Access to this full-text is provided by IOP Publishing.
Content available from Environmental Research Letters
This content is subject to copyright. Terms and conditions apply.
Environ. Res. Lett. 16 (2021) 044053 https://doi.org/10.1088/1748-9326/abe782
OPEN ACCESS
RECEIVED
23 September 2020
REVISED
15 February 2021
ACC EPT ED FOR PUB LICATI ON
18 February 2021
PUBLISHED
6 April 2021
Original content from
this work may be used
under the terms of the
Creative Commons
Attribution 4.0 licence.
Any further distribution
of this work must
maintain attribution to
the author(s) and the title
of the work, journal
citation and DOI.
LETTER
CMIP6 projects less frequent seasonal soil moisture droughts over
China in response to different warming levels
Sisi Chen and Xing Yuan
School of Hydrology and Water Resources, Nanjing University of Information Science and Technology, Nanjing, Jiangsu 210044, People’s
Republic of China
E-mail: xyuan@nuist.edu.cn
Keywords: seasonal drought, soil moisture, global warming, CMIP
Supplementary material for this article is available online
Abstract
Seasonal drought occurrences are found to increase across different regions over China under
global warming, but with large uncertainties among models. With ten selected Coupled Model
Intercomparison Project Phase 5 (CMIP5) climate models and seven CMIP6 models according to
their performances in reproducing historical drought trends (p< 0.1), here we show that future
seasonal soil moisture (SM) droughts over China projected by CMIP6 models are less frequent
than that by CMIP5 models. We find national mean seasonal drought frequency is projected to
increase by 28 ±4% based on CMIP5 models at 1.5 ◦C global warming level, but only increase by
18 ±6% based on CMIP6 models and 12 ±4% based on land surface model ensemble simulations
driven by downscaled CMIP5 models. Compared with CMIP6, CMIP5 projection suggests larger
increase in precipitation but also larger increase in evapotranspiration, leading to more frequent
seasonal SM droughts. Comparing the results at 3 ◦C global warming level with those at 1.5 ◦C,
drought frequency over China will increase further by 10 ±4%, but drought duration will decrease
by 6 ±4%, suggesting more frequent seasonal SM droughts with shorter durations will occur in a
warming future. The future increase in China drought frequency will reduce from 12%–45% based
on selected climate models to 3%–27% based on all available models (30 CMIP5 models and 31
CMIP6 models), which indicates that the model selection is critical for future drought projection.
Nevertheless, CMIP6 still projects less frequent seasonal SM droughts than CMIP5 even without
any model discriminations.
1. Introduction
Drought is a period with below-average precipitation
(P) and/or above-normal evaporation, all of which
reflected in the land surface with declined soil mois-
ture (SM). Drought caused by such SM deficit is often
referred to as agricultural drought, leading to reduced
crop production and plant growth (Dai 2011, Li et al
2018). Due to its slow development, it is difficult to
detect drought until it has caused devastating losses,
bringing out famine, migration, and potential con-
flict (Berg and Sheffield 2018). In a warming future,
droughts will become more frequently (Sheffield and
Wood 2008, Dai 2013, Zhao and Dai 2015, Liu et al
2018, Samaniego et al 2018, Yuan et al 2019, Gu
et al 2020), and approximately two thirds of global
population will experience a progressive increase in
drought conditions (Naumann et al 2018), leading to
large economic losses, ecosystem damages and water
shortages, and increasing the risk of wildfires and
heatwaves.
The current rate of warming is too fast for the
climate system to adapt to, leading to frequently
occurred extreme events in recent years. As a result,
the 2 ◦C warming level was first proposed in 1996 to
avoid irreversible risk from climate extremes includ-
ing droughts. The Copenhagen accord of 2009 expli-
citly accepted the 2 ◦C target. To further mitigate
the impact of global warming, the 2015 Paris Agree-
ment emphasized that ‘holding the increase in the
global average temperature to well below 2 ◦C above
pre-industrial levels and pursuing efforts to limit the
temperature increase to 1.5 ◦C above pre-industrial
levels’. Previous study suggests that the temperature
© 2021 The Author(s). Published by IOP Publishing Ltd
Environ. Res. Lett. 16 (2021) 044053 S Chen and X Yuan
in China is expected to increase by 1.3 ◦C–5 ◦C by
the end of this century, which is higher than the aver-
age warming level in the world (Chen and Sun 2015).
Hence, it is necessary to assess China drought changes
under different global warming levels.
A few studies focused on global drought changes
under different warming levels, but their conclusions
are not necessarily consistent over China. Based on
the Coupled Model Intercomparison Project Phase 5
(CMIP5) model outputs, Lehner et al (2017) found
that Palmer Drought Severity Index (PDSI) showed a
non-significant decreasing (drying) trend over South
China at the global warming levels of 1.5 ◦C and
2◦C, but Su et al (2018) found that the PDSI showed
a significant decreasing trend over South China and
a significant increasing trend over North China. By
using SM directly from the CMIP5 climate mod-
els, Lei et al (2019) found that seasonal droughts
over North China will increase at the warming level
of 1.5 ◦C, although the mean SM will increase. A
recent study from Cook et al (2020) showed that
the latest CMIP6 models projected a wetting SM
condition over North China at the end of the 21st
century, while the CMIP5 models projected a dry-
ing condition. Whether this affects the assessment of
seasonal SM drought changes over China needs fur-
ther investigation. In addition, many studies argued
that SM from CMIP models might have large uncer-
tainty due to the coarse resolution of climate mod-
els and simple representation of land surface pro-
cesses, so they usually downscaled climate output
from CMIP models to drive advanced land surface
models (LSMs) to provide more reliable estimation
of SM, and assessed the corresponding changes in
drought in the future (Samaniego et al 2018, Yuan
et al 2019). Again, whether the CMIPs and LSMs com-
binations provide significantly different SM drought
projection remains unknown.
Therefore, this work used SM from three sets
of data, i.e. CMIP5 climate models, CMIP6 climate
models, and a set of LSMs ensemble simulations
driven by CMIP5 models (CMIP5/LSM), to pro-
ject the future drought changes over China under
1.5 ◦C, 2 ◦C and 3 ◦C global warming scenarios sep-
arately, aiming to answer the following questions: (a)
how will seasonal SM droughts change over China
at 1.5 ◦C global warming level? (b) After exceed-
ing 1.5 ◦C warming target, are there any signific-
ant changes in seasonal SM droughts under further
warming (e.g. 2 ◦C and 3 ◦C warming levels)? (c)
Compared with the results from CMIP5, do CMIP6
and CMIP5/LSM provide any new evidence regard-
ing the drought changes over China?
2. Data and method
2.1. CMIP5 and CMIP6 data
In this work, we firstly compiled surface air temper-
ature, P, evapotranspiration (ET), meridional wind
at 850 hPa and total column SM data from 30
CMIP5 models for historical simulations (1961–
2005) and future projections (2006–2099) under
Representative Concentration Pathway 4.5 (RCP4.5)
and RCP8.5 due to data availability (table S1).
Given that the reference data (i.e. OBS/LSM, see
section 2.2 for details) showed a significant increas-
ing trend (p< 0.01) in national mean dry months
(SM lower than 20th percentile) during 1961–2005,
we finally chose ten CMIP5 models considering
their ability (p< 0.1) to reproduce this upward
trend (Top ten in table S1 and figure S1 (available
online at stacks.iop.org/ERL/16/044053/mmedia)).
Here, p< 0.1 instead of p< 0.01 was used to keep
enough climate models for further analysis. As for
the CMIP6 data, we compiled data from 31 climate
models (table S2) for historical simulations (1961–
2014) and future projections (2015–2100) under two
Shared Society Pathways (i.e. SSP245 and SSP585).
The SSP245 and SSP585 in CMIP6 future projections
are corresponding to the scenarios of RCP4.5 and
RCP8.5 in CMIP5 simulations, respectively. Based on
the performance regarding the upward trend for his-
torical droughts (p< 0.1), seven CMIP6 models were
selected for further analysis (Top seven models in
table S2 and figure S2). We used ALL experiments in
historical simulations, which were forced by both nat-
ural (e.g. solar and volcanic activities) and anthropo-
genic (e.g. anthropogenic greenhouse gases and aero-
sol emissions) forcings. All data were interpolated to
0.5◦bilinearly.
2.2. LSM ensemble simulations driven by
observations and CMIP5 climate data
To assess the performance of simulating seasonal SM
drought by using CMIP models, we used the bench-
mark data (OBS/LSM) provided by Yuan et al (2019),
where the meteorological observations were used to
drive three LSMs (i.e. variable infiltration capacity
(VIC; Xie et al 2007), Noah LSM with multiple para-
meterization options (Noah-MP; Niu et al 2011)
and Community Land Model version 4.5 (CLM4.5;
Oleson et al 2013)) to simulate SM. These three mod-
els have good performance for hydrological simula-
tions (Niu et al 2011, Zong et al 2011), and they have
been widely used and verified.
Yuan et al (2019) also selected 13 CMIP5 mod-
els, and bias corrected the simulated P and temperat-
ure through a quantile-mapping method (Wood et al
2002) and an equal distance cumulative distribution
function-matching method (Li et al 2010) during his-
torical and future periods, respectively. These bias-
corrected P and temperature simulations were then
used to drive the above three LSMs to provide SM
simulations (Yuan et al 2019), which are referred to as
CMIP5/LSM in this study. Again, only CMIP5/LSM
simulations that can capture the historical increasing
trend (p< 0.1) of national mean dry months (SM
lower than 20th percentile) were selected (table S3
2
Environ. Res. Lett. 16 (2021) 044053 S Chen and X Yuan
and figure S3). The difference between CMIP5 and
CMIP5/LSM not only originates from the bias correc-
tion of meteorological forcings, but also the advances
in three selected LSMs (i.e. VIC, Noah-MP, CLM4.5)
for representing land surface hydrology.
2.3. Identification of different warming periods
The period 1971–2000 was selected to represent
present-day conditions, and it corresponded to a
global warming of 0.46 ◦C to pre-industrial con-
ditions in 1881–1910 (Vautard et al 2014). Con-
sequently, we used this period as the baseline period,
and defined 1.5 ◦C, 2 ◦C and 3 ◦C global warm-
ing levels as the global surface air temperature
increases by 1.04 ◦C (=1.5 ◦C−0.46 ◦C), 1.54 ◦C
(=2◦C−0.46 ◦C), and 2.54 ◦C (=3◦C−0.46 ◦C)
compared with the baseline period, respectively (Jiao
and Yuan 2019). A 30 year moving window, which has
the same length as the baseline period, was used to
determine the first period reaching a specific warm-
ing level for a specific future scenario for CMIP5,
CMIP5/LSM and CMIP6 simulations.
2.4. Definition of seasonal SM drought
SM is a key state variable of the land surface, reflect-
ing complex interactions between the water, energy,
and carbon cycles (Berg and Sheffield 2018). This
work used the percentiles of SM to define drought
(Andreadis et al 2005). According to this index, a
seasonal drought event was defined as SM percentile
below 20% for no less than three consecutive months
(Yuan and Wood 2013). To characterize the seasonal
drought event, frequency was calculated as the aver-
age number of seasonal drought events that occurred
per year, duration was obtained by calculating the
average months that drought events last, and the
severity was the sum of deviation from 20% in per-
centiles averaged per events.
3. Results
The spatial distributions of frequency, duration and
severity of seasonal SM droughts in the baseline
period (1971–2000) are shown in figure S4. For each
dataset, the drought characteristics of each model
were calculated first and then averaged. Compared
with OBS/LSM, it can be found that CMIP5/LSM
simulates drought characteristics well, followed by
CMIP6 and CMIP5 (figure S4). All datasets show
a similar pattern where the frequency of seasonal
droughts is higher over South China and lower over
North and Northeast China (figures S4(a)–(d)). The
patterns for the mean duration and severity are differ-
ent from the drought frequency, where the droughts
over North and Northeast China have longer dura-
tions and larger severity than those over South China
(figures S4(e)–(l)). The results are consistent with
Wang and Yuan (2018), where SM from reanalysis
were used to estimate seasonal drought character-
istics. In fact, South China is located in a humid
region with a large climate variability, where sea-
sonal droughts would occur more frequently but with
shorter durations than those over North and North-
east China with semiarid and semi-humid climate.
With reasonable simulations of historical seasonal
droughts, these CMIP models were used for future
projections. Figure 1shows the percentage changes
of seasonal drought frequency under different warm-
ing levels as compared with the baseline period. With
1.5 ◦C warming, CMIP5 shows the largest area of
drought frequency increase, followed by CMIP6 and
CMIP5/LSM. All three datasets indicate that more
seasonal droughts would occur over South China,
with mixed changes over North and Northeast China
(figures 1(a), (d) and (g)). This suggests that the sea-
sonal drought-prone region (e.g. South China) will
face increasing drought frequency in the future. In
contrast to the CMIP6 results, CMIP5 not only pre-
dicts more seasonal drought over South China, but
also over parts of the North China (figures 1(a)–(c)).
For the changes in drought duration (figure S5) and
severity (figure S6), there is no clear North-South
contrast in the spatial distributions as the temperat-
ure increases.
Figure 2shows the changes in drought frequency
and duration averaged over China and three sub-
regions (figure 1(g)). CMIP5 suggests that the fre-
quency of seasonal SM drought over China will
increase by 28 ±4%, 31 ±8% and 45 ±13% at the
warming levels of 1.5 ◦C, 2 ◦C and 3 ◦C. However,
CMIP5/LSM suggests that the drought frequency will
only increase by 12 ±4%, 12 ±4% and 20 ±7%
(figure 2(a)). The results from CMIP6 are between
them, where the frequency will increase by 18 ±6%,
19 ±8% and 25 ±11%. The mean results from
three datasets suggest that the drought frequency over
China will increase by 17 ±3%, 18 ±4% and 26 ±7%
at 1.5 ◦C, 2 ◦C and 3 ◦C global warming levels. The
drought frequency increase projected by CMIP5 over
China is 34%–56%, 38%–60% and 44%–55% more
than CMIP5/LSM and CMIP6 under different warm-
ing levels. The seasonal drought durations averaged
over China also have increasing trends, and CMIP5
projects larger changes again (figure 2(b)).
The changes in regional mean results are elu-
sive as compared with the national mean res-
ults. For Northeast China (39–54◦N, 121–135◦
E), both CMIP5 and CMIP6 show increases in
drought frequency at 1.5 ◦C, 2 ◦C, 3 ◦C global
warming levels, but CMIP5/LSM shows insignificant
decreases (figure 2(c)). As the warming continues,
the uncertainties of the changes increase for all data-
sets (figure 2(c)). All three datasets show significant
increases in drought duration over Northeast China
at different warming levels, except for CMIP5/LSM
at 2 ◦C warming level (figure 2(d)). For North
China (34◦–42◦N, 98◦–121◦E), both CMIP6 and
3
Environ. Res. Lett. 16 (2021) 044053 S Chen and X Yuan
Figure 1. Projected percentage changes (%) in seasonal soil moisture drought frequency under 1.5 ◦C (left column), 2 ◦C (middle
column) and 3 ◦C (right column) global warming levels with respect to the reference period 1971–2000 (rows from top to bottom
are for simulations from CMIP5, CMIP5/LSM and CMIP6 respectively). The dotted areas indicate that at least 80% of the models
agree on sign of the changes.
CMIP5 show increases in drought frequency at dif-
ferent warming levels (figure 2(e)), but CMIP5/LSM
shows decreases (figure 2(e)). All three datasets show
increases in drought duration at different warming
levels over North China, but CMIP5 projects 46%–
88% and 49%–66% more increases than CMIP5/LSM
and CMIP6, respectively (figure 2(f)). For South
China (21◦–34◦N, 98◦–121◦E), all three datasets
project significant increases in both drought fre-
quency and duration (figures 2(g) and (h)), which
suggests that adaption for short-term droughts is
needed for humid regions. This is also consistent with
Yuan et al (2019), where South China shows a signific-
ant increase in flash droughts (with durations ranging
from 15 to 60 d) in the warming future.
To better understand the differences between
CMIP5 and the other two datasets, we investigated
changes in P, ET, P surplus (P-ET) and SM at 1.5 ◦C
warming level (figure 3). There are increases in P
over Northeast and North China, with different mag-
nitudes for different datasets (figures 3(a)–(c)). The
changes in P-ET over part of North China from
CMIP5 are opposite with that from CMIP5/LSM
and CMIP6 (figures 3(g)–(i)), suggesting that ET
modeling is quite different from different datasets
(figures 3(d)–(f)). As a result, SM increases over part
of North China based on CMIP5/LSM and CMIP6
data (figures 3(k) and (l)), leading to the decreases
or relatively small increase in drought frequency. For
South China, all three datasets project non-significant
changes in P and ET, except for CMIP5/LSM with
significant increases in ET (figures 3(a)–(f)). The
P-ET and SM decrease over parts of South China
for CMIP5 and CMIP5/LSM, while shows non-
significant change for CMIP6 (figures 3(g)–(l)).
Therefore, the reasons for the increases in drought
frequency over South China are different among three
datasets, where the decrease in P-ET (figures 3(g) and
(h)) and the consequently decreased SM (figures 3(j)
and (k)) are responsible for the drought increases for
CMIP5 and CMIP5/LSM. But for CMIP6, the mean
SM does not change significantly. To further explore
the reason, we calculated the PDF of annual mean
P, ET and SM percentile averaged over South China
(figure S9). We find that the PDF of ET skews towards
the right tail (figures S9(d)–(f)), while the SM per-
centile skews towards the left tail (figures S9(g)–(i)),
indicating that SM dynamics instead of mean SM
change might be related to the drought increase for
CMIP6 over South China (figure 3(l)). The changes
in these hydrometeorological variables at 2 ◦C warm-
ing level (figure S7) are similar to those at 1.5 ◦C
(figure 3). The P increases at 3 ◦C warming level are
more than at 1.5 ◦C warming level over most parts
of China (figures S8(a)–(c)), but ET also increases
(figures S8(d)–(f)).
4
Environ. Res. Lett. 16 (2021) 044053 S Chen and X Yuan
Figure 2. Projected percentage changes (%) in seasonal soil moisture drought frequency (left column) and duration (right
column) averaged over China and its three subregions as indicated in figure 1(g). The three subregions are Northeast China
(39–54◦N, 121–135◦E), North China (34◦–42◦N, 98◦–121◦E) and South China (21◦–34◦N, 98◦–121◦E). The bars show
5%–95% uncertainties which were estimated by using bootstrapping for 1000 times.
Previous study illustrated that a further increase
of 0.5 ◦C (global mean temperature rises from
1◦C to 1.5 ◦C with respect to the pre-industrial
period) posed greater risk than the previous increase
of 0.5 ◦C, and this ‘acceleration risk’ principle
may also be present in the next 0.5 ◦C addi-
tional warming (global mean temperature rises
from 1.5 ◦C to 2 ◦C relative to the pre-industrial
period) (Hoegh-Guldberg et al 2019). Therefore,
we compared changes in drought characteristics
under 2 ◦C and 3 ◦C warming levels relative to those
under 1.5 ◦C warming. For the average changes over
China, drought frequency will increase with addi-
tional warming of 0.5 ◦C or 1.5 ◦C (figures 4(a) and
(c)), but the change is not significant in CMIP5/LSM
and CMIP6 for additional warming of 0.5 ◦C
(figure 4(a)). Over Northeast and North China, all
three datasets show that P-ET will increase with
additional warming of 0.5 ◦C or 1.5 ◦C (figures
S10(a)–(c) and (g)–(i)), with a smaller magnitude for
5
Environ. Res. Lett. 16 (2021) 044053 S Chen and X Yuan
Figure 3. Projected percentage changes (%) in mean precipitation (P), evapotranspiration (ET), P surplus (P-ET) and soil
moisture (SM) under 1.5 ◦C warming level relative to the reference period 1971–2000 (columns from left to right are for
simulations from CMIP5, CMIP5/LSM and CMIP6 respectively). The dots indicate that at least 80% of models agree on sign of
the changes.
CMIP5. As a result, most datasets project a decrease
in both drought frequency and duration over North-
east and North China, although the changes are not
necessarily significant (figure 4). For South China,
all three datasets display a significant increasing
trend in drought frequency (figures 4(a) and (c)),
except for CMIP5/LSM at 0.5 ◦C additional warming
(figure 4(a)). The changes in drought duration over
South China with additional warming are not signi-
ficant. With additional warming of 1.5 ◦C, drought
frequency over China will increase by 10 ±4%, and
drought duration will decrease by 6 ±4%, suggesting
more frequent seasonal SM droughts with shorter
durations will occur. However, the effects of holding
temperature within 1.5 ◦C on seasonal SM droughts
are regionally dependent with large uncertainties.
4. Summary and discussion
This study explores the responses of seasonal SM
drought over China at different global warming levels
by using three datasets. At 1.5 ◦C warming level, the
average drought frequency over China will increase by
28 ±4%, 12 ±4% and 18 ±6% based on CMIP5,
CMIP5/LSM and CMIP6 datasets respectively. The
increases in drought frequency are 31 ±8%, 12 ±4%
and 19 ±8% for 2 ◦C warming, and 45 ±13%,
20 ±7% and 25 ±11% for 3 ◦C warming. Thus, hold-
ing the temperature at 1.5 ◦C has a positive effect on
controlling seasonal SM droughts over China.
Compared with CMIP5/LSM and CMIP6, CMIP5
shows 34%–56%, 38%–60% and 44%–55% larger
increases in drought frequency over China at 1.5 ◦C,
2◦C and 3 ◦C warming levels, which may be induced
by larger increase in ET projected by CMIP5. The
most significant increases in drought frequency and
duration occur over South China, where the decrease
in P-ET and the consequently decreased SM are
responsible for the drought increases for CMIP5 and
CMIP5/LSM, but SM dynamics might be related to
the drought increase for CMIP6 because there is no
significant changes in mean SM.
6
Environ. Res. Lett. 16 (2021) 044053 S Chen and X Yuan
Figure 4. Projected percentage changes in seasonal soil moisture drought frequency (left column) and duration (right column)
with continuous warming of 0.5 ◦C (top row) and 1.5 ◦C (bottom row) as compared with 1.5 ◦C warming. The bars show
5%–95% uncertainties which were estimated by using bootstrapping for 1000 times.
Given that East Asian summer monsoon (EASM)
greatly affects the P (drought) over China, it is neces-
sary to explore the change in EASM under global
warming. Here we use the meridional wind at 850 hPa
over East China (20◦–40◦N, 105◦–120◦E) to meas-
ure EASM (Jiang et al 2020), and find that both the
selected ten CMIP5 model ensemble and the seven
CMIP6 model ensemble show that EASM will be
intensified in response to different warming levels
(figure S11). A few studies have also shown that the
circulation of EASM is intensified by global warm-
ing (Lee and Wang 2014, Chen and Bordoni 2016, Jin
and Stan 2019, Park et al 2020). The enhancing mon-
soon will increase P over North China and decrease
P over South China, which is also consistent with
the P changes shown in figures 3(a) and (c). But
figures 3(d) and (f) show that ET also increases over
North China, which is not only caused by the increas-
ing P, but also due to increasing atmospheric water
demands with increasing temperature. Therefore, the
relation between EASM and SM drought is complic-
ated, where the changes in both P and ET are crit-
ical. Our results are consistent with previous stud-
ies, where they show that Asian monsoon region will
expose to more frequent seasonal drought conditions
despite the intensification of extreme rainfall events
(Ha et al 2020). Figure S11 also shows that CMIP6
projected monsoon wind increase is a little larger than
CMIP5, but again this does not necessarily suggest
more rainfall or less drought, as the change in sea-
sonal SM drought not only depends on the change in
rainfall seasonality, but also the change in ET. In addi-
tion, the warming of the equatorial eastern Pacific can
last from winter to spring-summer in the next year,
resulting in an enhanced delay relationship between
EASM and ENSO. Therefore, more droughts tend to
occur in South China in the years associated with the
summer of El Niño decay (Zhang and Zhou 2015).
In this study, we selected a subset of climate mod-
els according to their capability in reproducing his-
torical increase in SM droughts over China. If all
the available climate models were used for the ana-
lysis, the increase in drought would be reduced from
12%–45% to 3%–27% (figure S12). This is because
when we mix the models that have positive responses
(drought increase) to warming with those have neg-
ative (drought decrease) or insignificant responses,
the signal would be heavily smoothed. Therefore,
model selection is crucial for projecting drought in
the future.
Given that the 1.5 ◦C global warming target will
be reached in the near future, drought adaptation is
dependent on whether there is significant response
to additional warming (e.g. 2 ◦C or 3 ◦C warming
levels). Our results show that drought frequency over
China will increase by 10 ±4% with additional warm-
ing of 1.5 ◦C, but drought duration will decrease by
6±4%. The response is elusive over Northeast and
7
Environ. Res. Lett. 16 (2021) 044053 S Chen and X Yuan
North China (figure 4) due to large uncertainty in
modeling P and land surface hydrological processes.
However, it is found that South China, the hotspot
of the future increase in seasonal SM droughts, will
experience more frequent droughts but with shorter
durations (figure 4), although the reduction in dur-
ation is not statistically significant. This is also con-
sistent with the intensification of flash drought over
South China (Yuan et al 2019). Flash drought is
regarded as a subseasonal drought with rapid onset,
and it is also preferable over humid regions. This
raises a challenge for drought early warning, where
these short-term droughts are difficult to predict
because of the absence of oceanic anomaly. Under-
standing of the interactions among droughts at differ-
ent time scales (Yuan et al 2020) would be an urgent
need to interpret the projection of different types of
droughts and to facilitate developing early warning
systems.
In addition to the physical mechanism and
early warning of droughts, future research should
also concentrate on investigating the ecological and
physiological feedback between drought and veget-
ation. Recently, a study has shown that early spring
greening caused by warming can modulate the occur-
rence and magnitude of agricultural drought in sum-
mer, which is mainly due to the enhanced ET and
thus the loss of SM. Moreover, this additional soil dry-
ing may further exacerbate summer heatwaves (Lian
et al 2020). Faster SM losses, on the other hand, trig-
ger additional P by raising atmospheric humidity,
and partially compensate the soil water deficit caused
by early greening. But there exists another biophys-
ical feedback, where the soil water deficit results in
less incoming surface radiation and lower energy-
controlled ET, which may inhibit cloud development
and subsequent P. Besides, these interaction pro-
cesses can also be amplified or diminished by tele-
connection, where large scale atmospheric circula-
tion redistributes P over downwind regions (Zeng
et al 2019). Unfortunately, the overall impacts of these
processes are still unclear (Lian et al 2020). Con-
versely, drought affected by soil water deficit and the
consequent increase in summer extreme events may
inhibit vegetation growth and ecosystem production,
especially for the arid and semiarid ecosystems sensit-
ive to summer water supply. Given that South China
with humid climate and dense vegetation will exper-
ience more seasonal droughts in a warming future,
whether the early greening due to CO2fertilization
amplifies the drought condition by pumping more
water from soil, or the reduction in vegetation sto-
matal conductance resulted from increased CO2con-
centration decreases ET and alleviates the drought
condition (Ji et al 2020), is a key question. Moreover,
recent studies have shown that equilibrium climate
sensitivity of the latest generation of CMIP6 is lar-
ger than CMIP5 models (Meehl et al 2020, Zelinka
et al 2020), which means that CMIP6 models might
be is more sensitive to the increase of CO2. Thus, such
physiological effect is elusive and may affect future
drought projection at different warming levels (Dai
et al 2018).
Data availability statement
The CMIP data support the findings of
this study are openly available (https://esgf-
node.llnl.gov/search/cmip6/,https://esgf-node.llnl.
gov/search/cmip5).
The data that support the findings of this study are
available upon reasonable request from the authors.
Acknowledgments
We thank Professor Aiguo Dai for providing con-
structive comments. This work was supported by
National Natural Science Foundation of China
(41875105), National Key R&D Program of China
(2018YFA0606002), and the Startup Foundation for
Introducing Talent of NUIST.
ORCID iD
Xing Yuan https://orcid.org/0000-0001-6983-
7368
References
Andreadis K M et al 2005 Twentieth-century drought in the
conterminous United States J. Hydrometeorol. 6985–1001
Berg A and Sheffield J 2018 Climate change and drought: the
soil moisture perspective Curr. Clim. Change Rep.
4180–91
Chen H P and Sun J Q 2015 Changes in climate extreme events in
China associated with warming Int. J. Climatol.
35 2735–51
Chen J and Bordoni S 2016 Early summer response of the East
Asian summer monsoon to atmospheric CO2forcing and
subsequent sea surface warming J. Clim. 29 5431–46
Cook B I, Mankin J S, Marvel K, Williams A P, Smerdon G E and
Anchukaitis K J 2020 Twenty-first century drought
projections in the CMIP6 forcing scenarios Earth’s Future
8e2019EF001461
Dai A G 2011 Drought under global warming: a review Wiley
Interdiscip. Rev. Clim. Change 245–65
Dai A G 2013 Increasing drought under global warming in
observations and models Nat. Clim. Change 3171
Dai A G, Zhao T and Chen J 2018 Climate change and drought: a
precipitation and evaporation perspective Curr. Clim.
Change Rep. 4301–12
Gu L, Chen J, Yin J, Sullivan S C, Wang H-M, Guo S, Zhang L and
Kim J-S 2020 Projected increases in magnitude and
socioeconomic exposure of global droughts in 1.5 ◦C and
2◦C warmer climates Hydrol. Earth Syst. Sci. 24 451–72
Ha K, Moon S, Timmermann A and Kim D 2020 Future changes
of summer monsoon characteristics and evaporative
demand over Asia in CMIP6 simulations Geophys. Res. Lett.
47 e2020GL087492
Hoegh-Guldberg O et al 2019 The human imperative of stabilizing
global climate change at 1.5 ◦CScience 365(6459) eaaw6974
Ji P, Yuan X, Ma F and Pan M 2020 Accelerated hydrological cycle
over the Sanjiangyuan region induces more streamflow
extremes at different global warming levels Hydrol. Earth
Syst. Sci. 24 5439–51
8
Environ. Res. Lett. 16 (2021) 044053 S Chen and X Yuan
Jiang D B, Hu D, Tian Z P and Lang X M 2020 Differences
between CMIP6 and CMIP5 models in simulating climate
over China and the East Asian monsoon Adv. Atmos. Sci.
37 1102–18
Jiao Y and Yuan X 2019 More severe hydrological drought events
emerge at different warming levels over the Wudinghe
watershed in northern China Hydrol. Earth Syst. Sci.
23 621–35
Jin Y and Stan C 2019 Changes of East Asian summer monsoon
due to tropical air-sea interactions induced by a global
warming scenario Clim. Change 153 341–59
Lee J-Y and Wang B 2014 Future change of global monsoon in the
CMIP5 Clim. Dyn. 42 101–19
Lehner F, Coats S, Stocker T F, Pendergrass A G, Sanderson B M,
Raible C C and Smerdon J E 2017 Projected drought risk in
1.5 ◦C and 2 ◦C warmer climates Geophys. Res. Lett.
44 7419–28
Lei X, Chen N C and Zhang X 2019 Global drought trends under
1.5 and 2 ◦C warming Int. J. Climatol. 39 2375–85
Li H B, Sheffield J and Wood E F 2010 Bias correction of monthly
precipitation and temperature fields from
Intergovernmental Panel on Climate Change AR4 models
using equidistant quantile matching J. Geophys. Res.
115 D10101
Li Y H et al 2018 Mechanisms and early warning of drought
disasters: an experimental drought meteorology research
over China (DroughtEX_China) Bull. Am. Meteorol. Soc.
100 673–87
Lian X et al 2020 Summer soil drying exacerbated by earlier spring
greening of northern vegetation Sci. Adv. 6eaax0255
Liu W B, Sun F, Lim W H, Zhang J, Wang H, Shiogama H and
Zhang Y 2018 Global drought and severe drought-affected
populations in 1.5 ◦C and 2 ◦C warmer worlds Earth Syst.
Dyn. 9267–83
Meehl G A, Senior C A, Eyring V, Flato G, Lamarque J-F,
Stouffer R J, Taylor K E and Schlund M 2020 Context for
interpreting equilibrium climate sensitivity and transient
climate response from the CMIP6 Earth system models Sci.
Adv. 6eaba1981
Naumann G, Alfieri L, Wyser K, Mentaschi L, Betts R A,
Carrao H, Spinoni J, Vogt J and Feyen L 2018 Global
changes in drought conditions under different levels of
warming Geophys. Res. Lett. 45 3285–96
Niu G-Y et al 2011 The community Noah land surface model with
multiparameterization options (Noah-MP): 1. Model
description and evaluation with local-scale measurements J.
Geophys. Res. 116 D12109
Oleson K W et al 2013 Technical description of version 4.5 of the
community land model (CLM) Report No. NCAR/TN-503
+STR, 420 (Boulder, CO: National Center for Atmospheric
Research)
Park J, Kim H, Simon Wang S Y, Jeong J-H, Lim K-S, LaPlante M
and Yoon J-H 2020 Intensification of the East Asian summer
monsoon lifecycle based on observation and CMIP6
Environ. Res. Lett. 15 0940b9
Samaniego L, Thober S, Kumar R, Wanders N, Rakovec O, Pan M,
Zink M, Sheffield J, Wood E F and Marx A 2018
Anthropogenic warming exacerbates European soil moisture
droughts Nat. Clim. Change 8421–6
Sheffield J and Wood E F 2008 Projected changes in drought
occurrence under future global warming from multi-model,
multi-scenario, IPCC AR4 simulations Clim. Dyn.
13 79–105
Su B et al 2018 Drought losses in China might double between the
1.5 ◦C and 2.0 ◦C warming Proc. Natl Acad. Sci.
115 10600–5
Vautard R et al 2014 The European climate under a 2 ◦C global
warming Environ. Res. Lett. 9034006
Wang L Y and Yuan X 2018 Two types of flash droughts over
China and their connections with seasonal droughts Adv.
Atmos. Sci. 35 1478–90
Wood A W, Maurer E P, Kumar A and Lettenmaier D P 2002
Long-range experimental hydrologic forecasting for the
eastern United States J. Geophys. Res. 107 4429
Xie Z H, Yuan F, Duan Q, Zheng J, Liang M and Chen F 2007
Regional parameter estimation of the VIC land surface
model: methodology and application to river basins in
China J. Hydrometeorol. 8447–68
Yuan X, Ma F, Li H and Chen S S 2020 A review on multi-scale
drought processes and prediction under global change
Trans. Atmos. Sci. 43 225–37 (in Chinese)
Yuan X, Wang L Y, Wu P L, Peng J, Sheffield J and Zhang M 2019
Anthropogenic shift towards higher risk of flash drought
over China Nat. Commun. 10 4661
Yuan X and Wood E F 2013 Multimodel seasonal forecasting of
global drought onset Geophys. Res. Lett. 40 4900–5
Zelinka M D, Myers T A, McCoy D T, Po-Chedley S,
Caldwell P M, Ceppi P, Klein S A and Taylor K E 2020
Causes of higher climate sensitivity in CMIP6 models
Geophys. Res. Lett. 47 e2019GL085782
Zeng D W, Yuan X and Roundy J K 2019 Effect of teleconnected
land–atmosphere coupling on Northeast China persistent
drought in spring–summer of 2017 J. Clim. 32 7403–20
Zhang L X and Zhou T J 2015 Drought over East Asia: a review J.
Clim. 28 3375–99
Zhao T B and Dai A G 2015 The magnitude and causes of global
drought changes in the 21st century under a low-moderate
emissions scenario J. Clim. 28 4490–512
Zong L Y et al 2011 The community Noah land surface model
with multiparameterization options (Noah-MP): 2.
Evaluation over global river basins J. Geophys. Res.
116 D12110
9
Available via license: CC BY 4.0
Content may be subject to copyright.
