Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Jurnal 2

Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

Journal of Hydrology 617 (2023) 128999

Contents lists available at ScienceDirect

Journal of Hydrology
journal homepage: www.elsevier.com/locate/jhydrol

Research papers

Impacts of precipitation variations on agricultural water scarcity under


historical and future climate change
Jiongjiong Liu a, b, Zhonghao Fu a, b, Wenfeng Liu a, b, *
a
Center for Agricultural Water Research in China, College of Water Resources and Civil Engineering, China Agricultural University, Beijing 100083, China
b
National Field Scientific Observation and Research Station on Efficient Water Use of Oasis Agriculture in Wuwei of Gansu Province, Wuwei 733000, China

A R T I C L E I N F O A B S T R A C T

Keywords: Both blue water and green water contribute to agricultural water scarcity, which is subjected to impacts of
Blue water precipitation variations in the context of climate change. However, an explicit quantification of the possible
Green water effects of precipitation variations on agricultural water scarcity index (AWSI) under historical and future climate
Agricultural water scarcity index
is absent. In this study, we applied an integrated AWSI, which incorporates blue water and green water, to es­
Precipitation variations
timate agricultural water scarcity in China and to determine the association of AWSI with precipitation variations
Climate change
Irrigation efficiency over the historical (1971–2010) and future (2031–2070) periods. Precipitation variations have substantial effects
on changes in AWSI. Especially during extremely dry years with precipitation anomaly lower than − 2 standard
deviation, changes of AWSI from the multi-year average could be as high as 31%. Future climate change would
further intensify AWSI and amplify the effects of precipitation variations, particularly in Inner Mongolia with
changes of AWSI over 200%. Southwestern China could also experience expanding agricultural water scarcity
under future climate scenarios. Improving irrigation efficiency has potential to alleviate water scarcity by up to
30%. This study provides valuable information for scientists and policy decision makers to cope with the impacts
of precipitation variations on AWSI.

1. Introduction groundwater, streams, and lakes (Liu et al., 2017). Although green water
(the water store in soil aquifer and can be used as evapotranspiration by
Water scarcity is the status under which water requirements of crops) alone accounts for roughly 85 % of crop water demand across the
agricultural, economic, social, and environmental sectors cannot be world (Hoekstra and Mekonnen, 2012, Liu et al., 2009), it has often been
satisfied by available water resources (Oki and Kanae, 2006). It has ignored in water resources and scarcity evaluations (Liu et al., 2022a,
become an increasing threat to humans and ecosystems sustainability, Xie et al., 2020). However, evaluations based on only blue water may
and is expected to be more seriously under future climate change con­ not sufficiently reflect the realistic water scarcity conditions in major
ditions (Hoekstra, 2014, Richter et al., 2020, Schewe et al., 2014). crop production regions (Nunez et al., 2013, Quinteiro et al., 2018).
Agriculture is the world’s largest water user and is prominently That is because green water is the only water resources for rainfed
impacted by climate change and population growth (Gerten et al., 2020, cultivations and an important part of water resources for irrigated lands.
Ward and Pulido-Velazquez, 2008, Zhang et al., 2020). Continuously As a result, concurrent variations in both blue and green water avail­
increases in irrigation water withdrawal have led to growing water ability and consumption will affect water scarcity in agriculture (Liu
scarcity in many places of the world (Falkenmark, 2013, Postel et al., et al., 2017, Veettil and Mishra, 2016), which makes agriculture more
1996, Qin et al., 2019). As a consequence, aggravation of water scarcity sensitive to climatic change in many places. There have been some ef­
in agriculture affects food production, and thus threatens food security forts to incorporate blue and green water scarcity in an integrated water
(Huang et al., 2017, Liu et al., 2022a, Pastor et al., 2019). scarcity assessment. Liu et al. (2022a) developed a comprehensive
A proper assessment of water scarcity is the first step to address the agricultural water scarcity index (AWSI) that incorporates both green
water scarcity issues. Most previous researches on water scarcity have and blue water availability and consumption and compared it with
largely focused on blue water, i.e., available freshwater from conventional water scarcity index focusing only on blue water over the

* Corresponding author at: Center for Agricultural Water Research in China, College of Water Resources and Civil Engineering, China Agricultural University,
Beijing 100083, China.
E-mail addresses: wenfeng.liu@cau.edu.cn, wenfengliu@vip.sina.com (W. Liu).

https://doi.org/10.1016/j.jhydrol.2022.128999
Received 29 September 2022; Received in revised form 30 November 2022; Accepted 15 December 2022
Available online 19 December 2022
0022-1694/© 2022 Elsevier B.V. All rights reserved.
J. Liu et al. Journal of Hydrology 617 (2023) 128999

period 1971–2010 on a global scale. After that, Liu et al. (2022b) applied WAb = (Q − WW os − EFR) • IE (2)
this index to assess the possible impacts of future climate change on
AWSI. However, Liu et al. (2022b) only explored the average changes of where ETci is the unstressed water demand for crop i (m3 mon− 1), which
agricultural water scarcity during the period 2026–2050. The major was calculated by the product of potential evapotranspiration (PET) and
limitation of these integrated studies is that the impacts of precipitation crop-specific coefficients (Allen et al., 1998); n is the number of crops
variations on AWSI were overlooked for both historical and future considered, which equals to 19 in this study; WAg is agricultural green
climate conditions. water availability (m3 mon− 1), which includes water evaporated from
Renewable water resources will be influenced by expected changes soil moisture or rainwater from rainfed and irrigated cropland; WAb is
in climate, especially changes in precipitation, therefore leading to agricultural blue water availability (m3 mon− 1); IE is irrigation effi­
variations in water scarcity (Moursi et al., 2017). Recently, several ciency (dimensionless); Q is total blue water resources, estimated as
studies have investigated the effects of precipitation changes on water streamflow (m3 mon− 1); WWos is water withdrawal for other sectors
scarcity on different spatiotemporal scales (Gampe et al., 2016, Gao except irrigation, including domestic, industrial, and livestock (m3
et al., 2021). For example, Coffel et al. (2019) found that future hot and mon− 1); EFR is environmental flow requirements (m3 mon− 1), which is
dry climate could worsen blue water scarcity in the Nile river basin estimated by the variable monthly flow (VMF) method (Pastor et al.,
despite projected increases in future precipitation. Quinteiro et al. 2014). The VMF method is a valid hydrological method for estimating
(2019) assessed how climate in future could affect future green water environmental flow requirements on a large-scale, which uses simple
poverty under a reference period (1986–2005) and a future period algorithm and takes into account intra-annual variability. It improves
(2046–2065) climate in Portugal. Kummu et al. (2014) quantified the environmental flow calculations due to increased time resolution from
effects of hydroclimatic change on global blue and green water resources an annual to monthly basis and has been widely used (Jagermeyr et al.,
availability and requirements based on historical climate forcing data­ 2017, Pastor et al., 2019). The VMF method allocates 60 % of water
sets during the historical period 1977–2006. Besides, they also deter­ resources to EFR during low flow months (monthly streamflow: <0.4
mined the frequency of water scarcity by quantifying the number of times the long-term average); 45 % of water resources to EFR during
years falling in the water scarcity. Although it is already known that medium flow months (monthly flow: 0.4–0.8 times the long-term
precipitation has direct impact on water scarcity, the level to which average); and 30 % of water resources to EFR during high flow
precipitation variations would affect the integrated water scarcity months (monthly flow: > 0.8 times the long-term average). The esti­
incorporating blue and green water is still unclear, especially during mated EFR shows high value in the river channels, especially in the
extremely wet and dry years under both historical and future climate lower reach of the Yangtze River basin (Fig. S2).
conditions. This absent information impairs the formulation of robust In this study, we first estimated monthly AWSI and then obtained
policy decision for coping with climate change effects on AWSI. annual AWSI by averaging the monthly AWSI within each year. For
To address the above research gaps, this study applied an integrated every province, average AWSI for grid cells in that province was treated
AWSI via combining green water and blue water availability and use as as the provincial AWSI.
well as environmental flow requirements (EFR) (Liu et al., 2022a), to
assess historical and future changes in AWSI and the linkages to changes
2.2. Defining precipitation conditions
in precipitation. The mainland China was taken as the study area, given
that China is facing and will still experience severe water scarcity under
We applied the “standardized anomaly” method to define the extent
the current and future climate conditions (Cao et al., 2017, Liu et al.,
to which annual precipitation departs from its average condition, which
2019, Ren et al., 2021, Zhao et al., 2015, Zhu et al., 2019). The long-
has been previously used to define dry and wet years (Jones and Hulme
term changes in AWSI in China was verified for the historical
1996, Li et al., 2019). Standardized anomaly (SA) can be regarded as the
(1971–2010) and future (2031–2070) periods at the grid (0.5 arc de­
change of climate condition (precipitation in this study) for a given year
gree) and province levels (locations of each province and separation of
t (prt) from its multi-year average condition (pr), normalized by the
northern and southern provinces are provided in Fig. S1). Agricultural
standard deviation (σ):
blue and green water availability and use were simulated by a large-
scale hydrological model, namely PCR-GLOBWB (Wada et al., 2014). prt − pr
SAt = (3)
We considered two future Representative Concentration Pathway (RCP) σ
climate scenarios, i.e. low-emission scenario RCP2.6 and no-mitigation
scenario RCP6.0. This study explicitly and quantitatively determined where SAt is the standardized anomaly of precipitation in year t. We
the possible impacts of precipitation variations on AWSI under the his­ calculated the SA values of annual precipitation from 1971 to 2010 for
torical and future periods, which provides useful information for policy each province. Then annual precipitation was divided into different
decision makers to formulate robust measures to tackle the water scar­ categories based on the SA values at an interval of 0.5, i.e., (− ∞, − 2σ],
city issues. (− 2σ, − 1.5σ], (− 1.5σ, − 1.0σ], (− 1.0σ, − 0.5σ], (− 0.5σ, 0), (0, 0.5σ),
[0.5σ, 1.0σ), [1.0σ, 1.5σ), [1.5σ, 2σ), [2σ, 2.5σ), [2.5σ, +∞), according
2. Methods and data to Li et al., (2019). Dry (wet) years were defined as years with SA lower
than − 1σ (larger than +1σ). We considered extremely dry years with SA
2.1. Agricultural water scarcity index lower than − 2σ. Similarly, moderately dry (moderately wet) years were
defined as years with SA between − 1σ and 0 (between 0 and 1σ). Pre­
The agricultural water scarcity index (AWSI) was calculated as a cipitation standardized anomaly which approximates to 0 is considered
ratio of crop water demand (without water limitation) to total water to be more akin to normal (or average) condition. After this step, we
availability considering green water, blue water, and EFR (Liu et al., could identify the province-specific dry and wet conditions based on the
2022a). This index measures the degree to which total water availability local climate condition of each province.
(including blue and green water) could meet crop water demand. When
AWSI > 1, it indicates that the total water availability is not enough to 2.3. Impacts of different precipitation conditions on agricultural water
meet the crop water demand, otherwise there is no agricultural water scarcity index
scarcity. It could be estimated as:
∑n The impacts of precipitation variations on AWSI with a specific
i=1 ETci
AWSI = (1) province was quantified via comparing the average AWSI falling into a
WAg + WAb
specific SA category to the average AWSI over the long-term period (e.g.

2
J. Liu et al. Journal of Hydrology 617 (2023) 128999

1971–2010 for the historical period and 2031–2070 for the future
period):
ΔAWSIp,i = (AWSI p,i − AWSI p,m )/AWSI p,m • 100 (4)

where ΔAWSIp,i is the relative change of AWSI over years with SA falling
in the i SA category to multi-year average AWSI in province p (%);
AWSIp,i is the average AWSI over years with SA in the i category;
AWSIp,m is the average AWSI over the multi-year. This relative change
reflects the changes of AWSI from its long-term trend because of pre­
cipitation variation. After obtaining the relative change in AWSI and the
corresponding precipitation anomalies of each year in each province,
the national average effects of precipitation variations on ASWI can be
derived by integrating all province-year combinations with an area-
weight aggregation method (Li et al., 2019). In addition, we further
quantified the changes of AWSI over the dry, moderately dry, moder­
ately wet, and wet years from the multi-year average for each province.

2.4. Changes in future agricultural water scarcity index and effects of


irrigation efficiency

Future AWSI was estimated by using eq. (1) but with water avail­
ability and water demand data simulated by future climate data over the
period 2031–2070. Similar to historical period, we estimated the SA
values according to future precipitation and divided the SA into different
categories. We computed multi-year average AWSI over all future years
(2031–2070) (AWSIfut,m) and future dry years (AWSIfut,d) for all prov­
inces. The relative changes of AWSIfut,m and AWSIfut,d to multi-year
average AWSI over all historical years (1971–2010) (AWSIhist,m) and
historical dry years (AWSIhist,d) were quantified to indicate the impacts
of future climate change on AWSI:
ECCm,m = (AWSI fut,m − AWSI hist,m )/AWSI hist,m • 100 (5)
Fig. 1. Average agricultural water scarcity index (AWSI) over the period
ECCd,m = (AWSI fut,d − AWSI hist,m )/AWSI hist,m • 100 (6) 1971–2010 in China at the grid cell (a) and provincial (b) levels.

ECCd,d = (AWSI fut,d − AWSI hist,d )/AWSI hist,d • 100 (7) livestock density. More details of the model could be found in Wada
et al. (2014). This model has been widely used globally (Sutanudjaja
where ECCm,m is the effect of climate change on AWSI over all future et al., 2018, Wanders and Wada, 2015). These data were provided by the
years to all historical years (%); ECCd,m the effect of climate change over Inter-Sectoral Impact Model Intercomparison Project (ISIMIP) phase 2a
the future dry years to all historical years (%); ECCd,d the effect of (historical period) and 2b (future period). Over the historical period,
climate change over the future dry years to historical dry years (%); three different climate datasets were used to force the PCR-GLOBWB
AWSIhist,m and AWSIfut,m are the average AWSI for all historical and model, namely GSWP3 (Dirmeyer et al., 2006), WATCH-WFDEI (Wee­
future years; AWSIhist,d and AWSIfut,d are the average AWSI for the his­ don et al., 2014, Weedon et al., 2011), and Princeton (Sheffield et al.,
torical and future dry years. 2006) following the ISIMIP 2a protocol. For the future simulations, four
To isolate the effects of pure future climate change on AWSI, we used General Circulation Models (GCMs), i.e. GFDL-ESM2M, HadGEM2-ES,
the historical IE around the year 2010, which is the earliest available IPSL-CM5A-LR, and MIROC5, were used to drive the PCR-GLOBWB
province-specific IE data, to estimate historical and future AWSI. Be­ model following the ISIMIP2b protocol. Two RCPs, i.e. RCP2.6 and
sides, we also determined the possible mitigation on AWSI by increasing RCP6.0, were considered for future projections. ISIMIP2b was particu­
IE in 2010 to two proposed IE scenarios in 2030. We kept the future larly designed for a 1.5 degree warming scenario, therefore simulations
irrigation areas consistent with the historical period, given large un­ under scenario RCP8.5 were not provided. We used the average of
certainties in projecting future irrigation areas. different variables simulated by different climate forcing data for both
historical and future periods to reduce uncertainties associated with
2.5. Data source climate inputs.
In this study, we adopted the simulations of monthly streamflow (the
This study was conducted for the mainland China at grid cell (a variable “dis” in the ISIMIP protocol) to estimate blue water availability.
spatial resolution of 0.5 arc degree) and province levels for the historical The variable “arainfusegreen” for rainfed cropland and “airrusegreen”
(1971–2010) and future (2031–2070) periods. Historical and future for irrigated cropland from the ISIMIP archive were used to compute
monthly PET, water withdrawals, evaporated water from soil water or agricultural green water availability (Liu et al., 2022b). The ETc data
rainwater were simulated by the PCR-GLOBWB model. PCR-GLOBWB is were estimated by the method of Liu et al (2022a) for 19 crops, including
a grid-based large-scale hydrology and water resources model (Wada barley, cassava, citrus, cotton, grapes, groundnut, maize, millet, potato,
et al., 2014). PCR-GLOBWB was developed to estimate global daily pulses, rapeseed, rice, rye, sorghum, soybean, sugar beet, sugar cane,
surface water balance at a spatial resolution of 0.5 arc degree and to sunflower, and wheat. The MIRCA2000 dataset (Portmann et al., 2010)
represent different water withdrawals for meeting diverse water de­ was used to locate the irrigated and rainfed croplands, as well as the
mands. It can estimate water withdrawals for irrigation, industry, do­ growing period for each crop. Other information for estimating ETc, e.g.
mestic and livestock, which are influenced by irrigation area, crop water crop growth stage and crop-specific coefficient for scaling PET to ETc
demand, population, gross domestic product (GDP) per capita, and were obtained from Allen et al. (1998). The historical simulations by the

3
J. Liu et al. Journal of Hydrology 617 (2023) 128999

Fig. 2. Relationship between agricultural water scarcity index (AWSI) and precipitation over the period 1971–2010 in a northern province (Shandong, a)
and a southern province (Guangdong, b). For AWSI, dash line represents the mean value; for precipitation, the dotted line, dash line, and dash dot line represent
mean − 1σ (standard deviation), mean, and mean +1σ of multi-year precipitation, respectively. Red, light red, light blue, and blue colors indicate dry, moderately
dry, moderately wet, and wet years. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

PCR-GLOBWB model considered a continuous change in human activ­ 3. Results


ities, while the socioeconomic conditions (e.g., population, gross do­
mestic production, and land use) were fixed at the year 2005 for the 3.1. Water scarcity assessment during the historical period
future period, which was designed by the ISIMIP2b protocol. To keep the
historical and future AWSI comparable, i.e. both under the socioeco­ The average AWSI of per grid cell over the period 1971–2010 varies
nomic condition of 2005, we de-trended the historical AWSI by largely throughout the Mainland China (Fig. 1a). Regions with high
removing the multi-year linear trend and scaled it to the year 2005. AWSI values (AWSI > 1) are mainly distributed in the northwestern,
The IE data in the year 2010 obtained from the Bulletin of Water northern, and central eastern parts of China, particularly in the North
Resources in China at the provincial level (CPGPRC, 2012) was used for China Plain. On the other hand, AWSI is generally low in the southern
estimation of historical AWSI. We used the IE data in 2010 to estimate China. At the provincial level, the estimated average AWSI also changes
future AWSI for the purposes of isolating the pure effects of climate substantially between the southern and northern provinces over the
change on AWSI. In addition, projected IE in 2030 was used to reveal the historical period (Fig. 1b). Severe agricultural water scarcity is found in
impacts of improving IE on mitigating AWSI (Table S1). The Chinese Shandong, Tianjin, Hebei, Jiangsu, Henan, and Liaoning provinces.
government reported that national average IE in 2025 and 2030 would Average AWSI in the southern China is generally lower than 0.5, with
reach to 0.58 and 0.60, respectively, and provided province-specific IE AWSI of being lower than 0.2 in the southwestern part of China.
in 2025 (MWRC and NDRCC, 2022). We multiplied provincial IE in 2025
by the ratio of average IE in 2030 to average IE in 2025 (0.60/0.58) to
obtain the province-specific IE in 2030. Besides, we also used a high 3.2. Impacts of precipitation variations on agricultural water scarcity
national average IE (0.70) in 2030 to estimate future AWSI for the period index
2031–2070.
Fig. S3a shows that coefficient of variation (CV, the ratio of standard
deviation to mean) of precipitation is relatively large in northern China

4
J. Liu et al. Journal of Hydrology 617 (2023) 128999

Fig. 3. Relative changes of agricultural water scarcity index (AWSI) in different precipitation conditions compared to average AWSI in each province
during the 1971–2010 period. (a) dry years with precipitation anomalies lower than − 1σ; (b) moderately dry years with precipitation anomalies between − 1σ and
0; (c) moderately wet years with precipitation anomalies between 0 and 1σ; and (d) wet years with precipitation anomalies higher than 1σ.

provinces, especially in Xinjiang, Beijing, and Tianjin provinces. the ratios of province-year combinations in each precipitation anomaly
Meanwhile, there are strong negative correlations between annual pre­ ranging from dry to wet conditions and the changes of AWSI from
cipitation and AWSI over the period 1971–2010 (Fig. S3b). Unlike the provinces’ multi-year AWSI average. It is clear that (moderately) dry
CV, the correlations are more significant among the southern provinces. precipitation anomalies have significantly (p < 0.05) increased AWSI.
To explore the possible impacts of precipitation variations on AWSI, Especially for the province-year combinations with precipitation
we plotted the AWSI-precipitation interaction in a water scarce northern anomaly lower than − 2.0σ, average AWSI in this category is ~31 %
province (Fig. 2a, Shandong province) and a water abundant southern higher than the multi-year mean, although the sample ratio is only about
province (Fig. 2b, Guangdong province). The variations of annual pre­ 2 %. On the other hand, AWSI in (moderately) wet years is lower than
cipitation in Shandong and Guangdong are relatively large (with σ being the multi-year average, with ~29 % decrease for the precipitation
131 mm and 252 mm, respectively). For all dry years with precipitation anomaly higher than 2.5σ. Generally, AWSI during dry years could be
anomaly lower than –1σ over the 1971–2010 period, the AWSI values ~21 % higher than the multi-year average (Fig. S5).
are highest and above the multi-year average for both Shandong and
Guangdong provinces. With respect to wet years with precipitation 3.3. Changes in agricultural water scarcity index under future climate
anomaly higher than +1σ, the AWSI values are the lowest. change
For each province, AWSI during the historical dry years experienced
considerable positive changes from the multi-year average, particularly AWSI under future climate would have substantial changes
in Xinjiang, Shanxi, Beijing, and Zhejiang provinces with the changes compared to the historical AWSI (Figs. S6-S7). Future precipitation
over than 30 % (Fig. 3). On the other hand, during the wet years, AWSI variations will also induce strong influences on AWSI (Figs. S8-S9).
presents substantially negative changes from the multi-year average. Therefore, future climate change could have big effects on AWSI in
These positive and negative departures show large spatial variability terms of the relative changes of AWSI during all future years and future
over China. Generally, the northern provinces and eastern provinces dry years to those during all historical years and historical dry years
show larger changes compared to the southern and western provinces, (Fig. 5). The ECCm,m shows positive values in 19 provinces under the
which is consistent with spatial distribution of the CV of precipitation RCP2.6 scenario and 20 provinces under the RCP6.0 scenario (Fig. 5a
over the historical period (Fig. S3a). We further found that the anoma­ and b), indicating increases in future averaged AWSI. The Inner
lies of AWSI during dry or wet years present statistically significant (p < Mongolia province has the largest increases in AWSI over 200 %. On the
0.05) relationship with CV of precipitation (Fig. S4). Anomalies of AWSI other hand, future AWSI would have a decreasing trend over the prov­
tend to be large when CV of precipitation is high. inces with high historical AWSI, e.g. Tianjin, Hebei, Henan, and Liaon­
We aggregated percentage changes of AWSI from province-year ing provinces. Interestingly, we found that future climate change would
samples on the national scale to show water scarcity response to pre­ have more pronounced effects on changes of agricultural water scarcity
cipitation anomalies during the historical period in China. Fig. 4 exhibits compared to the effect of historical precipitation variations (Figs. 3 and

5
J. Liu et al. Journal of Hydrology 617 (2023) 128999

than 1 under the RCP2.6 scenario. Tibet has the highest decrease in
AWSI (about 20–30 %) by improving IE. These effects also hold for
average AWSI over all future years (Fig. S10).

4. Discussion

4.1. Impacts of precipitation variations on agricultural water scarcity

Precipitation variations could have direct impacts on the availability


of streamflow in the river systems (blue water) and soil moisture (green
water) for crop production, resulting in changes in agricultural water
scarcity. Although it was well recognized that precipitation variations
could have impacts on water scarcity (Coffel et al., 2019), there is no a
quantitative study to determine the changes of water scarcity from
multi-year average along different precipitation anomalies for both
historical and future climate. In this study, we applied the integrated
water scarcity index combining blue water and green water to estimate
AWSI in China and for this first time, to our knowledge, explicitly
revealed the percentage changes of AWSI relative to multi-year average
under different precipitation anomalies. Our study demonstrates that
precipitation variations under both current and future climate condi­
tions have substantial impacts on the AWSI. During the historical dry
years (precipitation anomaly lower than –1σ), AWSI is at least 20 %
higher than the multi-year average (Fig. 4 and Fig. S5). Especially under
historical extreme dry condition (precipitation anomaly lower than
–2σ), AWSI could be +30 % higher than the average, indicating strong
negative impacts of precipitation variations on AWSI. Future climate
Fig. 4. Changes of agricultural water scarcity index (AWSI) due to pre­ change would further amplify these impacts, with possible changes over
cipitation variations. (a) ratio (%) of province-year samples in each precipi­ than 200 % in Inner Mongolia and Qinghai provinces (Fig. 5).
tation anomaly to all province-year samples along the precipitation anomalies; AWSI exhibits distinct spatial patterns in responses to precipitation
(b) percentage change (%) of AWSI relative to multi-year average among
variations, which is mainly regulated by the CV of precipitation and the
different precipitation anomalies. In subplot b, each bar represents AWSI per­
correlation between precipitation and AWSI (Figs. S3 and S4). There­
centage change weighted by provincial area from province-year samples; error
bars represent 95% confidence interval estimated from 1,000 bootstrap esti­
fore, particular attentions should be paid to the provinces with strong
mations. Precipitation anomaly lower than − 1σ is defined as dry years, while negative correlation and high precipitation variations during the dry
higher than 1σ as wet years. The number of total province-year samples years, for example the provinces of Xinjiang, Shanxi, and Beijing, where
is 1,240. AWSI during dry years were 30 % higher than multi-year averages over
the historical period (Fig. 3). This study provides useful knowledge to
5). tackling future water scarcity by explicitly identifying these regions.
The AWSI of future dry years has more dramatic increases relative to
the AWSI of all historical years (Fig. 5c and d), with 24 and 26 provinces 4.2. Measurements to tackle the impacts of precipitation variations
showing increasing AWSI under RCP2.6 and RCP6.0 scenarios. In
addition, we found that the ECCd,m are even higher than 100 % in 8 and Future climate change would worsen agricultural water scarcity in
7 provinces under RCP2.6 and RCP6.0 scenarios, respectively. It means many provinces of China, for example Inner Mongolia and several
that AWSI over future dry years would more than double the current southwest provinces. Moreover, variations of future precipitation to­
average AWSI in those provinces. By comparing AWSI of dry years be­ wards dry condition further exacerbate the already water scarcity con­
tween future and historical periods, there are more provinces presenting dition there (Fig. 5). In such cases, different technologies and political
positive ECCd,d values than provinces with negative ECCd,d values measurements should be taken into consideration to alleviate the water
(Fig. 5e and f). It implies that future dry conditions would have more scarcity issues. Our study shows that improving IE is a useful approach
pronounced impacts on AWSI than historical dry conditions over the (Fig. 6). However, the effects of improving IE on decreasing AWSI is
provinces with positive ECCd,d values, especially in the southwest part of constrained in the regions with already high AWSI. This is mainly
China. However, the ECCd,d values are in quantity lower than ECCd,m. because IE is used to scale the blue water availability according to the
AWSI estimation method. With little blue water availability, improving
IE can only have minor effects on AWSI alleviation. Also, attentions
3.4. Impacts of irrigation efficiency on agricultural water scarcity should be paid to the unintended increases in overall water uses when
attempts to increase IE given the paradox of IE (Grafton et al., 2018).
Improving IE (irrigation efficiency) has potential to decrease the Redistribution of water resources between different seasons and
negative impacts of future dry precipitation condition on AWSI (Fig. 6). different years would partially offset the effects of hydrological varia­
For example, estimated AWSI during future dry years in Jiangsu prov­ tions but requiring smart water management technologies, e.g. reliable
ince using IE in 2010 has the highest value among the 31 provinces in precipitation forecast and rainwater harvest. In addition to the strategies
China. This high AWSI could decrease by 11 % and 20 % if China’s to increasing or stabilizing water resources, measurements to reducing
national average IE could increase to 0.6 and 0.7. Generally, improving crop water demand, e.g. sustainable intensification (Liu et al., 2018),
IE has stronger effects on declining AWSI in the provinces with lower reducing food losses and wastes (Xue et al., 2021), optimization of crop
AWSI. For example, AWSI in future dry years could decrease by 10 % cultivation areas (Davis et al., 2017), would also have high potentials to
and 18 % with national IE changing to 0.6 and 0.7 for the provinces with alleviate agricultural water scarcity.
AWSI (by using IE in 2010) lower than 1; while these decreases are only
7 % and 13 % for the province with AWSI (by using IE in 2010) higher

6
J. Liu et al. Journal of Hydrology 617 (2023) 128999

Fig. 5. Effects of future climate change on agricultural water scarcity index (AWSI) under scenarios RCP2.6 (left panel) and RCP6.0 (right panel) over the
period 2031–2070. Effect of climate change on AWSI over all future and historical years (ECCm,m, a, b); effect of climate change over the future dry years to all
historical years (ECCd,m, c, d); effect of climate change over the future and historical dry years (ECCd,d, e, f).

4.3. Future research dimension of integrated water scarcity and the AWSI using GSWP3 tends to be closer to the average value (Fig. S11).
climate change impacts Therefore, we used the average of hydrological simulations driven by
multi-climate datasets of historical and future periods to reduce possible
We applied the integrated AWSI, which incorporates blue water and uncertainties. More aspects of uncertainties should be included in future
green water in the estimation, to detect the impacts of precipitation research, e.g. from the choice of potential evapotranspiration estimation
variations under current and future climate conditions on water scarcity. method, global hydrological models, and land use scenarios. Besides, we
This integrated index is also with limitation. For example, it cannot demonstrate significant impacts of precipitation variations on AWSI and
detect water scarcity associated with water uses for other sectors than the impacts are associated with CV of precipitation and correlation be­
agriculture. Also, water quality and other resources, e.g. recycled water, tween precipitation and AWSI in China. However, our analyses were
and desalinated water and groundwater, should also be included in a conducted at the provincial level, mainly due to coarse resolution of
more comprehensive water scarcity index, considering water quality simulated water resources data from ISIMIP. However, the large size of
and other water resources could also have effects on water scarcity (van province ignores the spatial variability of green water, blue water, as
Vliet et al., 2021) and future climate would significantly affect water well as crop water demand within a province, which could dim the
quality and water resources (Liu et al., 2020). However, proposing such significance of the results. Future studies could apply the similar
a comprehensive water scarcity index is out of the scope of this study, approach to finer spatial scales, e.g. local river basin level, which will
but deserves attention in the future research. facilitate the detection of precipitation variations and correlation be­
In addition to providing a more general and comprehensive water tween precipitation and water scarcity, hence the impacts of precipita­
scarcity indicator, several other directions regarding the analyses of tion variations on agricultural water scarcity.
water scarcity index should be highlighted. The AWSI accounting covers
a few key processes of crop growth, human activities, and hydrological 5. Conclusions
cycles. It therefore is subjected to uncertainties from different resources,
e.g. from crop water use estimates (Liu et al., 2016) to water resource In this research, we evaluated an integrated agricultural water
calculation (Greve et al., 2018). It could be found that the PCR-GLBOWB scarcity index, i.e. AWSI, which incorporates green water and blue
simulated hydrological variables by using different climate forcing water, and linked the changes of AWSI to precipitation variations in
datasets could induce different estimations of AWSI, while the estimated China for the historical (1971–2010) and future (2031–2070) periods.

7
J. Liu et al. Journal of Hydrology 617 (2023) 128999

Fig. 6. Impacts of different irrigation efficiencies (IE) on agricultural water scarcity index (AWSI) for dry years over the period 2031–2070 under scenarios
RCP2.6 (a) and RCP6.0 (b). For each subplot, the upper panel shows the future average AWSI over future dry years with IE data in 2010; the lower panel shows the
relative changes of estimated AWSI over future dry years by using national average IE = 0.6 and national average IE = 0.7 to the estimated AWSI over future dry
years by using IE in 2010. Provinces are sorting by a descending order of AWSI under the RCP2.6 scenario.

We explicitly quantified the changes of AWSI from the long-term aver­ comprehensive water scarcity index and investigate the precipitation
ages. Annual precipitation shows high temporal variations, while AWSI variation impacts at finer spatial scales.
and precipitation presents strong negative correlations. AWSI during
extremely dry years (precipitation anomaly lower than –2σ) and dry CRediT authorship contribution statement
years (precipitation anomaly lower than –1σ) can be 31 % and 21 %
higher than the multi-year average, respectively. Future climate change Jiongjiong Liu: Formal analysis, Investigation, Visualization,
would further intensify agricultural water scarcity in many provinces, Writing – original draft. Zhonghao Fu: Software, Validation, Writing –
especially during future dry years with 24 and 26 provinces showing review & editing. Wenfeng Liu: Conceptualization, Funding acquisi­
increasing AWSI under RCP2.6 and RCP6.0. Qinghai and Inner Mongolia tion, Methodology, Supervision, Writing – review & editing.
provinces would have particularly high increases in AWSI under future
dry condition. Improving irrigation efficiency has potential to alleviate Declaration of Competing Interest
agricultural water scarcity, but with lower effects in the already water
scarce regions. The high positive changes of agricultural water scarcity The authors declare that they have no known competing financial
from multi-year average under both historical and future periods call for interests or personal relationships that could have appeared to influence
special attentions to deal with the impacts of precipitation variations on the work reported in this paper.
agricultural water scarcity. Future studies should propose more

8
J. Liu et al. Journal of Hydrology 617 (2023) 128999

Data availability Liu, W.F., Yang, H., Folberth, C., Muller, C., Ciais, P., Abbaspour, K.C., Schulin, R., 2018.
Achieving high crop yields with low nitrogen emissions in global agricultural input
intensification. Environ. Sci. Tech. 52 (23), 13782–13791.
Data will be made available on request. Liu, W.F., Ciais, P., Liu, X.C., Yang, H., Hoekstra, A.Y., Tang, Q.H., Wang, X.H., Li, X.D.,
Cheng, L., 2020. Global phosphorus losses from croplands under future precipitation
Acknowledgements scenarios. Environ. Sci. Tech. 54 (22), 14761–14771.
Liu, X.C., Tang, Q.H., Liu, W.F., Veldkamp, T.I.E., Boulange, J., Liu, J.G., Wada, Y.,
Huang, Z.W., Yang, H., 2019. A spatially explicit assessment of growing water stress
We are grateful to the Inter-Sectoral Impact Model Intercomparison in China from the past to the future. Earths Future 7 (9), 1027–1043.
Project (ISIMIP) for providing the global hydrological simulation data. Moursi, H., Kim, D., Kaluarachchi, J.J., 2017. A probabilistic assessment of agricultural
water scarcity in a semi-arid and snowmelt-dominated river basin under climate
This study was supported by the National Natural Science Foundation of change. Agric .Water Manag. 193, 142–152.
China (52109071 and 52239002) and the Chinese Universities Scientific MWRC & NDRCC 2022 “14th Five-Year Plan” double control target of total water
Fund (2022RC035). consumption and intensity. Ministry of Water Resources of China (MWRC), National
Development and Reform Commission of China (NDRCC). Beijing.
Nunez, M., Pfister, S., Roux, P., Anton, A., 2013. Estimating water consumption of
Appendix A. Supplementary data potential natural vegetation on global dry lands: Building an LCA framework for
green water flows. Environ. Sci. Tech. 47 (21), 12258–12265.
Oki, T., Kanae, S., 2006. Global hydrological cycles and world water resources. Science
Supplementary data to this article can be found online at https://doi. 313 (5790), 1068–1072.
org/10.1016/j.jhydrol.2022.128999. Pastor, A.V., Ludwig, F., Biemans, H., Hoff, H., Kabat, P., 2014. Accounting for
environmental flow requirements in global water assessments. Hydrol. Earth Syst.
Sci. 18 (12), 5041–5059.
References
Pastor, A.V., Palazzo, A., Havlik, P., Biemans, H., Wada, Y., Obersteiner, M., Kabat, P.,
Ludwig, F., 2019. The global nexus of food-trade-water sustaining environmental
Allen R.G., Pereira, L.S., Raes, D., Smith, M. 1998 Crop evapotranspiration: guidelines for flows by 2050. Nat. Sustainability 2 (6), 499–507.
computing crop water requirements FAO Irrigation and Drainage Paper 56 (Vol. 300). Portmann, F.T., Siebert, S., Doll, P., 2010. MIRCA2000-Global monthly irrigated and
FAO. rainfed crop areas around the year 2000: a new high-resolution data set for
Cao, X., Wu, M., Guo, X., Zheng, Y., Gong, Y., Wu, N., Wang, W., 2017. Assessing water agricultural and hydrological modeling. Glob. Biogeochem. Cycles 24, GB1011.
scarcity in agricultural production system based on the generalized water resources Postel, S.L., Daily, G.C., Ehrlich, P.R., 1996. Human appropriation of renewable fresh
and water footprint framework. Sci. Total Environ. 609, 587–597. water. Science 271 (5250), 785–788.
Coffel, E.D., Keith, B., Lesk, C., Horton, R.M., Bower, E., Lee, J., Mankin, J.S., 2019. Qin, Y., Mueller, N.D., Siebert, S., Jackson, R.B., AghaKouchak, A., Zimmerman, J.B.,
Future hot and dry years worsen Nile Basin water scarcity despite projected Tong, D., Hong, C.P., Davis, S.J., 2019. Flexibility and intensity of global water use.
precipitation increases. Earth’s Future 7 (8), 967–977. Nat. Sustain. 2 (6), 515–523.
CPGPRC, 2012. Opinions of the State Council on implementing the strictest water Quinteiro, P., Rafael, S., Villanueva-Rey, P., Ridoutt, B., Lopes, M., Arroja, L., Dias, A.C.,
resources management system. The Central People’s Government of the People’s 2018. A characterisation model to address the environmental impact of green water
Republic of China (CPGPRC), Beijing. flows for water scarcity footprints. Sci. Total Environ. 626, 1210–1218.
Davis, K.F., Rulli, M.C., Seveso, A., D’Odorico, P., 2017. Increased food production and Quinteiro, P., Rafael, S., Vicente, B., Marta-Almeida, M., Rocha, A., Arroja, L., Dias, A.C.,
reduced water use through optimized crop distribution. Nat. Geosci. 10 (12), 2019. Mapping green water scarcity under climate change: a case study of Portugal.
919–924. Sci. Total Environ. 696, 134024.
Dirmeyer, P.A., Gao, X.A., Zhao, M., Guo, Z.C., Oki, T.K., Hanasaki, N., 2006. GSWP-2 - Ren, D.D., et al., 2021. The Land-Water-Food-Environment nexus in the context of
Multimodel anlysis and implications for our perception of the land surface. Bull. Am. China’s soybean import. Adv. Water Resour. 151, 103892.
Meteorol. Soc. 87 (10), 1381–1398. Richter, B.D., et al., 2020. Water scarcity and fish imperilment driven by beef production.
Falkenmark, M., 2013. Growing water scarcity in agriculture: future challenge to global Nat. Sustain. 3 (4), 319–328.
water security. Philos. Trans. Royal Soc. A-Math. Phys. Eng. Sci. 371, 20120410. Schewe, J., et al., 2014. Multimodel assessment of water scarcity under climate change.
Gampe, D., Nikulin, G., Ludwig, R., 2016. Using an ensemble of regional climate models Proc. Natl. Acad. Sci. USA 111 (9), 3245–3250.
to assess climate change impacts on water scarcity in European river basins. Sci. Sheffield, J., Goteti, G., Wood, E.F., 2006. Development of a 50-year high-resolution
Total Environ. 573, 1503–1518. global dataset of meteorological forcings for land surface modeling. J. Clim. 19 (13),
Gao, T.G., Wang, X.M., Wei, D., Wang, T., Liu, S.W., Zhang, Y.L., 2021. Transboundary 3088–3111.
water scarcity under climate change. J. Hydrol. 598, 126453. Sutanudjaja, E.H., van Beek, R., Wanders, N., Wada, Y., Bosmans, J.H.C., Drost, N., van
Gerten, D., et al., 2020. Feeding ten billion people is possible within four terrestrial der Ent, R.J., de Graaf, I.E.M., Hoch, J.M., de Jong, K., Karssenberg, D., López
planetary boundaries. Nat. Sustain. 3 (3), 200–208. López, P., Peßenteiner, S., Schmitz, O., Straatsma, M.W., Vannametee, E., Wisser, D.,
Grafton, R.Q., et al., 2018. The paradox of irrigation efficiency. Science 361 (6404), Bierkens, M.F.P., 2018. PCR-GLOBWB 2: a 5 arcmin global hydrological and water
748–750. resources model. Geosci. Model Dev. 11 (6), 2429–2453.
Greve, P., et al., 2018. Global assessment of water challenges under uncertainty in water van Vliet, M.T.H., Jones, E.R., Florke, M., Franssen, W.H.P., Hanasaki, N., Wada, Y.,
scarcity projections. Nat. Sustain. 1 (9), 486–494. Yearsley, J.R., 2021. Global water scarcity including surface water quality and
Hoekstra, A.Y., 2014. Water scarcity challenges to business. Nat. Clim. Change 4 (5), expansions of clean water technologies. Environ. Res. Lett. 16 (2), 024020.
318–320. Veettil, A.V., Mishra, A.K., 2016. Water security assessment using blue and green water
Hoekstra, A., Mekonnen, M., 2012. The water footprint of humanity. Proc. Natl. Acad. footprint concepts. J. Hydrol. 542, 589–602.
Sci. USA 109, 3232–3237. Wada, Y., Wisser, D., Bierkens, M.F.P., 2014. Global modeling of withdrawal, allocation
Huang, J.P., Yu, H.P., Dai, A.G., Wei, Y., Kang, L.T., 2017. Drylands face potential threat and consumptive use of surface water and groundwater resources. Earth Syst. Dyn. 5
under 2 degrees C global warming target. Nat. Clim. Change 7 (6), 417–422. (1), 15–40.
Jagermeyr, J., Pastor, A., Biemans, H., Gerten, D., 2017. Reconciling irrigated food Wanders, N., Wada, Y., 2015. Human and climate impacts on the 21st century
production with environmental flows for Sustainable Development Goals hydrological drought. J. Hydrol. 526, 208–220.
implementation. Nat. Commun. 8, 15900. Ward, F.A., Pulido-Velazquez, M., 2008. Water conservation in irrigation can increase
Jones, P.D., Hulme, M., 1996. Calculating regional climatic time series for temperature water use. Proc. Natl. Acad. Sci. USA 105 (47), 18215–18220.
and precipitation: methods and illustrations. Int. J. Climatol. 16 (4), 361–377. Weedon, G.P., Gomes, S., Viterbo, P., Shuttleworth, W.J., Blyth, E., Osterle, H., Adam, J.
Kummu, M., Gerten, D., Heinke, J., Konzmann, M., Varis, O., 2014. Climate-driven C., Bellouin, N., Boucher, O., Best, M., 2011. Creation of the WATCH Forcing Data
interannual variability of water scarcity in food production potential: a global and its use to assess global and regional reference crop evaporation over land during
analysis. Hydrol. Earth Syst. Sci. 18 (2), 447–461. the twentieth century. J. Hydrometeorol. 12 (5), 823–848.
Li, Y., Guan, K.Y., Schnitkey, G.D., Delucia, E., Peng, B., 2019. Excessive rainfall leads to Weedon, G.P., Balsamo, G., Bellouin, N., Gomes, S., Best, M.J., Viterbo, P., 2014. The
maize yield loss of a comparable magnitude to extreme drought in the United States. WFDEI meteorological forcing data set: WATCH Forcing Data methodology applied
Glob. Chang. Biol. 25 (7), 2325–2337. to ERA-Interim reanalysis data. Water Resour. Res. 50 (9), 7505–7514.
Liu, J.G., et al., 2017. Water scarcity assessments in the past, present, and future. Earths Xie, P., Zhuo, L., Yang, X., Huang, H., Gao, X., Wu, P., 2020. Spatial-temporal variations
Future 5 (6), 545–559. in blue and green water resources, water footprints and water scarcities in a large
Liu, J.G., Zehnder, A.J.B., Yang, H., 2009. Global consumptive water use for crop river basin: a case for the Yellow River Basin. J. Hydrol. 590 (4), 125222.
production: the importance of green water and virtual water. Water Resour. Res. 45. Xue, L., Liu, X.J., Lu, S.J., Cheng, G.Y., Hu, Y.C., Liu, J.G., Dou, Z.X., Cheng, S.K., Liu, G.,
Liu, W.F., Liu, X.C., Yang, H., Ciais, P., Wada, Y., 2022a. Global water scarcity 2021. China’s food loss and waste embodies increasing environmental impacts. Nat.
assessment incorporating green water in crop production. Water Resour. Res. 58 (1). Food 2 (7), 519–528.
Liu, X.C., Liu, W.F., Tang, Q.H., Liu, B., Wada, Y., Yang, H., 2022b. Global agricultural Zhang, L., Chen, F., Lei, Y.D., 2020. Climate change and shifts in cropping systems
water scarcity assessment incorporating blue and green water availability under together exacerbate China’s water scarcity. Environ. Res. Lett. 15 (10), 104060.
future climate change. Earths Future 10 (4). Zhao, X., et al., 2015. Physical and virtual water transfers for regional water stress
Liu, W.F., Yang, H., Folberth, C., Wang, X.Y., Luo, Q.Y., Schulin, R., 2016. Global alleviation in China. Proc. Natl. Acad. Sci. USA 112 (4), 1031–1035.
investigation of impacts of PET methods on simulating crop-water relations for Zhu, W.B., Jia, S.F., Devineni, N., Lv, A.F., Lall, U., 2019. Evaluating China’s water
maize. Agric. For. Meteorol. 221, 164–175. security for food production: the role of rainfall and irrigation. Geophys. Res. Lett. 46
(20), 11155–11166.

You might also like