Anthropogenic and Climate-Driven Water Storage Variations on the Mongolian Plateau

Abstract

Evaluating the variations in terrestrial water storage anomalies (TWSA) associated with climate forcing and human activities in the Mongolian Plateau is crucial for assessing water scarcity and predicting potential pressures on water resources in the future. In this study, we assessed the impacts of climatic and anthropogenic drivers on the change in TWSA on the Mongolian Plateau by using the Independent Component Analysis (ICA) to examine Gravity Recovery and Climate Experiment (GRACE) based TWSA data and comparing the ICA modes with hydrometeorological data and statistical data related to human activities. The results showed that TWSA in the Mongolian Plateau has experienced significant depletion (−2.3 ± 0.62 mm/year) from 2002 to 2017, with a severe decline (−3.4 ± 0.78 mm/year) in Inner Mongolia, China, and a moderate depletion rate in Mongolia (1.44 ± 0.56 mm/year). The results of the statistical analysis indicated that climate change was the dominant driver for the decline in TWSA from 2002 to 2007, resulting in a decrease in TWSA in Mongolia and Inner Mongolia at rates of −5.17 ± 1.13 mm/year and −5.01 ± 2.0 mm/year, respectively. From 2008 to 2017, the intensity of human activities has increased in Mongolia, but climate-driven effects greatly offset the anthropogenic changes, leading to an increasing trend in TWSA in Mongolia during this period. Nevertheless, in Inner Mongolia, the anthropogenic water depletion, such as irrigation, coal mining, and grazing, outweighed the climate contributions on the variations in TWSA, causing the TWSA in Inner Mongolia to decline at a rate of 1.08 ± 0.54 mm/year during 2007–2011.

1. Introduction

The Mongolian Plateau (MP) covers an area of ~2,000,000 km², and administratively includes two parts: Inner Mongolia and Mongolia. The MP adjoins the east of the Eurasian grassland. Water security is a key issue in this plateau due to the global climate change situation. Intensive human activities and climate change have exacerbated the problem, including overexploitation of groundwater, grassland degradation and shrinking lakes in the Mongolian Plateau [1,2]. The plateau has become one of the most significant sources of sandstorms in China. These sandstorms pose a threat to agricultural productivity and the health and livelihoods of billions of people [3,4,5,6]. Therefore, it is essential to investigate the changes in terrestrial water storage anomalies (TWSA) and to explore the drivers causing such changes in the MP region in order to help the rational utilization and effective management of regional water resources and support environmentally sustainable development in this region. In-situ measurements are a feasible means to estimate the fine-scale local water storage changes. However, it is a significant challenge to accurately obtain regional water storage changes from in-situ measurements due to the poor spatiotemporal coverage [7]. Global hydrological models, amalgamating atmospheric forcing factors and terrestrial characteristics with some in situ observations, are utilized to calculate and holistically comprehend individual variables in water storage, encompassing groundwater and evapotranspiration [8]. These models offer the means to scrutinize and discern intricate interrelationships amongst and transitions of global hydrological signals. Nonetheless, Land Surface Models (LSMs) fall short of offering exhaustive information of TWSA, including surface water, groundwater storage, and anthropogenic effect of water storage, which are often overlooked [9,10].

The Gravity Recovery and Climate Experiment (GRACE) twin-satellite mission, which launched on 17 March 2002, has provided an unprecedented opportunity to quantitatively estimate the spatial and temporal variation of TWS in the hydrosphere, cryosphere, and oceans at spatial resolutions of approximately 300 km and more through measured tiny variations in the distance between two twin satellites [11,12]. Many studies have applied GRACE data to quantify the changes in water mass over scales ranging from river basins to the global scale, and to provide a new pathway to evaluate the droughts and floods that related to the significant decrease/increase in GRACE TWS [13,14,15].

Generally, the TWSA estimate provided by GRACE is comprehensively affected by human activities and climate change, and climate change may cause us to underestimate or overestimate the impact of human activities on TWSA in different periods [16,17,18,19,20,21,22,23,24,25]. It is difficult to isolate the impacts of climate change and human activities on the TWSA derived from GRACE observations.

Many researchers have attempted to assess the relative contribution of natural and anthropogenic effects on TWSA [1,21,23,26,27,28]. For example, Felfelani et al. (2017) isolated natural and human-induced changes on TWSA by comparing results from different hydrological model simulations and GRACE data over large basins. Yi et al. (2016) analyzed the impact of human activities and climate driving on TWSA by establishing a linear relationship between variation in precipitation and water storage, assuming that the impact of human activities on TWSA (HWSA) is constant over time. Zhong et al. (2019) quantified the contribution of climate-driven TWSA (CWSA) and its long-term trends based on the reconstruction of long-term climate-driven water storage variations. They also evaluated the HWSA by removing CWSA from TWSA. These studies were primarily conducted based on hydrological models or established statistical models that simulate natural dynamically driven changes in water storage to reconstruct climate-driven water storage anomalies. However, due to the uncertainty of state-of-art hydrological models and forcing datasets, the impact of human intervention on TWSA may be underestimated or overestimated [10].

The primary objective of this paper is to explore the potential driving factors for the change in TWSA. To do so, we conduct a comprehensive analysis of the TWSA in the Mongolian Plateau based on independent component analysis (ICA). First, we investigate the water storage changes in the MP from 2002 to 2017, then we decompose the TWSA using the ICA technique to illuminate the spatiotemporal patterns of water storage variations in MP. Finally, we compare the ICA modes with hydrological model data, ground statistics data (coal mining, irrigation, grazing, etc.), and simulated CWSA products derived from the statistical model proposed by Humphrey [8] to explain the meaningful of each dominant signal that separate through ICA results. This will improve understanding of the various factors affecting basin water storage, including the anthropogenic and climate-driven water storage variations.

2. Materials and Methods

2.1. Data

2.1.1. GRACE Data

The most recent GRACE mascon solutions (from April 2002 through June 2017), provided by the Jet Propulsion Laboratory (JPL), were used in our study [29]. The JPL mascon data were provided at 0.5 degrees at http://grace.jpl.nasa.gov/ (accessed on 1 March 2022), and all the appropriate corrections (C20, Degree-1, and Glacial Isostatic Adjustment (GIA)) were applied to the JPL GRACE mascon solutions [30,31,32]. The mean of the GRACE mascon solutions (from January 2004 through December 2009) is removed in order to obtain the regional TWSA.

2.1.2. Precipitation Data

Within the hydrological cycle, precipitation is the main component of water transport from the atmosphere to Earth’s surface. Precipitation varies strongly, and depends on geographical location, season, synopsis, and other meteorological factors. The supply of freshwater through precipitation is vital for many subsystems of the climate and the environment, but there are hazards related to extensive precipitation or the lack of precipitation [33]. Daily total precipitation data were taken from the ERA5 atmospheric reanalysis for the period 1979 to 2017. Reanalysis combines model data with observations (e.g., radiosondes, satellite observations, and buoys) from across the world into a globally complete and consistent dataset using the laws of physics.

2.1.3. Hydrological Model Simulations

Two hydrological models are used in our research: the Global Land Data Assimilation System (GLDAS) and the WaterGAP Global Hydrology Model (WGHM).

The GLDAS-Noah2.1 hydrological model [9], provided by NASA Goddard Space Flight Center (GSFC) and NOAA National Centers for Environmental Prediction, is used in our study. The GLDAS-Noah2.1 includes 4-layer soil moisture at depths of 0–200 cm, snow water equivalent, and canopy water storage at $1^{°}\times 1^{°}$ rectangular grids over global land area. Surface water and groundwater are not included in the GLDAS-Noah2.1 hydrological model.

The WGHM [34,35] simulates the terrestrial water cycle by using conceptual formulas for major hydrological processes, including fast surface and subsurface runoff, groundwater recharge, and river discharge, in order to track water storage variations in the canopy, snow, surface water, soil, groundwater, lakes, man-made reservoirs, wetlands, and rivers as a function of climate. Soil, land cover, relief, and observed river discharge, human activity processes, including the impact of human water use on river discharge, surface water, and groundwater storage for all land areas of the globe excluding Antarctica with a spatial resolution of ${0.5}^{°}\times {0.5}^{°}$ from 1979 to 2016 [36,37]. Anomalous fields are obtained by subtracting out the multiyear (from January 2004 through December 2009) mean field.

2.1.4. Data of Human Activities

We obtained grazing data (number of sheep and goats, 1979–2017), mining data (coal production, 1979–2017), and agricultural data (area of irrigated croplands from 1979 to 2017), with a temporal resolution of 1 year, documented from the Inner Mongolia Statistical Yearbook [38]. For Mongolia, those data were obtained from the Food and Agriculture Organization of the United Nations database (faostat.fao.org) and the United States Energy Information Administration (www.indexmundi.com/ (accessed on 1 March 2022).

2.2. Methodologies

2.2.1. TWSA and GWSA Estimation

The TWSA includes the surface water storage anomaly (SWSA, which mainly includes changes in lakes), soil moisture anomaly (SMA), groundwater storage anomaly (GWSA), canopy water storage anomaly (CWA), and snow water storage anomaly (SnWSA). These components satisfy the equation:

(1)$\mathrm{TWSA}=\mathrm{SWSA}+\mathrm{SMA}+\mathrm{GWSA}+\mathrm{CWA}+\mathrm{SnWSA}$

In this study, we obtained the GWSA by subtracting the SMA from the TWSA. We ignored the contributions of CWA and SnWSA to the TWSA, which is rather small in the MP.

2.2.2. Reconstruction of Climate−Driven Water Storage Anomalies

We reconstructed the CWSA in Mongolia and Inner Mongolia, based on GRACE products, precipitation, and temperature data, by using the data-driven statistical method proposed by Humphrey (CWSA_Humphrey) [18,19]. The CWSA_Humphrey is formulated as follows:

(2)${\mathrm{TWSA}}_{\mathrm{REC}}={\mathsf{\beta }}_1·{\mathrm{P}}_{\mathrm{inter}+\mathrm{subseas}}^{\mathsf{\tau }}+{\mathsf{\beta }}_2·{\mathrm{T}}_{\mathrm{inter}}+\mathsf{\epsilon }$

(3)${\mathrm{TWSA}}_{\mathrm{Humphrey}}={\mathrm{TWSA}}_{\mathrm{REC}}+{\mathrm{TWSA}}_{\mathrm{seasonal}}$

where CWSA_REC is the reconstructed TWSA without the trend and seasonal signals, and ${\mathrm{P}}_{\mathrm{inter}+\mathrm{subseas}}^{\mathsf{\tau }}$ is the deseasonal and detrended precipitation data. The decay parameter τ controls the steepness of the exponential decay filter applied to the daily precipitation time series before averaging to the monthly resolution. ${\mathrm{T}}_{\mathrm{inter}}$ corresponds to the monthly interannual temperature, parameters ${\mathsf{\beta }}_1$ and ${\mathsf{\beta }}_2$ correspond to calibrated scaling coefficients, and $\mathsf{\epsilon }$ denotes an error term. ${\mathrm{TWSA}}_{\mathrm{seasonal}}$ is the GRACE-estimated seasonal cycle.

2.2.3. ICA

To better illuminate the changes in complex hydrological signals, independent component analysis (ICA) was used to identify the main spatiotemporal variation characteristics of GRACE TWSA data.

Suppose there are m unknown source signals that form a column vector, $\mathrm{S}\left(\mathrm{t}\right)={[S_1\left(t_1\right),S_2(t_2)\cdots S_n(t_m\left)\right]}^T$, and A is an unknown mixing matrix representing the spatial distribution of the source signals [39,40,41]. Assuming n observation channels (in this case, TWSA time series at m grids in Mongolia/Inner Mongolia), the ICA can be formulated as follows:

(4)$\left[\begin{array}{ccc}{x}_1\left(t_1\right)& \cdots & x_1\left(t_m\right)\\ ⋮& \ddots & ⋮\\ x_n\left(t_1\right)& \cdots & x_n\left(t_m\right)\end{array}\right]=\mathrm{A}\left[\begin{array}{ccc}{s}_1\left(t_1\right)& \cdots & s_1\left(t_m\right)\\ ⋮& \ddots & ⋮\\ s_n\left(t_1\right)& \cdots & s_n\left(t_m\right)\end{array}\right]=\mathrm{A}\mathrm{S}(\mathrm{t})$

The goal of ICA is to estimate both the source signals S(t) and the mixing matrix A from the observed data X(t). ICA makes use of statistical independence as a criterion for separating the source signals.

The idea of ICA is to set up a linear decomposition matrix W, where X is transformed by W to obtain an n-dimensional random variable sequence $\mathrm{Y}\left(\mathrm{t}\right)={[y_1\left(t\right),y_1\left(t\right),\cdots y_n(t\left)\right]}^T$:

(5)$\mathrm{Y}\left(\mathrm{t}\right)=\mathrm{W}\mathrm{X}\left(\mathrm{t}\right)=\mathrm{W}\mathrm{A}\mathrm{S}\left(\mathrm{t}\right)$

ICA obtains the source signal s by optimizing the decomposition matrix W so that the independence between the source signals is strongest. When the best decomposition matrix W is determined, the mathematically defined Y(t) are referred to as independent components (ICs). In the case of the Mongolia/Inner Mongolia, ICA can be used to separate the source signals of TWSA. By comparing the source signals of the TWSA with the other hydrological products, it is possible to quantify the contributions of different drive factors to the changes in TWSA over time.

3. Results

3.1. Water Storage Changes in the MP from 2002 to 2017

Figure 1 shows the water storage variations of TWSA, SMA, and GWSA in the MP. There is a significant decline in TWSA with a rate of 2.3 ± 0.62 mm/year in the MP over the period 2002–2017. TWSA showed a decrease at a rate of 0.3 ± 0.15 mm/year during 2008–2017; this decrease is relatively smooth in comparison to the changes from 2002–2007, which showed a sharp decrease at a rate of 7.3 ± 2.4 mm/year.

As shown in Figure 1, the interannual changes in water storage in MP are mainly dominated by precipitation variation. Before 1998, the mean annual precipitation is above the average value. However, after 1998, there is significant decline in precipitation over the MP, especially during 2002–2007 (Figure 1a). For the same period, the MP experienced a rapid decline in TWSA (−6.1 ± 2.1 mm/year), and the decrease rates of SMA and GWSA were 2.16 mm/year and 3.9 mm/year, respectively. Two extreme droughts occurred in 2007 and 2011 in the MP; these were primarily caused by abnormal precipitation in that period.

We examined the variations in the TWSA of Mongolia and Inner Mongolia (Figure 2). The long-term changes of TWSA in Mongolia and Inner Mongolia shows significant differences during variable periods. For example, during 2002–2017, the linear trend of TWSA in Mongolia was −1.44 ± 0.56 mm/year (Figure 2A), and the TWSA decline rate of Inner Mongolia was 3.4 ± 0.78 mm/year (Figure 2B), which was higher than that in Mongolia.

From 2002–2007, Inner Mongolia and Mongolia showed a consistent decrease of TWSA, with a rate of 7.48 ± 2.5 mm/year and 5.3 ± 1.4 mm/year, respectively. However, during the period of 2008–2017, Mongolia shows positive TWSA trends (1.1 ± 0.4 mm/year), and a significant TWS depletion trend of −0.76 ± 0.12 mm/year was observed over the Inner Mongolia. The GWSA and TWSA showed consistent decline trend over each of the investigated region (Inner Mongolia and Mongolia) during the entire study period, but the SMSA experienced a nonsignificant trend (Figure 2).

3.2. TWSA Signal Decomposition Using ICA

We performed ICA on the TWSA and obtained two leading modes (the trend mode and the annual mode) in Mongolia and Inner Mongolia. Since the ranking of the components obtained by ICA corresponds to a random ordering instead of a reduction in variance, we calculate the average ratio of the contribution of each IC to observed unfiltered time series in order to reorder ICA components in ascending order, and the spatial portion of each ICA results are normalized by their maximum value (absolute value).

In Mongolia, the leading mode (IC1_M) is the annual mode, which explains 75.5% of the total variance of the filtered TWSA time series, followed by the trend mode that explains 14.6% of the total variance. As shown in Figure 3, the spatial weight of IC1_M shows a positive contribution to the corresponding time function in northern Mongolia and the spatial pattern of the IC1_M model shows a gradual increase in water storage toward the northwest. The spatial weights of IC2_M reflect the transition of the TWSA from northwest to southeast, with positive responses in central Mongolia and negative responses on the northwest tip and southeast side of Mongolia.

In Inner Mongolia, the first IC (IC1_IM) explains 73.46% of the total variance of TWSA, and the first two components make up 88.95%. The spatial patterns of IC1_IM and IC2_IM show the difference between east and west (Figure 3), which may be affected by the spatial distribution of land type (Figure S1), population density (Figure S2), and precipitation (Figure S3). In Inner Mongolia, there is a relatively high annual precipitation in the northeast region; in the southwest desert region, the annual precipitation is lower (Figure S3), which is consistent with the spatial distribution of IC2_IM. However, in some regions, the spatial pattern of IC (IC1_M and IC2_IM) and precipitation distribution exhibit systematic spatial variability that could potentially be related to that IC2_IM and IC1_M containing controlling factors, as well as snow, temperature and precipitation.

The IC2_M and IC1_IM represent a mostly linear trend that is representative of the TWSA. In Inner Mongolia, this mode occupies the main part, while in Mongolia, we see the opposite situation.

3.3. Comparison of the ICA Modes with Simulated CWSA and HWSA Products

The IC1_M and IC2_IM show the seasonal fluctuation in TWSA. Our results are generally consistent with the Humphrey result (TWSA_Humphrey) [18], which are affected by climate variability, with R² values of 0.88 (Mongolia) and 0.89 (Inner Mongolia) (Figure 4). This suggests that the CWSA in Mongolia and Inner Mongolia can be explained by the IC1_M and IC2_IM, respectively, and that ICA may provide a new method to isolate the impacts of climate variability and human activities on water storage variation. The seasonal fluctuation, with an averaged amplitude of CWSA derived from GRACE based on ICA, is 4.0 mm (Mongolia) and 4.3 mm (Inner Mongolia). In the GLDAS model, the maximum seasonal water mass changes are 2.9 mm in Mongolia and 1.9 mm in Inner Mongolia, and both are smaller than the CWSA derived from the GRACE based on the ICA. This discrepancy can be explained by different observation techniques and modeling approaches. The CWSA derived from GRACE, based on ICA, integrates all surface and subsurface water resources, whereas the GLDAS hydrological model only contains the soil moisture estimates from the surface to a depth of 2 m, snow water equivalent, and a canopy water component, without simulating the contributions of groundwater and reservoir water.

Mongolia and Inner Mongolia have experienced similar drying and warming climates [2], and the impacts of climate change in TWSA have good consistency. During 2002–2007, the climate-driven water storage loss rates in Mongolia and Inner Mongolia were 5.17 ± 1.13 mm/year and 5.01 ± 2.0 mm/year, respectively; this came in response to the precipitation reduction, which indicated that precipitation was the dominant driving force for the terrestrial water depletion in the two regions (Figure 4). However, the situation is markedly divergent between 2008 and 2017. The rainfall in Mongolia and Inner Mongolia increased, which caused the CWSA to increase in both regions at a rate of 2.67 ± 0.50 mm/year and 2.10 ± 0.96 mm/year, respectively.

WGHM was run with and without human intervention to further isolate the effects of human intervention on TWSA (HWSA_WGHM). In Inner Mongolia, the human-driven water storage estimated by WGHM, loss with a rate of −1.50 mm/year from 2002 to 2017, is consistent with the results of statistical data in the Yearbook but less than our estimation (−2.40 ± 0.24 mm/year) (Figure 5). In Mongolia, the trend of TWSA estimated from WGHM was near zero, which is smaller than that extracted from the IC2_IM (−0.97 ± 0.13 mm/year). The primary causes of these discrepancies are that the effects of human intervention on the TWSA, calculated by the WGHM model, are mainly based on irrigation data in Inner Mongolia, while other components, such as grazing and industrial water use, are not considered. Scanlon et al. (2018) also suggest that global models underestimate large decadal declining and rising water storage trends relative to GRACE satellite data. In these two regions, the conclusion was verified again. Therefore, IC2_M and IC1_IM may represent HWSA in Mongolia and Inner Mongolia, respectively.

4. Discussion

4.1. Human and Climate-Driven Water Storage Anomalies

From 2002 to 2017, both Mongolia and Inner Mongolia have experienced significant increases in mining, irrigation, and grazing intensity relative to pre–2002 activity (Figure 6). Overgrazing caused grassland degradation and resulted in a decrease in soil moisture retention capacity and groundwater depletion in Mongolia and Inner Mongolia [2].

In Inner Mongolia, huge irrigation expansions (1.18 million hectares (ha) in 1979 and 3.2 million hectares in 2017) have caused extensive groundwater withdrawals. For example, the groundwater depth in TongLiao city in southeastern Inner Mongolia decreased from 2.1 m in 1980 to 5.8 m in 2017 [2,23] (Figure S4).

Since the late 1990s, coal mining has been widely carried out in Inner Mongolia, especially in grassland areas [2]. The number of mining companies increased rapidly, from 156 in 2000 to 371 in 2017 (Inner Mongolia Statistical Yearbook), and coal production has increased from 22 Tg to 905 Tg (1 Tg = 10¹² g) (Figure 6A). In 2016, Inner Mongolia produced 838 million tons of coal, which accounted for 24.9% of the national total coal production (3.364 billion tons), and Inner Mongolia was the largest coal production province in China (China Statistical Yearbook). Coal mining is an extremely water intensive industry; each ton of coal mined consumes 2.54 m³ of water, which may cut off rivers and destroy underground aquifers [2,25].

Compared to the rapid expansion of farmland in Inner Mongolia since 1990, the agricultural area in Mongolia has been kept stable, and no obvious trend can be observed (Figure 6B); due to the sparse distribution of farmland in Mongolia, the impact of irrigation on GWSA can be negligible. In addition, the difference in population density between Inner Mongolia (20.3 people/km²) and Mongolia (1.92 people/km²) [2] might lead to a huge difference in the human intervention on water storage variation (Figure S2). The rate of increase in grazing numbers was also more rapid in Inner Mongolia than in Mongolia (Figure 6C).

In Inner Mongolia, the consumption of TWSA, calculated according to the Yearbook, accounts for 61% of IC1_IM; this is consistent with the result reported in the official water resources bulletin, which states that irrigation water accounts for approximately 55–65% of human water consumption. Figure 1b shows the temporal evolution of the TWSA exhibits two stages (2002–2007, 2008–2017) in the MP.

During 2002–2007, the TWSA in both Mongolia and Inner Mongolia showed a continuous decrease, with rates of 5.28 ± 1.7 mm/year and 7.48 ± 2.5 mm/year, respectively (Table 1). The decline rate of TWSA in Inner Mongolia was higher than that in Mongolia during this period, which can be attributed to the intensive human activities in Inner Mongolia. From 2008 to 2017, the CWSA, derived from GRACE, and based on ICA, showed an increasing trend in both Mongolia and Inner Mongolia, with rates of 2.67 ± 0.50 mm/year and 2.10 ± 0.96 mm/year, respectively. However, as a result of the human activity intensity increasing during this period, the HWSA decreased at rates of 1.64 ± 0.32 mm/year and 3.12 ± 0.57 mm/year (Table 1), respectively. In Mongolia, TWSA exhibits positive trends (1.12 ± 0.80 mm/year) over the investigated period (2008–2017), while Inner Mongolia experienced a significant negative TWSA.

The above analysis results indicate that although Inner Mongolia and Mongolia have experienced similar dry and wet climate scenarios, the TWSA trends in the two regions are different in different periods due to significant differences in human activity intensity.

4.2. Perspective

The underground mineral deposits of the MP are abundant, and in recent years, a significant number of mineral resources have been explored and mined. Mineral resources development has become the largest industry in Mongolia [42,43,44]. Moreover, China and Mongolia have been continually developing coal mining cooperation and Mongolia’s coal exports to China have skyrocketed, and it has become China’s largest coal supplier [2,44].

In wet years, the impact of human activities on land water storage may be offset by climate change; in dry years, intensive human activities may lead to further decrease in TWSA in the MP, which will exacerbate drought and cause several ecological problems, including a sharp decrease in surface runoff, shrinking lakes (Figure S5), and intensifying desertification. In recent years, the MP has also emerged as a significant source of sandstorms in northern China [3,4,45]. This is due to the soil moisture, the dominant components of land total water storage, and is one of key factors leading to sandstorms being produced. It is crucial to take effective measures to protect these precious land water resources and prevent their further deterioration.

This study provides a potential method to quantify the climate- and human-driven water storage variations in the MP. It is of great significance for helping people better understand the evolution of TWSA under the combined influence of climate change and anthropogenic activities and provides information for better protection and utilization of regional water resources. It also will help to facilitate the regional sustainable management and protection of water resources.

5. Conclusions

During 2002–2017, Mongolia and Inner Mongolia have experienced similar drought and warming climate scenarios, however, attributed to the combined effects of climate variability and intensive human activities, the changes in TWSA behaved significant differently in these two regions; TWSA in Mongolia decreased slightly at a rate of 1.44 mm/year, while the TWSA in Inner Mongolia decreased rapidly at a rate of 3.4 mm/year.

The temporal evolution of TWSA exhibits two stages in responding to precipitation fluctuation. During 2002–2007, climate change was the dominant driver for TWSA variation in the Mongolian Plateau. From 2008 to 2017, intensive human activities in Mongolia increased water consumption, however, climate driven increase in precipitation mitigated the anthropogenic depletion of TWSA, resulting to TWSA increase during this period consequently; therefore, climate change was the leading factor for the TWSA change. In contrast, human activities were the primary driving factor of TWSA in Inner Mongolia.

Under the background of intensive climate warming and human activities, quantifying water storage dynamic process is a key issue for ecologically fragile areas such as the Mongolian Plateau, this work implicated that it is possible to access the impact of climate change and human activities on TWSA changes through ICA method and multi-sources datasets. It can help manage and protect water resources and provide possible solutions for sustainable ecological and socio-economic development.

Author Contributions

Conceptualization, S.Z. and Z.Z.; methodology, S.Z.; software, S.Z.; validation, S.Z. and Z.Z.; formal analysis, S.Z.; investigation, S.Z.; resources, S.Z.; data curation, S.Z.; writing–original draft preparation, S.Z.; writing–review and editing, S.Z., Z.Z., Y.L., H.Y. and Z.S.; visualization, S.Z. and Y.L.; supervision, Z.Z.; project administration, Z.Z.; funding acquisition, Z.Z. and H.Y. All authors have read and agreed to the published version of the manuscript.

Data Availability Statement

The GRACE mascon solutions are downloaded from the Jet Propulsion Laboratory (JPL). Precipitation data is available at CMA. GLDAS data is available at https://disc.gsfc.nasa.gov/ (accessed on 1 March 2022).

Conflicts of Interest

The authors declare no conflict of interest.

Footnotes

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.

Figures and Table

Figure 1. Precipitation anomalies during 1979–2017 and TWSA during 2002–2017 on the Mongolian Plateau. (a) the precipitation anomalies deviate from the average over the whole period. The red curve in the upper plot is smoothed by 3 years. (b) the location of the Mongolian Plateau (red area) and its main topographic features. (c) changes in TWSA and its major components.

Figure 2. The area averaged TWSA derived from GRACE data, the SMA simulated by GLDAS models and estimated GWSA in Mongolia (A) and in Inner Mongolia (B).

Figure 3. (Top 2) the top two ICs of ICA solution for TWSA time series in Mongolia. (Bottom 2) The two ICs of the ICA solution for the TWSA time series in Inner Mongolia. The time series at the top of each panel shows the scaled IC of each component. The maps at the bottom show the normalized spatial weight. Multiplying the map percentage value of each grid with the time series produces the contribution of that IC to the TWSA.

Figure 4. Time series of precipitation and water storage changes in Mongolia and Inner Mongolia. (A) Comparison of annual precipitation anomaly (cyan bar), CWSAHumphrey (black curve), and TWSAGLDAS (blue curve) in Mongolia with IC1M (red curve). (B) Comparison of annual precipitation anomaly (cyan bar), CWSAHumphrey (black curve), and TWSAGLDAS (blue curve) in Inner Mongolia with IC2IM (red curve).

Figure 5. (A) comparison of monthly IC1M and HWSAWGHM time series in Mongolia. (B) comparison of monthly IC2IM and HWSAWGHM time series in Inner Mongolia. The red and cyan lines show the linear trend.

Figure 6. Changes in human activities in Inner Mongolia and Mongolia in the past decades. (A) mining production, (B) area of irrigated croplands, and (C) grazing in term of number of sheep and goats.

Table 1

Estimates of GRACE-derived, climate-driven, and human-induced TWSA trends in Mongolian and Inner Mongolia from 2002 to 2017 (unit: mm/year).

	Mongolia(2002–2007/2008–2017)	Inner Mongolia(2002–2007/2008–2017)
TWSA(GRACE)	5.28 ± 1.70/1.12 ± 0.80	−7.48 ± 2.5/−1.08 ± 0.54
CWSA(IC1M/IC2IM)	5.17 ± 1.13/2.67 ± 0.50	−5.01 ± 2.0/2.10 ± 0.96
HWSA(IC2M/IC1IM)	−0.11 ± 0.60/−1.64 ± 0.32	−1.65 ± 0.78/−3.12 ± 0.57
CWSAhumphrey	−2.66 ± 1.62/−1.87 ± 0.68	3.33 ± 2.26/3.41 ± 1.16

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/rs15174184/s1, Figure S1. Land use distribution map with spatial resolution of 30 m in Inner Mongolia in 2017; Figure S2. Population density estimates in Mongolian Plateau in 2020; Figure S3. Annual amplitude of precipitation from 2002 to 2017 over the Mongolian Plateau; Figure S4. Changes in groundwater depth in Tongliao City in southeastern Inner Mongolia; Figure S5. Lake level changes in the Tigris-Euphrates basin (red curves) and the change in TWSA in MP.

References

1. Huang, Y.; Salama, M.S.; Krol, M.S.; Su, Z.; Hoekstra, A.Y.; Zeng, Y.; Zhou, Y. Estimation of human-induced changes in terrestrial water storage through integration of GRACE satellite detection and hydrological modeling: A case study of the Yangtze River basin. Water Resour. Res.; 2015; 51, pp. 8494-8516. [DOI: https://dx.doi.org/10.1002/2015WR016923]

2. Tao, S.; Fang, J.; Zhao, X.; Zhao, S.; Shen, H.; Hu, H.; Tang, Z.; Wang, Z.; Guo, Q. Rapid loss of lakes on the Mongolian Plateau. Proc. Natl. Acad. Sci. USA; 2015; 112, pp. 2281-2286. [DOI: https://dx.doi.org/10.1073/pnas.1411748112]

3. Xia, X.; Yang, G. Dust Storms and its Control in Northwest China; Chinese Environmental Press: Beijing, China, 1996.

4. Wang, X.; Dong, Z.; Zhang, J.; Liu, L. Modern dust storms in China: An overview. J. Arid. Environ.; 2004; 58, pp. 559-574. [DOI: https://dx.doi.org/10.1016/j.jaridenv.2003.11.009]

5. Wada, Y.; Beek, L.; Kempen, C.; Reckman, J.; Vasak, S.; Bierkens, M. Global depletio-n of groundwater resources. Geophys. Res. Lett.; 2010; 37, 44571. [DOI: https://dx.doi.org/10.1029/2010GL044571]

6. Gao, S.; Tang, Y.; Tang, K. Groundwater vulnerability assessment in Tongliao Plain, Inner Mongolia. J. China Inst. Water Resour. Hydropower Res.; 2015; 13, pp. 261-269.

7. Nalbantis, I.; Tsakiris, G. Assessment of Hydrological Drought Revisited. Water Resour. Manag.; 2009; 23, pp. 881-897. [DOI: https://dx.doi.org/10.1007/s11269-008-9305-1]

8. Sun, Z.; Long, D.; Yang, W.; Li, X.; Pan, Y. Reconstruction of GRACE Data on Changes in Total Water Storage Over the Global Land Surface and 60 Basins. Water Resour. Res.; 2020; 56, e2019WR026250. [DOI: https://dx.doi.org/10.1029/2019WR026250]

9. Rodell, M.; Houser, P.R.; Jambor, U.; Gottschalck, J.; Mitchell, K.; Meng, C.J.; Arsenault, K.; Cosgrove, B.; Radakovich, J.; Bosilovich, M. et al. The global land data assimilation system. Bull. Am. Meteorol. Soc.; 2004; 85, pp. 381-394. [DOI: https://dx.doi.org/10.1175/BAMS-85-3-381]

10. Scanlon, B.R.; Zhang, Z.; Save, H.; Sun, A.Y.; Schmied, H.M.; van Beek, L.P.H.; Wiese, D.N.; Wada, Y.; Long, D.; Reedy, R.C. et al. Global models underestimate large decadal declining and rising water storage trends relative to GRACE satellite data. Proc. Natl. Acad. Sci. USA; 2018; 115, pp. E1080-E1089. [DOI: https://dx.doi.org/10.1073/pnas.1704665115] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29358394]

11. Tapley, B.D.; Bettadpur, S.; Ries, J.C.; Thompson, P.F.; Watkins, M.M. GRACE Measurements of Mass Variability in the Earth System. Science; 2004; 305, pp. 503-505. [DOI: https://dx.doi.org/10.1126/science.1099192]

12. Landerer, F.W.; Swenson, S.C. Accuracy of scaled GRACE terrestrial water storage estimates. Water Resour. Res.; 2012; 48, 4531. [DOI: https://dx.doi.org/10.1029/2011WR011453]

13. Scanlon, B.R.; Zhang, Z.; Rateb, A.; Sun, A.; Wiese, D.; Save, H.; Beaudoing, H.; Lo, M.H.; Müller-Schmied, H.; Döll, P. et al. Tracking Seasonal Fluctuations in Land Water Storage Using Global Models and GRACE Satellites. Geophys. Res. Lett.; 2019; 46, pp. 5254-5264. [DOI: https://dx.doi.org/10.1029/2018GL081836]

14. Tapley, B.D.; Watkins, M.M.; Flechtner, F.; Reigber, C.; Bettadpur, S.; Rodell, M.; Sasgen, I.; Famiglietti, J.S.; Landerer, F.W.; Chambers, D.P. et al. Contributions of GRACE to understanding climate change. Nat. Clim. Change; 2019; 9, pp. 358-369. [DOI: https://dx.doi.org/10.1038/s41558-019-0456-2]

15. Othman, A.; Abdelrady, A.; Mohamed, A. Monitoring Mass Variations in Iraq Using Time-Variable Gravity Data. Remote Sens.; 2022; 14, 3346. [DOI: https://dx.doi.org/10.3390/rs14143346]

16. Chen, J.; Jin, L.; Zhang, Z.; Ni, S. Long-term groundwater variations in Northwest In-dia from satellite gravity measurements. Glob. Planet. Change; 2014; 116, pp. 130-138. [DOI: https://dx.doi.org/10.1016/j.gloplacha.2014.02.007]

17. Chen, J.L.; Wilson, C.R.; Tapley, B.D.; Scanlon, B.; Guentner, A. Long-term groundwater storage change in Victoria, Australia from satellite gravity and in situ observations. Glob. Planet. Change; 2016; 139, pp. 56-65. [DOI: https://dx.doi.org/10.1016/j.gloplacha.2016.01.002]

18. Humphrey, V.; Gudmundsson, L.; Seneviratne, S.I. A global reconstruction of climate-driven subdecadal water storage variability. Geophys. Res. Lett.; 2017; 44, pp. 2300-2309. [DOI: https://dx.doi.org/10.1002/2017GL072564]

19. Humphrey, V.; Gudmundsson, L. GRACE-REC: A reconstruction of climate-driven wate-r storage changes over the last century. Earth Syst. Sci. Data; 2019; 11, pp. 1153-1170. [DOI: https://dx.doi.org/10.5194/essd-11-1153-2019]

20. Panda, D.K.; Wahr, J. Spatiotemporal evolution of water storage changes in India from the updated GRACE-derived gravity records. Water Resour. Res.; 2000; 52, pp. 135-149. [DOI: https://dx.doi.org/10.1002/2015WR017797]

21. Yi, S.; Sun, W.; Chen, J.; Feng, W. Anthropogenic and climate-driven water depletion in Asia. Pap. Present. Agu Fall Meet.; 2016; 43, pp. 9061-9069. [DOI: https://dx.doi.org/10.1002/2016GL069985]

22. Reager, J.T.; Gardner, A.S.; Famiglietti, J.S.; Wiese, D.N.; Eicker, A.; Lo, M.H. A decade of sea level rise slowed by climate-driven hydrology. Science; 2016; 351, 699. [DOI: https://dx.doi.org/10.1126/science.aad8386]

23. Zhong, Y.; Feng, W.; Humphrey, V.; Zhong, M. Human-Induced and Climate-Driven Contributions to Water Storage Variations in the Haihe River Basin, China. Remote Sens.; 2019; 11, 3050. [DOI: https://dx.doi.org/10.3390/rs11243050]

24. Wang, T.Y.; Wang, P.; Zhang, Y.C.; Yu, J.J.; Fang, Y.H. Contrasting groundwater depletion patterns induced by anthropogenic and climate-driven factors on Alxa Plateau, northwestern China. J. Hydrol.; 2019; 576, pp. 262-272. [DOI: https://dx.doi.org/10.1016/j.jhydrol.2019.06.057]

25. Liu, B.; Zou, X.; Yi, S.; Sneeuw, N.; Cai, J.; Li, J. Identifying and separating climate- and human-driven water storage anomalies using GRACE satellite data. Remote Sens. Environ.; 2021; 263, 112559. [DOI: https://dx.doi.org/10.1016/j.rse.2021.112559]

26. Fasullo, J.T.; Lawrence, D.M.; Swenson, S.C. Are GRACE-era Terrestrial Water Tren-ds Driven by Anthropogenic Climate Change?. Adv. Meteorol.; 2016; 2016, 4830603. [DOI: https://dx.doi.org/10.1155/2016/4830603]

27. Felfelani, F.; Wada, Y.; Longuevergne, L.; Pokhrel, Y.N. Natural and human-induced terrestrial water storage change: A global analysis using hydrological models and GRACE. J. Hydrol.; 2017; 553, pp. 105-118. [DOI: https://dx.doi.org/10.1016/j.jhydrol.2017.07.048]

28. Rodell, M.; Famiglietti, J.S.; Wiese, D.N.; Reager, J.T.; Beaudoing, H.K.; Landerer, F.W.; Lo, M.H. Emerging trends in global freshwater availability. Nature; 2018; 557, pp. 651-659. [DOI: https://dx.doi.org/10.1038/s41586-018-0123-1] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29769728]

29. Wiese, D.N. The JPL RL05M GRACE Mascon Solution: Status, Updates, and Future Prospects. AGU Fall Meet. Abstr.; 2016; 2016, G13A-1087.

30. Swenson, S.; Chambers, D.; Wahr, J. Estimating geocenter variations from a combination of GRACE and ocean model output. J. Geophys. Res. Solid Earth; 2008; 113, 5338. [DOI: https://dx.doi.org/10.1029/2007JB005338]

31. Cheng, M.; Ries, J.C.; Tapley, B.D. Variations of the Earth’s figure axis from satelli-te laser ranging and GRACE. J. Geophys. Res. Solid Earth; 2015; 116, 6265. [DOI: https://dx.doi.org/10.1093/nar/9.23.6265]

32. Peltier, W.R.; Argus, D.F.; Drummond, R. Comment on An Assessment of the ICE-6G_C (VM5a) Glacial Isostatic Adjustment Model. J. Geophys. Res. Solid Earth; 2018; 123, pp. 2019-2028. [DOI: https://dx.doi.org/10.1002/2016JB013844]

33. Ahmed, M.; Aqnouy, M.; El Messari, J.S. Sustainability of Morocco’s groundwater resources in response to natural and anthropogenic forces. J. Hydrol.; 2021; 603, 126866. [DOI: https://dx.doi.org/10.1016/j.jhydrol.2021.126866]

34. Döll, P.; Kaspar, F.; Lehner, B. A global hydrological model for deriving water availability indicators: Model tuning and validation. J. Hydrol.; 2003; 270, pp. 105-134. [DOI: https://dx.doi.org/10.1016/S0022-1694(02)00283-4]

35. Hunger, M.; Döll, P. Value of river discharge data for global-scale hydrological modeling. Hydrology and Earth System Sciences; 2007; 12, pp. 841-861. [DOI: https://dx.doi.org/10.5194/hess-12-841-2008]

36. Werth, S.; Güntner, A. Calibration analysis for water storage variability of the global hydrological model WGHM. Hydrol. Earth Syst. Sci. Discuss.; 2010; 6, pp. 59-78. [DOI: https://dx.doi.org/10.5194/hess-14-59-2010]

37. Zhang, Z.; Chao, B.F.; Chen, J.; Wilson, C.R. Terrestrial water storage anomalies of Yangtze River Basin droughts observed by GRACE and connections with ENSO. Glob. Planet. Change; 2015; 126, pp. 35-45. [DOI: https://dx.doi.org/10.1016/j.gloplacha.2015.01.002]

38. National Bureau of Statistics of China. Inner Mongolia Statistical Yearbook; China Statistics Press: Beijing, China, 2015.

39. Bingham, E.; Hyvärinen, A. A Fast Fixed-Point Algorithm for Independent Componet Analysis. Int. J. Neural Syst.; 2000; 10, pp. 1-8. [DOI: https://dx.doi.org/10.1142/S0129065700000028]

40. Hyvarinen, A. Fast and robust fixed-point algorithms for independent component analysis. IEEE Trans. Neural Netw.; 1999; 10, pp. 626-634. [DOI: https://dx.doi.org/10.1109/72.761722]

41. Hyvarinen, A.; Oja, E. Independent Component Analysis: Algorithms and Applications. Neural Netw.; 2000; 13, pp. 411-430. [DOI: https://dx.doi.org/10.1016/S0893-6080(00)00026-5]

42. Tsembelsuren, B.; Yu, C.H.; Rong, K. Study of Coal Trade between Mongolia and China, its Effect in Mongolian Coal Market. Eur. Res.; 2012; 19, pp. 379-382.

43. Purev, B.; He, W. Study on Competitiveness of Mongolian Coking Coal. Tech. Soc. Sci. J.; 2021; 19, 412.

44. Bulag, U.E. Mongolia in 2009 from landlocked to land-linked cosmopolitan. Asian Surv.; 2010; 50, pp. 97-103. [DOI: https://dx.doi.org/10.1525/as.2010.50.1.97]

45. Wang, H.Y.; Li, W.X.; Yang, Y.J.; Jiang, D.K. Groundwater dynamics in plain area in Tongliao, Inner Mongolia. Inn. Mong. Water Resour.; 2011; 4, pp. 111-113.