On-site effects of cassava cultivation on soil fertility

Many people are convinced that cassava production leads to soil degradation, and some governments do not encourage cassava cultivation in the belief that it causes serious erosion and nutrient depletion. In Vietnam, Thai Phien and Nguyen Tu Siem (1996) stated that, "as a direct consequence of planting upland rice and cassava for food self-sufficiency, more than one million ha have become eroded skeleton soils with no value for agriculture or for forestry". Figure 2 shows that after four years of continuous cassava cultivation without fertilizer inputs in Vietnam, yields declined from 19 to about 7 t/ha. However, in the same experiment, yields of upland rice declined from 2.5 t/ha to zero in the same time span. No soil data are available to indicate which crop caused greater nutrient depletion and/or soil degradation, but after four years cassava still produced a reasonable yield while rice produced none. There is little information in the literature on the relative rates of soil degradation and yield decline over time for different crops.

In south Vietnam, Cong Doan Sat and Deturck (1998) determined the physical and chemical characteristics of similar soils (Haplic Acrisols) that had been under continuous forest, rubber, cashew, sugarcane or cassava for many years. Table 6 and Figure 3 summarize their results. The data indicate that long-term cassava cultivation resulted in the lowest levels of organic C and total N (mainly due to frequent and intensive land preparation and weeding); the lowest CEC (associated with a decline in soil organic matter and preferential loss of clay due to erosion), and the lowest levels of K and Mg. P-levels, however, were higher than in cashew or forest, due to cassava's low off-take of P, and some application of P fertilizers to cassava, rubber and sugarcane. Figure 3 shows that continuous cassava cultivation resulted in a high bulk density, low water infiltration rate, low aggregate stability, low clay content and low water holding capacity (not shown). The conclusion is clear: continuous cassava cultivation under these conditions has a definite detrimental effect on both the physical and chemical properties of the soil^[14], and is thus unsustainable. Unfortunately, there are no comparisons with other annual food crops, which, like cassava, require intensive and frequent land preparation and weeding, and which leave the soil exposed to the impact of rainfal, leading to erosion during part of the year.

Figure 2. Yield reduction of upland rice and cassava due to fertility decline as a result of continuous cropping without fertilizer application. 100% corresponds to 18.9 t/ha of fresh cassava roots and 2.55 t/ha of rice.
Source: adapted from Nguyen Tu Siem, 1992.

There is no doubt that land clearing and conversion of forest to annual crops often lead to a rapid decline in OM content, in aggregate stability and nutrient supply, as well as a sharp increase in soil losses due to erosion. This decline is mainly due to exposure of the soil surface to high temperatures, leading to more rapid OM decomposition, as well as to the direct impact of rainfall, leading to the breakdown of soil aggregates, surface crusting, and erosion, with preferential losses of OM, clay and soil nutrients in the sediments and runoff. In addition, soil preparation by heavy machinery will increase bulk density and create hard pans, which further aggravates soil degradation. The question, however, remains whether cassava is worse in that respect than other annual crops, like cotton, soybean, maize, upland rice, peanut and mungbean.

Figure 4 is another example of the long-term effect of cassava cultivation on soil productivity in Thailand. In fertilizer trials conducted in different farmers' fields on three soil series since 1955, the yields of cassava decreased from initial levels of 26-28 t/ha to levels of 10-13 t/ha after 20-30 years of cropping without fertilizer inputs. This gradual yield decline was attributed both to nutrient depletion due to removal of nutrients in the harvested roots, as well as erosion. After the conversion of forest to cropland, the cultivation of cassava (or other crops) without nutrient inputs will inevitably result in a decline in soil productivity, even if crops are rotated or fields are fallowed occasionally. Again, the basic question is whether this yield decline observed in cassava is more or less rapid than that observed in other crops grown under similar conditions.

Figure 3. Bulk density (top), infiltration rate (middle) and aggregate stability (bottom) of Haplic Acrisols under different cropping systems in south Vietnam. Values are the average of at least 7 profiles per cropping system and three horizons per profile.
Note: Aggregate class stability: 1 = highly unstable, 2 = unstable, 3 = relatively stable, 4 = stable
Source: Cong Doan Sat and Pol Deturck (1998).

Table 5. Characteristics of cassava cropping systems and cultural practices used in major production zones in Asia.

1) Including family labor, harvest + transport
Source: Hershey et al., 1999.

Figure 4. Decline in fresh root yields due to continuous cultivation without fertilizers in three soil Series in Thailand.
Source: Sittibusaya, 1993.

Table 6. Chemical properties of various horizons of Haplic Acrisols that have been under different land use in southeastern Vietnam.¹⁾

¹⁾ Values are average of 6-10 profiles per cropping system. Within rows data followed by the same letter are not significantly different at 5% level by Tukey's Studentized Range Test.

Source: Cong Doan Sat and Deturck, 1998.

Nutrient removal in the crop harvest

Cassava grows relatively well on poor soils, which may result in a further reduction in soil fertility (Table 6). For that reason the crop has a reputation of removing large amounts of nutrients from the soil, leaving the soil depleted of nutrients and too infertile for further crop production. This may or may not be the case.

Table 7 shows the dry matter (DM) distribution and nutrient content in roots, stems and leaves for cassava harvested in Curvelo, Minas Gerais, Brazil, both in fertilized and non-fertilized plants (Paulo et al., 1983). In both cases, DM in the roots was about 50-55% of that in the total plant, but fertilized plants produced more than twice the root yield of unfertilized plants. Since fertilization not only increased the DM production but also increases the nutrient concentration in the tissue (Table 8), the nutrient content of the total plant and that of the roots were 2-3 times higher in the fertilized compared to the non-fertilized plants. If only roots were harvested and stems and leaves returned to the soil, the removal of nutrients from the field in this experiment was only 51 kg/ha of N, 2.1 kg of P, 20 kg of K, 6 kg of Ca and 3 kg of Mg. In fertilized plants it was 130 kg/ha of N, 9.3 kg of P, 80 kg of K, 15 kg of Ca and 12 kg of Mg. The removal of micronutrients was insignificant. Thus, in poor soils, the total amount of nutrients absorbed and that removed with the root harvest are relatively low when plants are not fertilized, but increase markedly with fertilization. The same has been reported by Howeler and Cadavid (1983) and by Howeler (1991a; 2001).

Table 7. Average dry matter and nutrient content in roots, stems and leaves of two cassava varieties (Riqueza and Branca de Santa Catarina) planted with or without fertilizers in Curvelo, Minas Gerais, Brazil in 1982.

Source: Paulo M.B. de et al.,1983.

In some countries, notably in northern Vietnam, southern China, Indonesia and India, farmers not only harvest the roots, but also the green leaves to feed cattle or fish, and the stems as fuel wood. In that case, nutrient removal would be 2-3 times higher, as about 60-65% of N, 50-60% of P, 40-50% of K, 85-90% of Ca and 70-80% of Mg are found in stems and leaves (Tables 7 and 9). Sometimes fallen leaves are collected as kindling. Nutrient removal, especially of N and Ca, would be even higher (Howeler, 1985). Thus, it is clear that nutrient removal by cassava depends on the fertility of the soil, the yield levels obtained, and whether only roots or other plant parts are removed from the field. Using data from many sources in the literature, Howeler (2001) calculated an "average" removal per tonne of fresh roots of 2.53 kg/ha of N, 0.37 kg of P, 2.75 kg of K, 0.44 kg of Ca and 0.26 kg of Mg if only roots are removed, but 6.68 kg/ha of N, 0.76 kg of P, 4.87 kg of K, 2.78 kg of Ca and 0.87 kg of Mg if the whole plants are removed (Table 9). If nutrient removal were proportional to yield, an average yield of 15 t/ha of fresh roots would remove "on average" 37 kg N, 6 kg P, 41 kg K, 6.6 kg Ca and 3.0 kg of Mg. Comparing nutrient removal data reported in the literature by various authors, Howeler (2001) reported that N removal was quite variable but more or less proportional to dry root yield, but that P and K removal increases more than proportionally with an increase in yield. According to the data in Figure 5, a fresh root yield of 15 t/ha would result in the removal of about 40 kg N, 3.5 kg P and 20 kg K/ha, considerably lower for P and K than previously estimated (Howeler, 1981). These relatively low levels of nutrient removal may explain why cassava yields of less than 10 t/ha do not seem to deplete the nutrients in the soil and that those low yields can be sustained for many years without application of fertilizers or manures as long as plant tops are re-incorporated into the soil (see Figure 8B).

Table 8. Nutrient concentrations in plant parts of fertilized and unfertilized cassava, cv. MVen 77, at 3-4 months after planting in Carimagua, Colombia.

Source: Howeler, 1985.

Table 9. Average nutrient removal (kg) per ton of harvested fresh cassava roots when only the roots or the whole plants are removed at harvest. Calculations are based on data from 14 experiments with root yields ranging from 6 to 65 t/ha. Numbers in parentheses indicate the proportion of each nutrient present in the roots.

¹⁾ Does not include fallen leaves
Source: Howeler, 2001.

It may be concluded that in a cassava root harvest considerable amounts of N and K are removed from the field, while the removal of P, Ca, and Mg is relatively low. The harvest of leaves and stems increases markedly the removal of mainly N and Ca.

Table 10 shows the nutrient removal in the harvested product of cassava and various other crops, calculated both in terms of nutrient removal per ha and per tonne of DM produced (Howeler, 1991a). In spite of a very high average cassava yield of 35.7 t/ha used in these calculations, the removal of N and P by cassava was similar to that of many other crops, while the removal of K was higher than in most other crops. Similar results have been reported by Prevot and Ollagnier (1958) and Amarasiri and Perera (1975). When calculated per tonne of DM produced, cassava removed much less N and P, and less or similar amounts of K compared with most other crops.

Cassava may be highly efficient in absorbing nutrients from poor soils, but the amounts of nutrients removed in the root harvest are rather low compared to other crops, with the possible exception of K. However, if stems and leaves are also removed from the field, then nutrient removal can be quite high and nutrient depletion of the soil can become a serious problem.

Figure 5. Relation between the N, P and K contents of cassava roots and dry root yield, as reported in the literature. Arrows indicate the approximate nutrient contents corresponding to afresh root yield of 15 t/ha.
Source: Howeler, unpublished. (see Appendix 2)

Figure 6. Effect of various levels of annual applications of N, P and K on cassava fresh root yield (A), and on the exchangeable K content of the soil (B) during eight consecutive cropping cycles in a long-term NPK trial conducted in CIAT-Quilicao, Colombia.
Source: Howeler and Cadavid, 1990.

Table 10. Average nutrient removal by cassava and various other crops, expressed in both kg/ha and kg/t harvested product, as reported in the literature.

¹⁾ Assuming cassava to have 38% DM, grain 86%, sweet potato 20%, sugarcane 26%, dry tobacco leaves 84%.
²⁾ Phaseolus vulgaris
Source: Howeler, 1991a.

Yield trends with and without fertilizer application

Figures 2 and 4 show the effect of continuous cassava production on yield in the absence of fertilizers. Figure 6A shows the long-term effect when various levels of nutrients were applied annually during eight years of cropping on an Andept of volcanic origin in CIAT-Quilichao, Colombia. Without fertilizer application, yields declined gradually from about 25 to 14 t/ha. With applications of only N or P a similar yield decline was observed. But when K or NPK were applied at sufficiently high rates (100 kg N, 200 P₂O₅ and 150 K₂O/ha) very high yields of 30-40 t/ha could be sustained, while the original level of exchangeable K could be maintained (Figure 6B). Higher rates of fertilization had no beneficial effect on cassava yields, but increased the P and K levels in the soil (Howeler and Cadavid, 1990).

Figure 7 shows similar results in a long-term fertility trial conducted on a sandy loam Ultisol in Khon Kaen, Thailand. In the first year of cropping yields were high (around 30 t/ha), and there was no response to fertilization. In the second year yields dropped precipitously to 10-15 t/ha in those treatments without K. With K, yields dropped to about 20 t/ha. In subsequent years, when no K was applied, the yields further declined to levels of 5-6 t/ha. With K application, even without any P, yields could be maintained between 20 and 25 t/ha for 19 years. Like in the case of Quilichao-Colombia (Figure 6), K became the most crucial nutrient for maintaining long-term productivity of cassava soils. The importance of K application to cassava for the maintenance of long-term soil productivity has also been observed in both northern and southern Vietnam (Nguyen Huu Hy et al., 1998), in three locations of China (Zhang Weite et al., 1998), as well as in southern Sumatra of Indonesia (Wargiono et al., 1998). The importance of K application for cassava and several other crops, like sweet-potato, banana, sugarcane, pineapple and oil palm, is due to the relatively high off-take of K in the harvested products of these crops.

Figure 7. Effects of annual application of various combinations of NPK (top) and crop residue management (bottom) on cassava yield during 23 consecutive crops grown in Khon Kaen, Thailand, from 1977 to 1999.
Source: C Nakviroj and K. Paisancharoen, DOA, Bangkok, (personal communication)

Figure 7B, for the same long-term fertilizer trial in Khon Kaen, also shows that if plant tops (leaves and stems) were reincorporated into the soil after each harvest, nutrient depletion was much less severe, and reasonable yields of 10-15 t/ha could be maintained over 19 years of continuous cropping without any outside nutrient input. This clearly indicates the importance of incorporating stems and leaves into the soil to prevent serious nutrient depletion.

Cassava is drought tolerant because it has the ability to regulate its water consumption by closing its stomata during periods of drought, so as to prevent excessive water use (El-Sharkawy, 1993); this allows slow but continuous growth without death of the plant until growth accelerates again when soil moisture conditions improve. Similarly, cassava adjusts its rate of growth to the nutrient supply of the soil, maintaining a level of productivity that can be sustained by the nutrient-supplying power of the soil. When fertility conditions improve, cassava responds quickly with increased growth and yield.

Nutrient balances

When we know the amount of nutrients that enter and leave the system, we can calculate the balance for each nutrient and determine whether this balance is positive (leading to accumulation), or negative (leading to depletion).

Table 11 shows a nutrient balance for cassava production in various regions of Vietnam, using officially published yield data for each region for 1991/92. The removal of N, P and K, when both roots and plant tops are removed from the field (as commonly practiced in Vietnam), was calculated from these yield data and "average" nutrient removal data shown in Table 9. Nutrients applied were calculated from the average fertilizer and manure input data obtained from over 1000 questionnaires in a cassava production survey conducted in 20 provinces in 1990/91 (Pham Van Bien et al., 1996). Finally, the nutrient balances were calculated by subtracting the outputs from the inputs. The results show that there was a positive P balance in four of the six regions, a positive N balance in three regions, and a positive K balance in only two regions. In four regions, the outflows of K were greater than the input, even when other losses of K, such as by leaching, erosion and runoff, were not considered. Thus, in most cassava growing areas of Vietnam, farmers are applying too much P but not enough K for the needs of the crop, resulting in a downward trend in soil fertility (see Table 6) and a decline in soil productivity. This is at least a partial explanation for the low cassava yields in Vietnam and the general perception that cassava is a soil degrading crop.