Journal of Cleaner Production 224 (2019) 375e383

Contents lists available at ScienceDirect

Journal of Cleaner Production

journal homepage: www.elsevier.com/locate/jclepro

Spatial-temporal assessment of water footprint, water scarcity and

crop water productivity in a major crop production region
Zhenci Xu a, Xiuzhi Chen b, Susie Ruqun Wu c, Mimi Gong a, Yueyue Du d, Jinyan Wang b,
Yunkai Li b, *, Jianguo Liu a, **
a
Center for Systems Integration and Sustainability, Michigan State University, East Lansing, 48823, USA
b
College of Water Resources and Civil Engineering, China Agricultural University, Beijing, 100083, China
c
Center for Global Change and Earth Observations, Michigan State University, East Lansing, MI, 48823, USA
d
Laboratory for Earth Surface Processes, Ministry of Education, College of Urban and Environmental Sciences, Peking University, Beijing, 100871, China

a r t i c l e i n f o a b s t r a c t

Article history: Irrigated agriculture has had an enormous influence on food security, water security and human well-
Received 26 March 2018 being. Water footprint (how much water is used), water scarcity (how scarce water is), and crop water
Received in revised form productivity (how much productivity irrigation adds) are important indicators for evaluating sustain-
13 December 2018
ability in irrigated agriculture. Yet these interrelated indicators have not been studied simultaneously at
Accepted 11 March 2019
the county level e the basic administrative unit of agricultural planning and water management in
Available online 19 March 2019
countries such as China, India and Japan. To fill this knowledge gap, we performed a demonstration in
China's major crop production region, the North China Plain (NCP)'s 207 counties from 1986 to 2010. The
Keywords:
Water footprint
results show that the irrigated agriculture's annual water footprint in the North China Plain increased
Water scarcity from 53 billion m3 in 1986 to 78 billion m3 in 2010. All counties faced water scarcity during 1986-2010
Crop water productivity even as the average crop water productivity increased from 0.90 kg m
3 to 1.94 kg m
3. There are 173
Irrigated agriculture NCP counties suffering severe water scarcity but still producing significant crop yield with a high water
Food security footprint, a red flag of unsustainable irrigated agriculture. This study has implications for revealing
Sustainability potential unsustainable conditions in irrigated agriculture worldwide.
China © 2019 Elsevier Ltd. All rights reserved.
County level
North China plain

1. Introduction worldwide (consumptive water use is water removed from avail-

able supplies without return to a water resource system) (Do €ll,
Global challenges involving food and water play significant roles 2009; Food and Agriculture Organization of the United Nations,
in sustainability and human well-being worldwide. The Earth's 2018). Forty percent of global agricultural production requires
freshwater resources have been facing tremendous pressure due to irrigation (Viala, 2008).
increasing consumptive use and water pollution (Steffen et al., Much effort has been made to improve irrigated agriculture's
2015; Mekonnen and Hoekstra, 2016). For example, global water performance on water consumption and crop yields for more sus-
withdrawal increased 630 percent during 1900e2010 (Food and tainable development. Many public policies have been applied and
Agriculture Organization of the United Nations, 2018). Global food billions of dollars spent to save water in irrigated agriculture (Ward
production also faces great challenges since by 2050, 9 billion and Pulido-Velazquez, 2008). The water footprint, water scarcity,
people would need to be fed (Godfray et al., 2010). and crop water productivity are used as indicators to assess water
Irrigated agriculture has important implications for both water and food sustainability. A product's water footprint (WF) is the total
security and food security. It accounts for more than 70% of the total volume of freshwater consumed to produce the product (Liu et al.,
water use, and more than 90% of total consumptive water use 2009; Mekonnen and Hoekstra, 2011). WF includes not only direct
water consumption of products, but also indirect water consump-
tion e water indirectly consumed and water polluted throughout
the production chain. Water scarcity shows a shortage of renewable
* Corresponding author. fresh water compared to water demand (Raskin et al., 1996;
** Corresponding author.
Damkjaer and Taylor, 2017). We measure agricultural water use
E-mail addresses: liyunkai@126.com (Y. Li), liuji@msu.edu (J. Liu).

https://doi.org/10.1016/j.jclepro.2019.03.108
0959-6526/© 2019 Elsevier Ltd. All rights reserved.
376 Z. Xu et al. / Journal of Cleaner Production 224 (2019) 375e383

against renewable agricultural water resources to represent the water management (e.g., sown area, planned total crop yield, and
extent of water scarcity in agriculture (Raskin et al., 1996; Damkjaer permits of water use) is done at the county level in countries such
and Taylor, 2017). Crop water productivity refers to the amount of as China, India, and Japan.
crop produced per unit of water used. China is challenged to in- To fill this knowledge gap, we chose the North China Plain (NCP),
crease crop water productivity to relieve pressures that agriculture with 207 counties, as a demonstration for integrated assessment.
puts on water resources while increasing crop production (Wang The NCP is the national agricultural base and main grain production
et al., 2014). Evaluating water footprints presents a comprehen- area in China. The region includes the plain of Beijing, Tianjin City,
sive picture of the relationship between water consumption and Hebei Province, and part of Henan and Shandong provinces with
human appropriation, because a water footprint includes both 133 million people (Zhang et al., 2012). Approximately 80% of the
direct water consumption of products and water indirectly seeded areas of all crops are grain areas, 96% of which are planted
consumed and polluted during production. Assessing the impacts with winter wheat and summer maize (Wang et al., 2001). From
of water scarcity helps pinpoint vulnerable hotspots for solving the 1986 to 2010, the total wheat production and maize production in
problem. Exploring crop water productivity can facilitate under- the NCP had increased from 1.58 and 1.07 to 2.49 and 2.97 million
standing the trade-offs between food production and water con- tons, respectively. While the NCP needs water for agriculture, the
sumption. Holistically, understanding all three variables can available freshwater per capita annually in the plain e 302 m3 per
illuminate pathways to alleviate conflicts between water security year (Zhang et al., 2011) e is less than 1/24 of the global average.
and food security. This is far below the international standard of freshwater resource
Many studies have focused on water footprint, water scarcity shortage with the 1000 m3 threshold (Kang et al., 2013). Using such
and crop water productivity separately (Hoekstra and Mekonnen limited water resources to support large amounts of agricultural
2011, 2012; Jaramillo and Destouni, 2015; Zhao et al., 2015; production and socioeconomic development is a great challenge,
Ashraf Vaghefi et al., 2017; Sun et al., 2017). Hoekstra and implicating significant impacts on national food security, water
Mekonnen (2012) has quantified and mapped the water footprint security, and sustainable development. Many policies and tech-
of humanity with high spatial resolution and found that agricul- nology investments have been applied in the NCP to solve the water
tural production accounted for almost 92% of global WF footprint crisis and ensure sustainable water use for food production, but the
during 1996e2005 (Hoekstra and Mekonnen, 2012). Jaramillo et al. outcome has not been assessed comprehensively. Exploring this
(2015) studied the global effects of flow regulation and irrigation on problem in the NCP can have implications for not only China, but
global freshwater conditions and revealed that the two can raise also other irrigated areas worldwide.
the global water footprint of humanity by approximately 18% The aim of this study was to assess the water footprint, water
(Jaramillo and Destouni, 2015). Hoekstra and Mekonnen (2011) scarcity and crop water productivity of irrigated agriculture at the
defined the blue water scarcity index as the ratio of blue water county level in the NCP from 1986 to 2010. We calculated the blue,
footprint to blue water availability, and applied this index in the green, and grey water footprint to illustrate the dynamics of total
world's major river basins (Hoekstra and Mekonnen, 2011). They water footprint (WFtotal) in the whole NCP; applied the water
found that the blue water scarcity level in 55% of the basins studied scarcity index to study the impacts of water consumption from
exceeded 100% at least one month of the year, meaning the blue irrigated agriculture on water scarcity in each county; and
water footprint surpassed available blue water in these study ba- measured the grain yield per unit water use to represent crop water
sins. Zhao et al. (2015) used the water scarcity index to investigate productivity (Mekonnen and Hoekstra, 2011).
impacts of interprovincial virtual water flow on trading provinces'
water scarcity, and found the virtual water flow could exacerbate 2. Materials and methods
trading provinces' water scarcity level (Zhao et al., 2015). Ashraf
Vaghefi et al. (2017) assessed the crop water productivity of irri- 2.1. Data sources
gated maize and wheat in Karheh River Basin by using a hydro-
logical model and a river basin water allocation model (Ashraf We compiled a set of data for our analyses, including agro-
Vaghefi et al., 2017). Their results indicated a close linear rela- meteorological data, basic agricultural data, and geographic infor-
tionship between crop water productivity and yield. Sun et al. mation system (GIS) data. We obtained the agrometeorological data
(2017) explored crop water productivity of wheat in the Hetao from the Meteorological Data Sharing Service System of National
irrigation district at the field scale and analyzed the impacts of Meteorological Information Center of China. These data covered 69
agricultural and climatic factors on crop water productivity (Sun meteorological stations in Beijing, Tianjin, Hebei Province, Shan-
et al., 2017). Their results showed that crop water productivity dong Province, and Henan Province and included average air
was highly sensitive to relative humidity, wind speed, and irriga- temperature, maximum air temperature, minimum air tempera-
tion efficiency, while less sensitive to sunshine hours and the ture, hours of sunshine, and daily precipitation data from 1986 to
amount of fertilizers used. 2010. These factors were used to calculate the reference evapo-
To our knowledge, water footprint, water scarcity, and crop transpiration (ET0) based on Penman-Monteith equation (Xu et al.,
water productivity have not been assessed simultaneously at the 2017). The ET0 was used for calculating water footprint. We also
county level in large plains over a temporal scale. Such information used data on the crop growth periods, estimation of accumulated
is urgently needed since the global irrigated agricultural area has temperature, and solar radiation of winter wheat and summer
nearly tripled from 1900 to 2005 amid growing population, water maize from the cited literature, to define crop water production
crisis and food shortage. Assessing them together can show a more function in different areas. We obtained basic county-level agri-
comprehensive interrelationship among food production, water cultural production data e the cultivated area, nitrogen use,
consumption, and water scarcity. This will help to construct tar- amount of production of winter wheat and summer maize e from
geted policies to achieve both food security and water security in the Agricultural Information Institute of Chinese Academy of
irrigated agriculture. Different from most water footprint studies at Agricultural Sciences, to help calculate effective rainfall and grey
coarse spatial scales (e.g., global and national scales) or focused on water footprint, and to explore the relationship between crop
geographic units (e.g., 5'  50 or 30'  300 grid), a study at the county production and water footprint. The empirically measured data of
level helps to better understand and manage water conservation ETc (crop evapotranspiration) were derived from Luancheng Agro-
and food production because much of agricultural planning and Eco-Experimental Station of the Chinese Academy of Sciences in
 Z. Xu et al. / Journal of Cleaner Production 224 (2019) 375e383 377

Shijiazhuang City, Hebei Province. We also acquired the digitized hygrometer constant.
soil organic matter map data (at a scale of 1:14,000,000) in 2005 For our proposed water consumption method, the actual crop
from the Data Center for Resources and Environmental Sciences of evapotranspiration in Eq. (1) was calculated from the crop water
Chinese Academy of Sciences, to help define crop water functions in production function (CWPF) shown below.
different areas. Furthermore, we received GIS shape files for prov-
inces, counties, main cities, and the Yellow River. Our unit of y ¼ aETa2 þ bETa þ c (4)
analysis was the county. For agrometeorological data that were not 0sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1
at the county level, we used the ordinary Kriging method to b y b2 c
interpolate data at agrometeorological stations to counties. Spe- ETa ¼ min@
± þ 2
A (5)
2a a 4a a
cifically, we converted vector data of stations into raster data and
then calculated the sum value by zonal statistics for each county in
ArcGIS (version 10.1 ESRI). Crop Water Production Function (CWPF) Where a,b,c are regression coefficients; and y(kg/ha) is unit area
is the mathematical expression that describes the relationship crop yield.
between water use and crop production for a certain kind of crop. Considering that the actual crop water consumption might
The function is mainly influenced by sunshine and heat factors such differ from the estimated amount in the conventional water
as photo-synthetically active radiation and effective accumulated requirement method, we proposed a new water consumption
temperature, and agricultural production factors such as soil method based on the crop water production function and
organic matter, crop types and varieties (see details in Supple- compared it with the conventional water requirement method. The
mentary Information). test showed that on average there was no significant difference
between the estimation from the proposed water consumption
method and the actual measurement of water use. However, the
2.2. WF assessment
estimation from the conventional water requirement method was
significantly higher than that of the water consumption method or
We assessed the WF for the entire grain production chain, which
the actual measurement at the Luancheng monitoring station as
included both the consumptive water usage for crop growth
shown in Supplementary Fig. 1.
(WFcons) and the fresh water needed to dilute associated pollutants
(WFgrey). According to the sources of water, WFcons were further
2.2.2. WFblue and WFgreen
divided into WFblue (the volume of surface water, shallow and deep
WFblue is the volume of consumed surface water and ground-
groundwater used for irrigation) and WFgreen (the volume of rain-
water to produce goods or delivering services. WFgreen is the vol-
water used for growing crops). More detailed procedures for the
ume of consumed rainwater during the production process. WFgreen
WF assessment methods (including all calculation equations) can
is particularly relevant for agricultural and forestry products,
be found in Supplementary Information.
including the total rainwater evapotranspiration (from fields and
plantations) plus the water incorporated into the harvested prod-
2.2.1. WFcons ucts. The WFblue and WFgreen were calculated by the following
WFcons of crop production is the total actual consumption of equations (Zhang et al., 2008):
water within its whole production chain. Often, it is difficult to
directly measure WFcons, thus the indirect water requirement WFcons ¼ WFblue þ WFgreen (6)
method is used. The crop water requirement is assumed to be the
needed water via crop evapotranspiration under optimal condi- ETblue  A  B
tions, which is calculated by multiplying the reference ETc with a WFblue ¼ (7)
Y
crop coefficient. Because actual crops are not always grown under
optimal conditions, actual evapotranspiration should be less than  
optimal crop evapotranspiration and thus a water stress coefficient ETblue ¼ max 0; ETa
Peff (8)
is introduced. The main factors that affected crop evapotranspira-
tion include precipitation, air temperature, pressure, sunshine ETgreen  A  B
hours, wind speed, crop type, soil condition, and planted time. The WFgreen ¼ (9)
Y
calculation functions are given below (Hoekstra et al., 2011).
 
1000  ET0  A ETgreen ¼ min ETa ; Peff (10)
WFcons ¼ (1)
Y
Peff ¼ sP (11)
ETc ¼ Kc  ET0 (2)
Where ETblue (mm) and ETgreen (mm) are evapotranspiration of blue
900 U ðe
e Þ
0:408dðRn
GÞ þ g Tþ273 and green water, respectively; Peff (mm) and P (mm) are effective
2 s a
ET0 ¼ (3) rainfall and total rainfall within crop growth period, respectively;
d þ gð1 þ 0:34U2 Þ
and s is the effective utilization coefficient of rainfall.
Where ETa (mm) is the actual crop evapotranspiration; A(km2) is
the total planation area; Y (kg) is the total crop yield; Kc is crop 2.2.3. WFgrey
coefficient comparing to reference crop evapotranspiration; ET0 The WFgrey is an indicator of freshwater pollution that is asso-
(mm) is reference crop evapotranspiration; Rn (MJ m
2 d
1) is net ciated with a product over its full production chain. It is calculated
radiation on surface of crop; G (MJ m
2 d
1) is soil heat flux; T ( C) as the volume of water required to dilute pollutants to meet water
is average air temperature; U2 (m s
1) is wind speed at 2 m above quality standards. We focused on the WFgrey of nitrogen because
ground; es (kPa) is saturation vapor pressure; ea (kPa) is measured fertilizers were used intensively in the NCP and potentially caused
vapor pressure; d (kPa  C
1) is the slope of the curve between the most severe pollution since nitrogen can easily be transported
saturation vapor pressure and temperature; and g (kPa  C
1) is in soil, surface water, and groundwater (Sun et al., 2018). Soil
378 Z. Xu et al. / Journal of Cleaner Production 224 (2019) 375e383

phosphorus often easily generates chemical reactions with other the change in crop water productivity.
soil minerals and produces chemical compounds that are not
readily soluble, resulting in less pollution. Potassium ions are 2.5. Statistics and mapping
attracted by soil colloids and thus not easily migrated. Therefore,
the pollution from phosphorus and potassium fertilizers can be To test whether or not WF and crop water productivity changed
ignored when assessing WFgrey. Supplementary Fig. 3 shows that significantly over time, we performed the statistical significance
the nitrogen application amount in the NCP, with an average test using the software SPSS Statistics 20 (Statistical Product and
amount of 435.99 kg ha
1 and a range from 179.76 to Service Solutions, IBM, USA). When P value < 0.05, it indicates a
879.11 kg ha
1. The spatial distributions of fertilizer application for significant change. We acquired GIS shape files for the NCP
winter wheat and summer maize are similar; however, on average, counties. We created the map of our study areas and mapped all the
the nitrogen application amount of winter wheat (223.56 kg ha
1) WF, water scarcity and productivity at the county level in ArcGIS.
is higher than that of summer maize (212.43 kg ha
1). The calcu-
lation functions for different types of grey WF are shown below: 3. Results

ðatotal  ARÞ=ðCmax
Cnat Þ Our results show the annual water footprint from irrigated
WFgrey ¼ (12)
y agriculture increased in almost all counties (Fig. 1). The southeast
NCP had a larger water footprint and the central part had a smaller
Where atotal is the total leaching fraction, measured to be 25.0% water footprint than other places in the NCP (Fig. 1). Also, the water
(Zhao et al., 2009); asurf and aground are run-off and leaching frac- footprint of southeast NCP increased most while that of the central
tions of applied chemicals for surface water and groundwater part increased the least over time (Fig. 1).
respectively, measured as 9.6% and 15.4%, respectively (Xu et al., The annual water footprint in all counties together increased
2013); AR (kg ha
1) is application amount of chemical fertilizers from 53 billion m3 in 1986 to 78 billion m3 in 2010 (Fig. 2a). For the
per hectare; and Cmax (g L
1) and Cnat (g L
1) are the maximum total amount of different types of WF, overall, there were statisti-
acceptable concentration and natural concentration of the chemical cally significant increases in WFtotal (F ¼ 17.97, p ¼ 0.0003), WFgreen
fertilizer, respectively. (F ¼ 22.17, p ¼ 0.0001), and WFgrey (F ¼ 21.88, p ¼ 0.0001) over time
(Fig. 2a; Table 1). The WFblue gradually increased from 1986, peaked
2.3. Water scarcity in 1997, then started to decline to the valley in 2003, and kept
relatively stable between 2004 and 2010 (Fig. 2a). Its overall tem-
The water scarcity index of grain production from a WF poral trend was not statistically significant (F ¼ 0.31, p ¼ 0.5821;
perspective can be reflected through the ratio of agricultural water Table 1). There are some potential reasons for the dynamics of the
use to renewable agricultural water resources (Itotal). The higher the WFblue. During 1986e1997, the rapid development of agriculture
water scarcity index, the less sustainable water use for grain pro- led to the WFblue increase. But during 1998e2003, the water-saving
duction. The water scarcity index can be calculated as follows: policies were implemented in the NCP to reduce planting area and
restrict the use of underground water and thus reduced the WFblue
Itotal ¼ WFgrain, total / WRagri, total (13) (Xie and Zhang, 2007; Liu et al., 2008; Hu et al., 2017). After 2004,
the increasing demand for crop production in the NCP compen-
Where Itotal is water scarcity due to agricultural use, Itotal >1 sated for the effects of water-saving policies (Xie and Zhang, 2007;
indicates water scarcity and Itotal >2.5 indicates severe water scar- Liu et al., 2008; Hu et al., 2017).
city due to grain production. WFgrain, total is the total WF for winter Except for annual WFblue, the whole NCP's annual irrigated
wheat and summer maize here; and WRagri, total refers to the agriculture WF, WFgrey, and WFgreen increased due to the overall
renewable agricultural water resources. increase of total crop production over time (Fig. 2b). By comparing
Fig. 2a and b, it is easy to observe that the temporal dynamics of
2.4. Crop water productivity total crop production (either winter wheat or summer maize) were
similar to the dynamics of WF of irrigated agriculture in NCP. The
Crop water productivity refers to the amount of crop produced overall temporal dynamics of different sources of WFblue fluctuated
per unit of water used. We divided the amount of crop production (Fig. 2c).
by its corresponding water footprint in each county in 1986 and Irrigated agriculture led to water scarcity in all evaluated
2010 to get the crop water productivity at the county level over counties (use intensity > 1 indicates water scarcity) (Fig. 3). Among
time. We also divided crop production by its corresponding water the NCP's 207 counties, 174 counties faced severe water scarcity
footprint in the whole NCP from 1986 to 2010 to obtain the average (use intensity > 2.5 indicates severe water scarcity). Our results
crop water productivity for the whole plain over time. To figure out showed that the average water scarcity for total available water for
to what extent increasing crop water productivity reduces the agricultural use was as high as 10.14, indicating an unsustainable
water footprint and water scarcity, we set the crop water produc- water usage pattern. There were 95.29% counties with severe water
tivity in 1986 (CWP1986) as the baseline, and recalculated the water scarcity over 5.0 for total available water for agricultural use.
scarcity and water footprint during 1987e2010 by multiplying the Overall, there were 46.5% of counties with total agricultural water
amount of crop production during 1987e2010 with the CWP1986 to use intensity over 10.0 for grain production, covering almost the
get the recalculated WF. Then we divided the recalculated WF from entire east of NCP (Fig. 3).
1987 to 2010 by the renewable agricultural freshwater resource to Crop water productivity increased in all counties, suggesting an
get the recalculated water scarcity from 1987 to 2010 (Note: the irony with rising water productivity coupled with severe water
amount of the renewable freshwater resource was kept constant scarcity (Figs. 4e5). The average crop water productivity increased
during 1987e2010 since we used the average renewable water from 0.90 kg m
3 in 1986 to 1.94 kg m
3 in 2010. The central and
resource value across years. The change in water scarcity was western parts of the NCP had higher crop water productivity while
determined by the change of WF). Then we compared the original its eastern part had lower crop water productivity. The central
WF and water scarcity with the recalculated WF and water scarcity part's crop water productivity increased the most, while the
to calculate the percent decrease in WF and water scarcity due to eastern crop water productivity increased the least.
 Z. Xu et al. / Journal of Cleaner Production 224 (2019) 375e383 379

Fig. 1. Spatial dynamics of total water footprint (billion m3) in irrigated agriculture from 1986 (a) to 2010 (b).

4. Discussion USD and annual transfer amount of 48.4 billion m3 water) has
mixed impacts. By transferring physical water from southern China
We find the increasing water footprint e worsening water to northern China, the SNWTP can help alleviate the water short-
scarcity while crop water productivity increased e in all 207 ages in northern China and indirectly enhance national food se-
counties of the North China Plain over 1986e2010. The results curity. But the environmental cost of SNWTP is also large (Yin et al.,
show that the improving crop water productivity had increasingly 2001; Shao et al., 2003; Zhang, 2009; Liu et al., 2016). Therefore, for
positive influences on reducing WF and water scarcity over time the NCP and China as a whole, the long-term water management
(Fig. 6). In 1987, an increase in crop water productivity dropped WF strategies should target controlling and reducing total water use by
and water scarcity 14.5%, and this number increased to 53.7% in improving water use efficiency rather than constructing more en-
2010 (Fig. 6). However, the total grain production WF was still high gineering projects to support the seemingly endless demand for
and led to water scarcity in all evaluated counties. The underlying water.
reasons for the persistence of water scarcity conditions include the There are also many other specific measures to reduce total
soaring grain production in the NCP driven by rapid economic water consumption from grain production in the NCP. For instance,
development and a growing national population, extensive decline the total WF of NCP can be reduced by importing grain and other
in arable areas in south China (Song et al., 2007). With the growing food products from water-abundant countries and reducing the
population, an increasing water crisis and anticipated food short- cultivation area in the NCP. One way to reduce total WF is to
ages in the future, the conflict in irrigated agriculture could be mitigate WFgrey, which is the highest priority since it accounts for a
exacerbated, posing threats to national sustainability. large percentage of the total WF. The primary reasons for high
Spatial variations in WF of irrigated agriculture across the NCP WFgrey in the NCP are overuse and low efficiency in applying
reveals hotspot areas requiring special management. For example, chemical fertilizers and pesticides. For example, the average per
the southeast part of the NCP showed higher WF than other parts, unit area amount of nitrogen applied (545 kg ha
1) in a wheat-
therefore more agricultural water management should be planned maize rotation system in the NCP during 1997e2005 was much
for this area. Southeast NCP produced more crops due to its higher higher than the nitrogen output within harvested crops of system
accumulated temperature, greater precipitation, and better soil (311 kg ha
1) (Zhao et al., 2009), meaning some nitrogen ended up
organic matter than other areas in the NCP (Foster et al., 2004). polluting rather than boosting crop growth. Many studies (Ju et al.,
Furthermore, because of its ineffective agricultural production 2003; Zhang et al., 2006; Zhang, 2011) suggest that the applied
management, excessive fertilization and waste of water, the amount of fertilizers and their use efficiency is negatively corre-
southeast's WF was much higher than that in other areas of the lated, and thus controlling the use of fertilizers and improving their
NCP. On the other hand, the comparatively slow-growing crop use efficiency are complementary. Using straw, livestock manure,
productions WF in the central NCP was less than other parts of NCP. biogas waste, and organic fertilizers instead of chemical fertilizers
Since 1980, the government controls the agricultural production in can not only reduce the applied amount of chemical fertilizers but
the central part of the NCP to limit the groundwater exploitation also increase crop yield in the NCP (Zhang et al., 2006). Many
because a groundwater funnel emerged (Wang et al., 2015). nutrient management techniques, such as balanced fertilization,
Moreover, since the precipitation in the central part of the NCP is soil testing and formulated fertilization, application of slow-release
smaller than that in other regions, the WF is much smaller because fertilizers, and selection of fertilization timing, can also improve the
only this available precipitation is consumed. use efficiency (Zhang et al., 2006; Quin ~ ones et al., 2007; Zhang,
Water transfer projects such as the South-North Water Transfer 2011).
Project (Liu and Yang, 2012) (SNWTP; the largest water transfer And crop production conditions can be altered (e.g., cultivar,
project in the world with a planned total investment of $80 billion water use efficiency, irrigation, and tillage methods) to change the
380 Z. Xu et al. / Journal of Cleaner Production 224 (2019) 375e383

Fig. 3. Spatial dynamics of water scarcity in irrigated agriculture from the water
footprint perspective. Index greater 1 indicates unsustainable water use.

Fig. 2. Temporal dynamics of different types of water footprint and annual total grain
yield from 1986 to 2010. (a) The different types of grain production water footprint; (b)
the total annual grain yield; and (c) the different sources of blue water footprint. (For
Fig. 4. Temporal dynamics of average crop water productivity (kg/m3) in the NCP from
interpretation of the references to colour in this figure legend, the reader is referred to
1986 to 2010.
the Web version of this article.)

Table 1
Temporal trend analyses of grain production water footprint from 1986 to 2010 based on regression lines.

WFtotal WFblue WFgreen WFgrey

Year 0.843*** (0.181)
0.021 (0.027) 0.161*** (0.023) 0.703*** (0.164)

Constant
1624.341*** (361.346) 47.299 (53.088)
306.696*** (45.378)
1364.944*** (327.980)
F statistics 21.65 0.62 49.89 18.29
R-Squared 0.44 0.01 0.49 0.49
N 25 25 25 25

Notes: Dependent variables are different types of grain production water footprint (109 m3) in average values for the 207 analyzed counties, respectively. Numbers outside and
inside parentheses are coefficients and robust standard errors, respectively. *p < 0.05; **p < 0.01; ***p < 0.001.
 Z. Xu et al. / Journal of Cleaner Production 224 (2019) 375e383 381

Fig. 5. Spatial dynamics of crop water productivity (kg/m3) in irrigated agriculture from 1986 (a) to 2010 (b).

improve soil moisture holding capacity, reduce ground infiltration,

and increase water use efficiency (Salem et al., 2015). Greenhouse
agriculture (e.g., covering the field with plastic films is much more
common than glass greenhouses in China) also helps reduce water
evaporation and water use (Chang et al., 2011).
Our work provides the first detailed and integrated assessment
that analyzes water footprint, water scarcity, and crop water pro-
ductivity at the county level in a large plain over long term. It re-
veals the serious unsustainable water use across all counties in the
NCP. The spatial variations of unsustainable water use, water
footprint, and crop productivity are disclosed. This information can
help the government make more holistic and better-targeted pol-
icies to manage crop production and water consumption more
sustainably in China's major crop production region. Future
research can focus on the interactions between irrigated agricul-
Fig. 6. Percent decrease in water footprint and water scarcity in the NCP due to in- ture in the NCP and the environmental and socioeconomic devel-
creases in crop water productivity from 1986 to 2010.
opment in the rest of China. Since much of the NCP harvest was
transferred to the rest of China to enhance food security, and since
WF. The WF can be reduced by increasing per unit area crop yield or much water was diverted from southern China through SNWTP to
decreasing actual crop evapotranspiration. Field experiments the NCP to alleviate the water shortage, the interactions between
confirm that using high yield cultivar improved crop water pro- these two systems are complex and have great impacts on both
ductivity and reduced water consumption (Zhang et al., 2010). systems. Cross-boundary studies can help get a comprehensive
Many techniques below are also documented to improve water use picture of drivers behind water use and thus provide holistic in-
efficiency and thus reduce actual crop evapotranspiration. formation for policy-making, therefore facilitating sustainable
Currently, the common irrigation approach in the NCP is still sur- development and improvement of human well-being (Liu 2017,
face irrigation with very low use efficiency of both water and fer- 2018).
tilizers. The combination of integrated irrigation and fertilization
technique with efficient water-saving irrigation systems (e.g.,
5. Conclusions
sprinkler irrigation, micro-irrigation) can reduce surface erosion,
retain fertilizers in the crop root zone, mitigate fertilizers leaching
In this paper, we quantified water footprint, crop water pro-
into underground (Liu and Kang, 2006; Man et al., 2014), and thus
ductivity, and water scarcity from irrigated agriculture in China's
reduce both WFgrey and WFblue. Research shows that the deficit
major crop production region, the North China Plain's 207 counties,
irrigation approach and appropriate reduction of irrigation times
from 1986 to 2010. Our results indicated that even though crop
for winter wheat can maintain or only slightly reduce crop yield but
water productivity grew over time, the water footprint in the NCP
largely increase water use efficiency (Yang et al., 2006; Zhang et al.,
due to crop production increased sharply from 53 billion m3 in
2008). Covering the soil with straw can reduce soil evaporation
1986 to 78 billion m3 in 2010, leading to water scarcity in all 207
while increasing rainfall infiltration and reducing surface runoff (Li
counties. This study revealed the unsustainable state of irrigated
et al., 2013). The use of soil tillage and subsoil tillage methods can
agriculture in China's major crop production region, which has
382 Z. Xu et al. / Journal of Cleaner Production 224 (2019) 375e383

implications for global irrigated agriculture at the county level. The of China. Agric. Water Manag. 116, 39e49.
Liu, H.-J., Kang, Y., 2006. Effect of sprinkler irrigation on microclimate in the winter
irrigated agriculture enhanced food security but increased the
wheat field in the North China Plain. Agric. Water Manag. 84, 3e19.
pressure on water use. There is tremendous pressure on water re- Liu, J., 2017. Integration across a metacoupled world. Ecol. Soc. 22.
sources due to a huge food demand on the NCP under the context of Liu, J., 2018. An integrated framework for achieving Sustainable Development Goals
a growing national population. Changing high water consumption around the world. Ecol. Econ. Soc. 1, 11e17.
Liu, J., Yang, W., 2012. Water sustainability for China and beyond. Science 337,
cropping systems, developing efficient water-saving irrigation 649e650.
technology and reducing cropland area should be considered for Liu, J., Zehnder, A.J., Yang, H., 2009. Global consumptive water use for crop pro-
the future agricultural management to help ensure water security duction: the importance of green water and virtual water. Water Resour. Res.
45, W05428.
and food security simultaneously. Liu, J.G., Yang, W., Li, S.X., 2016. Framing ecosystem services in the telecoupled
Anthropocene. Front. Ecol. Environ. 14, 27e36.
Liu, S.H., Cheng, W.Q., Song, S.G., 2008. Development trend and prospect of agri-
Acknowledgements
cultural water saving irrigation in the mountainous area of Hebei province.
Water saving irrig. 5, 47e50.
We acknowledge insightful editorial comments and careful Man, J., Yu, J., White, P.J., Gu, S., Zhang, Y., Guo, Q., Shi, Y., Wang, D., 2014. Effects of
language edits from Sue Nichols. We are grateful for financial supplemental irrigation with micro-sprinkling hoses on water distribution in
soil and grain yield of winter wheat. Field Crop. Res. 161, 26e37.
support from the National Science Foundation, Michigan State Mekonnen, M.M., Hoekstra, A.Y., 2011. National water footprint accounts: the green,
University, Michigan AgBioResearch, China Scholarship Council and blue and grey water footprint of production and consumption. In: Value of
National Natural Science Foundation of China (51621061, Water Research Report Series No.50. UNESCO-IHE.
Mekonnen, M.M., Hoekstra, A.Y., 2016. Four billion people facing severe water
51321001). scarcity. Sci. Adv. 2 e1500323.
Quin~ ones, A., Martínez-Alc antara, B., Legaz, F., 2007. Influence of irrigation system
Appendix A. Supplementary data and fertilization management on seasonal distribution of N in the soil profile
and on N-uptake by citrus trees. Agric. Ecosyst. Environ. 122, 399e409.
Raskin, P.D., Hansen, E., Margolis, R.M., 1996. Water and sustainability: global pat-
Supplementary data to this article can be found online at terns and long-range problems. In: Natural Resources Forum, vols. 1e15. Wiley
https://doi.org/10.1016/j.jclepro.2019.03.108. Online Library.
Salem, H.M., Valero, C., Mun  Rodríguez, M.G., Silva, L.L., 2015. Short-term
~ oz, M.A.,
effects of four tillage practices on soil physical properties, soil water potential,
Author contributions and maize yield. Geoderma 237, 60e70.
Shao, X., Wang, H., Wang, Z., 2003. Interbasin transfer projects and their implica-
tions: a China case study. Int. J. River Basin Manag. 1, 5e14.
Z.X., Y.L. and J.L. designed the research; Y.L., and X.C. contributed Song, Z., Sun, H., Dong, J., 2007. Solving China's North-South Grain Transfer Problem.
data; X.C., M.G., S.R.W., Y.D, J.W. and J.L. provided comments on the Steffen, W., Richardson, K., Rockstro €m, J., Cornell, S.E., Fetzer, I., Bennett, E.M.,
manuscript; Z.X., and X.C. analyzed the data and wrote the manu- Biggs, R., Carpenter, S.R., de Vries, W., de Wit, C.A., Folke, C., Gerten, D.,
Heinke, J., Mace, G.M., Persson, L.M., Ramanathan, V., Reyers, B., So € rlin, S., 2015.
script. All authors reviewed the manuscript. Planetary boundaries: guiding human development on a changing planet.
Science 347, 1259855.
Sun, J., Mooney, H., Wu, W.B., Tang, H.J., Tong, Y.X., Xu, Z.C., Huang, B.R., Cheng, Y.Q.,
Competing financial interests
Yang, X.J., Wei, D., Zhang, F.S., Liu, J.G., 2018. Importing food damages domestic
environment: evidence from global soybean trade. Proc. Natl. Acad. Sci. U.S.A.
The authors declare no competing financial interests. 115, 5415e5419.
Sun, S., Zhang, C., Li, X., Zhou, T., Wang, Y., Wu, P., Cai, H., 2017. Sensitivity of crop
water productivity to the variation of agricultural and climatic factors: a study
References of Hetao irrigation district, China. J. Clean. Prod. 142, 2562e2569.
Viala, E., 2008. Water for food, water for life a comprehensive assessment of water
Ashraf Vaghefi, S., Abbaspour, K.C., Faramarzi, M., Srinivasan, R., Arnold, J.G., 2017. management in agriculture. Irrig. Drain. Syst. 22, 127e129.
Modeling crop water productivity using a coupled SWATeMODSIM model. Wang, D., Zhang, G., Feng, H., Tian, Y., Wang, J., 2015. Spatial variation of ground-
Water 9, 157. water guarantee extent in typical well irrigation area of North China Plain.
Chang, J., Wu, X., Liu, A.Q., Wang, Y., Xu, B., Yang, W., Meyerson, L.A., Gu, B.J., South-to-North Water Transfers Water Sci. Technol. 13, 622e625 (in Chinese).
Peng, C.H., Ge, Y., 2011. Assessment of net ecosystem services of plastic Wang, H., Zhang, L., Dawes, W., Liu, C., 2001. Improving water use efficiency of
greenhouse vegetable cultivation in China. Ecol. Econ. 70, 740e748. irrigated crops in the North China Plaindmeasurements and modelling. Agric.
Damkjaer, S., Taylor, R., 2017. The measurement of water scarcity: defining a Water Manag. 48, 151e167.
meaningful indicator. Ambio 46, 513e531. Wang, Y., Wu, P., Engel, B., Sun, S., 2014. Application of water footprint combined
Do€ ll, P., 2009. Vulnerability to the impact of climate change on renewable with a unified virtual crop pattern to evaluate crop water productivity in grain
groundwater resources: a global-scale assessment. Environ. Res. Lett. 4, 035006. production in China. Sci. Total Environ. 497, 1e9.
Food and Agriculture Organization of the United Nations, 2018. http://www.fao.org/ Ward, F.A., Pulido-Velazquez, M., 2008. Water conservation in irrigation can in-
nr/water/aquastat/water_use/index.stm. crease water use. Proc. Natl. Acad. Sci. Unit. States Am. 105, 18215e18220.
Foster, S., Garduno, H., Evans, R., Olson, D., Tian, Y., Zhang, W., Han, Z., 2004. Qua- Xie, J.F., Zhang, J.G., 2007. Groundwater development model in the guantao county
ternary aquifer of the North China Plaindassessing and achieving groundwater of Hebei province. Gr. Water 6, 51e53.
resource sustainability. Hydrogeol. J. 12, 81e93. Xu, P., Hou, X., Zhou, F., Ying, N., Zhao, G., 2013. Estimation of nitrate leaching rate
Godfray, H.C.J., Beddington, J.R., Crute, I.R., Haddad, L., Lawrence, D., Muir, J.F., and loading deduced by agricultural fertilizers in the North China Plain. Acta
Pretty, J., Robinson, S., Thomas, S.M., Toulmin, C., 2010. Food security: the Scientiae Circumstantiae 33, 3173e3180.
challenge of feeding 9 billion people. Science 327, 812e818. Xu, J., Liu, X., Yang, S., Qi, Z., Wang, Y., 2017. Modeling rice evapotranspiration under
Hoekstra, A.Y., Mekonnen, M., 2011. Global Water Scarcity: the Monthly Blue Water water-saving irrigation by calibrating canopy resistance model parameters in
Footprint Compared to Blue Water Availability for the World's Major River the Penman-Monteith equation. Agric. Water Manag. 182, 55e66.
Basins. Yang, Y., Watanabe, M., Zhang, X., Zhang, J., Wang, Q., Hayashi, S., 2006. Optimizing
Hoekstra, A.Y., Mekonnen, M.M., 2012. The water footprint of humanity. Proc. Natl. irrigation management for wheat to reduce groundwater depletion in the
Acad. Sci. U. S. A. 109, 3232e3237. piedmont region of the Taihang Mountains in the North China Plain. Agric.
Hu, H.J., Wu, X.M., Qi, C.W., 2017. Using multi-objective optimization to restructure Water Manag. 82, 25e44.
agriculture with consideration of the constrains on groundwater extraction: Yin, K., Yuan, H., Liao, Q., Ruan, Y., 2001. Impact of middle route SN water transfer
taking hengshui as an example. J. Irrig. Drain. 10, 95e99. project on algae bloom in middle & lower reach of hanjiang river. Yangtze River
Jaramillo, F., Destouni, G., 2015. Local flow regulation and irrigation raise global 32, 31e50.
human water consumption and footprint. Science 350, 1248e1251. Zhang, F., 2011. Soil Testing and Fertilization Recommendation, vol. 296. China
Ju, X., Pan, J., Liu, X., Zhang, F., 2003. Study on the fate of nitrogen fertilizer in winter Agricultural University Press, Beijing.
wheat/summer maize rotation system in Beijing suburban. Plant Nutr. Fert. Sci. Zhang, F., Ma, W., Chen, X., 2006. Introduction to the Theory and Technology of
9, 264e270. Integrated Nutrient Management, vol. 329. China Agricultural University Press,
Kang, S., Yang, J., Pei, Y., 2013. Farmland Water Cycle and High-Efficient Water Use Beijing.
Mechanism in Agriculture of the Haihe River Basin, vol. 329. Science Press, Zhang, G., Fei, Y., Wang, J., Yan, M., 2012. Study on the Adaptativity of Groundwater
Beijing. and Irrigated Agriculture in North China Plain. Science Press, Beijing.
Li, S., Wang, Z., Li, S., Gao, Y., Tian, X., 2013. Effect of plastic sheet mulch, wheat Zhang, G., Lian, Y., Liu, C., Yan, M., Wang, J., 2011. Situation and origin of water
straw mulch, and maize growth on water loss by evaporation in dryland areas resources in short supply in North China Plain. J. Earth Sci. Environ. 33, 172e176.
 Z. Xu et al. / Journal of Cleaner Production 224 (2019) 375e383 383

Zhang, Q., 2009. the South-to-North Water transfer project of China: environmental Manag. 97, 1117e1125.
implications and monitoring strategy. J. Am. Water Resour. Assoc. 45, Zhao, R., Chen, X., Zhang, F., 2009. Nitrogen cycling and balance in winter-wheat-
1238e1247. summer-maize rotation system in Northern China Plain. Acta Pedol. Sin. 46,
Zhang, X., Chen, S., Sun, H., Pei, D., Wang, Y., 2008. Dry matter, harvest index, grain 684e697.
yield and water use efficiency as affected by water supply in winter wheat. Irrig. Zhao, X., Liu, J., Liu, Q., Tillotson, M.R., Guan, D., Hubacek, K., 2015. Physical and
Sci. 27, 1e10. virtual water transfers for regional water stress alleviation in China. Proc. Natl.
Zhang, X., Chen, S., Sun, H., Wang, Y., Shao, L., 2010. Water use efficiency and Acad. Sci. Unit. States Am. 112, 1031e1035.
associated traits in winter wheat cultivars in the North China Plain. Agric. Water