Reducing environmental risk by improving N management in intensive Chinese agricultural systems.
ABSTRACT Excessive N fertilization in intensive agricultural areas of China has resulted in serious environmental problems because of atmospheric, soil, and water enrichment with reactive N of agricultural origin. This study examines grain yields and N loss pathways using a synthetic approach in 2 of the most intensive double-cropping systems in China: waterlogged rice/upland wheat in the Taihu region of east China versus irrigated wheat/rainfed maize on the North China Plain. When compared with knowledge-based optimum N fertilization with 30-60% N savings, we found that current agricultural N practices with 550-600 kg of N per hectare fertilizer annually do not significantly increase crop yields but do lead to about 2 times larger N losses to the environment. The higher N loss rates and lower N retention rates indicate little utilization of residual N by the succeeding crop in rice/wheat systems in comparison with wheat/maize systems. Periodic waterlogging of upland systems caused large N losses by denitrification in the Taihu region. Calcareous soils and concentrated summer rainfall resulted in ammonia volatilization (19% for wheat and 24% for maize) and nitrate leaching being the main N loss pathways in wheat/maize systems. More than 2-fold increases in atmospheric deposition and irrigation water N reflect heavy air and water pollution and these have become important N sources to agricultural ecosystems. A better N balance can be achieved without sacrificing crop yields but significantly reducing environmental risk by adopting optimum N fertilization techniques, controlling the primary N loss pathways, and improving the performance of the agricultural Extension Service.
- SourceAvailable from: Fadong Li[Show abstract] [Hide abstract]
ABSTRACT: Irrigation projects have diverted water from the lower reaches of the Yellow River in China for more than 50 years and are unique in the world. This study investigated the effect of irrigation practices on the transfer and regional migration mechanisms of nitrate (NO3(-)) in surface water and groundwater in a Yellow River alluvial fan. Hydrochemical indices (EC, pH, Na(+), K(+), Mg(2+), Ca(2+), Cl(-), SO4(2-), and HCO(3-)) and stable isotopic composition (δ18O and δD) were determined for samples. Correlation analysis and principal component analysis (PCA) were performed to identify the sources of water constituents. Kriging was employed to simulate the spatial diffusion of NO3(-) and stable isotopes. Our results demonstrated that the groundwater exhibited more complex saline conditions than the surface water, likely resulting from alkaline conditions and lixiviation. NO3(-) was detected in all samples, 87.0% of which were influenced by anthropogenic activity. The NO3(-) pollution in groundwater was more serious than the common groundwater irrigation areas in the North China Plain (NCP), and was also slightly higher than that in surface water in the study area, but this was not statistically significant (p > 0.05). In addition, the groundwater sites with higher NO3(-) concentrations did not overlap with the spatial distribution of fertilizer consumption, especially in the central and western parts of the study area. NO3(-) distributions along the hydrogeological cross-sections were related to the groundwater flow system. Hydrochemical and environmental isotopic evidences indicate that surface water-groundwater interactions influence the spatial distribution of NO3(-) in the Piedmont of South Taihang Mountains.Environmental science. Processes & impacts. 10/2014;
- [Show abstract] [Hide abstract]
ABSTRACT: China's pig production has increased manifold in the past 50 years, and this has greatly affected the nitrogen (N) and phosphorus (P) use and losses in the pig production sector. However, the magnitude of these changes are not well-known. Here, we provide an in-depth account of the changes in pig production, and N and P use, and total N and P losses in the whole pig production chain during the period 1960-2010, through simulation modelling and using data from national statistics and farm survey. For the period 2010-2030, we explored possible effects of technological and managerial measures aimed at improving the performances of pig production via scenario analysis. We used and further developed the NUFER model to calculate the feed requirement and consumption, and N and P losses in different pig production systems for all the years. Between 1960 and 2010, pig production has largely shifted from the so-called backyard system to landless systems. The N use efficiencies at fattener level increased from 18 to 28%, due to the increased animal productivity. However, the N use efficiencies at the whole-system level decreased from 46 to 11% during this period, mainly due to the increase of land-less pig farms, which rely on imported feed and have no land-base for manure disposal. The total N and P losses were 5289 and 829 Gg in 2010, which is 30 and 95 times higher than in 1960. In the business as usual scenario, the total N and P losses were projected to increase by 25 and 55% between 2010 and 2030, respectively. Analyses of other scenarios indicate that packages of technological and managerial measures can decrease total N and P losses by 64 and 95%, respectively. Such improvements require major transition in the pig production sector, notably in of manure management, herd management and feeding practices.Environmental science & technology. 10/2014;
- [Show abstract] [Hide abstract]
ABSTRACT: Increased fertilizer input in agricultural systems during the last few decades has resulted in large yield increases, but also in environmental problems. We used data from published papers and a soil testing and fertilization project in Shaanxi province during the years 2005 to 2009 to analyze chemical fertilizer inputs and yields of wheat (Triticum aestivum L.) and maize (Zea mays L.) on the farmers' level, and soil fertility change from the 1970s to the 2000s in the Loess Plateau in China. The results showed that in different regions of the province, chemical fertilizer NPK inputs and yields of wheat and maize increased. With regard to soil nutrient balance, N and P gradually changed from deficit to surplus levels, while K deficiency became more severe. In addition, soil organic matter, total nitrogen, alkali-hydrolysis nitrogen, available phosphorus and available potassium increased during the same period. The PFP of N, NP and NPK on wheat and maize all decreased from the 1970s to the 2000s as a whole. With the increase in N fertilizer inputs, both soil total nitrogen and alkali-hydrolysis nitrogen increased; P fertilizer increased soil available phosphorus and K fertilizer increased soil available potassium. At the same time, soil organic matter, total nitrogen, alkali-hydrolysis nitrogen, available phosphorus and available potassium all had positive impacts on crop yields. In order to promote food safety and environmental protection, fertilizer requirements should be assessed at the farmers' level. In many cases, farmers should be encouraged to reduce nitrogen and phosphate fertilizer inputs significantly, but increase potassium fertilizer and organic manure on cereal crops as a whole.PLoS ONE 01/2014; 9(11):e112273. · 3.53 Impact Factor
Reducing environmental risk by improving
N management in intensive Chinese
Xiao-Tang Jua,1, Guang-Xi Xingb, Xin-Ping Chena, Shao-Lin Zhangb, Li-Juan Zhangc, Xue-Jun Liua, Zhen-Ling Cuia,
Bin Yinb, Peter Christiea,d, Zhao-Liang Zhub, and Fu-Suo Zhanga,1
aKey Laboratory of Plant and Soil Interactions, Ministry of Education, China, and College of Resources and Environmental Sciences, China Agricultural
University, Beijing 100193, China;bState Key Laboratory of Soil and Sustainable Agriculture, Institute of Soil Science, Chinese Academy of Sciences,
Nanjing 210008, China;cCollege of Agricultural Resources and Environmental Sciences, Hebei Agricultural University, Baoding 071001, China; and
dAgri-Environment Branch, Agri-Food and Biosciences Institute, Belfast BT9 5PX, United Kingdom
Communicated by G. David Tilman, University of Minnesota, St. Paul, MN, January 3, 2009 (received for review July 3, 2008)
Excessive N fertilization in intensive agricultural areas of China has
resulted in serious environmental problems because of atmospheric,
soil, and water enrichment with reactive N of agricultural origin. This
study examines grain yields and N loss pathways using a synthetic
waterlogged rice/upland wheat in the Taihu region of east China
compared with knowledge-based optimum N fertilization with 30–
60% N savings, we found that current agricultural N practices with
550–600 kg of N per hectare fertilizer annually do not significantly
environment. The higher N loss rates and lower N retention rates
indicate little utilization of residual N by the succeeding crop in
rice/wheat systems in comparison with wheat/maize systems. Peri-
odic waterlogging of upland systems caused large N losses by deni-
trification in the Taihu region. Calcareous soils and concentrated
summer rainfall resulted in ammonia volatilization (19% for wheat
and 24% for maize) and nitrate leaching being the main N loss
pathways in wheat/maize systems. More than 2-fold increases in
atmospheric deposition and irrigation water N reflect heavy air and
water pollution and these have become important N sources to
agricultural ecosystems. A better N balance can be achieved without
adopting optimum N fertilization techniques, controlling the primary
N loss pathways, and improving the performance of the agricultural
intensive agriculture ? synthetic N fertilizer ? denitrification ?
nitrate leaching ? N deposition
of the world was controlled by natural processes, but the expanded
production of synthetic N fertilizer and release of N from fossil fuel
combustion now matches the natural rate of formation of reactive
with half of the synthetic N fertilizer ever used having been applied
during the last 15 to 20 years (3). Furthermore, because of the
difficulty in accurately predicting N fertilizer requirements, rates
exceeding plant requirements are often applied, thus inducing
and emission of nitrous oxide and ammonia. This has become a
major concern for scientists, environmental groups, and agricul-
tural policymakers worldwide.
From 1977 to 2005, total annual grain production in China
increased from 283 to 484 million tons (a 71% increase) and the
average grain production per unit area increased from 2,348 to
4,642 (a 98% increase). However, synthetic N fertilizer application
same period. This resulted in a partial factor productivity from
applied N (PEPN) that decreased from 55 to 20 kg/kg (4, 5). Large
he last 40 years have seen an extraordinary change in the global
(6, 7). For example, the annual application rate of synthetic N for
conventional agricultural practices in east and southeast China
as well as the North China Plain now ranges from 550 to 600 kg
of N per hectare for typical double-cropping systems (5, 8).
The Taihu region and the North China Plain are 2 of the most
intensive agricultural regions in China and the most economically
developed areas (5, 9). Large inputs of synthetic N fertilizer, rapid
development of intensive livestock production systems, and rapidly
increasing consumption of fossil fuels have severely disturbed
regional biogeochemical N cycling and led to a series of environ-
pollution of groundwater, acid rain and soil acidification, green-
house gas emissions, and other forms of air pollution. There have
also been effects on human health and normal functioning of
ecosystems (9, 10), details of which are given in SI Text.
Here, we provide the actual grain yields and total N losses
associated with knowledge-based optimum fertilization strategies
compared with experience-based N management practices from
numerous on-farm field experiments. The ‘knowledge-based opti-
field-based academic research results or soil tests on quantity and
timing of synthetic N fertilizer practically, using regional mean
optimal N (RMON) application rate in rice/wheat systems and
in-season N management based on the soil mineral N test in
wheat/maize systems (5, 8, 12, 13). We use an integrated approach
to measure the fate of15N fertilizer and to document the major loss
pathways in situ rather than by using individual measurements of
loss processes reported in other studies to understand N behavior
in these 2 rotations and thus attempt to control the environmental
practiced in these intensive management systems and suggest
strategies for balancing N management by applying knowledge-
based optimum N fertilization techniques.
This study examines double-crop rotations, which are 2 of the
most intensive agricultural systems worldwide and are widely
practiced in Asian countries. The frequent alternation between
flooding and draining the rice/wheat rotations led to significant
changes in soil N transformations as compared with upland crop
Author contributions: X.-T.J., X.L.-Z., and F.-S.Z. designed research; X.-T.J., G.-X.X., X.-P.C.,
and X.-T.J. and P.C. wrote the paper.
The authors declare no conflict of interest.
Freely available online through the PNAS open access option.
1To whom correspondence may be addressed. E-mail: firstname.lastname@example.org or zhangfs@
This article contains supporting information online at www.pnas.org/cgi/content/full/
March 3, 2009 ?
vol. 106 ?
no. 9 ?
opportunity to study the effects of N on environmental degrada-
of these two intensive agricultural regions may also provide options
for more rational N management practices in other intensive
management cropping systems around the world.
Results and Discussion
We developed knowledge-based optimum N management tech-
in north China during the last decade. The regional mean optimal
N (RMON) application rate was calculated from the average of
economically optimum N rates (i.e., the point of the N rate where
the marginal grain production value is equal to the marginal N
fertilization cost) based on large numbers of on-farm field exper-
iments in the rice-based rotations and in-season N management
based on the soil mineral N (Nmin: NH4
previously developed for upland crop rotations. In principle,
fine tuning according to specific field conditions, taking into
consideration the relative uniformity of climate, soil N fertility, and
12). RMON gives the mean of economically optimum N rates and
can be used as a reference for extension technicians. This recom-
mendation method can be easily adapted to rural areas of China
index is high. The short time interval between harvest of the first
crop and sowing of the second makes soil testing impractical.
The improved Nmin method is based on synchronization of crop
N demand and soil Nmin supply, and is considered to be a soil N
supply index because of the high amounts of Nmin (mostly nitrate)
frequently found in the root zone and its high availability for
subsequent crop growth in the next phase of upland rotations. We
found that high potential rates of mineralization and nitrification
contribute to this high accumulation of nitrate in soils of the North
China Plain (14). This method has been well established in wheat/
maize rotations in recent years (8, 13, 15, 16) and can be easily
to perform in the laboratory and may even be feasible in the field
(17). In the present study, we also evaluated the agronomic per-
formance and potential environmental risk of these knowledge-
based strategies in comparison with the farmer’s experience-based
?-N ? NO3
N application rate in Taihu region and the improved Nmin method
on the North China Plain appeared to give slight increases in crop
no significant differences (P ? 0.05) were found, indicating that
because of increased soil indigenous N supply (18, 19). In contrast,
total N losses (including NH3 volatilization, denitrification and
leaching from the top 1 m of the soil profile) increased significantly
with increasing N inputs, indicating high environmental costs were
caused by exceeding optimum N fertilizer rates.
We used15N in some of the field experiments (Field Study 2) to
among the four crops. We found that N recovery rates decreased
with increasing N application rate for all of them (Fig. 1A). N rates
showed polynomial, linear, polynomial, and logarithmic relation-
ships with N recovery by rice, wheat (south), wheat (north) and
(P ? 0.01). At lower N input levels (?125 kg of N per hectare), the
N input levels (?125 kg of N per hectare) the order was: wheat
(north) ? rice ? maize ? wheat (south), indicating that rice
with increasing N application rate in all four crops (Fig. 1B). A
(south) in the Taihu region showed much higher N loss rates than
did wheat (north) and maize on the North China Plain. The trends
of the N retention rates with increasing N rate differed among the
four crops. The rice and wheat systems in the Taihu region showed
China Plain. Moreover, N retention rates in wheat (south) and
wheat (north) decreased first and then increased with further
increase in N rate, showing typical polynomial relationships (P ?
0.01). However, N retention rates in rice and maize decreased by
power function relationships when N rate increased (P ? 0.01).
These findings indicate that the upland wheat and maize rotation
system has significantly higher N recovery rates, higher N retention
rates, and lower N loss rates than the waterlogged rice and upland
farmers’ management practices. Moreover, on the North China
Plain wheat has significantly higher N recovery rates, higher N
retention rates, and lower N loss rates than maize, and in the Taihu
region, rice has significantly higher N recovery rates and lower N
loss rates than wheat. Rice/wheat rotations show high ‘leakage’ of
external N inputs with resulting pollution threats to water and air.
However, wheat/maize rotations show strong retention of external
N inputs that can remain available for use by subsequent crops.
Table 1. Average grain yields and total N losses of the optimum N fertilization (ON) compared with farmers’ N practices (FN) (Field
Study 1 and 2).
Crop and site of field experimentN fertilization
N rateGrain yieldTotal fertilizer N loss*
Rate, kg of
N per hectare
Ratio of FN
Ratio of FN
Total loss, kg of
N per hectare
Ratio of FN
Rice in Taihu (n ? 26) ON†
1.5 0.97 1.7
Wheat in Taihu (n ? 9)
Wheat in NCP‡(n ? 121)
Maize in NCP‡( n ? 148)
*Total fertilizer N losses calculated with the models of Fig. 1B simulated from15N field experiments.
†Regional mean optimal N application rate calculated from the mean of economically optimum N rates of field experiments in Taihu regain (5, 12).
‡Data including Field Study 1 and also summarized from ref. (15, 16); NCP, North China Plain.
§In-season nitrogen management based on soil Nmin test on the NCP (8, 13, 15, 16).
www.pnas.org?cgi?doi?10.1073?pnas.0813417106 Ju et al.
Retention of N is favorable for long-term improvement of soil N
fertility and further emphasizes the need to reduce N application
rates. It should be noted that N rates applied to wheat in the Taihu
region are similarly high to those on the North China Plain, but
lower N recovery rates (Table 1). One explanation is that farmers
in the Taihu region try to compensate for the unfavorable growing
due to high rainfall and groundwater) by increasing N fertilizer
application rates (20).
To further understand the different pathways of N loss from
field experiment (Field Study 3) on the North China Plain, mea-
suring in situ NH3volatilization, denitrification, N leaching, and
N2O emissions. We found that these N loss pathways differed
and management practices (Table 2). Ammonia volatilization
seems to be the main loss pathway for fertilizer N, with losses
accounting for 19.4 and 24.7% of applied N in wheat (north) and
maize seasons, respectively. Volatilization was much lower for rice
and wheat (south), with losses accounting for 11.6 and 2.1% of
applied N, respectively. Denitrification losses were quite small in
the wheat and maize on the North China Plain (0.1 and 3.3% of
fertilizer N loss in rice and wheat within the Taihu region, account-
ing for 36.4 and 43.5% of applied N, respectively. Leaching ac-
counted for only 0.3% of applied N from rice and 3.4% from wheat
(south) season, but accounted for 2.7 and 12.1% of applied N for
wheat and maize growing seasons on the North China Plain. We
may be overestimating the denitrification loss in rice/wheat systems
due to error accumulation with the different methods (5), but the
order of magnitude agrees with previous studies (5).
Alternating upland wheat and waterlogged rice led to high
denitrification of applied N and accumulated nitrate (after wheat
harvest) in the rice season (20–22). The traditional practice of
surface application of urea-N stimulated NH3volatilization from
rice (23, 24). After plowing of paddy soil, a hard layer that formed
below the plow layer due to clay deposition appeared to reduce
leakage from the rice crop, but resulted in higher leaching losses
during the wheat season when drainage prevails (25). The high
denitrification loss for wheat (south) is attributed to wet soil
conditions, higher temperatures and relatively high soil carbon (C)
in the soil, consequently becoming available for loss through
denitrification during the rice season (22). Thus the residual
crop as is the case in upland crop rotations. We speculate that the
very large N losses (corresponding with lower N retention rates)
and large C inputs in straw under the anaerobic conditions in
rice/wheat rotations may be another explanation for the structural
changes in soil organic matter and increased C-rich humic acids
found by Schmidt-Rohr et al. (28). On the North China Plain, with
its calcareous soils, typical pH 7.5 to 8.5 and predominant use of
?-N in the soil profile is not fully available to the subsequent
(C) with N application rate in 4 crops (Field Study 2). Vertical bars denote
standard deviation of the mean (average of 615N field experiments). Each
model fitting produced a highly significant model (P ? 0.01).
Relationships of N recovery rate (A), loss rate (B), and retention rate
Table 2. Different N loss pathways expressed as a percentage (mean ? SD) of N application rate in farmers’ N practices (Field Study
3, Lysimeter Study)
Taihu region North China Plain
Rice Wheat-southWheat-north Maize
N rate (kg of N per hectare)
Recovery rate (%)*
Retention rate (%)*
29.6 ? 4.9
21.7 ? 5.1
11.6 ? 4.7
0.3 ? 0.5
18.4 ? 6.3
28.5 ? 4.6
2.1 ? 1.4
3.4 ? 2.1
31.0 ? 3.6
45.7 ? 5.4
19.4 ? 5.2
2.7 ? 2.6
0.1 ? 0.04
25.5 ? 5.2
33.9 ? 2.3
24.7 ? 5.6
12.1 ? 8.5
3.3 ? 1.6
Leaching out of 1 m soil depth (%)
*Measured from corresponding15N field experiments.
†Calculated by difference method.
Ju et al.
March 3, 2009 ?
vol. 106 ?
no. 9 ?
urea or ammonium bicarbonate as fertilizer (accounting for 80%),
NH3volatilization is generally considered to be the major pathway
of N loss, especially in the hot maize growing season (29–31). Due
experimental results on the North China Plain have shown the
magnitude of N loss by denitrification is no more than 10 kg of N
season (30, 32). We found that nitrate accumulation in the soil
profile and gradual leaching below the root zone were the impor-
tant N loss pathways in wheat and maize rotations under conven-
tional agricultural N management practices due to concentrated
rainfall in summer and flood irrigation (8, 33), contradicted con-
ventional thinking that leaching losses are not an important path-
way for N loss in semi-humid upland agricultural systems of the
North China Plain.
Field Study 1 and the Lysimeter Study also showed that NH3
volatilization increased rapidly and represented the major pathway
for loss in response to increasing N rate, but N2O emission and
and accounted for most of the reactive N loss with increasing N
fertilizer rate, while N2O emissions and NH3 volatilization in-
creased very little in response to N rate during the wheat season in
the Taihu region (Fig. S1, wheat-south). Ammonia volatilization
and NO3-N leaching increased significantly with increasing N rate
and N2O emissions also showed an upward trend in the wheat
growing season on the North China Plain (Fig. S1, wheat-north).
During the maize growing season, NH3 volatilization, NO3-N
leaching, and N2O emissions all increased significantly, with N
leaching showing a particularly strong increase (Fig. S1, maize).
These results confirm that emissions of reactive N to the environ-
ment will increase after N fertilization exceeds the optimum rate.
However, we did not find large N2O emissions among the four
crops. The total N2O emissions in rice, wheat (south), wheat
(north), and maize with different N application rates were 0.13 to
0.50, 0.27 to 0.63, 0.13 to 0.24, and 0.16 to 0.65 kg of N per hectare,
respectively. Expressed as average emission factors, this accounted
for 0.20, 0.30, 0.12, and 0.23%, respectively, close the IPCC 2006
guideline of 0.3% for paddy rice soil and much lower than the
by the low ratio of N2O to N2O?N2from denitrification products
(36, 37) under waterlogged rice conditions, low available carbon
sources and low water-filled pore space (WFPS) for denitrification
in most semi-humid upland soils on the North China Plain (14).
To quantify atmospheric and irrigation water contributions
(hereafter ‘environmental N inputs’) to agricultural systems in the
Taihu region and the North China Plain, we measured deposition
and N inputs from irrigation water (Monitoring Study). Total
Plain. Comparisons with observations made from both regions in
the 1980s shows more than 2-fold greater environmental N inputs
and irrigation water N inputs from 15 to 56 kg of N per hectare in
Taihu region (Fig. S2) (38). During the same period, N deposition
increased significantly from 21 to 89 kg of N per hectare and
irrigation water N inputs from 8 to 15 kg of N per hectare (Fig. S2)
(38, 39) on the North China Plain. Inorganic N and soluble organic
N are the primary environmental N sources that can be used
directly by crops or after transformation in the soil (40). The
absence of crop yield response to N application in most field
to increased indigenous soil N supply, but also to high environ-
mental N inputs (41, 42). Moreover, high N deposition may also
contribute to surface water eutrophication, acid rain, and soil
acidification (43) in these regions. High total N concentrations (7
to 8 mg?N?L?1) in the irrigation surface water in the Taihu region
and high NO3?-N concentrations in the irrigation shallow ground-
water on the North China Plain also reflect the heavy N pollution
of water resources in both areas.
We calculated annual N balance for both rotations using two
scenarios: conventional N practice versus optimum N fertilization
(Tables S1 and S2). We computed an annual N surplus of 87 kg of
N per hectare for current practices with large losses by denitrifi-
cation for the rice/wheat system. Synthetic N fertilizer was the
primary input followed by N in irrigation water, biological N2
fixation, and N deposition. A better N balance can be achieved by
adopting optimum N fertilization strategies designed to maintain
relatively high yields but reduce environmental risk. However,
dinitrification could be further reduced by improving carbon man-
agement and controlling the water regime (25–27). We also com-
puted a 212 kg of N per hectare surplus for current practice with
large losses by NH3volatilization within the wheat/maize system. A
large proportion of surplus N accumulated as nitrate in the soil
profile after harvest (33), and partly existed as N in organic form
due to manure application. However, the unusually large quantity
of annual surplus N might also be caused by underestimating the
leaching loss due to drought conditions in our observation years
strong N leaching losses only occurred in some years with heavy
summer rainfall (33), leading to high nitrate accumulation in the
deep subsoil (8, 33) and the groundwater (43). In the future, a
slightly negative N balance could be achieved using an optimum N
rate of about 286 kg of N per hectare to maintain relatively high
yields (15, 16). The slightly negative balance would be conducive to
making sure the plants fully use accumulated nitrate and further
to maintenance of soil N fertility. Changing the N application
techniques and using deep placement of urea or ammonium
bicarbonate could substantially reduce NH3 volatilization losses
from calcareous soils (29).
A question that needs to be addressed is why farmers in China
1990s, scientists, government and extension staff encouraged farm-
ers to increase synthetic fertilizer inputs to increase yields and feed
an increasing population. As a result, synthetic fertilizer inputs
increased continually throughout the country, especially in the
Taihu region and North China Plain. By 2000, the rates were far
greater than crop demand and serious environmental degradation
had begun. Although this trend has been recognized by the
scientific community since the late 1990s, on-farm practices are
difficult to reverse after 10 to 15 years of effort (43). Persuading
farmers to limit fertilizer inputs is difficult because many of them
still hold to now traditional opinions that higher crop yield will be
include an extra ‘insurance’ component to prevent yield loss rather
than matching inputs to crop demand. The high off-farm incomes
and relatively low retail prices for N fertilizers (with government
subsidies for production and transportation) compared with U.S.
and European prices are also important factors (43, 44). A third
major reason is the poor infrastructure of the Extension Services
and poor transfer of knowledge to farmers (43). A critical objective
of the optimum N techniques developed using our studies are to
substantially lower fertilizer N application rates by accounting for
indigenous soil N and environmental N inputs for maintenance of
the yields needed to feed an increasing population. This strategy
also limits losses of total N to the water and atmosphere with the
aim of establishing sustainable agricultural systems (8, 11, 12).
too heavily on synthetic N fertilizer inputs and cereal monocrop-
ping, especially in Asian countries under pressure to feed large and
growing populations. It is now time to change this situation by
balancing yield and environmental consequences. Integrated man-
that include efficient recycling of manures and crop residues, the
www.pnas.org?cgi?doi?10.1073?pnas.0813417106Ju et al.
use of legume crops in rotations to increase internal N cycling and
further reduction in the reliance on synthetic N fertilizers (45).
It took nearly 50 years to achieve food sufficiency in China. An
unanticipated cost has been that massive fertilizer inputs have led
to significant environmental degradation (6). Over-fertilization is a
resulting in enrichment of reactive N in the air, soil and water with
consequent impairment of ecosystem services. Our studies show
that more efficient use of N fertilizer can allow current N applica-
yields and N balance in rotations, while substantially reducing N
an unnecessary economic expenditure for farmers. The new rec-
ommendations should fully take into account the N supplying
capacity of the soil and N deposited from air and irrigation water.
The characteristics of N behavior among the 4 crops were sharply
different depending on climatic, soil and management practices.
This must be taken into consideration to further reduce N losses.
Only by reducing fertilizer N inputs can degraded environments be
gradually restored, enhanced and protected. Although several
legislative controls in China equivalent to those in the European
Union (33). China would benefit from adopting and enforcing
relevant agricultural regulations (7). All these goals could be
achieved by removing government subsidies, introducing an N
for environmental awareness, and employing practices that avoid
serious environmental degradation (43).
Materials and Methods
Study Areas. Two different representative intensive agricultural regions were
China (32–41 °N, 113–120 °E) in the alluvial plain of the Yellow River (Fig. S3).
Details of the climate, soils, and crops are given in SI Text.
150, 200, and 250 kg of N per hectare. The plots (42 m2in area, 7 ? 6 m) were
On the North China Plain 6 winter wheat and 6 summer maize on-farm field
from October 2003 to October 2005 at Dongbeiwang near Beijing, Huimin
County in Shandong province, and Baoding County in Hebei province. The 5
the 6 summer maize N treatments were 0, 40, 80, 120, 160, and 240 kg of N per
hectare. The plot size was 63 m2(9 ? 7 m). The plots were arranged in a
types and application times are given in the SI Text. Except for fertilizer applica-
was harvested to determine grain dry matter yield.
Field Study 2. In the Taihu region, 6 rice (2 locations per year) and 6 wheat (2
locations per year) field experiments were selected to conduct15N studies in
China Plain, the15N study was carried out in microplots in the on-farm field
experiments (SI Text). In both rotations plant and soil sampling and analysis for
total N and15N abundance are described as in (44). The N recovery rate is
expressed as the percentage of applied15N fertilizer taken up by the
aboveground plant parts and the N retention rate as the percentage of applied
15N fertilizer recovered in the top 100 cm of the soil profile. The loss rate was
calculated by subtracting the recovery rate and retention rate from 100. The
mean and standard variation was calculated across all of the experimental years
and sites in same crop species.
Lysimeter Study. Twenty-four lysimeters containing undisturbed soil pro-
files were used at the research station at Changshu, Jiangsu province (31°
31.93? N, 120° 41.88? E) in the Taihu region. They contained a Typic
Epiaquept (46) formed on alluvial loess with a silty clay loam (46) texture.
A steel cylinder 1-m deep with 1.14-m inner diameter was pushed into the
soil, cut at the base, and then removed. After collection, the lysimeters
were prepared for free drainage by replacing approximately 0.08 m of soil
at the base of each column with a nylon mesh, a gravel layer, and a porous
plastic sheet (47). The lysimeters were then placed in plots permanently
installed belowground at the station. Each lysimeter was used to grow a
waterlogged summer rice/upland winter wheat rotation from October
2003 to October 2005 with 2 rice and 2 wheat growing seasons. Details of
treatments are given in SI Text. Water leaching through the soil columns
was collected in plastic containers placed in an underground measuring
station. The leachate volume was measured at 3-day intervals and sub-
samples were taken from the accumulated leachate and stored at ?18 °C
before chemical analysis. During the winter wheat growing seasons,
leachate could be collected only during rainfall periods. Leachate NO3
analyzer (TRAACS 2000, Bran and Luebbe Inc.). The amount of N leached at
each collection time was calculated by multiplying the volume and the N
concentration of leachate. The total amount of N leached over the whole
growing period was obtained by summing the individual amounts of N
leached at the different collection times. The N leached from fertilizer was
was calculated across 3 replicates over 2 years.
NH3volatilization and N2O measurements were conducted in both rice and
both wheat growing seasons from October 2003 to October 2005. Ammonia
volatilization rate was measured with a continuous airflow enclosure method
(48). The total amount of NH3volatilized over the whole growing period was
obtained by summing NH3volatilization measured each day. N2O fluxes were
closed chambers, and measurements were made more frequently after basal
fertilization and topdressing. Detail measurements and calculations are de-
scribed in (49). Similar NH3-volatilization and N2O measurements at different N
levels were conducted in the on-farm field experiment at Baoding County in
or N2O emission was calculated as the difference between volatized NH3or N2O
in fertilized and unfertilized treatments. The mean and standard variation of
each N treatment was calculated across 3 replicates over 2 years.
?, and total N concentrations were determined using a continuous flow
Field Study 3. A field experiment was conducted from September 2000 to
China Plain. Eight successive crops (4 winter wheat and 4 summer maize) were
experiment were reported previously (8).
15N studies in microplots were also conducted in the conventional N and
optimized N fertilization treatments from September 2002 to September 2004.
NH3volatilization was measured from September 2002 to September 2004 with
system using the C2H2-inhibition method (51) during the 2-year study from
September 2002 to September 2004 (32). Nitrate leaching was quantified using
TerrAquat passive samplers filled with ion-exchange resin. These were used to
ten replicates per plot (52). The mean and standard variation of NH3, N2O, and
nitrate leaching were calculated across 4 replicates over 2 years.
Monitoring Study. Measurements of wet and dry deposition were conducted
region, that is, Changshu, Nanjing, and Hangzhou. Three automatic wet-only
samplers (APS series, Wuhan Tianhong Inc.), which collected precipitation sam-
ples only during precipitation events (controlled by sensors) were separately
installed at the above 3 sites for the collection of wet-only deposition (53).
sampler (41). Before sunrise (about 6:00 AM in winter and around 5:00 AM in
summer), 250 mL of 2 M KCl solution was used to collect deposited dew and the
particles in the collector. In the present experiment, the N in dry deposition is
defined as only the part of the inorganic N adsorbed to particles and exchange-
able with 2 M KCl.
Precipitation samples were collected from 11 monitoring sites located on the
plus sedimenting dry deposition such as dust during dry periods) was collected
Ju et al.
March 3, 2009 ?
vol. 106 ?
no. 9 ?