Spatiotemporal Dynamics of Drought and the Ecohydrological Response in Central Asia

Abstract

Due to the influences of climate change and human activities, the resources and environments of the “One Belt and One Road” initiative are facing severe challenges. Using drought indicators, this study aimed to analyze the spatiotemporal characteristics of the drought environment and the response of vegetation cover in the area to drought conditions. The Gravity Recovery and Climate Experiment (GRACE) drought severity index (GRACE-DSI), GRACE water storage deficit index (GRACE-WSDI) and standardized precipitation index (SPI) were calculated to measure hydrological drought. Additionally, based on GRACE and Global Land Data Assimilation System (GLDAS) data, groundwater data in Central Asia was retrieved to calculate the groundwater drought index using the GRACE Standardized Groundwater Level Index (GRACE-SGI). The findings indicate that, from 2000, Central Asia’s annual precipitation decreased at a rate of 1.80 mm/year (p < 0.1), and its annual temperature increased slightly, at a rate of 0.008 °C/year (p = 0.62). Water storage decreased significantly at a rate of −3.53 mm/year (p < 0.001) and showed an increase-decrease-increase-decrease pattern. During the study period, the aridity in Central Asia deteriorated, especially on the eastern coast of the Caspian Sea and the Aral Sea basin. After 2020, most of Central Asia experienced droughts at both the hydrological and groundwater droughts levels and of varying lengths and severity. During the growing season, there was a substantial positive association between the Normalized Difference Vegetation Index (NDVI) and drought indicators such as GRACE-DSI and GRACE-WSDI. Nonetheless, the NDVI of cultivated land and grassland distribution areas in Central Asia displayed a strong negative correlation with GRACE-SGI. This study concludes that the arid environment in Central Asia affected the growth of vegetation. The ecological system in Central Asia may be put under additional stress if drought conditions continue to worsen. This paper explores the drought characteristics in Central Asia, especially those of groundwater drought, and analyzes the response of vegetation, which is very important for the ecological and environmental protection of the region.

1. Introduction

Drought is the most common and widespread natural disaster in the world [1,2], affecting sub-regions such as Europe [3], South Asia [4], and Northern China [5]. Drought has a profound impact on agricultural productivity, the ecological environment, and socioeconomic development [6]. The economic losses caused by global climatic disasters account for 70% of total losses involving natural disasters, while the losses resulting from drought account for 50% of the total losses incurred by climatic disasters [7]. Climate change is still a global challenge that is poorly addressed by governments, and global warming has been found to contribute to the more severe drought disasters and risks faced [8]. Unfortunately, as the frequency and intensity of extreme drought occurrences continue to rise, the occurrence of worldwide drought disasters has progressively grown. Drought disasters have become more prominent and more destructive, and their impacts lead to global consequences [9]. Droughts not only cause huge economic losses, but they also have a devastating influence on society and the ecological environment, especially in developing countries. In addition, human activities have contributed to accelerating droughts [10], especially in arid and semi-arid regions, where natural ecosystems are relatively vulnerable, with a combination of low rainfall, intensive evaporation, low vegetation covers, and scarce water resources. The drought occurrences in human-dominated environments should not be perceived as purely natural hazards since anthropogenic changes to the land surface alter hydrological processes and affect the development of drought [11]. Meteorological to hydrological drought propagation was influenced mostly by human activities, which has both positive and negative effects on the severity and duration of droughts [12].

Central Asia is the core area of “One Belt and One Road”, and is an inland arid and semi-arid region, with extremely unevenly distributed water resources [13,14]. Municipal and industrial water sources are mainly sourced from glacial meltwater, which somewhat alleviates the drought stress in Central Asia [15]. Due to significant geographic heterogeneity and complex climate system processes [16], ecosystems are very sensitive to global climate changes [17]. In the past 30 years, the temperature rise rate in Central Asia reached 0.4 °C/decade, higher than that of the northern hemisphere land and the surrounding areas (0.3 °C/decade). Moreover, rapid warming in the center of Central Asia has been reported, with a decreasing rate from the northwest to the southeast [18]. The changing climate impacts water resources in the region directly and rapidly. The total water storage over Central Asia, derived from the Gravity Recovery and Climate Experiment (GRACE) and Global Land Data Assimilation System (GLDAS), have shown a decreasing trend over the past 30 years [19]. The water resources diminished from 2003 to 2013 at a rate of −4.44 ± 2.2 mm/a [20].

Climate change and its environmental pressure on the unique temperate desert ecosystems of Central Asia threaten the ecological security of the core area of “One Belt and One Road”. In the context of global warming, there have been significant vegetation changes in Central Asia since the 21st century. While vegetation greening led to a significant temperature decrease, moderating regional warming throughout Central Asia [21], the arid environment in Central Asia is becoming drier [22]. Between 1966 and 2015, most drought events in Central Asia lasted from 3 to 5 months and tended to have an east–west trajectory [23]. Since the beginning of the 21st century, Kazakhstan suffered drought hazards every year. Droughts with different severities occurred in half of Central Asia in 2000, 2008, 2011, 2012, and 2014, and the most severe droughts took place in 2012 and 2014 [24]. Due to the combination of low rainfall, high temperature, and frequent natural disasters, western Central Asia’s desertification is intensifying [25]. In the next century, Central Asia is expected to become drier, especially in its western part, which includes Turkmenistan, Uzbekistan, and Kazakhstan, potentially impacting agriculture in the region [26]. However, there is currently insufficient research on the impact of drought on the ecological environment in Central Asia.

Precipitation and evapotranspiration data from hydro-meteorological observations have historically served as the primary and traditional basis for drought monitoring and evaluation methods. Due to the spatial heterogeneity of the natural environment and the scarcity of ground-based observations, the spatial distribution characteristics of droughts obtained through interpolation have large uncertainties [27], which necessitate further improvements. Additionally, it is infeasible to adequately assess the risk of drought disasters based on a single variable or indicator, such as precipitation, runoff, or soil water [28]. The causes and effects of droughts are complex, affecting not only surface water but also groundwater. Changes in regional-scale water storage include groundwater, soil moisture, surface water, ice, snow, biological water content, precipitation, evapotranspiration, and so on. Hence, drought indicators that take into account all of these changes can more comprehensively capture the characteristics of drought disasters.

GRACE satellites are used to monitor the dynamics of terrestrial water storage [29] and the hydrological variations [30]. The change in water storage derived by the GRACE satellites, including biological water content, surface water, soil moisture, ice/snow water, and groundwater, is the composite outcome of natural processes and human activities. The analysis of droughts from GRACE-based terrestrial water storage anomalies (TWSA) is more comprehensive than the traditional methods used [31]. Compared to other remote sensing data for monitoring the occurrence of droughts, GRACE assessments of water storage have the advantage of predicting hidden drought disasters. For example, while soil moisture in the root zone may be at suitable levels, irrigated agriculture consumes a large amount of groundwater. Thus, GRACE assessments of water storage are valuable for monitoring and predicting drought disasters. For example, drought characteristics and trends over ten major river basins in China were analyzed utilizing the GRACE drought severity index (DSI) [32]. And the spatiotemporal evolution of groundwater drought could also be observed using the GRACE data [33].

Based on the gridded hydrological drought index and groundwater drought index retrieved from GRACE satellite TWSA, this paper quantitatively evaluates the spatial and temporal distribution of hydrological and groundwater drought events in Central Asia from April 2002 to May 2023. The primary objectives were to (1) analyze the spatiotemporal distribution of climate, water resource, and drought conditions in Central Asia; (2) assess drought from multiple perspectives; and (3) investigate the response of vegetation to drought. Exploring the drought characteristics in Central Asia, especially groundwater drought, and analyzing the response of vegetation are critical for ecological and environmental protection. This study could provide scientific data for eco-environmental protection policies and practices.

2. Materials and Methods

2.1. Study Area

Central Asia (35.14°~55.44°N, 46.50°~87.35°E) includes the following five countries: Kazakhstan, Kyrgyzstan, Tajikistan, Uzbekistan, and Turkmenistan. It is located in the Eurasian continent’s hinterland, covering a vast area of about 4.19 million km² (Figure 1). The terrain is generally high in altitude in the southeast and low in the northwest, with a typical mountain basin structure. The region has the typical continental climate of temperate deserts and grasslands, with sparse and extremely uneven distributed precipitation (125 to 289 mm/year) [34], and with climate being drier in the east and west compared to the rest of the region. The average summer (July) temperature is between 26 °C and 32 °C, while temperature ranges from −20 °C at the northern end to 2 °C in the south of the basin. Regional evapotranspiration greatly exceeds precipitation. As shown in Figure 1, the major vegetation types in Central Asia are cropland (22%), grassland (24%), forest (2%), shrubs (7%), and sparse vegetation (17%). The southeast parts of Central Asia are mainly mountainous areas, with abundant water resources. Central Asia’s center is mainly covered by sparse vegetation; northern Central Asia is dominated by farmland and grassland, while grasslands and croplands are intermixed in the eastern part. The mountain–oasis–desert ecosystem in this area contributes to its unique and complex response to global climate changes [35].

Central Asia has rich land resources; however, local agriculture is not fully developed due to the limitation of water resources. The per capita arable land area of the five Central Asia countries is 0.52 hectares, which is about six-fold that of China. Agriculture is the primary sector in Central Asia; about 90% of water abstractions used for local farming, with the land undergoing widespread irrigation [36]. According to the Food and Agriculture Organization of the United Nations’ (FAO) statistics, irrigated agriculture prevails in Central Asia except for the northern part of Kazakhstan, which features rained agriculture (https://www.fao.org/aquastat/en/geospatial-information/global-maps-irrigated-areas/latest-version accessed on 12 December 2024). Given the necessity for irrigation across most of Central Asia, water resources are the most critical factor driving the economic and social development of the region [37].

2.2. Datasets

The GRACE satellites launched in 2002 provide a new method for monitoring drought/flood events. Terrestrial water storage changes can be captured by GRACE satellites observing Earth’s mass redistribution. The latest GRACE CSR RL06 Mascon products were retrieved via the GRACE website (https://www2.csr.utexas.edu/grace/RL06_mascons.html, accessed on 12 December 2024). This dataset has a monthly temporal resolution and a spatial resolution of 0.25°. Previous studies have shown that the accuracy of the data based on the mascon algorithm is higher [38]. The TWSA were calculated by removing the monthly average of raw data during the study period. Previous studies that focused on the spatiotemporal characteristics of TWSA of Central Asia [19,20] approved the effectiveness of GRACE data in this region. With the support of TWSA, we calculated two hydrological drought indicators, the GRACE drought severity index (GRACE-DSI) [39] and the GRACE water storage deficit index (GRACE-WSDI) [40].

The GRACE TWSA integrates five parts, including groundwater, soil water, surface water, ice and snow water, and biological water content in the vertical direction [41]. Biological water change is usually neglected due to its relatively small impact on deriving groundwater change [42]. Given the availability of the GRACE TWSA, the soil moisture, snow water equivalent, canopy water, and surface water data of the GLDAS Noah model, we derived the groundwater storage anomalies (GWSA) in Central Asia. Further, the GRACE standardized groundwater level index (GRACE-SGI) [43] was calculated accordingly.

The soil moisture (SM), snow water equivalent (SWE), and storm surface runoff were obtained from the Noah model of GLDAS-2.1 (https://disc.gsfc.nasa.gov/datasets/GLDAS_NOAH025_M_2.1/summary?keywords=GLDAS, accessed on 12 December 2024) [44], at a monthly time scale with a spatial resolution of 0.25° × 0.25°. The GLDAS model data have been widely used in various regions, including Central Asia [19,20]. All of the spatial data were reprojected to the WGS84 geographic coordinate system and were masked by the study area’s boundary. Based on the monthly soil moisture, snow water equivalents, and storm surface runoff, the monthly anomalies between 2002 and 2023 were calculated, which are the monthly values of soil moisture, snow water equivalent, or storm surface runoff minus the averages.

The climate distribution in Central Asia was analyzed utilizing the gridded precipitation and air temperature from ECMWF reanalysis v5 (ERA5). ERA5, with a higher spatiotemporal resolution, provided global climate data covering the period from January 1940 to the present. The monthly precipitation and air temperature with a spatial resolution of 0.25° from 2000 to 2023 were obtained from https://cds.climate.copernicus.eu (accessed on 12 December 2024). The Standard Precipitation Index (SPI) [45] with a spatial resolution of 0.25° × 0.25° at different periods, calculated utilizing the monthly precipitation of ERA5.

We utilized the L3 level normalized difference vegetation index (NDVI) at a 0.05° spatial resolution of the Terra Moderate Resolution Imaging Spectroradiometer (MODIS) sensor (MODIS/Terra Vegetation Indices Monthly L3 Global 0.05deg CMG) from 2002 to 2023, derived from the NASA Land Processes Distributed Active Archive Center (available at https://ladsweb.modaps.eosdis.nasa.gov/search/, accessed on 12 December 2024). The MODIS Reprojection Tools were used to implement the format conversion, splicing, and projection.

With the support of Python and the arcpy package, the 0.05°grid NDVI data were resampled to a 0.25° resolution.

2.3. Data Analysis

2.3.1. Groundwater Storage Anomalies

Based on the principle of groundwater-derived anomalies, the groundwater storage anomalies in the region were quantitatively retrieved using the terrestrial water storage anomalies, soil moisture water anomalies, snow water equivalent anomalies, and surface water anomalies expressed by Equation (1).

$$\mathrm{GWSA}=\mathrm{TWSA}-\mathrm{SMA}-\mathrm{SWEA}-\mathrm{SWA}$$

(1)

where GWSA represents the change in the groundwater storage, TWSA is the terrestrial water storage anomalies, SMA is the soil moisture anomalies, SWEA is the snow water equivalent anomalies, and SWA is the surface water anomalies.

Groundwater storage is associated with water security, which becomes more important during periods of drought. During drought, agricultural water demands, industry water demands, etc., are often met by groundwater storage. Therefore, groundwater anomalies are caused by climate changes and anthropogenic activities.

2.3.2. Drought Indexes

The GRACE-DSI, GRACE-WSDI, and GRACE-SGI were calculated based on the changes in the water content of a grid. For each grid, the standardized anomalies ($DS{I}_{i,j}$) of GRACE TWSA for month j and year i (ranging from 2002 to 2023) were calculated as follows:

$$DS{I}_{i,j}={\displaystyle \raisebox{1ex}{$TWS{A}_{i,j}-\overline{TWS{A}_j}$}\!\left/ \!\raisebox{-1ex}{${\sigma }_j$}\right.}$$

where ${\sigma }_j$ and $\overline{TWS{A}_j}$ are the standard deviation and average value of TWSA for month j, respectively [39]. The global GRACE-DSI distribution meets the non-standard state distribution.

The GRACE-WSDI [40] is able to identify various levels of hydrological drought events with following equation:

$${\mathrm{WSD}}_{i.j}=TWS{A}_{i.j}-\overline{TWS{A}_j}$$

(2)

$${\mathrm{WSDI}}_{i,j}={\displaystyle \frac{WS{D}_{i.j}-\mu }{\sigma }}$$

(3)

where ${\overline{TWSA}}_j$ refers to the average TWSA of months j from 2002 to 2023. In addition, $\mu \mathrm{and} \sigma$ refer to the mean value and standard error of the WSD time series data, respectively. Based on the standard normal distribution, the GRACE-WSDI values greater than 0, between −1.0 and 0, −2.0 and −1.0, −3.0 and −2.0, and less than −3.0 are classified as near normal, mild dry, moderate drought, severe drought, and extreme drought, respectively.

The GRACE-DSI and GRACE-WSDI are mainly applied to assess hydrological drought events. When an extreme hydrological drought occurs, groundwater drought events also occur when groundwater resources are severely depleted. Based on the GRACE inversion of grid-scale groundwater data, the groundwater drought assessment index GRACE-SGI [43] is calculated via Equation (4) as follows:

$${\mathrm{SGI}}_{i,j}={\displaystyle \frac{GWS{A}_{i,j}-\overline{GWS{A}_j}}{{\sigma }_{GWS{A}_j}}}$$

(4)

where i is the year ranging from 2002 to 2023;$GWS{A}_{i,j}$ refers to the groundwater storage anomalies of month j, year i; and $\overline{GWS{A}_j}$ and ${\sigma }_{GWS{A}_j}$ are the mean and standard deviation of groundwater storage anomalies in month j, respectively.

SPI is a commonly used indicator for assessing drought disasters based on precipitation alone [46]. Since the distribution of precipitation is asymmetric, when analyzing drought disasters based on precipitation, the gamma function $\Gamma$ describes the change in precipitation. After calculating the $\Gamma$ probability distribution in a certain period, standard normalization processing is performed, and the drought level is finally classified by the cumulative frequency distribution of precipitation. Assuming that the amount of precipitation in a certain period is $x$, its probability density function (PDF) is expressed by Equation (5) as follows:

$$f\left(x\right)={\displaystyle \frac{1}{{\beta }^{\alpha }\Gamma \left(x\right)}}{x}^{\alpha -1}{e}^{-{\displaystyle \frac{x}{\beta }}},x>0$$

(5)

where $\alpha$ and $\beta$ are the shape parameters and scales, respectively, $\alpha$ > 0, $\beta$ > 0; $x$ is the amount of precipitation; and $\Gamma \left(x\right)$ is the gamma function. In addition, $\alpha$ and $\beta$ can be obtained by the maximum likelihood estimation method, according to Equations (6) and (7) as follows:

$$\alpha ={\displaystyle \frac{1}{4A}}\left(1+\sqrt{1+{\displaystyle \frac{4A}{3}}}\right)$$

(6)

$$\beta ={\displaystyle \frac{\overline{x}}{\alpha }}$$

(7)

where $A=\ln \left(\overline{x}\right)-{\displaystyle \frac{\sum \ln \left(x\right)}{n}}$, n is the length of the measured precipitation and the specific calculation steps are in accordance with McKee et al. [45].

The monthly SPI is calculated based on the precipitation of the previous i months. The specific period length is i. Moreover, i could be equal to 1, 3, 6, 9, 12, 24, and 36. When i is 1 or 3, it is a short-term drought, 6 and 9 represent medium-term droughts, and 12, 24, and 36 mean long-term droughts. The short-term SPIs (SPI1, SPI3) reflect the drought effects on agriculture in the short term, such as soil moisture deficit, crop yield reduction, etc. When SPI is computed for medium accumulation periods (e.g., from 6 to 9 months), it can be used as an indicator for reduced stream flow and reservoir storage. Long-term SPIs (SPI12, SPI24, and SPI36) reflect the droughts in terms of hydrology and water supply, such as reduced groundwater, runoff, and reservoir water storage [47]. In this paper, SPI1, SPI3, SPI6, SPI12, and SPI24 were selected as short-term, medium-term, and long-term droughts.

Table 1. Dynamic range and corresponding categories for dry conditions of the GRACE-DSI, GRACE-WSDI, GRACE-SGI, and SPI.

Category	Description	GRACE-DSI	GRACE-WSDI	GRACE-SGI	SPI
D0	Near normal	0.49 to −0.49	0 or greater	−0.3 or greater	−0.5 or greater
D1	Mild drought	−0.50 to −0.79	−1.0 to 0	−0.6 to −0.3	−1.0 to −0.5
D2	Moderate drought	−0.80 to −1.29	−2.0 to −1.0	−0.9 to −0.6	−1.5 to −1.0
D3	Severe drought	−1.30 to −1.59	−3.0 to −2.0	−1.2 to −0.9	−2.0 to −1.5
D4	Extreme drought	−1.60 to −1.99	−3.0 or less	−1.5 to −1.2	−2.0 or less
D5	Exceptional drought	−2.0 or less		−1.5 or less

lists the classification of drought levels according to the GRACE-DSI, GRACE-WSDI, GRACE-SGI, and SPI [39,40,48,49].

2.3.3. Linear Regression Analysis

The spatiotemporal dynamics of the arid environment (drought index) and climate (precipitation and temperature) in Central Asia were analyzed using univariate linear regression. The rates of change in TWSA and drought indexes (i.e., DSI, WSDI, and SGI) were calculated, respectively, by univariate linear regression [50].

Many statistical methods are available for measuring time series trends. We used Pearson’s correlation coefficient to judge correlation, and we estimated the significance of change trends based on the p-value.

2.3.4. Correlation Analysis

The correlation coefficient can quantitatively describe the strength of the linear relationship between two variables. For two variables x and y, if their sample values are $x_i$ and $y_i$ (i = 1,2,..., n), the coefficient correlation between them is expressed by Equation (8) as follows:

$$r_{x,y}={\displaystyle \frac{{{\displaystyle \sum }}_{i=1}^n\left(x_i-\tilde{x}\right)\left(y_i-\tilde{y}\right)}{\sqrt{{{\displaystyle \sum }}_{i=1}^n{\left(x_i-\tilde{x}\right)}^2\times {{\displaystyle \sum }}_{i=1}^n{\left(y_i-\tilde{y}\right)}^2}}}$$

(8)

where $\tilde{x}$ and $\tilde{y}$ represent the average of the sample values of the two variables.

We calculated two correlation coefficients, one between the GRACE drought index and SPI, and the other between the GRACE and NDVI. The value of r can range from −1 to 1. Positive values indicate a positive correlation and vice versa; 0 means no correlation. The non-parametric Mann–Kendall trend test method [51] was applied to determine the changing trend at the confidence level of 95%.

The data processing flow is shown below (Figure 2).

3. Results

3.1. Spatiotemporal Distribution of the Climate in Central Asia

Central Asia is a climate-sensitive area. From 1901 to 2013, the seasonal precipitation in Central Asia concentrated in spring and summer, with spring precipitation being the most prominent [52]. The monthly precipitation in spring and winter increased, while the precipitation in summer and autumn gradually decreased [52]. Compared with previous studies, we found that the annual precipitation in Central Asia slightly and significantly decreased, while temperature slightly and non-significantly increased since 2000 (Figure 3). The annual precipitation decreased at a rate of 1.80 mm/year (p < 0.1) (Figure 3a), while the annual mean temperature increased at a rate of 0.008 °C/year (p = 0.62) (Figure 3c). However, strong inter-annual variation in both variables existed in the period. The maximum and minimum precipitations were recorded in 2016 and 2008, respectively, while the highest and lowest mean temperatures were recorded in 2013 and 2014, respectively. Figure 3b,d illustrates the seasonal distribution characteristics in the precipitation and temperature of the region. Precipitation concentrated in the summer (March–May), accounting for approximately 31.9% of annual precipitation. Peak precipitation occurred in March and the lowest was in August. The year’s highest temperature was recorded in July, while the coldest temperature was recorded in January.

The spatial distribution of precipitation varied greatly in Central Asia. As seen from Figure 3e, the annual average precipitation gradually increases from west to east in the whole region. The maximum annual precipitation is 2008.71 mm, and the minimum is 74.60 mm. The largest amounts of precipitation are in southeast and northeast central Asia (Tajikistan, east of Uzbekistan, west of Kyrgyzstan, and northeast of Kazakhstan), with annual average precipitation exceeding 1000 mm. The region with the least precipitation is mainly in central and southwest Central Asia, with annual average precipitation being less than 200 mm. A total of 1.5% of the regions received more than 1000 mm of annual average precipitation, 1.9% received 800–1000 mm, 2.7% received 600–800 mm, 7.0% received 400–600 mm, 46.4% received 200–400 mm, and 40.5% received less than 200 mm, respectively. Therefore, the majority of Central Asia has low precipitation, with less than 400 mm of precipitation (approximately 86.9% of the entire study area).

From Figure 3f, the annual mean temperature in Central Asia during 2000–2022 was generally high in the southwest and low in the southeast and northeast. The highest temperature was felt in southwestern Central Asia (Turkmenistan). The lowest temperature was felt in southeastern Central Asia (Tajikistan and Kyrgyzstan). The highest annual mean temperature in Central Asia was 17.73 °C and the lowest was −15.86 °C, respectively. A total of 4.2% of the regions had an annual mean temperature above 15 °C, 18.9% from 10 to 15 °C, 29.7% from 5 to 10 °C, 37.0% from 0 to 5 °C, and 10.2% below 0 °C, respectively.

At the pixel scale, most of Central Asia (approximately 77.6% of the regions) exhibited a negative linear trend in annual precipitation over the period 2000–2022 (Figure 3g). Regions with a high rate of temperature decrease (typically up to −33.68 mm/a) were primarily concentrated in the southeastern region of Central Asia. Spatially, the increasing trend of precipitation was mainly located in the central and southern parts (Figure 3g). The northern portions of Central Asia were becoming cooler, while most other regions were becoming warmer (approximately 62.5%) between 2000 and 2022 (Figure 3h). On the whole, precipitation in most regions of Central Asia showed a decreasing trend, while temperature displayed an upward trend. That is, the climate of Central Asia is warming and drying (Figure 3g,h).

3.2. Spatiotemporal Distribution of Water Resources in Central Asia

Since 2000, increased evapotranspiration, rising temperatures, and human interference have changed Central Asia’s water resources [20]. In the context of climate change and human activities, the water resources per capita in the five Central Asia countries were decreasing year by year. There were significant seasonal changes in TWSA in Central Asia (Figure 4); the peak was around April of each year, and the low value was around October of each year (Figure 4). That is, the total water storage in Central Asia in spring (March, April, May) is at its peak, while the total water storage in autumn (September, October, November) is at a trough, with significant losses. Comparing Figure 4 with Figure 3b demonstrates that the seasonality of the TWSA was consistent with precipitation, while TWSA lagged by a month. The quarterly dynamics of the water resources followed the variable precipitation. Monthly TWSA ranged from −139.17 mm to 101.32 mm, with a peak in April 2005 and a minimum in October 2022. From April 2002 to April 2005, TWSA showed an upward trend with a rate of 0.76 mm/month (p = 0.18). However, they showed a significant downward trend from April 2005 to December 2014 with a rate of −0.52 mm/month (p < 0.001). Then, from January 2015 to March 2017, it exhibited a significant upward trend with a change rate of 1.89 mm/month (p < 0.1). During April 2017~May 2023, TWSA in Central Asia displayed an obviously decreasing trend with a rate of −1.44 mm/month (p < 0.001). Since 2002, the Central Asian TWSA have fluctuated over time. Over the entire study period, the total regional water storage showed a significant downward trend with a rate of −3.53 mm/year (p < 0.001). Furthermore, there was a correlation between the decline in precipitation, increase in air temperature, and the reduction in Central Asia’s water resources.

Regarding the spatial change rate (Figure 5), the TWSA from 2002 to 2023 had evident spatial heterogeneity. The change rate of all grids varied between −55.36 mm/year to 5.21 mm/year, while 54.77% of the area showed a significant downward trend, mainly in the western and southeastern regions of Central Asia. The declining trend of the TWSA in the western region was the largest; 20.12% of the area showed a significant increasing trend, mainly in Central Asia’s northeast and southwest regions. The significant water storage reduction in the entire region was mainly due to the rapid decrease in the western region. The west became drier and the east became wetter, which was also found in other works [53]. Previous studies argued that human activity was the main reason for the rapid decline in water reserves in Central Asia’s western region [20].

3.3. Correlation Analysis Between GRACE Drought Indexes and SPI Indexes

Droughts are very likely to occur in the context of severe shortages of regional water resources in Central Asia. Previous studies showed that the SPI drought index is a significant indicator for monitoring long-term hydrological drought events [54]. We first investigated the relationship between the GRACE-DSI, GRACE-WDSI, GRACE-SGI, and SPI drought indicators. We utilized comparative analysis to identify drought events in Central Asia since 2002, and the dry/wet environment. The correlation coefficients between the average of the three indicators (GRACE-DSI, GRACE-WDSI, and GRACE-SGI) and the SPI for different periods were calculated for the entire region of Central Asia (

Table 2. Average correlations between GRACE DSI, WSDI, SGI, and SPI for different accumulation periods (** indicates the correlation is significant at the 0.01 level (double tail), * indicates the correlation is significant at the 0.05 level (double tail)).

SPI	DSI	WSDI	SGI
SPI1	0.22 **	0.23 **	−0.07
SPI3	0.47 **	0.48 **	−0.11
SPI6	0.53 **	0.54 **	−0.16 *
SPI12	0.69 **	0.70 **	−0.03
SPI24	0.80 **	0.81 **	0.13 *

). In Table 2, the correlation coefficients between the GRACE-DSI and SPI indicators were the same as the correlation coefficients between the GRACE-WDSI and SPI, indicating a similarity between GRACE-DSI and GRACE-WDSI. The correlation coefficients between GRACE-DSI, GRACE-WSDI, and SPI values ranged between 0.05 and 0.81 over different periods and gradually increased with time. The correlation coefficients of GRACE-DSI, GRACE-WSDI, and SPI-24 reached 0.81, indicating the GRACE hydrological drought indicators were sensitive to long-term hydrological drought events. The correlation coefficients between SPI and GRACE-DSI and GRACE-WSDI were larger than the correlation coefficient between SPI and GRACE-SGI, which was mainly distributed between −0.03 and 0.13, and only the correlation coefficient between GRACE-SGI and SPI-24 was positive, with a value of 0.13. Furthermore, the correlation between the GRACE drought indexes (DSI, WSDI, and SGI) and each SPI indicator were mostly significant at the 0.05 level.

Further analysis revealed the spatially distributed correlation coefficients between GRACE-DSI, GRACE-WSDI, GRACE-SGI and SPI-1, SPI-3, SPI-6, SPI-12, SPI-24. In Figure 6 and

Table 3. Percentage of pixels with positive or negative correlation and mean correlation between the GRACE drought indexes and SPI indexes.

SPI	DSI			WSDI			SGI
	>0	<0	Mean	>0	<0	Mean	>0	<0	Mean
SPI1	85.55	14.45	0.08	86.39	13.61	0.08	36.33	63.67	−0.03
SPI3	93.64	6.36	0.18	93.37	6.63	0.17	33.93	66.07	−0.05
SPI6	93.63	6.37	0.21	93.37	6.63	0.21	31.30	68.70	−0.07
SPI12	94.75	5.25	0.27	94.56	5.44	0.27	43.68	56.32	−0.03
SPI24	94.40	5.60	0.35	94.37	5.63	0.34	56.04	43.96	0.04

, the correlations between GRACE-DSI, GRACE-WSDI and SPI indicators show similar patterns, mainly with significant positive correlations. As the SPI time scale increases, the area with positive correlations increases, ranging from 85.55% to 94.40%. Strong positive correlation grids are shown to be concentrated in Tajikistan, Kyrgyzstan, and Kazakhstan’s western and southern regions. The negative correlation coefficient was relatively strong in Turkmenistan’s central region and Kazakhstan’s central region. Furthermore, with the increase in the SPI time scale, the spatial correlation coefficients showed an increasing trend.

Unlike the relationship between the GRACE-DSI and GRACE-WSDI indicators, the GRACE-SGI index and the SPI indexes were mainly negatively correlated. Among the correlation coefficients between the GRACE-SGI and SPI-3, 68.70% of the area was correlated negatively. Among the correlation coefficients between GRACE-SGI and SPI-24, 43.96% of the area was negatively correlated. The area with strong negative correlations (up to −0.66) was mainly distributed in Kazakhstan’s central and western regions and the central and eastern regions of Turkmenistan. The strong positive correlations (up to 0.67) were located mainly in Western Kazakhstan and Tajikistan. There were active dynamic changes in groundwater resources in Central Asia. The strong negative correlation between GRACE-SGI and SPI indicated that although regional precipitation showed an increasing trend, the groundwater resources were severely depleted, and their consumption was more extensive than the precipitation recharge. The strong positive correlation between GRACE-SGI and SPI, on the other hand, indicated that the groundwater storage changes were consistent with precipitation changes. According to the land use/land cover in Central Asia, the areas with a significant negative correlation between GRACE-SGI and SPI were mainly distributed in grassland areas, while the areas with a significant positive correlation were found to be areas of bare land. The findings imply that precipitation is insufficient to meet the water demand for grassland growth in Central Asia, requiring supplementary groundwater. On the other hand, the amount of precipitation in bare areas can fully recharge the groundwater resources. Previous studies also found that an increase in precipitation cannot counterbalance the aggravation of water shortage caused by the growing population, growing economy, and temperature rise in Central Asia [55], and human activities have a much bigger effect than climate change on water resource [56].

3.4. Spatiotemporal Distribution of Drought Conditions in Central Asia

The strong correlation between the GRACE drought indicators and the SPI drought indicators for different periods indicates the GRACE drought indicators’ effectiveness. According to the GRACE drought indicator category, this study only analyzed mild drought events that last for three months or more, and the drought disasters above moderate drought.

According to the GRACE drought indicators in Central Asia, since 2002, mild droughts have been the predominant events, moderate droughts have been relatively infrequent, and no extreme drought disasters occurred. Based on the GRACE-DSI, four light drought events and eleven moderate drought events occurred from April 2002 to May 2023. Based on the GRACE-WSDI, eight light droughts and two moderate droughts occurred. Based on the GRACE-SGI, there were six light droughts, two moderate droughts, and one severe drought. Although the results differ among indicators, the GRACE-DSI was similar to the GRACE-WSDI, which was sensitive to regional hydrological droughts. The GRACE-SGI was more sensitive to drought events due to insufficient groundwater resources. Central Asia has been relatively dry since 2020, and there were many consecutive hydrological droughts and groundwater droughts in different periods (

Table 4. Drought disasters in Central Asia obtained from the DSI, WSDI, and SGI.

Drought Indicators	No. of Events	Time Span of Each Event	Duration (Months)	Drought Classes
				Mild Drought	Moderate Drought	Severe Drought
DSI	15	December 2008~April 2009	5	✓
		May 2012~August 2012	4	✓
		June 2012	1		✓
		September 2014	1		✓
		December 2020~February 2021	3	✓
		January 2021	1		✓
		June 2021~May 2023	24	✓
		July 2021~August 2021	2		✓
		October 2021	1		✓
		December 2021	1		✓
		March 2022	1		✓
		May 2022	1		✓
		August 2022 ~September 2022	2		✓
		November 2022	1		✓
		May 2023	1		✓
WSDI	10	October 2007~February 2008	5	✓
		April 2008~December 2009	21	✓
		June 2010~November 2011	18	✓
		January 2012~February 2014	26	✓
		April 2014~October 2014	7	✓
		December 2014~October 2015	11	✓
		July 2019~October 2019	4	✓
		April 2020~May 2023	38	✓
		December 2021	1		✓
		March 2022	1		✓
SGI	9	November 2014~September 2015	11	✓
		January 2015~March 2015	3		✓
		November 2015 ~June 2016	8	✓
		January 2017~March 2017	3	✓
		September 2020~December 2020	4	✓
		March 2022~June 2022	4	✓
		November 2022	1			✓
		December 2022	1		✓
		January 2023. ~May 2023	5	✓

). The annual precipitation since 2020 continued to be deficient, while temperature continued to be high. The combination of low precipitation and high temperature in the period was the fundamental reason for the shortage of water resources and drought in Central Asia. Guo et al. [23] calculated the Standardized Precipitation Evapotranspiration Index (SPEI) in Central Asia from 1966 to 2015 and identified drought events in that period, which were consistent with our results using the GRACE-DSI, GRACE-WSDI, and GRACE-SGI.

Using the GRACE-DSI, GRACE-WSDI, and GRACE-SGI, we analyzed the wet and dry conditions from 2002 to 2023 in Central Asia. Figure 7 demonstrates the change rate of the drought indexes and the MK trend line. Both the GRACE-DSI and GRACE-WSDI showed a similar spatial change, with a decrease in most areas, including Uzbekistan, Kyrgyzstan, Tajikistan, the western regions Turkmenistan, and the southern and western regions of Kazakhstan. The significant reduction trend in these areas highlights the harsh arid conditions and the deterioration trend prevalent in Central Asia. Particularly, the arid condition in western Central Asia, the eastern coast of the Caspian Sea, and the Aral Sea basin had deteriorated severely. Studies argued that the lasting deterioration was mainly ascribed to human interference [25,57]. However, the hydrological drought indicators in northern Kazakhstan and eastern Turkmenistan showed a significant upward trend, indicating that the partial drought condition had been alleviated. Overall, the GRACE-SGI was consistent with the GRACE-DSI and GRACE-WSDI. However, the trend shown by the GRACE-SGI was different from those of the GRACE-DSI and GRACE-WSDI in northwestern Kazakhstan. The GRACE-SGI for the northwest of Kazakhstan increased significantly. This trend indicates that the amount of groundwater resources has been increasing in these regions during the past two decades.

3.5. Correlation Analysis Between Drought Indexes and Vegetation

The correlation coefficients between the growing season (April–September) NDVI of the Central Asian region from 2002 to 2023 and the GRACE drought indexes of the corresponding months were calculated. According to the correlation coefficients (Figure 8), the GRACE drought indexes were sensitive to regional water resources and affected the growth of regional vegetation. The correlation coefficients between the drought indexes and NDVI were high during the growing season, especially in the middle of the growing season (July). Zhao et al. [39] argued that the GRACE-DSI was able to capture the relationship between the vegetation–soil water correlation and the latitude. Central Asia also exhibited this regular distribution. The GRACE-DSI and NDVI in the high latitude regions were mainly positively correlated, while those in low latitudes were mainly negative. The correlation distribution between the GRACE-WSDI and NDVI was similar to that between the GRACE-DSI and NDVI, while the correlation between the GRACE-SGI and NDVI was different from the two above. A comparative analysis found that the correlation coefficient between the GRACE-SGI and NDVI during the growing season was mainly negatively correlated. Notably, it had a strong negative correlation in the cultivated land in northern Central Asia from July to September and in the grassland area in northern Central Asia from June to September, implying a decrease in groundwater resources and growing vegetation in this region. Therefore, vegetation growth in the areas depends not only on precipitation but also on groundwater consumption.

Our results revealed that vegetation exhibited the highest correlation with drought indexes during the growing season in Central Asia (Figure 8), which was consistent with the findings of previous studies [58]. Moreover, our results using the groundwater drought index showed an obviously negative correlation with the NDVI in the growing season, especially in Kazakhstan. Similar findings were reported in the Loess Plateau, China [59]. For cropland and grassland, drought due to groundwater scarcity inhibits vegetation growth (Figure 8). Liu et al. [60] also found that this was the case for groundwater-dependent ecosystems in central Kazakhstan.

4. Discussion

It is beneficial to utilize the GRACE drought indicators to explore the spatiotemporal distributions in a wide range of regions. The relationship derived between the GRACE-DSI, GRACE-WSDI, GRACE-SGI, and SPI (Table 2 and Table 3 and Figure 6) exhibited that the GRACE drought indicators were sensitive to hydrological drought events and groundwater drought. A series of droughts have been identified using the GRACE drought indicators (Table 4). And the drought events are consistent with other findings. For example, the 2008 and 2021 drought events were consistent with previous studies [61,62]. It was reported that 2008 was one of the driest years in Central Asia, particularly in Kyrgyzstan and Tajikistan, which were severely affected by the drought [61,63]. Northern, central, eastern, and southeastern Central Asia was frequently affected by flash droughts, while the western region experienced prolonged regular droughts with an insignificant weakening trend [64]. This study found that the GRACE-DSI, GRACE-WSDI, GRACE-SGI all exhibited significantly decreasing trends, particularly in the western regions, as well as in the eastern regions of Kyrgyzstan and Tajikistan (Figure 7). Based on the GRACE drought index, drought disasters of different degrees and lengths were found to be frequent after 2016, which is also consistent with other research results [65].

The drought had a significant detrimental effect on vegetation, leading to a reduction of approximately 10% in vegetation greenness and around 13% in productivity [62]. The impact of drought on vegetation may contribute to drought mitigation and land degradation measures in Central Asia under accelerating global warming [58]. During the middle of the growing season, the semi-arid areas of Central Asia were the most vulnerable regions, whereas most vegetation in the semi-arid zone was more susceptible to the negative impacts of drought events, particularly in croplands and grasslands, compared to other climate zones [66]. The effects of flash droughts on vegetation were more immediate in the comparatively arid region of western Central Asia [64]. There was a strong positive correlation between the GRACE-DSI, GRACE-WSDI, and NDVI and a strong negative correlation between the GRACE-SGI and NDVI during the growing season, especially in the middle of the growing season (July) (Figure 8). Vegetation growth in Central Asia depends not only on precipitation but also on groundwater consumption. Drought events in Central Asia are becoming increasingly frequent, causing further harm to the vegetation in this region.

5. Conclusions

Utilizing the data from the GRACE gravity satellites, GLDAS data, ERA5 data and MODIS NDVI, this study investigates hydrological drought and groundwater drought in Central Asia, while also exploring the spatiotemporal characteristics of the arid environment and its impact on the surface ecological environment. The principal findings are summarized as follows:

(1) Since 2000, Central Asia has experienced a trend of warming and drying. Annual precipitation has decreased marginally at a rate of 1.80 mm/year (p < 0.1), while temperature has increased slightly, at a rate of 0.008 °C/year (p = 0.62). Water storage has declined at a rate of −3.53 mm/year (p < 0.001), exhibiting an increase–decrease–increase–decrease process.

(2) The hydrological drought indicators derived from the GRACE data demonstrate a strong correlation with the traditional drought indicator SPI, indicating their effectiveness in assessing long-term comprehensive drought events.

(3) Since 2020, Central Asia has been classified as arid, experiencing varying degrees of hydrological and groundwater drought. The arid conditions have intensified in Uzbekistan, Kyrgyzstan, Tajikistan, western Kazakhstan, and western Turkmenistan since 2002, with particularly severe deterioration observed in the western regions of Central Asia and the Aral Sea basin.

(4) The arid environment adversely impacts vegetation growth, as plants in these regions rely heavily on groundwater resources. During the growing season, a significant correlation was found between the GRACE indexes and the NDVI, while cultivated land and grassland areas in Central Asia displayed a significantly negative correlation with the GRACE groundwater drought index.

Given the socio-economic development and the promotion of ecological civilization, there is an urgent need for effective drought prevention and mitigation strategies. Consequently, enhancing drought monitoring, prediction, and assessment is crucial for mitigating the impact of droughts. This paper quantitatively assessed drought dynamics in Central Asia from April 2002 to May 2023 through the analysis of GRACE data. The investigation into the spatiotemporal distribution of drought events and the vegetation’s response to drought provides valuable insights for understanding and adapting to the region’s ecological environment. The results of this study enhance the comprehension of droughts’ effects on ecosystems and should inform the formulation of strategies aimed at mitigating and adapting to drought events, thereby reducing their adverse impact on ecosystems.