Access to this full-text is provided by Wiley.
Content available from Earth's Future
This content is subject to copyright. Terms and conditions apply.
The Timing of Detectable Increases in Seasonal Soil Moisture
Droughts Under Future Climate Change
Sisi Chen
1,2
and Xing Yuan
1,2,3
1
Key Laboratory of Hydrometeorological Disaster Mechanism and Warning of Ministry of Water Resources/Collaborative
Innovation Center on Forecast and Evaluation of Meteorological Disasters, Nanjing University of Information Science and
Technology, Nanjing, China,
2
School of Hydrology and Water Resources, Nanjing University of Information Science and
Technology, Nanjing, China,
3
Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
Abstract Global warming exacerbates the increase of soil moisture drought by accelerating the water cycle,
posing potential threats to food security and ecological sustainability. The design of drought prevention and
mitigation policies should be based on the reliable detection of the future change signal in droughts, so it is
critical to know when the signal can be detected (Time of Emergence, ToE) in the background noise of the
climate system. While the ToE framework has been successfully applied for temperature‐related signal
detection, the ToE for changes in drought has not been well studied. Based on 66 Coupled Model
Intercomparison Project Phase 6 model ensemble members under four Shared Socio‐economic Pathways, we
conduct a global ToE analysis of seasonal soil moisture drought characteristics and discuss the impact of
different warming levels. Six subregions with robust increase in soil moisture droughts are identified. For
drought frequency, most of the subregion's ToE is centered around 2080, however for drought intensity it is
much earlier and can even reach around 2040 in AMZ. For drought frequency and drought intensity,
approximately 14%–22% and 47%–49% of global land areas would reach ToE in 21st century. The global land
areas with ToE of increasing droughts would increase by at least 1/5 when global warming level is kept to 2°C
rather than 1.5°C above pre‐industrial conditions. This suggests that limiting global warming can significantly
delay the emergence time of increases in seasonal soil moisture droughts, allowing additional adaptation time
for the drought‐related sectors.
Plain Language Summary Global warming has increased the frequency of droughts, especially soil
moisture droughts with widespread impacts. Thus, when would the soil moisture drought change exceed the
noise of the climate system and signal “emerge” (Time of Emergence, ToE) has far‐reaching implication for
relevant government departments and stakeholders to make corresponding adaptations. In this work, the future
changes in seasonal soil moisture droughts are projected by using Coupled Model Intercomparison Project
Phase 6 (CMIP6, multi‐model simulations of historical and future climate) data under four possible future
development scenarios (named Shared Socio‐economic Pathways), and the effects of controlling global
warming levels on the ToE of droughts are discussed. Results show that the frequency of seasonal soil moisture
droughts would increase over 42%–48% of global land areas in the future, and 14%–22% of global land areas
would show detectable signals of increased drought frequency before 2100. If the global warming level is
controlled within 2°C rather than 1.5°C, the areas with ToE would increase by at least one fifth, except for
drought intensity under the scenario of the most energy‐saving and emission reduction. This work highlights the
risk of rapid emergence of robust soil moisture drought increases under climate change.
1. Introduction
Climate change has caused a series of extreme events in many parts of the world (Collins et al., 2021; Mora
et al., 2017; Myhre et al., 2019; Pendergrass, 2018; Wang & Yuan, 2021; Yuan et al., 2023; Zhang et al., 2022),
and significantly affected the ecosystems, socio‐economics, and human well‐being (IPCC, 2013). Climate change
is becoming a major social issue of general concern. Some seemingly unrelated events are not as isolated as they
appear, and climate change is the common driver behind all these extreme events (Duan & Mu, 2018). Many
studies have focused on the risk assessment of future extreme events, especially the absolute magnitude of
extreme climate change. However, the absolute change may be within an adaptive range of the variable itself
(Rockström et al., 2009), and there may be a mechanism to deal with the change within this range (Mahlstein
et al., 2012). Only if it exceeds the variability of the climate system can it have a significant negative impact
RESEARCH ARTICLE
10.1029/2023EF004174
Key Points:
•The timing of detectable future
changes in seasonal soil moisture
droughts are identified through the
time of emergence framework
•Detectable increase signals in drought
frequency and intensity are predicted in
14%–22% and 47%–49% of land areas
by the end of this century
•The areas of detectable signals for
drought characteristics would increase
by 1/5 when limiting warming within
2°C compared to 1.5°C
Supporting Information:
Supporting Information may be found in
the online version of this article.
Correspondence to:
X. Yuan,
xyuan2@mail.iap.ac.cn
Citation:
Chen, S., & Yuan, X. (2024). The timing of
detectable increases in seasonal soil
moisture droughts under future climate
change. Earth's Future,12,
e2023EF004174. https://doi.org/10.1029/
2023EF004174
Received 7 OCT 2023
Accepted 8 MAY 2024
Author Contributions:
Conceptualization: Xing Yuan
Formal analysis: Sisi Chen, Xing Yuan
Funding acquisition: Xing Yuan
Investigation: Sisi Chen, Xing Yuan
Methodology: Sisi Chen
Project administration: Xing Yuan
Resources: Xing Yuan
Software: Sisi Chen
Supervision: Xing Yuan
Validation: Sisi Chen
Visualization: Sisi Chen, Xing Yuan
Writing – original draft: Sisi Chen
Writing – review & editing: Xing Yuan
© 2024. The Author(s).
This is an open access article under the
terms of the Creative Commons
Attribution‐NonCommercial‐NoDerivs
License, which permits use and
distribution in any medium, provided the
original work is properly cited, the use is
non‐commercial and no modifications or
adaptations are made.
CHEN AND YUAN 1 of 13
(Frame et al., 2017; Lobell & Burke, 2008). Therefore, the core issue of future climate change projection is to
distinguish between “signal” and “noise” (Hawkins & Sutton, 2012). When the change permanently exceeds the
background noise of natural variability, it can be said that it has “emerged,” and the year of its first emerge is
defined as Time of Emergence (ToE; Hawkins & Sutton, 2012; Lopez et al., 2018; Mahony & Cannon, 2018;
Silvy et al., 2020). ToE provides an indication of the time scale of climate change signals emergence and
associated risks. Its purposes are: (a) To provide a baseline of the temporal and spatial scales of monitoring
required for trend detection.This will inform the design of observing systems. (b) To inform impact studies. Once
anthropogenic trends exceed the range of natural variability to which biota are adapted, impacts on organisms and
ecosystems are likely to manifest themselves. Early detection of climate change signals is particularly important
and useful for vulnerability assessment, societal adaptation, and climate policy making. (c) To enable compar-
isons across variables, General Circulation Models (GCMs), and emission scenarios, anthropogenic responses
should be normalized to their natural internal variability. This will provide a framework for inter‐comparison of
models, scenarios, and impacts (Schlunegger et al., 2020).
The ToE framework has been applied to detect signals for meteorological variable such as temperature (Bador
et al., 2016; Hawkins et al., 2014; Ma et al., 2022; Raymond et al., 2020) and precipitation (Giorgi & Bi, 2009;
King et al., 2015; Landrum & Holland, 2020) as well as sea level rise (Lyu et al., 2014; Walker et al., 2022) and
carbon‐cycle (Lombardozzi et al., 2014; Schlunegger et al., 2019). In terms of drought extreme events that are
affected by decadal variability significantly, most studies focused on their changes in a certain period, while
nelgected the long‐term detectable signals based on the ToE framework. To the best of our knowledge, only few
studies focused on the ToE of drought events at the global scale (Satoh et al., 2022; Stevenson et al., 2022; Touma
et al., 2015). Touma et al. (2015) compared the Standardized Precipitation Index (SPI), the Standardized Runoff
Index (SRI), the Standardized Precipitation‐Evapotranspiration Index (SPEI) and the Supply Demand Drought
Index (SDDI) drought characteristics changes based on 15 Coupled Model Intercomparison Project Phase 5
(CMIP5) models under the Representative Concentration Pathway 8.5 (RCP8.5). It was found that the changes in
SPEI and SDDI‐based drought characteristics are more obvious than the changes in SPI and SRI‐based drought
characteristics due to the influence of temperature, and a permanent emergence of the spatial extent of drought in
some regions would occur during this century. Satoh et al. (2022) studied discharge‐based hydrological drought
changes under RCP2.6 and RCP8.5 scenarios and found that Southwestern South America, European Mediter-
ranean and North Africa would face detectable drought signals in the next 30 years, regardless of emission
scenario. Such drought conditions would be effectively mitigated if global warming is controlled. While both
efforts identify ToEs for hydrological and meteorological droughts, neither considers soil moisture drought,
which is a key factor affecting surface processes and may have greater extent and higher frequency compared to
other droughts (Cook et al., 2020; Orlowsky & Seneviratne, 2013). Stevenson et al. (2022) focused on mega‐
drought events based on soil moisture, that is, on a 15 years time scale, while studies on droughts at seasonal
scales are still missing. The changes in seasonal soil moisture droughts may have implications on vegetation
growth and carbon management. In addition, due to data limitations, most results are only based on the most
extreme scenario, that is, the CMIP5 RCP8.5 scenario. It is still uncertain whether there are differences in the
latest CMIP6 SSP585 scenario compared with CMIP5 RCP8.5 scenario, and how the results based on the more
likely development scenario (i.e., SSP245) will be.
In addition to drought events, a study combining the ToE of aridity and the warming levels proposed by the Paris
Agreement found that keeping global warming to 2°C compared to 1.5°C would prevent two‐thirds of the areas
from experiencing ToE of aridification (Park et al., 2018). However, all these previous studies were based on
CMIP5 data and no more than two emission scenarios, RCP8.5 and RCP2.6. Therefore, whether controlling
global warming is an effective way to reduce soil moisture drought “emerge" needs further investigation with new
CMIP models and more emission scenarios.
Therefore, this study used the CMIP6 multi‐model data to investigate future changes in seasonal soil moisture
drought characteristics, aiming to answer two questions: (a) when and where the signals of drought changes
would exceed the noise of natural climate variability? (b) Would there be an impact on the ToE of soil moisture
drought characteristics when global warming is controlled within a certain level?
Earth's Future
10.1029/2023EF004174
CHEN AND YUAN 2 of 13
2. Data and Method
2.1. CMIP6 Data
Considering the availability of monthly total column soil moisture (SM), surface temperature data, 2m air
temperature (T), precipitation (P) and evapotranspiration (ET), we compiled a total of 34 CMIP6 models with up
to 3 ensemble members per model for fairness (a total of 66 pool members; Table S1 in Supporting Informa-
tion S1) under both historical and four Shared Socio‐economic Pathways scenarios (i.e., SSP126, SSP245,
SSP370 and SSP585; Eyring et al., 2016). All data were interpolated to 1° bilinearly.
2.2. Definition of Seasonal Drought
The drought index for identifying seasonal soil moisture droughts is based on the percentile of the total column
soil moisture. For each month, we used soil moisture in the same month during 1986–2005 (thus 20 values) as the
baseline climatology for percentile calculation. Then, we determined drought characteristics based on a 240‐
month window (20 years*12 months). A seasonal drought event occurs when soil moisture percentiles are less
than a certain threshold for at least three 3 consecutive months (Yuan & Wood, 2013). In the main part of this
paper we focus on the 20% threshold, that is, severe drought. To test the sensitivity of ToE results to drought
thresholds, results for other thresholds, that is, 30% (moderate drought) and 10% (extreme drought), were pre-
sented in Supporting Information S1 (Figures S2–S5). We focused on three characteristics of seasonal drought
events: frequency, duration, and intensity of the drought. We considered the number of droughts that occur in each
year as frequency, the average number of months that each drought lasts as duration, and the average monthly
percentile change in each event as intensity.
2.3. Definition of Time of Emergence (ToE)
For the identification of ToE, we invoked the concept of signal‐to‐noise ratio, which was widely used in many
fields for signal detection and assessment of climate change (Diffenbaugh, et al., 2011; Giorgi & Bi, 2009; Ying
et al., 2022). We defined the anomaly in drought characteristics (i.e., drought frequency, drought duration and
drought intensity) with respect to its baseline as signal (S) and the internal variability of climate system as noise
(N). For the calculation of the internal variability, we used the method proposed by Hawkins and Sutton (2009) to
fit a fourth‐order polynomial to the original series and used its standard deviation of the residuals as the internal
variability. This method has been widely used to calculate the internal variability in precipitation (Hawkins &
Sutton, 2009,2012; Hodson et al., 2013; IPCC, 2013), temperature (Cai et al., 2021; Wu et al., 2020) and extreme
events (Chen & Yuan, 2022; Jiao & Yuan, 2019; Lu et al., 2019). For the threshold of S/N, we define as 1 as in
other studies focusing on extreme events (Frame et al., 2017; Giorgi & Bi, 2009; Nguyen et al., 2018). We define
ToE as the year in which signal first exceeded noise and lasted until the end of this century.
Accordingly, ToE is calculated for each model, and the multi‐model median ToE is used for the assessment. The
results for different models are divided into three categories, (a) S/N >1, (b) S/N <1, and (c) no emergence.
Only when more than 50% of the models satisfy the condition of “(a) S/N >1” and simultaneously less than one‐
third models satisfy the condition of “(b) S/N <1,” we consider it as a positive emergence. For models in
categories (b) and (c) that do not meet S/N >1, we set them with ToE =2101, and we then take the median ToE
value of all models as the final multi‐model ToE. Similarly, when the models in condition “(b) S/N <1” is more
than half and the models in condition of “(a) S/N >1” is less than 1/3, we can obtain a negative “emergence.”
Otherwise, there is no emergence (Park et al., 2018). Note that the probabilities for “less than 1/3” are derived
from the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC, 2013) and are
referred to as unlikely.
3. Results
Figure 1depicts the multimodel median changes in frequency, duration and intensity of seasonal soil moisture
droughts in the 21st century (2080–2100) relative to the 20th century under SSP245 and SSP585 scenarios
(SSP126 and SSP370 are shown in Figure 2). Approximately 42%–48% of the global land areas (excluding the
Antarctic, Greenland and deserts) show increasing changes in future drought frequency under different emission
scenarios. The regions with more pronounced increasing changes are located in North America (NAM), Amazon
(AMZ), southern Europe (SEU), southern Africa (SAF) and southern China (SC). As for drought duration, the
Earth's Future
10.1029/2023EF004174
CHEN AND YUAN 3 of 13
spatial distribution is broadly similar to drought frequency, except for some high latitude regions in Asia
(Figures 1c and 1d). It is worth noting, however, that as the external forcing strengthens, although there is a
concomitant increase in drought duration in the regions mentioned above, but the proportion of increased regions
will decrease accordingly (51%, 47%, 46%, 44% from SSP126 to SSP585). The drought intensity shows a robust
increasing trend almost globally, at least 60% of the models agree with the change (98% of the global land areas
under four SSPs; Figures 1e, 1f, 2e, and 2f). It suggests that even in areas where droughts would occur less
frequently and shorter duration in the future, the intensity of drought would actually increase.
In order to assess whether the changes in seasonal soil moisture drought characteristics would exceed the internal
variability of climate system and can be detectable in this century, we conducted the ToE framework on drought
characteristics under four different emission scenarios. For the drought frequency, the ToE can only be detected in
areas with robust changes in soil moisture. In some parts of the Amazon, ToE occurs between 2030 and 2050,
while in other areas, it is mainly concentrated between 2070 and 2080. And ToE would occur earlier as the
external forcing strengthens (Figures 3a, 3b, 4a, and 4b). For the drought duration, the ToE would occur later than
drought frequency and will be less widespread. The globally averaged ToE for drought intensity is earliest, which
is about 6–9 years earlier compared to ToE for drought frequency. Comparing different drought characteristics, it
can be found that although the signal of frequency of seasonal soil moisture drought does not “emerge” in some
regions of the 21st century, the intensity of each drought event is increasing, and the signal has already “emerged.”
Figure 1. Multimodel median changes in 2080–2100 relative to 1986–2005 in seasonal soil moisture drought characteristics (frequency, duration and intensity from top
to bottom) for SSP245 (left column) and SSP585 (right column) scenarios. The dotted areas indicate that the change is robust, that is, at least 60% of the models agree
with the change.
Earth's Future
10.1029/2023EF004174
CHEN AND YUAN 4 of 13
Considering the commonalities exhibited by different drought characteristics signals at the regional scale, we
selected 6 subregions (black boxes in Figure 3a): North America (NAM), Amazon (AMZ), southern Europe
(SEU), southern Africa (SAF), southern China (SC) and Australia (AUS). The results show that based on the
SSP245 scenario, the timing of detectable signal in drought frequency is mostly later than 2080, except for the
AMZ (Figure 5). In the higher emission scenario, the ToE for drought frequency in the SEU, SAF, and SC is also
earlier than 2080. However, under the SSP126 scenario with lower emission, the ToE of drought frequency for all
subregions is later than 2080 (Figure 6). The ToE for drought duration is later than 2080 for all subregions under
the SSP245 scenario, but varies significantly with higher emission scenarios, with the Amazon region ranging
from 2087 to 2071 (from SSP245‐SSP585), followed by the SEU region ranging from 2085 to 2078 (from
SSP245‐SSP585). The other subregions do not change significantly with the emission scenarios. Detectable
signals for drought intensity appear early in the three drought characteristics, with the earliest still occurring in the
AMZ and SEU, followed by SAF, SC, NAM, and AUS. Under the SSP245 scenario, detectable signals are
projected to appear in 2060 for the AMZ,SEU, and SAF, and roughly in the remaining subregions between 2070
and 2080 as the emission scenarios increased. The ToE of drought intensity in the AMZ is projected to advance by
about 30 years under SSP585 compared to lower emission scenarios, and other subregions are also expected to
advance to varying degrees. Therefore, mitigated emission scenarios can significantly delay the emergence of
seasonal soil moisture droughts in these regions. It is worth noting that, unlike the spread of drought frequency
and duration, drought intensity exhibits large spatial variations, and the relevant authorities should take into
Figure 2. The same as Figure 1, but for SSP126 & SSP370.
Earth's Future
10.1029/2023EF004174
CHEN AND YUAN 5 of 13
account the underlying surface factors that affecting drought development, such as vegetation, and develop
appropriate disaster prevention measures.
About 14%–22% (SSP126‐SSP585) of the land areas would reach the ToE for increased drought frequency by the
end of the century. In contrast, about 47% (SSP126), 47% (SSP245), 48% (SSP370), and 49% (SSP585) of the
land areas would reach the ToE for increased drought intensity, and about 4% (SSP126), 8% (SSP245), 11%
(SSP370), and 12% (SSP585) of the land areas would reach the ToE for increased drought duration. In view of the
warming levels proposed by the Paris Agreement, we are interested in whether controlling the warming level by
1.5°C or 2°C would postpone the ToE. For the year of specific warming level, we used a 20‐year moving window
to determine and then used the multi‐model median as the representative results. The year of specific warming
level of multi‐model median are shown in Table 1.
In general, the land areas with detectable drought signals increased with increasing temperature. At the global
scale, compared to 2°C global warming level, 1.5°C would reduce the ToE areas by approximately 1/3 for drought
frequency and duration, and by approximately 1/5 for drought intensity. In contrast, 2°C warming level compared
to 3°C would reduce the ToE area of all drought characteristics by at least about 1/4 (Table 1). Specifically,
controlling warming is also beneficial for reducing the areas with ToE of drought frequency in all subregions,
especially in SAF, SC and AUS (Figure 7). When the warming level is limited to 2°C, as opposed to 1.5°C, the
Figure 3. The multi‐model median time of emergence (ToE)for seasonal soil moisture drought characteristics for SSP245 (left column) and SSP585 (right column)
scenarios. The gray areas represent that drought characteristics show a ToE of decreasing changes, which are not considered in this study. The black boxes in
(a) represent subregions where seasonal soil moisture droughts increase significantly, that is, North America (NAM), Amazon (AMZ), southern Europe (SEU), southern
Africa (SAF), southern China (SC), and Australia (AUS).
Earth's Future
10.1029/2023EF004174
CHEN AND YUAN 6 of 13
areas of increase drought frequency signal in all subregions would increase by about 1/3, except for 23% in NAM
under SSP126 scenario. While when the warming level is limited to 3°C rather than 2°C, the areas with ToE
would at least 40%. For duration, the effect of controlling for warming is significant, except in SC, where there are
still very few areas experiencing ToE by the end of the century (Figure 2). In the other subregions, controlling for
warming up to 2°C instead of 1.5°C increases the areas with ToE by about 1/2. In contrast, when controlling
warming to 3°C instead of 2°C, the area experiencing ToE will at least double in most subregions. For drought
intensity, controlling warming does not make a large difference in the impacts of different subregions. Controlling
for warming up to 2°C instead of 1.5°C or 3°C instead of 2°C increases the areas with ToE by at least 1/5.
4. Discussion
In addition to the results based on 20% threshold shown in the main text, we also calculated droughts based on
10% and 30% thresholds, respectively. The spatial distribution of drought changes is similar, but slightly different
in magnitude, with the more extreme droughts having stronger future changes (Figure 1; Figures S2 and S4 in
Supporting Information S1). The ToE areas of drought frequency increase with decreasing drought threshold,
while drought duration is opposite (drought frequency: 9%–14% for threshold‐30%, 14%–22% for threshold‐20%,
16%–25% for threshold‐10; drought duration: 7%–16% for threshold‐30%, 4%–12% for threshold‐20%, 3%–12%
for threshold‐10; Figure 3; Figures S3 and S5 in Supporting Information S1).
Figure 4. The same as Figure 3, but for SSP126 & SSP370.
Earth's Future
10.1029/2023EF004174
CHEN AND YUAN 7 of 13
Based on the ToE framework, we focus on six subregions where detectable signals appear relatively early, which
are consistent with previous studies. Considering the differences in results among subregions, we further applied
the ToE framework to P, ET, precipitation surplus (P‐ET), and SM (Figure S6 in Supporting Information S1). The
results showed that in AMZ, SEU, and SAF, the decrease in P‐ET and hence decreased soil moisture may have
exacerbated the drought. However, for NAM, SC, and AUS, although there was no detectable decrease signal of
P‐ET, there was still a detectable decrease signal of soil moisture, which may have been influenced by runoff, for
example, In recent years, it also appears that studies have demonstrated that areas with earlier detectable signals
are experiencing more severe droughts. For example, in the 2000s, more than half of the Amazon experienced
droughts that were severe enough to damage forests (Duffy et al., 2015). From 1961 to 2016, flash droughts
increased substantially in South Africa (Yuan et al., 2018). Research has shown that the increased net radiation is
a key drying driver for NAM and SC, while the drying conditions for SEU and SAF are exacerbated by reduced
precipitation (Takeshima et al., 2020). It is worth noting that these hotspots exist in major countries with a total
population of approximately 3.8 billion, and most regions will experience an increase in the proportion of people
aged 65 or above (Desa, 2019). As a result, these densely populated and aging places are more vulnerable to the
effects of extreme events. Our results highlight the risk of drought signal deviating from the natural variability of
the climate system, and further provide time constraints for mitigation polices and disaster prevention. However,
further research is still needed to clarify the specific indication of ToE on human society or ecosystem, etc.
Figure 5. The multi‐model median time of emergence (ToE) for six subregion (indicated in Figure 3a) under SSP245 (left
column) and SSP585 (right column) scenarios. In each region, the value of ToE is derived from all grid points within this
region.
Earth's Future
10.1029/2023EF004174
CHEN AND YUAN 8 of 13
Our results have important implications for near‐ and long‐term climate risk management, revealing the emer-
gence of changes in drought characteristics, but our estimates are affected by model and internal variability. First,
although we considered the “equal opportunity criterion” of using all available models for the analysis to obtain a
more comprehensive understanding of the potential future drought scenarios and reduce the risk of biased pro-
jections, the ToE exhibits a large interquartile range in some grids due to the inherent uncertainty of the model
(Figures S7 and S8 in Supporting Information S1). Therefore, we must consider more advanced GCMs to account
for the uncertainty. On the other hand, due to the data limitations, the future projection is currently only up to
2100. Therefore, the ToE is set to 2101 in the model where the ToE does not occur even at the end of the century.
Although this may be a “pseudo emergence,” taking the median value for multi‐model reduces the impact of the
outlier projections after 2100. In addition, a study that used models with simulation up to 2300 found that 7/13 of
the models demonstrated similar or even earlier ToEs of temperature compared to models projected up to 2100
(Hawkins et al., 2014). However, due to the limitations of the soil moisture data, only seven models provide data
up to 2300 in a certain scenario with only one ensemble member, while three ensemble members are used in this
paper. And only one model provide soil moisture data up to 2300 under four SSP scenarios. Therefore, no relevant
conclusions can be drawn about soil moisture drought. Third, using the latest generation of CMIP data, we
conducted a comprehensive evaluation of the timing of future soil moisture drought change signals under four
different possible future development scenarios. However, studies have shown that compared to CMIP5, the
equilibrium climate sensitivity (ECS, a theoretical equilibrium value of CO
2
doubling‐induced global warming)
of CMIP6 models have a larger range, meaning that the models are more sensitive to the increase of CO
2
. This is
Figure 6. The same as Figure 5, but for SSP126 & SSP370.
Earth's Future
10.1029/2023EF004174
CHEN AND YUAN 9 of 13
related to the models of several outliers, namely CESM2, CanESM5, ESM‐1‐0, HadGEM3‐GC31‐LL,
HadGEM3‐GC31‐MM, INM‐CM4‐8, INM‐SM5‐0, UKESM1‐0‐LL (five of them and were used in this paper;
Table S1 in Supporting Information S1), with ECS exceeding 5°C or below 2°C (Meehl et al., 2020). Five out of
six models with higher ECS are considered to be related to cloud feedback and/or cloud aerosol interactions. Four
of these models have large negative present‐day aerosol effective radiative forcing ranging from 1.1 to
1.7 Wm
2
. Therefore, if these models want to reproduce historical temperature responses, they need to have
stronger model responses to CO
2
. We re‐identified ToE with exclusion of the “outlier model” data sets and found
that ToE with drought frequency, duration, and intensity can be identified in 13%–21%, 3%–12%, and 42%–46%
of the land areas (14%–22%, 4%–12%, and 47%–49% in the current version). In addition, although the ECS range
of CMIP6 is higher, the mean ECS in all generations of models is near 3.5 ±0.2°C (CMIP6: 3.7°C). And transient
climate response (TCR, surface temperature warming when CO2 doubles in a simulation of 1% annual CO2
increase) may be more relevant on the time scale in the coming decades, with its range relatively consistent with
CMIP5. Finally, the quantification of internal variability has always been one of the most data‐intensive chal-
lenges in climate science. Although the Multi‐Model Large Ensemble (LE) gives potential in simulating internal
variability reasonably, only seven GCMs currently participate the LE project due to enormous computational
costs and time‐consuming requirements, and very few LE models provide the soil moisture data (Chen &
Yuan, 2022). Therefore, reliable climate projections as well as ToE estimations still require the use of a large set
of GCM with soil moisture data available.
5. Conclusions
Based on CMIP6 multi‐model soil moisture data, this study investigated the future changes and ToE of three
characteristics of seasonal soil moisture droughts (i.e., frequency, duration and intensity), and further discuss the
influence of different global warming level on ToE. The main conclusions are as follows:
For seasonal soil moisture drought frequency, duration and intensity, about 42%–48%, 44%–51%, 98%–98% of
land areas display increasing changes and about 14%–22%, 4%–12% and 47%–49% of land areas would reach ToE
before the end of this century, respectively. The regions of increasing drought frequency are concentrated in
NAM, AMZ, SEU, SAF, SC and AUS. For the drought frequency and duration, no matter which SSP, ToEs are
latttaler than 2060. While for drought intensity, all subregions show ToE between 2040 and 2080, with AMZ and
SEU relatively earlier.
Furthermore, restricting the global warming has a significant effect on delaying the ToE of soil moisture drought.
At the global scale, restricting warming level to 1.5°C compared to 2°C or restricting warming level to 2°C
Table 1
The Multi‐Median Crossing Year of Reaching Specific Warming Levels and the Global Land Area Proportions (%) That
Reached the Time of Emergence (ToE) Before That Year Under Different Shared Socio‐Economic Pathways (SSPs)
Warming threshold Warming year Frequency Duration Intensity
SSP126 1.5°C 2023 3.64 0.99 12.55
2°C 2038 5.30 1.67 15.91
3°C NA NA NA NA
SSP245 1.5°C 2023 3.43 1.02 12.07
2°C 2037 4.97 1.94 15.06
3°C 2065 8.18 3.83 22.29
SSP370 1.5°C 2025 3.61 1.02 12.01
2°C 2038 5.13 1.84 14.70
3°C 2059 7.74 3.66 21.09
SSP585 1.5°C 2021 3.40 0.95 12.01
2°C 2033 4.79 1.63 14.68
3°C 2051 7.08 3.04 19.38
Note. The crossing year is obtained by a 20‐year moving window. The “NA” means that the model will not reach the specific
warming level throughout the 21st century.
Earth's Future
10.1029/2023EF004174
CHEN AND YUAN 10 of 13
Figure 7. The land area proportions (%) that reached the time of emergence (ToE) before specific warming thresholds for six
subregions (indicated in Figure 3a). The left column is drought frequency, the right column is drought intensity, and each row
represents a different subregion.
Earth's Future
10.1029/2023EF004174
CHEN AND YUAN 11 of 13
compared to 3°C would reduce the ToE area by at least 1/5. While on the regional scale, limiting warming to 2°C
would increase the ToE of drought frequency and duration by approximately 1/3 and the ToE of drought intensity
by approximately 1/5 across six subregions, compared to limiting warming to 1.5°C. However, if global warming
increased by 3°C instead of 2°C, the ToE for drought frequency and duration would increase by at least 40%,
while the ToE for drought intensity would increase by 20% across the six subregions.
Data Availability Statement
The CMIP6 data used in this study are available at https://esgf‐node.llnl.gov/search/cmip6/ (Eyring et al., 2016).
The data for different variables, models, emission scenarios, and time scales can be obtained by clicking on
“Variable,” “Source ID,” “Experiment ID,” and “Frequency” in the options on the right. Note that for the total
column soil moisture, surface temperature data, 2m air temperature, precipitation and evapotranspiration used in
this paper are correspond to “mrso,” “ts” “tas,” “pr” and “evspsbl” under “Variable,” respectively.
References
Bador, M., Terray, L., & Boé, J. (2016). Emergence of human influence on summer record‐breaking temperatures over Europe. Geophysical
Research Letters,43(1), 404–412. https://doi.org/10.1002/2015GL066560
Cai, Z. Y., You, Q. L., Wu, F. Y., Chen, H. W., Chen, D., & Cohen, J. (2021). Arctic warming revealed by multiple CMIP6 models: Evaluation of
historical simulations and quantification of future projection uncertainties. Journal of Climate,34(12), 4871–4892. https://doi.org/10.1175/
JCLI‐D‐20‐0791.1
Chen, S. S., & Yuan, X. (2022). Quantifying the uncertainty of internal variability in future projections of seasonal soil moisture droughts over
China. Science of the Total Environment,824, 153817. https://doi.org/10.1016/j.scitotenv.2022.153817
Collins, L., Bradstock, R. A., Clarke, H., Clarke, M. F., Nolan, R. H., & Penman, T. D. (2021). The 2019/2020 mega‐fires exposed Australian
ecosystems to an unprecedented extent of high‐severity fire. Environmental Research Letters,16(4), 044029. https://doi.org/10.1088/1748‐
9326/abeb9e
Cook, B. I., Mankin, J. S., Marvel, K., Williams, A. P., Smerdon, J. E., & Anchukaitis, K. J. (2020). Twenty‐first century drought projections in the
CMIP6 forcing scenarios. Earth's Future,8(6), e2019EF001461. https://doi.org/10.1029/2019EF001461
Desa, U. (2019). World Population Prospects 2019: Highlights (New York: United Nations Department for Economic and Social Affairs).
Diffenbaugh, N. S., Ashfaq, M., & Scherer, M. (2011). Transient regional climate change: Analysis of the summer climate response in a high‐
resolution, century‐scale ensemble experiment over the continental United States. Journal of Geophysical Research,116(D24), D24111.
https://doi.org/10.1029/2011jd016458
Duan, W. S., & Mu, M. (2018). Predictability of El Niño‐Southern Oscillation events. Oxford Research Encyclopedia of Climate Science.https://
doi.org/10.1093/acrefore/9780190228620.013.80
Duffy, P. B., Brando, P., Asner, G. P., & Field, C. B. (2015). Projections of future meteorological drought and wet periods in the Amazon.
Proceedings of the National Academy of Sciences of the United States of America,112(43), 13172–13177. https://doi.org/10.1073/pnas.
1421010112
Eyring, V., Bony, S., Meehl, G. A., Senior, C. A., Stevens, B., Stouffer, R. J., & Taylor, K. E. (2016). Overview of the Coupled Model Inter-
comparison Project Phase 6 (CMIP6) experimental design and organization [Dataset]. Geoscientific Model Development,9, 1937–1958.
https://doi.org/10.5194/gmd‐9‐1937‐2016
Frame, D., Joshi, M., Hawkins, E., Harrington, L. J., & de Roiste, M. (2017). Population‐based emergence of unfamiliar climates. Nature Climate
Change,7(6), 407–411. https://doi.org/10.1038/nclimate3297
Giorgi, F., & Bi, X. (2009). Time of emergence (TOE) of GHG‐forced precipitation change hot‐spots. Geophysical Research Letters,36(6),
L06709. https://doi.org/10.1029/2009gl037593
Hawkins, E., Anderson, B., Diffenbaugh, N., Mahlstein, I., Betts, R., Hegerl, G., et al. (2014). Uncertainties in the timing of unprecedented
climates. Nature,511(7507), E3–E5. https://doi.org/10.1038/nature13523
Hawkins, E., & Sutton, R. (2009). The potential to narrow uncertainty in regional climate predictions. Bulletin of the America Meteorology
Society,90(8), 1095–1107. https://doi.org/10.1175/2009BAMS2607.1
Hawkins, E., & Sutton, R. (2012). Time of emergence of climate signals. Geophysical Research Letters,39(1), 6. https://doi.org/10.1029/
2011gl050087
Hodson, D. L. R., Keeley, S. P. E., West, A., Ridley, J., Hawkins, E., & Hewitt, H. T. (2013). Identifying uncertainties in Arctic climate change
projections. Climate Dynamics,40(11–12), 2849–2865. https://doi.org/10.1038/550S53a
IPCC. (2013). Climate change 2013: The physical science basis. In T. F. Stocker, D. Qin, G. K. Plattner, M. Tignor, S. K. Allen, J. Boschung, et al.
(Eds.), Contribution of working group I to the fifth assessment report of the intergovernmental panel on climate change (p. 1535). Cambridge
University Press.
Jiao, Y., & Yuan, X. (2019). More severe hydrological drought events emerge at different warming levels over the Wudinghe Watershed in
northern China. Hydrology and Earth System Sciences,23(1), 621–635. https://doi.org/10.5194/hess‐23‐621‐2019
King, A. D., Donat, M. G., Fischer, E. M., Hawkins, E., Alexander, L. V., Karoly, D. J., et al. (2015). The timing of anthropogenic emergence in
simulated climate extremes. Environmental Research Letters,10(9), 094015. https://doi.org/10.1088/1748‐9326/10/9/094015
Landrum, L., & Holland, M. M. (2020). Extremes become routine in an emerging new Arctic. Nature Climate Change,10(12), 1108–1115. https://
doi.org/10.1038/s41558‐020‐0892‐z
Lobell, D. B., & Burke, M. B. (2008). Why are agricultural impacts of climate change so uncertain? The importance of temperature relative to
precipitation. Environmental Research Letters,3, 034007. https://doi.org/10.1088/1748‐9326/3/3/034007
Lombardozzi, D., Bonan, G. B., & Nychka, D. W. (2014). The emerging anthropogenic signal in land–atmosphere carbon‐cycle coupling. Nature
Climate Change,4(9), 796–800. https://doi.org/10.1038/nclimate2323
Lopez, H., West, R., Dong, S., Goni, G., Kirtman, B., Lee, S. K., & Atlas, R. (2018). Early emergence of anthropogenically forced heat waves in
the western United States and Great Lakes. Nature Climate Change,8(5), 414–420. https://doi.org/10.1038/s41558‐018‐0116‐y
Acknowledgments
This work was supported by National
Natural Science Foundation of China
(42330604), National Key R&D Program
of China (2022YFC3002803), Natural
Science Foundation of Jiangsu Province
for Distinguished Young Scholars
(BK20211540), the Major Science and
Technology Program of the Ministry of
Water Resources of China (SKS‐
2022019), and Jiangsu graduate scientific
research innovation program
(KYCX22_1158).
Earth's Future
10.1029/2023EF004174
CHEN AND YUAN 12 of 13
Lu, J., Carbone, G. J., & Grego, J. M. (2019). Uncertainty and hotspots in 21st century projections of agricultural drought from CMIP5 models.
Scientific Reports,9(1), 4922. https://doi.org/10.1038/s41598‐019‐41196‐z
Lyu, K., Zhang, X., Church, J. A., Slangen, A. B. A., & Hu, J. (2014). Time of emergence for regional sea‐level change. Nature Climate Change,
4(11), 1006–1010. https://doi.org/10.1038/nclimate2397
Ma, F., Yuan, X., Wu, P. L., & Zeng, Z. Z. (2022). A moderate mitigation can significantly delay the emergence of compound hot extremes.
Journal of Geophysical Research: Atmospheres,127(2), e2021JD035427. https://doi.org/10.1029/2021JD035427
Mahlstein, I., Hegerl, G., & Solomon, S. (2012). Emerging local warming signals in observational data. Geophysical Research Letters,39(21),
L21711. https://doi.org/10.1029/2012GL053952
Mahony, C. R., & Cannon, A. J. (2018). Wetter summers can intensify departures from natural variability in a warming climate. Nature Com-
munications,9(1), 783. https://doi.org/10.1038/s41467‐018‐03132‐z
Meehl, G. A., Senior, C. A., Eyring, V., Flato, G., Lamarque, J. F., Stouffer, R. J., et al. (2020). Context for interpreting equilibrium climate
sensitivity and transient climate response from the CMIP6 Earth system models. Science Advances,6(26), eaba1981. https://doi.org/10.1126/
sciadv.aba1981
Mora, C., Dousset, B., Caldwell, I. R., Powell, F. E., Geronimo, R. C., Bielecki, C. R., et al. (2017). Global risk of deadly heat. Nature Climate
Change,7(7), 501–506. https://doi.org/10.1038/nclimate3322
Myhre, G., Alterskjær, K., Stjern, C. W., Hodnebrog, Ø., Marelle, L., Samset, B. H., et al. (2019). Frequency of extreme precipitation increases
extensively with event rareness under global warming. Scientific Reports,9(1), 1–10. https://doi.org/10.1038/s41598‐019‐52277‐4
Nguyen, T.‐H., Min, S.‐K., Paik, S., & Lee, D. (2018). Time of emergence in regional precipitation changes: An updated assessment using the
CMIP5 multi‐model ensemble. Climate Dynamics,51(9–10), 3179–3193. https://doi.org/10.1007/s00382‐018‐4073‐y
Orlowsky, B., & Seneviratne, S. I. (2013). Elusive drought: Uncertainty in observed trends and short‐ and long‐term CMIP5 projections. Hy-
drology and Earth System Sciences,17(5), 1765–1781. https://doi.org/10.5194/hess‐17‐1765‐2013
Park, C. E., Jeong, S.‐J., Joshi, M., Osborn, T. J., Ho, C.‐H., Piao, S., et al. (2018). Keeping global warming within 1.5°C constrains emergence of
aridification. Nature Climate Change,8(1), 70–74. https://doi.org/10.1038/s41558‐017‐0034‐4
Pendergrass, A. G. (2018). What precipitation is extreme? Science,360(6393), 1072–1073. https://doi.org/10.1126/science.aat1871
Raymond, C., Matthews, T., & Horton, R. M. (2020). The emergence of heat and humidity too severe for human tolerance. Science Advances,
6(19), eaaw1838. https://doi.org/10.1126/sciadv.aaw1838
Rockstrom, J., Steffen, W., Noone, K., Persson, A., Chapin, F. S., Lambin, E. F., et al. (2009). A safe operating space for humanity. Nature,
461(7263), 472–475. https://doi.org/10.1038/461472a
Satoh, Y., Yoshimura, K., Pokhrel, Y., Kim, H., Shiogama, H., Yokohataet, T., et al. (2022). The timing of unprecedented hydrological drought
under. Climate Change Nature Communications,13(1), 3287. https://doi.org/10.1038/s41467‐022‐30729‐2
Schlunegger, S., Rodgers, K. B., Sarmiento, J. L., Frölicher, T. L., Dunne, J. P., Ishii, M., & Slater, R. (2019). Emergence of anthropogenic signals
in the ocean carbon cycle. Nature Climate Change,9, 719–725. https://doi.org/10.1038/s41558‐019‐0553‐2
Schlunegger, S., Rodgers, K. B., Sarmiento, J. L., Ilyina, T., Dunne, J. P., Takano, Y., et al. (2020). Time ofEmergence and Large Ensemble
intercomparison for ocean biogeochemical trends. Global Biogeochemical Cycles,34(8), e2019GB006453. https://doi.org/10.1029/
2019GB006453
Silvy, Y., Guilyardi, E., Sallée, J., & Durack, P. J. (2020). Human‐induced changes to the global ocean water masses and their time of emergence.
Nature Climate Change,10(11), 1030–1036. https://doi.org/10.1038/s41558‐020‐0878‐x
Stevenson, S., Coats, S., Touma, D., Cole, J., Lehner, F., Fasullo, J., & Otto‐Bliesner, B. (2022). Twenty‐first century hydroclimate: A continually
changing baseline, with more frequent extremes. Proceedings of the National Academy of Sciences of the United States of America,119(12),
e2108124119. https://doi.org/10.1073/pnas.2108124119
Takeshima, A., Kim, H., Shiogama, H., Lierhammer, L., ScinoccaSeland, J. F. O., & Mitchell, D. (2020). Global aridity changes due to differences
in surface energy and water balance between 1.5°C and 2°C warming. Enviromental Research Letters,15(9), 0940a7. https://doi.org/10.1088/
1748‐9326/ab9db3
Touma, D., Ashfaq, M., Nayak, M. A., Kao, S.‐C., & Diffenbaugh, N. S. (2015). A multi‐model and multi‐index evaluation of drought char-
acteristics in the 21st century. Journal of Hydrology,526, 196–207. https://doi.org/10.1016/j.jhydrol.2014.12.011
Walker, J. S., Kopp, R. E., Little, C. M., & Horton, B. P. (2022). Timing of emergence of modern rates of sea‐level rise by 1863. Nature
Communications,13(1), 966. https://doi.org/10.1038/s41467‐022‐28564‐6
Wang, Y., & Yuan, X. (2021). Anthropogenic speeding up of South China flash droughts as exemplified by the 2019 summer‐Autumn transition
season. Geophysical Research Letters,48(9), e2020GL091901. https://doi.org/10.1029/2020GL091901
Wu, F. Y., You, Q. L., Zhang, Z. Y., & Zhang, L. (2020). Changes and uncertainties of surface mean temperature over China under global
warming of 1.5 and 2°C. International Journal of Climatology,41(S1), E410–E427. https://doi.org/10.1002/JOC.6694
Ying, J., Collins, M., Cai, W., Timmermann, A., Huang, P., Chen, D., & Stein, K. (2022). Emergence of climate change in the tropical Pacific.
Nature Climate Change,12(4), 356–364. https://doi.org/10.1038/s41558‐022‐01301‐z
Yuan, X., Wang, L., & Wood, E. F. (2018). Anthropogenic intensification of southern African flash droughts as exemplified by the 2015/16
season. [in “explaining extreme events of 2016 from a climate perspective”]. Bulletin of the American Meteorological Society,99(1), S54–S59.
https://doi.org/10.1175/bams‐d‐17‐0077.1
Yuan, X., Wang, Y., Ji, P., Wu, P., Sheffield, J., & Otkin, J. (2023). A global transition to flash droughts under climate change. Science,380(6641),
187–191. https://doi.org/10.1126/science.abn6301
Yuan, X., & Wood, E. F. (2013). Multimodel seasonal forecasting of global drought onset. Geophysical Research Letters,40(18), 4900–4905.
https://doi.org/10.1002/grl.50949
Zhang, M., Yuan, X., Otkin, J., & Ji, P. (2022). Climate warming outweighs vegetation greening in intensifying flash droughts over China.
Environmental Research Letters,17(5), 054041. https://doi.org/10.1088/1748‐9326/ac69fb
Earth's Future
10.1029/2023EF004174
CHEN AND YUAN 13 of 13
