Nothing Special   »   [go: up one dir, main page]
Next Article in Journal
Mix MSTAR: A Synthetic Benchmark Dataset for Multi-Class Rotation Vehicle Detection in Large-Scale SAR Images
Next Article in Special Issue
Propagation of Meteorological Drought to Agricultural and Hydrological Droughts in the Tropical Lancang–Mekong River Basin
Previous Article in Journal
Driving Source of Change for Ionosphere before Large Earthquake -Vertical Ground Motion-
Previous Article in Special Issue
Research on Soil Moisture Inversion Method for Canal Slope of the Middle Route Project of the South to North Water Transfer Based on GNSS-R and Deep Learning
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Remote Sensing
Volume 15
Issue 18
10.3390/rs15184557
remotesensing-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Actual Evapotranspiration Dominates Drought in Central Asia

by
Zhuoyi Zhao
1,2,3,
Xingming Hao
1,3,*,
Xue Fan
1,2,3,
Jingjing Zhang
1,2,3,
Sen Zhang
1,2,3 and
Xuewei Li
1,2,3
1
State Key Laboratory of Desert and Oasis Ecology, Key Laboratory of Ecological Safety, Sustainable Development in Arid Lands, Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, Urumqi 830011, China
2
University of Chinese Academy of Sciences, Beijing 100049, China
3
Akesu National Station of Observation and Research for Oasis Agro-Ecosystem, Akesu 843017, China
*
Author to whom correspondence should be addressed.
Remote Sens. 2023, 15(18), 4557; https://doi.org/10.3390/rs15184557
Submission received: 8 August 2023 / Revised: 6 September 2023 / Accepted: 13 September 2023 / Published: 16 September 2023
(This article belongs to the Special Issue Exploring the Variations and Interplay of Drought and Heatwaves: Uncovering the Associated Mechanisms and Impacts)
Browse Figures
Graphical abstract
"> Figure 1
<p>Map of the study area. (<b>a</b>) Lakes, mountains, and administrative units in Central Asia. TS, KL, KAZ, UZB, TKM, KGZ, TJK, and ANW represent Tianshan Mountains, Kunlun Mountains, Kazakhstan, Uzbekistan, Turkmenistan, Kyrgyzstan, Tajikistan, and the arid regions of Northwestern China, respectively. (<b>b</b>) Aridity indexes, watersheds, and rivers in Central Asia. HAA, AA, SAA, SHA, and HA represent hyperarid, arid, semiarid, subhumid, and humid areas, respectively, and the numbers 1–11 represent the 11 watersheds located in the study area, which are the Ural River Basin, Uzboy River Basin, Irtysh River Basin, Syr Darya River Basin, Amu Darya River Basin, Balkhash Lake Basin, Manas River Basin, Baiyang River Basin, Tarim River Basin, Heihe River Basin, and Yellow River Basin. The boundary map was downloaded from <a href="http://bzdt.ch.mnr.gov.cn/index.html" target="_blank">http://bzdt.ch.mnr.gov.cn/index.html</a> with the approval number GS (2021)5453 accessed on 14 September 2023.</p> "> Figure 2
<p>Workflow of this study. TWS, LA, SM, ET, HD, AD, MD, SPAET, and LMG denote terrestrial water storage, lake area, soil moisture, actual evapotranspiration, hydrological drought, agricultural drought, meteorological drought, the spatial efficiency metric, and the Lindeman–Merenda–Gold method, respectively.</p> "> Figure 3
<p>(<b>a</b>–<b>f</b>) Spatiotemporal variations in meteorological, agricultural, and hydrologic drought in arid Central Asia in the period 2000–2021. Black points mark pixels with a significant linear trend (<span class="html-italic">p</span> &lt; 0.05) of annual indicators. P, SM, LA, and TWS represent precipitation, soil moisture, the lake area, and terrestrial water storage, respectively. The top right corners of show probability density histograms.</p> "> Figure 4
<p>Meteorological, agricultural, and hydrological drought parameters.</p> "> Figure 5
<p>Correlations between evapotranspiration (ET), precipitation (P), terrestrial water reserves (TWS), the lake area (LA), and soil moisture (SM) in the period 2000–2020. The blue line’s label is seen as the bottom blue labels and the ticks are seen as the left blue ticks in the top row; the orange line’s label is seen as the left orange labels and the ticks are seen as the right y-axis in the right-most column. The x-axis for both is the x-axis of the top row.</p> "> Figure 6
<p>Spatial efficiency of trends and characteristics of various types of drought in the period 2003–2020, where HD, AD, MD, SPAEF, α, β, and γ represent hydrological drought, agricultural drought, meteorological drought, spatial efficiency, Pearson’s correlation coefficient, the fraction of the coefficient of variation, and histogram match, respectively.</p> "> Figure 7
<p>Contributions of precipitation (P) (<b>a</b>,<b>b</b>), evapotranspiration (ET) (<b>c</b>,<b>d</b>), and runoff (<b>e</b>,<b>f</b>) to agricultural drought intensity (<b>a</b>,<b>c</b>,<b>e</b>) and hydrologic drought intensity (<b>b</b>,<b>d</b>,<b>f</b>). Black points indicate values of statistical significance (<span class="html-italic">p</span> &lt; 0.05).</p> "> Figure 8
<p>Precipitation (P) and actual evapotranspiration (ET) cross-wavelet (XWT) (<b>a</b>) and wavelet coherence (WTC) (<b>b</b>) spectra. In the figure, “→” indicates a positive correlation, “←” indicates a negative correlation, and “↓” indicates that the ET phase change exceeds the precipitation phase by 90° (i.e., it exceeds the precipitation phase by ¼ of a cycle in the time series). The time–frequency domain within the thick black solid line passes the significance test (<span class="html-italic">p</span> &lt; 0.05).</p> "> Figure 9
<p>Contributions of precipitation (P) (<b>a</b>,<b>b</b>), evapotranspiration (ET) (<b>c</b>,<b>d</b>), and runoff (<b>e</b>,<b>f</b>) to SPEI (<b>a</b>,<b>c</b>,<b>e</b>) and PDSI (<b>b</b>,<b>d</b>,<b>f</b>).</p> "> Figure A1
<p>Contributions of precipitation (<b>a</b>,<b>b</b>), evapotranspiration (<b>c</b>,<b>d</b>), and runoff (<b>e</b>,<b>f</b>) to soil moisture (<b>a</b>,<b>c</b>,<b>e</b>) and terrestrial water storage (<b>b</b>,<b>d</b>,<b>f</b>).</p> "> Figure A2
<p>Variations in evapotranspiration (ET) and its components in the period 2001–2019, spatial distributions of (<b>a</b>) ET, (<b>b</b>) vegetation transpiration (ETc), (<b>c</b>) evaporation (ETs) trends, and (<b>d</b>) the time series of ET and its components. ETi represents canopy interception evaporation.</p> ">
Versions Notes

Abstract

:
Central Asia is a drought-prone region that is sensitive to global climate change. The increased actual evapotranspiration intensifies the drought impacts in this area. However, little is known about the similarities and differences between various types of drought in Central Asia, as well as the relative importance of water income and consumption processes during drought events. Therefore, this study evaluates the trends and characteristics of meteorological, agricultural, and hydrological droughts in Central Asia using precipitation, soil moisture, and terrestrial water storage as indicators; explores the temporal correlation of and spatial similarity between various types of drought; and quantitatively assesses the contribution of water balance variables to drought intensity. The results indicate that drought has intensified in Central Asia, and the trends of precipitation, soil moisture, and terrestrial water storage in this region were −0.75 mm·yr−1 (p = 0.36), −0.0003 m3·m−3 yr−1 (p < 0.01), and −0.3742 cm·yr−1 (p < 0.001), respectively. Severe droughts are typically short in duration and high in intensity. Three various types of drought have low temporal correlation and spatial similarity. Furthermore, agricultural and hydrological droughts were primarily driven by actual evapotranspiration, accounting for relative contributions of 64.38% and 51.04% to these drought types, respectively. Moreover, the extent of increased actual evapotranspiration expanded to cover 49.88% of the region, exacerbating agricultural and hydrological droughts in 23.88% and 35.14% of the total study area, respectively. The study findings demonstrate that actual evapotranspiration plays a critical role in causing droughts. This study establishes a theoretical foundation to carry out drought assessment, the construction of multivariate drought indices, and water resource management in Central Asia.

Graphical Abstract">

Graphical Abstract

1. Introduction

Drought is the most widespread natural disaster, affecting the largest number of people, and it causes the greatest economic losses worldwide [1]. Drought decreases surface water and groundwater availability, reduces agricultural production, hampers economic and social activity, and damages the ecological environment [2,3]. Under the influence of climate change, extreme drought events have globally become more frequent [4], and droughts’ frequency, intensity, and duration have increased [5,6]; these trends are predicted to continue in the future [7,8]. Therefore, the occurrence and impact of drought under climatic changes is a critical research topic that has received heightened attention in the academic community [9,10]. Due to the complexity of drought and regional differences in drought assessment, a consensus regarding the definition of drought is lacking [11,12]. Drought is widely conceptualized in the academic community as a period of water scarcity relative to normal levels [7,11]. Drought can be defined as various types of moisture shortages. For example, a precipitation shortage is defined as a meteorological drought [12]; a soil moisture shortage is defined as an agricultural drought [1], a shortage of water in rivers, lakes, and terrestrial water storage can be defined as a hydrological drought [11]; and an imbalance between the water supply and human demand is defined as a socio-economic drought [12].
Central Asia is a ‘hotspot’ for global climate change, with temperature increases far exceeding averages for the Northern Hemisphere and the globe [13]. Climate change increases the atmospheric demand for moisture, leading to increases in actual evapotranspiration (ET) [12], which exacerbates drought events by increasing water consumption [13,14]. In addition, the shrinkage of the Aral Sea has led to an increase in the incidence of drought in the surrounding region, including Central Asia [15]. However, the regional characteristics of limited water resources, species poor status, and fragile ecosystems [13] make Central Asia highly vulnerable to the impacts of drought events [12,16]. For example, a severe agricultural drought occurred in Central Asia in 2021, causing a great number of deaths of crops and livestock such as cattle and sheep, which resulted in huge economic losses and severe ecological damage [17]. In addition, the lack of water resources has become a major constraint on Central Asia’s economic development and ecological security due to the combined effects of human activities and climate change. Human-induced water scarcity and naturally occurring droughts often simultaneously occur in water-scarce areas, and the occurrence of droughts will further exacerbate water scarcity and drought situations [12],making water scarcity, ecological degradation, and conflicts over water resources more pronounced in Central Asia [18,19]. Therefore, it is important to accurately assess the drought situation and its potential impacts on Central Asia.
Drought assessment has considerably progressed in recent decades, with methods related to both absolute [20,21] (indicators, e.g., precipitation or terrestrial water storage) and relative [22,23] (indices, e.g., Palmer Drought Severity Index, Standard Precipitation Index, etc.) value measurements being used to quantify and characterize drought [24]. However, there are substantial differences between drought characteristics in arid regions calculated using these two methods [12]. In recent years, most of the studies of drought in Central Asia have been conducted based on various drought indices [15,16], and fewer studies have been conducted using these indicators. However, scholars have indicated that existing drought analyses based on drought indices are characterized by substantial uncertainties [2] and a lack of consideration for actual drought conditions [13]; thus, the use of aridity indices may not be the ideal method of assessing the drought status of Central Asia [13]. Additionally, the absolute value method selects evaluation indicators based on the definition of drought, while drought indices often integrate multiple hydrological variables [3,10,25,26]. Consequently, the assessment results of the absolute value method are likely to provide a more accurate representation of the true conditions of various types of drought [27], avoiding the uncertainty associated with the categorization of the existing drought indices into a certain class of drought indices and facilitating the comparison between various types of drought and the assessment of drought propagation. Furthermore, drought is a shortage of water related to an imbalance in the water balance [2]. However, previous drought attribution studies have primarily focused on hydrothermal conditions and anthropogenic impacts [28,29,30], and fewer studies have been conducted regarding the relative contributions of water income and consumption to drought. In other words, uncertainty remains regarding whether insufficient moisture supply or excess evaporative demand is the predominant factor involved in drought events in Central Asia in the context of climate change.
This study aimed to take the following steps: (1) examine the trends and characteristics of meteorological, agricultural, and hydrological droughts in Central Asia, using precipitation, soil moisture, and terrestrial water storage as the assessment data; (2) explore the temporal correlation and spatial similarity of various types of drought in Central Asia; (3) quantitatively assess the relative contributions of precipitation, ET, and runoff to the intensity of agricultural and hydrological droughts from the perspective of the water cycle. Additionally, we investigate the impacts of ET on meteorological droughts. This study contributes to enabling a more comprehensive understanding of Central Asia’s drought development and characterization. Additionally, it provides a reference for the construction of drought indices, assessing drought, and managing water resources in Central Asia.

2. Data and Methods

2.1. Study Area

Central Asia is located in the Eurasian hinterland and spans six countries from east to west (Kazakhstan, Uzbekistan, Turkmenistan, Tajikistan, Kyrgyzstan, and the arid region of Northwestern China), covering an area of about 635.45 × 104 km2. The geographic location and prevailing westerly wind affect the regional climate of Central Asia. This region has little precipitation and very high atmospheric evaporation demand [31], and, thus, it is one of the driest regions in the world [13]. Based on the Climate Research Unit (CRU)’s precipitation and potential evapotranspiration data, hyper-arid, arid, semiarid, and subhumid regions comprise 11.34%, 45.96%, 35.77%, and 3.81% of Central Asia’s total area, respectively (Figure 1b).

2.2. Data Sources

In this study, data regarding precipitation, soil moisture, and terrestrial water storage were used to assess meteorological, agricultural, and hydrological drought, respectively. The CRU TS version 4.05 dataset used to study precipitation was developed by the Climate Research Unit (University of East Anglia) and the UK Met Office [32] (available at https://crudata.uea.ac.uk/cru/data/hrg/index.htm#current, 14 September 2023), and it is currently the most well-recognized and widely used precipitation data set related to Central Asia [13,15]. Soil moisture and runoff data were selected from the ERA5-Land dataset, which was produced by the Rebroadcast European Center for the Reanalysis of Medium-Range Weather Forecasts (available at https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-land-monthly-means?tab=overview, 14 September 2023). Compared to other widely used soil moisture data (e.g., GLDAS, GLEAM, ESA-CCI, SMAP, and SMOS), ERA5-Land soil moisture data have the highest correlation coefficients, as they contain measured soil moisture information and correlate well with arid and sparsely vegetated areas [33]. Since there are four layers of soil moisture in ERA5-Land, we weighted the soil moisture based on the thickness of the soil layer to obtain the soil moisture from 0 to 280 cm. Hydrologic drought-related data were used, including the total lake area of the reservoir and the lake surface time series dataset [34] (available at https://zenodo.org/record/6468209, 14 September 2023), as well as terrestrial water storage (TWS) anomaly data. TWS was calculated via the Gravity Recovery and Climate Experiment (GRACE) conducted by the Space Research Center (available at https://www2.csr.utexas.edu/grace/RL06_mascons.html, 14 September 2023). In this case, the total lake area refers to the total area of water in bodies larger than 1 km2 located in the study area. Due to GRACE terrestrial water storage data being missing from July 2017 to May 2018 and August 2018 to September 2018, annual terrestrial water storage was not included in the trend analysis for 2017 and 2018. Certain evapotranspiration data for Central Asia were erroneous or missing. For this reason, ET data were simulated using the Priestley–Taylor Diurnal Land Surface Temperature Range (PT-DTsR) model (see Section 2.3.1 for details) [35], and the simulated data were validated during our previous study [36]. The details of the original model profile are provided in [37]. As meteorological drought intensity is calculated based on precipitation (Table 1), it would be imprecise to analyze the contributions of water cycle elements to meteorological drought intensity because this method would introduce duplicate data; therefore, the effects of ET on meteorological drought were analyzed using wavelet analysis (see Section 2.3.4 for details) to account for the delayed correlation between ET and precipitation. The Standardized Precipitation Evapotranspiration Index (SPEI) dataset was acquired from the Global SPEI database (available at https://digital.csic.es/handle/10261/22363, 14 September 2023). The Palmer Drought Severity Index (PDSI) dataset was acquired from the Dai Global Palmer Drought Severity Index (available at https://rda.ucar.edu/datasets/ds299.0/dataaccess/, 14 September 2023). The aforementioned SPEI [38,39] and PDSI [22,40] datasets are already extensively used in drought monitoring.

2.3. Methods

2.3.1. PT-DTsR Model

The PT-DTsR model was used to calculate ET; the model was developed by Yao et al. [35] based on the Priestley–Taylor Jet Propulsion Laboratory (PT-JPL) model [41]. The PT-DTsR model replaced relative humidity and water vapor pressure differences with apparent thermal inertia obtained over time to calculate the SM constraint [35]. It simplified the calculation process, reduced data requirements, overcame the lack of ground data [42], and obtained ET data with broader coverage in arid Central Asia. In the PT-DTsR model, ET was split into four components, namely soil evaporation (Es), vegetation transpiration (T), canopy interception (Ei), and wetland surface evaporation (Ews); evaporation was considered to be the difference between evapotranspiration and transpiration (E = Es + Ei + Ews). The PT-DTsR model used the Priestley–Taylor equation to calculate the potential values for each component of ET. Potential values were corrected using environmental constraints (e.g., SM and plant temperature constraints) to obtain the actual values for each component. The details of the model are described in the study by Yao et al. [35].

2.3.2. Drought Parameters

Reliance on drought indices introduces substantial uncertainty into drought analyses [2,13]. Moreover, contradictory conclusions may be formed, even when the same drought index is applied to the same region and time period [22,23,25,43]. For these reasons, the present study used the absolute value to quantify and describe drought [2]. Furthermore, given this definition of drought [44] and the selection of thresholds [21], the present study set the multi-year mean of each month as the threshold. Since the present study was also concerned with horizontal comparisons between the three types of drought, the period used to perform the background threshold calculation was consistent. The observations of terrestrial water storage have been available since 2002, meaning that the mean of each month in the period 2002–2021 was set as the threshold. When the variables used to perform drought assessment were below the threshold, the month was classified as a drought month. The drought parameters included drought intensity (Id; deficit of the drought assessment variable in the current month), drought severity (Sd; cumulative deficit), drought duration (Dd; time below the threshold, as measured in months), maximum drought intensity (IdMAX; maximum deficit during this period), and mean drought intensity (IdMEAN; quotient of drought severity and drought duration) [1,24]. To assess agricultural drought, the parameters were calculated as follows:
S M i , y , m ′ = S M i , y , m − S M i , ∗ , m
where SMi,y,m represents the soil moisture in grid i in month m of year y, SMi,∗,m represents the multi-year average value of grid i soil moisture in month m, and S M i , y , m ′   represents the grid i soil moisture anomaly value in month m of year y. In this study, y was between 2000 and 2021, and m was between 1 and 12. For grid i, if the anomaly value was negative, it was defined as a drought month; the drought intensity of the month in question was the absolute value of the anomaly value, and the drought duration was one month. Otherwise, both parameters were designated as 0 according to the following equations:
I d , i , y , m = S M i , y , m ′ ,     S M i , y , m ′ < 0 0 ,     S M i , y , m ′ ≥ 0
d d , i , y , m = 1 ,     S M i , y , m ′ < 0 0 ,     S M i , y , m ′ ≥ 0
where Id,i,y,m represents the drought intensity of grid i in month m of year y, and dd,i,y,m represents the drought duration of grid i in month m of year y. Drought severity was the cumulative value of the drought intensity, while duration was the cumulative value of the drought duration, which were calculated using the following equations:
S d , i = ∑ 1 y ∑ 1 m I d , i , y , m
D d , i = ∑ 1 y ∑ 1 m d d , i , y , m
where Sd,i and Dd,i represent the drought severity and drought duration of grid i, respectively. To compare various drought durations, the average annual drought duration was defined as the drought frequency (FD; quotient of the drought duration and the observation year (M/12), while M was defined the total number of months, as shown in the following equation:
I d M A X = m a x I d , i , y , m
D d = S d i D d i
F d = 12 ∗ D d i M

2.3.3. The Spatial Efficiency (SPAEF) Metric

The SPAEF method is a multi-component performance evaluation metric used to assess the performance of spatial patterns in hydrological models. It consists of three equally weighted components: correlation, coefficient of variation, and histogram overlap. The advantage of using this method lay in its ability to comprehensively consider the relationship between simulation and observation, the spatial variability in the model, and the model’s histogram characteristics [45,46]. By adopting this multi-component approach, SPAEF could provide a more comprehensive comparison of the spatial patterns of different models. Furthermore, the three components of SPAEF were independent and could complement each other, providing more meaningful evaluation results. In this study, SPAEF was used to assess the spatial similarity of various types of drought in terms of trend and characterization.

2.3.4. Lindeman–Merenda–Gold Method

The Lindeman–Merenda–Gold method is an average ordering method that quantifies the contribution of each explanatory variable to the dependent variable by decomposing R2 into non-negative components [47]. This method has been widely used in previous studies and is based on the unweighted average of the sequential contributions of each variable to the ranking of all available regressors, avoiding regressor order effects [48,49]. It has previously been integrated into the R software version 4.3.1 with the package “relaimpo” [47]. In this study, the Lindeman–Merenda–Gold method was used to quantitatively assess the relative contributions of precipitation, ET, and runoff to agricultural and hydrological drought intensity during a drought on a monthly scale, working on a raster by raster basis. It should be noted that due to the inconsistent spatial resolution of different drought assessment data, while determining the actual ET contribution to drought intensity, the spatial resolution of ET, precipitation, and terrestrial water storage, it was resampled to 0.5° using a triple-convolution method.

2.3.5. Wavelet Transform

Wavelet analysis was used to identify and characterize major local variations in time series by decomposing the data into time and frequency space. Compared to Fourier analysis, continuous wavelet analysis has the advantages of self-adjusting temporal resolution, flexible mother wavelet, and applicability to non-stationary time series [15]. In this study, cross-wavelet and wavelet coherence in wavelet analysis were used to perform delayed correlation analysis of ET and precipitation. Cross-wavelet transform was developed via traditional wavelet analysis, and it can reflect the same high-energy region and phase relationships between two sequences, thus revealing their interactions in different time–frequency domains. Wavelet coherence relies on searching the frequency bands, as well as the co-existing time intervals between the two sequences, to identify their possible relationships and characterize the degree of consistency of the cross-wavelet transform in time and frequency space. Compared to the cross-wavelet transform, cross-wavelet coherence better compensates for the shortcomings of cross-wavelet power spectra identified in the low-energy region by exploring the significant correlation between two time series in the low-energy region [15,50]. The overall idea and workflow used in this study can be seen in Figure 2.

3. Results

3.1. Drought Assessment in Central Asia

Precipitation rates (P) in Central Asia were highly consistent between 2000 and 2021. These values slightly and non-significantly decreased (Figure 3a) at a rate of −0.75 mm·yr−1 (p = 0.360). Over the past 22 years, there was a high frequency of negative precipitation anomalies (14/22) with narrow fluctuations (−13.91 ± 11.02 mm·yr−1), as well as a low frequency of positive precipitation anomalies (8/22) with wide fluctuations (24.35 ± 17.88 mm·yr−1). The positive precipitation anomalies increased the mean annual precipitation. The proportions of the areas with positive and negative precipitation trends in Central Asia were 46.01% and 53.99%, respectively (Figure 3b). The areas with increasing precipitation were concentrated around the Heihe, Baiyang, Manas, and Irtysh Rivers and the southern area of the Tarim River Basin. The areas with decreasing precipitation were concentrated in the northern Tianshan Mountains, the southern Kunlun Mountains, and the coastal plains of the Caspian Sea.
Soil moisture (SM) in Central Asia decreased between 2000 and 2020 (−0.0003 m3·m−3 yr−1; p < 0.05) (Figure 3c). The SM decreased from 2000 to 2012, increased from 2012 to 2016, and decreased from 2016 to 2020. The rates of decrease before 2012 and after 2016 were −0.0011 m3·m−3yr−1 (p < 0.001) and −0.0015 m3·m−3yr−1 (p < 0.001), respectively. Soil drying (57.97%) and wetting (42.03%) were mixed in the study area (Figure 3d). The wetting areas were mainly distributed in the northern foothills of the Kunlun Mountains, the Qilian Mountains, the Kazakh Hills, and the area around the Aral Sea. The drying areas were mainly distributed in the Ural River Basin and nearby eastern areas, Turkmenistan, and the area around the Tianshan Mountains.
The lake areas (LA) and terrestrial water storage (TWS) in Central Asia both exhibited a significant downward trend, namely −12.3611 km2·yr−1 (p < 0.001) and −0.3742 cm·yr−1 (p < 0.001), respectively (Figure 3e). In Central Asia, the areas with increasing and decreasing TWS accounted for 25.84% and 74.16% of the total area, respectively. The most pronounced downward trends in TWS were observed in the Tianshan Mountains and the Caspian Sea coastal plain. The upward trends in TWS occurred in the Uzboy River Basin, the Kunlun Mountains, and the Ertysi River Basin.
These results reveal that agricultural and hydrological drought substantially intensified, while meteorological drought slightly intensified, in Central Asia during the study period. Moreover, the three drought types differed in terms of their spatial distributions. All three types only displayed a significant (p < 0.05) downward trend in the Ural River Basin.
The analysis of the drought characteristics of Central Asia based on drought parameters (Figure 4) revealed that the spatial distributions of the severity, maximum intensity, and average intensity were relatively similar for each drought type, but they differed between the various drought types. The results of the regions with severe meteorological, agricultural, and hydrological droughts were more consistent with the spatial distributions of the regions with decreasing precipitation, soil moisture, and terrestrial water storage, respectively (Figure 4), and only the western–central part of the Irtysh River Basin experienced more severe meteorological droughts occurring in conjunction with a slight increase in precipitation. This issue may be related to its drought characteristics of long duration and low intensity. In the severe drought region of the Irtysh River basin, more than seven months of the year were regarded as a period of low-intensity drought, the accumulated precipitation deficit was small, and the surplus of precipitation in other months contributed to a slight upward trend in annual-scale precipitation. In other regions with severe droughts, these events were characterized by short durations and high intensity. This outcome may be related to the frequent occurrence of Increased ET in Central Asia during droughts [51], which rapidly depleted water resources, increased drought intensity, and reduced drought duration.

3.2. The Spatiotemporal Correlation of Various Types of Drought

The correlations of the temporal variations in the four indicators were analyzed to assess drought and conduct preliminary exploration of the temporal correlations between actual evapotranspiration and the various types of drought (Figure 5). ET was significantly positively correlated with precipitation, TWS, LA, and SM [0.67 (p < 0.01), 0.47 (p < 0.1), 0.66 (p < 0.01), and 0.71 (p < 0.001), respectively]. However, the correlations between precipitation and SM, TWS, and LA were not significant [0.42 (p < 0.1), 0.20 (p = 0.459), and 0.23 (p = 0.402), respectively]. This phenomenon, when combined with the low contributions of precipitation to the changes in SM and TWS (Figure A1) and the differences in spatiotemporal variation between precipitation, SM, and TWS (Figure 3), indicated weak correlations between meteorological, agricultural, and hydrological drought, as well as the relatively weak influence of precipitation on agricultural and hydrological drought.
In this article, three combinations are designed based on the results shown in Figure 6 to evaluate the spatial similarity of various types of drought using spatial efficiency measures. The findings indicate a relatively low spatial similarity between the trends and characteristics of the three types of drought. In particular, the spatial similarity between agricultural drought and hydrological drought is the lowest (Figure 6a). Their spatial efficiencies in terms of trend, severity, average intensity, maximum intensity, and duration of drought are negative, being −0.74, −0.07, −0.03, −0.01, and −0.29, respectively. On the other hand, the spatial similarity between agricultural drought and meteorological drought is the highest of all similarity values (Figure 6c). Their spatial efficiencies in terms of trend, severity, average intensity, maximum intensity, and duration of drought are −0.21, 0.34, 0.46, 0.53, and −0.19, respectively.

3.3. Contributions of Precipitation, ET, and Runoff to Drought Intensity

3.3.1. Contributions of Precipitation, ET, and Runoff to Agricultural Drought Intensity

The relative contribution of ET to the agricultural drought intensity was highest in Central Asia (mean = 0.6138) (Figure 7b). The probability density gradually increased with the relative contribution and occupied a dominant position in most regions in the study area. The relative contributions of precipitation and runoff to agricultural drought intensity were small (mean = 0.1901) (Figure 7a). Their probability densities gradually decreased with the increase in relative contribution, and the latter aspect was only >0.5 in certain parts of the Irtysh River Basin. In the study area, runoff made a slightly greater relative contribution than precipitation to the agricultural drought intensity (mean = 0.1961) (Figure 7c). Its probability density decreased with the increase in relative contribution, and the latter aspect was only >0.5 in certain parts of the Tarim and Heihe river basins. The relative contributions of precipitation and runoff to agricultural drought were low, and ET dominated agricultural drought. Moreover, rainfed agriculture in Central Asia is mainly concentrated in the western part of the Irtysh River Basin and the northern part of the Ural River Basin. In the northern part of the Ural River Basin, agricultural droughts are severe and dominated by actual evapotranspiration. In contrast, there are no significant agricultural droughts in the Irtysh River Basin, and the intensity of droughts in different regions is determined based on different factors.

3.3.2. Contributions of Precipitation, ET, and Runoff to Hydrological Drought Intensity

ET made the highest relative contribution to hydrological drought intensity (mean = 0.5104) (Figure 7d). Its probability density did not change in tandem with the relative contribution, and it predominated in most regions in the study area. The contributions of precipitation and runoff to the hydrological drought intensity were relatively low (0.2263 and 0.2633, respectively) (Figure 7d,f). The probability density decreased with the increase in the relative contribution. For precipitation, the latter aspect was relatively high in parts of the Ural River Basin, the Uzboy River Basin, and the Amu Darya River Basin, whereas the relative contribution of runoff was relatively high in the Ural and Syr Darya river basins.

3.3.3. Contributions of Precipitation, ET, and Runoff to Meteorological Drought

The results of delayed correlation analysis between actual ET and precipitation on a monthly scale using wavelet analysis revealed that ET and precipitation had good resonance and correlation, and the period of precipitation lagged behind that of ET by 1/16–1/8 of a period (Figure 8). Therefore, the occurrence of ET likely recharged the atmospheric moisture and promoted the occurrence of precipitation, which, in turn, indirectly affected meteorological drought.

4. Discussion

4.1. Comparison of Relative Contributions Based on Index and Indicator Methods

The study employed a consistent methodology to examine the impacts of water balance variables on drought indices, namely the inclusion of SPEI and PDSI. The distinction lies in how drought months are determined; to identify the drought periods, SPEI <− 0.5 and PDSI <− 0.5 were adopted as threshold values, with the magnitude of the indices reflecting drought intensity. The findings indicate the significant role played by actual evapotranspiration in drought intensity (Figure 9). Moreover, precipitation, ET, and runoff made relative contributions of 0.4308, 0.3673, and 0.2019, respectively, to the intensity of SPEI drought. Regions in the western and northern parts of the study area exhibited high contributions from precipitation, whereas areas near the southeastern mountains predominantly showed high contributions from evapotranspiration. In terms of PDSI drought intensity, precipitation, evapotranspiration, and runoff made respective relative contributions of 0.1726, 0.6548, and 0.1726. Actual evapotranspiration exerted a dominant influence across the entire study area, occupying an intermediate position, with its spatial distribution of relative contribution to PDSI drought intensity spanning the spectrum between hydrological and agricultural droughts.
In this study, four approaches (two indicators and two indices) were utilized to calculate the contribution of water budget variables to drought intensity. The contributions and spatial distributions of hydrological elements varied between the various approaches used. We speculate that this issue might be associated with the drought assessment objects used in different approaches. In this study, agricultural drought was calculated based on soil moisture anomalies ranging from 0 to 280 cm, considering both the surplus and deficit of soil moisture. The calculation of hydrological drought utilized abnormal data of terrestrial water storage, reflecting the overall surplus or deficit of terrestrial water storage by observing gravity changes between the Earth’s surface and the atmosphere via satellites [52]. The calculation of SPEI includes precipitation and potential evapotranspiration [3], primarily focusing on above-ground climatic conditions. It has been widely regarded as a meteorological drought index [53]. On the other hand, PDSI is an indirectly calculated method based on the balance of water storage [13]. It is worth mentioning that the PDSI calculation includes recharge and involves the variation in groundwater storage. Therefore, the drought assessment object considered by the PDSI is more in depth than the soil moisture anomaly [25]. In conclusion, this study utilized four drought calculation methods, namely the SPEI, agricultural drought, the PDSI, and hydrological drought. The objects evaluated via these methods gradually extend from the surface to the underground, and the scope considered via the latter three methods gradually expands. Therefore, this study examined the contribution of water balance variables to the assessment of drought intensity through the application of four different methods. These methods can also be considered to evaluate the impacts of hydrological variables on spatial water deficits. Therefore, in this study, the contribution of water budget variables assessed via four different methods to drought intensity can also be regarded as the contribution of water budget variables to water deficits in different spatial areas. Based on the data and methods employed, actual evaporation plays a significant role in droughts across different spatial areas in Central Asia. It not only affects soil moisture in the shallow layers, but also has a further impact on groundwater and the overall terrestrial water storage, reaching its peak contribution in the calculation made using the Palmer Drought Severity Index (PDSI).
It should be noted that although this study consulted previous studies in the selection of data to reduce the impacts of uncertainties in input data on the results, inherent errors in the data themselves and errors arising from different data sources can still have impacts on the results. Additionally, the selection of the reference period has a significant influence on the assessment of drought events. Due to the limitations of available terrestrial water storage data, this study defined the reference period for meteorological, agricultural, and hydrological droughts as 2003–2020, which may deviate from the actual historical reference period in Central Asia.

4.2. Driving Forces of ET Variation in Arid Region

ET transports water from soil, plants, and water body surfaces to the atmosphere and represents the trade-off between surface water and thermal conditions. Evapotranspiration includes water evaporation from soil, plant canopies, and water bodies and transpiration from plant leaves as water moves from the soil into the roots and through the plant [54]. These processes deplete terrestrial water and affect drought. For example, ET exacerbates summer drought in Europe [14]. Furthermore, drought intensity increases in tandem with ET [13]. Additionally, increases in ET prolong the tropical dry season [55]. The present study quantified the relative contributions of ET to agricultural and hydrological drought intensity and demonstrated the dominant role played by ET in drought events in arid Central Asia. The contribution of ET to drought intensity was greater than those of changes in SM and TWS (Figure A1 and Figure 7). The high ET demand during drought periods increased water consumption and drought intensity and aggravated existing drought. In severe drought areas, the duration was short, and the intensity was high; therefore, ET may affect these characteristics.
ET replenishes atmospheric water, promotes precipitation, and indirectly affects meteorological drought (Figure 8). As it also influences the contribution of precipitation to agricultural and hydrological drought intensity, it may cause the contribution of precipitation to be overestimated. The trend in the distributions of ET and its components markedly differed from that of precipitation (Figure 3b and Figure A2). Therefore, precipitation change is a weak driver of ET change. In Central Asia, ET exacerbates drought by consuming water. It also affects drought characteristics and water income by recharging atmospheric water. Hence, ET is highly significant in drought events. The existing drought indices exhibit certain shortcomings. For this reason, scholars have attempted to establish multivariate drought indices by combining multiple drought-related variables based on the copula concept and information entropy [56,57]. These innovations have improved the characterization of the entire drought process [58]. ET has a vital impact on meteorological, agricultural, and hydrological drought. Consequently, water expenditure (ET) can be used to combine different drought-related variables, including precipitation, runoff, water vapor, glacial snow cover, SM, and TWS. The establishment of a multivariate drought index may substantially enhance the ability of the drought index to monitor and characterize droughts.

4.3. Effects of Precipitation and Runoff (Water Income) on Drought

The relative contributions of precipitation to agricultural and hydrological intensity were low in Central Asia (Figure 7). These findings contradict those previously reported [9,59]. Drought is a form of water shortage caused by an imbalance in the water balance. The impact of precipitation on drought is reflected in its ability to replenish terrestrial water. Precipitation does not meet atmospheric evaporation demand during the growing season in arid regions (Figure 3a). Furthermore, precipitation is often lower than ET in such regions, and after precipitation makes landfall, it is rapidly consumed by ET and cannot effectively replenish terrestrial water. The water resources in Central Asia have a diverse composition, and precipitation is not the sole water source in this region. The Tianshan Mountains are known as “Central Asia’s Water Tower”, as they have well-developed glaciers and widespread snow cover. Hence, they are the primary water source in Central Asia [60]. Runoff is replenished by melting ice and snow in the high mountains, precipitation in the middle mountain forests, and bedrock fissure water in the low mountains. Runoff then flows to plains, oases, deserts, and other water-consuming areas; infiltrates and transforms into surface water and groundwater; and becomes the main regional water supply [61]. In response to global warming, glaciers have receded, snow cover has melted, permafrost has thawed, and the water resource structure has changed in Central Asia [62]. In the Tianshan region, the decline in TWS has surpassed the precipitation rate (Figure 3), likely due to glacial melting. This change has amplified the impacts of runoff on changes in SM and TWS (Figure A1) and weakened the contribution of precipitation to the drought intensity in the study area [63]. As the water resources of the study area have a unique composition, the drought events in this region may not be adequately defined and assessed when using precipitation as the sole water source. Future regional drought assessments and drought index construction must consider multiple water sources.

4.4. Interactions between Evapotranspiration, Precipitation, and Runoff

A change in any element of the water cycle will affect the other elements [64]. However, the present study considered the three examined water cycle elements to be independent, and their interactions were not evaluated. Figure 7 shows that ET recharged the atmospheric water and formed precipitation, which implies that the contribution of precipitation to drought may have been overestimated. Terrestrial water recharge through runoff and precipitation may have also caused fluctuations in ET. The contribution of the latter aspect to drought intensity may also have been overestimated. Central Asia has complex topography, diverse water resources, complex flow–confluence mechanisms, and unique water cycle processes. The hydrological processes and the interactions between the elements therein require further clarification [60]. Future research should include systematic analyses of the water resource components, their changes, and the water cycle mechanisms in response to climate change. Quantitative analyses of these interactions between the water cycle elements are also necessary. The information derived from these studies may be crucial to future regional water resource planning and management in Central Asia.

5. Conclusions

This study examined the trends and characteristics of meteorological, agricultural, and hydrological droughts, as well as the spatiotemporal correlations between them. Indicators such as precipitation, soil moisture, and land water storage were used in the analysis. Additionally, this study investigated the impacts of precipitation, evapotranspiration (ET), and runoff on drought intensity. The primary findings were summarized as follows:
(1)
Drought has intensified in Central Asia. The trends observed in precipitation, SM, TWS, and total LA in this region were −0.75 mm·yr−1 (p = 0.36), −0.0003 m3·m−3 yr−1 (p < 0.05), −0.3742 cm·yr−1 (p < 0.001), and −12.3611 km2·yr−1 (p < 0.001), respectively. Severe droughts are typically characterized by short duration and high intensity.
(2)
Various types of drought display variations in both temporal correlation and spatial similarity. Only agricultural drought and hydrological drought demonstrate significant temporal correlation, while agricultural drought and meteorological drought show high spatial similarity.
(3)
Actual evapotranspiration played a dominant role in agricultural and hydrological drought in Central Asia, having relative contributions of 0.6438 and 0.5104 to agricultural and hydrological drought intensity, respectively.
This study provided comprehensive insights into the effects of evapotranspiration on drought in the region. The findings of this study offer a theoretical foundation for the assessment of drought, construction of drought indices, and planning and managing of water resources in Central Asia.

Author Contributions

All authors made significant contributions to this study. Conceptualization, Z.Z.; methodology, X.H.; data curation, Z.Z.; writing—original draft preparation, Z.Z.; formal analysis, X.F. and J.Z.; writing—review and editing, Z.Z., S.Z. and X.L.; project administration, X.H.; funding acquisition, X.H. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the Independent Deployment Project, the Xinjiang Institute of Ecology and Geography, the Chinese Academy of Sciences (E050010801), and the Science Fund for Distinguished Young Scholars of Xinjiang Autonomous Region (2022D01E02).

Data Availability Statement

Meteorological drought-related data, including precipitation and potential evapotranspiration data, were generated by the Climate Research Unit (University of East Anglia, UK) and the UK Met Office (https://crudata.uea.ac.uk/cru/, accessed on 14 September 2023). Soil moisture and runoff were based on the ERA5-Land dataset generated by the European Centre for Medium-Range Weather Forecasts (ECMWF) (https://cds.climate.copernicus.eu/, accessed on 14 September 2023). Hydrological drought-related data included the total reservoir lake area and lake surface time-series datasets (https://zenodo.org/record/, accessed on 14 September 2023), as well as the terrestrial water storage anomaly data provided by the Center for Space Research (https://www2.csr.utexas.edu/grace/, accessed on 14 September 2023). Evapotranspiration data were estimated using the PT-DTsR model [35], and the applicability of the model to Central Asia was validated in our earlier research. Catchment boundaries, river networks, and lakes were downloaded from the HydroSHEDS database (https://www.hydrosheds.org/, accessed on 14 September 2023).

Acknowledgments

We want to thank the editor and anonymous reviewers, whose comments have helped to substantially improve the quality of the manuscript.

Conflicts of Interest

The authors declare that they have no known competing financial interest or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A

Table A1. Statistical properties of used data, where CV, CI, STD, MIN, and MAX, respectively, represent the coefficient of variation (CV), confidence interval (CI), standard deviation (STD), minimum value (MIN), and maximum value (MAX).
Table A1. Statistical properties of used data, where CV, CI, STD, MIN, and MAX, respectively, represent the coefficient of variation (CV), confidence interval (CI), standard deviation (STD), minimum value (MIN), and maximum value (MAX).
PREPETGRACELAKESMROET
CV10.737782.644813−145.7215.8082062.03612314.809546.397844
CI[209.4703,
230.4125]		[1114.735, 1141.189][−3.03481, −0.48451][1094.199, 1164.093][0.201479, 0.205249][1.658742, 1.898547][145.2266, 154.192]
Mean219.94141127.962−1.759661129.1460.2033641.778645149.7093
Std23.6168229.832492.56420265.583120.0041410.2634099.578169
Min176.4051048.794−7.20911024.5140.196741.29871133.8691
Max277.5171190.7272.7584921219.3090.2099562.270297171.5517
Median211.97211128.332−2.591141130.0430.2036091.752532149.8471
Kurtosis3.3439214.0907352.5441941.50131.5798442.3333862.788438
Skewness0.655879−0.4297−0.06491−0.10801−0.005830.2484070.351814

Appendix A.1. Contributions of Precipitation, Evapotranspiration, and Runoff to Soil Moisture and Terrestrial Water Storage

The relative contributions of precipitation to soil moisture and terrestrial water storage were the lowest at 0.2160 and 0.1914, respectively. The relative contributions of evapotranspiration to soil moisture and terrestrial water storage were 0.3395 and 0.4236, respectively, and it had the highest contribution to terrestrial water storage. The relative contributions of runoff to soil moisture and terrestrial water storage were 0.4444 and 0.3849, respectively, and it had the highest contribution to terrestrial water storage. The spatial distribution of the relative contributions significantly varied. The changes in soil moisture in most of the study area were dominated by a combination of evapotranspiration and runoff, while the changes in terrestrial water storage were dominated by each of the three factors in one area.
Remotesensing 15 04557 g0a1
Figure A1. Contributions of precipitation (a,b), evapotranspiration (c,d), and runoff (e,f) to soil moisture (a,c,e) and terrestrial water storage (b,d,f).
Figure A1. Contributions of precipitation (a,b), evapotranspiration (c,d), and runoff (e,f) to soil moisture (a,c,e) and terrestrial water storage (b,d,f).
Remotesensing 15 04557 g0a1

Appendix A.2. Spatiotemporal Variations in ET

The three components of ET were vegetation transpiration, canopy interception, and evaporation. The latter component comprised soil and water body evaporation. Vegetation transpiration was generally elevated in Central Asia (Figure 4b), and its positive and negative trends accounted for 68.14% and 31.86% of the total land area, respectively. All regions with severe agricultural drought showed a decreasing trend in vegetation transpiration (Figure 3b). Vegetation activity was limited due to SM. Evaporation decreased overall (Figure 4c), and its positive and negative trends accounted for 36.67% and 63.33% of the total land area, respectively. Soil evaporation showed an increasing trend in the Tarim and Manas river basins, and SM in these areas showed an increasing trend in response to oasis irrigation agriculture (Figure 3d). Water body evaporation significantly decreased. In the study area, ET increased in the eastern areas and decreased in the western areas, and the positive and negative trends accounted for 49.88% and 50.12% of the total land area, respectively. The oasis in the arid region of Northwestern China had the most significant increase in ET. The Ural River Basin and its surrounding areas displayed the most significant decrease in ET, except for water body evaporation. The increase in ET intensified agricultural and hydrological drought in 23.88% and 35.14% of the total study area, respectively. The annual ET and its components had an inflection point in 2011 (Figure 3d). All three ET components declined before 2011. After that year, however, transpiration showed an increasing trend (2.46 mm·yr−1), whereas evaporation continued to decrease (−1.03 mm·yr−1). Transpiration dominated the ET trend (1.45 mm·yr−1).
Remotesensing 15 04557 g0a2
Figure A2. Variations in evapotranspiration (ET) and its components in the period 2001–2019, spatial distributions of (a) ET, (b) vegetation transpiration (ETc), (c) evaporation (ETs) trends, and (d) the time series of ET and its components. ETi represents canopy interception evaporation.
Figure A2. Variations in evapotranspiration (ET) and its components in the period 2001–2019, spatial distributions of (a) ET, (b) vegetation transpiration (ETc), (c) evaporation (ETs) trends, and (d) the time series of ET and its components. ETi represents canopy interception evaporation.
Remotesensing 15 04557 g0a2

References

  1. Mishra, A.K.; Singh, V.P. A review of drought concepts. J. Hydrol. 2010, 391, 202–216. [Google Scholar] [CrossRef]
  2. Trenberth, K.E.; Dai, A.; van der Schrier, G.; Jones, P.D.; Barichivich, J.; Briffa, K.R.; Sheffield, J. Global warming and changes in drought. Nat. Clim. Chang. 2013, 4, 17–22. [Google Scholar] [CrossRef]
  3. Vicente-Serrano, S.M.; Beguería, S.; López-Moreno, J.I. A Multiscalar Drought Index Sensitive to Global Warming: The Standardized Precipitation Evapotranspiration Index. J. Clim. 2010, 23, 1696–1718. [Google Scholar] [CrossRef]
  4. Trenberth, K.E.; Fasullo, J.T.; Shepherd, T.G. Attribution of climate extreme events. Nat. Clim. Chang. 2015, 5, 725–730. [Google Scholar] [CrossRef]
  5. Spinoni, J.; Naumann, G.; Carrao, H.; Barbosa, P.; Vogt, J. World drought frequency, duration, and severity for 1951-2010. Int. J. Climatol. 2014, 34, 2792–2804. [Google Scholar] [CrossRef]
  6. Piao, S.; Zhang, X.; Chen, A.; Liu, Q.; Lian, X.; Wang, X.; Peng, S.; Wu, X. The impacts of climate extremes on the terrestrial carbon cycle: A review. Sci. China Earth Sci. 2019, 62, 1551–1563. [Google Scholar] [CrossRef]
  7. Ault, T.R. On the essentials of drought in a changing climate. Science 2020, 368, 256–260. [Google Scholar] [CrossRef] [PubMed]
  8. Li, H.; Li, Z.; Chen, Y.; Xiang, Y.; Liu, Y.; Kayumba, P.M.; Li, X. Drylands face potential threat of robust drought in the CMIP6 SSPs scenarios. Environ. Res. Lett. 2021, 16, 114004. [Google Scholar] [CrossRef]
  9. Van Loon, A.F.; Laaha, G. Hydrological drought severity explained by climate and catchment characteristics. J. Hydrol. 2015, 526, 3–14. [Google Scholar] [CrossRef]
  10. Faiz, M.A.; Zhang, Y.; Tian, X.; Tian, J.; Zhang, X.; Ma, N.; Aryal, S. Drought index revisited to assess its response to vegetation in different agro-climatic zones. J. Hydrol. 2022, 614, 128543. [Google Scholar] [CrossRef]
  11. Van Loon, A.F. Hydrological drought explained. WIREs Water 2015, 2, 359–392. [Google Scholar] [CrossRef]
  12. Mukherjee, S.; Mishra, A.; Trenberth, K.E. Climate Change and Drought: A Perspective on Drought Indices. Curr. Clim. Chang. Rep. 2018, 4, 145–163. [Google Scholar] [CrossRef]
  13. Li, Z.; Chen, Y.; Fang, G.; Li, Y. Multivariate assessment and attribution of droughts in Central Asia. Sci. Rep. 2017, 7, 1316. [Google Scholar] [CrossRef] [PubMed]
  14. Teuling, A.J.; Van Loon, A.F.; Seneviratne, S.I.; Lehner, I.; Aubinet, M.; Heinesch, B.; Bernhofer, C.; Grünwald, T.; Prasse, H.; Spank, U. Evapotranspiration amplifies European summer drought. Geophys. Res. Lett. 2013, 40, 2071–2075. [Google Scholar] [CrossRef]
  15. Guo, H.; Bao, A.; Liu, T.; Jiapaer, G.; Ndayisaba, F.; Jiang, L.; Kurban, A.; De Maeyer, P. Spatial and temporal characteristics of droughts in Central Asia during 1966–2015. Sci. Total Environ. 2018, 624, 1523–1538. [Google Scholar] [CrossRef] [PubMed]
  16. Deng, H.; Yin, Y.; Han, X. Vulnerability of vegetation activities to drought in Central Asia. Environ. Res. Lett. 2020, 15, 084005. [Google Scholar] [CrossRef]
  17. Jiang, J.; Zhou, T. Agricultural drought over water-scarce Central Asia aggravated by internal climate variability. Nat. Geosci. 2023, 16, 154–161. [Google Scholar] [CrossRef]
  18. Li, Z.; Chen, Y.; Li, W.; Deng, H.; Fang, G. Potential impacts of climate change on vegetation dynamics in Central Asia. J. Geophys. Res. Atmos. 2015, 120, 12345–12356. [Google Scholar] [CrossRef]
  19. Chen, Y.; Li, W.; Deng, H.; Fang, G.; Li, Z. Changes in Central Asia’s Water Tower: Past, Present and Future. Sci. Rep. 2016, 6, 35458. [Google Scholar] [CrossRef]
  20. Arnell, N.W. Climate change and global water resources: SRES emissions and socio-economic scenarios. Glob. Environ. Chang. 2004, 14, 31–52. [Google Scholar] [CrossRef]
  21. Thomas, A.C.; Reager, J.T.; Famiglietti, J.S.; Rodell, M. A GRACE-based water storage deficit approach for hydrological drought characterization. Geophys. Res. Lett. 2014, 41, 1537–1545. [Google Scholar] [CrossRef]
  22. Dai, A. Increasing drought under global warming in observations and models. Nat. Clim. Chang. 2012, 3, 52–58. [Google Scholar] [CrossRef]
  23. Sheffield, J.; Wood, E.F.; Roderick, M.L. Little change in global drought over the past 60 years. Nature 2012, 491, 435–438. [Google Scholar] [CrossRef] [PubMed]
  24. Aksoy, H.; Cetin, M.; Eris, E.; Burgan, H.I.; Cavus, Y.; Yildirim, I.; Sivapalan, M. Critical drought intensity-duration-frequency curves based on total probability theorem-coupled frequency analysis. Hydrol. Sci. J. 2021, 66, 1337–1358. [Google Scholar] [CrossRef]
  25. Dai, A. Characteristics and trends in various forms of the Palmer Drought Severity Index during 1900–2008. J. Geophys. Res. 2011, 116, D12115. [Google Scholar] [CrossRef]
  26. Won, J.; Kim, S. Ecological Drought Condition Index to Monitor Vegetation Response to Meteorological Drought in Korean Peninsula. Remote Sens. 2023, 15, 337. [Google Scholar] [CrossRef]
  27. Iacobellis, S.F.; Steinemann, A.; Cayan, D.R. Developing and Evaluating Drought Indicators for Decision-Making. J. Hydrometeorol. 2015, 16, 1793–1803. [Google Scholar] [CrossRef]
  28. Vicente-Serrano, S.M.; Van der Schrier, G.; Beguería, S.; Azorin-Molina, C.; Lopez-Moreno, J.-I. Contribution of precipitation and reference evapotranspiration to drought indices under different climates. J. Hydrol. 2015, 526, 42–54. [Google Scholar] [CrossRef]
  29. Yang, M.; Mou, Y.; Meng, Y.; Liu, S.; Peng, C.; Zhou, X. Modeling the effects of precipitation and temperature patterns on agricultural drought in China from 1949 to 2015. Sci. Total Environ. 2020, 711, 135139. [Google Scholar] [CrossRef]
  30. Samaniego, L.; Thober, S.; Kumar, R.; Wanders, N.; Rakovec, O.; Pan, M.; Zink, M.; Sheffield, J.; Wood, E.F.; Marx, A. Anthropogenic warming exacerbates European soil moisture droughts. Nat. Clim. Chang. 2018, 8, 421–426. [Google Scholar] [CrossRef]
  31. Chen, F.-H.; Chen, J.-H.; Holmes, J.; Boomer, I.; Austin, P.; Gates, J.B.; Wang, N.-L.; Brooks, S.J.; Zhang, J.-W. Moisture changes over the last millennium in arid central Asia: A review, synthesis and comparison with monsoon region. Quat. Sci. Rev. 2010, 29, 1055–1068. [Google Scholar] [CrossRef]
  32. Harris, I.; Osborn, T.J.; Jones, P.; Lister, D. Version 4 of the CRU TS monthly high-resolution gridded multivariate climate dataset. Sci. Data 2020, 7, 109. [Google Scholar] [CrossRef] [PubMed]
  33. Beck, H.E.; Pan, M.; Miralles, D.G.; Reichle, R.H.; Dorigo, W.A.; Hahn, S.; Sheffield, J.; Karthikeyan, L.; Balsamo, G.; Parinussa, R.M.; et al. Evaluation of 18 satellite- and model-based soil moisture products using in situ measurements from 826 sensors. Hydrol. Earth Syst. Sci. 2021, 25, 17–40. [Google Scholar] [CrossRef]
  34. Khandelwal, A.; Karpatne, A.; Ravirathinam, P.; Ghosh, R.; Wei, Z.; Dugan, H.A.; Hanson, P.C.; Kumar, V. ReaLSAT, a global dataset of reservoir and lake surface area variations. Sci. Data 2022, 9, 1–12. [Google Scholar] [CrossRef]
  35. Yao, Y.; Liang, S.; Cheng, J.; Liu, S.; Fisher, J.B.; Zhang, X.; Jia, K.; Zhao, X.; Qin, Q.; Zhao, B.; et al. MODIS-driven estimation of terrestrial latent heat flux in China based on a modified Priestley–Taylor algorithm. Agric. For. Meteorol. 2013, 171–172, 187–202. [Google Scholar] [CrossRef]
  36. Hao, X.; Fan, X.; Zhao, Z.; Zhang, J. Spatiotemporal Patterns of Evapotranspiration in Central Asia from 2000 to 2020. Remote Sens. 2023, 15, 1150. [Google Scholar] [CrossRef]
  37. Yao, Y.; Liang, S.; Li, X.; Hong, Y.; Fisher, J.B.; Zhang, N.; Chen, J.; Cheng, J.; Zhao, S.; Zhang, X.; et al. Bayesian multimodel estimation of global terrestrial latent heat flux from eddy covariance, meteorological, and satellite observations. J. Geophys. Res. Atmos. 2014, 119, 4521–4545. [Google Scholar] [CrossRef]
  38. Vicente-Serrano, S.M.; Beguería, S.; López-Moreno, J.I.; Angulo, M.; El Kenawy, A. A New Global 0.5° Gridded Dataset (1901–2006) of a Multiscalar Drought Index: Comparison with Current Drought Index Datasets Based on the Palmer Drought Severity Index. J. Hydrometeorol. 2010, 11, 1033–1043. [Google Scholar] [CrossRef]
  39. Zhang, Z.; Ju, W.; Zhou, Y.; Li, X. Revisiting the cumulative effects of drought on global gross primary productivity based on new long-term series data (1982–2018). Glob. Chang. Biol. 2022, 28, 3620–3635. [Google Scholar] [CrossRef]
  40. Hulley, G.C.; Dousset, B.; Kahn, B.H. Rising Trends in Heatwave Metrics Across Southern California. Earths Future 2020, 8, e2020EF001480. [Google Scholar] [CrossRef]
  41. Fisher, J.B.; Tu, K.P.; Baldocchi, D.D. Global estimates of the land–atmosphere water flux based on monthly AVHRR and ISLSCP-II data, validated at 16 FLUXNET sites. Remote Sens. Environ. 2008, 112, 901–919. [Google Scholar] [CrossRef]
  42. Cao, M.; Wang, W.; Xing, W.; Wei, J.; Chen, X.; Li, J.; Shao, Q. Multiple sources of uncertainties in satellite retrieval of terrestrial actual evapotranspiration. J. Hydrol. 2021, 601, 126642. [Google Scholar] [CrossRef]
  43. van der Schrier, G.; Jones, P.D.; Briffa, K.R. The sensitivity of the PDSI to the Thornthwaite and Penman-Monteith parameterizations for potential evapotranspiration. J. Geophys. Res. 2011, 116. [Google Scholar] [CrossRef]
  44. Lloyd-Hughes, B. The impracticality of a universal drought definition. Theor. Appl. Climatol. 2013, 117, 607–611. [Google Scholar] [CrossRef]
  45. Demirel, M.C.; Mai, J.; Mendiguren, G.; Koch, J.; Samaniego, L.; Stisen, S. Combining satellite data and appropriate objective functions for improved spatial pattern performance of a distributed hydrologic model. Hydrol. Earth Syst. Sci. 2018, 22, 1299–1315. [Google Scholar] [CrossRef]
  46. Koch, J.; Demirel, M.C.; Stisen, S. The SPAtial EFficiency metric (SPAEF): Multiple-component evaluation of spatial patterns for optimization of hydrological models. Geosci. Model Dev. 2018, 11, 1873–1886. [Google Scholar] [CrossRef]
  47. Grömping, U. Relative Importance for Linear Regression inR: The Packagerelaimpo. J. Stat. Softw. 2006, 17, 1–27. [Google Scholar] [CrossRef]
  48. Fernández-Martínez, M.; Vicca, S.; Janssens, I.A.; Sardans, J.; Luyssaert, S.; Campioli, M.; Chapin Iii, F.S.; Ciais, P.; Malhi, Y.; Obersteiner, M.; et al. Nutrient availability as the key regulator of global forest carbon balance. Nat. Clim. Chang. 2014, 4, 471–476. [Google Scholar] [CrossRef]
  49. Yao, Y.; Wang, X.; Li, Y.; Wang, T.; Shen, M.; Du, M.; He, H.; Li, Y.; Luo, W.; Ma, M.; et al. Spatiotemporal pattern of gross primary productivity and its covariation with climate in China over the last thirty years. Glob. Chang. Biol. 2018, 24, 184–196. [Google Scholar] [CrossRef]
  50. Amantai, N.; Ding, J.; Ge, X.; Bao, Q. Variation characteristics of actual evapotranspiration and meteorological elements in the Ebinur Lake basin from 1960 to 2017. Acta Geogr. Sin. 2021, 76, 1177–1192. [Google Scholar] [CrossRef]
  51. Zhao, M.; Liu, Y.; Konings, A.G. Evapotranspiration frequently increases during droughts. Nat. Clim. Chang. 2022, 12, 1024–1030. [Google Scholar] [CrossRef]
  52. Save, H.; Bettadpur, S.; Tapley, B.D. High-resolution CSR GRACE RL05 mascons. J. Geophys. Res. Solid Earth 2016, 121, 7547–7569. [Google Scholar] [CrossRef]
  53. Svoboda, M.D.; Fuchs, B.A. Handbook of Drought Indicators and Indices; World Meteorological Organization: Geneva, Switzerland, 2016; Volume 2. [Google Scholar]
  54. Zhang, K.; Kimball, J.S.; Mu, Q.; Jones, L.A.; Goetz, S.J.; Running, S.W. Satellite based analysis of northern ET trends and associated changes in the regional water balance from 1983 to 2005. J. Hydrol. 2009, 379, 92–110. [Google Scholar] [CrossRef]
  55. Xu, H.; Lian, X.; Slette, I.J.; Yang, H.; Zhang, Y.; Chen, A.; Piao, S. Rising ecosystem water demand exacerbates the lengthening of tropical dry seasons. Nat. Commun. 2022, 13, 4093. [Google Scholar] [CrossRef] [PubMed]
  56. Hao, Z.; AghaKouchak, A. Multivariate Standardized Drought Index: A parametric multi-index model. Adv. Water Resour. 2013, 57, 12–18. [Google Scholar] [CrossRef]
  57. Kao, S.-C.; Govindaraju, R.S. A copula-based joint deficit index for droughts. J. Hydrol. 2010, 380, 121–134. [Google Scholar] [CrossRef]
  58. AghaKouchak, A.; Farahmand, A.; Melton, F.S.; Teixeira, J.; Anderson, M.C.; Wardlow, B.D.; Hain, C.R. Remote Sens. of drought: Progress, challenges and opportunities. Rev. Geophys. 2015, 53, 452–480. [Google Scholar] [CrossRef]
  59. Yang, X.; Wang, N.; Chen, A.A.; He, J.; Hua, T.; Qie, Y. Changes in area and water volume of the Aral Sea in the arid Central Asia over the period of 1960–2018 and their causes. Catena 2020, 191, 104566. [Google Scholar] [CrossRef]
  60. Chen, Y.; Li, W.; Fang, G. Changes of key hydrological elements and research progress of water cycle in the Tianshan Mountains, Central Asia. Arid. Land Geogr. 2022, 45, 1–8. [Google Scholar] [CrossRef]
  61. Chen, Y.; Li, Z.; Fan, Y.; Wang, H.; Deng, H. Progress and prospects of climate change impacts on hydrology in the arid region of northwest China. Environ. Res. 2015, 139, 11–19. [Google Scholar] [CrossRef]
  62. Wang, S.; Zhang, M.; Li, Z.; Wang, F.; Li, H.; Li, Y.; Huang, X. Glacier area variation and climate change in the Chinese Tianshan Mountains since 1960. J. Geogr. Sci. 2011, 21, 263–273. [Google Scholar] [CrossRef]
  63. Pritchard, H.D. Asia’s shrinking glaciers protect large populations from drought stress. Nature 2019, 569, 649–654. [Google Scholar] [CrossRef]
  64. Huntington, T.G. Evidence for intensification of the global water cycle: Review and synthesis. J. Hydrol. 2006, 319, 83–95. [Google Scholar] [CrossRef]
Remotesensing 15 04557 g001
Figure 1. Map of the study area. (a) Lakes, mountains, and administrative units in Central Asia. TS, KL, KAZ, UZB, TKM, KGZ, TJK, and ANW represent Tianshan Mountains, Kunlun Mountains, Kazakhstan, Uzbekistan, Turkmenistan, Kyrgyzstan, Tajikistan, and the arid regions of Northwestern China, respectively. (b) Aridity indexes, watersheds, and rivers in Central Asia. HAA, AA, SAA, SHA, and HA represent hyperarid, arid, semiarid, subhumid, and humid areas, respectively, and the numbers 1–11 represent the 11 watersheds located in the study area, which are the Ural River Basin, Uzboy River Basin, Irtysh River Basin, Syr Darya River Basin, Amu Darya River Basin, Balkhash Lake Basin, Manas River Basin, Baiyang River Basin, Tarim River Basin, Heihe River Basin, and Yellow River Basin. The boundary map was downloaded from http://bzdt.ch.mnr.gov.cn/index.html with the approval number GS (2021)5453 accessed on 14 September 2023.
Figure 1. Map of the study area. (a) Lakes, mountains, and administrative units in Central Asia. TS, KL, KAZ, UZB, TKM, KGZ, TJK, and ANW represent Tianshan Mountains, Kunlun Mountains, Kazakhstan, Uzbekistan, Turkmenistan, Kyrgyzstan, Tajikistan, and the arid regions of Northwestern China, respectively. (b) Aridity indexes, watersheds, and rivers in Central Asia. HAA, AA, SAA, SHA, and HA represent hyperarid, arid, semiarid, subhumid, and humid areas, respectively, and the numbers 1–11 represent the 11 watersheds located in the study area, which are the Ural River Basin, Uzboy River Basin, Irtysh River Basin, Syr Darya River Basin, Amu Darya River Basin, Balkhash Lake Basin, Manas River Basin, Baiyang River Basin, Tarim River Basin, Heihe River Basin, and Yellow River Basin. The boundary map was downloaded from http://bzdt.ch.mnr.gov.cn/index.html with the approval number GS (2021)5453 accessed on 14 September 2023.
Remotesensing 15 04557 g001
Remotesensing 15 04557 g002
Figure 2. Workflow of this study. TWS, LA, SM, ET, HD, AD, MD, SPAET, and LMG denote terrestrial water storage, lake area, soil moisture, actual evapotranspiration, hydrological drought, agricultural drought, meteorological drought, the spatial efficiency metric, and the Lindeman–Merenda–Gold method, respectively.
Figure 2. Workflow of this study. TWS, LA, SM, ET, HD, AD, MD, SPAET, and LMG denote terrestrial water storage, lake area, soil moisture, actual evapotranspiration, hydrological drought, agricultural drought, meteorological drought, the spatial efficiency metric, and the Lindeman–Merenda–Gold method, respectively.
Remotesensing 15 04557 g002
Remotesensing 15 04557 g003
Figure 3. (a–f) Spatiotemporal variations in meteorological, agricultural, and hydrologic drought in arid Central Asia in the period 2000–2021. Black points mark pixels with a significant linear trend (p < 0.05) of annual indicators. P, SM, LA, and TWS represent precipitation, soil moisture, the lake area, and terrestrial water storage, respectively. The top right corners of show probability density histograms.
Figure 3. (a–f) Spatiotemporal variations in meteorological, agricultural, and hydrologic drought in arid Central Asia in the period 2000–2021. Black points mark pixels with a significant linear trend (p < 0.05) of annual indicators. P, SM, LA, and TWS represent precipitation, soil moisture, the lake area, and terrestrial water storage, respectively. The top right corners of show probability density histograms.
Remotesensing 15 04557 g003
Remotesensing 15 04557 g004
Figure 4. Meteorological, agricultural, and hydrological drought parameters.
Figure 4. Meteorological, agricultural, and hydrological drought parameters.
Remotesensing 15 04557 g004
Remotesensing 15 04557 g005
Figure 5. Correlations between evapotranspiration (ET), precipitation (P), terrestrial water reserves (TWS), the lake area (LA), and soil moisture (SM) in the period 2000–2020. The blue line’s label is seen as the bottom blue labels and the ticks are seen as the left blue ticks in the top row; the orange line’s label is seen as the left orange labels and the ticks are seen as the right y-axis in the right-most column. The x-axis for both is the x-axis of the top row.
Figure 5. Correlations between evapotranspiration (ET), precipitation (P), terrestrial water reserves (TWS), the lake area (LA), and soil moisture (SM) in the period 2000–2020. The blue line’s label is seen as the bottom blue labels and the ticks are seen as the left blue ticks in the top row; the orange line’s label is seen as the left orange labels and the ticks are seen as the right y-axis in the right-most column. The x-axis for both is the x-axis of the top row.
Remotesensing 15 04557 g005
Remotesensing 15 04557 g006
Figure 6. Spatial efficiency of trends and characteristics of various types of drought in the period 2003–2020, where HD, AD, MD, SPAEF, α, β, and γ represent hydrological drought, agricultural drought, meteorological drought, spatial efficiency, Pearson’s correlation coefficient, the fraction of the coefficient of variation, and histogram match, respectively.
Figure 6. Spatial efficiency of trends and characteristics of various types of drought in the period 2003–2020, where HD, AD, MD, SPAEF, α, β, and γ represent hydrological drought, agricultural drought, meteorological drought, spatial efficiency, Pearson’s correlation coefficient, the fraction of the coefficient of variation, and histogram match, respectively.
Remotesensing 15 04557 g006
Remotesensing 15 04557 g007
Figure 7. Contributions of precipitation (P) (a,b), evapotranspiration (ET) (c,d), and runoff (e,f) to agricultural drought intensity (a,c,e) and hydrologic drought intensity (b,d,f). Black points indicate values of statistical significance (p < 0.05).
Figure 7. Contributions of precipitation (P) (a,b), evapotranspiration (ET) (c,d), and runoff (e,f) to agricultural drought intensity (a,c,e) and hydrologic drought intensity (b,d,f). Black points indicate values of statistical significance (p < 0.05).
Remotesensing 15 04557 g007
Remotesensing 15 04557 g008
Figure 8. Precipitation (P) and actual evapotranspiration (ET) cross-wavelet (XWT) (a) and wavelet coherence (WTC) (b) spectra. In the figure, “→” indicates a positive correlation, “←” indicates a negative correlation, and “↓” indicates that the ET phase change exceeds the precipitation phase by 90° (i.e., it exceeds the precipitation phase by ¼ of a cycle in the time series). The time–frequency domain within the thick black solid line passes the significance test (p < 0.05).
Figure 8. Precipitation (P) and actual evapotranspiration (ET) cross-wavelet (XWT) (a) and wavelet coherence (WTC) (b) spectra. In the figure, “→” indicates a positive correlation, “←” indicates a negative correlation, and “↓” indicates that the ET phase change exceeds the precipitation phase by 90° (i.e., it exceeds the precipitation phase by ¼ of a cycle in the time series). The time–frequency domain within the thick black solid line passes the significance test (p < 0.05).
Remotesensing 15 04557 g008
Remotesensing 15 04557 g009
Figure 9. Contributions of precipitation (P) (a,b), evapotranspiration (ET) (c,d), and runoff (e,f) to SPEI (a,c,e) and PDSI (b,d,f).
Figure 9. Contributions of precipitation (P) (a,b), evapotranspiration (ET) (c,d), and runoff (e,f) to SPEI (a,c,e) and PDSI (b,d,f).
Remotesensing 15 04557 g009
Table 1. All data used in this study.
Table 1. All data used in this study.
DataDescriptionData SourcesResolutionUnit
SpatialTemporal
PPrecipitationCRU0.5° × 0.5°1 monthmm
PETPotential EvapotranspirationCRU0.5° × 0.5°1 monthmm
ETActual EvapotranspirationPT-DTsR1 km × 1 km16 daysmm
SMSoil MoistureERA5_land0.1° × 0.1°1 monthm3 × m−3
RORunoffERA5_land0.1° × 0.1°1 monthm
TWSTerrestrial Water StorageGRACE0.25° × 0.25°1 monthcm
LALake AreaReaLSat300 m × 300 m1 yearkm2
PDSIPalmer Drought Severity Indexsc_PDSI2.5° × 2.5°1 month
SPEIStandardized Precipitation Evapotranspiration Index 0.5° × 0.5°1 month
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Zhao, Z.; Hao, X.; Fan, X.; Zhang, J.; Zhang, S.; Li, X. Actual Evapotranspiration Dominates Drought in Central Asia. Remote Sens. 2023, 15, 4557. https://doi.org/10.3390/rs15184557

AMA Style

Zhao Z, Hao X, Fan X, Zhang J, Zhang S, Li X. Actual Evapotranspiration Dominates Drought in Central Asia. Remote Sensing. 2023; 15(18):4557. https://doi.org/10.3390/rs15184557

Chicago/Turabian Style

Zhao, Zhuoyi, Xingming Hao, Xue Fan, Jingjing Zhang, Sen Zhang, and Xuewei Li. 2023. "Actual Evapotranspiration Dominates Drought in Central Asia" Remote Sensing 15, no. 18: 4557. https://doi.org/10.3390/rs15184557

APA Style

Zhao, Z., Hao, X., Fan, X., Zhang, J., Zhang, S., & Li, X. (2023). Actual Evapotranspiration Dominates Drought in Central Asia. Remote Sensing, 15(18), 4557. https://doi.org/10.3390/rs15184557

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop