Abstract

Anthropogenic forcing is anticipated to increase the magnitude and frequency of precipitation-induced extremes such as the increase in drought risks. However, the model-projected future changes in global droughts remain largely uncertain, particularly in the context of the Paris Agreement targets. Here, by using the standardized precipitation index (SPI), we present a multiscale global assessment of the precipitation-driven meteorological drought characteristics at the 1.5° and 2°C warming levels based on 28 CMIP5 global climate models (GCMs) under three representative concentration pathways scenarios (RCP2.6, RCP4.5, and RCP8.5). The results show large uncertainties in the timing reaching 1.5° and 2°C warming and the changes in drought characteristics among GCMs, especially at longer time scales and under higher RCP scenarios. The multi-GCM ensemble mean projects a general increase in drought frequency (Df) and area (Da) over North America, Europe, and northern Asia at both 1.5° and 2°C of global warming. The additional 0.5°C warming from 1.5° to 2°C is expected to result in a trend toward wetter climatic conditions for most global regions (e.g., North America, Europe, northern Asia, and northern Africa) due to the continuing increase in precipitation under the more intensified 2°C warming. In contrast, the increase in Df is projected only in some parts of southwest Asia, South America, southern Africa, and Australia. Our results highlight the need to consider multiple GCMs in drought projection studies under the context of the Paris Agreement targets to account for large model-dependent uncertainties.

1. Introduction

Global warming due to the increasing concentration of greenhouse gases is anticipated to have major impacts on the hydrologic cycle (Huntington 2006; Milliman et al. 2008; Arnell and Gosling 2013; Wu et al. 2018; Yeh and Wu 2018; Zhai et al. 2018), such as changing precipitation (PR) patterns and increasing risks of extreme hydrometeorological events (Sheffield and Wood 2008; Dai 2011, 2013; Sillmann et al. 2013; Prudhomme et al. 2014; Wu and Huang 2016; Wu et al. 2016; Dai and Zhao 2017; Zhao and Dai 2017; Xu et al. 2019a,b). To reduce the risks and impacts of climate change, the Paris Climate Agreement proposed for the first time a goal to limit global mean surface warming below 2.0°C and pursue for the target of within 1.5°C above the preindustrial levels (UNFCCC 2015). The political and socioeconomic achievability of these goals calls for advanced impact assessment studies of climate changes considering multiple sectors at these two specific warming levels (Stocker 2013; Schleussner et al. 2016b; Sanderson et al. 2017). Despite the urgent needs for climate change adaptation and mitigation, there still exist substantial knowledge gaps regarding the negative impacts of warming, which potentially can be alleviated by limiting the global warming target from the commonly recognized 2° to 1.5°C (Sanderson et al. 2017).

The general public usually perceives the adverse impacts of climate changes through the increasing frequency and severity of climate extremes such as flood, drought, and heat waves, all of which are closely tied to large economic losses and casualties (Wang et al. 2017). Among climate hazards, drought leads to major agricultural, economic, and environmental damages with significant global economic losses per year (Wilhite 2000). It is therefore important to investigate how the drought risks may change if global warming can be controlled to within 1.5° and 2°C, and whether significant differences in drought risk can be quantitatively projected between these two specific warming levels. Several recent climate modeling studies have projected the changes in climate extremes and water disasters under the 1.5° and 2°C warming at both global (Wang et al. 2017; Betts et al. 2018) and regional (Donnelly et al. 2017; King and Karoly 2017; Nangombe et al. 2018; Zhai et al. 2018; Li et al. 2019) scales. However, analyses of the changes in drought risk under 1.5° and 2°C warming are relatively scarce, with only few studies (e.g., Lehner et al. 2017; Su et al. 2018) examining the implications of mitigation levels over the continent. Lehner et al. (2017) projected a change in the global aridity and the associated risk in the occurrence of consecutive years of drought under 1.5° and 2°C warming using the multiple drought metrics and a set of simulations with the Community Earth System Model (CESM). Su et al. (2018) conducted an assessment on the drought losses in China under 1.5° and 2.0°C warming using 13 global climate models (GCMs) participated in the Coupled Model Intercomparison Project phase 5 (CMIP5). There are generally consistent findings from Su et al. (2018) and Lehner et al. (2017) in that both studies highlighted that constraining anthropogenic warming to 1.5°C can reduce future drought risk relative to 2.0°C warming, but more efforts using alternative climate models are necessary to ensure the robustness of these findings. Recent research (e.g., Wang et al. 2017) has demonstrated that the differences in the greenhouse gases (GHGs) emissions scenarios can significantly affect the global impact assessment of 1.5° and 2°C warming. However, our common knowledge on the sensitivity of drought risk to global 1.5° and 2°C warming under different GHGs emission scenarios is still limited. To enhance understanding on this critical issue, climate modeling studies aiming at limiting global warming to specific target levels require various combinations of GHG and aerosol forcing emission scenarios.

Here, using the standardized precipitation index (SPI) as the precipitation-driven drought index, this study presents a global assessment of drought characteristics at 1.5° and 2°C warming levels and the difference in the drought characteristics considered between the global 1.5° and 2°C warming levels based on the simulated global PR data from 28 GCMs participated in CMIP5 under three representative concentration pathway scenarios (RCP2.6, RCP4.5, and RCP8.5). Using a multimodel approach, we aim to explore globally the projected changes in drought characteristics under the future 2°C global warming, and whether these changes in characteristics can be alleviated by limiting the warming to 1.5°C, consistent with the primary concern and advocation of the Paris Agreement. Also, we assess the uncertainties in meteorological drought projections at 1.5° and 2°C warming as contributed by the alternative RCP scenarios and various GCMs used in this study.

In the following section 2, the details on the used global observation-based and CMIP5 GCM-simulated gridded datasets at 1.5° and 2°C global warming levels under three RCP scenarios are provided, and the definition of SPI as well as the details of bias-correction method used in this study are introduced. In section 3, the level of agreement on the GCM-projected drought frequency (Df) and area (Da), and the multimodel mean projected changes of Df and Da at the 1.5° and 2°C warming levels (as well as the effects due to the additional 0.5°C warming) are analyzed under the alternative emission scenarios. The main sources of uncertainties in projected future drought characteristics are discussed in section 4. Finally, the conclusions drawn from this study are summarized in section 5.

2. Dataset and methods

a. Global climate simulations at 1.5° and 2°C global warming levels

The monthly PR and temperature T data simulated by 28 CMIP5 GCMs under RCP2.6, RCP4.5, and RCP8.5 provided by the Canadian Climate Data and Scenarios (CCDS) are used in this study (see Table S1 in the online supplemental material). Only one ensemble member under each of three RCPs is selected for each GCM to be used in this study. The data were statistically downscaled onto a common 1° × 1° global grid by CCDS (available at http://climatescenarios.canada.ca/?page=main) and have been used in our previous study to assess the hydrologic impacts of climate change over China (Wu et al. 2018).

In this study, we define 1986–2005 as the “reference period” and the “20-yr future period” as the specific period within 2006–2100 with the mean global surface T reaching 1.5° and 2.0°C warming relative to the preindustrial levels. The simulated monthly T data by 28 GCMs were validated and corrected (see the details in section 2b) by referencing to the observation-based global 0.5° gridded dataset from the Climatic Research Unit (CRU TS 3.22, available at https://crudata.uea.ac.uk/cru/data/hrg/) of the University of East Anglia (Harris et al. 2014). The simulated monthly PR data by 28 GCMs were validated and corrected (see section 2b below) by referencing to the observation-based global 1° gridded dataset from Global Precipitation Climatology Centre (GPCC) V2018 (https://psl.noaa.gov/data/gridded/data.gpcc.html#detail). To ensure the consistency in the spatial resolution among the datasets used in this study, the CRU TS 3.22 gridded data were rescaled from the original global 0.5° to the 1° resolution using the bilinear interpolation.

There are multiple ways for defining the specific time reaching 1.5° or 2°C warming available in literature. The first one is when the long-term climate states (climatology) reach the stabilization at a targeted warming level, while another is when 1.5° or 2°C warming target is reached for the first time (Betts et al. 2018). We will adopt the first definition in this study. Following the methodology used in previous studies (Schleussner et al. 2016a; Wang et al. 2017; Leng 2018; Su et al. 2018; Li et al. 2019), the 1986–2005 is selected as the reference period when global mean surface T was 0.6°C above the preindustrial levels. To analyze the impacts of 1.5° and 2°C warming on projected drought characteristics, we extracted for each GCM the data for the identified 20-yr periods when the global mean surface T is 0.9° and 1.4°C above that in the reference period (1986–2005) under 3 RCPs. The identified 20-yr periods reaching 1.5° and 2.0°C warming for each GCM under all RCPs are summarized in Table 1, where it can be seen that for RCP2.6 not all GCMs can reach 1.5° or 2.0°C warming. Large differences in the crossing time of 1.5° and 2.0°C warming can be seen among GCMs and RCPs, exemplifying the large uncertainty generally found in climate change impact assessment under 1.5° and 2.0°C warming.

Table 1.

The projected 20-yr future periods for each of the 28 CMIP5 GCMs to reach the mean global warming levels of 1.5° and 2.0°C above the preindustrial level under the RCP2.6, RCP4.5, and RCP8.5 scenarios. The GCM number is the same as that given in Table S1. A “—” denotes there is no 20-yr period before 2100 identified with the projected global mean temperature reaching 1.5° or 2°C warming levels.

The projected 20-yr future periods for each of the 28 CMIP5 GCMs to reach the mean global warming levels of 1.5° and 2.0°C above the preindustrial level under the RCP2.6, RCP4.5, and RCP8.5 scenarios. The GCM number is the same as that given in Table S1. A “—” denotes there is no 20-yr period before 2100 identified with the projected global mean temperature reaching 1.5° or 2°C warming levels.
The projected 20-yr future periods for each of the 28 CMIP5 GCMs to reach the mean global warming levels of 1.5° and 2.0°C above the preindustrial level under the RCP2.6, RCP4.5, and RCP8.5 scenarios. The GCM number is the same as that given in Table S1. A “—” denotes there is no 20-yr period before 2100 identified with the projected global mean temperature reaching 1.5° or 2°C warming levels.

b. Bias correction of climate simulations

An appropriate bias correction for GCM-projected T and PR is necessary due to its significant influences on the hydrologic response. In this study, the simulated monthly T from 28 GCMs were bias corrected by using the “delta change” method (Wu and Huang 2016; Xu et al. 2019a). However, the following equidistant quantile matching method, which explicitly accounts for the distribution changes between the baseline and projection periods, was applied to correct the 28-GCM simulated PR due to large uncertainties in simulated PR data (Li et al. 2010; Lu et al. 2014):

 
x˜m,p.qm=xm,pFo,r1[Fm,p(xm,p)]/Fm,r1[Fm,p(xm,p)]
(1)

where F is the cumulative distribution function (CDF) of either observation o or model m for the 1986–2005 reference period r or the future 20-yr projection period p. A two-parameter Gamma probability distribution function (PDF) was chosen to fit monthly PR data, with the parameters estimated by the maximum likelihood method. Specifically, the following mixed-type Gamma distribution is used to account for the locations with zero PR:

 
G(x)=(1f)H(x)+fF(x),
(2)

where f is the percentage of months with PR > 0; H(x) is a step function with the value of 0 (PR = 0) and 1 (PR > 0).

The Mann–Kendall trend test (Kendall 1975) is used to estimate the trends of the 28-GCM ensemble mean annual PR under three RCPs for all subperiods within 2006–2100 with the minimum length of 20 years. By varying the starting and ending years of all possible subperiods [similar to that plotted in Fig. 4 of Yeh and Wu (2018)], a total of 2926 subperiods within 2006–2100 with the minimum length of 20 years are selected (i.e., 76+75+74++3+2+1=2926). The estimated PR trends are displayed in the contours in Fig. 1 for all subperiods for six global continents. As seen, both the magnitude and sign of estimated PR trends are sensitive to the various subperiods particularly under RCP2.6. Overall, an increased PR trend can be found for nearly all subperiods (including the identified periods reaching 1.5° and 2°C warming, see Table 1) in most Northern Hemisphere regions (i.e., North America, Asia, Europe, and Africa) under RCP4.5 and RCP8.5. In contrast, a decreased PR trend is found for most subperiods over the Southern Hemisphere regions (i.e., Oceania and South America) under all three RCPs, except for Oceania under RCP8.5. Under RCP2.6, the negative PR trends can be found for most subperiods near the end of twenty-first century over all continents (Fig. 1).

Fig. 1.

The contours of annual trend magnitudes (mm yr−1) of the 28-GCM ensemble mean PR under RCP2.6, RCP4.5, and RCP8.5 in six global continents for all subperiods within 2006–2100 with the minimum length of 20 years.

Fig. 1.

The contours of annual trend magnitudes (mm yr−1) of the 28-GCM ensemble mean PR under RCP2.6, RCP4.5, and RCP8.5 in six global continents for all subperiods within 2006–2100 with the minimum length of 20 years.

c. Standardized precipitation index

Droughts can be assessed differently from the meteorological, agricultural, hydrologic, and socioeconomic impacts, hence a decision on the choice of drought indices is required according to the purposes of studies (Orlowsky and Seneviratne 2013; Touma et al. 2015; Su et al. 2018). SPI is a widely used meteorological drought index that can quantify both wet and dry anomalous conditions based on the degree to which the accumulative PR for a specific time period deviates from the corresponding median (McKee et al. 1993). The first step in calculating SPI is to estimate the PDF of monthly PR (e.g., gamma distribution), then the cumulative probability of PR is computed and inversely transformed by the standard normal distribution with zero mean and unit variance (McKee et al. 1993). It can be used to quantify PR deficit at multiple time scales (e.g., 1, 3, 6, 12, 24 months). Short-time-scale SPIs (e.g., 1 and 3 months) are generally more sensitive to PR short-term changes, while long-time-scale SPIs (e.g., 6 and 12 months) reflect the long-term variations of dryness and wetness well. Overall, SPI offers several advantages over other commonly used indices, e.g., the unambiguous theoretical development, robustness, temporal flexibility, and the simplicity in the required data. More details on the SPI calculation can be found in McKee et al. (1993, 1995).

In this study, the gamma distribution was used for fitting monthly PR and then transformed for calculating the SPI (Rhee and Cho 2016). We used the baseline 1986–2005 as a reference period to fit the gamma distribution for calculating SPI. The parameters of the gamma distribution were estimated by the maximum likelihood method. Based on GCM-simulated monthly PR under RCP2.6, RCP4.5, and RCP8.5, the SPIs at the 3-, 6-, 12-, and 24-month time scales (termed as SPI3, SPI6, SPI12, and SPI24, respectively) are computed for each of the 28 CMIP5 GCMs for both 1986–2005 and the 20-yr periods corresponding to 1.5° and 2°C warming. The occurrence of meteorological drought is defined here as SPI ≤ −1 (see the classification of SPI in Table 2) following McKee et al. (1993). The frequency of drought occurrence (Df) at each grid is calculated as the percentage of months with SPI ≤ −1 within the given 20-yr period following that used in Duffy et al. (2015) and Rhee and Cho (2016). The drought area (Da) is quantified as the percentage of global grids with SPI ≤ −1.

Table 2.

The classification of SPI into seven categories.

The classification of SPI into seven categories.
The classification of SPI into seven categories.

3. Results

a. Evaluation of the bias-corrected CMIP5 model simulations

The comparisons of 1986–2005 monthly global-averaged PR and SPIs (SPI3, SPI6, SPI12, and SPI24) simulated by CMIP5 GCMs with the observed monthly GPCC data are plotted in Fig. 2. For monthly PR, the interannual variability of bias-corrected 28-GCM ensemble mean agrees well with GPCC observations (the correlation coefficient r = 0.654, Fig. 2a). However, the amplitude of the ensemble mean PR is slightly smaller than that of GPCC. For all global 1° grids, the mean bias-corrected PR during the reference period (1986–2005) are in excellent agreement with the GPCC data (r ~ 0.99, Fig. S1a), but with some underestimation in the standard deviation (STD) of monthly simulated PR (r = 0.91, Fig. S1b). For SPIs (Figs. 2b–e), although there are obvious discrepancies between the simulated ensemble mean and the corresponding GPCC in particular at shorter time scales, the differences are well within the uncertainties (envelops) among 28 GCMs. Furthermore, the correlation between the 28-GCM ensemble mean and GPCC becomes stronger at the longer time scales. Overall, the bias correction of CMIP5 model simulations can reproduce the range of SPIs during the historical period and is therefore considered appropriate for analyzing the projections of meteorological droughts at the global scale.

Fig. 2.

The comparisons of 1986–2005 simulated monthly global-mean PR and SPIs (SPI3, SPI6, SPI12, and SPI24) by CMIP5 GCMs with the observed monthly GPCC data.

Fig. 2.

The comparisons of 1986–2005 simulated monthly global-mean PR and SPIs (SPI3, SPI6, SPI12, and SPI24) by CMIP5 GCMs with the observed monthly GPCC data.

b. PDFs of SPI under 1.5° and 2°C warming levels

The PDFs of multimodel simulated global-mean SPIs and their ensemble means for the baseline (1986–2005) and 20-yr future periods reaching 1.5° and 2°C warming under RCP2.6, RCP4.5, and RCP8.5 are plotted in Fig. 3 for SPI3, SPI6, SPI12, and SPI24. Note that the PDF of 2°C warming under RCP2.6 is not included since only 14 of 28 GCMs can reach 2.0°C warming (see Table 1). A large spread in PDFs among GCMs for all SPIs (especially under RCP8.5) can be seen for all time scales considered. The most notable change in PDFs for both 1.5° and 2°C warming is that the peak Df is larger and shifted to the left relative to the baseline period, particularly at the longer time scales (SPI-24) and under RCP8.5. The pronounced shift to the left in the PDF peaks indicates the general increase in the global-mean Df for both 1.5° and 2°C warming. When comparing the PDF corresponding to two warming levels, a slight shift to the right in the SPI peaks can be found from 1.5° to 2°C warming under RCP4.5 and RCP8.5, suggesting an overall wetting trend due to the additional 0.5°C warming world.

Fig. 3.

The estimated PDFs of the multimodel simulated global-mean SPI3, SPI6, SPI12, and SPI24 and their ensemble means for the baseline (1986–2005) and the 20-yr future periods reaching 1.5° and 2°C warming under 3 RCPs.

Fig. 3.

The estimated PDFs of the multimodel simulated global-mean SPI3, SPI6, SPI12, and SPI24 and their ensemble means for the baseline (1986–2005) and the 20-yr future periods reaching 1.5° and 2°C warming under 3 RCPs.

c. Changes in Df under 1.5° and 2°C warming levels

1) Model agreement in projected changes in Df

Figure 4 plots globally the number of GCMs simulating the increased Df for the 3-month drought (SPI3 ≤ −1) during the 20-yr future period relative to the 1986–2005 baseline period, for both warming levels and all three RCPs considered. The large spatial variability seen in the number of GCMs with the increased Df (SPI3) indicates the large uncertainty among GCMs. Under RCP8.5 and RCP4.5, however, most GCMs (>18) projected an increased Df (SPI3) over north and south Asia, most of North America, and central Africa for both 1.5° and 2°C. More GCMs and a larger extent of areas are identified with an increased Df at 1.5°C than 2°C warming. In contrast, few GCMs (<7) are identified with an increased Df over southwest Asia, southern and northern Africa, most of South America, and most parts of Australia, indicating a higher probability toward wetting trend over these regions. Note that for the 2°C warming under RCP2.6 of which the half of GCMs cannot reach (Table 1), only <10 GCMs can be identified with an increased Df. At the longer time scales such as SPI6, SPI12, and SPI24 (see Figs. S2–S4, respectively), the broadly consistent global pattern in the change of future Df (SPI3) can be found, suggesting a relatively small uncertainty among GCMs across different time scales (SPI3, SPI6, SPI12, and SPI24) of droughts.

Fig. 4.

The global distribution of the number of GCMs identified with the increased frequency of the 3-month drought (SPI3 ≤ −1) relative to the baseline period (1986–2005) in the 20-yr future periods for both warming levels and all three RCPs considered.

Fig. 4.

The global distribution of the number of GCMs identified with the increased frequency of the 3-month drought (SPI3 ≤ −1) relative to the baseline period (1986–2005) in the 20-yr future periods for both warming levels and all three RCPs considered.

2) Model uncertainty and ensemble mean changes in projected Df

The global distributions of uncertainty range and multimodel ensemble mean changes (%) in projected Df relative to the baseline period (1986–2005) for both 1.5° and 2°C warming under all 3 RCPs are plotted in Fig. 5 (SPI3), Fig. S5 (SPI6), Fig. S6 (SPI12), and Fig. S7 (SPI24) in the supplemental material. Note that the 2°C warming under RCP2.6 are not shown due to the same reason mentioned above. As shown in Fig. 5a and Figs. S5a–S7a, the global distribution of model uncertainty is overall uniform, and consistent across different combinations of warming level and RCPs considered. However, the model uncertainty increases significantly from SPI3 to SPI24 (5%–96%), suggesting a larger overall uncertainty in the projected Df at longer time scales.

Fig. 5.

(a) The global distribution of the model uncertainty range and (b) the multimodel ensemble mean changes in the future 3-month Df (SPI3) relative to the baseline period 1986–2005 for both warming under and all RCPs considered. The model uncertainty is represented by the 5th and 95th percentiles of the projected changes in Df among GCMs (see Table 1).

Fig. 5.

(a) The global distribution of the model uncertainty range and (b) the multimodel ensemble mean changes in the future 3-month Df (SPI3) relative to the baseline period 1986–2005 for both warming under and all RCPs considered. The model uncertainty is represented by the 5th and 95th percentiles of the projected changes in Df among GCMs (see Table 1).

Regionally, larger model uncertainty can be found in northern Africa, northern South America, and southern North America for both SPI3 and SPI6, and the global areas with large model uncertainty (30%–60%) are larger for SPI6 than SPI3. At the longer time scale (SPI12 and SPI24), the model uncertainty is significantly larger in many global regions, especially northern Africa, northern South America, and southern North America. It is also found that the model uncertainty is larger for 1.5°C than 2°C warming particularly at longer time scales such as SPI24 (Fig. S7).

The ensemble mean changes in Df show highly similar drying patterns for all time scales considered (Fig. 5b and Fig. S7b), overall consistent with the numbers of GCMs projecting an increased Df (Fig. 4 and Figs. S2–S4). Specifically, a larger percentage increase in future Df is projected over most of North America, northern and southern Asia, southern South America, and some parts of central Africa at 1.5°C warming under all RCPs. In contrast, a decreased Df is projected mainly in southwest Asia, southern and northern Africa, most parts of South America and Australia, with the more pronounced decrease at 1.5°C than 2°C. In comparison, the projected increase in Df at both 1.5° and 2°C warming tends to become larger at longer time scales (e.g., SPI24).

Continental-scale uncertainty of the projected mean Df change (%) are presented are presented in Fig. 6 (SPI3, SPI6, SPI12, and SPI24) for both warming and all 3 RCPs considered. Large uncertainty in the projected Df changes can be seen especially in Europe, North America, and Oceania under higher RCPs and at longer time scales. The Df is projected to increase in Asia, Europe, and North America by most GCMs for both warming and all time scales considered, but with a more pronounced increase in North America (Fig. 6). A larger increase in Df is projected at longer time scales and under higher RCPs. The mean Df increase among 3 RCPs and four time scales considered ranges 2.4%–31.4%, 5.8%–51.1%, and 6.5%–54.7% in Asia, Europe, and North America, respectively (Table 3). The Df decrease is projected mainly in Oceania and South America ranging from −7.9% to 0.2% and from −4.0% to −1.3%, respectively (Table 3).

Fig. 6.

Boxplots of the continental-scale uncertainty of the projected mean Df change (SPI3, SPI6, SPI12, and SPI24) for both warming and all three RCPs considered: 1) Africa, 2) Asia, 3) Europe, 4) North America, 5) Oceania, 6) South America. The boxes denote the interquartile model spread (range between the 25th and 75th quantiles), with the horizontal line indicating the ensemble median and the whiskers showing the extreme range of the selected CMIP5 model simulations identified with the 1.5° and 2°C warming being reached (see Table 1).

Fig. 6.

Boxplots of the continental-scale uncertainty of the projected mean Df change (SPI3, SPI6, SPI12, and SPI24) for both warming and all three RCPs considered: 1) Africa, 2) Asia, 3) Europe, 4) North America, 5) Oceania, 6) South America. The boxes denote the interquartile model spread (range between the 25th and 75th quantiles), with the horizontal line indicating the ensemble median and the whiskers showing the extreme range of the selected CMIP5 model simulations identified with the 1.5° and 2°C warming being reached (see Table 1).

Table 3.

The multimodel ensemble mean changes (%) in the future Df (relative to the baseline period 1986–2005) for the six selected continents at 1.5° and 2°C warming levels and due to the additional 0.5°C warming from 1.5° to 2°C under RCP2.6, RCP4.5, and RCP8.5 scenarios. A “—” denotes there is no projection at 2°C warming level.

The multimodel ensemble mean changes (%) in the future Df (relative to the baseline period 1986–2005) for the six selected continents at 1.5° and 2°C warming levels and due to the additional 0.5°C warming from 1.5° to 2°C under RCP2.6, RCP4.5, and RCP8.5 scenarios. A “—” denotes there is no projection at 2°C warming level.
The multimodel ensemble mean changes (%) in the future Df (relative to the baseline period 1986–2005) for the six selected continents at 1.5° and 2°C warming levels and due to the additional 0.5°C warming from 1.5° to 2°C under RCP2.6, RCP4.5, and RCP8.5 scenarios. A “—” denotes there is no projection at 2°C warming level.

d. Changes in Da under 1.5° and 2°C warming levels

The 20-yr monthly relative changes in the 28-GCM ensemble mean globally averaged Da and the corresponding model uncertainty (i.e., the 5th and 95th percentile estimates) are plotted in Fig. 7 (SPI3) for both warming levels and all three RCPs considered. Similar plots for SPI6, SPI12, and SPI24 are presented in Figs. S8–S10, respectively. The plotted relative changes are the monthly changes of Da over the identified 20-yr future period relative to the 1986–2005 baseline period. Large uncertainty ranges among GCMs can be seen in projected Da particularly under higher RCPs and at longer time scales (e.g., SPI24). However, for all RCPs the uncertainty for both warming levels tend to decrease with time especially at longer time scales (e.g., SPI24 in Fig. S10). Most of the ensemble mean Da are larger than that in the baseline period (1986–2005) particularly for 1.5°C warming (see Figs. S10d,e), suggesting a more significant increase of Da by 1.5°C warming than 2°C. The multimodel ensemble mean global Da show an overall decreased trend over the identified 20-yr periods reaching 1.5° and 2°C warming under all RCP scenarios. In particular, by the end of 20-yr future periods the global meteorological Da tends to be smaller than that in the baseline period (1986–2005) at shorter time scale such as SPI3 (Figs. 7a–c) under RCP2.6 and RCP4.5.

Fig. 7.

The 20-yr monthly relative changes in the mean globally averaged Da (SPI3) and the corresponding model uncertainty (i.e., the 5th and 95th percentile estimates) for both warming levels and all three RCPs considered. The mean change is calculated as the percentage anomaly deviated from the global drought area in the reference period 1986–2005.

Fig. 7.

The 20-yr monthly relative changes in the mean globally averaged Da (SPI3) and the corresponding model uncertainty (i.e., the 5th and 95th percentile estimates) for both warming levels and all three RCPs considered. The mean change is calculated as the percentage anomaly deviated from the global drought area in the reference period 1986–2005.

e. Changes in Df and Da due to an additional 0.5°C warming

Figure 8 plots the global distribution of model uncertainty in Df due to the additional 0.5°C warming for all SPI3, SPI6, SPI12, and SPI24 under RCP4.5 and RCP8.5. As seen, model uncertainties increase considerably with the time scale from 3%–39% (SPI3, SPI6) to 75% (SPI24). At longer time scales (SPI12, SPI24) larger uncertainties are found mainly in North America, northern Asia, northern Europe, and central Africa.

Fig. 8.

The global distribution of model uncertainty in the future Df due to the additional 0.5°C warming for all SPI3, SPI6, SPI12, and SPI24 under RCP4.5 and RCP8.5. The model uncertainty range is represented by the difference between 5th and 95th percentile estimates of the projected changes in Df across the selected GCMs (see Table 1).

Fig. 8.

The global distribution of model uncertainty in the future Df due to the additional 0.5°C warming for all SPI3, SPI6, SPI12, and SPI24 under RCP4.5 and RCP8.5. The model uncertainty range is represented by the difference between 5th and 95th percentile estimates of the projected changes in Df across the selected GCMs (see Table 1).

Figure 9 plots the multimodel ensemble mean changes in Df due to the additional 0.5°C warming (from 1.5° to 2°C) for all SPI3, SPI6, SPI12, and SPI24 under RCP4.5 and RCP8.5. As shown, an overall significant reduction (up to ~70%) in future Df can be found over many global regions, mainly in the high-latitude regions of North America and northern Eurasia, while a significant increase (up to ~40%) over most of South America, southern Africa, southwest and eastern Asia, and some parts of Australia. Generally, the changes in future Df due to the additional 0.5°C warming for SPI24 is more significant than that at the shorter time scales such as for SPI3. For all SPIs, the Df changes are in general larger under RCP4.5 than RCP8.5 (Fig. 9).

Fig. 9.

The global distribution of the multimodel ensemble mean changes in Df due to the additional 0.5°C warming from 1.5° to 2°C for all SPI3, SPI6, SPI12, and SPI24 under both RCP4.5 and RCP8.5.

Fig. 9.

The global distribution of the multimodel ensemble mean changes in Df due to the additional 0.5°C warming from 1.5° to 2°C for all SPI3, SPI6, SPI12, and SPI24 under both RCP4.5 and RCP8.5.

Figure 10 presents the boxplot of the relative change (%) in Df due to the additional 0.5°C warming under RCP4.5 and RCP8.5 scenarios for six global continents. A similar pattern of changes in Df over six continents can be seen for all RCPs and SPIs considered in that considerably larger uncertainty in Df changes occurs in Europe and Oceania. The uncertainty ranges at longer time scales (SPI24) or under RCP8.5 are consistently higher than that at smaller time scale (SPI3) or under RCP4.5. For all SPIs and RCPs considered, the mean Df changes due to the additional 0.5°C warming are negative in Asia (from −21.6% to −7.2%), Europe (from −31.7% to −11.6%), North America (from −40.7% to −14.3%), and Oceania (from −4.69% to −1.19%), whereas the positive changes in Df can only be found in South America (except SPI6 under RCP8.5) and Africa (for SPI3 under RCP4.5) despite the magnitude is rather small (Table 3).

Fig. 10.

Boxplots of the relative change (%) in Df due to the additional 0.5°C warming from 1.5° to 2°C under RCP4.5 and RCP8.5 scenarios for six global continents: 1) Africa, 2) Asia, 3) Europe, 4) North America, 5) Oceania, 6) South America. The boxes denote the interquartile model spread (range between the 25th and 75th quantiles), with the horizontal line indicating the ensemble median and the whiskers showing the extreme range of the selected CMIP5 model simulations identified with 1.5° and 2°C warming levels (see Table 1).

Fig. 10.

Boxplots of the relative change (%) in Df due to the additional 0.5°C warming from 1.5° to 2°C under RCP4.5 and RCP8.5 scenarios for six global continents: 1) Africa, 2) Asia, 3) Europe, 4) North America, 5) Oceania, 6) South America. The boxes denote the interquartile model spread (range between the 25th and 75th quantiles), with the horizontal line indicating the ensemble median and the whiskers showing the extreme range of the selected CMIP5 model simulations identified with 1.5° and 2°C warming levels (see Table 1).

Figure 11 plots the 20-yr monthly relative changes in the multimodel ensemble mean globally averaged Da and the corresponding uncertainty range (i.e., the 5th and 95th percentiles) due to the additional 0.5°C warming for all SPIs under RCP4.5 and RCP8.5. The relative change is calculated as the percentage change of the global-average Da at 2°C relative to that at 1.5°C. As seen, the uncertainty (range between the 5th and 95th percentile estimates) in projected Da increases with time scales (>15% for SPI24). The Da difference between 1.5° and 2°C warming tends to increase with time, particularly under RCP8.5 and at longer time scale (SPI24). Overall, an additional 0.5°C warming results in a reduction in Da under both RCPs, particularly significant at longer time scales (e.g., SPI24). The mean decreases in global Da over the 20-yr period due to the additional 0.5°C warming range among 4 SPIs considered between 1.2% and 3.7% (1.5% and 4.6%) under RCP4.5 (RCP8.5).

Fig. 11.

The 20-yr monthly relative changes in the ensemble mean globally averaged Da and the corresponding uncertainty range (i.e., the 5th and 95th percentiles) due to the additional 0.5°C warming for all SPIs under RCP4.5 and RCP8.5. The model uncertainty range is represented by the 5th and 95th percentile estimates of the projected changes in Da across the selected GCMs (see Table 1).

Fig. 11.

The 20-yr monthly relative changes in the ensemble mean globally averaged Da and the corresponding uncertainty range (i.e., the 5th and 95th percentiles) due to the additional 0.5°C warming for all SPIs under RCP4.5 and RCP8.5. The model uncertainty range is represented by the 5th and 95th percentile estimates of the projected changes in Da across the selected GCMs (see Table 1).

4. Discussion

To evaluate different types of drought, various indices have been developed to describe the variations of dry and wet episodes. In this study, SPI is used to assess global meteorological droughts under future 1.5° and 2°C warming. The frequency of drought occurrence is quantified by the percentage of the months with SPI ≤ −1 (McKee et al. 1993) within a given period, as often adopted in previous studies (e.g., Duffy et al. 2015; Rhee and Cho 2016; Wu et al. 2016). Note that this method of definition of Df does not consider the actual drought duration. Therefore, as same as in the aforementioned studies, this study mainly focuses on the identification of drought occurrence with different magnitudes (moderate, severe, extreme, see Table 2) instead of the duration of droughts.

Our results suggest that future changes in droughts vary significantly among global regions. Overall, the multimodel ensemble mean projects an increasing Df under 1.5°C warming relative to the 1986–2005 baseline period for most mid- to high-latitude regions in the Northern Hemisphere (e.g., northern Asia, northern Europe, and North America, Fig. 5), and a general increase in the global Da particularly under RCP4.5 and RCP8.5 (Fig. 7). Under 2°C warming, however, the global patterns of the changes in Df and Da remain similar, but the magnitude is much smaller (e.g., Fig. 5). Thus, an additional 0.5°C warming from 1.5° to 2°C results in an overall reduction in both Df and Da under RCP4.5 and RCP8.5 over most global regions, particularly in the Northern Hemisphere and at longer time scales such as SPI12 and SPI24 (Figs. 9 and 11), which is overall consistent with the drought projection in China by Su et al. (2018) based also on SPI and projected decreasing mean coverage and frequency of drought events in China under an additional 0.5°C warming from 1.5° to 2°C due to the increased PR. The reduction in both Df and Da due to an additional 0.5°C warming under RCP4.5 and RCP8.5 (Figs. 9 and 11) can mainly be attributed to the overall increased PR trends over most Northern Hemisphere (Asia, Europe, Northern America, Fig. 1) for most 20-yr or longer subperiods within 2006–2100, consistent with the previous findings on the PR projections from CMIP5 GCM simulations (Lau et al. 2013; Taylor et al. 2013; Radić et al. 2014; Zhao and Dai 2015). Specifically, the increased Df in most parts of South America (under RCP4.5 and RCP8.5, Fig. 9) due to an additional 0.5°C warming can be mainly attributed to the decreased PR trend during the identified periods reaching 1.5° and 2°C warming (see Fig. 1 for South America).

Note that there are considerable uncertainties in projected spatiotemporal pattern and magnitude of Df (Figs. 6 and 9). The first uncertainty source is on the selected drought index SPI, which represents only the precipitation-driven meteorological drought with limited information on potential impacts on agriculture, ecosystems and societies. However, in addition to PR both evapotranspiration and soil moisture can have important influences in regional drought development (Su et al. 2018). The drought metrics related to soil moisture and runoff deficits [e.g., the Palmer drought severity index (PDSI)] are more sensitive to temperature change than SPI (Burke 2011; Dai 2013; Dai and Zhao 2017; Zhao and Dai 2017), since the warming increases evapotranspiration and leads to the deficit in soil moisture and reduction in runoff. It has been reported that the sign and magnitude of drought changes are highly dependent on the index definition (Burke and Brown 2008). Furthermore, large discrepancies between the projected SPI and other indices measuring of the atmospheric demand for moisture have been reported, such as little change in the proportion of the land surface in SPI-based drought (e.g., Burke and Brown 2008; Touma et al. 2015; Rhee and Cho 2016). Therefore, further evaluation of other drought indices sensitive to global warming (e.g., PDSI) is necessary in order to enhance the credibility of future drought projections.

The second uncertainty source is related to the 28-GCM simulations and three RCPs used in this study for the comparative analysis and systematical investigation of the uncertainty in projected future meteorological drought. Our results indicate the wide spread in the timing reaching global 1.5° and 2°C warming among GCMs under alternative RCPs (Table 1), and the model uncertainties significantly affect the projected spatial and temporal change in drought characteristics (e.g., Df) under 1.5° and 2°C warming as well as the additional 0.5°C warming especially at longer time scales and under higher RCPs (Figs. 5 and 8). GCMs and emission scenarios are commonly conceived as two main uncertainty sources in climate change projections, and the former is generally larger than the later (Kay et al. 2009; Giuntoli et al. 2015; Vetter et al. 2017; Xu et al. 2019b). Some recent studies comparing multiple GCM-projected PR have emphasized the large dependence on the models used (e.g., Mehran et al. 2014; Osuch et al. 2016; Xu et al. 2019b). Therefore, the results of impact assessment studies of 1.5° and 2°C warming should be interpreted with cautions if only few models or one emission scenario is considered. It should also be noted that the 20-yr window used for identifying the reach of 1.5° and 2.0°C warming may not be sufficiently long to capture the degree of freedom of T data, which may introduce the uncertainty into the timing reaching global 1.5° and 2°C warming.

The third potential uncertainty source is the use of bias-correction method. In this study, the equidistant quantile matching method was used for correcting the 28-GCM simulations to reduce the uncertainty in model physics and parameterization (Sheffield and Wood 2008). The results suggest that the envelope of the bias corrections of GCM simulations encompasses the GPCC observations well, a generally consistent finding with Nasrollahi et al. (2015). However, we found that the use of bias-correction method tends to result in an underestimation of simulated PR spatial variability (i.e., STD) (Fig. S1) as well as slightly smaller temporal variability of PR and SPIs than the observations (Fig. 2), suggesting that the use of bias correction methods may introduce additional uncertainty to model projections.

5. Conclusions

Based on the standardized precipitation index (SPI), this paper presents a global assessment of precipitation-driven meteorological drought frequency (Df) and areas (Da) at the multiple time scales (i.e., 3, 6, 12, and 24 months) under the 1.5° and 2°C global warming levels. We used the ensemble mean projection data simulated from 28 CMIP5 GCMs under three RCP scenarios (RCP2.6, RCP4.5, and RCP8.5) for a systematic analyses on the uncertainty in global drought projections, and test the sensitivity of future Df and Da changes to the projected 1.5° and 2°C warming targets corresponding to different time scales of drought occurrence.

Results show that there are large differences among 28 GCMs in the timing reaching 1.5° and 2°C warming targets under alternative RCP scenarios considered in this study. Model uncertainties significantly affect the projected spatial and temporal change in Df under 1.5° and 2°C global warming as well as due to the additional 0.5°C warming, particularly at longer time scales and under higher RCP scenarios. However, most GCMs project an increased trend of Df over most global regions, such as North America, Europe, northern Asia, and central Africa, at both 1.5° and 2°C warming under RCP4.5 and RCP8.5. Furthermore, the magnitude of the projected increase in Df tends to be larger under 1.5°C warming and at longer time scales such as SPI24. In contrast, a decrease in Df is projected mainly in southern and northern Africa, South America, and some parts of Australia. GCMs also project an overall increase in global Da relative to the baseline period (1986–2005) at both 1.5° and 2°C warming. However, the Da shows an overall decreased trend during the 20-yr future periods reaching 1.5° and 2°C warming due to the increased global precipitation.

An additional 0.5°C warming from 1.5° to 2°C leads to a decrease in Df in most global regions such as North America, Europe, northern Asia, and northern Africa, but increase in southwest Asia and some parts of the Southern Hemisphere (e.g., South America, southern Africa, and some parts of Australia). Similarly, there is a general decreased trend in global Da under the additional 0.5°C warming with a larger magnitude at longer time scales, mainly due to the increased precipitation under warming. Overall, this study underscores the significant spatiotemporal uncertainties in the projected change of drought characteristics at multiple time scales originated from different GCMs under alternative emissions scenarios considered, and highlights the need to consider multiple GCMs in drought projection studies under the context of the Paris Agreement targets.

Acknowledgments

This research was supported by funding from the National Natural Science Foundation of China (Grant 51909106, 51879108), the Natural Science Foundation of Guangdong Province, China (Grant 2018A030310653), The high-level talent project for the “Pearl River Talent Plan” of Guangdong Province (Grant 2017GC010397), the Youth Innovative Talents Project for Guangdong Colleges and Universities (Grant 2017KQNCX010), and the Fundamental Research Funds for the Central Universities (Grant 21617301). The CMIP5 model data used in this study are publicly available from the Canadian Climate Data and Scenarios (CCDS) (http://climatescenarios.canada.ca/?page=main). The CRU TS 3.22 global 0.5° gridded temperature dataset is obtained from the Climatic Research Unit (available at https://crudata.uea.ac.uk/cru/data/hrg/). The GPCC global 1° gridded precipitation dataset is obtained from the Global Precipitation Climatology Centre (https://psl.noaa.gov/data/gridded/data.gpcc.html#detail).

REFERENCES

REFERENCES
Arnell
,
N. W.
, and
S. N.
Gosling
,
2013
:
The impacts of climate change on river flow regimes at the global scale
.
J. Hydrol.
,
486
,
351
364
, https://doi.org/10.1016/j.jhydrol.2013.02.010.
Betts
,
R. A.
, and et al
,
2018
:
Changes in climate extremes, fresh water availability and vulnerability to food insecurity projected at 1.5 °C and 2 °C global warming with a higher-resolution global climate model
.
Philos. Trans. Roy. Soc.
,
376A
, 20160452, https://doi.org/10.1098/rsta.2016.0452.
Burke
,
E. J.
,
2011
:
Understanding the sensitivity of different drought metrics to the drivers of drought under increased atmospheric CO2
.
J. Hydrometeor.
,
12
,
1378
1394
, https://doi.org/10.1175/2011JHM1386.1.
Burke
,
E. J.
, and
S. J.
Brown
,
2008
:
Evaluating uncertainties in the projection of future drought
.
J. Hydrometeor.
,
9
,
292
299
, https://doi.org/10.1175/2007JHM929.1.
Dai
,
A.
,
2011
:
Drought under global warming: A review
.
Wiley Interdiscip. Rev.: Climate Change
,
2
,
45
65
, https://doi.org/10.1002/wcc.81.
Dai
,
A.
,
2013
:
Increasing drought under global warming in observations and models
.
Nat. Climate Change
,
3
,
52
58
, https://doi.org/10.1038/nclimate1633.
Dai
,
A.
, and
T.
Zhao
,
2017
:
Uncertainties in historical changes and future projections of drought. Part I: Estimates of historical drought changes
.
Climatic Change
,
144
,
519
533
, https://doi.org/10.1007/s10584-016-1705-2.
Donnelly
,
C.
,
W.
Greuell
,
J.
Andersson
,
D.
Gerten
,
G.
Pisacane
,
P.
Roudier
, and
F.
Ludwig
,
2017
:
Impacts of climate change on European hydrology at 1.5, 2 and 3 degrees mean global warming above preindustrial level
.
Climatic Change
,
143
,
13
26
, https://doi.org/10.1007/s10584-017-1971-7.
Duffy
,
P. B.
,
P.
Brando
,
G. P.
Asner
, and
C. B.
Field
,
2015
:
Projections of future meteorological drought and wet periods in the Amazon
.
Proc. Natl. Acad. Sci. USA
,
112
,
13 172
13 177
, https://doi.org/10.1073/pnas.1421010112.
Giuntoli
,
I.
,
J. P.
Vidal
,
C.
Prudhomme
, and
D. M.
Hannah
,
2015
:
Future hydrological extremes: The uncertainty from multiple global climate and global hydrological models
.
Earth Syst. Dyn.
,
6
,
267
285
, https://doi.org/10.5194/esd-6-267-2015.
Harris
,
I.
,
P. D.
Jones
,
T. J.
Osborna
, and
D. H.
Lister
,
2014
:
Updated high-resolution grids of monthly climatic observations –the CRU TS3.10 Dataset
.
Int. J. Climatol.
,
34
,
623
642
, https://doi.org/10.1002/joc.3711.
Huntington
,
T. G.
,
2006
:
Evidence for intensification of the global water cycle: Review and synthesis
.
J. Hydrol.
,
319
,
83
95
, https://doi.org/10.1016/j.jhydrol.2005.07.003.
Kay
,
A. L.
,
H. N.
Davies
,
V. A.
Bell
, and
R. G.
Jones
,
2009
:
Comparison of uncertainty sources for climate change impacts: Flood frequency in England
.
Climatic Change
,
92
,
41
63
, https://doi.org/10.1007/s10584-008-9471-4.
Kendall
,
M. G.
,
1975
:
Rank Correlation Methods. 4th ed. Charles Griffin, 202 pp
.
King
,
A. D.
, and
D. J.
Karoly
,
2017
:
Climate extremes in Europe at 1.5 and 2 degrees of global warming
.
Environ. Res. Lett.
,
12
, 114031, https://doi.org/10.1088/1748-9326/aa8e2c.
King
,
A. D.
,
D. J.
Karoly
, and
B. J.
Henley
,
2017
:
Australian climate extremes at 1.5 °C and 2 °C of global warming
.
Nat. Climate Change
,
7
,
412
416
, https://doi.org/10.1038/nclimate3296.
Lau
,
W. K. M.
,
H. T.
Wu
, and
K. M.
Kim
,
2013
:
A canonical response of precipitation characteristics to global warming from CMIP5 models
.
Geophys. Res. Lett.
,
40
,
3163
3169
, https://doi.org/10.1002/grl.50420.
Lehner
,
F.
,
S.
Coats
,
T. F.
Stocker
,
A. G.
Pendergrass
,
B. M.
Sanderson
,
C. C.
Raible
, and
J. E.
Smerdon
,
2017
:
Projected drought risk in 1.5°C and 2°C warmer climates
.
Geophys. Res. Lett.
,
44
,
7419
7428
, https://doi.org/10.1002/2017GL074117.
Leng
,
G.
,
2018
:
Keeping global warming within 1.5° C reduces future risk of yield loss in the United States: A probabilistic modeling approach
.
Sci. Total Environ.
,
644
,
52
59
, https://doi.org/10.1016/j.scitotenv.2018.06.344.
Li
,
H.
,
J.
Sheffield
, and
E. F.
Wood
,
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
, https://doi.org/10.1029/2009JD012882.
Li
,
Z.
,
X.
Guo
,
Y.
Yang
,
Y.
Hong
,
Z.
Wang
, and
L.
You
,
2019
:
Heatwave trends and the population exposure over China in the 21st century as well as under 1.5 °C and 2.0 °C global warmer future scenarios
.
Sustainability
,
11
,
3318
, https://doi.org/10.3390/su11123318.
Lu
,
G. H.
,
Y.
Yang
,
Z. Y.
Wu
,
H.
He
, and
H.
Xiao
,
2014
:
Temporal and spatial variations of snow depth in regions of the upper reaches of Yangtze River under future climate change scenarios: A study based on CMIP5 multi-model ensemble projections (in Chinese)
.
Shui Kexue Jinzhan
,
25
,
484
493
.
McKee
,
T. B.
,
N. J.
Doesken
, and
J.
Kleist
,
1993
:
The relationship of drought frequency and duration to time scales. Eighth Conf. on Applied Climatology, Anaheim, CA, Amer. Meteor. Soc., 179–184
.
McKee
,
T. B.
,
N. J.
Doesken
, and
J.
Kleist
,
1995
:
Drought monitoring with multiple time scales. Ninth Conf. on Applied Climatology, Dallas, TX, Amer. Meteor. Soc., 233–236
.
Mehran
,
A.
,
A.
AghaKouchak
, and
T. J.
Phillips
,
2014
:
Evaluation of CMIP5 continental precipitation simulations relative to satellite-based gauge-adjusted observations
.
J. Geophys. Res. Atmos.
,
119
,
1695
1707
, https://doi.org/10.1002/2013jd021152.
Milliman
,
J. D.
,
K. L.
Farnsworth
,
P. D.
Jones
,
K. H.
Xu
, and
L. C.
Smith
,
2008
:
Climatic and anthropogenic factors affecting river discharge to the global ocean, 1951–2000
.
Global Planet. Change
,
62
,
187
194
, https://doi.org/10.1016/j.gloplacha.2008.03.001.
Nangombe
,
S.
,
T.
Zhou
,
W.
Zhang
,
B.
Wu
,
S.
Hu
,
L.
Zou
, and
D.
Li
,
2018
:
Record-breaking climate extremes in Africa under stabilized 1.5 °C and 2 °C global warming scenarios
.
Nat. Climate Change
,
8
,
375
380
, https://doi.org/10.1038/s41558-018-0145-6.
Nasrollahi
,
N.
,
A.
AghaKouchak
,
L.
Cheng
,
L.
Damberg
,
T. J.
Phillips
,
C.
Miao
,
K.
Hsu
, and
S.
Sorooshian
,
2015
:
How well do CMIP5 climate simulations replicate historical trends and patterns of meteorological droughts?
Water Resour. Res.
,
51
,
2847
2864
, https://doi.org/10.1002/2014WR016318.
Orlowsky
,
B.
, and
S. I.
Seneviratne
,
2013
:
Elusive drought: Uncertainty in observed trends and short- and long-term CMIP5 projections
.
Hydrol. Earth Syst. Sci.
,
17
,
1765
1781
, https://doi.org/10.5194/hess-17-1765-2013.
Osuch
,
M.
,
R. J.
Romanowicz
,
D.
Lawrence
, and
W. K.
Wong
,
2016
:
Trends in projections of standardized precipitation indices in a future climate in Poland
.
Hydrol. Earth Syst. Sci.
,
20
,
1947
1969
, https://doi.org/10.5194/hess-20-1947-2016.
Prudhomme
,
C.
, and et al
,
2014
:
Hydrological droughts in the 21st century, hotspots and uncertainties from a global multimodel ensemble experiment
.
Proc. Natl. Acad. Sci. USA
,
111
,
3262
3267
, https://doi.org/10.1073/pnas.1222473110.
Radić
,
V.
,
A.
Bliss
,
A. C.
Beedlow
,
R.
Hock
,
E.
Miles
, and
J. G.
Cogley
,
2014
:
Regional and global projections of twenty-first century glacier mass changes in response to climate scenarios from global climate models
.
Climate Dyn.
,
42
,
37
58
, https://doi.org/10.1007/s00382-013-1719-7.
Rhee
,
J.
, and
J.
Cho
,
2016
:
Future changes in drought characteristics: Regional analysis for South Korea under CMIP5 projections
.
J. Hydrometeor.
,
17
,
437
451
, https://doi.org/10.1175/JHM-D-15-0027.1.
Sanderson
,
B. M.
, and et al
,
2017
:
Community climate simulations to assess avoided impacts in 1.5 and 2 °C futures
.
Earth Syst. Dyn.
,
8
,
827
847
, https://doi.org/10.5194/esd-8-827-2017.
Schleussner
,
C. F.
, and et al
,
2016a
:
Differential climate impacts for policy-relevant limits to global warming: The case of 1.5 C and 2 C
.
Earth Syst. Dyn.
,
7
,
327
351
, https://doi.org/10.5194/esd-7-327-2016.
Schleussner
,
C. F.
, and et al
,
2016b
:
Science and policy characteristics of the Paris Agreement temperature goal
.
Nat. Climate Change
,
6
,
827
835
, https://doi.org/10.1038/nclimate3096.
Sheffield
,
J.
, and
E. F.
Wood
,
2008
:
Projected changes in drought occurrence under future global warming from multi-model, multi-scenario, IPCC AR4 simulations
.
Climate Dyn.
,
31
,
79
105
, https://doi.org/10.1007/s00382-007-0340-z.
Sillmann
,
J.
,
V. V.
Kharin
,
F. W.
Zwiers
,
X.
Zhang
, and
D.
Bronaugh
,
2013
:
Climate extremes indices in the CMIP5 multimodel ensemble: Part 2. Future climate projections
.
J. Geophys. Res. Atmos.
,
118
,
2473
2493
, https://doi.org/10.1002/jgrd.50188.
Stocker
,
T. F.
,
2013
:
The closing door of climate targets
.
Science
,
339
,
280
282
, https://doi.org/10.1126/science.1232468.
Su
,
B.
, and et al
,
2018
:
Drought losses in China might double between the 1.5° C and 2.0° C warming
.
Proc. Natl. Acad. Sci. USA
,
115
,
10 600
10 605
, https://doi.org/10.1073/pnas.1802129115.
Taylor
,
I. H.
,
E.
Burke
,
L.
McColl
,
P. D.
Falloon
,
G. R.
Harris
, and
D.
McNeall
,
2013
:
The impact of climate mitigation on projections of future drought
.
Hydrol. Earth Syst. Sci.
,
17
,
2339
2358
, https://doi.org/10.5194/hess-17-2339-2013.
Touma
,
D.
,
M.
Ashfaq
,
M. A.
Nayak
,
S. C.
Kao
, and
N. S.
Diffenbaugh
,
2015
:
A multi-model and multiindex evaluation of drought characteristics in the 21st century
.
J. Hydrol.
,
526
,
196
207
, https://doi.org/10.1016/j.jhydrol.2014.12.011.
UNFCCC
,
2015
:
Adoption of the Paris agreement. United Nations Framework Convention on Climate Change, 32 pp.
, http://unfccc.int/resource/docs/2015/cop21/eng/l09r01.pdf.
Vetter
,
T.
, and et al
,
2017
:
Evaluation of sources of uncertainty in projected hydrological changes under climate change in 12 large-scale river basins
.
Climatic Change
,
141
,
419
433
, https://doi.org/10.1007/s10584-016-1794-y.
Wang
,
Z.
,
L.
Lin
,
X.
Zhang
,
H.
Zhang
,
L.
Liu
, and
Y.
Xu
,
2017
:
Scenario dependence of future changes in climate extremes under 1.5 °C and 2 °C global warming
.
Sci. Rep.
,
7
,
46432
, https://doi.org/10.1038/srep46432.
Wilhite
,
D.
,
2000
:
Drought as a natural hazard: Concepts and definitions. Drought: A Global Assessment, D. A. Whilhite, Ed., Routledge, 3–18
.
Wu
,
C. H.
, and
G. R.
Huang
,
2016
:
Projection of climate extremes in the Zhujiang River basin using a regional climate model
.
Int. J. Climatol.
,
36
,
1184
1196
, https://doi.org/10.1002/joc.4412.
Wu
,
C. H.
,
Z. Y.
Xian
, and
G. R.
Huang
,
2016
:
Meteorological drought in the Beijiang River basin, South China: Current observations and future projections
.
Stochastic Environ. Res. Risk Assess.
,
30
,
1821
1834
, https://doi.org/10.1007/s00477-015-1157-7.
Wu
,
C. H.
,
B. X.
Hu
,
G. R.
Huang
,
P.
Wang
, and
K.
Xu
,
2018
:
Responses of runoff to historical and future climate variability over China
.
Hydrol. Earth Syst. Sci.
,
22
,
1971
1991
, https://doi.org/10.5194/hess-22-1971-2018.
Xu
,
K.
,
C. H.
Wu
, and
B. X.
Hu
,
2019a
:
Projected changes of temperature extremes over nine major basins in China based on the CMIP5 multimodel ensembles
.
Stochastic Environ. Res. Risk Assess.
,
33
,
321
339
, https://doi.org/10.1007/s00477-018-1569-2.
Xu
,
K.
,
B. B.
Xu
,
J. L.
Ju
,
C. H.
Wu
,
H.
Dai
, and
B. X.
Hu
,
2019b
:
Projection and uncertainty of precipitation extremes in the CMIP5 multimodel ensembles over nine major basins in China
.
Atmos. Res.
,
226
,
122
137
, https://doi.org/10.1016/j.atmosres.2019.04.018.
Yeh
,
P. J.-F.
, and
C. H.
Wu
,
2018
:
Recent acceleration of the terrestrial hydrologic cycle in the U.S. Midwest
.
J. Geophys. Res. Atmos.
,
123
,
993
3008
, https://doi.org/10.1002/2017jd027706.
Zhai
,
R.
,
F.
Tao
, and
Z.
Xu
,
2018
:
Spatial–temporal changes in runoff and terrestrial ecosystem water retention under 1.5 and 2 °C warming scenarios across China
.
Earth Syst. Dyn.
,
9
,
717
738
, https://doi.org/10.5194/esd-9-717-2018.
Zhao
,
T.
, and
A.
Dai
,
2015
:
The magnitude and causes of global drought changes in the twenty-first century under a low–moderate emissions scenario
.
J. Climate
,
28
,
4490
4512
, https://doi.org/10.1175/JCLI-D-14-00363.1.
Zhao
,
T.
, and
A.
Dai
,
2017
:
Uncertainties in historical changes and future projections of drought. Part II: Model simulated historical and future drought changes
.
Climatic Change
,
144
,
535
548
, https://doi.org/10.1007/s10584-016-1742-x.
For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).