Economic and environmental impact of improved nitrogen management in Asian rice-farming systems - R.N. Roy ^a and R.V. Misra ^b

^a Senior Officer and ^b Consultant, Land and Plant Nutrition Management Service, FAO, Rome, Italy

INTRODUCTION

More than 40 percent of the world’s population depends on rice for its major source of food. In order to feed a growing population, annual world rice production needs to increase from 586 million tonnes (Mt) (2001) to 756 Mt by 2030 (FAO, 2000a). The need for such an increase is reflected in efforts to: increase the acreage under high-yielding varieties (HYV); develop hybrids and super rice; evolve more appropriate and efficient crop, soil, water and nutrient management technologies; and accelerate technology transfer.

Rice is grown in various climates ranging from tropical and subtropical to temperate, with the tropics the major area of cultivation. Within their seven broadly-defined agro-ecozones, rice systems are further characterized and quantified in terms of their rice-phase water regimes: irrigated (56%); rainfed lowland (26%); flood-prone/swampland (9%); and rainfed upland (9%). The irrigated sector generates more than 75 percent of total rice production. The estimated global rice acreage is about 152 million ha (Mha) (Table 1) - an increase of 3.5 percent since 1990.

TABLE 1
Rice area, production and yield

Rice is Asia’s main crop, and the region accounts for 90 percent of the world’s acreage and produces 91 percent of the world’s rice production. About 250 million households in Asia (70% of the world’s farming households) cultivate rice and support 2 600 million rice consumers.

Within Asia, rice occupies about 25 percent of all land under arable and permanent crops. Per caput rice area in Asia fell from 0.12 ha in 1961 to 0.07 ha in 1998. The average holding size also declined during this period, partly due to land reform/redistribution programmes, and partly due to inheritance-related over-fragmentation of smallholdings. National average yields range from 2.6 t/ha in Thailand to 6.4 t/ha in China (2001).

RICE PRODUCTION PROJECTIONS TO 2015 AND 2030

FAO estimates of crop yields and area for 2015 and 2030 rely on country forecasts of demand and supply variables. The most important demand determinants are population, income, commodity prices and dietary tastes and preferences. Food supply is assumed to be mainly a function of cropped area and yield, but crop and input prices and other agricultural policies can play a determining role in planted acreage and output yield.

In Asia, the potential for expanding crop area is limited in most countries (Table 2). Moreover, there has been a decline in the rate of increase in irrigated area. The overall area is expected to increase from 135 Mha in 1995-97 (average taken as base year) to 147 Mha in 2030. Although a reduction in rice acreage is projected for Japan, the Republic of Korea, Malaysia, Thailand and the Philippines, overall rice production needs to increase from 516 Mt in 1995-97 to 645 Mt in 2015 and 694 Mt in 2030. Production is expected to rise in all countries (with the exception of Japan, the Republic of Korea and Malaysia). The projected slowdown in overall production growth for 2015-2030 reflects the anticipated deceleration in population growth and the change towards a dietary preference for meat.

TABLE 2
Rice area, yield and production projections for 2015 and 2030

As in Japan, the Republic of Korea, Malaysia, Thailand and the Philippines, other countries are likely to transfer more land currently under rice (especially marginal land and areas vulnerable to degradation) to other uses. Yield increases in favourable areas will be the main element for output growth. Hence, current yield potentials and yield gaps warrant careful analysis.

Gaps between potential and current yield levels are substantial (Table 3). The irrigated ecology offers the greatest scope for narrowing the substantial gaps between potential and current yield levels; indeed, the projected demand and production potential for 2030 may be within reach (Figure 1).

TABLE 3
Rice production potential and yield gaps in selected countries

Source: Pingali et al., 1997 (adapted).

FIGURE 1
Rice production potential in selected countries

The removal of constraints on realizing yield potentials would enable India, Thailand, Bangladesh and the Philippines to achieve substantial surpluses (with China breaking even). The realization of yield potentials in these countries could lead to the elimination of marginal lands and enable crop diversification and other suitable land uses.

Lin and Shen (1996) analysed yield gaps under favourable conditions in China. Technical constraints accounted for between 41 and 45 percent of the yield differences, with deficiencies in N, P, K, soil organic matter and micronutrients the most important. This situation also reflected the long history of intensive cropping, soil fertility depletion and an inability to recover through natural processes.

In India, the major issues and challenges for bridging yield gaps (Siddiq, 2000) pertain to:

• yield plateauing in high productivity areas;

• declining productivity due to inappropriate soil management practices in major rice-based cropping systems under irrigated ecologies;

• continued imbalanced fertilizer use;

• widespread deficiencies in micro- and secondary nutrients, such as zinc and sulphur;

• low input management vis-à-vis risk of crop losses; and

• lack of varietal solutions to problematic soils.

Thus, improved nutrient management is essential for the further exploitation of the genetic yield potentials of HYVs.

NUTRIENT REQUIREMENTS FOR PROJECTED PRODUCTION TARGETS

FAO uses relatively straightforward techniques to obtain mineral fertilizer use estimates (FAO, 2000b). The increased nutrient-use efficiency scenario is probably more appropriate for rice because of the immense scope for improving N-use efficiency. Ideally, there should be a crop-based model for each country, as climate, resource, production and other factors vary considerably. However, such an endeavour would be fraught with data limitations.

In the absence of known crop production functions, estimates of future fertilizer application rates and nutrient-use efficiencies were obtained by quantifying the relationship between production and total fertilizer application rates that existed in the period 1995-97 for all producing countries. This global crop-specific fertilizer response coefficient was used to compute the future fertilizer application rates required to attain the projected yield increases.

Nitrogen use projections for 2015 and 2030

Projections for N fertilizer requirements are given in Table 4. The projections reveal that N use per hectare would increase in all Asian countries over the years. Total N consumption is also projected to increase in all countries, except in China, Japan, the Republic of Korea and Malaysia, where a significant decrease in total consumption is expected, mainly due to the reduction in area. The aggregated projected nitrogen use figures for Asia for 2015 and 2030 are 16.0 and 16.6 Mt, respectively.

TABLE 4
Projections for nitrogen fertilizer use

Nitrogen-use efficiency

The calculation of N-use efficiency assumes that 20 kg of N is needed for the production of 1 tonne of rice and that applied fertilizer provides 50 percent of this requirement. Average N-use efficiency in Asia is projected to rise from 35 percent in 1995-97 to 40.3 percent by 2015 and 41.8 percent by 2030 (Table 5). Although N-use efficiency is less than 35 percent in almost all Asian countries, except for Japan, Philippines, Thailand and Republic of Korea (Figure 2), there is ample scope for improvement through the pooling of research and development and technology transfer resources.

TABLE 5
Estimates of N-use efficiency in selected countries

FIGURE 2
Nitrogen-use efficiency in selected countries, 1995/97

A major cause of low N-use efficiency is its loss from the plant-soil system through gaseous emissions, runoff, erosion and leaching. The relative importance of each of these pathways varies from site to site. The adjustment of N fertilizer inputs in accordance with plant demand and appropriate application methods reduces such losses. Gaseous N loss processes include volatilization and denitrification, resulting in the release of NH₃, and NO, N₂O and N₂ to the atmosphere.

Table 6 summarizes the estimated NH3-N loss in Asia’s wetland rice regions. Factors such as climate and soil-crop-water management contribute to regional variations. For Asia, the estimated total NH3 loss from mineral N application in wetland rice for 1995 was 2.2 Mt.

TABLE 6
Estimated NH3 volatilization loss from mineral N applications in wetland rice in Asia, 1995

Source: FAO/IFA, 2001.

The estimated total fertilizer N loss from all the rice ecologies in Asia is 9.6 Mt/year (1995-97). Thus, 65 percent of applied mineral N (worth US$3 120 million) is lost to the environment, with adverse environmental impacts such as greenhouse effects, diminishing stratospheric ozone, acid rain, changes in the global N cycle and nitrate pollution of surface water and groundwater.

In terms of prevailing crop yields and nitrogen-use efficiency levels, an interesting picture emerges (Table 7). Most countries, with the exception of the Republic of Korea, Japan and China, fall within the medium- and low-yield groups, indicating considerable scope for raising yields (Table 7). Of the various N-use efficiency combinations, M and H appear to be the most promising options for minimizing the yield gap while enhancing benefits in terms of dosage economies and reduced environmental threat. Although the level of N use in the Republic of Korea (178 kg/ha) is double that in Japan (88 kg/ha), the yield difference is small (0.2 t/ha) as a result of the efficiency factor. Enhanced N-use efficiency in the Republic of Korea may lead to optimized N use while maintaining yield levels similar to those in Japan. For countries such as China, Indonesia and Iran, improved N-use efficiency accompanied by optimization of N-use levels would be a suitable approach for closing the yield gap. The possibilities for raising yields in Viet Nam, Sri Lanka, Malaysia, India and Pakistan remain high, provided that the prevailing low N-use efficiency levels can be improved. The enhancement of N use to medium levels, coupled with efficiency improvement measures, is important for the Philippines, Bangladesh and Thailand.

TABLE 7
Categorization of selected countries based on rice yield, nitrogen use and nitrogen-use efficiency, 1995-97

^a Yield: H (high) = 5.5 t/ha; M (medium) = 3.6-5.5 t/ha; L(low) = 3.5 t/ha.
^b N use: H (high) = 120 kg/ha; M (medium) = 81-120 kg/ha; L (low) = 80 kg/ha.
^c N-use efficiency: H (high) =55%; M (medium) = 36-55%; L (low) = 35%.

MANAGEMENT PRACTICES FOR INCREASING NITROGEN-USE EFFICIENCY

The required nutrient management procedures are well understood in general terms. Moreover, methodologies (and associated diagnostics) are emerging to facilitate site-specific nutrient management. FAO promotes an integrated nutrient management approach, which uses manufactured (mineral) fertilizers in appropriate combinations with organic and biological sources.

Nitrogen is the nutrient required in the largest quantities while urea is the principal nitrogenous fertilizer for rice production. However, N from urea is subject to considerable losses to the atmosphere and runoff water in the rice ecosystem, especially where urea is broadcast on standing water. On a global level, more than 55 percent of the N applied through fertilizers to irrigated rice is not taken up (the figure is even higher for rainfed lowland rice).

Incorporating urea into the water-saturated soil at a depth of 7 to 10 cm (a reduced zone) minimizes N losses by preventing N from being dissolved in floodwater or oxidized in soil near the surface. This practice is environmentally friendly and enables higher yields with less fertilizer N.

Extensive on-farm trials suggest that the quantification of limiting factors and the adoption of appropriate crop and nutrition management practices can minimize the effects of diminishing returns at increasing N application rates with considerable agronomic, economic and environmental benefits. The (ranked) limiting factors that smallholder rice farmers using prill (or granular) urea can address immediately are: i) too few splits (divided applications of a particular seasonal total fertilizer N) resulting in substantial N losses and consequent inadequate N supply to meet crop requirements at various growth stages; ii) cultivars that may be insufficiently N responsive; and iii) inadequate initial plant population.

Research suggests that farmers and extension agencies can improve N-use efficiency, especially in irrigated rice farming, by using technologies and procedures that can be either incorporated immediately into ongoing national extension programmes or brought rapidly to a state of applicability through relatively small programmes of ecozone-specific adaptive agronomic investigation by extension-cum-research teams.

A multilocation on-farm trial/demonstration project on irrigated rice (1995-98), funded by Japan and implemented by FAO in Indonesia, the Philippines and Malaysia, obtained the following results:

• For a representative grain yield, a total of 58 kg/ha N applied in five equal splits could achieve the same response to that obtained from a seasonal total of 70 kg/ha N applied in three equal splits (a saving of 17%) (Table 8). Leaf-colour chart (for chlorophyll content) guided top-dressing applications showed good promise.

• Deep-placed Indonesian tablet (super granule) urea enabled a 21 percent N saving (compared to 70 kg/ha N applied as prill urea in three splits) and was found to be economically sound.

• Substitution of one-quarter of 70 kg/ha prill urea N by commercial organic fertilizer N resulted in a saving of 8 kg/ha N (11%) of prill urea.

• Banded and paired-row systems of rice establishment, with a corresponding concentration of fertilizer N within these bands and paired rows, resulted in N savings of up to 20 percent (compared to square-array transplanting).

• Establishing an initial rice-plant population density that would result in 40 000 spikelets/m² at harvest achieved an N saving of 25 percent as compared to typical suboptimal populations; the option for increased plant population (perhaps by transplanting 4-5 rather than 2-3seedlings per hill, and at 0.20 x 0.20 m rather than at 0.25 x 0.25 m spacing) proved worthwhile, particularly as the influence of diminishing return to N was less than for other technology options.

• A re-analysis (using Mitscherlich procedures) of existing data sets for ecozone-specific N responses of various farmer-preferred rice cultivars could help identify cultivars with high grain yield responses to realistic N applications in particular ecozones.

• At prevailing prices, prill urea (especially with five splits), deep-placed tablet urea and prill urea in combination with plant-derived N (organic fertilizers) were the most cost-effective of the N formulations and types evaluated.

• For the more efficient technologies, grain yield enhancements were typically 30 to 60 percent higher in more N-responsive than in less N-responsive ecozones, and benefit-cost ratios were 12: 4 as compared to 6: 7.

• Technologies identified as appropriate for immediate adoption could be applied in combination. Deep placement of tablet urea in banded or in paired row configurations yielded savings of about 30 percent. However, there was no conclusive information on the practicability and costs of mechanizing such systems. There is a need to determine the N savings that might result where other individual technologies are combined.

TABLE 8
Economic implications of yield increase estimates for technology components in more responsive situations

The International Fertilizer Development Centre, with funding from the International Fund for Agricultural Development, is engaged in on-farm testing of deep placement of super/mega urea granules in Bangladesh, Nepal and Viet Nam. At village level, locally manufactured machines produce briquettes of about 3 g which are hand-placed at a depth of between 7 and 10 cm in transplanted irrigated rice. Table 9 presents a summary of the trial results (Roy and Nagy, 2002).

TABLE 9
Urea deep placement vs broadcast application

^a 77 kg N/ha.
^b 52 kg N/ha as UDPvs 80 kg N/ha as broadcast prill urea (farmers’ practice).
^c 77 kg N/ha as UDPvs 110 kg N/ha broadcast prill urea (farmers’ practice).
^d N application rates of UDP, on average 35% less than in farmers’ practice.
^e N application rates of UDP, on average 44% less than in farmers’ practice.

In Bangladesh, where urea deep placement (UDP) technology is adopted on about 10 percent of the land under rice cultivation (Table 10), paddy yields have increased by 16 to 25 percent with UDP compared to broadcast urea in three split applications. The additional cost - of making the briquettes (US$10/t of urea) and of labour to place the briquettes (6-8 workdays/ha) - was about US$13/ha. The increase in net revenue was between US$75 and US$148 per ha. These revenue increments result from the use of less fertilizer and the increase in the per-hectare yield in relation to farmers’ practices. The higher yields from UDPwere also due to other improved management practices, such as line planting with optimum space and plant population, better water management and plant protection. In Viet Nam, the consequent yield increase of 14 percent resulted in a net revenue gain of between US$62 and US$78 per ha.

TABLE 10
Contribution of UDP to rice yield, income and employment generation in Bangladesh

^a 8 days/ha; 1 person-year = 300 person-days.

The project concludes that farmers have begun to adopt knowledge concerning the advantages of a long-researched practice. Enabling NGOs to work alongside extension officers facilitates the extension of new adaptable cultivation practices to farmers. The UDP

technology could be readily extended to other countries in the region.

Urea coated with Nimin, a commercial extract from neem (Azadirechta indica) seed, has been widely tested, especially in India. This reasonably inexpensive biological product shows great promise for resource-poor farmers, with an average yield increase of between 5 and 10 percent over uncoated prill urea. Super granules made with Nimin-coated urea and placed deep show a slight improvement over the UDP technology.

At national level, the widespread adoption of suitable combinations of more efficient technologies could generate savings of about 30 percent in the total of prill urea applied to rice. This would imply an upgrading of the average N efficiency from 35 to 50 percent. This could generate additional annual savings of about 1.3 to 1.6 Mt (worth about US$500 million). Such efficiency improvements may lead to a reduction in future manufacturing investments of some US$1 500 million, and would save local economies the cost of remedying water pollution. Furthermore, where there are subsidies on urea production and distribution, it would also lessen the burden on the public purse.

RESEARCH AND DEVELOPMENT ISSUES

The prime concerns relating to rice production are: yield gaps, yield stagnation in irrigated ecology, declining productivity in intensively cultivated areas (localized), low fertilizer efficiency (especially N) and the environmental consequences of input use. Inadequate crop management practices are a major constraint to achieving yield potentials. These inadequacies also lead to inefficient input use. Improved crop management technologies need to be transferred and adapted to local conditions. The current deficiencies are the result of inadequate development and extension. Furthermore, there is a need for strong policy support in fostering farmer-participatory, environmentally-friendly and efficient nutrient management programmes (Roy et al., 2002).

Projections to 2015 and 2030 suggest substantial increases in the rates of N applied to rice, with major environmental implications. The relatively low price of urea N fertilizer (compared to other major nutrients) prevents the widespread adoption of efficient, balanced and environmentally friendly nutrient management. The environmental pricing of urea N, coupled with integrated plant nutrient management, may be an appropriate procedure for fostering N-use efficiency in Asian rice systems.

Nutrient balance studies in many ecosystems have exposed the inadequacies relating to nutrient replenishment and the consequent mining of soil nutrient reserves. These failures need to be rectified through the balanced and efficient use of organic and inorganic plant nutrient sources and through knowledge-intensive integrated soil and nutrient management practices. Compared with input-intensive technologies, such practices are more cost-effective, sustainable and environmentally benign.

Inadequate information on quantified N-uptake efficiency for the various component technologies has been a major constraint to on-farm adoption. Precise quantification is needed for: the major ecozones; varied application rates; farmer-preferred rice cultivars; and transplanted and direct-seeding systems. Furthermore, there needs to be an assessment of the efficiency (and N application decreases) that would result from combinations of the more efficient individual N technology components.

As N application in accordance with the nutrient requirements at the various crop-growth stages is important for fertilizer-use efficiency technologies, colour-card guided top-dressing of N holds good promise. However, such colour cards need to be calibrated for specific ecoregions.

Legume green manures for subsequent rice crops offer considerable potential as partial or total substitutes for N fertilizers, because they can also bring back the lost nutrients from the subsoil to the surface. However, the adoption of green manuring technology has been poor in many areas (especially in countries with limited land resources and high subsidies on mineral N fertilizers), because farmers perceive it to be laborious and uneconomical. As farmers look for multiple benefits from their investments, green manure systems will need to provide additional benefits, such as food, fodder, fuel and commercial products (Garrity and Becker, 1994).

The induction of nitrogen-fixing capabilities in rice through bioengineering processes, such as induced symbiosis and Nif-gene transfer, offers significant potential for reducing the need for mineral N applications. Furthermore, free-living blue green algae and Azolla with its symbiont Anabaena azollae are principal biological nitrogen fixation (BNF) agents in rice fields. Evidence suggests that algal inoculations can yield a benefit equivalent to the application of 25 to 30 kg/ha N. However, their adoption has been constrained by several factors, including maintenance of Azolla inocula between cropping seasons. Measures, such as the selection of superior germplasm, the development of improved Azolla hybrids, and efforts to induce sporulation, should enhance the adoption of this technology (Ladha and Reddy, 2001).

CONCLUSIONS

Growing concerns about the environmental consequences of mineral N use and its future costs have highlighted three approaches:

• reduce N losses by adopting efficient fertilizer use technologies;

• recycle animal and plant residues; and

• develop BNF within the framework of an integrated nutrient management approach.

Even with reduced reliance on mineral N, a firm commitment and the provision of adequate resources from all stakeholders (governments, fertilizer sector, NGOs and international agencies) in transferring improved sound N management technologies as part of a bottom-up participatory approach, can produce more rice, increase farmers’ incomes and reduce environmental pollution.

REFERENCES

FAO. 2000a. Agriculture towards 2015/30. Technical interim report, April 2000. Rome.

FAO. 2000b. Fertilizer requirements in 2015 and 2030. Rome.

FAO/IFA. 2001. Global estimates of gaseous emissions NH3, NO and N2O from agricultural land. Rome.

Garrity, D.P. & Becker, M. 1994. Where do green manures fit in Asian rice systems? In Ladha, J.K. & Garrity, D.P. eds. Green manure production systems for Asian ricelands, p. 1-10. Manila, IRRI.

Ladha, J.K. & Reddy, P.M. 2001. State of knowledge on nitrogen fixation in rice and future prospects. Paper presented at FAO technical expert meeting on increasing the use of biological nitrogen fixation in agriculture, Rome, 13-15 Mar. 2001.

Lin, Y.F. & Shen, M.G. 1996. Rice production constraints in China. In Herdt, R.W. & Hossain, M. eds. Rice research in Asia: progress and priorities, p. 161-178. London, CAB International and Philippines, IRRI.

Pingali, P.L., Hossain, M. & Gerpacio, R.V. 1997. The Asian rice bowls: the returning crisis? London, CAB International and IRRI, Makati, Philippines.

Roy, R.N. & Nagy, J. 2002. Adapting nutrient management technologies (ANMAT): Participatory evaluation, adaptation and adoption of environmentally friendly nutrient management technologies for resource-poor farmers. External review report.

Roy, R.N., Misra, R.V. & Montanez, A. 2002. Decreasing reliance on mineral nitrogen - yet more food. Ambio, 31(2): 177-183.

Siddiq, E.A. 2000. Bridging the rice yield gap in India. In Bridging the rice yield gap in the Asia-Pacific region. Bangkok, FAO-RAP.