Report

# Integration of Environmental, Agronomic, and Economic Aspects of Fertilizer Management

See allHide authors and affiliations

Science  03 Apr 1998:
Vol. 280, Issue 5360, pp. 112-115
DOI: 10.1126/science.280.5360.112

## Abstract

Nitrogen fertilization is a substantial source of nitrogen-containing trace gases that have both regional and global consequences. In the intensive wheat systems of Mexico, typical fertilization practices lead to extremely high fluxes of nitrous oxide (N2O) and nitric oxide (NO). In experiments, lower rates of nitrogen fertilizer, applied later in the crop cycle, reduced the loss of nitrogen without affecting yield and grain quality. Economic analyses projected this alternative practice to save 12 to 17 percent of after-tax profits. A knowledge-intensive approach to fertilizer management can substitute for higher levels of inputs, saving farmers money and reducing environmental costs.

Agricultural intensification through the use of high-yielding crop varieties, chemical fertilizers and pesticides, irrigation, and mechanization—known as the “Green Revolution”—has been responsible for dramatic increases in grain production in developing countries over the past three decades. At the same time, intensification has had environmental consequences such as leaching of nitrate and pesticides, and emissions of environmentally important trace gases. We evaluated the economic and agronomic consequences, and the effects on N trace gas, of fertilizer management in irrigated spring wheat systems in the Yaqui Valley, Sonora, Mexico. This region is one of Mexico's major breadbaskets, so agricultural production and its environmental consequences are regionally important. In addition, as the “home” of the Green Revolution for wheat, the pattern of increasing fertilizer use in the Yaqui Valley provides a gauge of what is likely to occur in other high-productivity irrigated cereal systems of the developing world (1,2).

Globally, application of fertilizer nitrogen (N) has increased rapidly in the last several decades, from 32 Tg N (32 million metric tons) in 1970 to around 80 Tg in 1990 (1 Tg = 1012 g); it is expected to increase to 130 to 150 Tg year−1 by 2050, with two-thirds of that application in developing countries (3). Among the consequences of this change are increased losses of nitrate from soils to freshwater and marine systems and of N-containing gases to the atmosphere (4). Fertilized agriculture is the single most important anthropogenic source of N2O, accounting for over 70% of the anthropogenic sources of this accumulating greenhouse gas (5, 6). Likewise, fertilization results in elevated emissions of NO, a chemically reactive gas that regulates tropospheric ozone production and is a precursor to acid precipitation (7). Research in industrialized countries has shown that management practices can be used to control losses of N (6-9). However, integrated assessments of management alternatives in terms of their ability to reduce N trace gas fluxes and yet be feasible agronomically and attractive economically are wholly lacking. We carried out such an evaluation in the Yaqui Valley (10).

Using daily to weekly sampling frequencies during the 1994/1995 and 1995/1996 wheat cycles, we evaluated changes in soil nutrients and gas fluxes before and after fertilizer additions in experimental plots at the International Maize and Wheat Improvement Center (CIMMYT) field station (11). Several experimental conditions were studied: the conventional farmers' practice for the valley, as determined by farm survey (12); three alternative practices that were based on agronomist recommendations and that added less fertilizer N or fertilizer later in the crop cycle, or both (13); and a nonfertilized control. In our treatment that simulated the farmers' practice, 187 kg N/ha of urea were applied to dry soils 1 month beore planting, followed by preplanting irrigation; an additional 63 kg N/ha of anhydrous ammonia were applied ∼6 weeks after planting.

After the soil was wetted by preplanting irrigation, ammonium (NH4) levels increased markedly to over 600 μg/g (weighted average of bed and furrow positions) and then diminished to near zero as the microbially mediated process of nitrification converted NH4 to nitrate (NO3) (14,15). By the 1994 planting date, 116 kg/ha of NO3-N were left in the top 15 cm of soil, with very little remaining in the NH4 form. A similar pattern of transformation and loss was evident in the 1995/1996 wheat season.

Changes in N trace gas fluxes mirrored changes in the soil pools of inorganic N. The farmers' practice resulted in very large emissions of N2O and NO in both years (Fig.1), with preplanting gas fluxes summing to 5.6 and 4.6 kg N/ha in the 1994/ 1995 and 1995/1996 wheat cycles, respectively, and crop cycle fluxes summing to 6.61 and 11.3 kg N/ha, respectively (16, 17). In the 1994/1995 study, average fluxes at midday in the bed positions (where most of the fertilizer was located) ranged up to 650 ng cm 2 hour 1 for N2O-N and 300 ng cm 2hour 1 for NO-N in the period before planting (15). In 1995/1996, which had less rainfall during the preplanting period, average N2O and NO fluxes in the beds ranged up to 100 and 550 ng cm 2hour 1, respectively, during the same period (15). These values are among the highest ever reported (6-8).

All but one of the alternative practices evaluated during the two study years resulted in significant reductions in crop-cycle N gas emissions as compared with the farmers' practice; the alternative in which 250 kg/ha N were applied, with 33% preplanting, 0% at planting, and 67% after planting, lost 6.93 kg/ha of N2O plus NO-N, in contrast to 6.61 kg/ha lost in the farmers' practice. In both years, the alternative in which 250 kg/ha N were added, with 33% at planting and 67% after planting, lost at least 50% less N gas than the farmers' practice. In the “best” alternative with respect to reduced N2O and NO emissions, a total of 180 kg N/ha were applied, with 33% at planting and 67% 6 weeks afterplanting (Fig. 1). In this treatment, total fluxes of N2O and NO, summed over the 1995/1996 wheat cycle, were 0.74 kg N/ha.

The interplay between the timing of fertilization and irrigation was critical to inorganic N transformations and gas losses in this site. When fertilizer was added to dry soils, only very small changes in inorganic N concentrations or N gas emissions were measured. With irrigation, however, rapid conversion of urea to NH4 was followed by nitrification of NH4 to NO3. High losses of N2O occurred soon after irrigation, largely resulting from denitrification under waterlogged conditions (18). As the soils dried, NO emissions increased, produced during nitrification (14). Both N2O and NO emissions dropped substantially by planting (Fig. 1) (15), when denitrification apparently was limited by a well-aerated soil environment and nitrification was limited by the low availability of NH4. Process studies with 15N-labeled NO3 and NH4 confirmed these patterns and controls (19).

Emissions of N2O and NO under the farmers' practice were large relative to those observed in many other studies. However, they represent just two of several important pathways by which N can be lost from terrestrial ecosystems; others include ammonia volatilization, nitrate leaching, and dinitrogen gas flux (20). As seen by farmers, total loss of N is of more interest than specific trace gas losses, as it represents wasted fertilizer. To determine the total loss of fertilizer N, we applied 15N-labeled urea (in place of the fertilizer) in isolated plots that were otherwise treated like the experimental plots; at the end of the crop cycle, the isolated plots were harvested and 15N recovery in soil and plant components was measured (21). In the farmers' practice and in our best alternative, proportional recovery of the applied N in plants was 46 and 57%, respectively, and recovery in soil to a meter depth was 26 and 16%, respectively. Because less N was added in the alternative (180 kg/ha), quantitatively less N was lost than in the farmers' practice (48 kg/ha lost in the alternative versus 70 kg/ha in the farmers' practice).

Fertilizer use and loss are just one component of farm budgets, and farmers typically focus not only on costs but on the balance between costs and expected income under some degree of price and production uncertainty. For wheat farmers in the Yaqui Valley, yield of good-quality wheat provides the essential income. Yields reported in our socioeconomic surveys in 1994/1995 and 1995/1996 ranged from 3.1 to 7.3 tons/ha, with average values of 4.9 and 5.3 tons/ha for the two seasons, respectively (12). Mean yields in our simulated farmer practice were 6.08 ± 0.18 and 6.07 ± 0.28 tons/ha in 1994/1995 and 1995/1996, respectively (22). Our best alternative, in which 180 kg N/ha were added as compared with 250 kg N/ha in the farmers' practice, resulted in yields that were not significantly different (6.16 ± 1.3 tons/ha). Likewise, grain quality (estimated as the protein concentration in grain) in the alternative was not significantly different from the farmers' practice (14.87 versus 14.83%, respectively) (22).

As in many high-productivity agricultural systems of the developing world, the dissemination of Green Revolution technologies initially provided farmers in the Yaqui Valley with modern seed varieties and highly subsidized N fertilizers. In recent years, however, the reduction of subsidies (in real terms) has been dramatic (23). Our economic analysis of farmers' costs and returns for both 1994/1995 and 1995/1996 wheat seasons indicates that fertilization has now become the highest direct production cost in the Yaqui Valley farm budgets (Table 1). During the 2 years of our study, fertilization exceeded even the costs of land preparation, which traditionally have represented the largest cost category in this highly mechanized system; just 5 years ago, the cost of land preparation was 50% higher than that of fertilization (2).

Table 1

Costs and returns for the 1994/1995 and 1995/1996 wheat cycles. Data were collected during on-farm surveys (12).

View this table:

Given the importance of fertilizer in the Yaqui Valley farm budgets, we evaluated the extent to which increased fertilizer efficiency represented a significant budgetary savings to the farmers (15). In contrasting the farmers' practice with our best alternative, we found that the alternative resulted in a savings of new pesos (N$) 414 to N$571/ha, or U.S.$55 to U.S.$76/ha at then-existing exchange rates. These values, which resulted from lower fertilizer applications and reduced loss of fertilizer, were equivalent to 12 to 17% savings of after-tax profits from wheat farming in the Yaqui Valley. Such potential cost savings may over time induce a shift in technology and management toward fertilization later in the wheat cycle; indeed, our on-going surveys indicate that some farmers are now postponing their first fertilizer application until planting. However, farmers may also face greater risks of low yields with our best alternative, particularly in years when late rains delay the second fertilizer application beyond the point of optimal plant response (24). The importance of such real or perceived risks, and the development of recommendations that are sensitive to them, are topics of our current research.

Our results demonstrate that alternative fertilizer practices can reduce trace gas and total losses of fertilizer and maintain yields. These alternatives, which require greater knowledge about efficient use of nutrients, can substitute for higher levels of those inputs and might ultimately allow Yaqui Valley farmers to remain competitive in an era of economic liberalization and expanding free trade. At the same time, they reduce the environmental costs of agriculture, some of which are directly felt in the Yaqui Valley, and others of which are globally important. An integration of agronomic knowledge of practical alternatives, economic analysis of their on-farm costs and benefits, and biogeochemical analysis of their consequences in soils and the atmosphere can provide the basis for the identification or development of win-win solutions.

View Abstract