sustainability

Article
Comprehensive Assessment of Water Footprints and
Water Scarcity Pressure for Main Crops in Shandong
Province, China
Mengran Fu 1,† , Bin Guo 1, * , Weijiao Wang 1,† , Juan Wang 2,† , Lihua Zhao 3 and Jianlin Wang 4, *
1 College of Geomatics, Shandong University of Science and Technology, Qingdao 266590, China;
fumengran41@hotmail.com (M.F.); wwjsmz@163.com (W.W.)
2 College of Science and Information, Qingdao Agricultural University, Qingdao 266109, China;
juanwang@qau.edu.cn
3 Hebei Institute of Water Science, Shijiazhuang 050057, China; 13932156919@163.com
4 College of Agronomy, Qingdao Agricultural University, Qingdao 266109, China
* Correspondence: guobin07@mails.ucas.ac.cn (B.G.); wangjianlin@qau.edu.cn (J.W.);
Tel.: +86-532-8068-1170 (B.G.)
† These authors contributed equally to this work.




Received: 31 January 2019; Accepted: 23 March 2019; Published: 28 March 2019


Abstract: Rapid economic development has posed pressure on water resources, and the potential
for a water crisis has become an important obstacle to the sustainable development of society.
Water footprint theory and its applications in agriculture provide an important strategic basis for
the rational utilization and sustainable development of water resources. Based on the monthly
meteorological observations and agricultural data of Shandong Province, CROPWAT 8.0 and
Geographic Information System (GIS) technology, the green, blue and grey water footprints of
wheat, maize, cotton and groundnut from 1989 to 2016 were calculated and the spatial variations
of water footprints for crops in different rainfall years were analyzed. Additionally, assessment
of water stress for agricultural productions was conducted in this study. The results showed that
the average water footprints of wheat, maize, cotton and groundnut were 2.02 m3 /kg, 1.24 m3 /kg,
7.29 m3 /kg and 1.75 m3 /kg, respectively in Shandong Province. A large amount of the average total
water footprint was calculated for wheat (420.59 × 108 m3 /yr), maize (222.16 × 108 m3 /yr), cotton
(72.70 × 108 m3 /yr) and groundnut (50.07 × 108 m3 /yr). The average total water footprint of the
four crops was 765.52 × 108 m3 /yr (29.98% blue) and exhibited a gradual decreasing trend over time.
Specifically, the total water footprint of wheat was the highest among four main crops in Shandong
Province and exhibited a decreasing trend during 1989–2016. The maize was ranked in the second
place, and was the only crop still increasing rapidly. The spatial and temporal changes of water
footprints for crops were obvious in different rainfall years. Additionally, agricultural productions
in most regions were facing the threat of water scarcity. Therefore, the scientific planning of crop
planting structures and rational control of sown areas of crops with large water footprints should
be implemented in severely water-scarce regions. This study can give some suggestions on the
adjustment of planting structure for the sustainable development of agriculture and the realization of
efficient utilization of water resources.

Keywords: water footprint; virtual water; water scarcity; spatiotemporal variation

1. Introduction
Water is a key element to human sustenance and socioeconomic activities [1]. However, water
scarcity is one of the major problems facing many societies across the world [2]. More than 2.3 billion

Sustainability 2019, 11, 1856; doi:10.3390/su11071856 www.mdpi.com/journal/sustainability

Sustainability 2019, 11, 1856 2 of 18

people will live in severely water-stressed environment by the year 2050 [3]. In order to alleviate the
shortage of water resources, the concept of water footprint (WF) [4] was proposed to quantify water
resources used. The water footprint founded on the concept of “virtual water, VW” [2] is recognized
as a suitable indicator of human occupying freshwater resources and is becoming widely applied to
get better awareness of sustainable water use [5,6]. The WF of crop production is the total amount of
freshwater that is consumed and used for diluting pollutants during the crop growing process, and it
includes green, blue and grey WFs [2,7,8].
WFs within the agricultural sector have been extensively studied. Most studies were mainly
focusing on the WF of crop production, at scales from an irrigation district [9–11], a city level
region [11–17], a river basin [12,18–22], and a country [23–30], to the global perspective [31–35]. Among
previous studies, Gobin et al. [36] conducted a calibrated model on wheat, barley, grain maize, oilseed
rape, potato and sugar beet. They found that the WF of cereals could be up to 20 times larger than
the WF of tuber and root crops; the largest share was attributed to the green WF. Chapagain et al. [37]
assessed the water footprint of worldwide cotton consumption. The worldwide consumption of cotton
products demanded 256 Gm3 water per year, out of which about 42% was blue WF, 39% green WF
and 19% dilution WF. Chen et al. [38] utilized an interregional input-output model to estimate the WF
of each province in China and to quantify the inter-provincial transfer of virtual water. The results
showed that provinces with larger populations and greater Gross Domestic Product (GDP) had larger
WFs, and developed regions had higher proportions of external WFs. Landon and Megan [39] analyzed
drought impacts on the WF from pre-drought conditions (2011) through three years of exceptional
drought (2012–2014). The results indicated that drought may strengthen the telecoupling between
groundwater withdrawals and distant consumers of agricultural commodities. WF is now mostly
being studied at country-level and calculated using average annual climate data provided by FAO’s
CLIMWAT database.
Agricultural water use varies widely from country to country. At the global scale, it is responsible
for about 70% of the human water use [40]. In China, agricultural water use is about 3770 × 108 m3 ,
accounting for 62.32% of the total water consumption [41]. Shandong Province is one of the major
agricultural provinces in China. However, water scarcity is a long-standing and widespread problem
in this region, and available water resources per capita volume in Shandong are only 334 m3 , which
is less than 1/6 of China’s average [42]. At the same time, rapid population growth and increasing
agricultural activities have greatly increased water demand for agricultural productions recently.
Therefore, it makes Shandong Province a good test-bed for comprehensive assessment of water
footprint and water scarcity pressure for main crops.
In this study, we adopted the monthly climate data for each year from 22 meteorological stations
during 1989–2016 in Shandong Province to accurately calculate WFs of crops and to analyze the
spatiotemporal variation characteristics of WFs for wheat, maize, cotton and groundnut in different
rainfall years. Therefore, the primary objectives of this study are: (1) to comprehensively assess the
green, blue, grey and total WFs in the period of 1989–2016; and (2) to estimate the pressure of water
scarcity on agriculture productions by combining with blue water pressure index (WPIblue ), then to
provide valuable suggestions for adjustment of production layout to reduce agricultural water use in
Shandong Province. The results of this present study will provide a baseline for the sustainable use of
water resources in this typical water scarce region.

2. Materials and Methods

2.1. Study Area

Shandong Province is located in eastern China (34◦ 22.90 –38◦ 24.010 N, 114◦ 47.50 –122◦ 42.30 E) and
is divided into 17 prefecture-level regions (Figure 1). This province has a warm temperate monsoon
climate. The annual mean temperature is 11.2–14.4 ◦ C [43]. The average annual precipitation is
550–950 mm, of which summer rain contributes more, and decreases gradually from southeast to
Sustainability 2019,
2019, 11, x FOR PEER REVIEW
1856 33of
of 19
18

most widely planted crops in Shandong Province, namely, wheat, maize, cotton and groundnut,
Sustainability 2019, 11, x FOR PEER REVIEW 3 of 19
northwest
were across the province [44]. In this study, the WFs of four kinds of main crops which were
analyzed.
the most
Using widely plantedannual
the planted
average crops in Shandong Province,
precipitation of 679.5 namely,
mm wheat, maize,
inwheat,
Shandong cotton
Province as and groundnut,
a groundnut,
baseline [45],
most widely crops in Shandong Province, namely, maize, cotton and
were
years analyzed.
in which
were analyzed. precipitation was >20% higher were humid years, years in which precipitation was
Using
>20% lower the
Using were average
drought
the average annual precipitation
years,precipitation
annual and years in of 679.5
of which
679.5 mmmm in Shandong
precipitation
in Shandong Province
wasProvince
higher or as a
as lower baseline
within
a baseline [45],
[45],10%
years
were
years in which
normal
in which precipitation
years was
[46]. According
precipitation >20%
was >20% higher
to the were
precipitation
higher humid
were humid years,
characteristicsyears
years, yearsand in which precipitation
the available
in which waswas
data sources
precipitation
>20% WFlower
for>20% lower were
weredrought
calculation, drought years,
years 1990 andmm),
(934.9
years, and years2002
years inwhich
in which
(417.4precipitation
mm), and 2016
precipitation was
was higher
(693.9
higher or
mm)
or lower
were
lower within
10%10%
categorized
within
were
aswere normal
humid,
normal years
drought [46].
yearsand According
[46].normal to the precipitation
years respectively
According characteristics
(Figure
precipitation 2).
characteristics and the available data
and the available data sourcessources
for
forWF
WFcalculation,
calculation,years years1990
1990(934.9
(934.9mm),
mm),2002
2002(417.4
(417.4 mm),
mm), and and 2016
2016 (693.9 mm) werewere categorized
categorized as
humid,
as humid,drought andand
drought normal years
normal respectively
years respectively (Figure 2).2).
(Figure

Figure
Figure 1.1.Digital
Digital Elevation
Elevation ModelModel
(DEM) (DEM) and of
and locations locations of meteorological
meteorological stations and
stations and prefecture-level
Figure 1. Digital
prefecture-level Elevation
regions Model (DEM)
in Shandong Province. and locations of meteorological stations and
regions in Shandong Province.
prefecture-level regions in Shandong Province.

Figure 2.
Figure
Figure 2. Selection
Selectionof
2. Selection oftypical
of typicalyears
typical yearsinin
years inthis study.
this
this study.
study.
Sustainability 2019, 11, 1856 4 of 18

2.2. Data
The data used to calculate the WF values for crops growth included meteorological observations
and agricultural data. The meteorological observations of different regions during 1989–2016
were obtained from the National Meteorological Information Center of China Meteorological
Administration [47], including the monthly average max and min temperature, relative humidity,
wind speed, sun hours and precipitation. We selected data from 22 meteorological stations to calculate
WF, and selected one for each region. When the meteorological data was incomplete in a region, it was
supplemented by the observations of nearby meteorological station. If there was no meteorological
station in one region, the meteorological data from a nearby region was selected for calculation.
The sown areas, yields and consumption of chemical fertilizers for four main crops were obtained from
field investigations, literatures, and Shandong Statistical Yearbook (1990–2017) [48]. Crops growth
periods were derived from FAO’s crops database [49]. Water resources data obtained from Shandong
Province Water Resources Bulletin were also used to estimate the pressure of water scarcity on
agriculture productions [45].

2.3. Methods
WF of a crop production, defined as the ratio of the crop water requirement to crop yield, was
calculated based on the framework from Chapagain and Hoekstra (Figure 3) [8]. The yield considers an
average quantity of crop produced over the full life span of the crop. WF consists of three components:
green (WFgreen , rainwater), blue (WFblue , irrigation water) and grey (WFgrey , freshwater pollution in
the production process) [23,37]. The value of WF is the sum of WFgreen , WFblue and WFgrey [46]:

WF = WFgreen + WFblue + WFgrey (1)

where WF is the water footprint of crop production (m3 /kg); WFgreen is the green water footprint
(m3 /kg); WFblue is the blue water footprint (m3 /kg) and WFgrey is the grey water footprint (m3 /kg).
WFgreen and WFblue are computed as follows:

WFgreen = 10 × ETgreen /Y (2)

WFblue = 10 × ETblue /Y (3)

where ETgreen is WFgreen evapotranspiration (mm); ETblue is WFblue evapotranspiration (mm); Y is

the crop yield (kg/ha) and the factor of 10 converts water depth (mm) into water volume per area
(m3 /ha).
WFgrey is calculated as follows [7]:

WFgrey = [(α × AR)/(cmax − cnat )]/Y (4)

where AR is the rate of chemical application to the field per hectare (kg/ha); α is the leaching run-off
fraction; cmax is the maximum acceptable concentration (10 mg/L) and cnat is the concentration in
natural water, assumed to be 0 mg/L [7,50].
The volume of polluted water depends on both the pollutant load and the adopted permissible
limit [46]. Considering the availability of local data, this study selected nitrogen as a representative
pollutant, and it was assumed that the nitrogen fraction that reached free flowing water bodies through
leaching or runoff equaled 25% of the application rate [51].
Subsequently, green and blue ETs are calculated as follows:

ETgreen = min (ETc , Peff ) (5)

ETblue = max (0, ETc − Peff ) (6)

Sustainability 2019, 11, 1856 5 of 18

where ETc is daily crop evapotranspiration (mm); Peff is the effective precipitation over the crop
growing period (mm), which is estimated by USDA S.C. Method [49].
Next, daily reference evapotranspiration (ET0 ) is multiplied by the crop coefficient (Kc )
Sustainability 2019, 11, x FOR PEER REVIEW 5 of 19
during the corresponding periods which are obtained from CROPWAT 8.0 to calculate the daily
evapotranspiration
Next, dailyof crop (ET
reference c ) [46]:
evapotranspiration (ET0) is multiplied by the crop coefficient (Kc) during
the corresponding periods which are ET c = Kc ×
obtained ET CROPWAT 8.0 to calculate the daily (7)
from
evapotranspiration of crop (ETc) [46]:
Finally, daily reference evapotranspiration is calculated by using the software CROPWAT 8.0 [49]
based on the Penman–Monteith model [52]: ETc = Kc × ET (7)
Finally, daily reference evapotranspiration is calculated by using the software CROPWAT 8.0
[49] basedET = [0.408∆
on0the (Rn − G) +
Penman–Monteith γ900(e
model a − ed )]/(T + 273)}/[∆ + γ (1 + 0.34U2 )]
[52]: (8)
ET0 = [0.408∆ (Rn − G) + γ900(ea − ed)]/(T + 273)}/[∆ + γ (1 + 0.34U2)] (8)
where ET0 is reference crop evapotranspiration (mm/d); Rn is the daily net radiation (MJ/m2 /d); G is
where
the daily ETheat
soil 0 is reference crop2 /d);
flux (MJ/m evapotranspiration (mm/d);
T is the daily mean temperature (◦ C);
Rn is the daily netUradiation (MJ/m2/d); G is
2 is the daily wind speed at
the daily soil heat flux (MJ/m 2/d); T is the daily mean temperature (°C); U2 is the daily wind speed at
2 m height (m/s); ea is the saturation vapor pressure per day (kPa); ed is the actual vapor pressure per
2 m height (m/s); ea is the saturation vapor pressure per day (kPa); ed is the actual vapor pressure per
day (kPa); ∆ is the slope of the saturation vapor pressure versus air temperature curve (kPa/◦ C) and γ
day (kPa); Δ is the slope of the saturation vapor pressure versus air temperature curve (kPa/°C) and
is the hygrometer constant (kPa/◦ C).
γ is the hygrometer constant (kPa/°C).

Figure
Figure 3. Framework
3. Framework to to calculatethe
calculate thewater
water footprint
footprint (WF)
(WF)ofofcrop production
crop in this
production study.
in this study.

3. Results
3. Results

3.1. Temporal
3.1. Temporal Changes
Changes of WF
of WF for for Main
Main Crops
Crops

3.1.1. 3.1.1. Temporal

Temporal Change
Change of WF
of WF forfor Wheat
Wheat
WheatWheat is main
is the the main crop
crop in in ShandongProvince.
Shandong Province. The
The sown
sownarea
areawas
wasabout 35%
about of the
35% of total area area
the total of of
crops in 2016 [48]. Over the period of 1989–2016, the average WF for wheat production was 2.02 3
crops in 2016 [48]. Over the period of 1989–2016, the average WF for wheat production was 2.02 m /kg
m3/kg in Shandong Province, and WFgreen, WFblue and WFgrey contributed 14.85%, 36.14% and 49.01%,
in Shandong Province, and WFgreen , WFblue and WFgrey contributed 14.85%, 36.14% and 49.01%,
respectively. In the humid year of 1990, the WF was 2.58 m3/kg, which was the highest during the
3 /kg, which was the highest during
respectively. In the humid year of 1990, the WF was 2.58 m
study period, with the WFgreen, WFblue and WFgrey accounting for 20.40%, 29.10% and 50.50%,
the study period,Inwith
respectively. the WFyear
the drought , WF
greenof blue
2002, and
the WFWF 2.50accounting
wasgrey for 20.40%,
m3/kg, comprising 12.92%29.10%
WFgreen,and 50.50%,
37.64%
respectively. In the
WFblue, and drought
49.44% WFgreyyear of 2002,
. Finally, the
in the WF was
normal year2.50 m3 /kg,
of 2016, comprising
the WF was 1.50 12.92% WFgreen , 37.64%
m3/kg, consisting of
17.56% WFgreen, 39.91% WFblue and 42.53% WFgrey.
Sustainability 2019, 11, 1856 6 of 18

WF 3
blue , and 2019,
Sustainability 49.44%
11, xWF
FORgrey . Finally,
PEER REVIEWin the normal year of 2016, the WF was 1.50 m /kg, consisting 6 of 19 of
17.56% WFgreen , 39.91% WFblue and 42.53% WFgrey .
Lessrainfall
Less rainfallresulted
resultedin inmore
moreirrigation
irrigation (WF
(WFblue),),while more rainfall resulted in more WFgreen.
blue while more rainfall resulted in more WFgreen .
Therefore, the WF blue and WFgreen were the highest in the drought year of 1989 and the humid year of
Therefore, the WFblue and WFgreen were the highest in the drought year of 1989 and the humid year
of1990,
1990,repectively.
repectively.The TheWFWFblue and WFgreen were the lowest in 2003 and 2010. Besides, the WFgrey was
blue and WFgreen were the lowest in 2003 and 2010. Besides, the WFgrey
the lowest in 2015 and highest in 1993. The WFblue and WFgrey reduced obviously, while WFgreen
was the lowest in 2015 and highest in 1993. The WFblue and WFgrey reduced obviously, while WFgreen
decreased slowly (Figure 4a). The WFblue and WFgrey of wheat decreased from 1.14 m3/kg and 3 1.29
decreased slowly (Figure 4a). The WF blue and WFgrey of wheat decreased from 1.14 m /kg and
m3/kg3 in 1989 to 0.60 m3/kg and 0.64 m 3/kg in 2016, respectively. The sharp reduction of the WFs was
1.29 m /kg in 1989 to 0.60 m3 /kg and 0.64 m3 /kg in 2016, respectively. The sharp reduction of the
a result of improved wheat yields. Meanwhile, the WFgrey of wheat fell due to the efficient use of
WFs was a result of improved wheat yields. Meanwhile, the WFgrey of wheat fell due to the efficient
fertilizers.
use of fertilizers.

Figure4.4.Temporal
Figure Temporal changes of the
changes of the WF
WFgreen
green , WF
, WF , WF
blueblue
, WF grey and
grey and yieldsyields
for (a)for (a) wheat,
wheat, (b) maize,
(b) maize, (c)
(c)cotton
cottonand
and(d)
(d) groundnut
groundnut during
during 1989–2016
1989–2016 in Shandong
in Shandong Province.
Province.

3.1.2. Temporal Change of WF for Maize

3.1.2. Temporal Change of WF for Maize
Maize,
Maize,with witha asown
sownarea
areaofofabout
about29%29%ofofthethetotal area
total areaofofcrops
crops in in
2016,
2016,is is
thethesecond
second major
major crop
incrop
Shandong
in Shandong Province [48]. In the period of 1989–2016, the average WF for maize productionwas
Province [48]. In the period of 1989–2016, the average WF for maize production
1.24 3
m 1.24
/kgm in3/kg
Shandong Province, and WF
was in Shandong Province, WF, green
andgreen WF, blue andand
WFblue WFWF greygreycontributed
contributed 28.22%,
28.22%,20.91%
20.91%and
50.87%, respectively. In the humid year of 1990, the WF was 1.38 m 3 /kg,3 with the WF , WF and
and 50.87%, respectively. In the humid year of 1990, the WF was 1.38 m /kg, with the green WFgreen, WF blueblue
WFand
greyWFaccounting
grey accountingfor 35.77%, 20.81%
for 35.77%, andand
20.81% 43.42%43.42%respectively.
respectively. In Inthethe drought
drought year
yearofof2002,
2002,thetheWF
was
WF1.54was m31.54
/kg,mcomprising 17.26%17.26%
3/kg, comprising WFgreenWF, 31.96% WFblue
green, 31.96% WFand
blue 50.78%
and 50.78% WFgrey WF . Finally, in the
grey. Finally, innormal
the
year of 2016,
yearthe of WF
2016,was
the1.13
WF m 3 /kg,
1.13consisting of 30.11% of WF , 19.36% WF and 50.53% WF
normal was m3/kg, consisting 30.11%
green WF green, 19.36% blueWF blue and 50.53% grey .
WFThe
grey. WFgreen was the highest in 1990. The WFblue and WFgrey were the highest in 1997 during the

TheThe
28 years. WFWF green wasand
blue the WF
highest
greenin 1990.
were The
the WFblueinand
lowest 2003WF and were the
grey 2013. highest inthe
In addition, 1997 WF during
grey was thethe
28 years.
lowest The There
in 2011. WFblue was
and aWF clear were the lowest
green decreasing tendencyin 2003 andWF
in the 2013.
blue
In
overaddition,
time the
while WFthe WF
grey was the
green and
lowest in 2011. There was a clear decreasing tendency in the WF
WFgrey were reducing gradually in the study period (Figure 4b). The WFblue of maize decreased
blue over time while the WF green andfrom
WF 3
grey were reducing gradually
3
0.50 m /kg in 1989 to 0.22 m /kg in 2016. in the study period (Figure 4b). The WF blue of maize decreased from

0.50 m3/kg in 1989 to 0.22 m3/kg in 2016.

3.1.3. Temporal Change of WF for Cotton

Cotton is a major economic crop in Shandong Province. The sown area was about 4% of the
total area of crops in 2016 [48]. In general, cotton needs a lot of water during the growing season.
Sustainability 2019, 11, 1856 7 of 18

3.1.3. Temporal Change of WF for Cotton

Cotton is a major economic crop in Shandong Province. The sown area was about 4% of the
total area of crops in 2016 [48]. In general, cotton needs a lot of water during the growing season.
Therefore, the adjustment of planting structure and improvement of irrigation technology are helpful
to save water. During the period of 1989–2016, the average WF for cotton production was 7.29 m3 /kg
in Shandong Province, and WFgreen , WFblue and WFgrey contributed 50.00%, 35.48% and 14.52%,
respectively. In the humid year of 1990, the WF was 9.68 m3 /kg, with the WFgreen , WFblue and WFgrey
accounting for 56.50%, 25.23% and 18.27%, respectively. In the drought year of 2002, the WF was
6.79 m3 /kg, comprising 31.60% WFgreen , 55.16% WFblue and 13.24% WFgrey . Finally, in the normal year
of 2016, the WF was 4.49 m3 /kg, consisting of 61.76% WFgreen , 31.00% WFblue and 7.24% WFgrey .
The overall results indicated that the highest WFgreen was 6.13 m3 /kg in 1994. The highest WFblue
was 7.26 m3 /kg and WFgrey was 3.58 m3 /kg in 1992. The trends of WFgreen , WFblue and WFgrey all
decreased from 1989 to 2016 due to the increase in yields (Figure 4c). The WFgreen and WFblue of cotton
decreased from 2.92 m3 /kg and 5.14 m3 /kg in 1989 to 2.77 m3 /kg and 1.39 m3 /kg in 2016, respectively.

3.1.4. Temporal Change of WF for Groundnut

Groundnut is one of the main economic crops in Shandong Province. And the sown area was about
7% of the total area of crops in 2016 [48]. Over the period of 1989–2016, the average WF for groundnut
production was 1.75 m3 /kg in Shandong Province, and WFgreen , WFblue and WFgrey contributed
54.47%, 31.21% and 14.32%, respectively. If the 28 years were divided into three periods: period I
(1989–1998), period II (1999–2008) and period III (2009–2016), the average WFgreen was approximately
1.13 m3 /kg, 0.90 m3 /kg and 0.79 m3 /kg, respectively, for the three periods. Due to the higher effective
rainfall in period I, the average WFgreen in period I was greater than that in period III. The average
WFblue of groundnut was 0.75 m3 /kg, 0.46 m3 /kg and 0.40 m3 /kg in period I, period II and period
III, respectively. And WFblue in period I was 1.88 times greater than that in period III. The average
WFgrey of groundnut was 0.26 m3 /kg, 0.31 m3 /kg and 0.17 m3 /kg in period I, period II and period III,
respectively. The WFgrey in period I was 1.53 times greater than that of period III. Figure 4d showed
that the decline in the value of WFgreen and WFblue was greater than that of WFgrey . The WFgreen and
WFblue of groundnut decreased from 1.12 m3 /kg and 1.60 m3 /kg in 1989 to 0.87 m3 /kg and 0.33 m3 /kg
in 2016 respectively.

3.2. Spatial Distribution of WF for Main Crops in Different Rainfall Years

3.2.1. Spatial Distribution of WF for Wheat

Figure 5 showed the spatial distribution of WFgreen , WFblue and WFgrey for wheat in different
rainfall years in Shandong Province. It could be seen that the WFgreen of wheat displayed a spatial
aggregation pattern (Figure 5a–c). The WFgreen was the highest in eastern Shandong while it was the
lowest in western Shandong because the local precipitation in the east region was more than that in
the west region. In the different rainfall years, the responses of WFgreen varied in different regions [46].
In the humid year of 1990, which had the highest precipitation among the three years, the WFgreen
was the highest. The areas with higher values were mainly distributed in Heze, Linyi and Zaozhuang.
The region with lower WFgreen mainly distributed in Dezhou, with the value of 0.35 m3 /kg. In the
drought year of 2002, which had the lowest precipitation among the three years, the WFgreen of all
regions displayed decreasing trends compared with that in 1990. The regions with higher values
mainly included Weihai and Linyi. The lowest WFgreen was 0.16 m3 /kg, which was found in Binzhou.
Compared with 2002, the precipitation increased in the normal year of 2016, but the WFgreen was the
lowest among the three years. The area with high value was concentrated in Weihai. However, Dezhou
showed the lowest WFgreen level in 2016.
Sustainability 2019, 11, 1856 8 of 18
Sustainability 2019, 11, x FOR PEER REVIEW 8 of 19

Figure 5. Spatial distribution of (a–c) WFgreen , (d–f) WFblue and (g–i) WFgrey for wheat in 1990, 2002
and 20165.inSpatial
Figure Shandong Province.of (a–c) WFgreen, (d–f) WFblue and (g–i) WFgrey for wheat in 1990, 2002
distribution
and 2016 in Shandong Province.
The spatial distribution of WFblue was also assumed to be influenced by the rainfall pattern.
It mainly
The showed dynamic changes
spatial distribution of WFamong
blue was the
alsothree yearsto
assumed (Figure 5d–f). Inby
be influenced thethe
humid yearpattern.
rainfall of 1990,It
due to low
mainly yield in
showed the north
dynamic central
changes region,
among thethe areas
three with(Figure
years high values
5d–f).were
In themainly
humid distributed in
year of 1990,
Jinan,
due toDongying
low yieldand Weifang.
in the In the drought
north central region, theyear of 2002,
areas withthehighWF blue was
values themainly
were highestdistributed
of the threein
years.
Jinan,In the normal
Dongying andyear of 2016,Inthe
Weifang. theareas with year
drought high of
values
2002,decreased
the WFbluecompared with those
was the highest in 2002
of the three
and distributed
years. in all regions.
In the normal year of 2016, the areas with high values decreased compared with those in 2002
andThe spatial distribution
distributed in all regions.of the WFgrey followed the pattern of the planting area. In different
rainfallThe spatial high
years, the values areas
distribution of theappeared in different
WFgrey followed theregions
pattern (Figure
of the5g–i). Specially,
planting area. low WFgrey
In different
was found
rainfall in thethe
years, middle
high part of Laiwu.
values The reasoninwas
areas appeared that theregions
different wheat planting area was
(Figure 5g–i). less in this
Specially, low
region. Meanwhile, the WF
WFgrey was found in the middle grey was lower in the east region in 2016 because of the higher
part of Laiwu. The reason was that the wheat planting area was less wheat yield
and the efficient
in this utilization ofthe
region. Meanwhile, fertilizers.
WFgrey was Thelower
spatial
invariation of WFgrey
the east region in 2002
in 2016 was more
because obvious
of the higher
than
wheatthatyield
in 1990
andandthe2016. In summary,
efficient utilizationthe of
WFfertilizers.
of wheat wasThe higher
spatialinvariation
Weihai, Yantai
of WFand Linyi
grey in while
2002 was
itmore
was the lowest in Laiwu.
obvious than that in 1990 and 2016. In summary, the WF of wheat was higher in Weihai,
Yantai and Linyi while it was the lowest in Laiwu.
3.2.2. Spatial Distribution of WF for Maize
3.2.2.
TheSpatial Distribution
spatial of WF
distribution forgreen
of WF Maize, WFblue and WFgrey for maize in 1990, 2002 and 2016 in
Shandong Province was shown in Figure
The spatial distribution of WFgreen, WF 6. In general, the WF maize
blue and WFgrey for green
of in
maize displayed
1990, 2002 anda spatial
2016 in
aggregation pattern, with highest in the east regions and lowest in the west regions
Shandong Province was shown in Figure 6. In general, the WFgreen of maize displayed a spatial(Figure 6a–c).
The areas with higher values included Dongying, Binzhou, Yantai, Laiwu and Linyi. The
aggregation pattern, with highest in the east regions and lowest in the west regions (Figure 6a–c).regions
with
The lower
areas WF green
with mainly
higher distributed
values includedin the west inland
Dongying, areas, Yantai,
Binzhou, such as Laiwu
Taian, Liaocheng
and Linyi.and
TheDezhou.
regions
with lower WFgreen mainly distributed in the west inland areas, such as Taian, Liaocheng and
Sustainability 2019, 11, 1856 9 of 18
Sustainability 2019, 11, x FOR PEER REVIEW 9 of 19

Dezhou.
The The
average WFaverage WFgreen of Shandong Province 3was 0.50 m3/kg in 1990, which was higher3than
green of Shandong Province was 0.50 m /kg in 1990, which was higher than 0.27 m /kg
0.27
in m3and
2002 /kg in 2002
0.34 m3and
/kg 0.34 m3/kg in 2016.
in 2016.

Figure 6. Spatial distribution of (a–c) WFgreen , (d–f) WFblue and (g–i) WFgrey for maize in 1990, 2002
Figure 6. Spatial distribution of (a–c) WFgreen, (d–f) WFblue and (g–i) WFgrey for maize in 1990, 2002 and
and 2016 in Shandong Province.
2016 in Shandong Province.

WFblue of maize mainly showed obvious changes among the three years (Figure 6d–f). In the
WFblue of maize mainly showed obvious changes among the three years (Figure 6d–f). In the
humid year of 1990, larger WFblue was found in Binzhou, Dongying and Jinan. This was due to low
humid year of 1990, larger WFblue was found in Binzhou, Dongying and Jinan. This was due to low
yields in the north regions. The WFblue in the drought year of 2002 was the highest among the three
yields in the north regions. The WFblue in the drought year of 2002 was the highest among the three
years in most regions except Dongying and Yantai. In the normal year of 2016, the WFblue was the
years in most regions except Dongying and Yantai. In the normal year of 2016, the WFblue was the
lowest among the three years, though the WFblue in the east regions was relatively high.
lowest among the three years, though the WFblue in the east regions was relatively high.
The spatial distribution of the WFgrey of maize decreased from north to south (Figure 6g–i).
The spatial distribution of the WFgrey of maize decreased from north to south (Figure 6g–i). In
In the humid year of 1990, regions with larger WFgrey were located, for example, in Zibo, Binzhou
the humid year of 1990, regions with larger WFgrey were located, for example, in Zibo, Binzhou and
and Dongying. In 2002, the highest values areas were concentrated in the north part of Shandong.
Dongying. In 2002, the highest values areas were concentrated in the north part of Shandong. The
The areas with high values were different in 2002 and 2016. In summary, the WF of maize was higher
areas with high values were different in 2002 and 2016. In summary, the WF of maize was higher in
in Qingdao, Binzhou and Jinan while Taian had the lowest WF.
Qingdao, Binzhou and Jinan while Taian had the lowest WF.
3.2.3. Spatial Distribution of WF for Cotton
3.2.3. Spatial Distribution of WF for Cotton
For cotton, the results showed that the WFgreen was highest in the middle regions and lowest in
For cotton,
the eastern the(Figure
regions results7a–c).
showedIn that the WFyear
the humid green was highest in the middle regions and lowest in
of 1990, the WFgreen was the highest among the
the eastern regions (Figure 7a–c). In the humid
three years. The areas with higher values were mainlyyear of 1990, the in
distributed WF green was the highest among the
Taian, Jinan and Linyi. However,
Rizhao and Weihai had the lowest WFgreen , which was 0 m /kg. The reasonTaian,
three years. The areas with higher values were mainly 3 distributed in was thatJinan and
there Linyi.
were no
However, Rizhao and Weihai had the lowest WF green, which was 0 m3/kg. The reason was that there
plantings in these regions. In the drought year of 2002, the WFgreen of all regions displayed decreasing
were no
trends plantings
except in these
Rizhao. regions.
Compared In2002,
with the drought
WFgreenyear of 2002,
increased the WF
in most green of all regions displayed
regions in 2016 and the areas
decreasing
with trends were
higher values except Rizhao. Compared
concentrated with regions.
in the middle 2002, WFgreen increased in most regions in 2016
and the areas with higher values were concentrated in the middle regions.
Sustainability 2019, 11, 1856 10 of 18
Sustainability 2019, 11, x FOR PEER REVIEW 10 of 19

Figure 7. Spatial distribution of (a–c) WFgreen , (d–f) WFblue and (g–i) WFgrey for cotton in 1990, 2002
Figure 7. Spatial distribution of (a–c) WFgreen, (d–f) WFblue and (g–i) WFgrey for cotton in 1990, 2002
and 2016 in Shandong Province.
and 2016 in Shandong Province.
The WFblue was highest in Binzhou and Dongying, where cotton was mainly grown. That was
becauseTheagriculture
WFblue waswas highest in Binzhou and Dongying,
a high-water-consuming where
sector, with highcotton was
water mainly grown.
consumption Thatwith
in areas was
because
large cropagriculture
production was a high-water-consuming
[53]. In 1990 and 2002, the sector, with higher
areas with high water
valuesconsumption
of WFblue were in areas
mainlywith
large crop production [53]. In 1990 and 2002, the areas with higher values
distributed in Binzhou and Jinan. This was due to low yields in the two regions. Compared with 1990 of WF blue were mainly

distributed
and 2002, theinspatial
Binzhou and Jinan.
discrepancy of This
WFblue waswas due to low yields
relatively small in in2016
the two regions.
(Figure 7d–f).Compared with
1990Obviously,
and 2002, thethe spatial
WFgrey discrepancy
of cotton wasoflow WFin blue was relatively small in 2016 (Figure 7d–f).
eastern Shandong due to the limited planting area of
Obviously, the WF grey of cotton was low in eastern Shandong due to the limited planting area
cotton (Figure 7g–i). However, it was relatively larger in Dongying and Binzhou in different rainfall
of cotton
years. (Figure 7g–i).
In summary, the WFHowever,
of cotton it was was relatively
higher larger in
in Dongying, Dongying
Binzhou and Binzhou
and Dezhou while in different
Yantai had
rainfall years.
the lowest WF. In summary, the WF of cotton was higher in Dongying, Binzhou and Dezhou while
Yantai had the lowest WF.
3.2.4. Spatial Distribution of WF for Groundnut
3.2.4. Spatial Distribution of WF for Groundnut
We also calculated the WFgreen , WFblue and WFgrey of groundnut in each region of Shandong
We in
Province also calculated
1990, 2002 andthe 2016.WFThe green,results
WFblueshowed
and WF that of groundnut
grey WF in each
green was higher in the region
northofandShandong
central
Provincewhile
regions, in 1990, 2002
it was and in
lower 2016.
the The
east results
regionsshowed
(Figurethat WFIn
8a–c). thewas
green higher
humid yearin ofthe1990,
north WFand central
green was
regions, while it was lower in the east regions (Figure 8a–c). In the humid
the highest among the three years. The areas with higher values were mainly distributed in Dongying year of 1990, WF green was
the highest among the three years. The areas with higher values
and Binzhou. However, Rizhao and Qingdao had the lowest WFgreen , with the values of 0.99 m /kg were mainly distributed 3 in
Dongying
and 1.02 m and
3 /kg,Binzhou. However,
respectively. Rizhao
The reason was and Qingdao
that had higher
there were the lowest WFin
yields , with
these
green the values
regions. In theof
0.99 m 3/kg and 1.02 m3/kg, respectively. The reason was that there were higher yields in these
drought year of 2002, WFgreen was the lowest and the areas with higher values were concentrated in
regions.
the middle Inregions.
the drought year ofwith
Compared 2002, WFWF
2002, green wasof
green the alllowest
regionsand the areas
showed with higher
an upward tendencyvalues were
in 2016.
concentrated in the middle regions. Compared with 2002, WFgreen of all regions showed an upward
tendency in 2016.
Sustainability 2019, 11, 1856 11 of 18
Sustainability 2019, 11, x FOR PEER REVIEW 11 of 19

Figure 8. Spatial distribution of (a–c) WFgreen , (d–f) WFblue and (g–i) WFgrey for groundnut in 1990,
Figure 8. Spatial distribution of (a–c) WFgreen, (d–f) WFblue and (g–i) WFgrey for groundnut in 1990,
2002 and 2016 in Shandong Province.
2002 and 2016 in Shandong Province.
WFblue of groundnut presented obvious spatial changes among the three years (Figure 8d–f).
In theWF blue of groundnut presented obvious spatial changes among the three years (Figure 8d–f). In
humid year of 1990, WFblue was lower in northeast regions than that in southwest regions.
the humid
In the drought year of 1990,
year of 2002,WFWF blue was lower in northeast regions than that in southwest regions. In the
blue was the highest among the three years. In the normal year of 2016,
drought year of 2002, WF blue was the highest among the three years. In the normal year of 2016, the
the highest WFblue was mainly distributed in east regions while Taian had the lowest WFblue .
highest
The WF blue was mainly distributed in east regions while Taian had the lowest WFblue.
spatial distribution of the WFgrey of groundnut decreased from east to west in different
The spatial
rainfall years (Figure distribution
8g–i). It wasof the WFgrey of
consistent groundnut
with the plantingdecreased
structure from east to west
of groundnut in in different
Shandong
rainfall years (Figure 8g–i). It was consistent with the planting
Province. As can be seen from the figure, there were obvious differences in WFgrey of groundnut structure of groundnut in Shandong
Province.
between As can
regions in be seen from
Shandong. In the figure,
general, thethere
WF ofwere obviouswas
groundnut differences
higher inin WFgrey ofYantai
Dongying, groundnut
and
between regions in Shandong.
Weihai while Qingdao had the lowest WF. In general, the WF of groundnut was higher in Dongying, Yantai and
Weihai while Qingdao had the lowest WF.
3.3. Trends of Annual Total WF and WFblue for Main Crops
3.3. Trends of Annual Total WF and WFblue for Main Crops
The total WF can reflect types and quantities of water needed for crops growth. During the
study The
period,totalthe WF can reflect
average annual types
totalandWF quantities
was 420.59of × water
108 m3needed
/yr requiredfor crops growth.
for wheat During the
production in
study period,
Shandong the average
Province, of which annual
14.38% total
wasWFWFwas green 420.59
(60.48 ×× 10
108 m
8 3
m /yr
3 required
/yr), 35.34% for
waswheat
WF production
blue (148.64 ×in
Shandong
8 3 Province, of which 14.38% was WF
8 3 (60.48 × 10 8 m3/yr), 35.34% was WFblue (148.64 × 108
10 m /yr), 50.28% was WFgrey (211.47 × 10 m /yr), respectively. The total WF and WFblue showed
green

m3/yr), 50.28%
downward trends was
overWF the (211.47
greyyears × 109a).
(Figure 8 m3/yr), respectively. The total WF and WFblue8 showed
The average annual total WF was 222.16 × 0 m3 /yr
downward
required trendsproduction
for maize over the years (Figure Province,
in Shandong 9a). The average
and WFgreen annual total
(61.54 × WF
108 m was 222.16
3 /yr), WFblue× 08(44.52
m3/yr
×required
108 m3 /yr) for maize
and WF production
grey (116.10 in×Shandong
108 m3 /yr) Province,
contributed and 27.70%,
WFgreen (61.54
20.04%× and
108 m 3/yr), WFblue (44.52 ×
52.26%, respectively.
10 total
The 8 m /yr)
3
WF and WFgreyan
showed (116.10 × 10 trend
increasing 8 m /yr)and
3 contributed
the WFblue27.70%,appeared 20.04% and 52.26%,
a fluctuating trendrespectively.
over the yearsThe
total WF
(Figure 9b).showed
The averagean increasing
annual total trend
WF and the WF
for cotton was appeared
blue72.70 × 108 m a 3fluctuating trend Province,
/yr in Shandong over the with
years
(Figure
the green9b).(30.44The × average annual
108 m3 /yr), bluetotal WF
(23.79 for8 m
× 10 cotton
3 /yr)wasand 72.70
grey (18.47× 108 × m310
/yr
8m in3 /yr)
Shandong Province,
WF accounting
with the green (30.44 × 10 m /yr), blue (23.79 × 10 m /yr) and grey (18.47 × 10 m3/yr) WF
8 3 8 3 8

accounting for 41.87%, 32.72% and 25.41%, respectively. The total WF and WFblue showed
Sustainability 2019, 11, 1856 12 of 18

Sustainability 2019, 11, x FOR PEER REVIEW 12 of 19

for 41.87%, 32.72% and 25.41%, respectively. The total WF and WFblue showed decreasing trends over
the years (Figure
decreasing trends9c).overObviously,
the yearsthere was9c).
(Figure a sharp descending
Obviously, theretrend
was because
a sharp ofdescending
the reduction in
trend
cotton sown area during 1993–1999. The average annual total WF was 8 3
× 10annual
because of the reduction in cotton sown area during 1993–1999. The50.07
average m /yr required
total WF wasfor
groundnut production in Shandong Province, comprising 51.23% WF (25.65 × 108 m3 /yr), 25.60%
50.07 × 10 m /yr required for groundnut production in Shandong Province, comprising 51.23%
8 3 green
WF
WFgreen (25.65 ×
blue (12.82 1088 m
× 10 m33/yr),
/yr)25.60%
and 23.17%
WFblueWF grey (11.60
(12.82 × 310
× 108 m
8
/yr)m
3 /yr), respectively. The total 8WF3 and
and 23.17% WFgrey (11.60 × 10 m /yr),
WF blue showed
respectively. The decreasing
total WF and trends
WFin volatility
blue showed (Figure 9d).trends in volatility (Figure 9d).
decreasing

Figure 9.
Figure Temporal changes
9. Temporal changes of
of the
the annual
annual total
total WF
WF and
and WF
WFblue for (a)
blue for (a) wheat,
wheat, (b)
(b) maize,
maize, (c)
(c) cotton
cotton and
and
(d) groundnut
(d) groundnut during
during 1989–2016
1989–2016 in
in Shandong
Shandong Province.
Province.

3.4. Pressure of Water Scarcity on Agriculture Productions

3.4. Pressure of Water Scarcity on Agriculture Productions
Figure 10 showed the trends of the total annual WF, WF for four main crops and total
Figure 10 showed the trends of the total annual WF, WFblue bluefor four main crops and total blue
blue water resources in Shandong Province. From 1989 to 2016, the average annual total water
water resources in Shandong Province. From 1989 to 2016, the average annual total water footprint
footprint consumption for four crops in Shandong Province was 765.52 × 108 m3 /yr, of which WFblue
consumption8 for3 four crops in Shandong Province was 765.52 × 108 m3/yr, of which WFblue (229.53 ×
(229.53 × 10 m /yr) accounted for 29.98%. In addition, the total annual WF and WFblue for four crops
10 m /yr) accounted for 29.98%. In addition, the total annual WF and WFblue for four crops
8 3
presented decreasing trends over the years. The decreasing trends were due to the improvement of
presented decreasing trends over the years. The decreasing trends were due to the improvement of
irrigation management and fertilizer use efficiency in Shandong Province in recent years. Furthermore,
irrigation management and fertilizer use efficiency in Shandong Province in recent years.
a reasonable planting structure could also result in the decreasing of total annual WF and WF .
Furthermore, a reasonable planting structure could also result in the decreasing of total annual bule WF
In some drought years, the total volume of water resources was obviously less than the total annual
and WFbule. In some drought years, the total volume of water resources was obviously less than the
WFblue for four main crops, indicating that blue water crisis existed in these years in Shandong.
total annual WFblue for four main crops, indicating that blue water crisis existed in these years in
The blue water pressure index (WPIblue ), namely the ratio of total WFblue to total blue water
Shandong.
resources, was used to evaluate the degree of blue water scarcity, which could reflect the status of water
resources in the region. When WPIblue is greater than 0.4, the region is recognized as severe water
scarce region [54]. And if WPIblue is greater than 1, this region is facing very serious water scarcity.
Based on the results of the total WFblue for four main crops in Shandong Province and combined with
the total blue water resources in different regions, we calculated the WPIblue and analyzed the spatial
variations of the blue water risk of the main crops in Shandong Province. The data of total water
resources for each region was only available in 2015 and 2016, so we analyzed the spatial distribution
of WPIblue in these two years.
As is shown in Figure 11, the WPIblue presented obvious spatial differences in Shandong Province.
In 2015, the areas with higher values were mainly distributed in Liaocheng and Qingdao which were
Sustainability 2019, 11, 1856 13 of 18

5.17 and 3.05. Because of the severe drought in these two areas in 2015, water resources could not
meet the needs of agricultural production. The WPIblue was greater than one, which accounted for
58.82% of the whole province, indicating that most regions in the province were facing the threat
of water resources shortage. And Laiwu had the lowest value of 0.51, which was also higher than
the warning line of 0.4. In 2016, WPIblue was higher in the northeast than that in other areas. The
areas with high values of more than 1.5 were mainly distributed in Dezhou, Liaocheng and Qingdao.
Obviously, the drought was still serious in the three regions in 2016. The WPIblue of all regions was
Sustainability
greater than 0.4 2019, 11, x FOR
except PEER REVIEW
Laiwu and Linyi. 13 of 19

Figure 10.10.Temporal
Figure Temporalchanges
changesofofthe
thetotal
total annual
annual WF,
WF, WF
WFblue forfour
bluefor fourmain
maincrops
cropsand
andtotal
totalwater
water
Sustainability 2019, 11, x FOR PEER REVIEW 14 of 19
resources
resources in in Shandong
Shandong Province.
Province.

The blue water pressure index (WPIblue), namely the ratio of total WFblue to total blue water
resources, was used to evaluate the degree of blue water scarcity, which could reflect the status of
water resources in the region. When WPIblue is greater than 0.4, the region is recognized as severe
water scarce region [54]. And if WPIblue is greater than 1, this region is facing very serious water
scarcity. Based on the results of the total WFblue for four main crops in Shandong Province and
combined with the total blue water resources in different regions, we calculated the WPIblue and
analyzed the spatial variations of the blue water risk of the main crops in Shandong Province. The
data of total water resources for each region was only available in 2015 and 2016, so we analyzed the
spatial distribution of WPIblue in these two years.
As is shown in Figure 11, the WPIblue presented obvious spatial differences in Shandong
Province. In 2015, the areas with higher values were mainly distributed in Liaocheng and Qingdao
which were 5.17 and 3.05. Because of the severe drought in these two areas in 2015, water resources
could not meet the needs of agricultural production. The WPIblue was greater than one, which
accounted for 58.82% of the whole province, indicating that most regions in the province were
facing the threat of water resources shortage. And Laiwu had the lowest value of 0.51, which was
also higher than the warning line of 0.4. In 2016, WPIblue was higher in the northeast than that in
Figure
other 11. Spatial
areas. distribution of the blue water pressure index (WPIblue ) in (a) 2015 and (b)
in2016 in
Figure 11.The areas
Spatial with high
distribution values
of the of more
blue water thanindex
pressure 1.5 (WPI
were mainly
blue distributed
) in (a) 2015 Dezhou,
and (b) 2016 in
Shandong
Liaocheng Province.
and Qingdao. Obviously, the drought was still serious in the three regions in 2016. The
Shandong Province.
WPI blue of all regions was greater than 0.4 except Laiwu and Linyi.
In summary, judging from the distribution of the WPIblue in Shandong Province in 2015 and
In summary, judging from the distribution of the WPIblue in Shandong Province in 2015 and
2016, agricultural productions were facing the threat of water scarcity in most regions of Shandong
2016, agricultural productions were facing the threat of water scarcity in most regions of Shandong
Province. Especially in the last two years, the drought in Qingdao had been very serious. And the
Province. Especially in the last two years, the drought in Qingdao had been very serious. And the
Chanzhi Reservoir, which was the second largest reservoir of Shandong Province, had nearly dried
Chanzhi Reservoir, which was the second largest reservoir of Shandong Province, had nearly dried
up.
up.Therefore,
Therefore,ititwaswasurgent
urgentto to adjust
adjust the
the planting structureofofagriculture
planting structure agriculturereasonably
reasonably according
according to to
the local water resources. In case of insufficient of water, the cotton planting area should be
properly controlled in water scarce regions since it needs a lot of water during the growing season.

4. Discussion
Sustainability 2019, 11, 1856 14 of 18

the local water resources. In case of insufficient of water, the cotton planting area should be properly
controlled in water scarce regions since it needs a lot of water during the growing season.

4. Discussion
The main water resources for crops growth are rainfall and irrigation. Rainfall impacts the
amount of water footprint and has both annual and inter-annual variations. Therefore, study on
the temporal variation trends of crops water footprints and exploration of the distinctions of water
footprints between different rainfall years are very important to take advantage of rainfall and to adjust
irrigation accordingly. This present study provided more accurate assessment of water consumption
and laid a foundation for the adjustment of crop planting structure and the determination of reasonable
irrigation quotas.
In the past, several studies had estimated WFs of wheat, maize, cotton and groundnut in Shandong
Province. Three previous studies were selected for comparison (Figure 12). Comparing the WFgreen-blue
of wheat in the present study with those from previous studies, the WFgreen-blue in our estimation
was smaller than that of Cheng et al. [55], while it was larger than that of Li [56] and Gao [44].
The WFgreen-blue of maize in our study was slightly smaller than that of Gao [44] and Cheng et al. [55]
while it was almost the same as Li [56]. Additionally, the WFgreen-blue of cotton in this study was much
larger than that of Cheng et al. [55]. The WFgreen-blue of groundnut in current study was extremely
close to that of Cheng et al. [55] while it was much smaller than that of Gao [44]. Compared with
previous studies, this current study also estimated WFgrey of main crops in Shandong Province. At the
same time, the results showed that the WFgrey of wheat and maize was larger than WFblue and WFgreen ,
which was the most important part of water use during crop production. In addition, monthly
meteorological observations for each year from 1989–2016 were adopted in this study instead of the
Sustainability
multi-year 2019, 11,average
monthly x FOR PEER REVIEW
data. 15 of 19

Figure 12. Comparison of estimated WFs of crops with the results from previous studies in Shandong
Figure 12. Comparison of estimated WFs of crops with the results from previous studies in
Province. Each data point refers to the WF of a crop. Period: 1995–2008 for Gao, 2005 for Li and
Shandong Province. Each data point refers to the WF of a crop. Period: 1995–2008 for Gao, 2005 for Li
1978–2014 for Cheng
and 1978–2014 et al. et al.
for Cheng

These discrepancies
These discrepanciesofofWFs WFswere
were resulted fromthe
resulted from thedifferences
differences in calculation
in calculation models,
models, studystudy
periods and the various environments including climatic conditions and soil texture,
periods and the various environments including climatic conditions and soil texture, etc. The main etc. The main
objective of this
objective study
of this is istotoevaluate
study evaluate water scarcitypressure
water scarcity pressure adopting
adopting the the widely
widely usedused
methodmethod
in in
Shandong
Shandong Province.
Province. However,the
However, theresults
results of
of WF
WFgreen
greenand
andWF blue calculated
WF blue calculatedby CROPWAT
by CROPWAT8.0 may
8.0 may
havehave ledunderestimation
led to to underestimation or or overestimationofofWF
overestimation WFdue due to
to the
the lack
lackof
ofKKc cvalues
values and accurate
and dates
accurate dates of
of growing stages for crops. Additionally, the WF grey may also be underestimated or overestimated
growing stages for crops. Additionally, the WFgrey may also be underestimated or overestimated in
in this study due to the wide variety of fertilizers applied in farmland throughout Shandong
Province, the difference in maximum allowable concentrations and the difficulty in collecting
accurate data on the amount of fertilizer used. Therefore, it is not enough to give a single value of
WF without providing an uncertainty range [5]. Furthermore, sensitivities and uncertainties in the
assessment of WF were not provided in this present study. Despite these uncertainties, the results
are helpful to assess the problem of water scarcity and to develop wise water management strategies
Sustainability 2019, 11, 1856 15 of 18

this study due to the wide variety of fertilizers applied in farmland throughout Shandong Province,
the difference in maximum allowable concentrations and the difficulty in collecting accurate data
on the amount of fertilizer used. Therefore, it is not enough to give a single value of WF without
providing an uncertainty range [5]. Furthermore, sensitivities and uncertainties in the assessment of
WF were not provided in this present study. Despite these uncertainties, the results are helpful to
assess the problem of water scarcity and to develop wise water management strategies in Shandong
Province [57].
Three suggestions are put forward based on the results of this present study. First of all, the
government should strengthen macro-control and policy guidance on crop production. Total WF is
closely related to sown area which is influenced by macro-control. And reasonable prediction of prices
of crops can control the sown area, and then total WF will be reduced under the condition of food
security. The second, drought and water scarcity cause losses in the agricultural sector, and insurance
is a possible instrument to recovery the losses. Therefore, the government ought to encourage farmers
to take agricultural insurance [58,59]. The third, agricultural planting structure and distribution
are urgent to be optimized. According to the regional differences in pressure of water scarcity, the
scientific planning of crop planting structure and rational control of sown area of crops with large
water footprints should be implemented in severely water-scarce regions of Shandong. For example,
Weihai and Yantai are not suitable for planting wheat, because the WF of wheat in the regions is greater
than that of other regions. Qingdao is not suitable for planting maize due to the higher WF of maize in
this region. The WFs of crops can provide an important guidance for the adjustment of agricultural
structure in water scarce areas. Some other influential factors should also be considered in deciding the
planting structures besides water factor, like planting conditions, farmers’ habits, cultural and climatic
change [60–62]. Therefore, the above suggestions from the perspective of the water footprint strategy
may be limited.
At the same time, the pressure of water scarcity on agriculture productions should be evaluated
more accurately [63]. However, in this current study, we only calculated annual WPIblue at prefecture
level due to the lack of data. In our future work, we will adopt remote sensing technology to assess
monthly WPIblue at a higher spatial resolution.

5. Conclusions
Wheat, maize, cotton and groundnut are the main crops in Shandong Province, and the growth of
these four crops consumes a great deal of freshwater. This study calculated the WFs of four crops from
1989 to 2016 in Shandong Province and analyzed the spatial distribution of WFs in different rainfall
years. The results showed that cotton had the highest WF among the selected crops in 1989–2016.
The total annual WF of four crops exhibited a slight decrease over time. Different types of rainfall year
had great influences on WFs. During the three years, the areas with higher WF for wheat were found
in Weihai, Yantai and Linyi. Therefore, the planting area of wheat should be reduced in these regions.
The WF of maize was higher in Qingdao, Binzhou and Jinan. Similarly, the planting area of maize
in these regions should also be controlled. The areas with higher WF for cotton were distributed in
Dongying, Binzhou and Dezhou. In addition, Dongying, Yantai and Weihai were higher value areas
of groundnut WF. Finally, the spatial distribution of WPIblue indicated that agricultural productions
were facing the threat of water scarcity in most regions of Shandong Province. Therefore, in order
to realize efficient utilization of water resources, it was urgent to reasonably adjust the planting
structure of agriculture according to the local water resources. However, it should be noted that several
uncertainties remain in this current study, which may be resulted from the calculation models, study
periods, climatic conditions and soil texture, etc. Despite these uncertainties, the results will help to
assess the problem of water scarcity and to provide valuable suggestions for sustainable utilization of
water resources in Shandong Province. In our future study, we will adopt remote sensing technology
and develop a comprehensive evaluation method to estimate water footprint more accurately on a
large scale and study the flow of water footprint among different regions.
Sustainability 2019, 11, 1856 16 of 18

Author Contributions: M.F., B.G., W.W. and J.W. carried out the calculation, result analysis and drafted the
manuscript, which was revised by all authors. All authors gave their approval of the version submitted
for publication.
Funding: This work was supported by National Natural Science Foundation of China (41807170, 31371574),
Natural Science Foundation of Shandong Province (ZR2017BD021), SDUST Research Fund (2014TDJH101) and
the Doctor Star-up Foundation of Qingdao Agricultural University (663/1117012).
Acknowledgments: We appreciate the editors and the reviewers for their constructive suggestions and insightful
comments, which helped us greatly to improve this manuscript.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Guo, B.; Chen, Y.N.; Shen, Y.J.; Li, W.H.; Wu, C.B. Spatially explicit estimation of domestic water use in the
arid region of northwestern China: 1985–2009. Hydrol. Sci. J. 2013, 58, 162–176. [CrossRef]
2. Shrestha, S.; Pandey, V.P.; Chanamai, C.; Ghosh, D.K. Green, Blue and Grey Water Footprints of Primary
Crops Production in Nepal. Water Resour. Manag. 2013, 27, 5223–5243. [CrossRef]
3. OECD. OECD Environmental Outlook to 2050: The Consequences of Inaction; OECD: Paris, France, 2012.
4. Chapagain, A.K.; Hoekstra, A.Y. Water Footprints of Nations; Value of Water Research Report Series No. 16;
UNESCO-IHE: Delft, The Netherlands, 2004.
5. Zhuo, L.; Mekonnen, M.M.; Hoekstra, A.Y. Sensitivity and uncertainty in crop water footprint accounting:
A case study for the Yellow River basin. Hydrol. Earth Syst. Sci. 2014, 18, 2219–2234. [CrossRef]
6. Miglietta, P.P.; Leo, F.D.; Toma, P. Environmental Kuznets curve and the water footprint: An empirical
analysis. Water Environ. J. 2016, 31, 20–30. [CrossRef]
7. Hoekstra, A.Y.; Chapagain, A.K.; Aldaya, M.M.; Mekonnen, M.M. The Water Footprint Assessment Manual:
Setting the Global Standard; Earthscan: London, UK, 2011.
8. Hoekstra, A.Y.; Chapagain, A.K. The blue, green and grey water footprint of rice from production and
consumption perspectives. Ecol. Econ. 2011, 70, 749–758.
9. Sun, S.K.; Wu, P.T.; Wang, Y.B.; Zhao, X.N. The virtual water content of major grain crops and virtual water
flows between regions in China. J. Sci. Food Agric. 2013, 93, 1427–1437. [CrossRef]
10. Cao, X.C.; Wu, P.T.; Wang, Y.B.; Zhao, X.N. Assessing blue and green water utilisation in wheat production
of China from the perspectives of water footprint and total water use. Hydrol. Earth Syst. Sci. 2014, 18,
3165–3178. [CrossRef]
11. Suttayakul, P.; Aran, H.; Suksaroj, C.; Mungkalasiri, J.; Wisansuwannakorn, R.; Musikavong, C. Water
footprints of products of oil palm plantations and palm oil mills in Thailand. Sci. Total Environ. 2016, 542,
521–529. [CrossRef] [PubMed]
12. Sun, S.; Wu, P.; Wang, Y.; Zhao, X.; Liu, J.; Zhang, X. The impacts of interannual climate variability and
agricultural inputs on water footprint of crop production in an irrigation district of China. Sci. Total Environ.
2013, 444, 498–507. [CrossRef]
13. Xu, Y.; Huang, K.; Yu, Y.; Wang, X. Changes in water footprint of crop production in Beijing from 1978 to
2012: A logarithmic mean Divisia index decomposition analysis. J. Clean. Prod. 2015, 87, 180–187. [CrossRef]
14. Marano, P.R.; Filippi, A.R. Water Footprint in paddy rice systems. Its determination in the provinces of Santa
Fe and Entre Rios, Argentina. Ecol. Indic. 2015, 56, 229–236. [CrossRef]
15. Su, M.H.; Huang, C.H.; Li, W.Y.; Tso, C.T.; Lur, H.S. Water footprint analysis of bioethanol energy crops in
Taiwan. J. Clean. Prod. 2015, 88, 132–138. [CrossRef]
16. Lu, Y.; Zhang, X.; Chen, S.; Shao, L.; Sun, H. Changes in water use efficiency and water footprint in grain
production over the past 35 years: A case study in the North China Plain. J. Clean. Prod. 2016, 116, 71–79.
[CrossRef]
17. Chu, Y.M.; Shen, Y.J.; Yuan, Z.J. Water footprint of crop production for different crop structures in the Hebei
southern plain, North China. Hydrol. Earth Syst. Sci. 2017, 21, 3061–3069. [CrossRef]
18. Martinezaldaya, M.; Llamas, M.R. Water Footprint Analysis (Hydrologic and Economic) of the Guadiana River
Basin; Value of Water Research Report 35, No. 3; UNESCO-IHE Institute for Water Education: Delft,
The Netherlands, 2009.
Sustainability 2019, 11, 1856 17 of 18

19. Zeng, Z.; Liu, J.; Koeneman, P.H.; Zarate, E.; Hoekstra, A.Y. Assessing water footprint at river basin level:
A case study for the Heihe River Basin in northwest China. Hydrol. Earth Syst. Sci. 2012, 16, 2771–2781.
[CrossRef]
20. Bocchiola, D.; Nana, E.; Soncini, A. Impact of climate change scenarios on crop yield and water footprint of
maize in the Po valley of Italy. Agric. Water Manag. 2013, 116, 50–61. [CrossRef]
21. Denis, D.M.; Kumar, M.; Srivastava, S.; Suryavanshi, S.; Denis, A.F.; Singh, R.; Yadav, A.; Mishra, H. A High
Resolution Assessment of Water Footprint of Wheat to Understand Yield and Water Use Heterogeneity.
Water Resour. Manag. 2016, 30, 2641–2649. [CrossRef]
22. Roux, B.L.; Laan, M.V.D.; Vahrmeijer, T.; Bristow, K.L.; Annandale, J.G. Establishing and testing a catchment
water footprint framework to inform sustainable irrigation water use for an aquifer under stress. Sci. Total
Environ. 2017, 599–600, 1119–1129. [CrossRef]
23. Jing, M.; Hoekstra, A.Y.; Hao, W.; Chapagain, A.K.; Dangxian, W. Virtual versus real water transfers within
China. Philos. Trans. R. Soc. B-Biol. Sci. 2006, 361, 835–842.
24. Bulsink, F.; Hoekstra, A.Y.; Booij, M.J. The water footprint of Indonesian provinces related to the consumption
of crop products. Hydrol. Earth Syst. Sci. 2010, 14, 119–128. [CrossRef]
25. Wang, W.; Yu, Z.; Zhang, W.; Shao, Q.; Zhang, Y.; Luo, Y.; Jiao, X.; Xu, J. Responses of rice yield, irrigation
water requirement and water use efficiency to climate change in China: Historical simulation and future
projections. Agric. Water Manag. 2014, 146, 249–261. [CrossRef]
26. Wang, Y.B.; Wu, P.T.; Engel, B.A.; Sun, S.K. Application of water footprint combined with a unified virtual
crop pattern to evaluate crop water productivity in grain production in China. Sci. Total Environ. 2014,
497–498, 1–9. [CrossRef] [PubMed]
27. Cao, X.; Pute, W.U.; Wang, Y.; Zhao, X. Water Footprint of Grain Product in Irrigated Farmland of China.
Water Resour. Manag. 2014, 28, 2213–2227. [CrossRef]
28. Wang, Y.B.; Wu, P.T.; Engel, B.A.; Sun, S.K. Comparison of volumetric and stress-weighted water footprint of
grain products in China. Ecol. Indic. 2015, 48, 324–333. [CrossRef]
29. Zhuo, L.; Mekonnen, M.M.; Hoekstra, A.Y.; Wada, Y. Inter- and intra-annual variation of water footprint
of crops and blue water scarcity in the Yellow River basin (1961–2009). Adv. Water Resour. 2016, 87, 29–41.
[CrossRef]
30. Lamastra, L.; Miglietta, P.P.; Toma, P.; Leo, F.D.; Massari, S. Virtual water trade of agri-food products:
Evidence from italian-chinese relations. Sci. Total Environ. 2017, 599–600, 474–482. [CrossRef]
31. Doll, P.; Siebert, S. Quantifying blue and green virtual water contents in global crop production as well as
potential production losses without irrigation. J. Hydrol. 2010, 384, 198–217.
32. Mekonnen, M.M.; Hoekstra, A.Y. The green, blue and grey water footprint of crops and derived crop
products. Hydrol. Earth Syst. Sci. 2011, 15, 1577–1600. [CrossRef]
33. Mekonnen, M.M.; Hoekstra, A.Y. Water footprint benchmarks for crop production: A first global assessment.
Ecol. Indic. 2014, 46, 214–223. [CrossRef]
34. Hoekstra, A.Y.; Mekonnen, M.M. The water footprint of humanity. Proc. Natl. Acad. Sci. USA 2012, 109,
3232–3237. [CrossRef]
35. Lovarelli, D.; Bacenetti, J.; Fiala, M. Water Footprint of crop productions: A review. Sci. Total Environ. 2016,
548–549, 236–251. [CrossRef]
36. Gobin, A.; Kersebaum, K.C.; Eitzinger, J.; Trnka, M.; Hlavinka, P.; Takáč, J.; Kroes, J.; Ventrella, D.; Marta, A.D.;
Deelstra, J.; et al. Variability in the water footprint of arable crop production across European regions. Water
2017, 9, 93. [CrossRef]
37. Chapagain, A.K.; Hoekstra, A.Y.; Savenije, H.H.G.; Gautam, R. The water footprint of cotton consumption:
An assessment of the impact of worldwide consumption of cotton products on the water resources in the
cotton producing countries. Ecol. Econ. 2006, 60, 186–203. [CrossRef]
38. Chen, W.; Wu, S.; Lei, Y.; Li, S. China’s water footprint by province, and inter-provincial transfer of virtual
water. Ecol. Indic. 2017, 74, 321–333. [CrossRef]
39. Marston, L.; Konar, M. Drought impacts to water footprints and virtual water transfers of the Central Valley
of California. Water Resour. Res. 2017, 53, 5756–5773. [CrossRef]
40. Shen, Y.J.; Oki, T.; Utsumi, N.; Kanae, S.; Hanasaki, N. Projection of future world water resources under
SRES scenarios: Water withdrawal. Hydrolog. Sci. J. 2008, 53, 11–33. [CrossRef]
Sustainability 2019, 11, 1856 18 of 18

41. Ministry of Water Resources of the People’s Republic of China. Available online: http://www.mwr.gov.cn/
sj/tjgb/szygb/201811/t20181116_1055003.html (accessed on 10 March 2019).
42. Chinese Government Network. Available online: http://www.gov.cn/guoqing/2018-01/17/content_
5257529.htm (accessed on 10 March 2019).
43. Yu, C. Study on Regional Difference of Production Water Footprint of Main Crop Based on Cropwat in
Shandong Province. Master’s Thesis, Shandong Normal University, Jinan, China, 2014.
44. Gao, T. Industrial Structure Adjustment of Agriculture in Shandong Province Based on Virtual Water Theory.
Master’s Thesis, Nanjing Agricultural University, Nanjing, China, 2010.
45. Water Resources Department of Shandong Province. Available online: http://www.sdwr.gov.cn/zwgk/
ggsjkf/sjkfml/szygb/ (accessed on 10 March 2019).
46. Li, H.; Qin, L.; He, H. Characteristics of the water footprint of rice production under different rainfall years
in Jilin Province, China. J. Sci. Food Agric. 2018, 98, 3001–3013. [CrossRef] [PubMed]
47. National Meteorological Information Center of China Meteorological Administration. Available online:
http://data.cma.cn/ (accessed on 10 March 2019).
48. Shandong Provincial Bureau of Statistics. Shandong Statistical Yearbook (1990–2017); China Statistics Press:
Beijing, China, 2017.
49. FAO Database. Available online: http://www.fao.org/nr/water/infores_databases_cropwat.html (accessed
on 10 March 2019).
50. EPA. List of Drinking Water Contaminants: Ground Water and Drinking Water; US Environmental Protection
Agency: Washington, DC, USA, 2005.
51. Zhao, R.; Chen, X.; Zhang, F. Nitrogen cycling and balance of winter wheat-summer maize rotation system
in North China. Acta Pedol. Sin. 2009, 46, 684–697.
52. Allen, R.G.; Pereira, L.S.; Raes, D.; Smith, M. Crop Evapotranspiration: Guidelines for Computing Crop Water
Requirements; FAO Irrigation and Drainage Paper No. 56; FAO: Rome, Italy, 1998.
53. Ma, X.; Ma, Y. The spatiotemporal variation analysis of virtual water for agriculture and livestock husbandry:
A study for Jilin Province in China. Sci. Total Environ. 2017, 586, 1150–1161. [CrossRef] [PubMed]
54. Hoekstra, A.Y.; Mekonnen, M.M.; Chapagain, A.K.; Mathews, R.E.; Richter, B.D. Global Monthly Water
Scarcity: Blue Water Footprints versus Blue Water Availability. PLoS ONE 2012, 7, e32688. [CrossRef]
55. Cheng, Y.; Cao, S.; Yang, Y.; Li, X. Analysis of agricultural water footprint cycle and trend change in Shandong
Province. South-to-North Water Transf. Water Sci. Technol. 2018, 16, 169–175.
56. Li, S. Virtual Water and the Optimization of Industrial Structure of Agriculture in Shandong Province.
Master’s Thesis, Shandong Agricultural University, Tai’an, China, 2007.
57. Pietrucha-Urbanik, K.; Tchórzewska-Cieslak, B. Water supply system operation regarding consumer
safety using kohonen neural network. In Safety, Reliability and Risk Analysis: Beyond the Horizon;
Steenbergen, R.D.J.M., van Gelder, P.H.A.J.M., Miraglia, S., Vrouwenvelder, A.C.W.M., Eds.; Taylor &
Francis Group: London, UK, 2014; pp. 1115–1120.
58. Fusco, G.; Miglietta, P.P.; Porrini, D. How Drought Affects Agricultural Insurance Policies: The Case of Italy.
J. Sustain. Dev. 2018, 11, 1. [CrossRef]
59. Pietrucha-Urbanik, K.; Żelazko, A. Approaches to Assess Water Distribution Failure. Period. Polytech. Civ.
2017, 61, 632–639. [CrossRef]
60. Cai, Y.L. Vulnerability and adaptation of Chinese agriculture to global climate change. Chin. Geogr. Sci. 1997,
7, 289–301. [CrossRef]
61. Ye, Y.; Fang, X.Q. Boundary shift of potential suitable agricultural area in farming-grazing transitional zone
in northeastern China under background of climate change during 20th century. Chin. Geogr. Sci. 2013, 23,
655–665. [CrossRef]
62. Guo, B.; Li, W.H.; Guo, J.Y.; Chen, C.F. Risk assessment of regional irrigation water demand and supply in
an arid inland river basin of northwestern China. Sustainability 2015, 7, 12958–12973. [CrossRef]
63. Mekonnen, M.M.; Hoekstra, A.Y. Four billion people facing severe water scarcity. Sci. Adv. 2016, 2, e1500323.
[CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).