Trends in groundwater storage
Figure 1 illustrates the spatial distribution of the mean annual groundwater storage change trend from 2003 to 2022. The following regions were excluded: Antarctica, Greenland, the Gulf of Alaska coast, the Canadian Archipelago, and the Patagonian ice fields. The primary reason is that changes in water storage in these regions are mainly due to ice-sheet and glacier ablation caused by a warming climate, rather than changes in groundwater reserves (Rodell et al. 2018; Velicogna et al. 2014).
Over the past two decades, groundwater storage has exhibited spatial heterogeneity, with most depletion occurring within the Earth's mid-latitudes. With a 95% confidence level, approximately 81.15% of global regions, excluding ice melt regions, have experienced a significant trend in groundwater storage. Of these regions, 48.37% have witnessed a decline in groundwater levels, whereas 51.63% have experienced an increase. Notably, a significant decline in groundwater storage (> 20 mm/year) was observed in northern India, eastern Brazil, the Middle East, and areas surrounding the Caspian Sea. Among these, the most severe groundwater depletion occurred in the regions around the Aral, the Caspian Sea and northern India, with an average annual decline exceeding 30 mm over the past 20 years. These areas are characterized by sparse vegetation and fragile ecological environments. Approximately 3.2 billion people live in regions where groundwater levels have significantly declined over the past 20 years. Examples of areas with rising groundwater storage (GWS) trends include West and Central Africa, the Amazon rainforest, the northern United States, and Central Canada, with groundwater increases exceeding 20 mm annually. This significant increase in groundwater storage is likely due to high average precipitation, dense surface vegetation, and prevailing wetting climate trends. Additionally, groundwater depletion predominantly occurs in areas with deep water tables (Fan et al. 2013), which can make groundwater discharge more difficult and may require the construction cost of deeper wells to access groundwater (Jasechko & Perrone, 2021).
GWS and climate
Aridity index
The climatological mean Aridity index (AI), representing the climatological state for the period from 2003 to 2022, is shown in Extended Data Fig. 1. Since groundwater storage largely depends on the climatic region, the regionally averaged groundwater storage trend plotted against of the climatological mean aridity index is presented in Fig. 2. The analysis shows that the most significant groundwater decline occurred within the AI range of 0.1 to 0.5, peaking at 0.1 to 0.2. This indicates that arid and semi-arid areas, rather than hyper-arid zones, are experiencing the greatest depletion, a fact that has also been found by Scanlon et al. (2023). Contrary to popular belief, the Sahara Desert, which has an AI below 0.1 and lacks vegetation, does not significantly contribute to this extreme phenomenon. Excluding the Sahara Desert from the analysis did not substantially alter the groundwater storage trends, suggesting it is not a primary factor in substantial groundwater depletion, a fact that has also been found by Cuthbert et al. (2019). In arid and semi-arid regions, more than 60% of areas showed groundwater decline trends, indicating that these regions play a dominant role in both the area and intensity of groundwater depletion.
Overall, the observed groundwater distribution trends indicate that groundwater storage are declining in arid and semi-arid regions, while increasing in humid regions (AI ≥ 0.8). This disparity may lead to a more uneven distribution of groundwater.
Precipitation
The correlation between precipitation and groundwater storage has been confirmed in multiple studies (Thomas & Famiglietti, 2019; Russo & Lall, 2017). As shown in Fig. 3, precipitation trends exhibit an uneven spatial distribution. The highest significant increases in precipitation are observed in the eastern United States, northern Russia, central Africa, and southern China, while marked decreases are primarily concentrated in South America. Despite the notable reduction in precipitation in South America, most regions do not show a corresponding decline in groundwater storage. For instance, in the Amazon rainforest, even though precipitation is decreasing, high average temperatures and high average precipitation have not prevented the trend of rising groundwater storage (Heerspink et al. 2020). Although changes in precipitation can affect groundwater recharge, variations in groundwater storage do not always directly correspond to changes in precipitation. Therefore, the increase or decrease in precipitation is not the sole factor influencing groundwater storage. Long-term precipitation changes may impact global groundwater storage due to the slow response to recharge, and many regions experience groundwater decline due to excessive extraction (Herrera-Pantoja & Hiscock 2008; Thomas & Famiglietti 2019). Increased evapotranspiration due to climate change may prevent groundwater reserves from being replenished by precipitation, even if precipitation increases (Green et al. 2011).
SPEI Drought index
The climatological mean Standardized Precipitation-Evapotranspiration Index (SPEI), representing meteorological drought conditions for the period from 2003 to 2022, is shown in the Extended Data Fig. 2. The spatial distribution of the trend of the global SPEI trend over the past 20 years is shown in Fig. 4. Compared to the trend of precipitation map in Fig. 3, the changes in groundwater storage are more closely related to the SPEI trend distribution, a fact that has also been found by Dai (2013) and Wang et al. (2014). This is likely because the SPEI is influenced by multiple factors including precipitation, temperature, evapotranspiration, and other climatic conditions. Van Loon (2015) noted that groundwater storage fluctuates in response to meteorological trends, increasing during wetter periods and decreasing during drier ones. However, not every meteorological drought leads to a decline in groundwater levels. This is mainly because short-term meteorological droughts do not impact groundwater, and only prolonged droughts cause significant changes in groundwater storage (Li & Rodell 2015; Barichivich et al. 2022). Additionally, groundwater droughts exhibit a delayed effect compared to meteorological droughts, as it takes time for precipitation to recharge groundwater, and a reduction in precipitation leading to meteorological drought does not immediately cause groundwater depletion (Han et al. 2019; Gates et al. 2011).
GWS and human factors
Regarding the impact of human activity on groundwater storage trends, we examined four human activity related factors as illustrated in Fig. 5. These factors are: (a) area actually irrigated (%), (b) areas equipped for irrigation with groundwater (%), (c) annual mean net abstraction from groundwater (mm/a), and (d) population density (per/km²). For Fig. 5(a), when the proportion of the actually irrigated area is below 50%, the trend in groundwater storage is not conspicuous. However, as the proportion increases from 50–90%, a discernible decline in groundwater levels becomes evident in more and more areas, with many of them experiencing a decrease of approximately 10 mm per year. Upon surpassing the 90% threshold, more areas exhibit a pronounced downward trend in groundwater levels, with many regions experiencing declines exceeding 10 mm per year. This also includes the regions with the most pronounced declines of more than 20 up to 50 mm/a. Figure 5(b) shows that many regions with varying proportions of irrigation using groundwater all exhibit significant areas with declines in groundwater storage, most of them in regions with 20% or more area equipped for irrigation with groundwater, while there are also some regions with high declines and no or little irrigation with groundwater. This may be related to regional recharge and discharge rates, making the relationship between areas equipped for irrigation with groundwater and groundwater storage not so clear. For Fig. 5(c), the data indicate that net recharge to groundwater (negative values of net abstraction, caused by irrigation return flow in areas irrigated by surface water) does not necessarily lead to an increase in groundwater storage, but excessive groundwater abstraction most often results in a decrease in groundwater storage. Different to irrigation and abstraction factors, population density (Fig. 3 (d)) does not have a direct relationship with groundwater storage. Thus, high population density does not necessarily lead to reduced groundwater storage. This confirms that abstractions connected with high population density, e.g. for drinking water or industrial purposes, have a minor effect on GWS, compared to agriculture, which is the cause for the majority of abstracted groundwater and is more likely to be found in less populated regions.
GWS and Vegetation Cover
To analyze the potential impacts of vegetation cover and its changes, particularly due to rainforest deforestation, on groundwater storage (GWS), we compared GWS trends with the mean leaf area index (LAI) (Extended Data Fig. 4) and LAI trends from 2003 to 2022 (Extended Data Fig. 5). In recent years, researchers have observed a global greening trend in vegetation using earth observation technologies and remote sensing imagery (Keenan & Riley, 2018; Song et al., 2018), similar to the greening trend observed in the LAI trend discussed in this paper. However, on a global scale, these patterns are not immediately apparent. Regions exhibiting increasing or decreasing GWS trends are found in areas with both high and low LAI, as well as in regions with both increasing and decreasing LAI trends (Tao et al., 2020). Therefore, we discuss the potential influences in selected regions below.
Regional groundwater storage changes and attributions
To better analyze the driving factors of regional groundwater storage changes, we identified 23 regions where groundwater storage significantly changed over the past 20 years (2003–2022), as shown in Fig. 6. To infer the reasons behind these changes, we compiled several factors that may influence groundwater storage in these regions, as detailed in Table 1. Based on these factors, we attributed the trends in groundwater storage changes to human impact, climate impact, or a combination of both. Groundwater storage trends and extremes value were obtained by calculating the area-weighted mean and extreme values for each region (maximum value for regions with groundwater rising and minimum value for regions with groundwater declining). The proportion of irrigated area was calculated by area-weighting the pixel values of irrigation intensity within each study region. The spatial variation maps for these factors are as follows: precipitation trends in Fig. 3, SPEI trends in Fig. 4, mean leaf area index (LAI) and its trends in the Appendix Fig. A4 and A5, average population density in the Appendix Fig. A6, and annual average net groundwater extraction in the Appendix Fig. A7.
Table 1
Statistics of groundwater storage trends and possible influencing factors for the selected regions in Fig. 6. Supposed main driving forces are not directly concluded from the statistics but discussed in the text.
Region No.
|
Location
|
Area (km2)
|
Mean GWS trend (mm/a)
|
GWS trend extreme value (mm/a)
|
Areas equipped for irrigation (%)
|
Area equipped for irrigation with groundwater (%)
|
Mean Precipitation trend (mm/a)
|
Mean Leaf area index mean (%)
|
Mean Leaf area index trend (%)
|
SPEI trend
|
Population density (people/km²)
|
Annual net abstraction from groundwater max value (mm/a)
|
Supposed main driving force
|
1
|
Western America
|
4,241,546
|
-3.0
|
-12.7
|
49.7
|
30.2
|
0.00
|
0.72
|
0.0025
|
-0.017
|
63
|
177.2
|
climate + human impact
|
2
|
Central Canada
|
1,525,571
|
-1.9
|
-6.3
|
3.7
|
0.1
|
-0.03
|
0.59
|
0.0088
|
0.008
|
2
|
3.2
|
climate impact
|
3
|
Eastern United States
|
906,647
|
5.4
|
16.4
|
37.9
|
32.5
|
0.46
|
1.00
|
0.0000
|
0.013
|
94
|
35.4
|
climate impact
|
4
|
Amazon
|
2,716,443
|
4.7
|
14.1
|
12.0
|
4.2
|
-0.53
|
3.60
|
-0.0008
|
-0.013
|
7
|
3.9
|
climate + human impact
|
5
|
Eastern Brazil
|
1,831,986
|
-5.2
|
-17.9
|
87.5
|
18.9
|
-0.15
|
1.64
|
0.021
|
-0.016
|
23
|
32.2
|
climate impact
|
6
|
Southern Brazil-Paraguay region
|
1,553,784
|
5.0
|
12.3
|
42.0
|
13.3
|
-0.17
|
1.49
|
0.0205
|
-0.010
|
20
|
24.6
|
human impact
|
7
|
Central Argentina
|
369,510
|
-3.5
|
-11.4
|
12.6
|
1.6
|
0.01
|
0.33
|
0.0052
|
-0.009
|
8
|
0.5
|
climate impact
|
8
|
Central Europe
|
3,711,950
|
-1.9
|
-5.0
|
23.4
|
16.5
|
0.00
|
1.14
|
0.063
|
-0.013
|
119
|
244.4
|
climate + human impact
|
9
|
Northern Africa
|
2,273,262
|
-1.6
|
-7.3
|
9.4
|
6.2
|
-0.08
|
0.04
|
0.0003
|
-0.011
|
21
|
139.9
|
climate impact
|
10
|
Western Africa
|
2,477,929
|
4.9
|
14.8
|
4.0
|
0.9
|
-0.01
|
0.40
|
0.0026
|
0.003
|
80
|
22.4
|
human impact
|
11
|
Nile headwaters
|
3,451,624
|
6.9
|
28.2
|
2.8
|
0.3
|
0.20
|
2.26
|
0.032
|
0.012
|
78
|
3.3
|
climate + human impact
|
12
|
Southeastern Africa
|
1,802,587
|
-2.1
|
-7.2
|
7.8
|
0.6
|
0.00
|
0.90
|
0.0060
|
0.005
|
32
|
2.4
|
climate impact
|
13
|
Aral Sea and Caspian Sea
|
5,708,583
|
-5.8
|
-40.4
|
22.4
|
13.3
|
-0.09
|
0.18
|
0.0035
|
-0.025
|
65
|
101.9
|
climate + human impact
|
14
|
Northwestern China
|
898,833
|
-3.0
|
-14.7
|
20.5
|
4.4
|
-0.06
|
0.16
|
0.0000
|
-0.020
|
20
|
19.8
|
climate impact
|
15
|
Tibetan Plateau
|
230,001
|
5.5
|
10.7
|
0.0
|
0.0
|
0.00
|
0.02
|
0.0000
|
0.001
|
1
|
0.0
|
climate impact
|
16
|
North China Plain
|
1,004,133
|
-3.4
|
-16.3
|
43.7
|
30.2
|
0.02
|
0.38
|
0.015
|
-0.008
|
208
|
290.3
|
human impact
|
17
|
Northeast China
|
966,799
|
-1.7
|
-5.0
|
61.1
|
36.4
|
0.12
|
0.99
|
0.0075
|
0.009
|
141
|
68.8
|
human impact
|
18
|
Arabian Peninsula
|
1,754,517
|
-2.9
|
-9.3
|
7.4
|
9.2
|
0.00
|
0.02
|
0.0012
|
-0.017
|
62
|
169.3
|
climate + human impact
|
19
|
Northern India
|
1,532,713
|
-9.9
|
-54.8
|
59.8
|
36.5
|
0.08
|
0.53
|
0.038
|
0.012
|
363
|
251.6
|
human impact
|
20
|
Southern India
|
434,225
|
-4.8
|
-11.7
|
77.1
|
57.9
|
0.08
|
0.75
|
0.052
|
0.017
|
447
|
173.9
|
human impact
|
21
|
South Asia
|
1,198,228
|
-5.6
|
-25.2
|
49.5
|
19.9
|
-0.02
|
1.10
|
0.076
|
0.012
|
399
|
238.4
|
human impact
|
22
|
Eastern Central China
|
461,212
|
3.6
|
5.4
|
71.3
|
0.1
|
0.47
|
1.12
|
0.15
|
0.023
|
295
|
-0.1
|
climate + human impact
|
23
|
Mainland Southeast Asia
|
710,066
|
-3.3
|
-9.7
|
49.8
|
3.0
|
0.18
|
2.05
|
0.059
|
0.012
|
145
|
0.5
|
climate impact
|
North America
1 Western America
A historically severe drought that began in 2011–2014 (Griffin & Anchukaitis 2014) and has persisted to the present is centered in southern California and extends into Mexico. This drought has exhibited a trend of decreasing SPEI values, indicating progressive aridification (Alam et al. 2021). This has resulted in a mean groundwater storage (GWS) decline at a rate of -3 mm/a and an increased groundwater demand, with net abstraction reaching a maximum of 177 mm/a. Approximately 50% of the surface area is agricultural, with 30% of this region irrigated using groundwater and
268 million people. Consequently, the decline in groundwater levels is due to a combination of drought and agricultural water demands that exceed renewable water resources.
2 Central Canada
From 2003 to 2022, overall precipitation has been slightly decreasing, leading to an average GWS decline of -1.9 mm/a in Central Canada. This loss of water aligns with a recent study concluding that Canada's subarctic lakes are vulnerable to drying when snow cover declines (Bouchard et al. 2013). Moreover, there is hardly any abstraction from groundwater connected to irrigation. Therefore, the changes are supposed to be climate-related.
3 Eastern United States
The increasing precipitation trend in the Eastern United States has led to a significant rise in groundwater storage (Rateb et al. 2020), with an average increase of 5.4 mm/a and a maximum increase of 16.4 mm/a. In this region, 38% of the land is used for irrigation, mostly relying on groundwater. However, the increase in groundwater levels due to the wetter climate persists despite the extensive use of groundwater for irrigation.
South America
4 Amazon
Despite the decreasing trend in precipitation and significant deforestation in the Amazon over the past few decades (Lin et al. 2016), the region still receives an average annual rainfall of 2362 mm, making it one of the most precipitation-rich areas globally. During these two decades, GWS in central and western Brazil and its neighboring areas increased at an average rate of 4.7 mm/a, with a maximum increase of 14.1 mm/a. This trend is attributed to changes of the Amazon water cycle, which has been significantly altered by changes in climate, land cover (especially deforestation, as supported also by a negative trend in the LAI), sea surface temperature and precipitation patterns since the 1980's (Heerspink et al. 2020; Satizábal-Alarcón et al. 2024).
5 Eastern Brazil
Eastern Brazil has recently experienced a severe drought (Lima et al. 2022), characterized by a significant decrease in the SPEI values and reduced precipitation. This phenomenon is likely associated with a strong El Niño event (Santos et al. 2021). In fact, 2015 was the driest year in the past 37 years (Marengo et al. 2017). Consequently, GWS over the past 20 years has decreased by 5.2 mm/a, with a maximum decline of 17.9 mm/a. This region is not primarily attributed to human influence due to its relatively low maximum groundwater abstraction of 32 mm/a, accounting for 19% of irrigation. Agriculture covers 88% of the surface area, predominantly reliant on rainfed agriculture.
6 Southern Brazil-Paraguay regions
Groundwater storage has shown significant increases over the past two decades in southern Brazil and adjacent regions in Paraguay. On average, GWS has increased by 5 mm/a, with a maximum increase of 12.3 mm/a. Despite a decreasing trend in precipitation, the SPEI values indicate a gradual shift towards a drought climate. Consequently, GWS has recovered from its earlier drought (Rodell et al., 2018), exhibiting a substantial yet brief upward trend. It is noteworthy that southern Brazil is a focal point for dam construction (Hu et al. 2017), suggesting that reservoir impoundment may have influenced this upward trend.
7 Central Argentina
Groundwater storage in Central Argentina had previously been declining at a rate of -3.5 mm/a, with the maximum decline reaching − 11.4 mm/a. Vegetation in central Argentina is sparse, with a mean Leaf Area Index (LAI) of only 0.33%. A multi-year drought that began in 2009 has persisted to the present (Müller et al. 2014). Due to limited human activities and agriculture in the study area, we suppose the driving factors of declining GWS to be mainly climatic.
Europe
8 Central Europe
An analysis of the mean annual trends over the study period reveals a slight decrease in GWS in many European countries, with an average value of -1.9 mm/a for Central Europe. The maximum annual net abstraction from groundwater reached 244 mm/a, which supports 16% of agricultural activities and 443 million people. Concurrently, extreme drought events over the past 20 years (e.g., 2003, 2011–2013, 2015–2016) (Ionita et al., 2021; Van Lanen et al. 2016) have also influenced the distribution of groundwater reserves. That is depicted also in negative trend of the SPEI in parts of the region, e.g. northern Germany. The results clearly indicate medium-term groundwater stress in most European countries (Xanke & Liesch 2022).
Africa
9 Northern Africa
During the study period, a severe drought occurred in Northern Africa (Spinoni et al. 2019), closely aligning with areas experiencing groundwater depletion. The region exhibited a weak negative trend with an average decrease of -1.6 mm/a. Sparse abstraction regions are concentrated in densely populated areas, with the highest abstraction from groundwater at 140 mm/a, supporting 6% of irrigation from groundwater and 48 million people. Therefore, the primary cause of groundwater depletion is attributed to meteorological drought.
10 Western Africa
Groundwater storage has been increasing in the tropical climate of Western Africa at an average rate of 4.9 mm/year. Over the past 20 years, the region's SPEI values and precipitation trends have remained relatively stable. The maximum groundwater abstraction of 22 mm/year supports a population of 199 million people and a small amount of agricultural land. The construction of numerous dams in this region is likely a major factor contributing to the rising groundwater levels (Scanlon et al., 2022; Zarfl et al., 2015).
11 Nile headwaters
The groundwater level in Nile headwaters has experienced the largest increase globally, with a maximum groundwater storage (GWS) rise of 28.2 mm/a and mean rise of 6.9 mm/a. Despite a large population of 270 million, the maximum abstraction amount is only 3.27 mm/a. This suggests that increased precipitation (positive P trend of 0.2 mm/a and
wetter climate) is the primary driver of GWS variations. Additionally, the management of large lakes and dam construction (Ahmed et al. 2014) in the northern part of the region could also contribute to these changes (Kebede et al. 2017).
12 Southeastern Africa
The negative trend along the southeastern coast of Africa has resulted in an average groundwater storage decline of -2.1 mm/a. Over the past 20 years, the regions experiencing groundwater depletion partially correspond to areas affected by this event. This event was associated with a pronounced north–south dipole pattern of positive or negative rainfall and water balance anomalies, typical of the El Niño–Southern Oscillation (ENSO) teleconnection to the region (Kolusu et al. 2019). Therefore, it is likely that this trend is primarily climatic (Scanlon et al. 2022).
Asia
13 Aral Sea and Caspian Sea
The demise of the Aral Sea has been extensively documented in numerous studies (Qadir et al. 2009; Hu et al. 2022). Our estimates indicate a mean groundwater storage (GWS) depletion rate of -5.8 mm/year, with a maximum recorded rate of -40.4 mm/year. This area represents the second-largest groundwater depletion region globally, characterized by both its vast expanse and significant depth of the groundwater depression. The primary climate type in this region is arid, with desert precipitation. Concurrently, the area experiences a high abstraction rate, reaching up to 102 mm/a to support approximately 369 million people. Precipitation in this region is minimal and continues to decrease (as depicted by a negative P trend), contributing to worsening drought conditions. Approximately 22% of the land is equipped for irrigation, with 13% of the irrigation relying on groundwater. Therefore, the decline in groundwater levels is attributed to both climate change and human activity.
14 Northwestern China
During the study period, the groundwater storage in Northwestern China decreased by 3 mm/a, with a maximum decrease of 14.7 mm/a. This decline may be related to ice melt from the Tien Shan mountains (Jacob et al. 2012). In this region, 20% of the area consists of arable land that primarily relies on rainfed agriculture. The annual average precipitation is low, and the trend in precipitation is also declining, with a low leaf area index (LAI). Over the past 20 years, the SPEI Index has shown a drought trend towards. Thus, the decrease in groundwater is primarily due to the impact of climate change.
15 Tibetan Plateau
The majority of lakes in the Tibetan Plateau have increased in water level and extent during the 2000s due to a combination of elevated precipitation rates and increased glacier-melt flows, which are difficult to disentangle (Zhang et al. 2013). Over the past 20 years, groundwater has been rising at a mean rate of 5.5 mm/a. There is no abstraction for agricultural irrigation in this region, thus the rise in GWS can be attributed to climate.
16 North China Plain
Numerous studies indicate a gradual GWS decline in the North China Plain (e.g. Gong et al. 2018; Su et al. 2021). This region supports a large population of 209 million, with an average annual groundwater abstraction rate of 290 mm/a, primarily for irrigation purposes. The region surrounding Beijing heavily relies on extensive agriculture, which constitutes 44% of the area. A significant portion of the water used for agriculture (30%) comes from groundwater. All evidence suggests that this trend is human-induced and likely to continue until groundwater becomes scarce or regulations are implemented to reduce consumption rates. Therefore, the decrease in groundwater is primarily due to human activity.
17 Northeast China
61% of Northeast China covered by agriculture, 36% of irrigation comes from groundwater. The net abstraction from groundwater reaches a maximum of 69 mm/a, primarily supporting 137 million people and intense irrigation. During the GRACE period under consideration, the estimated mean rate of groundwater storage change was − 1.7mm/a, with a maximum of -5 mm/a. During this period, there was an observed increase in precipitation and a wetting trend. All available evidence suggests that the negative GWS trend is attributable to intensity agriculture and overload abstraction form groundwater (Sun et al. 2022).
18 Arabian Peninsula
Decreasing water storage in the Arabian Peninsula has been quantified using GRACE in previous studies (Joodaki et al. 2014; Voss et al. 2013). Over the past 20 years, most of the Arabian Peninsula has shown an increasing trend of aridity. The average groundwater storage (GWS) in this region is -2.9 mm/year, while precipitation, characteristic of a desert climate, averages 114 mm/year. Groundwater abstraction reaches a maximum of 169 mm/year, thus by far exceeding recharge rates and being not sustainable. The decline in groundwater storage results from recent drought and excessive groundwater extraction.
19 Northern India
Over the past 20 years, Northern India has experienced the most severe groundwater depletion globally (Asoka et al. 2017; Bhanja & Mukherjee 2019). Despite an upward trend in precipitation and wetting trend in SPEI values, the substantial groundwater depletion trend persists. The average groundwater storage (GWS) in this region has decreased by -9.9 mm/a, with a maximum decline of -54.8 mm/a. The contribution of Himalayan glacier mass loss to this regional trend is minor (Tiwari et al. 2009; Rodell et al. 2018). Agriculture occupies 60% of the area, with 37% of irrigation relying on groundwater. Annual net abstraction from groundwater has reached 252 mm/a, exceeding the recharge rate, to support 566 million people and irrigation.
20 Southern India
The groundwater storage in India exhibit distinct regional trends: a significant decline in the north, a rise in the central part, and a decreasing trend in Southern India. In the study by Asoka et al. (2017), southern India showed an increasing trend from 2002 to 2013. However, over the 20-year period from 2003 to 2022, groundwater storage (GWS) has exhibited a sharp decline, with an average decrease of -4.8 mm/a and a maximum decrease of -11.7 mm/a. In this region, groundwater abstraction reaches a maximum of 174 mm/a, with 58% of the area relying on groundwater for irrigation, supporting 194 million population. The primary factors contributing to the groundwater decline include intensive agriculture (Salmon et al. 2015), significant groundwater extraction, and high population density, all of which result in excessive groundwater withdrawal.
21 South Asia
Approximately half of the South Asia is devoted to irrigation. The region's highest net abstraction from groundwater is 238 mm/a, supporting irrigation agriculture and exceeding 478 million population and persistent groundwater withdrawal has led to a downward trend in groundwater levels. Although the area receives relatively abundant rainfall, averaging 1200 mm/a, the precipitation trend is slightly decreasing. However, the excessive extraction driven by intense agriculture and population is the primary factor for this decline. The fact that extractions already exceed recharge during years of normal precipitation indicates a concerning outlook for groundwater availability during future drought (Aadhar & Mishra 2017).
22 Eastern Central China
Despite the predominance of rainfed agriculture (71%) in Eastern Central China, groundwater storage has increased on average by 3.6 mm/a, with a maximum rise of 5.4 mm/a. Simultaneously, this region has experienced the largest upward trend in precipitation globally. The increase in precipitation, along with the construction and filling of the Three Gorges Dam (Yang & Lu 2014), has significantly contributed to the rise in groundwater storage (Chao et al. 2020; Yin et al. 2021).
23 Mainland Southeast Asia
In Mainland Southeast Asia, 50% of the area is dedicated to agriculture, yet the water source for irrigation is rarely from groundwater. Overall, the region experiences abundant rainfall, averaging 1847 mm/a, with a rising precipitation trend at the rate of 2.12 mm/a over the past 20 years. Despite this, the mean groundwater storage (GWS) has the mean negative trend of -3.3 mm/a, with a maximum decline of -9.7 mm/a. This decline may be associated with climatic factors, such as the El Niño-Southern Oscillation (ENSO) (Sulaiman et al. 2023; Le et al. 2021).