G Model
JGEC-1116; No. of Pages 10
Global Environmental Change xxx (2013) xxx–xxx
Contents lists available at SciVerse ScienceDirect
Global Environmental Change
journal homepage: www.elsevier.com/locate/gloenvcha
Nitrate in groundwater of China: Sources and driving forces
Baojing Gu a,b, Ying Ge b,c, Scott X. Chang d, Weidong Luo a,c, Jie Chang b,c,*
a
College of Economics, Zhejiang University, Hangzhou 310027, PR China
College of Life Sciences, Zhejiang University, Hangzhou 310058, PR China
c
Research Center for Sustainable Development, Zhejiang University, Hangzhou 310058, PR China
d
Department of Renewable Resources, University of Alberta, Edmonton, Alberta, Canada T6G 2E3
b
A R T I C L E I N F O
A B S T R A C T
Article history:
Received 3 October 2012
Received in revised form 5 May 2013
Accepted 7 May 2013
Identifying the sources of reactive nitrogen (N) and quantifying their contributions to groundwater
nitrate concentrations are critical to understanding the dynamics of groundwater nitrate contamination.
Here we assessed groundwater nitrate contamination in China using literature analysis and N balance
calculation in coupled human and natural systems. The source appointment via N balance was well
validated by field data via literature analysis. Nitrate was detected in 96% of groundwater samples based
on a common detection threshold of 0.2 mg N L 1, and 28% of groundwater samples exceeded WHO’s
maximum contaminant level (10 mg N L 1). Groundwater nitrate concentrations were the highest
beneath industrial land (median: 34.6 mg N L 1), followed by urban land (10.2 mg N L 1), cropland
(4.8 mg N L 1), and rural human settlement (4.0 mg N L 1), with the lowest found beneath natural land
(0.8 mg N L 1). During the period 1980–2008, total reactive N leakage to groundwater increased about
1.5 times, from 2.0 to 5.0 Tg N year 1, in China. Despite that the contribution of cropland to the total
amount of reactive N leakage to groundwater was reduced from 50 to 40% during the past three decades,
cropland still was the single largest source, while the contribution from landfill rapidly increased from 10
to 34%. High reactive N leakage mainly occurred in relatively developed agricultural or urbanized regions
with a large population. The amount of reactive N leakage to groundwater was mainly driven by
anthropogenic factors (population, gross domestic product, urbanization rate and land use type). We
constructed a high resolution map of reactive N source appointment and this could be the basis for future
modeling of groundwater nitrate dynamics and for policy development on mitigation of groundwater
contamination.
ß 2013 Elsevier Ltd. All rights reserved.
Keywords:
Nitrate
Source
High resolution
Landfill
Driving force
Pollution
1. Introduction
Groundwater nitrate (NO3 ) contamination is a threat to
human health (Bryan and Loscalzo, 2011). About 5% of ingested
nitrate is converted by bacteria in the digestive system to nitrite,
which then forms N-nitrosamines and N-nitrosamides that
damage DNA (Davidson et al., 2012). High nitrate concentration
in drinking water can usually induce birth defects and cancers,
which have been the subject of epidemiological studies (Johnson
et al., 2010), particularly in rural agricultural areas where shallow
groundwater is often used for domestic water supplies (Burow
et al., 2010). Nitrate concentrations above the World Health
Organization’s (WHO’s) maximum contaminant level (MCL,
10 mg N L 1) are relatively common in some regions, especially
* Corresponding author at: Department of Biological Science, College of Life
Sciences, Zhejiang University, 866 Yuhangtang Road, Hangzhou 310058, PR China.
Tel.: +86 571 8820 6465; fax: +86 571 8820 6465.
E-mail address: [email protected] (J. Chang).
in the emerging developing countries (Townsend et al., 2003;
Burow et al., 2010).
Previous studies have identified the overuse of nitrogen (N)
fertilizer to be one of the main sources for groundwater nitrate (Li
et al., 2007; Kaushal et al., 2011). Meanwhile, groundwater nitrate
can also be from other sources, such as sewer leakage and landfill
leachate, which are of growing importance alongside urbanization
(Mor et al., 2006; Gu et al., 2011a, 2012a). However, considerable
uncertainty remains in our knowledge of the magnitude and
spatiotemporal changes of groundwater nitrate concentrations
owing to the many sources involved (Galloway et al., 2008; Burow
et al., 2010). Therefore, understanding the sources and implementing source control are key issues on mitigating groundwater nitrate
pollution.
Experimental sampling of groundwater on a large scale is
usually lacking, and local scale experiments are difficult to be
scaled up to regional and continental levels (Kaushal et al., 2011;
Stigter et al., 2011). Thus, attaining a high resolution in the source
appointment becomes an important approach in groundwater
nitrate research. There are a large number of factors, such as
0959-3780/$ – see front matter ß 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.gloenvcha.2013.05.004
Please cite this article in press as: Gu, B., et al., Nitrate in groundwater of China: Sources and driving forces. Global Environ. Change
(2013), http://dx.doi.org/10.1016/j.gloenvcha.2013.05.004
G Model
JGEC-1116; No. of Pages 10
2
B. Gu et al. / Global Environmental Change xxx (2013) xxx–xxx
climate (mean annual temperature, precipitation), human activities (urbanization, industrialization), economy and land use types,
that affect the reactive N (Nr) leakage to groundwater (Birch et al.,
2011; Gu et al., 2011a). Different Nr leaching sources may have
different sensitivities to different driving factors. Understanding
and quantifying human and natural factors that contribute to
changes in groundwater nitrate concentrations are critical in
prioritizing effective strategies for nitrate reduction from different
sources.
China has been experiencing rapid industrialization and
urbanization since the reform and opening up in the late 1970s.
Currently, China consumes about 30% of the global anthropogenic
N fixation on only 7% of the world’s land base, and has been a
hotspot of global N cycle (Townsend and Howarth, 2010; Cui et al.,
2013). Serious groundwater nitrate pollution has occurred
accompanying the tremendous socioeconomic development (Fu
et al., 2007). China’s national water quality assessment program
has been monitoring groundwater quality in 182 main metropolitans and has found that the water quality of 57% (MEP, 2010) of all
monitoring wells does not meet the clean groundwater standard
(<20 mg N L 1 nitrate concentration, MEP, 1993). This has exposed
residents to the risk of high nitrate concentration of groundwater
in China (Liu and Diamond, 2008; Han et al., 2009a). However,
currently, we do not know the overall patterns and dynamics of
groundwater nitrate in China, especially in the rural area, and we
do not fully understand how groundwater nitrate concentrations
change with the rapid urbanization and economic development. In
order to protect groundwater resources, the State Council of China
has officially approved the ‘‘National Groundwater Pollution
Prevention Plan (NGPPP)’’ on 24th August 2011. The NGPPP would
invest about 6 billion US dollars to monitor groundwater pollution
sources and remediate polluted groundwater before 2020. Thus,
comprehensive quantification of the sources of Nr loading to
groundwater in China is essential for the development of effective
management practices for the most vulnerable areas.
The primary purposes of this study were to investigate the
changes in the sources of groundwater nitrate on the spatiotemporal scale and the factors impact groundwater nitrate concentrations in China. To achieve these goals, we first conducted a
literature analysis to understand the current status of groundwater
nitrate pollution in China and the effect of land use type on
groundwater nitrate concentration. Second, we carried out a
source appointment analysis based on the coupled human and
natural systems (CHANS) approach to identify and quantify
specific sources of groundwater nitrate both on temporal (from
1980 to 2008) and spatial (provincial) scales. Third, we quantified
the driving forces of changes in groundwater nitrate concentrations by considering natural and human factors. Finally, we
estimated the spatial distribution of the sources at the county level,
which was further used to compare with the literature analysis
results of the current status of groundwater nitrate pollution, to
assess the accuracy of source appointment analysis. This study
produced a high resolution map of source appointment of
groundwater nitrate concentrations in China that can be used
for the modeling of groundwater nitrate dynamics in the future.
concentration was determined; (ii) locations for the sampling sites
were provided; (iii) the land use type, such as urban (nonindustrial region), industrial, rural, cropland, or natural, of the
sampling sites was clearly indicated. The urban area includes
urban human settlements, commercial areas, parks, etc., but not
including industrial regions. For the industrial area, samples are
taken in factories or very close to the factories. For the rural area,
samples usually are taken from villages. For cropland, samples are
taken in cropland or very close to the cropland. The natural area
generally includes natural forest, grassland, desert, etc.; (iv) the
sampling date was after 1990. All data before 1990 were excluded
since the intensity of human activities would be very different
before and after 1990 in China and the data would be too old to be
of use to represent the current groundwater nitrate status.
2. Methods
Owing to the large amount of data required for source
appointment using the CHANS approach, the spatial resolution
of our calculation can only reach provincial level. Thus, regression
models were used in this study to simulate the Nr loading to
groundwater at the county level that could support a higher
resolution assessment of the environmental and health effects.
Although there are many sources of Nr (both natural and
anthropogenic) that could potentially lead to the pollution of
groundwater with nitrate, anthropogenic sources are the ones that
most often cause the amount of nitrate to rise to a dangerous level
2.1. Literature analysis of groundwater nitrate
To understand the current status of groundwater nitrate
concentrations in China, we reviewed more than 1000 published
papers searched from Web of Science and China National
Knowledge Infrastructure (CNKI) (2000–2012) and chose 108 of
them for this analysis (Supplementary data Table S1) based on the
following criteria: (i) publications in which groundwater nitrate
2.2. CHANS approach on source appointment
The CHANS approach covers and integrates all Nr fluxes and
their interactions that can identify the specific sources of Nr to the
environment (Alberti et al., 2011). Thus, the CHANS approach has
recently been widely used on the studies of pollution source
appointment (Werner and McNamara, 2007; Gu et al., 2011b,
2012a, b). The approach is useful in identifying the components
and flows in crucial systems, and the linkages among subsystems,
as well as analyzing the role of driving factors (Alberti et al., 2011).
In this study, the CHANS is divided into four functional groups (Fig.
S1; Table S2): processors, consumers, removers and life supporters,
based on the mutual services among these groups, and each
functional group includes several subsystems (14 subsystems in
total). External N inputs first go through the processors (e.g.,
cropland, industry, etc.), and are then transferred to consumers for
consumption, and on to removers (wastewater and garbage
treatment) for Nr inactivation, and ultimately become an output
from the system (Fig. S2, Gu et al., 2012b). The Nr from the
processer, consumer and remover functional groups would leak to
the life-supporter group (atmosphere, surface water and groundwater), contributing to the accumulation of Nr in that group (Gu
et al., 2012a).
A mass balance approach was used to quantify the N fluxes for
each subsystem in a CHANS of China (SI text) with over 6000 N
flows (Gu et al., 2012b). Data were mainly derived from Chinese
governmental statistical yearbooks and bulletins (e.g., NBS, 1981–
2009) that supplied the best available data for the quantification of
N fluxes and from published papers that were used for comparison
(Table S3). On the basis of the N balance of the CHANS, we
extracted all the Nr fluxes that were leaked from different
subsystems to the groundwater. The specific sources of each Nr
item were identified and quantified for different regions. We used
the N Cycling Network Analyzer (NCNA) model to compile the
dataset and to calculate all N fluxes (Min et al., 2011). This model
can standardize the parameter collections for the N flux calculations, and automatically calculate the N fluxes and their relationships based on the mass balance approach (Fig. S3).
2.3. Estimates of Nr leakage at county level
Please cite this article in press as: Gu, B., et al., Nitrate in groundwater of China: Sources and driving forces. Global Environ. Change
(2013), http://dx.doi.org/10.1016/j.gloenvcha.2013.05.004
G Model
JGEC-1116; No. of Pages 10
B. Gu et al. / Global Environmental Change xxx (2013) xxx–xxx
(Burow et al., 2010; Bryan and Loscalzo, 2011). In this study, 6 main
sources were identified with the CHANS approach, including
cropland, livestock, wastewater, landfill, grassland, and urban
lawn. These sources were affected by both natural and anthropogenic factors (Gu et al., 2012b). Therefore, a regression model was
developed to estimate the magnitude of Nr leakage to groundwater
considering both natural and anthropogenic factors. Based on the
available data for both provinces and counties in China, a range of
parameters including population (total, rural and urban), gross
domestic product (GDP), per capita GDP, land use (cropland,
grassland and forest), annual mean temperature and precipitation
were examined for their quantitative relationships with various
sources of Nr leakage to groundwater. Most parameters were not
employed in the final model because of an absence of quantitative
relationships. Both linear and nonlinear, and univariate and
multivariate regression models were tested for quantitative
relationships. Generally, a univariate linear model was preferred
for simplicity (details can be found in Section 3.3). Although
regression models were developed by trial-and-error, the final
format could be interpreted reasonably well (details can be found
in Section 3.3). Regressions and statistical analysis were conducted
using the R statistical software.
In this study, 2311 county units were identified in 34 provincial
units combining the administrative division and the availability of
socioeconomic data in China. Data, including populations, GDP, per
capita GDP, annual mean temperature and precipitation were
assembled at county and provincial levels (NBS, 2009; CMA, 2013).
County boundaries of 2008 (at a scale of 1:4,000,000) were used for
mapping (IGSNRR, 2012). The county area was obtained from the
function of ‘calculate geometry’ of ArcGIS 9.3, and then the real
total area of China was used to revise the county area. The land use
of different counties, including cropland, grassland, forest, water
bodies, were derived from the polygon in polygon calculation
between a land use map and the county map in ArcGIS 9.3.
The Nr leakage to groundwater was calculated from an
optimized regression model for all counties in China for 6 main
sources. The total leakage of each county was presented as a sum of
Nr from all sources. Since regression models were derived from
3
provincial data as applied to counties, the rationality of the
interpolation had to be justified. Models were calibrated by
comparing provincial leakage to the sums of the county estimates
before validation. The uncertainties of estimates were characterized by generating probability distributions of the leakages from
various sources. The distributions of activity rates were generated
from normal distributions with a fixed coefficient of variation (5%)
of a number of independent variables including population (total,
rural and urban), GDP, cropland area and grassland area. A total of
10,000 runs were conducted for the Monte Carlo simulation.
3. Results and discussion
3.1. Occurrence of nitrate in groundwater
In the literature analysis, a total of 628 valid data of
groundwater nitrate concentrations were obtained (Fig. 1), the
median nitrate concentration was 4.0 mg N L 1, lower than the
WHO’s MCL of 10 mg N L 1 (Townsend et al., 2003). Nitrate was
detected in 96% of groundwater samples based on a common
detection threshold of 0.2 mg N L 1, and 83% of groundwater
samples had a nitrate concentration above 1.0 mg N L 1, much
higher than the proportion found in the US (about 50%). Nitrate
exceeded the WHO’s MCL in more than 28% of groundwater
samples, and exceeded the Chinese drinking water standard of
20 mg N L 1 (MEP, 1993), in about 15% of sampling sites (Fig. 1).
Generally, nitrate concentrations are lower in deep groundwater
where groundwater is reduced, or where groundwater is older and
hence nitrate concentrations reflect historically low N application
rates (Rupert, 2008). In this study, we also found that nitrate
concentrations were higher in shallow groundwater, but the
sampling depth only explained about 12% of total variation of
nitrate concentrations (R2 = 0.12, p < 0.05), which indicates that
there was other factor affecting the groundwater nitrate concentration. The sampling sites with high nitrate concentrations were
mainly from regions with developed agriculture or industry with
high population density. For samples with groundwater nitrate
concentrations exceeding 10 and 20 mg N L 1, more than 64% and
Fig. 1. Groundwater nitrate concentrations in different sampling sites during 2000–2012. The green color of background represents the N input intensity (e.g., fertilizer,
manure, deposition, biological N fixation) to the land surface. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this
article.)
Please cite this article in press as: Gu, B., et al., Nitrate in groundwater of China: Sources and driving forces. Global Environ. Change
(2013), http://dx.doi.org/10.1016/j.gloenvcha.2013.05.004
G Model
JGEC-1116; No. of Pages 10
B. Gu et al. / Global Environmental Change xxx (2013) xxx–xxx
4
Fig. 3. Changes of Nr leakage to groundwater from different sources in China from
1980 to 2008.
Fig. 2. Comparison of measured groundwater nitrate concentration with Nr leakage
to groundwater among different land use types in China. (a) Groundwater nitrate
concentration in 2000s; (b) Nr leakage to groundwater; (c) comparison of Nr
leakage with nitrate concentration. National maximum contaminant level (MCL)
represents the maximum contaminant level of groundwater for drinking defined by
Ministry of Environmental Protection in China (20 mg NO3 N L 1, MEP, 1993);
WHO MCL represents the maximum contaminant level of groundwater defined by
World Health Organization (10 mg NO3 N L 1, Townsend et al., 2003). Different
letters on each box represent a significant difference across land use type (p < 0.05).
74% were from North China Plain (the old industrial and
agricultural bases in China), respectively. This indicates that
human activities had dramatically affected groundwater nitrate
concentrations.
In order to further explore the effects of human activities on
groundwater nitrate, we evaluated nitrate concentrations under
different land use types (Fig. 2). Groundwater beneath industrial
land has the highest median nitrate concentration, 34.6 mg N L 1,
in China. Although the point source pollution of industrial
wastewater has been substantially reduced in developed countries
(Grimm et al., 2008), it seems that industrial wastewater still has a
large effect on local groundwater quality in industrial areas in
China. Groundwater nitrate concentrations beneath industrial land
in China were about a magnitude higher than that in the USA
(about 1.6 mg N L 1, Burow et al., 2010). Groundwater beneath
urban human settlement ranked second in terms of the median
nitrate concentration (Fig. 2), at 10.2 mg N L 1, which was close to
the WHO’s MCL. For cropland and rural human settlement, the
median groundwater nitrate concentrations were similar, at 4.8
and 4.0 mg N L 1, respectively, slightly higher than that in the USA
(3.1 mg N L 1, Burow et al., 2010). Groundwater beneath natural
land had the lowest median nitrate concentration, about
0.8 mg N L 1, but still higher than the common detection threshold
of 0.2 mg N L 1, indicating that natural ecosystems had also been
affected by human activities in China, such as via elevated N
deposition (Liu et al., 2011).
3.2. Source appointment
Through source appointment using the CHANS approach, we
found that the annual total of Nr leakage to groundwater increased
about 1.5 times, from 2.0 Tg N year 1 in 1980 to 5.0 Tg N year 1 in
2008 (Fig. 3). Cropland was the largest source of Nr leaching to
groundwater, although its contribution was reduced from 50% in
the 1980s to 40% in the 2000s. Fertilizer application was the main
process that led to the cropland being the leading contributor to
groundwater nitrate pollution. There is about 5–10% of applied
fertilizer that would enter into groundwater (Ju et al., 2009; Gu
et al., 2012a). The annual growth of Nr leakage to groundwater
from cropland was as high as 12% before 1995, much higher than
the average value, 7%, for the recent 30 years, mainly because of the
rapidly increased usage of synthetic fertilizer before 1995 (NBS,
1981–2009; Gu et al., 2012b). The Nr leakage to groundwater from
cropland remained stable after 1995 owing to two reasons: first,
the growth rate of synthetic fertilizer application was relatively
low; and second, the manure applied to cropland was reduced
accompanying urbanization that decoupled cropland and excretion, and more excretions were treated through wastewater
treatment plants (Naylor et al., 2005; Gu et al., 2012a).
Landfill leakage is the fastest growing source of nitrate to
groundwater in China (Fig. 3). Its contribution increased from 10%
in 1980 to 34% in 2008, becoming the second largest source to
groundwater after cropland. Rapid urbanization and industrialization are the main driving forces for the increase of landfill leakage
(Mor et al., 2006). The main sources of landfill are food waste,
discarded commodity, lawn waste and pet excretion. Food waste in
rural area generally is used for raising livestock in China; on the
contrary, it is sent to landfill in the urban area (Zhang et al., 2010).
Thus, the urbanization increases the amount of food waste going to
landfill. Meanwhile, urbanization largely contributed to the
increase of urban greenland and number of pets (NBS, 1981–
2009), which led to a 5.1 and 1.3 times of increase of lawn waste
and pet excretion, respectively, over the past 30 years. The
industrialization process largely promoted the usage of N-containing products, such as synthetic fiber, which has increased 13.4
times over the past 30 years, becoming one of the primary landfill
components in China (Gu et al., 2012c).
Domestic wastewater leakage via septic systems in rural areas
and sanitary sewer systems in urban areas and the industrial
wastewater leakage increased over 40% during the past 30 years,
from 0.4 to 0.6 Tg N year 1 (Fig. 3). However, the contribution
fraction of wastewater was reduced from 22% to 12% during this
period owing to the rapid increase of Nr leaking from landfill.
Domestic waste leakage is primarily driven by population growth
(Gu et al., 2012a), while industrial wastewater leakage is mainly
determined by economic development and technological improvement (Kapley and Purohit, 2009). Industrial wastewater only
Please cite this article in press as: Gu, B., et al., Nitrate in groundwater of China: Sources and driving forces. Global Environ. Change
(2013), http://dx.doi.org/10.1016/j.gloenvcha.2013.05.004
G Model
JGEC-1116; No. of Pages 10
B. Gu et al. / Global Environmental Change xxx (2013) xxx–xxx
5
Fig. 4. Variations of Nr leakage to groundwater from different sources across China’s provinces in 2008. (a) Total Nr leakage to groundwater; (b) total Nr leakage to
groundwater on the base of per liter groundwater. X-axis represents different provinces, sorted by per capita GDP from low to high.
accounts for less than 5% of total Nr leakage to groundwater from
wastewater (including industrial and domestic wastewater).
However, industrial wastewater always has a concentrated dose
of Nr that leaches to a small area and deteriorates groundwater
quality (Fig. 2). The Nr leakage to groundwater from livestock and
grassland doubled over the past 30 years, mainly attributed to the
growth of the livestock industry that satisfied the increase of
human’s needs on animal protein in China (NBS, 1981–2009).
Urban lawn generally contributes less than 0.3% of total Nr input to
groundwater.
On the spatial basis, the high Nr leakage mainly occurred in
developed agricultural or industrialized provinces with large
populations (Fig. 4), such as Shandong, Henan, Sichuan, Hubei,
Guangdong and Jiangsu. Cropland was a major source in
agricultural provinces, such as Nr leakage from cropland contributes over 50% of total Nr leakage to groundwater in Shandong and
Henan. However, in industrialized provinces, landfill was the
major source that could contribute over 50% of total Nr leakage to
groundwater, such as in Guangdong and Jiangsu. In some provinces
in Western China with less developed agriculture or industry, such
as in Tibet, Nr leakage from grassland was the main source.
Efforts have been made to understand how different factors
(e.g., redox conditions, soil properties, iron and dissolved oxygen
concentrations) affect the final nitrate concentration in the
groundwater (Burow et al., 2010; Li et al., 2010; Kaushal et al.,
2011). However, it is difficult to alter these factors, such as
dissolved oxygen concentrations, to mitigate nitrate pollution
(Bryan and Loscalzo, 2011; Stigter et al., 2011). Therefore, it is
important to analyze the driving forces on Nr leaching sources that
can help regulate groundwater nitrate contamination. On the basis
of per hectare land area and per liter groundwater, the driving
forces of changes in nitrate leakage across different provinces were
analyzed. Results show that the amount of nitrate leakage were
generally related to socioeconomic factors, such as per capita GDP,
urbanization and population density (Fig. 5), indicating that high
nitrate concentrations usually occur in relatively developed
Fig. 5. Relation between anthropogenic (per capita GDP (PGDP), urbanization and population density (PD)) and natural factors (mean annual temperature (MAT) and
precipitation) and Nr leakage to groundwater on the basis of per hectare land area and per liter groundwater.
Please cite this article in press as: Gu, B., et al., Nitrate in groundwater of China: Sources and driving forces. Global Environ. Change
(2013), http://dx.doi.org/10.1016/j.gloenvcha.2013.05.004
G Model
JGEC-1116; No. of Pages 10
B. Gu et al. / Global Environmental Change xxx (2013) xxx–xxx
6
regions, such as Beijing, Shanghai, Tianjin. Natural factors (mean
annual temperature and precipitation) were not related to the
variations in Nr release to the groundwater. Furthermore, landfill
was one of the most important sources in these developed
provinces or cities, consistent with the plight of being ‘‘junk
besieged’’ metropolitans in China (Zhang et al., 2010).
3.3. Factors dominate different sources
The Nr leakage from cropland was very well predicted by
cropland area in each province except three outliers (Fig. 6a).
Although the N application rate (including all Nr inputs to
cropland, e.g., fertilizer, manure, straw, deposition, etc.) can also
affect the Nr leakage to groundwater, its contribution to total Nr
leaching in each province was much smaller than that of the
cropland area. In fact, fertilizer application rate was higher in
eastern China than that in western China, but, the reverse was true
for manure application rate (Gu et al., 2011b). Therefore, except the
three outliers (Fig. 6a), the total Nr application rates to cropland
were similar among the provinces in China. Unlike other provinces,
Henan and Shandong are agricultural provinces that are relatively
more developed and have a large population (NBS, 2009). To
increase crop yield and meet human’s food demand, overfertilization is common in both provinces (Ju et al., 2009), leading
to the discrepancy between predicted and measured values of Nr
leakage to groundwater. On the contrary, the predicted values in
Heilongjiang were much higher than the measured values.
Heilongjiang is one of the main regions of soybean production,
accounting for about 10% of total production in China (NBS, 1981–
2009). There is almost no need for anthropogenic Nr input to
soybean cultivation owing to its symbiotic N fixation (Galloway
et al., 2008), which largely lower the per area Nr leakage to
groundwater. If the three outliers (Henan, Shandong, and
Heilongjiang) are removed, the R2 value is 0.72 for the relationship
between N leakage and cropland area (Fig. 6a). For other provinces,
although the relatively longer growing season in southern China
may result in larger Nr input to cropland (Huang et al., 2007), the
higher annual mean temperature and precipitation may also lead
to more Nr loss through volatilization or surface water runoff, not
leaching to groundwater (Han et al., 2009b). This explains that the
Nr leakage to groundwater from per unit cropland area was similar
in different provinces except the outliers.
The Nr leakage from livestock was related to rural population
except two outliers, Tibet and Jiangsu (Fig. 6b). Generally, the
majority of China’s livestock is raised by small holder farmers, and
breeding has been one of the main sources of income in rural areas
(Li, 2009). This indicates that the size of the rural population was
highly related to the number of livestock raised. Tibet has a small
population while abundant grassland that supports a large per
capita number of animals raised, about 50% larger than China’s
average value (NBS, 1981–2009). While for Jiangsu province, the
development of industry in the townships attracts most of the rural
labors that reduces the number of per capita livestock raised.
Without the two outliers, the R2 value of the model was as high as
0.77.
Similar to cropland, the Nr leakage to groundwater from
grassland is also significantly related to grassland area in each
province (Fig. 6c). If the outlier (i.e., Sichuan) was removed, the R2
was 0.81. Generally, the Nr leakage to groundwater in grassland is
mainly determined by the Nr input (e.g., N deposition) and removal
(animal grazing). Compared to other grassland provinces (e.g.,
Inner Mongolia, Xinjiang), Sichuan is a province with relatively
developed agriculture and industry that could result in higher Nr
deposition (Gu et al., 2012b). Meanwhile, the grassland in Sichuan
belongs to the alpine-bush grassland type that is inconvenient for
grazing, which reduces the Nr removal and increases Nr leakage
from grassland. Therefore, the actual Nr leakage to groundwater
from grassland was higher than the prediction.
Domestic wastewater accounted for over 95% of the total
wastewater Nr leaching to groundwater in China. Generally
speaking, the per capita Nr leakage via excretion is stable because
the human’s basic metabolic need on protein is stable (Bilsborough
and Mann, 2006). Therefore, it was expected that the Nr leakage to
groundwater from wastewater is highly correlated to human
population (Fig. 6d). The per capita Nr leakage to groundwater
via domestic wastewater is smaller in rural than in urban areas
due to the execration recycled to cropland in rural areas (Li, 2009).
Due to the development of industries in small townships, the
Fig. 6. Different sources of Nr leakage to groundwater as functions of anthropogenic factors. (a) Nr leakage from cropland as a function of cropland area; (b) Nr leakage from
livestock as a function of rural population; (c) Nr leakage from grassland as a function of grassland area; (d) Nr leakage from wastewater as a function of population; (e) Nr
leakage from landfill as a function of GDP; (f) Nr leakage from urban lawn as a function of urban population. The red dots were considered as outliers and not included in the
regression analysis.
Please cite this article in press as: Gu, B., et al., Nitrate in groundwater of China: Sources and driving forces. Global Environ. Change
(2013), http://dx.doi.org/10.1016/j.gloenvcha.2013.05.004
G Model
JGEC-1116; No. of Pages 10
B. Gu et al. / Global Environmental Change xxx (2013) xxx–xxx
industrialization reduced the human excretion recycled to
cropland in rural areas in Jiangsu province (NBS, 1981–2009).
Therefore, the predicted values were lower than the measured
values for Jiangsu province. If Jiangsu is removed from consideration, the R2 reached 0.93 for the model.
Accompanying the socioeconomic development, food waste
and discarded commodity rapidly increased to be the majority
components of landfill (Zhang et al., 2010; Gu et al., 2012c),
accounting for over 93% of total Nr going to landfills in China. We
found that Nr release to groundwater from landfill was significantly related to GDP (R2 = 0.96) without outliers occurring among
all the provinces in China (Fig. 6e). It revealed that the sole driving
force on the amount of landfill was socioeconomic development.
For urban lawn, urban population was a good index for Nr leakage
to groundwater from lawn if Guangdong was excluded (Fig. 6f).
Generally, the urban green space is correlated with urban
population. However, Guangdong is located in southern China
with high mean annual temperature and precipitation (NBS, 1981–
2009), which is suitable for rapid plant growth. Furthermore, it is
one of the most developed provinces that can also afford the
inclusion of more green space in urban areas. Therefore,
Guangdong has a higher per capita green space. If Guangdong
was excluded from the dataset, the R2 of the model increased to
0.75.
3.4. Nr leakage to groundwater at the county level
The total Nr leakage calculated for all counties in China was
4932.2 Gg N year 1, which was very close to the provincial-level
estimation of 5013.0 Gg N year 1 in 2008. Similar estimation at
county level based on the regression between total Nr leakage and
population was 5071.0 Gg N year 1 in 2008. The overall prediction
was further evaluated by plotting the predictions directly at
provincial level against those calculated at county level and
summed for each province; all data points fall around the 1:1 line
(Fig. 7), showing no systematic error in the results. The final
estimates of Nr leakage for individual counties were rectified based
on the provincial level estimation, bringing all data points in Fig. 7
right onto the 1:1 line.
7
Fig. 7. Comparison of the total Nr leakage to groundwater derived from provincial
level and county level estimations. Method 1 represents the county level estimation
derived from the sum of Nr leakage from different sources. Method 2 represents the
county level estimation derived from direct correlation with population.
Owing to large differences in local socioeconomic conditions
(NBS, 2009), total Nr leakages differed considerably from one
county to another in many provinces (Fig. 8). As indicated by Fig. 4,
Nr leakage densities of the southeastern provinces were generally
higher than those of other provinces. When the leakage is
presented at the county level resolution, detailed differences
within various provinces were revealed. For all provinces, major
cities stand out as the leakage centers, even for provinces in the
west with lower leakage rates (Fig. 8). Capital cities can be clearly
pinpointed with notably higher leakages than those of surrounding
areas. For the 324 largest cities with a population greater than 1
million in China, the average leakage density was 25 kg N ha 1 in
2008, over 4 times higher than the national average of 5 kg N ha 1.
Although these cities occupy only 8% of the total territory, they
contributed to 40% of the total leakage. Geographical variations
Fig. 8. Geographical distribution of Nr leakage intensity to groundwater in China.
Please cite this article in press as: Gu, B., et al., Nitrate in groundwater of China: Sources and driving forces. Global Environ. Change
(2013), http://dx.doi.org/10.1016/j.gloenvcha.2013.05.004
G Model
JGEC-1116; No. of Pages 10
8
B. Gu et al. / Global Environmental Change xxx (2013) xxx–xxx
within individual provinces are noteworthy. For example, leakages
averaged 6 kg N ha 1 in Sichuan close to the average for all
provinces. Leakages were high, >15 kg N ha 1, for the counties in
Sichuan Basin (the eastern part of Sichuan), but <1 kg N ha 1 for
the counties in the mountain region (the western part of Sichuan).
Such high intra-provincial variations were also remarkable in
Heilongjiang, Inner Mongolia, Gansu, Xinjiang, Jilin, Yunnan, and
some other provinces.
For different sources, large variations were also observed across
counties (Fig. 9). The Nr leakage from cropland and grassland was
related to the size of cropland and grassland in each county; thus,
the distribution of Nr leakage from cropland and grassland
exhibited a contiguous change. For example, Nr leakage from
cropland in North China Plain, Sichuan Basin, the Mid-Lower
Yangtze Plain was as high as 7–10 kg N ha 1 year 1, much higher
than in other regions. Similarly, the hotspots of Nr leakage from
grassland were mainly located in Western Tibet, Eastern Inner
Mongolia, and so on. Leakage intensity of Nr from livestock was
related to rural population, leading to the hotspots of Nr leakage
from livestock mainly occurring in rural areas, and urban areas
turned out to be the cold spots of Nr leakage from livestock. On the
contrary, Nr leakages from urban lawn, wastewater and landfill
were highly related to urban development; therefore, scattered
hotspots of Nr leakage were observed in each province, especially
in capital regions in each province.
On the basis of population density, urbanization level and
cropland area ratio in each county, we classified all the counties
into five classes: industrial, urban, rural, cropland and natural land
(Fig. 2). The Nr releases to the five different land uses were highly
consistent with the sampling data of nitrate concentration. It
Fig. 9. Distributions of Nr leakage to groundwater from cropland, livestock, grassland, urban lawn, wastewater, and landfill in China.
Please cite this article in press as: Gu, B., et al., Nitrate in groundwater of China: Sources and driving forces. Global Environ. Change
(2013), http://dx.doi.org/10.1016/j.gloenvcha.2013.05.004
G Model
JGEC-1116; No. of Pages 10
B. Gu et al. / Global Environmental Change xxx (2013) xxx–xxx
suggests that our source appointment and high resolution of Nr
leakage worked well on assessing groundwater nitrate pollution
(Fig. 2). Therefore, these factors dominated different sources could
be used as potential indicators for groundwater nitrate concentrations.
3.5. Uncertainty analysis
In this study, the uncertainty was mainly associated with
source appointment and simulation of spatial distribution at the
county level. Uncertainties of source appointment were mainly
caused by obtaining data from Chinese official publications for
information such as population, GDP, fertilizer usage and the
derivation of the N cycle related variables. The publication of
official data started from 1980 when a sound system of statistical
collection was established, and the trends of changes over time
and comparisons of spatial distributions were used; therefore,
the official statistics (NBS, 1981–2009) were relied on as the
source of data in this study. Uncertainties that result from data
derived from official data, e.g., statistical yearbooks, fall within
the range of approximately 5% since they use an identical system
for statistical analysis (NBSC, 1981–2009). Variables used for N
flux estimation in this study were taken from the existing
literature, which would also introduce potential uncertainties.
The 6 main sources of groundwater nitrate were identified based
on the N processes in the whole CHANS. During the N balance
calculation in the CHNAS, all the N fluxes smaller than 1 Gg
(0.001 Tg) were ignored. On this basis, six sources of Nr leaching
to groundwater were identified at the subsystem level. There
must be other sources with N fluxes smaller than 1 Gg, which
were not considered in this study. Meanwhile, there also might
be other processes that can lead to Nr leaching to groundwater,
e.g., Nr leakage from rivers, but we still lack the knowledge of
these processes, which may also lead to uncertainties on source
appointment of groundwater nitrate. Further research is
required in these areas to better understand the sources and
degrees of uncertainty so as to improve the accuracy of nitrate
flux estimation.
For the uncertainties of simulation of spatial distribution at the
county level, although the significances of regression models
reached the level of p < 0.001, the coefficient of determinations
(R2) of the models were somewhat low, ranging between 0.72 and
0.96, and still carried some uncertainties. In particular, several
outliers were found for the Nr leakages from cropland and
livestock. We characterized the uncertainties of the modeling
through generating probability distributions of the leakages from
various sources. The distributions of socioeconomic factors were
generated from normal distributions with a fixed coefficient of
variation (5%) of a number of independent variables including
population, land area, and GDP. We found that the uncertainties of
Nr leakages from different sources varied, but all less than 1%,
which means that the socioeconomic factors brought in small
uncertainties and the majority of the uncertainties were attributed
to variables that were related to N cycle.
4. Concluding remarks
Human activities have intensively altered the processes
contributing to Nr leakage to groundwater, and a large amount
of nitrate has accumulated in groundwater in China, a trend that is
accelerating. Agriculture is the largest source for groundwater
nitrate accumulation, although its contribution is decreasing. This
results in a large risk to the health of rural populations, the
majority of which still directly use the shallow groundwater as
drinking water in China. Meanwhile, landfill has rapidly grown to
become the second largest source of groundwater nitrate, which is
9
critical in affecting groundwater quality in urban areas. The point
source Nr leakage (landfill, wastewater) to groundwater usually
occurs with more concentrated doses of release in a small area, and
leads to the much higher accumulation of groundwater nitrate in
urban regions, especially in industrial regions. The high resolution
spatial distribution map of Nr sources to groundwater indicates
scattered hotspots of Nr leakage in urban areas, especially for
provincial capital cities. Although natural factors may also affect
the groundwater nitrate concentrations, anthropogenic factors
(growth of population and GDP, urbanization and land use change)
can explain the majority of Nr leakages to groundwater.
Acknowledgments
We thank Du Y., Fan X., Gong X., Li D., and Liu Y. for the help on
nitrate data collection, and Yang G. for the help on GIS analysis.
This study was supported by National Science Foundation of China
(Grant nos. 41201502 and 31170305) and China Postdoctoral
Science Special Foundation (Grant no. 2012T50508).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.gloenvcha.2013.05.004.
References
Alberti, M., Asbjornsen, H., Baker, L.A., Brozovic, N., Drinkwater, L.E., Drzyzga, S.A.,
Jantz, C.A., Fragoso, J., Holland, D.S., Kohler, T.A., Liu, J., McConnell, W.J.,
Maschner, H.D.G., Millington, J.D.A., Monticino, M., Podestá, G., Pontius, R.G.,
Redman, C.L., Reo, N.J., Sailor, D., Urquhart, G., 2011. Research on coupled
human and natural systems (CHANS): approach, challenges, and strategies.
Bulletin of the Ecological Society of America 92, 218–228.
Bilsborough, S., Mann, N., 2006. A review of issues of dietary protein intake in
humans. International Journal of Sport Nutrition and Exercise Metabolism 16,
129–152.
Birch, M.B.L., Gramig, B.M., Moomaw, W.R., Doering III, O.C., Reeling, C.J., 2011. Why
metrics matter: evaluating policy choices for reactive nitrogen in the Chesapeake Bay watershed. Environmental Science and Technology 45, 168–174.
Bryan, N.S., Loscalzo, J., 2011. Nitrite and Nitrate in Human Health and Disease.
Humana Press, New York.
Burow, K.R., Nolan, B.T., Rupert, M.G., Dubrovsky, N.M., 2010. Nitrate in groundwater of the United States, 1991–2003. Environmental Science and Technology 44,
4988–4997.
China Meteorological Administration (CMA), 2013. Dataset of Surface Climate Data
in China. China Meteorological Data Sharing Service System, http://cdc.cma.gov.cn/home.do.
Cui, S., Shi, Y., Groffman, P.M., Schlesinger, W.H., Zhu, Y.-G., 2013. Centennial-scale
analysis of the creation and fate of reactive nitrogen in China (1910–2010).
Proceedings of the National Academy of Sciences 110, 2052–2057.
Davidson, E., David, M., Galloway, J., Goodale, C., Haeuber, R., Harrison, J., Howarth,
R., Jaynes, D., Lowrance, R., Nolan, T., Peel, J., Pinder, R., Porter, E., Snyder, C.,
Townsend, A., Ward, M.H., 2012. Excess nitrogen in the U.S. environment:
trends, risks, and solutions. Issues in Ecology 15, 1–16.
Fu, B., Zhuang, X., Jiang, G., Shi, J., Lu, Y., 2007. Environmental problems and
challenges in China. Environmental Science and Technology 41, 7597–7602.
Galloway, J.N., Townsend, A.R., Erisman, J.W., Bekunda, M., Cai, Z., Freney, J.R.,
Martinelli, L.A., Seitzinger, S.P., Sutton, M.A., 2008. Transformation of the
nitrogen cycle: recent trends, questions, and potential solutions. Science
320, 889–892.
Grimm, N.B., Faeth, S.H., Golubiewski, N.E., Redman, C.L., Wu, J., Bai, X., Briggs, J.M.,
2008. Global change and the ecology of cities. Science 319, 756–760.
Gu, B., Dong, X., Peng, C., Luo, W., Chang, J., Ge, Y., 2012a. The long-term impact of
urbanization on nitrogen patterns and dynamics in Shanghai, China. Environmental Pollution 171, 30–37.
Gu, B., Ge, Y., Luo, W., Du, Y., Yang, G., Chang, J., 2012c. Rapid growth of industrial
nitrogen fluxes in china: driving forces and consequences. Science China Earth
Science, http://dx.doi.org/10.1007/s11430-012-4556-3.
Gu, B., Ge, Y., Ren, Y., Xu, B., Luo, W., Jiang, H., Gu, B., Chang, J., 2012b. Atmospheric
reactive nitrogen in china: sources, recent trends, and damage costs. Environmental Science and Technology 46, 9420–9427.
Gu, B., Liu, D., Wu, X., Ge, Y., Min, Y., Jiang, H., Chang, J., 2011b. Utilization of waste
nitrogen for biofuel production in China. Renewable and Sustainable Energy
Reviews 15, 4910–4916.
Gu, B., Zhu, Y., Chang, J., Peng, C., Liu, D., Min, Y., Luo, W., Howarth, R.W., Ge, Y.,
2011a. The role of technology and policy in mitigating regional nitrogen
pollution. Environmental Research Letters 6, 014011.
Please cite this article in press as: Gu, B., et al., Nitrate in groundwater of China: Sources and driving forces. Global Environ. Change
(2013), http://dx.doi.org/10.1016/j.gloenvcha.2013.05.004
G Model
JGEC-1116; No. of Pages 10
10
B. Gu et al. / Global Environmental Change xxx (2013) xxx–xxx
Han, H., Allan, J.D., Scavia, D., 2009b. Influence of climate and human activities on
the relationship between watershed nitrogen input and river export. Environmental Science and Technology 43, 1916–1922.
Han, J., Xu, Z., Xing, H., Tan, J., Chen, F., Wang, M., Kong, F., Zhang, S., Wu, X., Li, X., Li,
J., Guo, P., 2009a. The causes and preventive measures in China’s 966 counties
having a high incidence of esophageal cancer. Henan Journal of Preventive
Medicine 20, 1–4.
Huang, Y., Zhang, W., Sun, W., Zheng, X., 2007. Net primary production of Chinese
croplands from 1950 to 1999. Ecological Applications 17, 692–701.
Institute of Geographical Sciences, Natural Resources Research Chinese Academy of
Sciences (IGSNRR), . Thematic database for human-earth system. http://
www.naturalresources.csdb.cn/.
Johnson, P.T.J., Townsend, A.R., Cleveland, C.C., Glibert, P.M., Howarth, R.W., McKenzie, V.J., Rejmankova, E., Ward, M.H., 2010. Linking environmental nutrient
enrichment and disease emergence in humans and wildlife. Ecological Applications 20, 16–29.
Ju, X.-T., Xing, G.-X., Chen, X.-P., Zhang, S.-L., Zhang, L.-J., Liu, X.-J., Cui, Z.-L., Yin, B.,
Christie, P., Zhu, Z.-L., Zhang, F.-S., 2009. Reducing environmental risk by
improving N management in intensive Chinese agricultural systems. Proceedings of the National Academy of Sciences 106, 3041–3046.
Kapley, A., Purohit, H.J., 2009. Diagnosis of treatment efficiency in industrial
wastewater treatment plants: a case study at a refinery ETP. Environmental
Science and Technology 43, 3789–3795.
Kaushal, S.S., Groffman, P.M., Band, L.E., Elliott, E.M., Shields, C.A., Kendall, C., 2011.
Tracking nonpoint source nitrogen pollution in human-impacted watersheds.
Environmental Science and Technology 45, 8225–8232.
Li, P., 2009. Exponential growth, animal welfare, environmental and food safety
impact: the case of China’s livestock production. Journal of Agricultural and
Environmental Ethics 22, 217–240.
Li, S., Liu, C., Lang, Y., Zhao, Z., Zhou, Z., 2010. Tracing the sources of nitrate in karstic
groundwater in Zunyi, Southwest China: a combined nitrogen isotope and
water chemistry approach. Environmental Earth Sciences 60, 1415–1423.
Li, X., Masuda, H., Koba, K., Zeng, H., 2007. Nitrogen isotope study on nitratecontaminated groundwater in the Sichuan basin, China. Water, Air, and Soil
Pollution 178, 145–156.
Liu, J., Diamond, J., 2008. Revolutionizing China’s environmental protection. Science
319, 37–38.
Liu, X., Duan, L., Mo, J., Du, E., Shen, J., Lu, X., Zhang, Y., Zhou, X., He, C., Zhang, F.,
2011. Nitrogen deposition and its ecological impact in China: an overview.
Environmental Pollution 159, 2251–2264.
Min, Y., Gong, W., Jin, X., Chang, J., Gu, B., Han, Z., Ge, Y., 2011. NCNA: integrated
platform for constructing, visualizing, analyzing and sharing human-mediated
nitrogen biogeochemical networks. Environmental Modelling and Software 26,
678–679.
Ministry of Environmental Protection (MEP), 1993. Quality Standard for Groundwater. MEP, Beijing.
Ministry of Environmental Protection (MEP), 2010. China Environment Yearbook.
China Environment Yearbook, Beijing.
Mor, S., Ravindra, K., Dahiya, R., Chandra, A., 2006. Leachate characterization and
assessment of groundwater pollution near municipal solid waste landfill site.
Environmental Monitoring and Assessment 118, 435–456.
National Bureau of Statistics (NBS), 1981–2009. China Statistical Yearbook. National
Bureau of Statistics of China, Beijing.
National Bureau of Statistics (NBS), 2009. China Statistical Yearbook for Regional
Economy 2009. China Financial and Economic Publishing House, Beijing.
Naylor, R., Steinfeld, H., Falcon, W., Galloway, J., Smil, V., Bradford, E., Alder, J.,
Mooney, H., 2005. Losing the links between livestock and land. Science 310,
1621–1622.
Rupert, M.G., 2008. Decadal-scale changes of nitrate in ground water of the United
States, 1988–2004. Journal of Environmental Quality 37 (Suppl. 5) 240–248.
Stigter, T.Y., Carvalho Dill, A.M.M., Ribeiro, L., 2011. Major issues regarding the
efficiency of monitoring programs for nitrate contaminated groundwater.
Environmental Science and Technology 45, 8674–8682.
Townsend, A.R., Howarth, R.W., 2010. Fixing the global nitrogen problem. Scientific
American 302, 64–71.
Townsend, A.R., Howarth, R.W., Bazzaz, F.A., Booth, M.S., Cleveland, C.C., Collinge,
S.K., Dobson, A.P., Epstein, P.R., Holland, E.A., Keeney, D.R., Mallin, M.A., Rogers,
C.A., Wayne, P., Wolfe, A.H., 2003. Human health effects of a changing global
nitrogen cycle. Frontiers in Ecology and the Environment 1, 240–246.
Werner, B.T., McNamara, D.E., 2007. Dynamics of coupled human-landscape systems. Geomorphology 91, 393–407.
Zhang, D.Q., Tan, S.K., Gersberg, R.M., 2010. Municipal solid waste management in
China: status, problems and challenges. Journal of Environmental Management
91, 1623–1633.
Please cite this article in press as: Gu, B., et al., Nitrate in groundwater of China: Sources and driving forces. Global Environ. Change
(2013), http://dx.doi.org/10.1016/j.gloenvcha.2013.05.004