998).

Green manures, alley cropping and intercropping

Numerous experiments have been conducted to study the effects of green manuring, alley cropping and intercropping on cassava yields and on soil fertility. Green manures are usually planted in the early wet season and are either mulched or incorporated into the soil before planting cassava. Experiments conducted in CIAT-Quilichao, Colombia, indicate that green manuring cassava with kudzu (Puerarea phaseoloides), zornia (Zornia latifolium), or peanut (Arachis hypogea) had a significant beneficial effect on the subsequent yield of cassava in the absence of fertilizers. On sandy soils in Media Luna on the north coast of Colombia, Canavalia ensiformis or natural weeds were most effective (Howeler et al., 1999a). Similar experiments conducted in Thailand (Tongglum et al., 1992; 1998; Sittibusaya et al., 1995; Howeler, 1995; Howeler et al. 2001) indicate that green manuring with Crotalaria juncea was most effective, followed by Canavalia ensiformis, Mucuna fospeada and pigeon pea. However, in Thailand this use of green manures is not practical, as cassava planted late in the rainy season, following the incorporation of green manures, produced very low yields. The system might be acceptable if farmers would leave cassava in the ground for another wet season, harvesting only after 18 months. This combination of green manuring followed by an 18-month crop of cassava in a 2-year cycle produced high cassava yields, but is unlikely to be popular with farmers as it produces an income only every other year (Tongglum et al., 1998).

At Ibadan, Nigeria, Hahn et al. (1993) showed that rotating cassava in alternate years with either a green manure (Mucuna pruriens) or weed fallow could sustain the yield of an improved cassava line at about 20 t/ha, and that of a local variety at 11 t/ha, for 18 years without any fertilizer inputs.

Few alley cropping experiments have been conducted with cassava. On a very N-deficient soil in Malang, Indonesia, alley cropping with Leucaena leucocephala, Gliricidia sepium and Flemingia macrophylla markedly increased cassava yields and reduced erosion in the fourth year of cropping. Intercropping with peanut had a similar beneficial effect (Wargiono et al., 1995; 1998). A similar trial conducted on a rather fertile Oxisol near Ho Chi Minh city, Vietnam, also showed that alley cropping with Gliricidia significantly increased cassava yields, but only in the seventh year of cropping. Except for Indonesia, where farmers plant hedgerows of Leucaena or Gliricidia for animal feed along plot borders, and in some parts of north Vietnam where farmers have planted Tephrosia candida for erosion control, alley cropping is seldom adopted by farmers. This is because the hedgerows occupy considerable space in the field, they require labor for regular pruning, and the beneficial effect is generally long-term rather than immediate.

Intercropping experiments conducted in Colombia (Leihner, 1983), Nigeria (Okeke, 1984), Thailand (Tongglum et al., 1992; 1998), and Indonesia (Wargiono et al, 1992; 1995; 1998) usually show that intercropping cassava with maize, cowpea or peanut, slightly decreases cassava yields, but increases the gross and net income, as well as the land equivalent ratio (LER), indicating that intercropping makes more efficient use of the land than growing each crop separately. Many trials have been conducted to optimize the system in terms of intercrop selection, varieties, interrow and interplant spacing, relative time of planting, fertilization and weed control (Leihner, 1983; 1999). Intercropping can also provide crop residues, which, when incorporated, add nutrients and organic matter to the soil. The long-term effect of this on soil fertility and productivity has been much speculated about, but has not been well quantified. Tongglum et al. (1993) showed that intercropping with peanut could maintain the OM and P contents of the soil, but that the K content decreased from 0.19 in the first to 0.12 me K/100g in the 18^th year of continuous cropping in Rayong, Thailand. Planting cassava in monoculture, or intercropping with sweet corn, mungbean or soybean were less effective than peanut in maintaining soil fertility, even with annual applications of 46 kg each of N, P₂O₅ and K₂O/ha. Intercropping with sweet corn generally produced the highest gross income (Tongglum et al., 1993).

Mulch

Application of mulch of local weeds, of cut-and-carry grass or of crop residues, such as maize stalks, rice straw, or stems of grain legumes, can also improve soil fertility, reduce the soil surface temperature, conserve soil moisture and control weeds and erosion. In Africa, application of mulch, especially that of leguminous species, increased cassava yields on acid sandy soils (Ofori, 1973; Hulugalle et al., 1991), but had no significant effect on a gravelly Alfisol. In sandy soils on the north coast of Colombia, Cadavid et al. (1998) found that annual mulch applications of 12 t/ha of dry grass of Panicum maximum significantly increased cassava yields during eight years of continuous cropping, especially in the absence of chemical fertilizers. Mulch application also increased the root dry matter and decreased the HCN content. Available soil P and exchangeable soil K gradually increased, while mulch application prevented the decline in soil Ca and Mg. In addition, mulching reduced significantly the surface soil temperature, which enhanced the maintenance of soil C. Thus, where available, application of mulch can help maintain or improve both the physical and chemical conditions of the soil to increase yields.

Fertilization i. Responses to fertilizer applications

While cassava grows better than most crops on very acid and infertile soils, this does not mean that the crop does not need, or does not respond to, fertilizer application. Like most other crops cassava grows and yields better on more fertile soils and responds well to fertilizer applications when grown in poor soils. Table 17 shows the average yield response of cassava in comparison with that of other crops, as determined from thousands of fertilizer trials and demonstrations conducted by FAO throughout the world. Cassava trials were conducted in Ghana and Nigeria in West Africa, and in Brazil and Indonesia in Latin America and Asia, respectively. It is clear that in all three continents cassava responded as much as, or even more than, other crops to fertilizer application.

Table 17. Average response to fertilizer application of cassava and other major crops in fertilizer trials conducted by FAO from 1961 to 1977. Numbers in parentheses indicate the number of countries in which trials were conducted.

Source: Richards, 1979.

In Africa relatively few fertilizers trials have been conducted with cassava, mainly because fertilizers are not readily available or are too costly for most poor cassava farmers. Okogun et al. (1999) reported that in West Africa cassava responded most frequently to N. Responses to P were reported in Ghana (Stephens, 1960; Takyi, 1972) and in Madagascar (Cours et al., 1961), while responses to K were reported mainly for acid sandy soils of southwest Nigeria (Kang and Okeke, 1984; Odurukwe and Oji, 1984) and for strongly acid soils in eastern Nigeria (Okeke, unpublished).

In Latin America, principally Brazil and Colombia, short-term fertilizer trials showed mainly a response to P in acid Oxisols, Ultisols and Inceptisols, which are extremely low in P and have a high P fixing capacity. Available P (Bray II) levels were often below 1-2 ppm. In that case, P was the main limiting nutrient and yields could be doubled or tripled by P application. Responses to K were found at intermediate frequencies, with significant responses to K observed in 9 out of 48 trials in Brazil and 6 out of 22 trials in Colombia. In Latin America, responses to N were the least frequent, with significant responses to N reported in only 5 out of 41 trials conducted in Brazil (Gomes, 1998) and in 5 out of 22 trials conducted in Colombia (Howeler and Cadavid, 1990). However, in sandy soils of the eastern coast of Santa Catarina, Brazil, and the north coast of Colombia, the crop responded markedly to application of N (Moraes et al., 1981; Howeler and Cadavid, 1990).

In nearly 100 short-term NPK trials conducted by FAO on farmers' fields in Thailand, there was mainly a response to N, followed by K and P (Hagens and Sittibussaya, 1990). Similar results were obtained in 69 trials conducted in Indonesia (FAO, 1980). Trials conducted more recently in four other countries in Asia (Table 18) show an initial response mainly to N and some response to K; in most locations the crop did not respond significantly to application of P, as the P content of the soil was usually above 5-6 ppm. After several years of continuous cropping, however, the responses to K increased markedly, those to N to a lesser extent, while those to P increased only slightly over time (Howeler, 1992; 1995; 1998). Similar results were obtained in long-term experiments conducted in Colombia (Howeler and Cadavid, 1990; Howeler, 1991a; CIAT, 1995), in Thailand (Figure 7), in India (Kabeerathumma et al., 1990) and in Malaysia (Chan, 1980), always showing that K became the most limiting nutrient after several years of continuous cassava cultivation.

In most low-fertility soils cassava responds to applications of fertilizers in the ratio of N-P₂O₅-K₂O of 2-1-2, 3-1-2 or 2-1-3, but in some soils very low in P (<2mgP/g) an initial application of 100-200 kg/ha P₂O₅ may be required to overcome acute P-deficiency. When grown continuously for many years on the same land, cassava responds mainly to applications of K.

Table 18. Response of cassava to annual application of N, P and K during the first year and after several years of continuous cropping in long-term fertility trials conducted in various locations in Asia.

¹⁾ NS = no significant response
* = significant response (p< 0.05)
** = highly significant response (p<0.01)

ii. Organic vs inorganic nutrient sources

If chemical fertilizers are not available or too expensive, farmers often apply organic manures, or to a lesser extent, composts to maintain soil fertility, especially if these sources are available on their own farm. Thus, north Vietnamese cassava farmers usually apply 2-7 t/ha of pig manure, while Indian farmers apply up to 20 t/ha of cattle manure (Table 5); in the mountainous areas of southwest Colombia, cassava farmers like to apply chicken manure, which, they claim, gives better results than chemical fertilizers. These manures are excellent sources of both macro- and micro-nutrients and may also improve the physical condition of the soil. However, organic manures and composts have very low nutrient contents as compared to chemical fertilizers, so large amounts are required to have the desired effect; this may result in high costs of transport and application. Most manures are relatively low in K and high in P to be suitable for cassava production. Thus, a combination of organic manures and NK or NPK fertilizers often gives the best results (Kabeerathumma et al., 1990; Nguyen The Dang et al., 1998).

Measures to reduce erosion

Soil erosion control measures can be separated into two broad categories, i.e. engineering structures and vegetative techniques. In many cases both are used at the same time.

Engineering structures

These include land leveling, the construction of contour earth banks or bunds, hillside ditches and various types of terraces. Although these structural solutions were emphasized in the past, their cost effectiveness was found to be rather poor. This is because of their high cost of installation ($400-1000/ha for terraces), as well as high cost of maintenance (Magrath and Doolette, 1990). If terraces or contour banks are not well designed or maintained they can easily collapse, causing severe loss of land. Moreover, drainage ways need to be constructed and maintained, to safely conduct the water down slope. Besides the loss of productive land by terrace risers, there is an additional loss of land of 3-5% for drainage ways. Also, depending on slope and depth of top soil, there may be considerable exposure of infertile subsoil, resulting in low yields or increased fertilizer requirements during the first few years after terrace construction. If terraces are built with heavy machinery, this may lead to soil compaction and extremely high rates of erosion during and shortly after construction.

Vegetative techniques

These include various crop and soil management practices that will provide a vegetative cover of the soil to reduce the impact of raindrops and increase infiltration, or to provide barriers to reduce the speed of runoff. Some examples of these techniques are:

• Fertilization is one of the most effective ways to enhance early canopy closure, which protects the soil against rainfall impact and reduces erosion (Howeler and Cadavid, 1984; Jantawat et al., 1994)

• Closer plant spacing (0.8 x 0.8 m) will also speed up cassava canopy closure and is often very effective in reducing erosion (Jantawat et al., 1994).

• Vertical or inclined planting will result in more rapid germination and canopy closure than horizontal planting (Zhang Weite et al., 1998).

• Varieties with high early vigor may provide a more rapid cover of the soil surface (Howeler et al., 1993).

• Intercropping cassava with maize, grain legumes, melons or pumpkin will help to cover the soil between cassava plants during the first 1-2 months after cropping (Aina et al., 1979) and thus reduce erosion (Tongglum et al., 1992; Jantawat et al., 1994).

• Contour cultivation is one of the most effective ways to reduce runoff and erosion, capture soil moisture and increase yields. On moderate slopes (up to 15%) this can be done by tractor, although it may take more time than up-and-down tillage. On steeper slopes (up to 50%) land can be prepared with animal-drawn equipment. A reversible plow, utilized in the Andean zone of Colombia was found to be very effective in contour plowing of steep slopes (Howeler et al., 1993).

• Minimum tillage and/or stubble mulching can be very effective in reducing runoff and erosion. In loose and friable soil, seeds can be planted directly using a pointed stick to make holes while cassava can be planted by pushing the stakes directly into the soil. In compacted soil or in weedy plots it may be necessary to prepare individual planting spots with a hoe. Another form of minimum tillage is to either reduce the intensity of tillage (one plowing instead of various passes with plow or harrow) or in the area to be tilled, alternating contour strips of tilled and untilled soil. While minimum tillage can decrease erosion significantly, it often leads to a reduction in yield due to soil compaction, weed competition and reduced efficiency of fertilizers when these are left on the soil surface (Ruppenthal, 1995). When soils are compacted or the surface is sealed by heavy rains, runoff and erosion may actually increase and water infiltration decrease (Jantawat et al., 1994).

• Contour ridging was found to be very effective in reducing runoff and erosion on gentle slopes and in stable soils; it often also increases yields by concentrating topsoil in the ridge, increasing rooting depth and conserving soil moisture. However, on steep slopes or with unstable soils, too much water accumulating behind the ridges may cause them to break, resulting in concentrated water flow and gulley erosion. In poorly drained soil, contour ridges may keep the soil too wet, resulting in poor growth and/or root rot.

• Mulching with crop residues or grass on the soil surface greatly improves water infiltration, protects the soil from direct raindrop impact and reduces runoff and erosion. Annual application of grass mulch has been shown to more than double the yields of cassava in the absence of fertilizers (Cadavid et al., 1998) by supplying nutrients and reducing surface soil temperatures. However, sufficient mulching materials are often not available or their collection and transport is costly. For that reason, in-situ production of mulch by rotating or intercropping cassava with leguminous cover crops may be another solution (Tongglum et al., 1998).

• Cover crops or "live mulches" of Calopagonium, Pueraria phaseoloides or Macroptillium atropurpureum have been used successfully for erosion control under perennial trees like rubber or oilpalm. Attempts to use perennial legumes as cover crops in cassava have been less successful due to severe competition of the cover crops with cassava (Muhr et al., 1995), especially once the cover crops are well established (Howeler et al., 1999a, 1999b). Cassava yields were reduced on average 20-50% by nine cover crop species in Thailand (Howeler, 1992) and by 40% or more in Colombia (Leihner et al., 1996).

• Vegetative barriers may include:

  a. Contour strips of cut-and-carry grasses such as elephant or napier grass (Pennisetum purpureum), king grass (Saccarum sinense), Bermuda or Bahama grass (Cynadon dactylon), Imperial grass (Axonopus scoparius), Paspalum atratum and Setaria sphacelata. These have been used successfully to reduce runoff and erosion and to supply cut-and-carry feed for cattle or water buffaloes (CIAT, 1995; Howeler et al., 1999b). Contour grass strips of about 50-100 cm width are usually planted at 1-2 m vertical intervals. The drawback of this system is that 15-25% of the land must be taken out of crop production, the grass trimming is labor-intensive, feed production is often more than the family can use, and the grass stolons or feeder roots can seriously reduce yields of adjacent rows of food crops (Leihner et al., 1996; Tscherning et al., 1995).

  b. Contour hedges of vetiver grass (Vetiveria zizaniodes) are very effective in reducing runoff and erosion and may increase crop yields by improved water conservation and reduced nutrient loss (Howeler et al., 1998). Single-row hedges of about 30-50 cm width are generally sufficient, taking less than 10% of land out of production on gentle slopes. Moreover, the deep vertical root system of this grass does not compete much with adjacent crops (Tscherning et al., 1995). However, the low forage quality of the grass is a serious drawback for those farmers that need to produce animal feed. Lemon grass (Cymbopogon citratus) is intermediately effective in controlling erosion, but may be favored by farmers if it has commercial value.

  c. Contour strips of native grasses or weeds. In this system contour-plowed strips of 4-10 m width are alternated with 50-100 cm wide unplowed strips. These unplowed strips often have native grasses such as cogon grass (Imperata cylindrica) and Paspalum conjugatum, which can be cut regularly for animal feed (Fujisaka 1993). Depending on the dominant grass species and its management, the competition with the adjacent crop can be light or quite substantial. Competition from native grass strips of Paspalum notatum caused a reduction in cassava yield of about 13% in Quilichao, but strips of less aggressive native grasses had no significant effect on cassava yields in Mondomo, Colombia (Reining, 1992). In this system about 15-25% of land area is taken out of crop production, which may be a serious drawback in areas where land is scarce. By making the unplowed strips a little wider, these "macro-contour lines" can be used to plant fruit trees or coffee and provide enough fodder for ruminants, which in turn produce manure to improve soil fertility in the cropped areas (Basri et al., 1990; Garrity et al., 1998).

  d. Hedgerows of leguminous trees. This system is also called "alley cropping" and consists of planting fast-growing leguminous tree species such as Leucaenae leucocephlala, Gliricidia sepium or Tephrosia candida in single or double contour rows about 5-10 m apart. Crops are grown in the space between the hedgerows. To prevent light competition the trees need to be pruned regularly to about 30-50 cm height and the prunings can be used as either animal feed (Leucaena and Gliricidia) or placed between the hedgerows as a mulch and source of nutrients (Tephrosia). While rather labor intensive and slow to establish, this system can eventually be effective in forming terraces, reducing erosion and increasing yields (Wargiono et al., 1995). Initially, however, yields may be reduced due to substantial competition for water and nutrients from the hedgerows.

The advantages of these various vegetative techniques are:

• low cost of installation: barriers of vetiver grass cost only $16 per ha compared with $21-80/ha for construction of earthen bunds in India (Margrath, 1990);

• adaptability: allows for flexible management and does not require much expertise;

• greater farmer control;

• less area out of production: about 15-25% for hedgerows in alley cropping systems, but less than 10% for vetiver hedgerows;

• no need for water disposal systems; better water retention;

• natural terrace formation by such practices as alley cropping and contour grass barriers;

• may provide animal fodder by hedgerow trees or grass barriers, or additional income from perennials grown in contour strips; and

• usable for a wide range of land tenure situations.

Many of these vegetative techniques can be applied solely or in combination and in many cases they act synergistically to increase productivity as well as reduce erosion. However, each technique has its own benefits and its own limitations, which may require certain trade-offs.

To be effective and acceptible to farmers these techniques must:

· produce direct and tangible benefits to farmers in the form of increased productivity or income;

· require few outside inputs and have low labor requirements for installation and maintenance;

· be simple and not require expensive machinery nor expert advice;

· be adapted to the local conditions of soil and climate as well as the availability of necessary inputs or markets for outputs; and

· be effective in soil and water conservation.

Enhancing the adoption of soil-conserving practices by farmers

Upland ecosystems in which cassava is grown are highly diversified and each region has a unique set of physical and socio-economic conditions as well as a unique complex of interrelated problems, such as poverty, lack of secure tenure, land degradation, sedimentation, irregular stream flow etc. Therefore, solutions to these problems are obviously site-specific and must take into account both the physical conditions of the site and the resources and needs of the local population. Solutions may have to be highly diversified and include agriculture, forestry, animal husbandry and vegetable or fruit production. Attempts in the past at implementing soil conservation measures have often failed because the proposed solutions were inappropriate for the site, they were too narrowly focused on a particular crop or soil conservation practice and/or they did not take into account the indigenous knowledge nor the needs of the population.

Because of the diversity of upland ecosystems, the technology development must start with adequate problem identification through Rapid Rural Appraisal (RRA) techniques, farmer participatory planning, the description of land classes and uses, the determination of causes of land degradation and a definition of local needs, objectives and possible solutions. The principle objective is to improve upland agricultural productivity and income, while a secondary objective is environmental improvement. These two objectives can best be achieved through simple low-cost innovations in crop/soil management so as to conserve moisture and maintain soil fertility. A reduction in soil erosion will enhance moisture conservation and increase moisture- and fertilizer-use efficiency, thus increasing yields and farm income. To be acceptable the proposed solutions must address immediate and short-term needs and must be based on existing practices rather than on the introduction of a "package" of new practices.

Because problems and solutions are site-specific, the technology development can best be done by multidisciplinary teams through networking and improvement of research capabilities in local and national institutes. The right choice of options could involve some trade-offs and thus requires a thorough knowledge of local conditions and needs; this is best left to local researchers. Once promising technology components have been identified, linkages must be established with development-oriented and extension institutes for the technology adaptation and dissemination. This may best be done through farmer surveys, on-farm trials and demonstrations and Farmer Participatory Research (FPR) so as to obtain feedback about the farmers' priorities and the effectiveness and acceptability of various options. From a menu of many options farmers will be able to select the most appropriate ones and then experiment with these on their own farm, with the help of researchers and extensionists trained in farmer participatory research approaches. After determining which practices are most beneficial, these can be further adapted and then adopted by the farmers. Finally, they are disseminated to others through regular extension methodologies and farmer participatory extension. Spontaneous adoption of soil conservation practices can only be expected if farmers themselves and the whole community are involved in their development, the practices are of low risk and satisfy immediate needs for food and income. In other cases, some incentives may be necessary, including public investments in soil erosion control measures, especially if these have to be implemented at the community or watershed level.

Table 19 summarizes the effect of various soil/crop management practices in cassava-based cropping systems on erosion control, terrace formation and cassava yield, and indicates their relative costs and labor requirements as well as their long-term benefits and limitations. Obviously, the benefits and limitations depend to a certain extent on local conditions of climate, slope, soil, prices of inputs and products, farm size and resources availability. For that reason it is unlikely that any one practice would be universally effective and attractive to farmers. In order to develop effective soil and crop management practices for a particular region and to achieve their adoption, a strategy is required that links the research and extension institutions with farmers' organizations, and considers both socio-economic as well as technical factors. Figure 14 shows the activities and possible outputs of such a strategy.

Besides basic and applied research for the development of a menu of effective technology components, it is necessary to test these components in a particular region by conducting simple soil management and erosion control trials (Howeler, 1987; 1996a), using treatments selected in consultation with farmers. These trials can serve as demonstration plots so farmers can clearly observe the effect of certain practices on erosion and yield. Once the effectiveness of various practices has been determined, then farmers can be asked to select those practices that best satisfy their needs. These practices can be further tested on farms with direct farmer and community participation (Howeler, 1999). By actually collecting and measuring the soil loss due to erosion, the problem becomes visible and farmers will become more aware of the seriousness of soil erosion, and more willing to adopt better management practices that not only increase yields but also reduce soil losses. At the same time, researchers and extension officers should be trained, not only in conducting better soil conservation trials, but also in farming systems research and farmer participatory research methodologies, so as to increase their knowledge of, and sensitivity to, farmer's priorities and needs. Only when those needs are taken into account in the development and selection of improved soil and water conservation practices, might we reasonably expect some adoption of these practices, which in turn, will have a positive impact on the environment. In addition, cassava is not necessarily the only or the best option available to farmers to increase income and preserve their soil resources. Ecosystem research teams, comprising both agricultural and social scientists, should study a wide range of possible technical as well as policy options to determine the best land use that will improve the income and well-being of farmers, while at the same time prevent any further degradation of soil and water resources. To be successful, this will require increased collaboration between various national and international research and development-oriented institutes, so as to tap the expertise available in each institute and make maximum use of scarce financial resources.

Figure 14, Strategy to improve the sustainability of cassava-based cropping systems.
Source: Howeler, unpublished

Table 19. Effect of various soil/crop management practices on erosion and yield, as well as on labor and monetary requirements and long-term benefits in cassava-based cropping systems.

+ = effective, positive or high
- = not effective, negative or low
¹⁾ = value added in terms of animal feed, staking material or fuel wood
Source: Howeler, 1994.

Gaps in knowledge

Intensive research on cassava nutrition and fertilization, on crop/soil management practices to maintain soil fertility and reduce erosion, has been conducted over the past 25 years in both Latin America and Asia, and to a lesser extent in Africa. Much is known both about the principal causes of soil erosion and fertility decline and about specific measures to be taken to mitigate against these problems. Gaps in our knowledge may still exist on:

1. Effect of continuous production of cassava in comparison with that of other annual food crops on crop yields, and on the chemical, physical and biological conditions of the soil.

2. Long-term effect of crop rotations and intercropping on crop yields and the chemical, physical and biological conditions of the soil.

3. Nutrient balances at several levels of scale

4. Nutrient losses in runoff and eroded sediments.

5. Effect of various tillage practices on cassava yield, production costs and erosion.

6. Diagnostic criteria and effective methods to control micro-nutrient deficiencies, especially in calcareous or organic soils.

This knowledge is important for improving our understanding of erosion processes and may help to predict the effect of changing land-use practices on soil and water resources; it may also help to develop more efficient management practices to increase farm income and protect the soil from degradation. However, the greatest gap in our knowledge is not in the technical aspects but in the social and economic aspects.

By far the greatest challenge is to develop soil conservation practices that are not only technically effective but also economically efficient and acceptable to farmers, and to develop methodologies, strategies and institutional arrangements that will effectively disseminate these practices and facilitate their adoption.

Conclusions and recommendations

1. Nutrient removal by cassava depends mainly on the yield level obtained and on the plant parts removed from the field. When only roots are removed and yield levels are relatively low, cassava removes less N, P and K than most other crops; when yield levels are relatively high, cassava removes less N and P and similar amounts of K as other crops. However, when leaves and stems are also removed from the field, nutrient removal (especially of N and Ca) increases substantially, and in that case cassava may remove more N and K than most other crops.

2. When cassava is grown continuously on the same soil, without nutrient replenishment, the crop tends to deplete soil-K first, followed by N, Mg and Ca (the latter two nutrients mainly if plant tops are removed from the field).

3. Cassava is highly dependent on mycorrhiza for P uptake, but in practically all soils cassava roots become infected with native soil mycorrhiza resulting in a highly effective symbiosis. This enables the crop to take up P from soils with low levels of available P. Thus, the critical soil-P level for cassava is much lower than for most other crops.

4. Cassava is extremely tolerant of low pH and high levels of Al. For those reasons the crop is well-adapted to acid soils and seldom requires lime applications.

5. When grown on infertile soils, cassava responds as well as other crops to fertilizer application. In some soils with extremely low levels of available P (<2 µg P/g), the crop responds initially mainly to P applications. In most other soils, however, it responds mainly to applications of K and N. Compound fertilizers with N, P₂O₅ and K₂O ratios of 2-1-2, 3-1-2 or 2-1-3 will generally give the best results. Of the micronutrients Zn is the most important.

6. Cassava is extremely tolerant of drought and can be grown in areas with less than 500 mm annual rainfall. Once established, the crop will survive long (6-8 months) periods of drought.

7. Cassava adjusts its rate of growth to the moisture and nutrient conditions of the soil. When these conditions are unfavorable the growth rate slows, and yields are low. But plants generally survive, and when conditions improve the growth rate increases, resulting in higher yields. In the case of drought, the crop consumes little water; in the case of poor soils, the crop absorbs few nutrients and is able to sustain low but stable yields without serious soil nutrient depletion.

8. When grown on slopes, cassava causes more erosion than most other crops, due to slow initial growth and wide plant spacing. Erosion is particularly serious on sandy, low-OM soils on which cassava is often grown. These soils have poor aggregate stability and become particularly susceptible to erosion if water infiltration is limited by a hardpan in the subsoil.

9. When cassava is grown on slopes, farmers must take measures to reduce erosion. These can include common cultural practices, such as fertilization, reduced tillage, contour ridging, closer plant spacing, intercropping and chemical weed control; or it may include special measures, like planting contour barriers or making trash lines, stone walls or terraces. On gentle but long slopes, the speed of runoff should be reduced by contour grass strips or hedgerows, or by contour land preparation and ridging.