IFA International Workshop on Enhanced-Efficiency Fertilizers

Frankfurt, Germany, 28-30 June 2005

NITROGEN USE EFFICIENCY – STATE OF THE ART

A. DOBERMANN
University of Nebraska, USA

University of Nebraska-Lincoln
P.O. Box 830915, Lincoln, NE 68583-0915, USA
Tel: +1 (402) 472 1501 Fax: +1 (402) 472 7904 Email: [email protected]
 IFA International Workshop on Enhanced-Efficiency Fertilizers
Frankfurt, Germany, 28-30 June 2005

“Nitrogen Use Efficiency – State of the Art”

A. Dobermann
University of Nebraska, USA

Reactive nitrogen and the need to increase fertilizer nitrogen use efficiency
Nitrogenous fertilizers have contributed much to the remarkable increase in food production that has
occurred during the past 50 years (Smil, 2001). Globally, however, N fertilizers also account for 33% of
the total annual creation of Nr or 63% of all anthropogenic sources of reactive nitrogen (Nr) (Table 1).
Reactive nitrogen is defined as all biologically, photochemically, and/or radiatively active forms of N --
a diverse pool of nitrogenous compounds that includes organic compounds (e.g. urea, amines, proteins,
amides), mineral N forms, such as NO3- and NH4+ as well as gases that are chemically active in the
troposphere (NOx, NH3, N2O) and contribute to air pollution and the greenhouse effect (Galloway et al.,
1995). Asia alone accounts for more than 50% of the global N fertilizer consumption as well as 37% for
the global Nr creation. Smil (1999) estimated that only about half of all anthropogenic N inputs to
cropland are taken up by harvested crops and their residues, with the remainder contributing significantly
to Nr enrichment of the atmosphere, ground and surface waters.

Table 1: Global creation of reactive N from anthropogenic and natural sources in the mid 1990s (Boyer et al., 2004).

Total
Anthropogenic (million t/yr) Natural (million t/yr)
Region Fertilizer BNF Import Depos. Total BNF Lightng. Total
Africa 2.1 1.8 0.5 2.9 7.3 25.9 1.4 27.3 34.6
Asia 44.2 13.7 2.3 3.8 64.0 21.4 1.2 22.6 86.6
Europe + FSU 12.9 3.9 1.0 2.9 20.7 14.8 0.1 14.9 35.6
Latin America 5.1 5.0 -0.9 1.8 11.0 26.5 1.4 27.9 38.9
N. America 12.6 6.0 -2.9 2.7 18.4 11.9 0.2 12.1 30.5
Oceania 0.7 1.1 -0.3 0.3 1.8 6.5 0.2 6.7 8.5
Total 77.6 31.5 -0.3 14.4 123.2 107.0 4.5 111.5 234.7

It is widely believed that accumulation of excessive amounts of Nr in terrestrial and aquatic ecosystems as
well as in the troposphere leads to significant costs to society that occur through direct and indirect
negative effects on environmental quality, ecosystem services, biodiversity, and human health (Pretty et
al., 2000; Schweigert and van der Ploeg, 2000; Townsend et al., 2003). Such estimates are not very
precise, however, and it is not clear whether they place an appropriate value on the large positive impact
of N fertilizer on ensuring food security and adequate human nutrition.

1
Environmental benefits also accrue from fertilizer use by avoiding expansion of agriculture into natural
ecosystems and marginal areas that cannot sustain crop production and provide critical habitat for
protecting biodiversity (Cassman et al., 2003). Regardless of what the true societal costs of accumulation
of Nr in cultivated and natural ecosystems are, it is clear that Nr creation associated with human activities
must slow down. Mitigation options include:
(i) Reduction of Nr emissions from fossil fuel combustion,
(ii) Transformation of Nr to non-reactive N forms (e.g., denitrification to N2 or sequestration of N in
soil organic matter),
(iii) Changes in human diet and associated changes in food, feed, and fertilizer demand, and
(iv) Improvements in fertilizer nitrogen use efficiency (NUE) in agricultural systems: less N fertilizer
per unit food produced.
Many of these mitigation strategies are of long-term nature and they are closely linked to policy decisions
that need to be made. However, improving NUE in agriculture has been a concern for decades and
numerous new technologies have been developed in recent years to achieve this. Therefore, fertilizers and
their management will be at forefront of measures to improve the global N balance over both the short-
and long-term.
This paper addresses three issues: (i) definition and measurement of NUE, (ii) global status of NUE in
agriculture, and (iii) a brief outline of major technology options for increasing NUE. I primarily focus on
cereals because they account for nearly 60%% of global N fertilizer use (IFA, 2002) and represent 20% of
the global annual creation of Nr.

Definition and measurement of nitrogen use efficiency

Nitrogen budgets
Nitrogen budgeting approaches are often used to evaluate system-level N use efficiency and to understand
N cycling by estimates of input, storage and export processes by mass balance. A surplus or deficit is a
measure of the net depletion (output > input) or enrichment (output < input) of the system, or simply of
the ‘unaccounted for’ N. This approach is used in research studies that aim at the identification of the fate
of N surpluses or in long-term assessment of N flows and their respective impact and soil and the
environment in managed or natural ecosystems. Unlike many of the agronomic indices of NUE described
below, N budgeting approaches are also suitable for systems that are not at a relative equilibrium in terms
of N, i.e., systems in which either significant accumulation or losses of N from indigenous sources occur.
Nitrogen budgets can be constructed for a different time periods at any scale, ranging from an agricultural
management unit to regional and continental scales. The degree of detail depends on the purpose of
budgeting and on the resources available to collect the information. For example, partial budgets that do
not include all inputs can be used to estimate the N balance, provided that the major fluxes in and out of
the ecosystem are accounted for. Budgets constructed for the purpose of guiding agricultural management
or government policy decisions often consist of simple mass balances.
A more complete budget analysis quantifying the relative role of various N inputs and outputs and the
distribution and turnover of N among internal compartments is necessary in order to gain mechanistic
understanding of the ecosystem. This limits the use of such approaches in evaluating fertilizer
management strategies, new fertilizer products, or other technologies. Nitrogen budgets may also be
difficult to compare because of different purposes and approaches used for making N budgets. Therefore,
methodologies used must be clearly described and N budgets should include statements about scales and
uncertainties associated with the estimates.

2
Agronomic indices of nitrogen use efficiency
Various indices are commonly used in agronomic research to assess the efficiency of applied N (Novoa
and Loomis, 1981; Cassman et al., 2002), mainly for purposes that emphasize crop response to N (Table
2). In field studies, these indices are either calculated based on differences in crop yield and total N uptake
with aboveground biomass between fertilized plots and an unfertilized control (‘difference method’), or by
using 15N-labeled fertilizers to estimate crop and soil recovery of applied N. Time scale is usually one
cropping season. Spatial scale for measurement is mostly a field or plot. Because each of the indices in
Table 2 has a different interpretation value, research on fertilizer-N efficiency should include
measurements of several indices in order to assess causes of variation in NUE.
The agronomic framework is most useful for understanding the factors governing N uptake and fertilizer
efficiency, to compare short-term NUE in different environments, and to evaluate different N management
strategies or technologies. The ‘difference method’ is simple and cost-efficient, which makes it
particularly suitable for on-farm research. However, measurement of NUE requires careful
experimentation and interpretation must consider potentially confounding factors. Agronomic efficiency
(AEN) and recovery efficiency (REN) are not appropriate indices of NUE when comparing cropping
practices such as crop establishment methods or different water management regimes when the crop yield
in control treatments (Y0) differs significantly because of these management practices. In these instances,
PFPN is a more appropriate index for making comparisons. Comparisons of REN and physiological
efficiency (PEN) among genotypes should use agronomically fit varieties and avoid comparisons with
‘inferior germplasm’ not adapted to the particular growth conditions. Caution is required when using AEN,
REN or PEN for assessing trends in NUE in long-term experiments because depletion of indigenous soil N
in permanent 0-N plots will lead to overestimation of the true NUE in fertilized plots. Results obtained
with the ‘difference method’ may also be confounded by added-N interactions, i.e., differences in N
mineralization rates from soil organic matter and crop residues between +N and 0-N plots. Since many of
the indices in Table 2 are ratios of several measurements, sampling and/or measurement errors can cause
significant errors.
Agronomic NUE indices only provide accurate assessment of NUE for systems that are at a relatively
steady-state with regard to soil organic N content and where differences in root systems between
unfertilized and fertilized crops are relatively small. Nitrogen in roots as well as any net accumulation of
N from fertilizer in soil organic matter and its effect on the indigenous soil N supply for subsequently
grown crops cannot be easily accounted for. This may lead to an underestimation of the overall system
level efficiency of applied N inputs. Therefore, N budgeting or 15N methods should be used to assess the
fate of N in the entire soil-crop systems over longer time periods and across different spatial scales.
Compared to the difference method, no 0-N plot is required for estimating REN using 15N, but costs tend to
be higher and a generally higher level of sampling and measurement quality is required. This limits the
use of this method in on-farm studies. In general, 15N methods tend to produce results that are similar to
those obtained with the difference method, but the relationships between REN values obtained with both
methods is often quite scattered (Krupnik et al., 2004). Overall, REN values obtained with 15N are often
slightly lower than those estimated with the difference method because of confounding effects related to
pool substitution, i.e., immobilization of 15N fertilizer in microbial biomass and initial release of
microbial-derived 14N. Ladha et al. (2005) estimated an average worldwide REN for cereal research trials
of 51% measured with the difference method as compared to 44% measured with the 15N method.
However, their estimates were not based on paired comparisons at the same sites.

3
Table 2: Agronomic indices of N use efficiency and their typical ranges in cereals.
NUE index Calculation Interpretation Common values
• Most important for farmers because it
PFPN - Partial factor PFPN = YN/FN integrates the use efficiency of both 40–70 kg grain kg-1
productivity of applied indigenous and applied N resources: N
N (often simply called PFPN = (Y0/FN) + AEN
>70 kg kg-1 at low
nitrogen use efficiency • Increasing indigenous soil N
rates of N or in very
or NUE) (Y0) and the efficiency of
efficiently managed
applied N (AEN) are equally systems
(kg harvest product per
kg N applied) important for improving PFPN
• Limited potential for identifying
see Fig. 1a insert specific constraints or promising
management strategies .
• AEN is the product of the
AEN = Agronomic AEN = (YN – Y0)/FN 10–30 kg grain kg-1
efficiency of N recovery from
efficiency of applied N N
applied N and the efficiency
(kg yield increase per with which the plant uses each
>30 kg kg-1 in well-
kg N applied) additional unit of N acquired:
managed systems or
AEN = REN x PEN
see Fig. 1a at low levels of N use
• AEN can be increased by N, or low soil N supply
crop, and soil management
practices that affect REN, PEN,
or both.
• REN depends on the congruence
REN = Crop recovery REN = (UN – U0)/FN 0.30–0.50 kg kg-1
between plant N demand and the
efficiency of applied N
quantity of N released from 0.50–0.80 kg kg-1 in
(kg increase in N
applied N. well-managed
uptake per kg
• REN is affected by the N systems or at low
N applied)
application method (amount, levels of N use or low
see Fig. 1c
timing, placement, N form) as soil N supply
well as by factors that determine
the size of the crop N sink
(genotype, climate, plant
density, abiotic/biotic stresses).
• PEN represents the ability of a
PEN = Physiological PEN = (YN – Y0)/(UN – 30–60 kg kg-1
plant to transform N acquired
efficiency of applied N U0)
from fertilizer into economic >60 kg kg-1 in well-
(kg yield increase per yield (grain). managed systems or
kg increase in N uptake • PEN depends on genotypic at low levels of N use
from fertilizer) characteristics (e.g., harvest or low soil N supply
see Fig. 1d index), environmental and
management factors, particularly
during reproductive growth.
• Low PEN suggests sub-optimal
growth (nutrient deficiencies,
drought stress, heat stress,
mineral toxicities, pests).
-1
FN – amount of (fertilizer) N applied (kg ha )
YN – crop yield with applied N (kg ha-1)
Y0 – crop yield (kg ha-1) in a control treatment with no N
UN – total plant N uptake in aboveground biomass at maturity (kg ha-1) in a plot that received N
U0 – the total N uptake in aboveground biomass at maturity (kg ha-1) in a plot that received no N
4
For the same soil and cropping conditions, NUE generally decreases with increasing N rate (Fig. 1). Crop
yield (Y) and plant N accumulation (U) typically increase with increasing N rate (F) and gradually
approach a ceiling (Figures 1a and 1c). The level of this ceiling is determined by the site yield potential.
At low levels of N supply, rates of increase in yield and N uptake are large because N is the primary factor
limiting crop growth and final yield. As the N supply increases, incremental yield gains become smaller
because yield determinants other than N become more limiting as the maximum yield potential is
approached.

Gross return above N cost (GRF, $ ha-1)

90 1200 5

Incremental GRF (dGRF/dF, $/ha kg-1)

17 (a) (b)

Incremental AE (dY/dF, kg kg-1)

Y=f(F) 80
Grain yield (Y, 1000 kg ha-1)

1150 GRF=f(F) 4
16 70
AE=31 AE=23 AE=16
1100 3
15 AE=44 60
50 1050 2
600
PFP (kg kg-1)

14 500
max. profit
400 40 1000 1
300
13 200 30
100 950 0
12 0 20 dGRF/dF
0 50 100 150 200 250 300
900 -1
dY/dF 10
11
0 850 -2
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Fertilizer N (F, kg ha-1) Fertilizer N (F, kg ha-1)

320 1.0
(c) 17 (d)

Incremental PE (dY/dU, kg kg-1)

100
Incremental RE (dU/dF, kg kg-1)

300 RE=0.41 PE=48

Grain yield (Y, 1000 kg ha )

U=f(F)
-1
Plant N uptake (U, kg ha-1)

0.8 16 PE=39
RE=0.49 PE=49
280 80
RE=0.62
15 Y=f(U) PE=58
260 0.6
RE=0.77 60
14
240 0.4
13 40
220 dU/dF
0.2 12 dY/dU
20
200
11
180 0.0 0
0 50 100 150 200 250 300 180 200 220 240 260 280 300 320
Fertilizer N (F, kg ha-1) Plant N uptake (U, kg ha-1)

Figure 1: Response of irrigated maize to N application at Clay Center, Nebraska, 2002: (a) relationship between
grain yield (Y) and N rate (F) and the incremental agronomic N efficiency (AE, kg grain yield increase per kg N
applied); (b) relationship between gross return above fertilizer cost (GRF) and N rate and the incremental GRF
(dGRF/dF); (c) relationship between plant N accumulation (U) and N rate and the incremental recovery efficiency of
fertilizer N (RE, kg increase in N uptake per kg N applied), (d) relationship between grain yield and plant N
accumulation (U) and the incremental physiological efficiency of fertilizer N (PE, kg increase in grain yield per kg N
taken up). Dashed lines indicate where maximum profit occurred. Measured values of AE (a), RE (c) and PE (d)
calculated by the difference method are shown for the four N rates used. The insert in graph (a) shows the decline in
PFPN (ratio Y/F) with increasing N rate (Dobermann and Cassman, 2004).

5
The broadest measure of NUE is the ratio of yield to the amount of applied N, also called the partial factor
productivity [PFPN] of applied N, which declines with increasing N application rates (Figure 1a insert).
The PFPN is an aggregate efficiency index that includes contributions to crop yield derived from uptake of
indigenous soil N, N fertilizer uptake efficiency, and the efficiency with which N acquired by the plant is
converted to grain yield. In addition to N uptake by the crop and N losses, a portion of the N applied is
retained in soil as residual inorganic N (either ammonium or nitrate) or incorporated into various organic
N pools—including microbial biomass and soil organic matter. Such retention should be considered a
positive contribution to N input efficiency only when there is a net increase in total soil N content.
Because more than 95% of total soil N is typically found in organic N pools, an increase in soil organic
matter (i.e. carbon sequestration) is required to achieve increases in total soil N. Sustained increases in
organic matter in cropping systems practiced on aerated soils (e.g. maize- and wheat-based systems
without irrigated rice) result in greater indigenous N supply from decomposition of the organic N pools,
which can reduce N fertilizer requirements to maintain yields and thereby increase PFPN (Bell, 1993;
Kolberg et al., 1999). In contrast, greater soil organic matter in continuous irrigated rice systems does not
necessarily result in an increase in N mineralization because there is little relationship between soil
organic matter content and indigenous soil N supply in anaerobic soils (Cassman et al., 1996a; Dobermann
et al., 2003). For cropping systems in which soil organic matter is declining over time, there is an
additional loss of N above that from applied N fertilizer and organic N sources. This additional loss of N
reduces PFPN and greater amounts of applied N are required to maintain yields.
Figure 1 also illustrates how, alternatively to calculating NUE indices for few fixed levels of N application
only, continuous response functions between yield, plant N uptake, and fertilizer N input can be fitted to
more accurately quantify the curvilinear nature of crop response to N application. The incremental yield
increase that results from N application at any point along the N response curve is the first derivative of
the fitted model describing the relationship between yield and N rate, which we may also call the
incremental agronomic efficiency from applied N (AEi = dY/dF in Fig. 1a). Likewise, the AEi is the
product of the efficiency of N recovery from applied N sources (incremental recovery efficiency, REi =
dU/dF in Fig. 1c) and the efficiency with which the plant uses each unit of N acquired from applied N to
produce grain (incremental physiological efficiency, PEi = dY/dU in Fig. 1b). The REi largely depends on
the degree of congruence between plant N demand and the available supply of N from applied fertilizer or
organic N sources. Consequently, optimizing the timing, quantity, and availability of applied N is the key
to achieving high REi.

Global status of N use efficiency

World consumption of N fertilizers has averaged 83 million metric tons (Mt) in recent years, of which
about 47 Mt is applied to cereal crops (Table 3). The share of total N fertilizer consumption that is applied
to cereals ranges from a low of 32% in Northeast Asia to more than 71% in SE Asia. At a global scale,
cereal production (slope = 31 x 106 Mg yr-1), cereal yields (slope = 45 kg yr-1), and fertilizer N
consumption (slope = 2 Mt yr-1) have increased in a near-linear fashion during the past 40 years. However,
significant differences exist among world regions, particularly with regard to N use efficiency. On a global
or regional scale, PFPN is the only index of NUE that can be estimated reasonably well, although not very
precisely because of uncertainties about the actual N use by different crops. Because PFPN is a ratio, it
always declines from large values at small N application rates to smaller values at high N application
rates. Thus, differences in the average cereal PFPN among world regions depend on which cereal crops are
grown, their attainable yield potential, soil quality, amount and form of N application, and the overall
timeliness and quality of other crop management operations.

6
Table 3: Current levels of cereal production, nitrogen fertilizer use on cereals, and cereal nitrogen use efficiency
by world regions. Values shown represent annual means for the 1999 to 2002/03 period.

Developed Transitional/Developing World

North NE W E Europe Ocean. Africa W Asia South SE East Latin
America Asia Europe C Asia NE Africa Asia Asia Asia America

Cereal prod. (Mt) 377 19 208 216 34 98 81 307 141 447 144 2072
-1
Cereal yield (t ha ) 5.1 6.1 5.5 2.1 1.9 1.1 2.3 2.4 3.2 4.8 2.9 3.1
1
Total N use (Mt) 12.5 0.9 9.5 4.9 1.3 1.4 4.2 14.6 4.0 24.9 5.1 83.2
2
Cereal share N (%) 66 32 45 51 67 56 56 50 71 58 53 57
N use cereals (Mt) 8.3 0.3 4.3 2.5 0.9 0.8 2.4 7.3 2.8 14.5 2.7 46.7
-1 3
N rate (kg N ha ) 112 89 113 25 48 9 68 58 65 155 55 70
PFPN (kg kg-1)4 45 71 59 90 46 123 34 44 53 32 55 44
5
Relative PFP 1.0 1.6 1.4 2.1 1.1 2.8 0.8 1.0 1.2 0.7 1.3 1.0
1
Total fertilizer N consumption by all crops (FAO, 2004).
2
Estimated share of cereal N use of total N consumption, calculated as weighted average of country-specific
estimates of fertilizer use by crops (IFA, 2002). Weights were proportional to N use by countries.
3
Estimated average N application rate on all cereal crops.
4
Average partial factor productivity of applied N = kg grain yield per kg N applied.
5
PFPN relative to world average (World = 1).

At global level, PFPN in cereal production has decreased from of 245 kg grain kg-1 N in 1961/65, to 52 kg
kg-1 in 1981/85, and is currently about 44 kg kg-1. This decrease in PFPN occurs as farmers move yields
higher along a fixed response function unless offsetting factors, such as improved management that
remove constraints on yield, shift the response function up. In other words, an initial decline in PFPN is an
expected consequence of the adoption of N fertilizers by farmers and not necessarily bad within a systems
context.
In developing regions, N fertilizer use was small in the early 1960s and increased exponentially during the
course of the Green Revolution. Although the growth rate in N consumption has slowed substantially in
recent years, it still averaged 1.45 Mt N yr-1 (3.2% yr-1) during the past 20 years. The large increase in N
use since the 1960s resulted in a steep decrease in PFPN in all developing regions (Fig. 2). However,
average regional N rates on cereals range from less than 10 kg N ha-1 in Africa to more than 150 kg N ha-1
in East Asia (Table 3) and, with the exception of Africa, PFPN continues to decline in all developing
regions at rates of –1 to –2% yr-1 (Fig. 2). The low PFPN in East Asia, which is dominated by China, is of
particular concern for the global Nr budget because this region uses the greatest amount of N fertilizer
(Table 1). Declines in PFPN on cereal production in developing countries will likely continue without
greater investment in research and extension to reverse this trend.
In developed regions, excluding Eastern Europe/Central Asia, cereal yields have continued to increase in
the past 20 years without significant increases in N fertilizer use. As a consequence, average PFPN has
remained virtually unchanged at 49 kg kg-1 since the early 1980s. Trends of increasing PFPN have
occurred in some regions (Fig. 3), e.g., Western Europe (mostly rainfed wheat with high yields) and
Northeast Asia (irrigated rice).

7
In North America, average cereal PFPN has changed little because of low PFPN in dryland wheat areas
with low and variable yields, while PFPN of maize has increased substantially (Dobermann and Cassman,
2002). At present, average cereal yields in North America, Western Europe, and East Asia are 60 to 100%
above the world average, even though the N rates applied are only 30 to 60% above world average rates
(Table 3). High yields and high PFPN in these regions result from a combination of fertile soils, favorable
climate, and improved crop and soil management practices, including N fertilizer management. Trends of
increasing PFPN are likely to continue in developed countries because they primarily result from
investments in research and extension on crop improvement, new fertilizer products, and better
management technologies by both public and private sectors, at levels that greatly exceed those currently
available in the developing world.

Developed Developing
N America S Asia
400 400
W Europe SE Asia
300 300
PFPN (kg grain/kg N)

E Europe + C Asia E Asia

NE Asia W Asia + N Africa
200 Oceania 200 Africa
Latin America

100 100
70 70
50 50
40 40
30 30

20 20
1960 1970 1980 1990 2000 1960 1970 1980 1990 2000
5
5
4
Annual growth rate (%/yr)

4 1961/65-1981/85 1961/65-1981/85
3 1981/85-2001/02 3 1981/85-2001/02
2 2
1 1
0 0
-1 -1
-2 -2
-3 -3
-4 -4
-5 -5
-6 -6
-7 -7
-8 -8
-9 -9
a e sia ia ia sia ia sia ica ica ica
ric rop As ean As Afr Afr Amer
Ame -Eu e/ CA NE Oc SA SE EA E
N W
uro
p ia/N Lat
in
EE As
W

Figure 2: Regional trends in nitrogen use efficiency in cereals. Note: a logarithmic scale was used for the NUE axis.

8
The very high PFPN in Africa (123 kg kg-1) and Eastern Europe/Central Asia (90 kg kg-1) are indicative of
soil N mining. Fertilizer use in Africa has lagged behind other world regions and is a major reason for the
low cereal yields in this region (Table 3). In Eastern Europe and countries of the former Soviet Union
(FSU), N fertilizer use on cereals dropped drastically in the late 1980s as a result of political and economic
turmoil. Consequently, PFPN doubled from 1988 to 2000 without improvements in yield potential or
major changes in N management. Because these trends of increasing PFPN in both Africa and Eastern-
Central Europe are likely associated with a mining of soil N resources, they are not sustainable over the
long-term and we would expect yields to stagnate or even decline unless greater amounts of N fertilizer
are used in cereal production.
The trends shown in Figure 2 depend on the reliability of the aggregate data on crop yields and fertilizer
use. Both are difficult to validate. Data on fertilizer use by individual crops within countries and regions
are notoriously difficult to obtain and we do not have reliable series. For many countries, the values used
were derived from estimated total N fertilizer use and expert estimates of the average N fertilizer use by
crop (IFA, 2002). Very few countries collect more detailed information. Despite these caveats, there are
several pieces of supporting evidence. One assumption we made in calculating trends in NUE (Fig. 2) is
that the share of total N fertilizer consumption by cereals within a region has not changed substantially
since the early 1960s. In the USA, for example, surveys of cropping practices are annually conducted with
sample sizes of several thousand farmers (http://www.ers.usda.gov). Those data indicate that the cereal
share of total N consumption has remained virtually unchanged since the mid 1960s. In our approach,
average PFPN for rice grown worldwide was estimated at 44 kg kg-1 (data not shown). This value is in
reasonable agreement with an average PFPN of 46 kg kg-1 as directly measured in on-farm studies
conducted on 400 farmers’ fields in South Asia, East Asia, Southeast Asia and West Africa (Adhikari et
al., 1999; Wopereis et al., 1999; Haefele et al., 2001; Dobermann et al., 2002)
The relationship between the mean national cereal yield and the mean rate of N fertilizer applied to cereal
crops on a country-by-country basis is linear and it provides an estimate of the ‘global’ average AEN in
cereals (Fig. 3). On a global basis, the slope of the regression suggests that global cereal production will
increase by 30 kg ha-1 for each kg of additional N fertilizer. The slopes and intercepts (yield at zero N
applied), however, differ significantly among crops (Cassman et al., 2003). Rice, for example, often yields
more with no N fertilizer applied than wheat or maize because of greater N supply from indigenous soil
resources. Thus the slope of the regression is lower for rice (26 kg kg-1) than for wheat and maize (36-41
kg kg-1, not shown). Actual N response within countries or at farm level varies widely due to differences
in climate, soil fertility and the technological sophistication of crop management.
Figure 3 also illustrates the potential global impact of increasing NUE in agricultural systems. If losses of
cereal cropping area continue at present rates and fertilizer-N efficiency cannot be increased substantially,
a 60% increase in global N consumption by cereals or 74% increase in average N rates per ha would be
required to meet the predicted 38% increase in cereal demand by 2025 (Scenario 1). Such a large increase
in N consumption would have major environmental consequences at local, regional, and global scales
through continued accumulation of different forms of Nr. On the other hand, the predicted cereal demand
can be met by only a 30% increase in global N fertilizer use on cereals if the incremental cereal yield
response to applied N can be increased by about 20% within a period of 20 years (Dobermann and
Cassman, 2005). Such a level of increase in NUE is well within the scatter of the present ‘global N
response curve’, i.e., there are many countries in which even higher NUE has already been achieved.

9
Ladha et al. (2005) provide a summary of published literature data on fertilizer-N efficiency in cereal
crops. In their analysis, the average REN in aboveground biomass (grain+straw) in research plots was 44%
in rice, 54% in wheat and 63% in maize (Table 4). Recovery in grain alone averaged 35 to 44% for the
three major cereals, which is significantly higher than the crude global estimate (33%) suggested by Raun
and Johnson (1999). Not included in this is fertilizer-N recovered in roots, N recovered in subsequently
grown crops, and N that remains in the soil N.

9
Yield N rate AEN PFPN
80 countries, 1994-97 level (t/ha) kg/ha kg/kg kg/kg
Mean cereal yield 1994-97 (t/ha)

8
Current ∆Y/∆Nrate Present average
7 (Y = 1.0 + 0.030 N) 3.1 70 30 44
+20% in ∆Y/∆Nrate 2025 scenario 1
(Y = 1.0 + 0.036 N) 4.6 121 30 38
6 2025 scenario 2
4.6 101 36 46
5

2
Present 2025 2025
1 average scenario scenario
2 1
0
0 20 40 60 80 100 120 140 160
Cereal fertilizer N rate (kg N/ha harvested area)
Figure 3: Global relationships between average cereal yields and average fertilizer-N use for 81 countries during
the late 1990s. The solid line indicates the present average N response of all cereals to fertilizer N application. The
dashed line indicates a possible increase in NUE due to a 20% increase in the slope of the average N response
( Y/ Nrate), but no change in the intercept. Drop lines and values in the table show the effect of different N
response on present and required N rates and NUE at global yield levels for two future scenarios in which cereal
harvest area continues to decline slowly until 2025, but NUE either increases or decreases: [1]: No change in the
global N response function. Yield increases are mainly associated with increasing N rates (move along the current N
response function); [2]: A 20% increase in the slope of the global N response function. Yield increases are
associated with increasing N rate and increasing NUE (Dobermann and Cassman, 2005).

In field studies with rice and dryland systems, average 15N fertilizer recovery was 3.3% in the 1st
subsequent crop, 1.3% in the 2nd subsequent crop, 1.0% in the 3rd subsequent crop, 0.4% in the 4th
subsequent crop, and 0.5% in the 5th subsequent crop, or 6.5% in total (IAEA, 2003; Krupnik et al., 2004).
Thus, together with an average first-crop REN of 51% (difference method) or 44% (15N method), total crop
N recovery from a one-time application of N averages about 50 to 57% in research trials with cereals. The
remainder is either stored in soil organic matter pools or lost from the cropping system. In the IAEA trials,
the average amount of 15N fertilizer recovered in soil after five growing season was 15 %, suggesting that,
under research conditions, about 30 to 35% of the fertilizer-N applied is typically lost from the system.

10
Detailed research studies provide valuable insides into N pathways and the processes that lead to N losses
in agricultural systems. However, results from research plots cannot be extrapolated to obtain estimates of
NUE at regional or global scales because N losses in farmers’ fields are often much larger. Unfortunately,
little is known about the current level of NUE in key cropping systems of the world at the scale of typical
production fields. This shortage of information reflects the logistical difficulty and high cost of obtaining
direct on-farm measurements and the lack of funding for what appear to be routine on-farm evaluations
(Cassman et al., 2002).

Table 4. Average apparent first-crop recovery efficiency of applied fertilizer-N in cereals (REN = fertilizer-N
recovery in above-ground biomass).

Crop Region Average N rate REN

-1
(no. of observations) (kg N ha ) (%)

Maize, research trials1 World (36) 102 63

Maize, on-farm2 USA (55) 103 37
Rice, research trials1 World (307) 113 44
Rice, on-farm3 Asia (179) 117 31
Wheat, research trials1 World (507) 117 54

Average research trials1 World (850) - 51

1 Ladha et al. (2005)

2 Cassman et al. (2002)
3 Dobermann et al. (2002)

The few available on-farm studies generally suggest a greater disconnection between the amount of
fertilizer N applied by farmers and the crop yield that is achieved, resulting in often low and highly
variable NUE among and within farmers’ fields. Irrigated rice is the only cropping system for which
systematic on-farm measurements of NUE have been conducted for numerous regions in Asia and West
Africa (Cassman et al., 1996b; Dobermann et al., 2002; Haefele et al., 2003). Average REN in irrigated
rice fields in Asia was 31% as compared to 44% in research trials (Table 4). Similarly, whereas Ladha et
al. (2005) cited an average AEN in rice of 21.6 kg kg-1 and average PFPN of 63.2 kg kg-1, measured on-
farm averages in south and southeast Asia were 11.5 kg kg-1 and 49.2 kg kg-1, respectively (Dobermann et
al., 2002). Major conclusions drawn from the on-farm studies with rice were (Olk et al., 1999; Dobermann
et al., 2003; Dobermann et al., 2004):
(i) Large spatial and temporal variability exists among fields with regard to indigenous N
supply, fertilizer use, crop yields, NUE, and marginal return from N fertilizer;
(ii) Grain yield obtained by farmers is closely correlated with plant N uptake, but not
with fertilizer N use;
(iii) NUE varies widely and is often not related to N rates or the supply of N from soil;
(iv) Climate, the supply of other essential nutrients, disease, insect pest, and weed
pressure, stand establishment, water management and N management technology
(timing, forms, placement, etc.) have large effects on REN and PEN and, therefore, the
overall crop response to N fertilizer, and
(v) It is difficult to predict the dynamic N supply from indigenous sources using simple
assessment methods such as soil tests.

11
Extensive on-farm studies of similar kind and nearly global scope have not been conducted in other
environments or for other major cereal crops. This makes it difficult to judge whether the findings made
for rice systems are applicable to other crops and cropping systems. However, there is some evidence that
this may be the case for wheat grown in rice-wheat systems of south Asia and maize grown in rainfed and
irrigated systems of the USA Corn Belt (Adhikari et al., 1999; Cassman et al., 2002). On-farm studies
with maize in the U.S. Corn Belt also showed much lower average REN of 37% (Table 4) than the ‘global’
average of 63% cited for maize in Ladha et al. (2005). A similar discrepancy occurs for PFPN in maize,
with a computed research trial average of PFPN of 69.9 kg kg-1 (Ladha et al., 2005) as opposed to an
average value of 58 kg kg-1 estimated for maize in the USA (Dobermann and Cassman, 2002). The latter
was estimated at national scale based on crop yield statistics and large annual surveys of farmers’ fertilizer
use.
Lower NUE in farmers’ fields is usually explained by a lower level of management under practical
farming conditions and greater spatial variability of factors controlling REN and other indices of NUE
(Cassman et al., 2002). Considering this, NUE achieved in research trials is a good indicator of what can
be targeted with good management, but farm-level NUE is always lower. It is reasonable to assume that,
on a global scale, at least 50% of the fertilizer-N applied is lost from agricultural systems and most of
these losses occur during the year of fertilizer application. It has also been demonstrated, however, how 30
to 50% increases in NUE in rice can be achieved through field-specific management approaches
(Dobermann et al., 2002).

What can be done to increase nitrogen use efficiency?

Is a 20% increase in the incremental yield response to applied N (Fig. 3) achievable at the global scale
over a time period of about 20 years? To answer this question it is important to re-iterate that a 20%
increase in this incremental N efficiency, as estimated by the slope of the regression line in Figure 3, is not
equivalent to a 20% increase in the overall NUE (or PFPN). On a global scale, higher cereal yields are
likely to be achieved through a combination of increased N applications in regions with low N fertilizer
use, such as Africa and parts of Asia and Latin America, and improved N fertilizer efficiency in countries
where current N fertilizer use is already high. For example, the global PFPN in cereals only needs to
increase at a rate of 0.1 to 0.4% yr-1 to meet cereal demand in 2025 (Dobermann and Cassman, 2005).
Such rates have been achieved in some developed regions in the past 20 years (Fig. 2), and far greater
rates of increase have been achieved in several countries.
In the UK, our estimates suggest an average cereal NUE of 36 kg kg-1 in 1981/85, which increased to 44
kg kg-1 by 2001/02 (+23%, 1.1% yr-1). In the USA, annual surveys of cropping practices indicate that
NUE in maize increased from 42 kg kg-1 in 1980 to 57 kg kg-1 in 2000 (+36%, 1.6% yr-1)(Dobermann and
Cassman, 2002). In Japan, NUE of irrigated rice remained unchanged at about 57 kg kg-1 from 1961 to
1985, but it increased to more than 75 kg kg-1 (+32%, 1.8% yr-1) in since 1985 (Suzuki, 1997; Mishima,
2001). In each of these countries, key factors that contributed to this improvement included: (i) increased
yields and more vigorous crop growth associated with greater stress tolerance of modern cultivars, (ii)
improved management of production factors other than N, and (iii) improved N fertilizer management.
The latter may include use of better fertilizers and NUE-enhancing products as well as better application
strategies and methods. The combination of these measures allowed achieving higher yields with either
stagnating (USA) or declining N use (UK, Japan).

12
These improvements were achieved without general restrictions or regulations on N fertilizer use. They
were driven by investments in public and private sector research and extension. Because of the large
differences in NUE among countries, regions, farms, and fields within a farm, policies that focus only on
increasing or decreasing N fertilizer use at a state or national level would have a widely varying impact on
yields, farm profitability, and environmental quality. Instead, achieving greater NUE at state or national
levels will require policies that favor increases in NUE at the field scale with emphasis on technologies
that can achieve greater congruence between crop N demand and N supply from all sources—including
fertilizer, organic inputs, and indigenous soil N (Cassman et al., 2002).
Most of the fertilizer-N is lost during the year of application. Consequently, N and crop management must
be fine-tuned in the cropping season in which N is applied in order to maximize system-level NUE.
Numerous concepts and tools needed to increase NUE have been developed. These technologies can be
divided into (1) those that enhance crop N demand and uptake (genetic improvements, management
factors that remove restrictions on crop growth and N demand) and (2) management options that influence
the availability of soil and fertilizer-N for plant uptake. The latter primarily include more efficient
fertilizers (new N forms, modified fertilizers & inhibitors that lead to slow/controlled release), more
efficient N application methods, and various forms of site-specific N management. It is important to
understand, however, that many of the technology options have different effects on crop yield response to
N and that it is often the combination of measures that leads to the greatest benefit (Fig. 4).

D
Crop yield

C A

0 high
Fertilizer N rate
Figure 4: Generalized changes in crop yield response to fertilizer N application as affected by improvements in
crops and/or crop and fertilizer management (Giller et al., 2004).
A: Average N response function with low to medium fertilizer N efficiency.
B: Shift in the curvature (slope) of the N response function due to increased fertilizer N efficiency. Measures to
achieve this can include improved general crop management (plant density, irrigation, pest control, etc.) or improved
N management technologies (placement, timing, modified fertilizers, inhibitors, etc.).
C: Upward-shifted N response function, i.e., increase in the intercept (yield at zero N rate) but no change in the
curvature because there is no increase in fertilizer-N efficiency. An increase in the 0-N yield may be due to an
improved variety with greater N acquisition or greater internal N utilization, amelioration of constraints that
restricted uptake of indigenous N, or other measures that increase the indigenous N supply.
D: Shift in the intercept and curvature of the N response function, i.e., increase in both 0-N yield and slope through a
combination of measures. Full exploitation of yield potential is achieved by implementation of a site-specific,
integrated crop management approach, in which an advanced genotype is grown with near-perfect management,
closely matching crop N demand and supply. As a result, both profit and fertilizer N use efficiency are highest.

13
It is beyond the scope of this paper to discuss specific technologies in more detail and the reader is
referred to the recent literature on this (Schroeder et al., 2000; Cassman et al., 2002; Dobermann and
Cassman, 2004; Giller et al., 2004; Ladha et al., 2005). Modern N management concepts usually involve a
combination of anticipatory (before planting) and responsive (during the growing season) decisions.
Improved synchrony, for example, can be achieved by more accurate N prescriptions based on the
projected crop N demand and the levels of mineral and organic soil N, but also through improved rules for
splitting of N applications according to phenological stages, by using decision aids to diagnose soil and
plant N status during the growing season (models, sensors), or by using controlled-release fertilizers or
inhibitors. The latter have a theoretical advantage over other, more knowledge-intensive forms of fined-
tuned N management in a sense that the knowledge is ‘embedded’ in the product to be applied. As
experience with seeds shows, embedded knowledge can lead to high adoption rates by farmers, provided
that the benefit : cost ratio is high.
Important prerequisites for the adoption of advanced N management technologies are that they must be
simple, provide consistent and large enough gains in NUE, involve little extra time and be cost-effective
(Giller et al., 2004). If a new technology leads to at least a small and consistent increase in crop yield with
the same amount or less N applied, the resulting increase in profit is usually attractive enough for a
farmer. This is particularly relevant for developing countries or large-scale grain farms in North and South
America or in Australia, where there is still potential and need to produce more food and feed. Where
yield increases are more difficult to achieve, where increasing crop yield is of less priority, or where
reducing the creation of Nr in agriculture is the top societal priority, adoption of new technologies that
increase NUE but have little effect on farm profit may need to be supported by appropriate technology
incentives.

Summary
Quantifying the status of NUE in agriculture is a difficult task because (i) definitions used in research
papers and interpretation of different NUE indices vary and (ii) reliable data needed to compute NUE
indices are often not available, particularly at national, regional and global scales. Worldwide, crops do
not directly utilize about half of the applied N and the overall NUE has declined with increasing N
fertilizer use. This trend seems to continue in many developing countries. In many industrialized countries
NUE has been increased, even at high levels of cropping intensity and fertilizer use. Interventions to
increase NUE and reduce N losses to the environment must be accomplished at the farm level through a
combination of improved technologies and carefully crafted local policies that promote the adoption of
improved N management practices while sustaining yield increases. Improved fertilizer products play an
important role in the global quest for increasing NUE, but their relative importance varies by regions and
cropping systems.

14
References
Adhikari, C., K.F. Bronson, G.M. Panaullah, A.P. Regmi, P.K. Saha, A. Dobermann, D.C. Olk, P. Hobbs, and E.
Pasuquin. 1999. On-farm soil N supply and N nutrition in the rice-wheat system of Nepal and Bangladesh.
Field Crops Res. 64:273-286.
Bell, M.A. 1993. Organic matter, soil properties, and wheat production in the high valley of Mexico. Soil Sci.
156:86-93.
Boyer, E.W., R.W. Howarth, J.N. Galloway, F.J. Dentener, C. Cleveland, G.P. Asner, P. Green, and C. Vörösmarty.
2004. Current nitrogen inputs to world regions. p. 221-230. In A.R. Mosier et al. (ed.) Agriculture and the
nitrogen cycle: assessing the impacts of fertilizer use on food production and the environment. SCOPE 65.
Island Press, Washington,D.C.
Cassman, K.G., A. Dobermann, P.C. Sta.Cruz, H.C. Gines, M.I. Samson, J.P. Descalsota, J.M. Alcantara, M.A.
Dizon, and D.C. Olk. 1996a. Soil organic matter and the indigenous nitrogen supply of intensive irrigated rice
systems in the tropics. Plant Soil 182:267-278.
Cassman, K.G., A. Dobermann, and D.T. Walters. 2002. Agroecosystems, nitrogen-use efficiency, and nitrogen
management. Ambio 31:132-140.
Cassman, K.G., A. Dobermann, D.T. Walters, and H.S. Yang. 2003. Meeting cereal demand while protecting natural
resources and improving environmental quality. Annu. Rev. Environ. Resour. 28:315-358.
Cassman, K.G., H.C. Gines, M. Dizon, M.I. Samson, and J.M. Alcantara. 1996b. Nitrogen-use efficiency in tropical
lowland rice systems: Contributions from indigenous and applied nitrogen. Field Crops Res. 47:1-12.
Dobermann, A., and K.G. Cassman. 2002. Plant nutrient management for enhanced productivity in intensive grain
production systems of the United States and Asia. Plant Soil 247:153-175.
Dobermann, A., and K.G. Cassman. 2004. Environmental dimensions of fertilizer N: What can be done to increase
nitrogen use efficiency and ensure global food security? p. 261-278. In A.R. Mosier et al. (ed.) Agriculture
and the nitrogen cycle: assessing the impacts of fertilizer use on food production and the environment.
SCOPE 65. Island Press, Washington,D.C.
Dobermann, A., and K.G. Cassman. 2005. Cereal area, yield and nitrogen use efficiency are drivers for future
nitrogen fertilizer consumption. Science in China (in press).
Dobermann, A., C. Witt, S. Abdulrachman, H.C. Gines, R. Nagarajan, T.T. Son, P.S. Tan, G.H. Wang, N.V. Chien,
V.T.K. Thoa, C.V. Phung, P. Stalin, P. Muthukrishnan, V. Ravi, M. Babu, G.C. Simbahan, M.A.A. Adviento,
and V. Bartolome. 2003. Estimating indigenous nutrient supplies for site-specific nutrient management in
irrigated rice. Agron. J. 95:924-935.
Dobermann, A., C. Witt, and D. Dawe. 2004. Increasing productivity of intensive rice systems through site-specific
nutrient management. Science Publishers, Inc., International Rice Research Institute, Enfield, NH (USA) and
Los Baños (Philippines).
Dobermann, A., C. Witt, D. Dawe, G.C. Gines, R. Nagarajan, S. Satawathananont, T.T. Son, P.S. Tan, G.H. Wang,
N.V. Chien, V.T.K. Thoa, C.V. Phung, P. Stalin, P. Muthukrishnan, V. Ravi, M. Babu, S. Chatuporn, M.
Kongchum, Q. Sun, R. Fu, G.C. Simbahan, and M.A.A. Adviento. 2002. Site-specific nutrient management
for intensive rice cropping systems in Asia. Field Crops Res. 74:37-66.
FAO. 2004. FAOSTAT Database--Agricultural Production. http://apps.fao.org. Food and Agriculture Organization
of the United Nations, Rome.
Galloway, J.N., W.H. Schlesinger, H. Levy, A. Michaels, and J.L. Schnoor. 1995. Nitrogen fixation: atmospheric
enhancement - environmental response. Global Biogeochemical Cycles 9:235-252.
Giller, K.E., P.M. Chalk, A. Dobermann, L.C. Hammond, P. Heffer, J.K. Ladha, P. Nyamudeza, L.M. Maene, H.
Ssali, and J.R. Freney. 2004. Emerging technologies to increase the efficiency of use of fertilizer nitrogen. p.
35-52. In A.R. Mosier et al. (ed.) Agriculture and the nitrogen cycle: assessing the impacts of fertilizer use on
food production and the environment. SCOPE 65. Island Press, Washington,D.C.
Haefele, S.M., M.C.S. Wopereis, C. Donovan, and J. Maubuisson. 2001. Improving the productivity and
profitability of irrigated rice production in Mauritania. Eur. J. Agron. 14:181-196.

15
Haefele, S.M., M.C.S. Wopereis, M.K. Ndiaye, S.E. Barro, and M. Ould Isselmo. 2003. Internal nutrient
efficiencies, fertilizer recovery rates and indigenous nutrient supply of irrigated lowland rice in Sahelian West
Africa. Field Crops Res. 80:19-32.
IAEA. 2003. Management of crop residues for sustainable crop production. IAEA TECHNDOC-1354. International
Atomic Energy Agency, Vienna.
IFA. 2002. Fertilizer use by crop. 5 ed. International Fertilizer Industry Association (IFA), International Fertilizer
Development Centre (IFDC), International Potash Institute (IPI), Potash and Phosphate Institute (PPI), Food
and Agriculture Organization (FAO), Rome.
Kolberg, R.L., D.G. Westfall, and G.A. Peterson. 1999. Influence of cropping intensity and nitrogen fertilizer rates
on in situ nitrogen mineralization. Soil Sci. Soc. Am. J. 63:129-134.
Krupnik, T.J., J. Six, J.K. Ladha, M.J. Paine, and C. van Kessel. 2004. An assessment of fertilizer nitrogen recovery
efficiency by grain crops. p. 193-207. In A.R. Mosier et al. (ed.) Agriculture and nitrogen cycle: assessing the
impact of fertilizer use on food production and the environment. SCOPE, Paris.
Ladha, J.K., H. Pathak, T.J. Krupnik, J. Six, and C. van Kessel. 2005. Efficiency of fertilizer nitrogen in cereal
production: retrospects and prospects. Adv. Agronomy (in press).
Mishima, S. 2001. Recent trend of nitrogen flow associated with agricultural production in Japan. Soil Sci. Plant
Nutr. 47:157-166.
Novoa, R., and R.S. Loomis. 1981. Nitrogen and plant production. Plant Soil 58:177-204.
Olk, D.C., K.G. Cassman, G.C. Simbahan, P.C. Sta.Cruz, S. Abdulrachman, R. Nagarajan, P.S. Tan, and S.
Satawathananont. 1999. Interpreting fertilizer-use efficiency in relation to soil nutrient-supplying capacity,
factor productivity, and agronomic efficiency. Nutr. Cycling Agroecosyst. 53:35-41.
Pretty, J., C. Brett, D. Gee, R.E. Hine, C.F. Mason, J.I.L. Morison, H. Raven, M.D. Rayment, and G. van der Bijl.
2000. An assessment of the total external costs of UK agriculture. Agric. Syst. 65:113-136.
Schroeder, J.J., J.J. Neeteson, O. Oenema, and P.C. Struik. 2000. Does the crop or the soil indicate how to save
nitrogen in maize production? Reviewing the state of the art. Field Crops Res. 66:151-164.
Schweigert, P., and R.R. van der Ploeg. 2000. Nitrogen use efficiency in German agriculture since 1950: facts and
evaluation. Berichte über Landwirtschaft 80:185-212.
Smil, V. 1999. Nitrogen in crop production: An account of global flows. Global Biogeochemical Cycles 13:647-662.
Smil, V. 2001. Enriching the earth: Fritz Haber, Carl Bosch, and the transformation of world food production. The
MIT Press, Cambridge, MS, London.
Suzuki, A. 1997. Fertilization of rice in Japan. Japan FAO Association, Tokyo, Japan.
Townsend, A.R., R.B. Howarth, F.A. Bazzaz, M.S. Booth, C.C. Cleveland, S.K. Collinge, A.P. Dobson, P.R.
Epstein, E.A. Holland, D.R. Keeney, M.A. Mallin, C.A. Rogers, P. Wayne, and A.H. Wolfe. 2003. Human
health effects of a changing global nitrogen cycle. Frontiers Ecol. Environ. 1:240-246.
Wopereis, M.C.S., C. Donovan, B. Nebie, D. Guindo, and M.K. N'Diaye. 1999. Soil fertility management in
irrigated rice systems in the Sahel and Savanna regions of West Africa. Part I. Agronomic analysis. Field
Crops Res. 61:125-145.

16