Journal Pre-proofs

Research papers

Evolution characteristics of groundwater and its response to climate and land-

cover changes in the oasis of dried-up river in Tarim Basin

Wanrui Wang, Yaning Chen, Weihua Wang, Junxin Jiang, Ming Cai,
Yongjun Xu

PII: S0022-1694(20)31105-7
DOI: https://doi.org/10.1016/j.jhydrol.2020.125644
Reference: HYDROL 125644

To appear in: Journal of Hydrology

Received Date: 23 July 2020

Revised Date: 18 September 2020
Accepted Date: 11 October 2020

Please cite this article as: Wang, W., Chen, Y., Wang, W., Jiang, J., Cai, M., Xu, Y., Evolution characteristics of
groundwater and its response to climate and land-cover changes in the oasis of dried-up river in Tarim Basin,
Journal of Hydrology (2020), doi: https://doi.org/10.1016/j.jhydrol.2020.125644

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2020 Elsevier B.V. All rights reserved.

Evolution characteristics of groundwater and its response to climate
and land-cover changes in the oasis of dried-up river in Tarim Basin

Wanrui Wang a,b , Yaning Chen a,b,*, Weihua Wang a,b，Junxin Jiang c, Ming Cai
c, Yongjun Xu c
a State Key Laboratory of Desert and Oasis Ecology, Xinjiang Institute of Ecology and
Geography, Chinese Academy of Sciences, Urumqi 830011, China
b University of Chinese Academy of Sciences, Beijing 100000, China
c Xinjiang Tarim River Basin Aksu Management Bureau, Aksu 843000, China

* Corresponding author: Yaning Chen

State Key Laboratory of Desert and Oasis Ecology, Xinjiang Institute of Ecology

and Geography, Chinese Academy of Sciences, Urumqi 830011, China.

E-mail address: chenyn@ms.xjb.ac.cn (Y. Chen).

Abstract:

Groundwater is essential for agricultural, ecological and domestic water use, and its

depletion under climatic and anthropogenic impacts can impose challenges on

regional water resources and ecological security, particularly in arid regions.

Spatiotemporal variation and drivers of groundwater are critical to understand

groundwater cycling and restore groundwater, while they are poorly understood in

arid oasis region due to sparse field monitoring. To address this, we examined the

spatiotemporal variations and controlling factors of groundwater level and

mineralization degree using data collected at 22 wells in the Wei-Ku Oasis of Tarim

Basin during 2000-2018, a typical arid inland river basin in northwest China. Results

showed that obviously large spatial and temporal variations existed in both DGL

(depth to groundwater level) and DM (degree of mineralization). Groundwater level

was the shallowest in spring and deepest in autumn. DGL exhibited a markedly

increasing trend (2.05 m per decade), while DM marginally decreased (-0.21 g/L per

1
decade) (p < 0.001) during 2000-2018. The spatial distributions of DGL and DM were

complex. Larger variation of DGL/DM generally occurred at the areas with greater

DGL/DM. DM in the northwestern part was obviously smaller than that in the

southeastern part. The variation of DGL in the eastern region was noticeably larger

than that in the western region. Human activity was the main driving factor of DGL

increasing, climate change had a dominant role in DM decreasing, and the

relationship between controlling factors and groundwater varied in space and season.

Decreasing riverbank leakage and increasing volume of evapotranspiration,

groundwater pumping and water drainage led to decreasing groundwater level, while

increasing infiltration/leakage from irrigated water, canal and reservoir resulted in

increasing groundwater level. The results could help to advance the understanding of

groundwater cycling mechanism, and be useful in improving sustainable groundwater

management and restoring ecosystem in arid region.

Keywords:

Groundwater dynamics; Climate variability; Land use change; Agricultural

irrigation; Dried-up river oasis; Tarim Basin

1. Introduction

Groundwater is essential for agricultural, ecological and domestic water use

(Orellana et al., 2012; Taylor et al., 2013; Erler et al., 2019), and is extremely

important in arid regions due to low rainfall and insufficient surface water (Zhang et

al., 2014a; Wang et al., 2019b; Chen et al., 2019; Jia et al., 2020), including the

quantity and quality of groundwater (Ma et al., 2013; Liu et al., 2018b; Porhemmat et

al., 2018; Zou et al., 2019). Owing to agricultural development, urban land expansion,

increasing population and climate change, the locations, amounts and timing of

groundwater discharge and recharge are being modified worldwide (Han et al., 2017;

Manna et al., 2019; Riley et al., 2019), which are also currently leading to an

unforeseen consequence for groundwater quality (Tweed et al., 2011; Wang et al.,

2
2013; Bekele et al., 2019; Riedel, 2019). Groundwater depletion caused primarily by a

growing influence of human activities during the past decades can impose challenges

on regional water resources and ecological security in arid areas worldwide (Zhang et

al., 2014a; Liu et al., 2018a; Graaf et al., 2019), particularly in arid inland river basin

with large irrigated agriculture (Ji et al., 2006; White et al., 2014; Cheng et al., 2014).

The most affected areas include the Northwestern India (Asoka et al., 2017), Middle

East (Joodaki et al., 2014), California and Southern High Plains in the North America

(Scanlon et al., 2012), and North China Plain (Feng et al., 2013). Groundwater level is

one of the most important parameters for studying arid ecosystems (Soylu et al.,

2011), and its continuous decline mainly due to groundwater overexploitation for

irrigation supplement in recent decades has influenced hydrological and ecological

processes, biogeochemical cycling and plant communities in arid regions of the world

(Huang and Pang, 2010; Zhang et al., 2014). Groundwater level fluctuation can reflect

the state of connectivity between river and aquifer in semiarid and arid areas

(McCallum et al., 2013; Fuchs et al., 2019; Riley et al., 2019). Furthermore,

groundwater salinization is linked with the recharge, discharge and migration

processes of groundwater (Tweed et al., 2011; Liu et al., 2018b), and can impact

public health, soil fertility, vegetation growth and renewable fresh groundwater

resources (Manchanda and Garg, 2008; Liu et al., 2018b; Jia et al., 2020). Spatial

distribution of groundwater salinity can reflect groundwater salinization process (Xie

et al., 2013; Cary et al., 2015), and thus in turn could provide insight into

hydrogeochemical evolution as well as the interactions between local environment

and groundwater in arid regions (Edmunds, 2009; Liu et al., 2018b). Therefore, it is

critical to evaluate the dynamics and drivers of groundwater under climatic and

anthropogenic impacts for effectively understanding, modelling and managing

groundwater systems and improving ecosystem in these arid regions (Han et al.,

2017).

Numerous studies have been carried out to analyze groundwater hydrological

3
processes for reasonable groundwater management in arid areas via various

techniques, e.g. geostatistical analysis (Hu et al., 2012), traditional field experiments

(Wang et al., 2014; Jiang et al., 2015), water chemistry (Huang and Pang, 2010),

tracer experiments (Guo et al., 2019; Ma et al., 2019a), remote sensing inversion

(Zhao et al., 2013; Wang et al., 2019b) and hydrologic modeling (Liu et al., 2018).

Currently, studies of groundwater dynamic are mainly based on two data sources,

including local in-situ data (borehole data) and satellite-based GRACE (Gravity

Recovery and Climate Experiment) data combined with hydrological models (Tweed

et al., 2007; Strassberg et al., 2009; Hughes et al., 2012; Chen et al., 2016; Hu et al.,

2019). Previous studies on groundwater in arid regions mainly focused on the

changing trend and rate of groundwater level on annual or interannual scale (Ma et al.,

2005; Yang and Li, 2011; Hu et al., 2019), groundwater recharge and discharge in

different aquifers (Scanlon et al., 2006; Pang et al., 2010; Tweed et al., 2011; Guo et

al., 2019; Ma et al., 2019a), groundwater salinization and hydrogeochemical evolution

(Feng et al., 2004; Wang et al., 2013; Liu et al., 2018), the interaction between surface

water and groundwater (McCallum, et al., 2013; Yao et al., 2015; Xi et al., 2018), and

the relationships between groundwater dynamics and driving factors (Wang et al.,

2014; Zhang et al., 2014a; Zhou et al., 2018; Liu et al., 2018a; Wang et al., 2019b).

Generally, the hydrological processes that control regional phreatic fluctuation are

mainly river water leakage, little rainfall infiltration, irrigation return flow, lateral

groundwater flow, groundwater evapotranspiration and pumping in arid regions

(Healy and Cook, 2002; Abdalla, 2008; Wang et al., 2014; Guo et al., 2019; Liu et al.,

2018b). The spatio-temporal dynamic of groundwater salinity in arid inland river

basin is influenced by several mechanisms, including evaporation, transpiration of

deep-rooted plants, dilution, mixing with saline water, mineral dissolution and

leaching, and agricultural irrigation (Herczeg et al., 2001; Han et al., 2011; Tweed et

al., 2011; Farid et al., 2015; Sun et al., 2016; Liu et al., 2018b; Jia et al., 2020).

However, although the importance of phreatic fluctuation as a mechanism that

4
influences soil desertification and ecosystem restoration is well recognized,

groundwater level dynamics and its associated hydrological processes have not been

studied extensively in arid areas (Wang et al., 2014).

Changes in the groundwater level and quality could be caused by many factors,

including human activities and climate variability (Wang et al., 2014; Hu et al., 2019;

Fu et al., 2019). Recently, anthropogenic activities (e.g. agricultural irrigation) have

exerted great pressure on natural hydrological cycle in arid regions (Liu et al., 2018a),

and even have greater effects than climate change (Taylor et al., 2013). Previous

studies investigated the impacts of anthropogenic activities and climate change on

groundwater via building a geostatistical relationship based on historical observations

or using a hydrological model (Fu et al., 2019; Wang et al., 2019b). Generally,

groundwater level and quality in arid regions could be affected by topographic

characteristics, geology, aquifer lithology, climatic factors, land use pattern,

vegetation, soil texture, agricultural activities, etc. (Wang et al., 2014; Jiang, et al.,

2015; Liu et al., 2018b; Liu et al., 2018a; Porhemmat et al., 2018; Segura et al., 2019;

Wang et al., 2019b). The combined effects of these control factors on groundwater are

complicated (Bekele et al., 2019). Chen et al. (2009) found that groundwater

extraction can alter the recharge and discharge processes of groundwater, and lead to

increasing salinity in aquifer. Porhemmat et al. (2018) reported that water saving

irrigation systems reduced potential groundwater recharge. Agam et al. (2012) and

Zhang et al. (2014a) found that groundwater table declined with increasing

agricultural land in arid and semi-arid regions. However, the controls of shallow

groundwater level and quality have not been thoroughly examined in arid regions,

especially in the arid inland river basins.

The Tarim River Basin is the largest inland river basin and socioeconomically

important region in China, and is highly stressed in terms of water scarcity and the

water use conflicts between eco-environment and agriculture due to uncontrolled

cultivated land expansion (Pang et al., 2010; Zhang et al., 2016; Chen et al., 2019). In

5
the last 30 years, the exploitation of water and land resources associated with

population growth in the Tarim Basin have resulted in great changes in water regime

and sharply increased water consumption (Zhang, et al., 2015; Wang et al., 2019a),

and the irrigated agriculture relied heavily on groundwater due to scarce surface water

resources, which induced groundwater overexploitation for irrigation (Zhang et al.,

2014b; Chen et al., 2019). Some studies, e.g., Chen et al. (2019), Bai et al. (2020b),

and Wang et al. (2019a), reported that groundwater level is decreasing in Tarim Basin

due to the increased demands for water usage on the one hand, and increase in

evaporation on the other hand. The current decline and extraction of groundwater are

not sustainable, i.e., human abstraction exceeds natural groundwater recharge (Chen

et al., 2019). Anthropogenic impacts and climate variability are exerting pressure on

the groundwater in Tarim Basin (Bai et al., 2020b), which has caused some serious

hydrological and ecological issues in the middle and lower reaches, such as

dramatical groundwater level decline, natural vegetation degradation (especially the

riparian ecosystem), soil salinization and desertification, and drying of rivers (Huang

and Pang, 2010; Bai et al., 2020b; Chen et al., 2019). Therefore, a thorough study of

long-term variations of groundwater level and quality affected by agricultural

activities in the Tarim Basin is quite necessary to understand groundwater cycling and

salt migration mechanism, and also is the basis for sustainable groundwater

management to meet ecosystem and human needs in such a water-limited arid

environment (Scanlon et al., 2006; Liu et al., 2018a; Hu et al., 2019; Wang et al.,

2019b). However, the evolution characteristic of groundwater and its response to

climate variability and enhanced anthropogenic activities were poorly understood in

this region due to relatively few hydrological monitoring stations and limited field

observation data (Chen et al., 2019). This had resulted in uncertainties when

simulating and predicting the regional hydrological and biogeochemical cycling, and

imposed challenges on the sustainable management of regional water resources and

appropriate policy formulation in this arid inland river basin.

6
 The Wei-Ku Oasis, a typical dried-up river oasis of the Tarim Basin in

northwestern China was selected as the study area to explore the variations and

drivers of groundwater. The oasis has experienced intensive cultivated area

expansion, rapidly growing population, and increasing extreme weather events in

recent decades (Chen et al., 2019). Previous studies using multiple temporal scales to

examine oasis groundwater dynamics revealed obvious inter-annual variability of

groundwater level (Zhang et al., 2014b). However, the evolution characteristics of

groundwater and their response to climate and land-cover changes in the oasis were

not well understood. We analyzed the spatio-temporal variations and determinants of

groundwater over the Wei-Ku Oasis using long-term situ groundwater monitoring

data collected at 22 wells over the past 19 years (2000-2018) (monthly water level and

quality). The main objectives of this study were: (1) to assess climate variability and

land-cover change in the oasis, (2) to examine the long-term spatio-temporal

variations of groundwater level and quality on annual scale and seasonal scale, and (3)

to evaluate the influence of environmental drivers on groundwater level and quality

over time (i.e. air temperature, precipitation, streamflow, actual evapotranspiration,

normalized difference vegetation index, and cropland area). The results would be

expected to improve our understanding of groundwater cycling process in the oasis of

dried-up river basin in arid areas, and provide theoretical foundation for sustainable

groundwater resource management in the Tarim Basin.

2. Study area

The Tarim River Basin (73.40-93.65 ºE and 34.80-43.35 ºN) is the largest inland

river basin in the arid region of northwest China (Fig. 1a). The Basin is flanked by the

Kunlun Mountains to the south and Tianshan Mountains to the north, with an area of

1.02 × 106 km2 (Fig. 1b). Our study area (Wei-Ku Oasis) is located in the middle and

lower reaches of the Weigan-Kuqa River basin in northern Tarim Basin (Fig. 1b), and

natural desert ecosystem and agricultural economic development are heavily

dependent on groundwater in the Oasis (Chen et al., 2019). The Weigan-Kuqa River

7
basin, which originated in the south of Tianshan Mountains, is mainly composed of

the Weigan River and Kuqa River, and finally disappeared near the northern bank of

Tarim River. The Weigan River has five tributaries: the Muzati River, Kapusilang

River, Televichuk River, Karasu River and Kizil River (Fig. 1b). The annual

streamflow is about 26.6×108 m3 for the Weigan River and 4.6×108 m3 for the Kuqa

River, and the flood season is from June to September, accounting for 56.3% of the

total.

The Wei-Ku Oasis (82.09-83.47 ºE and 40.90-41.85 ºN) consists of Kuqa, Shaya

and Xinhe counties, and covers approximately 7104 km2 (Fig. 1c). The elevation

declines from the north to the south, and ranges from 945 to 1147 m above sea level.

This area is characterized by a typical temperate arid continental climate. The mean

annual air temperature is 11.4 ℃ and mean annual precipitation is 74.6 mm (data

from the Chinese National Meteorological Centre), and mean annual potential

evaporation is about 2401 mm with a strong seasonal dynamic (Chen et al., 2019).

Precipitation in this area has an uneven seasonal distribution, and more than 76% of

the overall precipitation occurs from May to September. The annual surface runoff

into the oasis is 31.2 × 108 m³, and more than 90% of the river water is diverted for

irrigation (Zhang et al., 2014b), causing dried-up natural river courses in recent

decades. Groundwater level and quality data were collected from 22 groundwater

observation wells on the Wei-Ku Oasis, and the spatial distribution and geographic

information of the 22 monitoring wells are shown in Fig. 1c and Table 1. The depth to

groundwater level in this oasis ranges from 2 to 6 m, and the phreatic water

evaporation is strong, causing severe soil salinization. Based on the groundwater level

contours (Fig. 1d), groundwater levels generally decline gradually from northwestern

(higher elevation) to southeastern (lower elevation), indicating that the groundwater in

the Wei-Ku Oasis flows from the north to the south, and finally into the Tarim River.

The Wei-Ku Oasis is covered by the Quaternary sediments of various

hydrogeologic units (Fig. 1e), including fine sandstone, pebbly sandstone, silty fine

8
sandstone, fine siltstone, sandy gravel and sandstone. The Weigan and Kuqa river

transported large amount of eroded materials from the Tianshan Mountains into the

plain, and lead to the formation of alluvial fan. The southeastern part of the Wei-Ku

Oasis is alluvial sediment, mainly covered by fine sandstone with a thickness of 40-60

m, with shallow groundwater level and moderate water-abundance (Fig. 1e). The

northeastern part and northern margin of the oasis are proluvial sediments, mainly

covered by pebbly sandstone (very strong water-abundance) and sandstone (moderate

water-abundance), with deep groundwater level. The central part of the oasis is

alluvial-proluvial sediment, the grain size of sediments is decreasing from pebbly

sandstone to fine siltstone, and correspondingly the water abundance of aquifer

systems is decreasing from very strong water-abundance to weak water-abundance,

with shallow groundwater level (Fig. 1e). Moreover, in the vertical layers, coarse

material and fine material in the sediments were alternately deposited due to the

runoff variations of the Weigan River and Kuqa River (Fig. 1f). The spatial

heterogeneity of hydrogeologic conditions over the oasis impacts the occurrence,

reserve, distribution, and flow of subsurface water (Wang et al., 2013). Furthermore,

the Wei-Ku Oasis is dominated by agriculture, and the main crops are cotton, corn

and winter wheat (Chen et al., 2019). In the oasis, the dominated natural plants are

Populus euphratica, Tamarix spp., Halostachys caspica, Phragmites communis,

Alhagi pseudalhagi, Karelinia caspica, Glycyrrhiza, and Sophora alopecuroides

(Zhao, 2006). In 2018, the cropland, grassland, forest, urban areas and bare areas

accounted for 62.5%, 17.9%, 5.1%, 1.1% and 13.0%, respectively, of the total area of

this oasis (Table 3).

9
Fig. 1 The location of (a) the Tarim Basin, (b) study region and (c) groundwater monitoring well

sites in the Wei-Ku Oasis, and (d) groundwater level contours based on its average value during

2000-2018 (in meters), (e) regional hydrogeological map, and (f) the log of typical borehole. A:

fine sandstone, moderate water-abundance; B: silty fine sandstone, weak water-abundance; C:

pebbly gravel, very strong water-abundance; D: sandy gravel and sandstone, strong

10
water-abundance; E: fine sandstone, moderate water-abundance; F: fine siltstone, weak

water-abundance; G: pebbly sandstone, very strong water-abundance; H: sandy gravel and

sandstone, moderate water-abundance; I: silty fine sandstone, weak water-abundance.

Table 1

Basic information of the 22 groundwater monitoring wells on the Wei-Ku Oasis in the Tarim

Basin.
Lon Lat Alt Depth MAT MAP AET NDVI LCT DGL DM
Well ID
(ºE) (ºN) (m) (m) (℃) (mm) (mm) (m) (g/L)

W1 82.753 41.653 1021.85 8.50 9.9 116.8 115.7 0.42 Crop 4.47 3.86

W2 82.417 41.547 1005.24 8.50 9.3 131.5 127.4 0.57 Crop 3.95 3.80

W3 82.529 41.550 1010.75 13.20 9.9 116.8 111.4 0.52 Crop 3.24 3.50

W4 82.618 41.549 1014.17 13.00 9.9 116.8 111.4 0.65 Urban 5.00 3.28

W5 82.919 41.569 1006.56 8.70 9.9 116.8 115.7 0.66 Crop 3.75 3.30

W6 82.998 41.593 1004.50 9.00 9.9 116.8 115.7 0.31 Crop 2.97 3.19

W7 82.419 41.487 1006.82 8.50 11.4 70.0 90.3 0.44 Crop 4.59 3.33

W8 82.532 41.528 1008.25 13.50 9.9 116.8 111.4 0.62 Crop 5.68 3.26

W9 82.253 41.358 987.59 11.50 11.4 70.0 90.3 0.65 Crop 3.26 3.17

W10 82.428 41.429 996.48 9.00 11.4 70.0 90.3 0.54 Crop 3.41 3.08

W11 82.614 41.427 997.85 13.40 11.7 62.8 98.2 0.59 Crop 3.16 3.33

W12 82.786 41.438 991.00 8.50 11.7 62.8 93.6 0.70 Crop 5.21 3.16

W13 83.135 41.496 979.77 9.00 12.1 56.1 93.3 0.70 Crop 4.06 3.27

W14 83.197 41.404 971.18 8.50 12.1 56.1 93.3 0.64 Crop 5.53 3.28

W15 83.051 41.371 981.86 8.70 12.1 56.1 93.3 0.72 Crop 4.14 3.10

W16 83.032 41.267 968.02 8.70 12.1 56.1 93.3 0.14 Grass 5.26 3.13

W17 82.481 41.253 981.76 8.00 11.4 70.0 90.3 0.15 Grass 3.45 3.17

W18 82.725 41.231 981.47 8.70 11.7 62.8 95.4 0.57 Crop 3.82 3.19

W19 82.928 41.253 976.06 9.00 11.7 62.8 93.6 0.63 Crop 2.67 3.07

W20 82.818 41.115 971.95 8.50 11.7 62.8 85.0 0.64 Crop 2.27 3.86

W21 82.746 41.068 968.91 9.00 11.7 62.8 95.4 0.72 Crop 2.08 3.80

W22 82.903 41.067 966.74 8.50 11.7 62.8 85.0 0.67 Crop 3.22 3.50

Lon: longitude; Lat: latitude; Alt: altitude; Depth: well depth; MAT: mean annual air temperature; MAP: mean annual

precipitation; AET: actual evapotranspiration; NDVI: normalized difference vegetation index; LCT: land cover type; Crop:

cropland; Grass: grassland; Urban: urban areas; DGL: depth to groundwater level; DM: degree of mineralization. The

statistical period for all the variables was from 2000 to 2018.

3. Methodology

3.1. Data

In this study, we concentrate on the time period from January 2000 to December

11
2018. Data used in the study included hydrological observations (streamflow,

groundwater level and mineralization degree), meteorological data (air temperature,

precipitation, actual evapotranspiration), vegetation cover data (NDVI), and land use

and land cover data (LULC).

3.1.1. Hydrological data

Daily streamflow data of the two tributaries (Weigan River, Kuqa River) in the

upper reached of the Weigan-Kuqa River basin from 2000 to 2018 were obtained

from the Weigan River Basin Authority. Streamflow was measured daily by

hydrological observational stations at the outlets of the mountainous sub-basins. In

addition, monthly groundwater level and quality data from 2000 to 2018 for the

Wei-Ku Oasis were collected from the Water-salt Monitoring Station of the Weigan

River Basin Authority, China. Groundwater was observed from dates throughout the

year. For each groundwater monitoring well (Fig. 1c), groundwater level was

manually observed three times on the 5th, 15th and 25th of each month, and

groundwater samples were collected once every season. The water quality

measurements of samples were undertaken at the Water-salt Monitoring Station of the

Weigan River Basin Authority, and the analytical detection limit for ions was < 0.1

mg/L. The basic information of the 22 groundwater monitoring wells on the oasis was

shown in Table 1, which observed only the water level and quality dynamics for

shallow groundwater. In this study, the groundwater quality parameter of interest was

mineralization degree, which represented salts. The groundwater level and

mineralization degree data were interpolated into a 250 m resolution grid data using

the inverse distance weighting (IDW) method to keep the spatial resolution identical

with other data.

3.1.2. Meteorological data

Monthly mean air temperature and monthly precipitation data were all obtained

from the Chinese National Meteorological Information Center

(http://cdc.cma.gov.cn), which provided China datasets of monthly precipitation and

12
temperature from 1961 to present with a spatial resolution of 0.5º × 0.5º and a

monthly temporal resolution. This data was validated by error analysis and

cross-validation with gauge-based observations, indicating good quality. The above

temperature and precipitation grid data in the Wei-Ku Oasis from 2000 to 2018 were

used in this study.

The amount of actual evapotranspiration (AET) strongly depends on vegetation

covers, climatic variables and climate forcing factors (e.g., greenhouse gases, ozone,

CO2) (Liu et al., 2016). Long-term climate variability, LULC change and related

management strategies are expected to have an influence on regional AET, thus

affecting groundwater recharge and discharge (Berihun et al., 2019). Therefore,

evaluating long-term effects of AET on groundwater is extremely important. In this

study, monthly AET data in the Wei-Ku Oasis during 2000 to 2018 were sourced

from the GLEAM (Global Land Evaporation Amsterdam Model,

http://www.gleam.eu) v3.3a dataset with a spatial resolution of 0.25º and a monthly

temporal resolution, which has been verified to agree well with land surface flux

observation results in China, particularly in arid northwestern China (both for the

areas with/without vegetation) (Ma et al., 2019b; Wu et al., 2020). The datasets are

based on reanalysis radiation and air temperature, reanalysis and satellite-based

precipitation, a combination of gauge-based, and satellite-based vegetation optical

depth, which have higher temporal and spatial resolutions and match better with data

requirements in large-scale research compared with other AET data products (Wu et

al., 2020). Monthly AET data was accumulated into the seasonal AET and annual

AET from 2000 to 2018. In order to maintain consistency with other data used in this

study, all the temperature, precipitation and AET data were resampled into a spatial

resolution of 250 m using the bilinear interpolation method by ArcGIS 10.5 software.

3.1.3. Vegetation data

NDVI (normalized difference vegetation index) was used in this study to

describe vegetation cover due to its good correlation with vegetation cover (An et al.,

13
2015). NDVI data for the study area from 2000 to 2018 were derived from the

MODIS NDVI product (MOD13Q1) provided by NASA (National Aeronautics and

Space Administration, http://ladsweb.nascom.nasa.gov/data/search.html), with a

spatial resolution of 250 m and a 16-day temporal resolution, orbital number h24v04.

This dataset has been widely used in ecohydrological studies in semi-arid and arid

regions of northern China due to its finer resolution and high reliability (An et al.,

2015; Mo et al., 2019). In this study, the maximum value composite (MVC) method

was used to determine the monthly NDVI and annual NDVI values for the 16-day

MODIS NDVI product (2 images per month, 23 images per year), which represents

the growth of vegetation in the month or year. Then, the seasonal NDVI was

determined by averaging the monthly NDVI for the three months in each season,

which represents the growth of vegetation in the season.

3.1.4. Land use data

The annual LULC (land use and land cover) data used in this study were derived

from the ESA CCI LC products, 300 m resolution, 2000 to 2015

(http://maps.elie.ucl.ac.be/CCI/viewer/), and the C3S Global Land Cover products,

300 m resolution, 2016 to 2018 (http://cds.climate.copernicus.eu/), which is a spatial

data layer of land use types denoted in grid. The ESA CCI LC products were released

by the European Space Agency (ESA) Climate Change Initiative (CCI), which was an

annual global land-cover time-series from 1992 to 2015 with a 300 m resolution. This

dataset has fine classification, long time series and high spatial resolution, and is

suitable for the large-scale and long-term land cover change study, which was found

to have good applicability in China (Yang et al., 2017). Further, the C3S global land

cover products for the years 2016-2018 were released by the Copernicus Climate

Change Service (C3S) in 2019, and described the land surface into 22 classes (United

Nations Land Cover Classification System, LCCS), which were consistent with the

ESA CCI LC products to ensure continuity. Then, the 22 LCCS land use classes were

grouped into 6 IPCC land categories for ear year, including cropland, forest,

14
grassland, urban areas, bare areas and water bodies. Finally, the LULC type maps for

the years 2000-2018 were produced for further analysis.

3.2. Statistical analyses

Data statistical analyses were performed using the ArcGIS (version 10.5), IBM

SPSS software package (version 22.0) and MATLAB (version R2018a), and the

statistical significance level was 5% or 1%. The Sen’s slope method was used to

evaluate the variation trends and amplitudes of groundwater, hydrometeorological and

vegetation variables. The inverse distance weighting (IDW) method was used to

interpolate the borehole observations of groundwater level and mineralization degree

into a 250 m resolution grid data. Pearson’s correlation analysis was conducted to

evaluate the correlations between groundwater and forcing factors. Multiple linear

stepwise regression analysis was used to determine the key controlling factors of the

groundwater level and mineralization degree. Spatial multiple linear regression

analysis was used to identify the dominant factors of groundwater level and

mineralization degree for each pixel.

Percentage variation of individual LULC types was evaluated to describe the

extent of LULC variation between two time points using Eq. (1) (Long et al., 2009):

Percent variation (%) = (A2 ― A1

A1 ) × 100 (1)

where A1 and A2 are the area of a given LULC type in year 1 and year 2,

respectively.

The spatial and temporal variations of hydroclimatic and vegetation variables

were evaluated using zonal statistics and the Sen’s slope method (Sen, 1968), which

has been widely applied to analyze variation of long-term hydrometeorological time

series (Berihun et al., 2019). The Sen’s slope method can evaluate the variation trend

and amplitude of time-series data based on the median of the series of slopes, and

could avoid or reduce the influence of data omissions and anomalies (Wu et al.,

2020). The formula is given as Eq. (2) (Sen, 1968):

15
 (Xj ― Xi)
Senij = MEDIAN (2)
(j ― i)
where Senij is the Sen’s slope; Xi and Xj are the sequential values corresponding to

times i and j, respectively, where 1 < i < j < n, n is the length of the time series.

When Sen > 0, the time series show an increasing trend, vice versa.

Pearson’s correlation analysis was conducted to assess the relationship between

environmental factors and oasis groundwater level and mineralization degree. For two
elements x and y, if their sample values are 𝑥𝑖 and 𝑦𝑖 (i = 1, 2, ..., n), the correlation

coefficient between them is defined as (Eq. (3)):

𝑛
∑𝑖 = 1(𝑥𝑖 ― 𝑥)(𝑦𝑖 ― 𝑦)
r𝑥𝑦 = 𝑛 𝑛
(3)
∑𝑖 = 1(𝑥𝑖 ― 𝑥)2 ∑𝑖 = 1(𝑦𝑖 ― 𝑦)2

where 𝑟𝑥𝑦 is the correlation coefficient between x and y; 𝑥 is the average value of

the sample values for x; 𝑦 is the average value of the sample values for y.

The key controlling factors of oasis groundwater level and mineralization degree

were evaluated using the multiple linear stepwise regression analysis by setting

averages of driving factors as independent variables at the monthly, seasonal and

annual time scales. For the reasonable regression fittings, Akaike information

criterion (AIC) values were calculated step by step under various parameter settings to

evaluate the reliability of model capacity and parameter selection (Wu et al., 2020).

During the process of stepwise regression analysis, this study used the whole time

series data (2000-2018) for exploring the driven module (Eqs. (4) and (5)), and the

raw data was standardized firstly:

DGL = (∑b ∗ X ) + ε
i i (4)

DM = (∑b ∗ X ) + ε
i i (5)

where DGL and DM are the depth to groundwater level and mineralization degree,
respectively; 𝑏𝑖 is the regression coefficient; 𝑋𝑖 is the key controlling factor

(climatic, hydrological and vegetation variables); 𝜀 is a constant term.

Standardized partial regression coefficient is a relative number without unit,

16
could be used to estimate the direct effect of independent variables on dependent

variables, and compare the relative importance of independent variables. The formula

is given as Eq. (6):

𝜎𝑋𝑖
𝑘𝑖 = 𝑏𝑖 ∗ 𝜎𝑌
(6)

where 𝑘𝑖 is the standardized partial regression coefficient; b𝑖 is the partial

regression coefficient; 𝜎𝑋𝑖 is the standard deviation of independent variables

(climatic, hydrological and vegetation variables); 𝜎𝑌 is the standard deviation of

dependent variables (DGL and DM).

Inverse distance weighting (IDW) method was used to generate a continuous

surface of groundwater level and quality based on the borehole observations data,

which is one of the most important interpolation methods recognized in geostatistical

analysis (Haldar et al., 2020; Nistor et al., 2020). The IDW is based on the first law of

geography, and assumes the proportionality of similarity and correlation between

neighbors and the distance between them. The cell values are evaluated by linear

weighted combination of sample point dataset, and the weighting is a function of

inverse distance, that is, the weighting of farther known points is less than that for

nearer known points. The accuracy of the IDW is influenced by the power parameter

p (Burrough and McDonnell,1998), a lower p value has more effect on the

surrounding region of the farther points, while a higher p value can define a detailed

surface on the nearest points. The formulas for the IDW are given as Eqs. (7)-(9)

(Nistor et al., 2020):

𝑁

𝑍= ∑𝑤 𝑍
𝑖=1
𝑖 𝑖 (7)

𝑑𝑖―𝑝
𝑤𝑖 = 𝑁
(8)
∑𝑖 = 1𝑑𝑖―𝑝
𝑁
∑ 𝑖=1
𝑤𝑖 = 1 (9)

where Z is the estimation value for the unknown point; 𝑍𝑖 is the sample value in

17
known point i; N is the amount of known points; 𝑤𝑖 is the weighting of each known

points; 𝑑𝑖 is the distance from each known points to the unknown point; p is the

power, and generally assumed as 2. In this study, N was 12, and p was 2.

4. Results

4.1. Variations of climate and land-cover

An overview of the temporal analysis was shown in Fig. 2, presenting annual and

monthly averages of air temperature, precipitation, and streamflow in the oasis from

2000 to 2018. The annual mean air temperature had a decreasing trend with a mean

rate of about -0.12 ℃ /10a over the period from 2000 to 2018, while annual

precipitation and streamflow exhibited an increasing trend with a mean rate of about

12 mm/10a and 0.30 × 108 m³/10a during this period, respectively (Fig. 2 and Table

2). In addition, the variation trends of hydrometeorological factors (air temperature,

precipitation, and streamflow) differed between the four seasons (Table 2). Air

temperature significantly decreased in summer (-0.33 ℃ /10a) and autumn (-0.57 ℃

/10a) from 2000 to 2018, while increased in spring (0.20 ℃ /10a). Precipitation

increased in summer (10.31 mm/10a) and autumn (3.27 mm/10a), while decreased in

spring (-0.86 mm/10a) and winter (-1.00 mm/10a). Streamflow increased in spring

(0.11 × 108 m³/10a) and summer (1.04 × 108 m³/10a), while decreased in autumn

(-0.47 × 108 m³/10a) and winter (-0.53 × 108 m³/10a), due to the regulation of

mountain reservoirs (Chen et al., 2019). Furthermore, there were diverse intra-annual

variation characteristics for the three hydrometeorological factors in the oasis during

2000 to 2018 (Fig. 2g, h, and i). As shown in Fig. 2i, the monthly streamflow of the

Weigan-Kuqa river showed some large fluctuations, the highest value appeared in

July (6.19×108 m³), and the lowest value occurred in January (0.5×108 m³). The

intra-annual distribution of streamflow showed three peaks, occurred in March, July

and November, respectively, due to reservoir regulation. The peak air temperature

occurred in July (25.4℃), and the lowest value occurred in January (-8.1℃) (Fig. 2g).

The maximum value of monthly precipitation occurred in June (15.2 mm), and the

18
minimum value occurred in March (0.8 mm) (Fig. 2h).

Fig. 2 Inter-annual variations (a-f) and intra-annual distributions (g-l) of air temperature (a and g),

precipitation (b and h), streamflow (c and i), NDVI (d and j), DGL (e and k), and DM (f and l) in

the Wei-Ku Oasis (2000-2018). NDVI: normalized difference vegetation index; DGL: depth to

groundwater level; DM: degree of mineralization.

Table 2

Sen’s slope tested variation of air temperature, precipitation, streamflow, NDVI, depth to

groundwater level, and degree of mineralization in the Wei-Ku Oasis from 2000 to 2018.
Annual Spring Summer Autumn Winter

T (℃/10a) -0.12 0.20 -0.33+ -0.57* 0.00

19
 P (mm/10a) 12.00 -0.86 10.31 3.27 -1.00

Q (108 m³/10a) 0.30 0.11 1.04 -0.47 -0.53*

NDVI (/10a) 0.10*** 0.03** 0.10*** 0.06*** 0.03**

DGL (m/10a) 2.05*** 2.15*** 1.96*** 1.94*** 2.01***

DM ((g/L)/10a) -0.21*** -0.41*** -0.28* -0.17 -0.22+

Asterisks indicate level of significance: P<0.1(+); P<0.05(*); P<0.01(**); P<0.001(***). T: air temperature; P: precipitation; Q:

streamflow; NDVI: normalized difference vegetation index; DGL: depth to groundwater level; DM: degree of mineralization.

The spatial distributions of LULC for 2000, 2010, and 2018 were prepared for

the Wei-Ku Oasis (Fig. 3a, b, and c). As shown in Fig. 3, cropland was mainly

distributed in the west and northeast of the oasis, grassland mainly in the middle and

southeast, forest mainly on the edge of the cropland (mainly as artificial forest), and

bare land mainly on the edge of the oasis (i.e. the transition zone between oasis and

desert) (Fig. 3a, b, and c). In general, during the entire study period, the area of

cropland, forest and urban land gradually increased, while the area of grassland, bare

land and water bodies gradually decreased as time goes on. Cropland, forest and

urban areas continuously increased to 62.5%, 5.1% and 1.1%, respectively, in 2018

(Table 3). Conversely, grassland, bare areas and water bodies continuously decreased

to 17.9%, 13.0% and 0.4%, respectively, in 2018 (Table 3). Furthermore, the

dominant LULC type was cropland over the entire period, covering more than half of

the oasis area (55.9%, 60.6%, and 62.5% in 2000, 2010 and 2018, respectively),

resulting in high water demand. Grassland was second in dominance (20.3%, 18.9%,

and 17.9% in 2000, 2010 and 2018, respectively), followed by bare areas

(20.2%,15.2%, and 13.0% in 2000, 2010 and 2018, respectively) (Fig. 3 and Table 3).

Compared to other LULC types, urban areas accounted for the least coverage (0.3%)

in 2000, but showed a dramatic increase during 2000 to 2018. The highest percentage

change was observed in urban areas, which increased by 137.0% between 2000 and

2010 and by 68.1% between 2010 and 2018 (Table 3). On the other hand, bare areas

decreased markedly between 2000 and 2018, by 35.4%, and grassland underwent a

general decrease by 11.9% between 2000 and 2018, mainly due to farmland

expansion in the study area. In contrast, cropland exhibited a persistent increase by

20
8.4% from 2000 to 2010 and by 11.9% from 2000 to 2018 (Table 3), indicating that a

considerable area of bare land and grassland was converted into cropland and urban

land in the oasis throughout the entire study period (Zhang et al., 2014b). However,

the expansion of farmland had obviously slowed by 2018 in the Wei-Ku Oasis (Table

3). Forest area also underwent a conspicuous increase by 72.1% from 2000 to 2018 in

the oasis, including natural forest and artificial forest (artificial ecological forest and

economic forest). Water bodies had remained unchanged from 2000 to 2010, but

underwent a general decrease by 8.3% between 2010 and 2018.

Fig. 3 Distributions of annual averaged LULC, NDVI, DGL and DM across the Wei-Ku Oasis:

(a-c) LULC, (d-f) NDVI, (g-i) DGL, and (j-l) DM for 2000 (a, d, g, and j), 2010 (b, e, h, and k),

21
 and 2018 (c, f, i, and l), respectively. LULC: land use and land cover; NDVI: normalized

difference vegetation index; DGL: depth to groundwater level; DM: degree of mineralization.

Table 3

Area of different land cover types in the Wei-Ku Oasis from 2000 to 2018 (unit: km2).

Variation
LULC 2000 2010 2018
In 2000-2010 In 2010-2018 In 2000-2018

Cropland 3969.3 (55.9%) 4302.8 (60.6%) 4442.0 (62.5%) +333.5 (8.4%) +139.2 (3.2%) +472.7 (11.9%)

Forest 211.0 (3.0%) 303.0 (4.3%) 363.1 (5.1%) +92.0 (43.6%) +60.1 (19.8%) +152.1 (72.1%)

Grassland 1441.2 (20.3%) 1344.6 (18.9%) 1270.1 (17.9%) -96.6 (6.7%) -74.5 (5.5%) -171.1 (11.9%)

Urban areas 18.9 (0.3%) 44.8 (0.6%) 75.3 (1.1%) +25.9 (137.0%) +30.5 (68.1%) +56.4 (298.4%)

Bare areas 1432.7 (20.2%) 1077.8 (15.2%) 925.1 (13.0%) -354.9 (24.8%) -152.7 (14.2%) -507.6 (35.4%)

Water bodies 31.4 (0.4%) 31.4 (0.4%) 28.8 (0.4%) 0.0 (0.0%) -2.6 (8.3%) -2.6 (8.3%)

LULC: land use and land cover.

Annual mean NDVI in the oasis exhibited a significant increasing trend at the

0.1% significance level with a mean rate of about 0.10 per decade over the period

from 2000 to 2018 (Fig. 2d and Table 2). For the oasis, NDVI showed increasing

trend for all seasons at least at the 1% significance level (Table 2), the rate of NDVI

increase was the largest in summer (0.10 /10a), followed by autumn (0.06 /10a), and

the smallest in spring and winter (0.03 /10a). As shown in Fig. 2j, NDVI in the study

area displayed a noticeable seasonal distribution. The maximum value of monthly

NDVI appeared in August (0.42), the minimum value occurred in February (0.10),

and the values from December to March were very close (< 0.15). In addition, Fig. 3

used colour gradients to exhibit details in the spatial distribution of NDVI over the

oasis for 2000, 2010 and 2018 (Fig. 3d, e, and f). The figures exhibited a decreasing

gradient for NDVI from north-west to the south-east for almost all years, and the

sparsely vegetated area was mainly located on the edge of the oasis and in certain

localized areas of the central oasis (mainly as bare land). Vegetation cover in the

northwestern and central parts changed more obviously than that in the southeastern

part. Furthermore, Fig. 4 presented the spatiotemporal variation of NDVI across the

oasis for annual average and four seasons during 2000 to 2018 (Fig. 4k, l, m, n, and

o). Overall, NDVI increased significantly in most parts of the entire oasis (p < 0.05),

22
while decreased noticeably in sporadic areas of the northeastern and southeastern

parts (Fig. 4k). For most of the study area, there was significant spatiotemporal

variation in the NDVI in all seasons (p < 0.05) (Fig. 4l, m, n, and o). In summer, the

LULC type with the largest increase of NDVI was mainly cultivated land, with a rate

of (0.25 - 0.55)/10a (Fig. 4m). In spring and winter, the LULC types with the

significant changes of NDVI were mainly cultivated land and bare land, with a rate of

(-0.05 - 0.10)/10a (Fig. 4l and o). In autumn, the variation of NDVI in the

southeastern part was significantly smaller than that in other regions (Fig. 4n).

23
Fig. 4 Sen’s slope tested variation of DGL (a-e), DM (f-j), and NDVI (k-o) across the Wei-Ku

24
Oasis for annual average (a, f, and k), spring (b, g, and l), summer (c, h, and m), autumn (d, i, and

n), and winter (e, j, and o) during 2000 to 2018. DGL: depth to groundwater level; DM: degree of

mineralization; NDVI: normalized difference vegetation index.

4.2. Spatio-temporal variations of groundwater

The distributions of DGL and DM across the Wei-Ku Oasis for annual average,

spring, summer, autumn, and winter during 2000 to 2018 were shown in Fig. 5. Both

DGL and DM showed significant spatial variability in the oasis, and the spatial

distribution pattern of DGL and DM varied among the four seasons. DGL ranged

from 0.4 to 9.6 m, the maximum value was observed at well W8 in the summer of

2017, and the minimum value was observed at well W21 in the spring of 2009, maybe

due to underlying surface conditions, hydrometeorological conditions and human

activities (Wang et al., 2014).The DM for the four seasons spanned a relatively wide

range (from 0.4 to 16.0 g/L) during 2000 to 2018, the maximum and minimum DM

were observed at well W22 in the autumn of 2017 and W1 in the winter of 2002,

respectively. The normal operation of the Paman Reservoir caused a visible decrease

in DGL near well W21, which was 0.3-1.2 m smaller than the DGL of ambient wells

(W20 and W22) in all seasons. Across most of the oasis, the DGL was greater than

3.0 m; however, there were many areas where it was between 3.5 and 5.0 m (Fig. 4a).

In general, in the east-central region of the oasis there were several sizable areas

where the DGL exceeded 5.0 m and reached a maximum depth of about 6.5 m in

some locations, while in the southwest there was a sizable area where the DGL was

below 3.0 m and reached a minimum depth of about 1.5 m in some locations due to

reservoir influence. Overall, there appeared to be a pattern by which the regions with

the deepest groundwater level also experienced the largest seasonal variations in the

DGL (Fig. 5a, b, c, d, and e). For the oasis, DGL was the smallest in spring, followed

by winter, and the largest in summer and autumn (Fig. 5b, c, d, and e). Moreover,

across most of the oasis, the DM was less than 4 g/L, and was relatively lower in the

northwestern and central parts of the oasis (below 2 g/L) (Fig. 5k). Lowest

25
concentrations of DM were in the northwestern part of the oasis (0.4 g/L), and the

highest concentrations were in the southeastern part (up to 13 g/L; Fig. 5k) during all

seasons. The DM in spring was significantly lower than that in the other seasons for

the northwestern and central parts, while the differences of DM between the four

seasons were not significant for the other regions of our study area (Fig. 5l, m, n, and

o).

Fig. 5 (a-e) depth to groundwater level (DGL), (f-j) DGL and DM, and (k-o) degree of

mineralization (DM) across the Wei-Ku Oasis for annual average (a, f, and k), spring (b, g, and l),

summer (c, h, and m), autumn (d, i, and n), and winter (e, j, and o) during 2000 to 2018.

The trend statistics of groundwater level and quality were shown in Fig. 2. The

DGL in the Wei-Ku Oasis showed a significant continuously increasing trend of about

4.1 m on average across the oasis from 2000 to 2018 (Fig. 2e), while the DM showed

a noticeable decreasing trend of about 0.42 g/L on average during the same period

(Fig. 2f). However, variations of DGL and DM were not significant after 2015. In the

seasonal analysis, 22 wells showed a downward trend for groundwater level and

quality in all seasons, and the decreasing trends between the four seasons were close.

In Fig. 2, it was clear that the degree of mineralization and groundwater level showed

the most significant decrease in the spring months (April and May), decreasing by

26
about 0.82 g/L and 4.3 m, respectively; while the decrease was the least in the autumn

months (October and November). Furthermore, the mean annual DGL varied from 2.4

to 5.7 m with mean value of 4.0 m, and DM varied from 3.1 to 3.9 g/L with mean

value of 3.3 g/L during 2000-2018 period (Fig. 2e and f). Moreover, the intra-annual

distributions of the DGL and DM averaged over the entire oasis were shown in Fig.

2k and l. The DGL developed a pronounced seasonal cycle in the oasis, while the DM

exhibited a noticeably weaker season cycle. In general, the intra-annual distribution of

DGL showed two peaks, the first peak appeared in April and the other in December

(Fig. 2k). The groundwater level remained too shallow in spring (3.6 m in April),

while too deep in autumn (4.2 m in October). The DM in groundwater was lower in

winter (3.2 g/L) than that in the other seasons, but the difference of DM was not

significant between seasons (Fig. 2l).

Values of DGL and DM in our study area significantly varied in space and time

(Fig. 3). It was evident that the DGL across this oasis was greater in 2018 (Fig. 3i)

than that in 2010 (Fig. 3h) or 2000 (Fig. 3g). In contrast, the DM of groundwater was

lower in 2018 (Fig. 3l) than that in 2010 (Fig. 3k) or 2000 (Fig. 3j). Overall, from

2000 to 2018, there appeared to be a pattern by which the areas with the smallest

DGL experienced the smallest increase in the DGL, and vice versa (Fig. 3g, h, and i);

while the areas with the highest DM had the largest decrease in the DM, and vice

versa (Fig. 3j, k, and l). During the study period (2000-2018), the DGL increased

significantly in most areas of the oasis (increased by about 4.1 m or more), and the

increase in the southwest was significantly smaller than that in other regions. In 2000,

the spatial difference of DGL was not obvious, and the DGL was relatively small

across the entire oasis (about below 3.0 m), except for a sizable area in the northwest

(Fig. 3g). In 2010, DGL presented significant spatial differences, and the lower value

mainly occurred in the southwest part, the higher value in the east and central parts

(Fig. 3h). Furthermore, the southwestern part had consistently higher DM compared

to other parts in the oasis during the entire period (Fig. 3j, k, and l). Overall, DM

27
displayed obvious spatial variability for 2000, 2010, and 2018, and decreased

significantly in most parts of the entire oasis. In 2000, the spatial difference of DM

was significant, the minimum value of DM was in the northwest and central parts of

the oasis, while the maximum value in the southwest (Fig. 3j). Likewise, in 2010, the

maximum value of DM was in the southwest and northeast, the lower value in the

northwest and central areas (Fig. 3k).

The spatiotemporal variations of DGL and DM across the oasis for annual

average and four seasons from 2000 to 2018 were shown in Fig. 4 and Table 2.

During the period from 2000 to 2018, the average rate of DGL variation (Sen’s slope

estimator) in the whole oasis was about 2.05 m per decade (p < 0.001), with the

greatest rate in spring (2.15 m/10a) and the smallest in autumn (1.94 m/10a) (Table

2). Sen’s slope tested variation showed a noticeable increase in the DGL across the

entire oasis at the 5% significance level, with changes of more than 3 m/10a in areas

where the groundwater level was deep (> 5 m), while small increase (< 1 m/10a)

where the DGL was shallow (Fig. 4a, b, c, d, and e). That is, the regions that are the

most sensitive and experience the largest changes were the regions where the DGL

was already relatively great, and vice versa (Fig. 4 and 5). Overall, the variation of

DGL in the eastern region of our study area was significantly larger than that in the

western region (Fig. 4a). For the seasonal analysis of groundwater level, 22 wells

showed a downward trend in all seasons at the 5% significance level. The increase of

DGL was noticeably greater in spring and winter than that in summer and autumn for

the west part (W9, W17, and W18), was the greatest in spring and smallest in winter

for the east part (W14 and W15), and was the largest in spring and smallest in

summer and autumn for the northeastern part (W6) (Fig. 4b, c, d, and e).

Furthermore, the variation of DM across the study area during 2000 to 2018

presented great spatial heterogeneity, with no clear pattern emerging, maybe due to

the impact of irrigation and drainage network (Fig. 4f, g, h, i, and j). As shown in

Table 2, over the entire study period, the mean decreasing rate of DM (Sen’s slope

28
estimator) in the whole oasis was about -0.21 g/L per decade at the 0.1% significance

level, with the greatest in spring (-0.41 g/L per decade) and the smallest in autumn

(-0.17 g/L per decade). In summer and winter, the mean decreasing rate of 22 wells

was -0.28 g/L and -0.22 g/L per decade, respectively. Overall, the rate of DM change

at the northwest and southwest parts was positive (increasing trend) in four seasons,

whereas it in the western and eastern parts was negative (decreasing trend) in four

seasons, while the decreasing trend was not significant in the remaining sporadic

areas during the study period (Fig. 4f). It was evident that the area with significant

DM changes was noticeably larger in spring than that in the other seasons (Fig. 4g, h,

i, and j). In the analysis on seasonal scale, the decrease of DM was noticeably greater

in spring than that in other season for the western part of our study area (W9, W10,

and W17), while was greater in spring and autumn than that in summer and winter for

the south part (W22). Likewise, the increase of DM was significantly smaller in

spring than that in other seasons for the southwest and northwest parts (W3, W8, and

W18), while the decrease of DM was obviously smaller in winter than that in other

seasons for the northeast part (W6).

The relationship between groundwater level and mineralization degree over the

study period was prepared for the Wei-Ku Oasis (Fig. 5f, g, h, i, and j). As shown in

Fig. 5, a negative correlation (p < 0.01) between DGL and DM was observed across

the oasis for annual average (Fig. 5f) and four seasons (Fig. 5g, h, i, and j). That is,

the regions with a higher average groundwater level showed a noticeably greater DM,

and vice versa. On the annual scale, the fitting equation between DGL and DM across

the entire oasis was described as 𝐷𝑀 = 1.448 + 6.066𝐷𝐺𝐿 ―1

(𝑛 = 418, 𝑅2 = 0.081,𝑝 < 0.01) (Fig. 5f). However, this relationship during the study
period was generally weak in the study area, mainly due to the influence of strong

evaporation and drainage channels (Zhang et al., 2014b). Furthermore, in the analysis

on seasonal scale, the correlation coefficient between DGL and DM also varied

among the four seasons, and was greater in spring than that in the other seasons (Fig.

29
5g, h, i, and j).

4.3. Correlations between groundwater and climate and land-cover change

Pearson’s correlation analysis results showed that the environmental factors had

greater effects on DGL than DM, and the effects exhibited significant seasonal

heterogeneity in the study area during 2000 to 2018 (Table 4). DGL was significantly

positively correlated with precipitation, actual evapotranspiration, and cropland area,

while significantly negatively related to streamflow. Similarly, DM had a significant

positive relation with air temperature, while showed significant negative correlations

with precipitation and actual evapotranspiration. DM in all seasons had no statistical

correlation with streamflow and cropland area. Overall, compared with climate

drivers, anthropogenic activities had greater impacts on the decline of groundwater

level during the study period (Table 4). Among hydrometeorological factors, the

important factor was actual evapotranspiration (r=0.374, p<0.001, n=418) for DGL,

and air temperature (r=0.147, p<0.01, n=418) for DM. Among anthropogenic activity

factors, cropland area (r=0.602, p<0.001, n=418) was important factor for DGL.

Furthermore, cropland area in all seasons was significantly correlated with DGL, but

had no significant relationship with DM (Table 4). Streamflow was significantly

correlated with DGL in autumn and winter, but had no significant relationship with

DM in all seasons. NDVI was significantly correlated with DGL in summer and

winter, while had significant relationship with DM in spring and winter. Actual

evapotranspiration was significantly correlated with DGL in all seasons except winter,

but was significantly correlated with DM in spring and summer. Precipitation showed

a significant relationship with DGL in summer and winter, whereas was significantly

correlated with DM in all seasons except winter. Air temperature showed a significant

relationship with DGL in summer and autumn, while was significantly correlated with

DM in all seasons.
Table 4

Pearson’s correlation coefficients of groundwater and climate and land-cover factors.

30
 T Precipitation Streamflow AET NDVI Cropland
N
℃ mm 108 m³ mm km2
DGL Annual 418 / 0.17*** -0.09* 0.37*** / 0.60***
Spring 418 / / / 0.23*** / 0.60***
Summer 418 -0.10** 0.27*** / 0.41*** 0.09* 0.59***
Autumn 418 -0.19*** / -0.11** 0.33*** / 0.59***
Winter 418 / -0.12** -0.26*** / 0.23*** 0.58***
DM Annual 418 0.15*** -0.13*** / -0.09* / /
Spring 418 0.10** -0.09* / -0.09* -0.13** /
Summer 418 0.16*** -0.10** / -0.08* / /
Autumn 418 0.13*** -0.10* / / / /
Winter 418 0.08* / / / -0.10** /
A backslash (/) means not significant. N: number of samples; T: air temperature; AET: actual evapotranspiration; NDVI: normalized
difference vegetation index; Cropland: area of cropland; DGL: depth to groundwater level; DM: degree of mineralization.
* P < 0.1 (indicates level of significance).
** P < 0.05 (indicates level of significance).
*** P < 0.01 (indicates level of significance).

Furthermore, the stepwise linear regression analysis was used to identify the

major controlling factors for the decreasing groundwater level and mineralization

degree, and their seasonal differences in the Wei-Ku Oasis (Table 5). The stepwise

linear regression models for the annual mean DGL and DM across the whole oasis

were expressed as: DGL_Annual = ―21.12 + 0.54CROPLAND + 0.14AET (R2 =

0.38, p < 0.001, n = 418) and DM_Annual = ―1.54 + 0.15TEMPERATURE (R2 =

0.02, p < 0.005, n = 418). The cropland area and actual evapotranspiration were the

main controlling factors affecting groundwater level among the climate and

anthropogenic drivers, which both had positive effects. Nevertheless, air temperature

was the major factor of the mineralization degree of groundwater, which had a

positive effect. In addition, cropland area and actual evapotranspiration explained

important proportions of the DGL in spring and summer, while cropland area played

an important role in the DGL in autumn and winter (Table 5). The correlations

between groundwater and environmental variables confirmed that anthropogenic

activities had strong influences on the groundwater level in our study area (Table 4

and 5).
Table 5

Summary of stepwise linear regression of groundwater level and mineralization degree with

environmental factors in the Wei-Ku Oasis.

31
 Coefficients
B R Square p-value N
Cropland AET T

DGL_Annual -21.124 0.541 0.143 0.379 <0.001 418

DGL_Spring -25.050 0.577 0.145 0.378 <0.001 418

DGL_Summer -19.811 0.519 0.147 0.369 <0.001 418

DGL_Autumn -22.029 0.592 0.351 <0.001 418

DGL_Winter -22.807 0.582 0.338 <0.001 418

DM_Annual -1.544 0.147 0.022 <0.005 418

DGL: depth to groundwater level; DM: degree of mineralization; Cropland: area of cropland; AET: actual evapotranspiration; T:

air temperature; N: number of samples.

Moreover, the dominant factor influencing groundwater level and mineralization

degree also differed spatially in the Wei-Ku Oasis during the study period (Fig. 6). As

shown in Fig. 6, the dominant factor affecting groundwater level variation was NDVI

or actual evapotranspiration for most areas of the oasis, while was precipitation or air

temperature for very small areas (Fig. 6a). In contrast, the spatial difference of the

dominant factor of mineralization degree change was more significant. The dominant

factor of DM change was precipitation or actual evapotranspiration or NDVI for most

regions of the oasis, while was air temperature for some sporadic small areas (Fig.

6b). Generally, for the regions with the largest increase in NDVI (cropland, forest and

bare areas), NDVI was the dominant factor causing DGL change (Fig. 4k and Fig.

6a), probably due to artificial oasis expansion leading to more water consumption. For

the areas with insignificant variation in NDVI (bare land in the west and southeast

parts, cultivated land near the river channel), actual evapotranspiration was the

dominant factor of DGL change (Fig. 1b, Fig. 3a-3c, Fig. 4k, and Fig. 6a).

Precipitation was the dominant factor of DGL variation for the grassland in southeast

part (Fig. 3a-3c, Fig. 6a). For the areas with the largest decrease in NDVI (urban land

in the northern part), air temperature was the dominant factor of DGL change (Fig.

3a-3c, Fig. 4k, Fig. 6a). Furthermore, for the areas with very small DM (mainly in the

northwestern part), NDVI was the dominant factor of DM change, maybe due to the

impact of drainage channels (Fig. 1, Fig. 3l, and Fig. 6b). For the regions with the

largest increase in NDVI (cropland, bare land and forest in the northwest), actual

evapotranspiration was the dominant factor of DM (Fig. 3c, Fig. 4k, and Fig. 6b).

32
Precipitation was the dominant factor of DM for the areas with insignificant variation

of NDVI (grassland and bare land in the southeast) (Fig. 3c，Fig. 4k and 4o), while air

temperature was the dominant factor of DM for some sporadic areas in the northwest

part (urban areas and some grassland).

Fig. 6 The dominant factor of groundwater level (a) and mineralization degree (b) across the

Wei-Ku Oasis during 2000 to 2018. NDVI: normalized difference vegetation index; AET: actual

evapotranspiration.

5. Discussion

5.1. Variation characteristics of groundwater

The average annual DGL in the present study across the whole oasis varied from

2.4 to 5.7 m with mean value of 4.0 m, and DM varied from 3.1 to 3.9 g/L with mean

value of 3.3 g/L during 2000 to 2018. These data were consistent with the results by

Zhang et al. (Zhang et al., 2014b) for the Wei-Ku Oasis. However, the DGL values in

the present study were higher than those in the lower reaches of the Tarim River in

recent years (Sun et al., 2020). This could be explained by the rapid farmland

expansion in the oasis and the continuous ecological water conveyance in the lower

reaches of Tarim River (Chen et al., 2019). The rapid expansion of cultivated area

across the oasis in recent decades has led to a continuous decline of the groundwater

level, since the sharp increase of regional water consumption could cause

33
groundwater overexploitation (Chen et al., 2019). In contrast, since 2000, the

continuous ecological water conveyance to the lower reaches of Tarim River has

increased the recharge from the river channel to groundwater, which in turn has

caused a substantial increase of groundwater level in the downstream (Chen et al.,

2019).

Groundwater level and mineralization degree significantly decreased with time

(Fig. 2e and f). Our results exhibited that the annual DGL had a significant increasing

trend (about 2.05 m per decade, Table 2) across the whole oasis from 2000 to 2018,

while the annual DM showed a decreasing trend (about -0.21 g/L per decade, Table

2). Several studies in the Tarim Basin reported similar groundwater depletion (Chen

et al., 2019). This was consistent with the observations in the oases of the Hexi

Corridor in China (Feng et al., 2005; Chen, 2010; Huang and Pang, 2010) and other

arid regions around the world (Scanlon et al., 2006; Joodaki et al., 2014; Asoka et al.,

2017). Hu et al. (2019) reported that groundwater table declined with a rate of 13

mm/month in the western Australia during 1980 to 2015, and Ma et al. (2005, 2013)

discovered groundwater level decreased by 35 m in the Minqin Oasis and by 10 m in

the Dunhuang Oasis in northwest China in the past decades. The falling of

groundwater level in the study area could be explained by five aspects. First, climate

change plays an important role in groundwater dynamics (Beal et al., 2019;

Wittenberg et al., 2019). Elevated air temperature could enhance the duration and

intensity of evaporation, which in turn lead to lower groundwater level due to the

increased evaporation of soil water and groundwater (Wang et al., 2019b); While

elevated air temperature also could increase the melting water of glacier and snow in

mountains by enhancing melting period (Hori et al., 2017), and thus results in higher

groundwater level due to the supply from meltwater and river water (Chao et al.,

2020). The amount and spatio-temporal distribution of precipitation would determine

the quantity and distribution of regional water resources by affecting runoff formation

process, which in turn lead to groundwater level fluctuation owing to variations of

34
groundwater recharge (Zhang et al., 2014b; Eamus et al., 2015). Second, LULC

changes greatly impact groundwater dynamics in arid regions due to the great

differences in evapotranspiration and water consumption of different vegetation (Bai

et al., 2020a), and cultivated area expansion could lead to a dramatic increase in

irrigation water from river water and groundwater, thus causing groundwater level

decline due to reduced river infiltration and groundwater overexploitation (Chen et al.,

2019). Third, agricultural irrigation activities in arid regions could greatly impact

groundwater level and quality by affecting groundwater cycle process (Wang et al.,

2014). Field irrigation usually enhances the recharge from infiltrating irrigation water

to groundwater, increased channel water utilization coefficient could reduce channel

water leakage, and gradually improved drainage systems could reduce the

groundwater level (Porhemmat et al., 2018). Fourth, groundwater overexploitation

could lead to the falling of groundwater level and the depletion of groundwater

storage, mainly due to the direct pumping of shallow and deep groundwater to meet

regional water demand in arid regions (Asoka et al., 2017; Wang et al., 2019b). Fifth,

soaring population could greatly enhance domestic water consumption, and thus

results in the continuous falling of groundwater level due to limited surface water

resources and increased groundwater exploitation in the arid oasis area (Scanlon et al.,

2006).

Groundwater level and mineralization degree exhibited significant seasonal

differences in the arid oasis (Fig. 2 and 5). These were consistent with the results by

Bai et al. (Bai et al., 2020b) for the upper Tarim River and Wang et al. (2014) for the

lower Heihe River in northwestern China. These were reasonable due to the seasonal

deviations in climatic and anthropogenic factors. In fact, great differences of driving

factors exist among the seasons, such as climatic conditions (temperature,

precipitation, evapotranspiration, and streamflow), hydrological connectivity,

vegetation coverage, soil water content, groundwater extraction, and agricultural

irrigation activities (irrigation methods, irrigation water requirement, agricultural

35
drainage) (Fig. 2; Taylor et al., 2013; Porhemmat et al., 2018). Furthermore,

groundwater level was the shallowest in spring, and exhibited the most significant

decrease in spring (Fig. 2k and Table 2). This might be contributed to less

evapotranspiration due to lower air temperature, more recharge from snowmelt runoff,

and infiltration recharge due to spring flood irrigation from pumping deep

groundwater (Zhang et al., 2014b; Wang et al., 2014; Hu et al., 2019). In contrast,

groundwater level was the deepest in summer and autumn, and presented the least

decrease in autumn (Fig. 2k and Table 2), probably due to strong evapotranspiration,

agricultural drainage, groundwater extraction, increased crop water requirements

during the growing season, decreased irrigation water seepage by water-saving

irrigation, and reduced recharge from river water resulting from surface water transfer

for irrigation (Han et al., 2011; Porhemmat et al., 2018). Additionally, the

mineralization degree of groundwater in our study area was the highest in autumn and

lowest in winter (Fig. 2l). These could be related to the different influence of

evaporation and transpiration, dilution, mineral dissolution and leaching, and

agricultural irrigation among the seasons (Han et al., 2011; Liu et al., 2018b; Jia et al.,

2020). Transpiration of deep-rooted plants and evaporation of soil moisture and

shallow groundwater could affect groundwater quality by redistributing moisture and

salinity within the soil profile (Herczeg et al., 2001; Tweed et al., 2011; Liu et al.,

2018b). The dilution caused by precipitation, river water and irrigation water could

reduce groundwater ion concentration due to lower salinity of precipitation and

streamflow compared to shallow groundwater (Tweed et al., 2011; Wang et al., 2013;

Jia et al., 2020). Soil mineral dissolution and leaching could increase the salts in

groundwater, since the downward migration of soil water could also cause the

downward migration of salinity dissolved in soil moisture into groundwater (Su et al.,

2009; Tweed et al., 2011; Farid et al., 2015; Liu et al., 2018b). Liu et al. (2018b)

reported that saline soil in unsaturated zone and evaporites in fine-grained sediment

were widely distributed in the arid regions, and subsurface water could efficiently

36
dissolve evaporates during vertical migration processes. Agricultural irrigation

activities greatly impact groundwater hydrochemistry in arid regions, since

agricultural drainage systems could reduce groundwater salinity by lowering

groundwater level (discharging of subsurface water with high salinity) and decreasing

subsurface water evaporation, and the quality of irrigation water also could affect

groundwater salinity (Han et al., 2011; Liu et al., 2018a). However, the difference of

DM was not significant between seasons in the oasis (Fig. 2l and Fig. 5). This may be

attributed to the complex coupled effects of climate and human activities on

groundwater quality (Zhang et al., 2014b), and might imply that anthropogenic

activities play a great role in groundwater quality in the arid oasis area (Liu et al.,

2018b).

The spatial distributions and variations of groundwater level and mineralization

degree both exhibited significant spatial heterogeneity, and a spatial trend was not

obvious (Fig. 4 and 5). These could be related to the great spatial heterogeneity of

environmental conditions and anthropogenic activities among the monitoring wells,

such as topography, climatic conditions, hydrogeological condition, soil properties,

LULC types, vegetation, agricultural irrigation and drainage activities, and

groundwater exploitation (Fig.1 and Fig. 3, Table 1; Zhang et al., 2014b). The random

spatial trend of DGL and DM revealed the complex interaction between groundwater

and environmental factors, and suggested that the effects of climatic and topographic

conditions on groundwater level and quality were not significant in the Tarim Basin

(Zhang et al., 2014b; Chen et al., 2019). Furthermore, the regions with a higher

groundwater level exhibited a higher mineralization degree (Fig. 5), probably due to

greater salt accumulation caused by greater evaporative enrichment at the areas with

higher water level (Liu et al., 2018b). In addition, larger variation of DGL during the

study period was found at the areas with greater DGL (Fig. 4 and 5), probably due to

more stable recharge sources for the regions with smaller DGL. Smaller DGL were

generally found near river courses or cultivated land, which was mainly recharged by

37
infiltrating water from river water or irrigation water (from river water transfer or

pumping deep groundwater), leading to smaller variation of shallow groundwater

level (Han et al., 2011). Moreover, the variation of groundwater level in the eastern

region was significantly larger than that in the western region (Fig. 4a), probably due

to weaker anthropogenic interference in the eastern part (Wang et al., 2014).

Compared to the western part, the land use types in the east was dominated by

grassland and bare land (Fig. 3 a), and thus shallow groundwater recharge was more

unstable due to less infiltration water from almost dried-up river channels and little

agricultural irrigation, leading to greater interannual fluctuation of groundwater level

due to the effects of climate change (Zhang et al., 2014b). Furthermore, the lowest

values of DM were found in the northwestern part, while the highest values were

found in the southeastern part (Fig. 5k). These could be related to the different

recharge mechanisms, hydrogeochemical evolution, hydrogeological conditions

(hydraulic gradient, drainage conditions), evaporation effect, field irrigation and

drainage network, and groundwater level (Ma et al., 2013; Porhemmat et al., 2018).

Compared to the southeastern part, the higher terrain in the northwest could lead to

better drainage conditions due to greater hydraulic gradient and improved agricultural

drainage systems (Fig. 1), thus making less salt accumulate in shallow groundwater

due to lower water level, less evaporation and more salt excretion by drainage channel

(Zhang et al., 2014b).

5.2. Effects of climate and land-cover changes on groundwater

Climatic factors interact to affect groundwater level and quality (Table 4 and Fig.

6). Our results showed that DGL was significantly positively correlated to

precipitation and actual evapotranspiration, while negatively related to air temperature

(Table 4). DM was significantly positively related to air temperature, while

significantly negatively related to precipitation (Table 4). Higher air temperature

could decrease groundwater level by increasing evapotranspiration (Berihun et al.,

2019; Fu et al., 2019), but could also stimulate glacier and snow melting and thus lead

38
to higher groundwater level due to greater supply from meltwater (Ma et al., 2005;

Segura et al., 2019; Lin et al., 2020). Furthermore, increased precipitation during the

wet season could promote vegetation growth by enhancing soil moisture content

(Riley et al., 2019), and thus lead to lower groundwater level owing to the absorption

of soil water and groundwater by deep-rooted vegetation (Wang et al., 2014; Liu et

al., 2018b). This effect is limited in the arid oasis area due to scarce precipitation,

intensive evaporation, and high groundwater depths (Table 1; Ma et al., 2013; Wang

et al., 2014; Zhang et al., 2014a; Liu et al., 2018b). Groundwater variations in arid

oasis region mainly depend on surface water infiltration (river water and irrigation

water) and lateral groundwater flow rather than rainfall (Wang et al., 2013; Guo et al.,

2019), and high evaporation rate makes it hard for precipitation to become

groundwater in arid area (Hu et al., 2019). In addition, actual evapotranspiration was

one of the major controlling factors for groundwater level (Table 5). Elevated actual

evapotranspiration could lead to more water dissipation from river, soil water and

shallow groundwater due to increased temperature (Berihun et al., 2019; Folton et al.,

2019), and thus decrease groundwater level due to less recharge (Ma et al.,

2013;Wang et al., 2014; Liu et al., 2018b; Fu et al., 2019). Moreover, air temperature

was the dominant factor of groundwater mineralization degree (Table 5). Elevated

temperature could increase salt evaporative enrichment during evaporation process,

which in turn result in more salt accumulation in soil profile (Chen et al., 2018; Rose

et al., 2005), leading to higher mineralization degree for the shallow groundwater due

to the downward migration of salt dissolved in soil moisture into aquifer (Tweed et

al., 2011; Wang et al., 2013; Liu et al., 2018b). Furthermore, temperature is one of the

controlling factors for the rate of mineral dissolution and leaching which exhibited a

positive correlation (Welch and Ullman, 2000), and thus higher temperature can lead

to large salt leaching from soil into soil water and moving downward into aquifers,

resulting in higher groundwater salinity (Kindler et al., 2011; Riedel, 2019).

Additionally, elevated rainwater could dilute the salt concentration in soil water and

39
shallow groundwater during the wet season due to greater recharge from precipitation

infiltration, which has lower salinity compared to subsurface water (Huang et al.,

2017; Jia et al., 2020). Moreover, actual evapotranspiration was not significantly

related to groundwater mineralization degree (Table 4), indicating that other factors

might exert strong influences on shallow groundwater salts (e.g., agricultural

activities) (Liu et al., 2018b).

Streamflow had great effects on groundwater (Table 4). Streamflow was

significantly negatively related to DGL, while positively correlated with DM (Table

4). Greater streamflow could increase groundwater level by enhancing groundwater

recharge from river water (Wang et al., 2014; Gu et al., 2016a; Guo et al., 2019). For

example, Zhang et al. (Zhang et al., 2014b) reported that the upward trend of

groundwater level in 2002 and 2003 for the Wei-Ku Oasis was mainly caused by the

catastrophic flood on July 22 to 23, 2002. Moreover, the relationship between DGL

and streamflow was significant in autumn and winter, while not significant in spring

and summer (Table 4). One possible explanation was that the reservoirs in the Tarim

Basin can regulate the annual distribution of surface water (Chen et al., 2019), and

there is a lag between groundwater level fluctuation and river water infiltration during

the flood period (Wang et al., 2014; Zhang et al., 2014b; Erler et al., 2019). Greater

streamflow during flood period (with 61% occurring from May to September; Fig. 2i)

could increase groundwater level due to large river water seepage into subsurface

water (Wang et al., 2014; Gu, et al., 2016b). In addition, streamflow showed not

significant relation with groundwater mineralization degree (the correlation was not

shown), possibly due to limited surface water and intensified anthropogenic activities

in the study area (Wang et al., 2013; Liu et al., 2018b). The direct impact of river

water on groundwater quality was limited due to dried-up channels in the oasis.

Surface water transfer for irrigation could enhance salt evaporative enrichment by

increasing evaporation in the arid region, and increased irrigation water from river

water could promote mineral dissolution and leaching due to high soil salts (Liu et al.,

40
2018b). This in turn results in higher groundwater salinity due to more salt moving

downward with soil water into aquifer during the irrigation season (Wang et al., 2013;

Han et al., 2011). Nevertheless, variations in groundwater level and quality may

respond slowly to climate change and anthropogenic forcing (Erler et al., 2019; Riley

et al., 2019), so that the effect of forcing differences in spring was still seen in

summer and even in autumn (Hu et al., 2019; Wang et al., 2014). Zhang et al. (2014b)

observed a lag between groundwater level and runoff over the Wei-Ku Oasis, and

pointed out that the high groundwater level in 2003 was affected by the catastrophic

flood in 2002.

LULC types and vegetation coverage may exert a strong influence on

groundwater level and mineralization degree (Table 4 and 5, Fig. 6). Land-cover

change controls on the overall water balance (water demand and supply) and

redistributes water resources (modifying groundwater recharge locations, rates, and

mechanisms) (Han et al., 2017; Berihun et al., 2019; Folton et al., 2019), thus

affecting groundwater dynamics (Zhang et al., 2014b; Chen et al., 2019; Riley et al.,

2019; Scanlon et al., 2006). The cropland area was significantly positively related to

DGL, while negatively related to DM (Table 4 and 5). NDVI was positively

correlated with DGL, while significantly negatively related to DM in spring and

winter (Table 4). The cropland area was one of the main controlling factors for

groundwater level (Table 5). Elevated cropland area could enhance agricultural water

consumption by increasing evaporation and transpiration (Zhang et al., 2014a; Shukla

et al., 2018), which in turn leads to greater groundwater exploitation and surface water

transfer to meet irrigation water demands, resulting in lower groundwater level in arid

regions (Feng et al., 2005; Liu et al., 2018a). Zhang et al. (2014b) reported that the

effects of anthropogenic activities on groundwater change were roughly proportional

to cultivated area in the Tarim Basin. Berihun et al. (2019) reported that cultivated

land expansion could continuously deteriorate soil structural qualities by tillage, thus

affecting regional hydrological process. In addition, elevated cropland area can reduce

41
groundwater level due to more complete agricultural irrigation and drainage systems

(Castellano et al., 2019), which in turn leads to less salt accumulation in the shallow

groundwater owing to weaker evaporative enrichment and larger salt discharge by the

drainage channel, resulting in lower groundwater salinity (Table 6; Zhang et al.,

2014b). Riedel (2019) reported that agriculturally used land could receive large

amounts of nutrients (including N, P, and K), which eventually accumulate in soil and

aquifer, thus influencing groundwater salinity. Moreover, vegetation type also

contributed to the variation of groundwater level (Table 4 and Fig. 6a). Vegetation

type could influence evaporation and transpiration rates, root uptake of water, and

intercepting rainfall, leading to different fluctuation of shallow groundwater (Levia

and Frost, 2003; Jobbágy and Jackson, 2004; Berihun et al., 2019; Riley et al., 2019).

Increased NDVI due to farmland expansion (cropland and artificial forest) could

change hydrological process by enhancing crop water demand and evapotranspiration

in arid region (Ma et al., 2005; Huang et al., 2013), leading to lower groundwater

level due to greater groundwater exploitation and higher water loss via groundwater

transpiration in the growing season (Wang et al., 2014; Chen et al., 2019). Increased

NDVI could also dry and disconnect stream to groundwater due to most of surface

water transfer for irrigation, which in turn leads to lower water level owing to little

supply from river water (Zhang et al., 2014b). On the other hand, higher vegetation

coverage could make the upper soil layer have greater water holding capacity, thus

slower the infiltration of rainfall and irrigation water through the shallow soil layer to

the deep layer, resulting in phreatic decline (Yinglan et al., 2019). Furthermore,

higher NDVI in winter and spring (e.g. winter wheat) could increase the discharge of

water and salt by drainage channels due to flood irrigation, leading to less salt

migration and accumulation into shallow groundwater (Chen et al., 2019).

Agricultural irrigation activities significantly affected groundwater level and

quality (these correlations were not displayed). Agricultural irrigation canal can

redistribute surface water resources in time and space, thus influencing groundwater

42
level and its distribution (Liu et al., 2018b; Zhang et al., 2018). With the expansion of

oasis, a large number of irrigation canals were constructed to directly transport the

upstream river water to the irrigated fields (more than 90% of the inflow) (Zhang et

al., 2014b; Sun et al., 2018), which in turn reduced river water leakage due to the

dried-up river channel, resulting in lower groundwater level (Chen et al., 2019). The

gradually increased channel water utilization coefficient could also reduce channel

water leakage, leading to decreased groundwater level owing to little recharge from

surface water (Zhang et al., 2014b; Chen et al., 2019; Liu et al., 2018b). Liu et al.

(2018a) and Porhemmat et al. (2018) reported that water saving irrigation methods

could reduce deep percolation and thus decrease potential groundwater recharge in

arid and semiarid regions. Liu et al. (2018b) noted that irrigation water infiltration and

channel leakage could promote soil salinization by elevating shallow groundwater

level, thus enhancing the direct evaporation effect on groundwater salinization.

Furthermore, agricultural drainage systems also affected groundwater dynamics

((Zhang et al., 2014b). Gradually improved drainage system could discharge a certain

amount of water and solute salt from irrigated fields (Table 6; Castellano et al., 2019),

and thus can reduce the groundwater level and salinity due to weaker subsurface

water evaporation and less salt accumulation into unconfined aquifer (Zhang et al.,

2014b). Zhang et al. (2014b) reported that the peaks of water and salt discharge by

drainage canals mostly occurred after a large amount of irrigation, indicating that

drainage and irrigation are inseparable. In addition, groundwater overexploitation

carried out in arid region could modify the regional natural hydrological regime

(groundwater recharge and discharge processes, flow path) (Liu et al., 2018b; Liu et

al., 2018a; Mancuso et al., 2020), thus affecting groundwater storage and phreatic

fluctuation (Scanlon et al., 2012; Asoka et al., 2017; Hu et al., 2019). Farmland

expansion and population growth could lead to a dramatic increase in water

consumption (Shukla et al., 2018), which in turn increased the intensity of

groundwater extraction to meet regional water demand due to limited surface water

43
and increased groundwater usage in arid regions (Table 6), resulting in groundwater

depletion and phreatic decline (Chen et al., 2019; Wang et al., 2019b; Riley et al.,

2019; Liu et al., 2018b). Zhang et al. (2014b) found that the decline of groundwater

level was also related to the sharp increase of electromechanical wells in irrigation

areas, and the increased electromechanical wells could accelerate the exploitation of

groundwater resources. Riley et al. (2019) reported that increasing groundwater

extraction would reduce groundwater storage in the High Plains.

Table 6

The basic information related to agricultural activities, population and streamflow in the Wei-Ku

Oasis between 2000 and 2015.

Parameters 2000 2005 2010 2015

Electromechanical well numbers 283 761 2265 —

Cropland areas (km2) 3969.3 4065.0 4302.8 4393.0

Total population (104 persons) 73.14 78.56 89.22 96.28

Total water discharge (104 m3) 31141.7 37619.8 22873.3 —

Total salt discharge (104 tons) 170.6 225.3 150.9 —

Annual streamflow (108 m3) 26.76 32.32 38.58 30.18

Depth to groundwater level (m) 2.67 3.14 4.01 5.67

Total water discharge: total water discharge by agricultural drainage canals; Total salt discharge: total salt discharge by

agricultural drainage canals.

Other various environmental factors, such as population, policies, physiographic

characteristics (topography, lithology, geology) and extreme weather conditions,

could also impact groundwater level and quality directly or indirectly. For example,

Shukla et al. (2018) reported that population growth would influence groundwater

table and quality by enhancing groundwater exploitation and increasing domestic

waste discharge (contaminating groundwater by direct discharge or leaching) in arid

regions (Table 6). Chen et al. (2019) noted that policies significantly influenced the

groundwater dynamics in arid oasis region, and the decelerating of phreatic decline in

the Tarim Basin in recent years was most likely due to the implementation of the

“strictest water resource management system” in Xinjiang. Fu et al. (2019) found that

soil types and land attributes were the most significant factors influencing spatial

variation of groundwater recharge. Erler et al. (2019) reported that topographic relief

44
and hydrostratigraphy significantly affected groundwater table. It has been widely

known that groundwater level decreased significantly with altitude decrease during

average conditions (Fig. 1d; Bekele et al., 2019). However, our study did not address

these aspects, further studies on these topics are required.

5.3. Implications for regional groundwater cycling process

The above analyses revealed that spatial and temporal distributions of

groundwater level and salinity may provide insight into regional groundwater cycling

and hydrogeochemical evolution. Phreatic decline was caused by an imbalance

between recharge and discharge effects (Healy and Cook, 2002). Groundwater

hydrological cycle in arid regions has been altered significantly by climate change and

human activities, shifting groundwater recharge patterns (sources, locations, rates,

mechanisms), discharge pathways and salts dissolved in water (Han et al., 2017; Liu

et al., 2018). Moreover, anthropogenic activities seem to have stronger impacts on

groundwater after 2000 compared to climatic factors (Hu et al., 2019). Before the

oasis expansion, the river eventually flowed into the Tarim River, and shallow

groundwater system was characterized by simple recharge sources (from river and

lateral groundwater flow) and discharge (mainly via evapotranspiration) (Wang et al.,

2014). The dominant hydrological processes were riverbank leakage and groundwater

evaporation for the riparian zone, lateral groundwater flow and groundwater

evaporation for the desert, and irrigation water infiltration, lateral groundwater flow

and groundwater evapotranspiration for the cultivated land along the river.

Precipitation has little or no direct recharge for shallow aquifer (Seiler and Gat, 2007;

Hu et al., 2019), because the scarce rainfall has to satisfy soil and plant transpiration

first before infiltrating into aquifers (Alley, 2009; Liu et al., 2018b). After the oasis

expansion, groundwater recharge and discharge processes have been disturbed by

anthropogenic activities (Wang et al., 2014). For the riparian zone, the discharge

includes groundwater evaporation, and shallow aquifer is recharged by little riverbank

leakage due to the dried-up river courses, because more than 90% of the river water

45
was diverted for irrigation (plain reservoir regulation) (Zhang et al., 2014b). For the

desert (bare areas decreasing to 13.0% in 2018), the dominant hydrological processes

are lateral groundwater flow, groundwater evaporation and transpiration by

deep-rooted plants. For the cultivated land (cropland areas increasing to 62.5% in

2018), shallow aquifer is mainly recharged by massive irrigation water infiltration

(traditional flood irrigation), leakage from canal and reservoir, and lateral

groundwater flow, while the discharge includes groundwater evapotranspiration,

groundwater extraction and water discharge by agricultural drainage. Excessive

pumping from aquifers has caused local and rapid fall in groundwater level in the

Weigan-kuqa oasis (Chen et al., 2019). In addition, groundwater cycle and salt

migration are affected by many environmental factors, such as climate change, LULC

change, agricultural irrigation activities (e.g., irrigation and drainage), groundwater

exploitation, and population growth. Fig. 7 summarizes the variations of groundwater

recharge and discharge due to climate change and anthropogenic activities before and

after the oasis expansion in the Tarim Basin.

46
Fig. 7 Conceptual diagram depicting groundwater variation due to climate change and

anthropogenic activities under (a) pre the oasis expansion and (b) post the oasis expansion in the

typical oasis of Tarim Basin.

As pointed out previously, population growth and cultivated land expansion has

caused a sharp increase in water consumption in the arid region. On the one hand,

excessive groundwater abstraction for irrigation, domestic and industrial use may

47
reduce groundwater storage. On the other hand, massive irrigation water infiltration

could increase the recharge of shallow aquifers, which could counteract part of the

negative effects of groundwater overexploitation (Han et al., 2017). However,

groundwater level continued to decrease due to extensive development of agricultural

irrigation by pumping groundwater, increased irrigation water infiltration would be

primarily adding to vadose zone storage rather than aquifer recharge, thus preventing

any meaningful groundwater level recovery (Cao et al., 2016). This triggered a series

of hydrological and ecological problems including dried-up river courses, soil

salinization and natural desert vegetation degradation. Thus, excessive farmland

expansion was an unsuitable use of surface water and groundwater resources, which

has retarded the restoration of groundwater level and ecosystem in the Tarim Basin.

The decreasing trend of groundwater level was still foreseen if land use, crop pattern

and irrigation water were maintained as present, which could strongly challenge water

resource security and ecosystem health as well as sustainable economic development

in the arid oasis (Liu et al., 2018a; Wang et al., 2019b). Therefore, it is still necessary

to reasonably control or even reduce the irrigated area, supervise groundwater

pumping and/or close part of wells, and implement the water-saving practices in the

arid oasis of Tarim Basin. We should scientifically integrate surface water and

groundwater management in the oasis scale to balance the recharge and discharge of

aquifers, restrict non-ecological water use and optimize the water allocation ratio

between agriculture and ecology to promote groundwater level recovery (Guo et al.,

2019).

6. Conclusions

Spatio-temporal variations and determinants of groundwater level and

mineralization degree were examined using groundwater monitoring data collected at

22 wells in the Wei-Ku Oasis of Tarim Basin during the period of 2000-2018. The

results showed that the regions with a higher groundwater level generally exhibited a

higher mineralization degree. Obviously large spatial and temporal variations existed

48
in the DGL and DM, and the spatial trend was not obvious. DGL greatly increased

with time (2.05 m per decade) (p < 0.001) during the period due to climate change,

LULC change, agricultural irrigation activities, groundwater overexploitation and

population growth, while DM at most sites was characterized by a decreasing trend

(-0.21 g/L per decade) (p < 0.001) due to evaporative enrichment, mineral dissolution

and leaching, dilution and agricultural drainage. Groundwater level exhibited

significant seasonal differences, and was the shallowest in spring and deepest in

autumn. Moreover, there appears to be a pattern by which larger variation of

DGL/DM during the study period generally occurred at the areas with greater

DGL/DM. The lowest values of DM were in the northwestern part, and the highest

values were in the southeastern part. The variation of DGL in the eastern region was

noticeably larger than that in the western region. Furthermore, climate conditions,

streamflow, LULC types, vegetation coverage, agricultural irrigation activities, and

groundwater exploitation combined to affect the dynamic variations of DGL and DM,

and the effects exhibited obvious seasonal heterogeneity. Human activity was the

major influencing factor for DGL, and climate change for DM. The spatial and

temporal distributions of groundwater level and salinity may provide insight into

groundwater cycling and hydrogeochemical evolution. A simple conceptual diagram

was drawn to systematically depict the variations of groundwater recharge and

discharge associated with climatic and anthropogenic impacts in the arid oasis of

Tarim Basin. With the continuous expansion of oasis, anthropogenic activities have

contributed more and more to groundwater dynamics, and further study is needed to

quantitatively distinguish the hydrological impacts of climate change and

anthropogenic activities on groundwater dynamics in these arid regions, which is

critical to sustainable groundwater resources management.

CRediT authorship contribution statement

Wanrui Wang: Investigation, Methodology, Writing-original draft. Yaning

Chen: Supervision, Formal analysis, Conceptualization, Writing-review & editing.

49
Weihua Wang: Investigation, Formal analysis. Junxin Jiang: Writing-review &

editing. Ming Cai: Writing-review & editing. Yongjun Xu: Writing-review &

editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or

personal relationships that could have appeared to influence the work reported in this

paper.

Acknowledgement

We would like to thank the Water-salt Monitoring Station of the Weigan River

Basin Authority. We also appreciate the editor and the five anonymous reviewers for

their time and constructive comments on the manuscript. This study was supported by

the Strategic Priority Research Program of Chinese Academy of Sciences

(XDA20100303) and the Key Research Program of the Chinese Academy of Sciences

(ZDRWZS-2019-3).

References
Abdalla, O. A., 2008. Groundwater discharge mechanism in semi‐arid regions and
the role of evapotranspiration. Hydrol. Process. 22(16), 2993-3009.
Agam, N., Evett, S.R., Tolk, J.A., Kustas, W.P., Colaizzi, P.D., Alfieri, J.G., McKee,
L.G., Copeland, K.S., Howell, T.A. and Chávez, J.L., 2012. Evaporative loss
from irrigated interrows in a highly advective semi-arid agricultural area. Adv.
Water Resour. 50, 20-30.
Asoka, A., Gleeson, T., Wada, Y., Mishra, V., 2017. Relative contribution of
monsoon precipitation and pumping to changes in groundwater storage in India.
Nat. Geosci. 10 (2), 109-117.
Bai, P., Liu, X., Zhang, Y. Liu, C., 2020a. Assessing the impacts of vegetation
greenness change on evapotranspiration and water yield in China. Water Resour.
Res.
Bai, Y. F., Wang, Y., Chen, Y. N., Zhang, L. Y., 2020b. Probabilistic analysis of the
controls on groundwater depth using Copula Functions. Hydrol. Res. 51(3),
406-422.
Beal, L.K., Wong, C.I., Bautista, K.K., Jenson, J.W., Banner, J.L., Lander, M.A.,
Gingerich, S.B., Partin, J.W., Hardt, B. van Oort, N.H., 2019. Isotopic and
geochemical assessment of the sensitivity of groundwater resources of Guam,

50
 Mariana Islands, to intra-and inter-annual variations in hydroclimate. J. Hydrol.
568, 174-183.
Bekele, E., Zhang, Y., Donn, M., Don, M., 2019. Inferring groundwater dynamics in a
coastal aquifer near wastewater infiltration ponds and shallow wetlands
(Kwinana, Western Australia) using combined hydrochemical, isotopic and
statistical approaches. J. Hydrol. 568, 1055-1070.
Berihun, M.L., Tsunekawa, A., Haregeweyn, N., Meshesha, D.T., Adgo, E., Tsubo,
M., Masunaga, T., Fenta, A.A., Sultan, D., Yibeltal, M., Ebabu, K., 2019.
Hydrological responses to land use/land cover change and climate variability in
contrasting agro-ecological environments of the Upper Blue Nile basin, Ethiopia.
Sci. Total Environ. 689, 347-365.
Cao, G., Scanlon, B.R., Han, D. ， Zheng, C., 2016. Impacts of thickening
unsaturated zone on groundwater recharge in the North China Plain. J. Hydrol.
537, 260-270.
Cary, L., Petelet-Giraud, E., Bertrand, G., Kloppmann, W., Aquilina, L., Martins, V.,
Pierre, D., 2015. Origins and processes of groundwater salinization in the urban
coastal aquifers of Recife (Pernambuco, Brazil): A multi-isotope approach. Sci.
Total Environ. 530, 411-429.
Castellano, M.J., Archontoulis, S.V., Helmers, M.J., Poffenbarger, H.J., Six, J., 2019.
Sustainable intensification of agricultural drainage. Nat. Sustain. 2(10), 914-921.
Chao, N., Chen, G., Li, J., Xiang, L., Wang, Z., Tian, K., 2019. Groundwater storage
change in the Jinsha River basin from GRACE, hydrologic models, and in situ
data. Groundwater.
Chen, J. L., Wilson, C. R., Tapley, B. D., Scanlon, B., Güntner, A., 2016. Long-term
groundwater storage change in Victoria, Australia from satellite gravity and in
situ observations. Glob. Planet. Change. 139, 56-65.
Chen, J. Y., 2010. Holistic assessment of groundwater resources and regional
environmental problems in the North China Plain. Environ. Earth Sci. 61(5),
1037-1047.
Chen, T. D., Chen, X. G., Wang, W. K., Liu, B., Jiang, S. B., Zeng, Y. G., … Yang,
Z. Y. 2009. Investigation and assessment of groundwater resources and its
environmental issues in the Junggar Basin. Beijing: Geological Publishing
House. pp. 25-79 (in Chinese).
Chen, W., Jin, M., Ferré, T., Liu, Y., Xian, Y., Shan, T., Ping, X. 2018. Spatial
distribution of soil moisture, soil salinity, and root density beneath a cotton field
under mulched drip irrigation with brackish and fresh water. Field Crop. Res.
215, 207-221.
Chen, Y.N., Hao, X.M., Chen, Y.P., Zhu, C.G., 2019. Study on water system
connectivity and ecological protection countermeasures of Tarim River Basin in
Xinjiang. Bull. Chin. Acad. Sci. 34, 1156-1164 (in Chinese).
Cheng, G.D., Li, X., Zhao, W.Z., Xu, Z.M., Feng, Q., Xiao, S.C., Xiao, H.L., 2014.
Integrated study of the water-ecosystem-economy in the Heihe River basin. Natl.

51
 Sci. Rev. 1 (3), 413-428.
Eamus, D., Zolfaghar, S., Villalobos-Vega, R., Cleverly, J. Huete, A., 2015.
Groundwater-dependent ecosystems: recent insights from satellite and
field-based studies. Hydrol. Earth Syst. Sci.
Edmunds, W. M., 2009. Geochemistry’s vital contribution to solving water resource
problems. Appl. Geochem. 24(6), 1058-1073.
Erler, A. R., Frey, S. K., Khader, O., d'Orgeville, M., Park, Y. J., Hwang, H. T., ... &
Sudicky, E. A. 2019. Evaluating Climate Change Impacts on Soil Moisture and
Groundwater Resources Within a Lake-Affected Region. Water Resour. Res.
55(10), 8142-8163.
Farid, I., Zouari, K., Rigane, A., Beji, R., 2015. Origin of the groundwater salinity and
geochemical processes in detrital and carbonate aquifers: Case of Chougafiya
basin (Central Tunisia). J. Hydrol. 530, 508-532.
Feng, Q., Liu, W., Su, Y. H., Zhang, Y. W., Si, J. H., 2004. Distribution and evolution
of water chemistry in Heihe River basin. Environ. Geol. 45(7), 947-956.
Feng, W., Zhong, M., Lemoine, J. M., Biancale, R., Hsu, H. T., Xia, J., 2013.
Evaluation of groundwater depletion in North China using the Gravity Recovery
and Climate Experiment (GRACE) data and ground ‐ based measurements.
Water Resour. Res. 49(4), 2110-2118.
Folton, N., Martin, E., Arnaud, P., L'hermite, P., Tolsa, M., 2019. A 50-year analysis
of hydrological trends and processes in a Mediterranean catchment. Hydrol.
Earth Syst. Sci. 23(6), 2699-2714.
Fu, G. B., Crosbie, R. S., Barron, O., Charles, S. P., Dawes, W., Shi, X., ... Li, C.,
2019. Attributing variations of temporal and spatial groundwater recharge: A
statistical analysis of climatic and non-climatic factors. J. Hydrol. 568, 816-834.
Fuchs, E. H., King, J. P., Carroll, K. C., 2019. Quantifying disconnection of
groundwater from managed-ephemeral surface water during drought and
conjunctive agricultural use. Water Resour. Res. 55.
Graaf, I. E., Gleeson, T., van Beek, L. R., Sutanudjaja, E. H., Bierkens, M. F. 2019.
Environmental flow limits to global groundwater pumping. Nature, 574(7776),
90-94.
Gu, X. H., Zhang, Q., Singh, V. P., Chen, Y. D., Shi, P. J., 2016a. Temporal
clustering of floods and impacts of climate indices in the Tarim River basin,
China. Glob. Planet. Change 147, 12-24.
Gu, X. H., Zhang, Q., Singh, V. P., Liu, L., 2016b. Nonstationarity in the occurrence
rate of floods in the Tarim River basin, China, and related impacts of climate
indices. Glob. Planet. Change 142, 1-13.
Guo, X. Y., Feng, Q., Si, J. H., Xi, H. Y., Zhao, Y., Deo, R. C., 2019. Partitioning
groundwater recharge sources in multiple aquifers system within a desert oasis
environment: Implications for water resources management in endorheic basins.
J. Hydrol. 579, 124212
Han, D., Currell, M. J., Cao, G., Hall, B. 2017. Alterations to groundwater recharge

52
 due to anthropogenic landscape change. J. Hydrol. 554, 545-557.
Han, D., Song, X., Currell, M., Cao, G., Zhang, Y., Kang, Y., 2011. A survey of
groundwater levels and hydrogeochemistry in irrigated fields in the Karamay
Agricultural Development Area, northwest China: Implications for soil and
groundwater salinity resulting from surface water transfer for irrigation. J.
Hydrol. 405, 217-234.
Healy, R. W., Cook, P. G., 2002. Using groundwater levels to estimate recharge.
Hydrogeol. J. 10: 91-109.
Herczeg, A. L., Dogramaci, S., Leaney, F. W., 2001. Origin of dissolved salts in a
large, semi-arid groundwater system: Murray Basin, Australia. Mar. Freshw.
Res. 52, 41-52.
Hori, M., Sugiura, K., Kobayashi, K., Aoki, T., Tanikawa, T., Kuchiki, K., Niwano,
M., Enomoto, H., 2017. A 38-year (1978–2015) Northern Hemisphere daily
snow cover extent product derived using consistent objective criteria from
satellite-borne optical sensors. Remote Sens. Environ. 191, 402-418.
Hu, K. X., Awange, J. L., Kuhn, M., Saleem, A., 2019. Spatio-temporal groundwater
variations associated with climatic and anthropogenic impacts in South-West
Western Australia. Sci. Total Environ. 696, 133599.
Huang, T.M., Pang, Z.H., 2010. Changes in groundwater induced by water diversion
in the Lower Tarim River, Xinjiang Uygur, NW China: evidence from
environmental isotopes and water chemistry. J. Hydrol. 387, 188-201.
Hughes, J., Petrone, K., Silberstein, R., 2012. Drought, groundwater storage and
stream flow decline in southwestern Australia. Geophys. Res. Lett. 39 (3),
L03408.
Ji, X.B., Kang, E.S., Chen, R.S., Zhao, W.Z., Zhang, Z.H., Jin, B.W., 2006. The
impact of the development of water resources on environment in arid inland river
basins of Hexi region, northwestern China. Environ. Geol. 50 (6), 793-801.
Jia, H., Qian, H., Zheng, L., Feng, W. W., Wang, H. K., Gao, Y. Y., 2020. Alterations
to groundwater chemistry due to modern water transfer for irrigation over
decades. Sci. Total Environ. 717: 137170.
Jiang, Y., Xu, X., Huang, Q., Huo, Z., Huang, G., 2015. Assessment of irrigation
performance and water productivity in irrigated areas of the middle Heihe River
basin using a distributed agro-hydrological model. Agric. Water Manage. 147,
67-81.
Jobbágy Esteban G., Jackson Robert B. 2004. Groundwater use and salinization with
grassland afforestation. Glob. Change Biol. 10, 1299-1312.
Joodaki, G., Wahr, J., Swenson, S., 2014. Estimating the human contribution to
groundwater depletion in the Middle East, from GRACE data, land surface
models, and well observations. Water Resour. Res. 50, 2679-2692.
Kindler, R., Siemens, J.A.N., Kaiser, K., Walmsley, D.C., Bernhofer, C., Buchmann,
N., Cellier, P., Eugster, W., Gleixner, G., Grũnwald, T., Heim, A., 2011.
Dissolved carbon leaching from soil is a crucial component of the net ecosystem

53
 carbon balance. Glob. Change Biol. 17(2), 1167-1185.
Levia Jr, D.F., Frost, E.E., 2003. A review and evaluation of stemflow literature in the
hydrologic and biogeochemical cycles of forested and agricultural ecosystems. J.
hydrol. 274(1-4), 1-29.
Lin, L., Gao, M., Liu, J., Wang, J., Wang, S., Chen, X., Liu, H., 2019. Quantifying
streamflow and active groundwater storage in response to climate warming in an
alpine catchment, upper Lhasa River. Hydrol. Earth Syst. Sci. Discussions, 1-41.
Liu, M., Jiang, Y., Xu, X., Huang, Q., Huo, Z., Huang, G. 2018a. Long-term
groundwater dynamics affected by intense agricultural activities in oasis areas of
arid inland river basins, Northwest China. Agric. Water Manage. 203, 37-52.
Liu, S., Chen, M. Zhuang, Q., 2016. Direct radiative effects of tropospheric aerosols
on changes of global surface soil moisture. Clim. Change 136(2), 175-187.
Liu, Y. L., Jin M. G., Wang J. J., 2018b. Insights into groundwater salinization from
hydrogeochemical and isotopic evidence in an arid inland basin. Hydrol. Process.
32(20): 3108-3127.
Ma, B., Jin, M., Liang, X., Li, J., 2019a. Application of environmental tracers for
investigation of groundwater mean residence time and aquifer recharge in
fault-influenced hydraulic drop alluvium aquifers. Hydrol. Earth Syst. Sci. 23(1).
Ma, J. Z., He, J. H., Qi, S., Zhao, W., Edmunds, W. M., Zhao, Y. P. 2013.
Groundwater recharge and evolution in the Dunhuang Basin, northwestern China.
Appl. Geochem. 28: 19-31.
Ma, J. Z., Wang, X. S., Edmunds, W. M., 2005. The characteristics of ground-water
resources and their changes under the impacts of human activity in the arid
Northwest China-a case study of the Shiyang River Basin. J. Arid. Environ.
61(2), 277-295.
Ma, Z., Yan, N., Wu, B., Stein, A., Zhu, W., Zeng, H., 2019b. Variation in actual
evapotranspiration following changes in climate and vegetation cover during an
ecological restoration period (2000–2015) in the Loess Plateau, China. Sci. Total
Environ. 689, 534-545.
Manchanda, G., Garg, N., 2008. Salinity and its effects on the functional biology of
legumes. Acta Physiol. Plant. 30, 595-618.
Mancuso, M., Santucci, L., Carol, E., 2020. Effects of intensive aquifers exploitation
on groundwater salinity in coastal wetlands. Hydrol. Process. 34(11), 2313-2323.
Manna, F., Murray, S., Abbey, D., Martin, P., Cherry, J., Parker, B. 2019. Spatial and
temporal variability of groundwater recharge in a sandstone aquifer in a semiarid
region. Hydrol. Earth Syst. Sci. 23(4).
McCallum, A. M., Andersen, M. S., Giambastiani, B. M. S., Kelly, B. F. J., Acworth,
R. I., 2013. River-aquifer interactions in a semi-arid environment stressed by
groundwater abstraction. Hydrol. Process. 27(7), 1072-1085.
Orellana, F., Verma, P., Loheide, S. P., Daly, E. 2012. Monitoring and modeling
water-vegetation interactions in groundwater-dependent ecosystems. Rev.
Geophys. 50(3).

54
Pang, Z. H., Huang, T. M., Chen, Y. N., 2010. Diminished groundwater recharge and
circulation relative to degrading riparian vegetation in the middle Tarim River,
Xinjiang Uygur, Western China. Hydrol. Process. 24, 147-159.
Porhemmat, J., Nakhaei, M., Dadgar, M. A., Biswas, A. 2018. Investigating the
effects of irrigation methods on potential groundwater recharge: A case study of
semiarid regions in Iran. J. hydrol. 565, 455-466.
Riedel, T., 2019. Temperature-associated changes in groundwater quality. J. hydrol.
572, 206-212.
Riley, D., Mieno, T., Schoengold, K., Brozović, N. 2019. The impact of land cover on
groundwater recharge in the High Plains: An application to the Conservation
Reserve Program. Sci. Total Environ. 696, 133871.
Rose, D., Konukcu, F., Gowing, J. 2005. Effect of water table depth on evaporation
and salt accumulation from saline groundwater. Soil Res. 43, 565-573.
Scanlon, B. R., Faunt, C. C., Longuevergne, L., Reedy, R. C., Alley, W. M., McGuire,
V. L., McMahon, P. B., 2012. Groundwater depletion and sustainability of
irrigation in the US High Plains and Central Valley. Proc. Natl. Acad. Sci. U. S.
A. 109, 9320-9325.
Scanlon, B.R., Keese, K.E., Flint, A.L., Flint, L.E., Gaye, C.B., Edmunds, W.M.
Simmers, I., 2006. Global synthesis of groundwater recharge in semiarid and arid
regions. Hydrol. Process. 20(15), 3335-3370.
Segura, C., Noone, D., Warren, D., Jones, J. A., Tenny, J., Ganio, L. M., 2019.
Climate, landforms, and geology affect baseflow sources in a mountain
catchment. Water Resour. Res. 55(7), 5238-5254.
Seiler KP, Gat JR. 2007. Mechanisms and processes of recharge. In Groundwater
Recharge from Run-off, Infiltration and Percolation. Springer: Netherlands;
31-68.
Shukla, A.K., Ojha, C.S.P., Mijic, A., Buytaert, W., Pathak, S., Garg, R.D. Shukla, S.,
2018. Population growth, land use and land cover transformations, and water
quality nexus in the Upper Ganga River basin. Hydrol. Earth Syst. Sci. 22(9).
Soylu ME, Istanbulluoglu E, Lenters JD, Wang T. 2011. Quantifying the impact of
groundwater depth on evapotranspiration in a semi-arid grassland region.
Hydrol. Earth Syst. Sci. 15: 787-806.
Strassberg, G., Scanlon, B.R., Chambers, D., 2009. Evaluation of groundwater storage
monitoring with the GRACE satellite: Case study of the High Plains aquifer,
central United States. Water Resour. Res. 45(5).
Su, Y. H., Zhu, G. F., Feng, Q., Li, Z. Z., Zhang, F. P., 2009. Environmental isotopic
and hydrochemical study of groundwater in the Ejina Basin, northwest China.
Environ. Geol. 58(3), 601-614.
Sun, H. T., Chen, Y. P., Chen, Y. N., Zhang, Y., He, Z. L., 2020. Groundwater
evapotranspiration in desert riparian forest in the lower reaches of the Tarim
River. Arid Zone Res. 37(1), 116-125 (in Chinese).
Sun, P., Zhang, Q., Yao, R., Singh, V. P., Song, C., 2018. Low flow regimes of the

55
 Tarim River Basin, China: probabilistic behavior, causes and implications. Water
10, 470.
Sun, Z., Ma, R., Wang, Y., Hu, Y., Sun, L., 2016. Hydrogeological and
hydrogeochemical control of groundwater salinity in an arid inland basin:
Dunhuang Basin, northwestern China. Hydrol. Process. 30, 1884-1902.
Taylor, R. G., Scanlon, B., Doll, P., Rodell, M., van Beek, R., Wada, Y.,
Longuevergne, L., Leblanc, M., Famiglietti, J. S., Edmunds, M., Konikow, L.,
Green, T. R., Chen, J. Y., Taniguchi, M., Bierkens, M. F. P., MacDonald, A.,
Fan, Y., Maxwell, R. M., Yechieli, Y., Gurdak, J. J., Allen, D. M.,
Shamsudduha, M., Hiscock, K., Yeh, P. J. F., Holman, I., Treidel, H. 2013.
Ground water and climate change. Nat. Clim. Chang. 3, 322-329.
Tweed, S., Leblanc, M., Cartwright, I., Favreau, G., Leduc, C. 2011. Arid zone
groundwater recharge and salinisation processes; an example from the Lake Eyre
Basin, Australia. J. hydrol. 408(3-4), 257-275.
Tweed, S.O., Leblanc, M., Webb, J.A. and Lubczynski, M.W., 2007. Remote sensing
and GIS for mapping groundwater recharge and discharge areas in salinity prone
catchments, southeastern Australia. Hydrogeol. J. 15(1), 75-96.
Wang, F., Chen, Y.N., Li, Z., Fang, G.H., Li, Y.P., Xia, Z.H., 2019a. Assessment of
the irrigation water requirement and water supply risk in the Tarim River Basin,
Northwest China. Sustainability. 11, 4941.
Wang, P., Yu, J. J., Pozdniakov, S. P., Grinevsky, S. O., Liu, C. M., 2014. Shallow
groundwater dynamics and its driving forces in extremely arid areas: a case study
of the lower Heihe River in northwestern China. Hydrol. Process. 28(3),
1539-1553.
Wang, P., Yu, J. J., Zhang, Y. C., Liu, C. M., 2013. Groundwater recharge and
hydrogeochemical evolution in the Ejina Basin, northwest China. J. hydrol. 476,
72-86.
Wang, T. Y., Wang, P., Zhang, Y. C., Yu, J. J., Du, C. Y., Fang, Y. H., 2019b.
Contrasting groundwater depletion patterns induced by anthropogenic and
climate-driven factors on Alxa Plateau, northwestern China. J. hydrol. 576,
262-272.
Welch, S.A., Ullman, W.J., 2000. The temperature dependence of bytownite feldspar
dissolution in neutral aqueous solutions of inorganic and organic ligands at low
temperature 5-35 °C. Chem. Geol. 167, 337-354.
White, C.J., Tanton, T.W., Rycroft, D.W., 2014. The impact of climate change on the
water resources of the Amu Darya basin in central Asia. Water Resour. Manag.
28 (15), 5267-5281.
Wittenberg, H., Aksoy, H., Miegel, K., 2019. Fast response of groundwater to heavy
rainfall. J. hydrol. 571, 837-842.
Wu, L., Li, C. B., Xie X. H., He, Z. B., Wang, W. R., Zhang, Y., Wei, J. M., Lv, J. N.,
2020. The impact of increasing land productivity on groundwater dynamics: a
case study of an oasis located at the edge of the Gobi Desert. Carbon Balance

56
 and Manag. 15, 1-13.
Xi, H., Feng, Q., Zhang, L., Si, J., Yu, T., 2018. Groundwater storage changes and
estimation of stream lateral seepage to groundwater in desert riparian forest
region. Hydrol. Res. 49(3), 861-877.
Xie, X., Wang, Y., Li, J., Su, C., Duan, M., 2013. Hydrogeochemical and Isotopic
Investigations on Groundwater Salinization in the Datong Basin, Northern
China. J. Am. Water Resour. Assoc. 49(2), 402-414.
Yang, J., Li, Z., 2011. The dynamic trends and prevention countermeasures
ofgroundwater in Danghe irrigation, Dunhuang basin. Water Resource and
Environmental Protection (ISWREP), 2011 International Symposium on. IEEE
4, 3227-3230.
Yang, Y., Xiao, P., Feng, X., Li, H. 2017. Accuracy assessment of seven global land
cover datasets over china. ISPRS-J. Photogramm. Remote Sens. 125:156-173.
Yao, Y., Zheng, C., Liu, J., Cao, G., Xiao, H., Li, H., Li, W., 2015. Conceptual and
numerical models for groundwater flow in an arid inland river basin. Hydrol.
Process. 29(6), 1480-1492.
Yinglan, A., Wang, G., Liu, T., Shrestha, S., Xue, B. and Tan, Z., 2019. Vertical
variations of soil water and its controlling factors based on the structural
equation model in a semi-arid grassland. Sci. Total Environ. 691, pp.1016-1026.
Zhang, F., Guo, Y. C., Jiang, H. T., Zhang, Y. J., Zhou, M., 2014b. Impacts of climate
and land use/cover change on water resources in the delta oasis of Weigan river
-Kuqa river. Hubei Agric. Sci. 53(23), 5716-5723 (in Chinese).
Zhang, M., Wang, S., Fu, B., Gao, G., Shen, Q., 2018. Ecological effects and potential
risks of the water diversion project in the Heihe River Basin. Sci. Total Environ.
619, 794-803.
Zhang, Q., Gu, X. H., Singh, V. P., Sun, P., Chen, X. H., Kong, D. D., 2016.
Magnitude, frequency and timing of floods in the Tarim River basin, China:
changes, causes and implications. Glob. Planet. Change 139, 44-45.
Zhang, Q., Sun, P., Li, J. F., Xiao, M. Z., Singh, V. P., 2015. Assessment of drought
vulnerability of the Tarim River basin, Xinjiang, China. Theor. Appl. Climatol.
121, 337-347.
Zhang, X. F., Zhang, L. H., He, C. S., Li, J. L., Jiang, Y. W., Ma, L. B., 2014a.
Quantifying the impacts of land use/land cover change on groundwater depletion
in Northwestern China-a case study of the Dunhuang oasis. Agric. Water
Manage. 146, 270-279.
Zhao, R., Chen, Y., Shi, P., Zhang, L., Pan, J., Zhao, H., 2013. Land use and land
cover change and driving mechanism in the arid inland river basin: a case study
of Tarim River, Xinjiang, China. Environ. Earth Sci. 68(2), 591-604.
Zhao, Z. Y., 2006. Landscape pattern changes and its ecological response in delta
oasis of Weigan River. University of Chinese Academy of Sciences, China.
Zhou, Y., Li, X., Yang, K., Zhou, J., 2018. Assessing the impacts of an ecological
water diversion project on water consumption through high-resolution

57
 estimations of actual evapotranspiration in the downstream regions of the Heihe
River Basin, China. Agric. For. Meteorol. 249, 210-227.
Zou, S., Jilili, A., Duan, W. L., Maeyer, P. D., de Voorde, T. V., 2019. Human and
natural impacts on the water resources in the Syr Darya River Basin, Central
Asia. Sustainability 11, 3084.

Abstract:

Groundwater is essential for agricultural, ecological and domestic water use, and

its depletion under climatic and anthropogenic impacts can impose challenges on

regional water resources and ecological security, particularly in arid regions.

Spatiotemporal variation and drivers of groundwater are critical to understand

groundwater cycling and restore groundwater, while they are poorly understood in

arid oasis region due to sparse field monitoring. To address this, we examined the

spatiotemporal variations and controlling factors of groundwater level and

mineralization degree using data collected at 22 wells in the Wei-Ku Oasis of Tarim

Basin during 2000-2018, a typical arid inland river basin in northwest China. Results

showed that obviously large spatial and temporal variations existed in both DGL

(depth to groundwater level) and DM (degree of mineralization). Groundwater level

was the shallowest in spring and deepest in autumn. DGL exhibited a markedly

increasing trend (2.05 m per decade), while DM marginally decreased (-0.21 g/L per

decade) (p < 0.001) during 2000-2018. The spatial distributions of DGL and DM were

complex. Larger variation of DGL/DM generally occurred at the areas with greater

DGL/DM. DM in the northwestern part was obviously smaller than that in the

southeastern part. The variation of DGL in the eastern region was noticeably larger

than that in the western region. Human activity was the main driving factor of DGL

increasing, climate change had a dominant role in DM decreasing, and the

relationship between controlling factors and groundwater varied in space and season.

Decreasing riverbank leakage and increasing volume of evapotranspiration,

groundwater pumping and water drainage led to decreasing groundwater level, while

increasing infiltration/leakage from irrigated water, canal and reservoir resulted in

58
increasing groundwater level. The results could help to advance the understanding of

groundwater cycling mechanism, and be useful in improving sustainable groundwater

management and restoring ecosystem in arid region.

CRediT authorship contribution statement

Wanrui Wang: Investigation, Methodology, Writing-original draft. Yaning

Chen: Supervision, Formal analysis, Conceptualization, Writing-review & editing.

Weihua Wang: Investigation, Formal analysis. Junxin Jiang: Writing-review &

editing. Ming Cai: Writing-review & editing. Yongjun Xu: Writing-review &

editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or

personal relationships that could have appeared to influence the work reported in this

paper.

Highlights:
 Variations and drivers of groundwater in the oasis of dried-up river in Tarim
Basin were examined.
 Groundwater level and mineralization degree decreased with time in research
area.
 Larger variation of DGL/DM generally occurred at the areas with greater
DGL/DM.
 Human activity was the main driving factor of DGL increasing, climate change
had a dominant role in DM decreasing.

59
60
61
62
63
64
65
66