Journal of Hydrology 620 (2023) 129416

Contents lists available at ScienceDirect

Journal of Hydrology
journal homepage: www.elsevier.com/locate/jhydrol

Research papers

Estimating the spatio-temporal assessment of GRACE/GRACE-FO derived

groundwater storage depletion and validation with in-situ water quality
data (Yazd province, central Iran)
Vahab Amiri a, *, Shoaib Ali b, c, Nassim Sohrabi d
a
Department of Geology, Yazd University, Yazd, Iran
b
Department of Earth and Space Sciences, Southern University of Science and Technology, Shenzhen 518005, China
c
Department of Irrigation and Drainage, Faculty of Agricultural Engineering and Technology, University of Agricultural Faisalabad, Faisalabad, Pakistan
d
Yazd Regional Water Authority, Yazd, Iran

A R T I C L E I N F O A B S T R A C T

Keywords: Due to the lack of surface water availability in Yazd province, groundwater has been overexploited as a key
GRACE/GRACE-FO source of water supply, resulting in the overexploitation of groundwater resources for agricultural activities. This
Groundwater storage study employed the data of the Gravity Recovery and Climate Experiment (GRACE)/GRACE Follow-On (GFO)
Groundwater depletion
satellite to determine terrestrial water storage (TWS) and deriving groundwater storage (GWS) variations using
Salinity
Yazd province
global land data assimilation system (GLDAS) water storage components in Yazd province from 2003 to 2020.
The time series results showed that the annual changes of GWS are positive from 2003 to 2007, and continue
negative with a decrease in precipitation significantly causing depletion of GWS in the aquifers. The Pearson
correlation coefficient is close to 0.58 between observed and GRACE-derived GWS on yearly basis and both show
a decreasing GWS trend. The results of calculating the groundwater balance for aquifers showed that the
renewable water resources have decreased from 786 to 250 million cubic meters (MCM) between the water year
2009–2010 and 2019–2020. This has caused the electrical conductivity (EC) to increase from about 8500 to
12,500 μS/cm. The correlation coefficient between EC and observed and GRACE-derived GWS was around − 0.50
and − 0.79, respectively. GRACE-derived and analytically calculated groundwater storage deficit (GSD) is around
2491 and 2775 MCM showing a similar downward trend. These results, while confirming GRACE-derived GWS,
point out the need to revise the analytical calculation of the groundwater balance based on the unmodified
aquifer domains and the non-uniform distribution of observation wells in most aquifers to reduce the difference
between observed and GRACE-derived GWS anomalies.

1. Introduction areas under water stress (de Graaf et al. 2019). This situation is even
more severe in the Middle East with its dry climate. Nearly 85% of water
The lives of about half of the world’s people depend on groundwater resources in the Middle East region are consumed in the agricultural
reserves (Han et al. 2021). Groundwater, as the largest source of sector (Joodaki et al. 2014).
freshwater, supplies about 36 and 40–42% of the water needed in the Iran is one of the countries in the Middle East region with limited
drinking and agricultural sectors around the world, respectively (Ashraf water resources, which has faced many problems in this regard in recent
et al. 2021; Ali et al. 2022; Zhang et al. 2022b). Nevertheless, population years. Ashraf et al. (2021) showed that in almost 77% of Iran’s land area,
growth and the need to increase the rate of water extraction from these groundwater pumping and depletion occur at a rate higher than the
vital sources for irrigation, shortages of surface water sources, socio- recharging capacity, safe yield, and sustainability of aquifers. In addi­
economic development, and climate changes have caused the rate of tion to anthropogenic factors, Vaghefi et al. (2019) by predicting the
groundwater extraction to be higher than the rate of recharge (Smith climatic conditions from 2022 to 2045 showed that Iran, especially its
et al. 2016; Gleeson et al. 2020; Ali et al. 2022; Amiri et al. 2022a). This arid regions, will face an increase in temperature and a decrease in
has caused that today about one-third of the world’s population lives in precipitation, and as a result, a decrease in the groundwater recharge

* Corresponding author.
E-mail addresses: v.amiri@yazd.ac.ir (V. Amiri), shoaib@sustech.edu.cn (S. Ali), n.sohrabi1986@gmail.com (N. Sohrabi).

https://doi.org/10.1016/j.jhydrol.2023.129416
Received 7 January 2023; Received in revised form 22 February 2023; Accepted 17 March 2023
Available online 22 March 2023
0022-1694/© 2023 Elsevier B.V. All rights reserved.
V. Amiri et al. Journal of Hydrology 620 (2023) 129416

rate. Noori et al. (2021) showed that the annual extraction of nearly 5.4 groundwater quality. The relationship between measured and GRACE-
km3 of groundwater in Iran has caused a decrease of 10 to 100 cm (with derived GWS and electrical conductivity (EC) of groundwater re­
an average of 49 cm/year) in the groundwater level. This can lead to an sources is also evaluated to confirm the results.
increase in land subsidence and a decrease in groundwater quality
(Motagh et al. 2008; Herrera-Garcia et al. 2021; Fatolazadeh and Goïta 2. Materials and methods
2022).
Despite the importance of groundwater in guaranteeing human life 2.1. Study area
in a considerable part of the world, sufficient attention has not been paid
to groundwater management (Famiglietti and Rodell, 2013). By moni­ Yazd province with an area of about 74,645 km2 is located in the
toring groundwater storage and providing a sustainable management central plateau of Iran and is the eighth largest province of Iran (Fig. 1).
plan, which has become one of the major global concerns today (Long The average annual precipitation is about 95 mm, which is about 38% of
et al. 2020; Amiri et al. 2021c; Saemian et al. 2022), it is possible to the long-term average precipitation in Iran (250 mm). The average
achieve a relative balance between the water supply and demand (Wu annual volume of precipitation in Iran is about 7 billion cubic meters
et al. 2019; Wang et al. 2022; Ahamed et al. 2022). (BCM). The minimum and maximum temperatures of this province are
Today, changes in the groundwater level and storage of aquifers are around –23 and 48 ◦ C, respectively. The average annual evaporation of
recorded using a monitoring network consisting of piezometers and nearly 3300 mm causes a significant share of water resources to be
observation wells. Despite the significant advantage of direct data unavailable due to evaporation (YRWA 2022). Fig. 2.
recording, this approach has limitations such as gaps in the dataset, Water resources are exploited by about 7300 water supply sources,
inappropriate distribution, and an insufficient number of piezometers including wells (59%), springs (5%), and qanats (36%). Nearly 85% of
(Srivastava and Dikshit 2022). On the other hand, creating an optimal pumped water is used for irrigation, 11% for drinking, and 4% for in­
monitoring network, especially in large aquifers, can be very expensive. dustry (YRWA 2022). The average annual water level drop in the
Therefore, it is necessary to use low-cost methods such as remote sensing aquifers of this province is between 1.03 and 0.02 m. The long-term
techniques that can provide reliable information on the changes in average deficit of the groundwater balance is about 243 million cubic
groundwater storage (Taylor and Alley 2001; Mishra et al., 2015; meters (MCM), and this has increased in recent years, reaching about
Chawla et al., 2020). 384 MCM in the water year (or hydrological year) 2020–2021. The
Since the mission of Gravity Recovery and Climate Experiment investigation of the calculated balance of groundwater in this province
(GRACE, April 2002-June 2017) and GRACE Follow-On (GFO, June shows that the reduction of the cumulative volume of the groundwater
2018-present) satellites have been started (Arshad et al., 2022), valuable reservoirs of the province during the years 1976–2021 was about 9 BCM
information on the changes in total water storage (TWS) with global (YRWA 2022).
coverage every 30 days has been provided to researchers (Tapley et al.,
2019; Landerer et al., 2020; Zhang et al., 2022a). These satellites can 2.2. TWS data from GRACE/GFO satellite
determine changes in the gravity field and directly provide mass redis­
tribution and changes in the volume of water resources, including soil The GRACE/GFO satellite, a collaboration between the National
moisture, water from snow and ice, and groundwater on large scales Aeronautics and Space Administration (NASA) and the German Aero­
(Castellazzi et al., 2018; Chen et al., 2019; Mehrnegar et al., 2021). The space Center (DLR), was launched in April 2002. The CSR-mascon TWS
information on these satellites has a wide application in water resources data was used which is provided by the GRACE-GFO satellite and has a
(Jacob et al., 2012; Joodaki et al., 2014; Tangdamrongsub et al., 2020; spatial and temporal resolution of 0.25×0.25◦ and monthly, respec­
Liu et al. 2021; Akhtar et al., 2022; Li et al., 2022), climate (Houborg tively. The missing data were filled in using the linear interpolation
et al., 2012; Freedman et al., 2014; Yan et al., 2021; Khorrami and method (Ali et al. 2021). The gap between GRACE and GRACE-FO of 11
Gunduz 2021), natural disasters (de Linage et al. 2009; Sjöberg and months was un-filled in this study. The data can be obtained using the
Bagherbandi 2020; Xiong et al., 2021; Yan et al., 2022), and surveying GRACE Data Analysis Tool from https://grace.jpl.nasa.gov/data/get
and navigation (Chambers, 2006; Swenson et al., 2008; Nagale et al., -data/.
2022) studies.
One of the most important and practical capabilities of the data 2.3. GLDAS water storage components
collected by GRACE/GFO satellites is the calculation of groundwater
storage (GWS). So far, various researchers around the world (such as GLDAS hydrological model uses NASA’s ground and space observa­
Rodell et al. 2007; Rodell et al., 2009; Chen et al., 2019; Tangdam­ tion system to determine near real-time data of land surface changes.
rongsub et al. 2020; Ali et al. 2021; Ali et al., 2022a; Ali et al., 2022b; This model can provide important information such as evapotranspira­
Srivastava and Dikshit, 2022) have used this data to study groundwater tion, precipitation, soil moisture storage, air, and land surface temper­
balance. Despite the great importance of identifying groundwater re­ ature, and snow water equivalent through simulation (Wang et al.,
sources and preparing a sustainable management plan in Iran with 2022). The GLDAS model, as one of the most important data sources for
critical conditions in water resources, few studies have focused on the evaluating the water cycle and studying water storage changes, has four
estimation of groundwater resources variations based on remote sensing models including Mosaic, variable infiltration capacity (VIC), commu­
methods (e.g., Forootan et al., 2014, 2017; Fatolazadeh et al. 2016; nity land model (CLM), and Noah. Due to the high accuracy of the
Khaki et al. 2018). GLDAS hydrological model, it can be used for the verification of GRACE
In this study, GRACE/GFO satellite products have been used to results (Cao et al., 2015). Due to the ability of the GLDAS Noah model in
determine TWS and GWS variations in Yazd province one of the driest providing the best spatial resolution (0.25×0.25◦ ), it has been used to
regions in central Iran. However, due to the low spatial resolution of determine the GWS from GRACE TWS.
GRACE/GFO products (Wang et al. 2022), the spatiotemporal changes of
GRACE/GFO-derived GWS from 2003 to 2020 have been validated ac­ 2.4. TRMM precipitation data
cording to changes in groundwater level and hydrograph of aquifers
(Zhong et al. 2018; Neves et al. 2020). Also, considering the role of The Tropical Rainfall Measuring Mission (TRMM), a major source of
precipitation in the recharging of groundwater resources, the relation­ grid precipitation data used in meteorological, hydrological, ecological,
ship between analytically calculated and GRACE-derived GWS has been and other fields of research (Huffman et al., 2007, Cheema and Bas­
investigated. In addition, the change in recharging rate as well as the tiaanssen, 2012), provides the most commonly used precipitation data
increase in groundwater level drop may lead to a decrease in at a geographic resolution of 0.25◦ × 0.25◦ . The TRMM data (Version 7,

2
V. Amiri et al. Journal of Hydrology 620 (2023) 129416

Fig. 1. Location of Iran in West Asia (a), DEM map (b), and aquifers of Yazd province (c).

3B43) is available from http://disc.sci.gsfc.nasa.gov/precipitation/. monitoring network. The time range and density of measurement points
Table 1 summarizes the data used in this study. are generally varied according to the importance and extent of aquifers.
For example, in aquifers such as Bafgh, Yazd-Ardakan, and Abarkouh,
2.5. In-situ groundwater level/EC data EC changes have been measured for 30, 27, and 26 years, respectively.

Trusting the results of this study and using them for water resources
management depends on their validation based on in-situ measurements 2.6. Calculation of GWS using GRACE and GLDAS data
and accuracy in the analysis (Neves et al. 2020). In this study, ground­
water level changes measured in 512 piezometers and observation wells In this study, GRACE/GFO satellite capability is used to estimate
are used to compare the results with the calculations based on GRACE/ GWS changes. The main component for these calculations is the TWS
GFO satellite products. These monthly changes are recorded regularly data received from GRACE/GFO. In addition, soil moisture storage
and are used to calculate the rate of water level drop, determine the (SMS), surface water storage (SWS), and snow water equivalent (SWE)
volume of groundwater withdrawal from the aquifer and formulate anomalies were also obtained from the Global Land Data Assimilation
aquifer management plans. It should be noted that in different aquifers, System (GLDAS) (Fang et al., 2009; Beaudoing and Rodell 2019).
these measurements were recorded in different periods. However, the The value of TWS as an important hydrological component is
data recorded between 2003 and 2020 have been used so that it is calculated as follows:
possible to compare the results with the calculations based on GRACE/
TWS = GWS + SMS + SWS + SWE
GFO satellite products.
On the other hand, groundwater EC changes have also been used as a The anomalies recorded by GLDAS are subtracted from the TWS
tool to validate the results. In each aquifer, EC is measured regularly in a value, which finally provides the GWS value as follows.

3
V. Amiri et al. Journal of Hydrology 620 (2023) 129416

considering the median of the time series can be used as an alternative

method for parametric least-squares regression in determining the trend
of univariate time series. Sen’s slope estimator (Sen 1968) was used to
examine the trend of different datasets.

2.8. Analytically estimation of GSD

It has been tried to calculate the GSD analytically and compare it

with the GRACE/GFO-derived results. For this purpose, changes in the
groundwater level have been recorded in the monitoring networks. In
addition, the specific yield (Sy) in aquifers has been determined using
pumping tests. The changes in the groundwater volume of these aquifers
are obtained according to the total fluctuation of the groundwater level
and multiplying it by the aquifer’s area and Sy. Since the volume of
groundwater extraction is higher than the annual recharging rate, this
leads to the GSD and its accumulation in consecutive years.

3. Results and discussion

3.1. Temporal variations analysis of GWS

The temporal changes of GWS from 2003 to 2020 (Fig. 3) show that
the maximum GWS was recorded in April, March, and May. Generally,
the GWS anomaly varies between 60.15 mm (recorded in May 2003) and
− 145.25 mm (recorded in October 2018). This can be justified consid­
ering the precipitation patterns that occur mainly in winter and early
spring (YRWA 2022). The monthly precipitation changes show that the
Fig. 2. The general workflow of the study. GSD is the groundwater stor­ highest precipitation was recorded in April, February, and March.
age deficit.
However, this pattern has a very high agreement with TWS fluctuations,
GLDAS anomalies, and GWS (Fig. 3).
GWS = TWS − [SMS + SWS + SWE] As shown in Fig. 3, the changes of TWS, GLDAS anomalies, and GWS
Annual, seasonal, and monthly changes in GWS evaluated concern­ have a very good agreement with the monthly observed precipitation
ing stresses such as precipitation. In arid and semi-arid areas (such as changes recorded in this region. Although, precipitation has a 2–3
this study area), which lack significant and perennial surface flows, it is months lag in recharging of GWS. Therefore, according to the significant
possible to ignore SWS to calculate GWS (Yang et al. 2015; Deng and impact of precipitation as the most important factor of aquifer recharge
Chen 2017; Frappart and Ramillien 2018; Hu et al. 2019). The flow in this province, it is expected that the calculations and interpretations
process of the methodology is shown in Fig. 1. based on GRACE-derived data have acceptable accuracy.
GLDAS anomalies have a moderate positive correlation coefficient
with GWS (0.61) and a strong positive correlation with TWS (0.81). The
2.7. Trend analysis of GWS reduction of the correlation coefficient between GWS and GLDAS can be
seen in 2003 and especially in the last 3 years of the time series. It is
Mann-Kendall (MK) test has been used to analyze the trend of GWS because of the lag between the recharge of water storage components to
anomalies. This method, which has been used by various researchers GWS. Nevertheless, the GLDAS hydrological model can be used as a
(such as Ali et al., 2022a; Ali et al., 2022b), can evaluate the possibility reliable tool for the verification of GRACE results due to its high accu­
of a monotonic increasing or decreasing trend in GWS changes based on racy (Wang et al., 2022; Yang and Chen, 2015).
the correlation coefficient between sequences and ranks of a time series.
(Akhtar et al. 2022). Details of the MK method can be found in Mann 3.2. Seasonal variations analysis of GWS
(1945) and Kendall (1975).
In linear trends, the slope can be calculated by considering the According to the water year (October-September) as well as the wet
weighted mean and using linear regression. However, this method is and dry months in this region, GWS changes have been evaluated in two
very sensitive to outliers. Therefore, Sen’s slope estimator by dry periods (April-September) and wet periods (October-March) (Fig. 4).

Table 1
Summary of the data utilized sources and timespan.
Dataset Resolution Unit Data source Data span

TWS Monthly, 0.25◦ × 0.25◦ cm GRACE CSR 2003–2020

http://www2.csr.utexas.edu/grace
Soil moisture storage Monthly, 0.25◦ × 0.25◦ Kg/m2 GLDAS 2003–2020
https://disc.gsfc.nasa.gov/datasets/GLDAS_NOAH025_M_2.1
Snow water equivalent Monthly, 0.25◦ × 0.25◦ Kg/m2 GLDAS 2003–2020
https://disc.gsfc.nasa.gov/datasets/GLDAS_NOAH025_M_2.1
Precipitation Monthly, 0.25 × 0.25
◦ ◦
mm TRMMhttps:// 2003–2020
disc.gsfc.nasa.gov/
In-situ groundwater level Monthly m YRWAhttps:// 2003–2020
www.yzrw.ir/
In-situ groundwater EC Monthly µS/cm YRWAhttps:// 2003–2020
www.yzrw.ir/

4
V. Amiri et al. Journal of Hydrology 620 (2023) 129416

Fig. 3. The monthly time series of observed precipitation, TWS, GWS, and GLDAS anomalies.

Like monthly and annual trends, in this case, GWS values are positive in 2007 onwards, the negative changes in GWS are very significant, so it
first five years of the study period, and then GWS values are negative continues until the end of the study period. The increase in groundwater
until the end of this period. The results show that in the first 5 years, the extraction along with the decrease in precipitation has caused the
distribution of precipitation in the last months of the winter and espe­ depletion of aquifers. This reduction trend is observed at different rates
cially spring has increased groundwater storage. in the unit hydrograph of aquifers (Supplementary Fig. S1).
Considering the high thickness of the unsaturated zone in most
aquifers (for example, up to 120 m in the Yazd-Ardakan aquifer, which is 3.3. Annual variations analysis of GWS
the largest aquifer in this province), these changes can be due to a delay
of about 3–4 months between the occurrence of precipitation and The annual changes of GWS show that in the first five years from
recharge to groundwater. In addition, despite the dry climate of this 2003 to 2007, the value of GWS is positive (Fig. 5). The highest amount
region, part of the precipitation in winter is in the form of snowfall. From (about 309.73 mm) was recorded in 2005. On the other hand, the

Fig. 4. The seasonal time series of GRACE-derived GWS (mm).

5
V. Amiri et al. Journal of Hydrology 620 (2023) 129416

decreasing trend starts after 2007 and the largest decrease in GWS factors, e.g., 1) the inherent leakage errors of GRACE data; 2) the errors
(1029.91 mm) for 2020 indicates a significant discharge of the aquifer in introduced by data post-processing; 3) the insufficient consideration of
this year. It should be noted that the increase in GWS in 2017 is due to a non-GWS components in the separation of GRACE-derived total water
missing gap in the measurements. The results show that without storage changes. Besides, GRACE monitors regional-scaled GWS
considering the missing gap in the dataset from July 2017 to May 2018, changes, while wells observe GWS changes at the point scales. They are
the net depletion of groundwater in this area was about − 7065 mm. inconsistent in measurement scales and observed variables.
Therefore, the annual average net loss is around 372 mm. It should be The trends of seasonal time series of GRACE-derived and observed
noted that based on the annual changes in precipitation in this area GWS also do not show the same pattern despite the expected fluctuations
(Fig. 5), it is expected that there would be an increase in the depletion of due to precipitation and aquifer recharging or increasing extraction
groundwater in the missing gap section because a decrease in precipi­ from aquifers (Fig. 8). The results show that in recent years, the differ­
tation has been recorded in the three years from 2016 to 2018. This ence between these datasets is increasing, and although both observed
means that with any delay time between precipitation and groundwater and GRACE-derived GWS datasets have a downward trend, the calcu­
recharge, GWS will decrease. lated Sen’s slope shows that the trend of GRACE-derived GWS (with test
Z = − 2.74 and significance level of α = 0.01) decreases with a steeper
slope. On the other hand, despite the increase in GWS value in the rainy
3.4. Comparison of GRACE-derived GWS with in-situ measurements seasons of recent years, the trend of the observed GWS data is also
downward (with test Z = − 0.31 and a significance level of α = 0.01).
After examining the trend of GWS changes, it has been tried to This has caused that, like the trend of monthly changes, the correlation
compare GRACE-derived GWS with in-situ measurements. As shown in of seasonal trends is also weak (Table 2).
Fig. 6, a grid of pixels with monitoring wells is selected for this com­ Annual observed and GRACE-derived GWS (Fig. 9a) has a higher
parison. After evaluating the measured dataset of the groundwater level correlation compared to monthly and seasonal time series. The results
in each of the aquifers, October 2009 to the end of 2020 was considered show that these time series have a Pearson correlation coefficient close
to compare the results. It should be noted that the largest number of to 0.58 (Table 2). If the peak created in GRACE-derived GWS in 2017
piezometers and observational wells are located in important aquifers and also the role of data gaps in the first months of 2018 (red dot-line)
such as Yazd-Ardakan, Bafqh, Abarkouh, and Aghda with areas of are not considered, their correlation coefficient would increase. As
approximately 2620, 1460, 930, and 920 Km2, respectively. Therefore, shown in Fig. 9a, both observed and GRACE-derived time series show
it was expected that the comparison results in pixels related to these negative GWS anomalies. The results of calculating the groundwater
aquifers would be more accurate. balance for the aquifers located in this region show that the renewable
The evaluation of the monthly time series of GRACE-derived and water resources in this province have decreased from 786 to 250 MCM
observed GWS for the validation stage shows that similar patterns can be between the water year 2009–2010 and 2019–2020 (YRWA, 2021a). In
seen in some months at the beginning of the considered period (Fig. 7). addition, the parameters considered to calculate the observed GWS also
However, according to the uncertainty range drawn based on the stan­ play a significant role in the difference with the GRACE-derived GWS
dard deviation of the GRACE-derived GWS dataset, for many months the dataset. Evaluations show that the domain of aquifers, as one of these
calculated and observed groundwater level changes are close to each important parameters in calculating the groundwater balance, has not
other. Nevertheless, the correlation coefficients determined based on all been updated for several years. In some cases, it can be seen that there
months of these time series show that the highest correlation coefficient are observation wells outside the current extent of the aquifer. There­
(0.27) is provided by Spearman’s method, which of course is a weak fore, this can cause errors in calculations, especially in groundwater
correlation (Table 2). It should be noted that the gap in the GRACE- storage. In addition, the non-uniform distribution of observation wells in
derived GWS dataset can be one of the reasons for reducing the corre­ some aquifers leads to an incorrect generalization of the groundwater
lation of these coefficients (Table 2). Furthermore, it is reasonable that level to large areas without monitoring networks. Considering the role of
GRACE GWS exhibits high-frequency variations. GRACE-based esti­ groundwater level in GWS calculation, this can be a source of error.
mates of GWS changes are noisy because of the influence of a range of

Fig. 5. The annual time series of TWS, precipitation, GWS, and GLDAS anomalies (mm).

6
V. Amiri et al. Journal of Hydrology 620 (2023) 129416

Fig. 6. The distribution of monitoring wells on the selected pixels.

Fig. 7. The monthly time series of observed and GRACE-derived GWS. The shaded area indicates the uncertainty of GWS determined by its standard deviation.

On the other hand, Specific yield (Sy) is one of the most important
Table 2
parameters that, along with the groundwater level, has a significant role
Nonparametric correlation of GRACE-derived and observed GWS.
in estimating GWS based on observational data. Technical reports show
Method Monthly Seasonal Annual that despite many methods including laboratory (such as simple satu­
Kendall’s tau_b 0.17 0.14 0.41 ration and drainage, correlation with particle size, and centrifuge
Spearman’s rho 0.27 0.19 0.55 moisture equivalent (CME)) (Johnson et al. 1963; Robson 1993) and
Pearson 0.17 0.21 0.58
field methods (such as field saturation and drainage, sampling after
lowering of the water table, recharge method, and pumping method) to
estimate Sy (Maliva 2016; Khadri and Moharir 2016; Lv et al. 2021;

7
V. Amiri et al. Journal of Hydrology 620 (2023) 129416

Fig. 8. Comparison of seasonal observed and GRACE-derived GWS; the gray lines represent the Sen’s slopes.

Amiri et al. 2022b), the pumping test method has been used in all water quality due to a factor such as climate change or overexploitation
aquifers. Although this method has been used by various researchers, of groundwater can lead to the occurrence of some diseases and increase
nevertheless, if the aquifers are heterogeneous and anisotropic, the treatment costs (Mousavi et al. 2020). Meanwhile, the study of 206
number and distribution of pumping tests must be carefully determined groundwater samples collected in Yazd province in 2018 showed that 33
to obtain accurate results. Otherwise, the results will not be generaliz­ and 18% of the samples in wet and dry seasons, respectively, are not
able to other places. Investigations show that one of the most important suitable for children to drink (Amiri et al. 2021b). Today, a significant
limitations for the numerical modeling of aquifers in this region is the part of drinking water in some cities, especially Yazd the capital of this
lack of access to the hydrodynamic parameters of the aquifer. This has province, is supplied from nearby catchment basins (Amiri et al., 2021a;
caused the uncertainty of the results to be mentioned in various reports. Amiri et al., 2021b).
In Fig. 9b, the monthly average of the observed and GRACE-derived
GWS dataset is presented. The results show that the pattern of GRACE-
derived GWS changes is more consistent with meteorological condi­ 3.5. Spatio-temporal changes analysis of TWS and GWS
tions (i.e. rainy and dry months). This pattern has been observed in some
studies (such as Akhtar et al., 2022, Ali et al., 2023). Changes in TWS show that from 2003 to 2007 water sources have
Changes in salinity can be suitable indices for determining and favorable conditions (Supplementary Fig. S2). Meanwhile, from 2007
validating the process of storage changes and groundwater fluctuations. onwards, TWS gradually decreases in the southern parts. The largest
The results show that between 2009 and 2021, the electrical conduc­ spatial decrease was recorded in the southern region of the Yazd in
tivity (EC) increased from about 8500 to 12,500 μS/cm (Fig. 10). To 2016, which can be justified considering the annual changes in precip­
calculate the representative EC of each year, the sum of the weighted itation (Fig. 9a). The changes in GWS also have the same trend as TWS,
seasonal values of EC in each aquifer was used. The weight of each so after the first few years with a relatively good situation in terms of
measurement is based on the ratio of the area of each aquifer to the total groundwater storage, GWS has experienced a sharper decline since 2008
area of 26 aquifers in the study area. The correlation coefficient between (Supplementary Fig. S3). Likewise, the southern and eastern aquifers
EC and observed GWS is around − 0.50, which indicates a negative have experienced the greatest decrease in GWS. Of course, it should be
correlation. On the other hand, regardless of the missing data gap, EC noted that some of the most important aquifers in this region have faced
and GRACE-derived GWS show a strong negative correlation (r = a smaller decrease in GWS in some years due to their convenient
− 0.79). This means that old high-density water may be pumped as the geographical location and the possibility of natural recharging through
groundwater level decreases and reaches higher depths. On the other precipitation events on the plains or the surrounding mountains (YRWA,
hand, Thaw et al. (2022) showed that excessive pumping of ground­ 2021b).
water can transport polluted modern surface and groundwaters to The evaluation of TWS distribution from 2003 to 2020 shows that its
greater depths and increase their pollution and salinity. Therefore, it value decreases from north to south. In other words, these changes
seems that despite the location of some parts of the selected pixels in directly affect the value of GWS and it was expected that a similar trend
areas without water sources such as mountains and deserts, and also the would be observed for it. Based on this, the highest values of TWS are
lack of identification of these sources by field studies and analy­ seen in aquifers such as Aghda, Rigzarrin, and Siahkouh, and the lowest
tical–numerical methods, this strong negative correlation coefficient can values are seen in southern aquifers such as Chahak Shahriari, Marvast,
verify the results obtained based on the GRACE data. Herat, and Bahadoran (Fig. 11a). On the other hand, the amount of TWS
An increase of nearly 47% in EC has led to a decrease in the quality of increases from east to west. In this case, eastern aquifers such as Bafgh
groundwater for many uses. Therefore, it is impossible to use it in some and Bahabad have lower TWS values compared to western aquifers such
areas without desalination. On the other hand, the poor quality and as Abarghu, Dehshir, and Ernan-Dahaj. These changes can be justified
significant drop in groundwater level have caused the cost of using these according to the precipitation pattern (YRWA, 2021b) as well as the
groundwater resources to increase. On the other hand, the decrease in topography of the region.
The TWS trend is in good agreement with the GWS (Fig. 11b). This is

8
V. Amiri et al. Journal of Hydrology 620 (2023) 129416

Fig. 9. Comparison of observed and GRACE-derived GWS at (a) annual and (b) mean monthly.

not far from expected considering the strong positive correlation coef­ the increase in population, the development of cities and villages, as well
ficient of around 0.97 of these time series. The climate reports show that as the necessity of supplying part of the water needs of the agricultural
the average maximum and minimum temperatures have increased by 1 and industrial sectors. However, Scanlon et al. (2023) also recently
to 2.5 ◦ C, respectively in the last 50 years. At the same time, the rate of showed that climate variability causes some changes in water storage,
evaporation has increased by about 3 mm/year. Ashraf Vaghef et al. but human interventions, particularly overexploitation of groundwater
(2019) also predicted the future climatic conditions of Iran, especially in for irrigation have a more important role. Similar results were reported
the central arid regions, and showed that dry periods are expected to be by Nouri et al. (2021) and indicated that groundwater decline due to
longer and intermittent heavy rainfalls will lead to an increase in the overexploitation of groundwater and increasing salinity are documented
probability of floods. In addition, except for some mountainous areas, in almost all aquifers.
the average long-term annual precipitation in most parts of this region is
less than 150 mm (YRWA, 2021b). This is consistent with Mansouri 3.6. Trend analysis of GWS
Daneshvar et al. (2019) who believe that climate change due to green­
house gas emissions caused by the production of oil, gas, and rapid ur­ The results show that the annual minimum, maximum, and average
banization will increase the temperature and reduce rainfall by nearly GWS estimated from 2003 to 2020 based on Sen’s slope method are
35%. However, the decrease in GWS and TWS can be mainly due to the about − 972.59 mm, 182.64 mm, and − 404.88 mm, respectively. These
overexploitation of groundwater for various purposes, especially agri­ trends are significant at α = 0.001 level. In addition, the value of Test Z is
culture, and the lack of significant surface water resources in this region. equal to − 4.17 mm, which shows a downward trend (Fig. 12). The ab­
Amiri et al. (2021b) showed that from 2006 to 2018, the extraction of solute value of test statistic Z is compared to the standard normal cu­
groundwater for drinking experienced a 250% increase. This is due to mulative distribution to define if there is a trend or not to the selected

9
V. Amiri et al. Journal of Hydrology 620 (2023) 129416

Fig. 10. Comparison of the trends of annual changes of EC and GWS. The red dot line indicates incomplete data in the corresponding year. (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of this article.)

level α of significance. A positive (negative) value of Z indicates an (11865 km2/42530 km2). The results show that the GRACE-derived GSD
upward (downward) trend. It should be noted that the trend of TWS with is around 2228 MCM (Fig. 13). It should be noted that if the average
Test Z equal to − 4.24 mm and a significance level of α = 0.001 has the GRACE-derived GSD of the same period before and after the missing gap
same changes as GWS. However, the general trend of GLDAS values in the dataset from July 2017 to May 2018 (approximately − 265 MCM)
decreases with a lower slope (Test Z = − 2.88 mm, significance level of α is added to it, the GSD will increase to about 2491 MCM.
= 0.01). One of the most important reasons for the difference between the
The reported groundwater storage deficit (GSD) in the study period groundwater storage deficit calculated based on in-situ measurements
shows the same downward trend, but its rate is smaller than GWS, which and the GRACE dataset is the lack of a standard groundwater level
was discussed above (see Fig. 5). According to Sen’s slope estimator, the monitoring network. This has caused the calculation of the water bal­
minimum, maximum, and average of GSD trend (with a significance ance in these aquifers and subsequently the GSD in Yazd province to be
level of α = 0.1) are about − 0.27, − 20 and − 0.23 BCM (from 2003 to uncertain. It should be noted that the area of aquifers announced by
2020), respectively. Therefore, regardless of various natural (such as Yazd Regional Water Company was used to calculate the GRACE-
changing climatic and hydrological components) and anthropogenic derived GSD. However, it seems that in many cases, the aquifer do­
(such as increasing groundwater extraction) changes, the groundwater mains should be determined again according to the distribution of
balance is negative. It should be noted that GWS changes were based on pumping wells and piezometers. Because in some cases, they are outside
GRACE satellite measurements from all regions, but changes in the aquifer boundary. Another reason for the difference between the
groundwater balance deficit were done analytically only for aquifers calculations is the Sy values. The inadequacy and non-uniform distri­
(see the next section). bution of pumping tests in some aquifers cause the estimated Sy to have
Depletion of aquifers, regardless of the rates determined by mea­ high uncertainty.
surements or calculated values based on GRACE data, has led to the
deterioration of groundwater quality and also the occurrence of land 4. Conclusion
subsidence in some areas. For example, Mirzadeh et al. (2021) showed
that in some parts of the Yazd-Ardakan aquifer, the rate of subsidence The results show that the annual changes of GWS are positive from
caused by groundwater depletion between 2003 and 2020 was even up 2003 to 2007, and after that, it has decreased significantly, so that it
to 15 cm/year. continues almost until the end of the study period. The increase in
groundwater extraction along with the decrease in precipitation has
caused the depletion of aquifers. The southern and eastern aquifers in
3.7. Estimation of GSD the Yazd region have experienced the greatest decrease in GWS. The
minimum, maximum, and average GWS estimated based on Sen’s slope
The changes in the aquifer storage were evaluated between October method, which represents the general trend of the annual data, are about
2009 and December 2020. The evaluation of groundwater storage − 972.59, 182.64, and − 404.88 mm, respectively. The results of the
changes based on the water level records shows that the GSD is around analytically calculated GSD in this period show the same downward
2775 MCM (Fig. 13). This is relatively consistent with the water balance trend, but its rate is smaller than GWS. The minimum, maximum, and
calculation report (approximately 3050 MCM until October 2021) average of this data set are about − 0.27, − 20, and − 0.23 BCM,
(YRWA 2022). The difference between the calculated values in this respectively. Therefore, regardless of various natural changes and
study and the 3050 MCM reported by the Yazd Regional Water Company anthropogenic, the balance of groundwater resources is negative and
is due to not considering the changes caused by the 6 months of April to this has led to a sharp decrease in groundwater reserves. The annual
October 2021. observed and calculated GWS anomalies have a Pearson correlation
On the other hand, the GRACE-derived GSD was calculated by coefficient close to 0.58, which is higher than the coefficients obtained
multiplying GWS anomaly by 0.28 (aquifers area/GRACE net area

10
V. Amiri et al. Journal of Hydrology 620 (2023) 129416

Fig. 11. The net changes of (a) TWS (mm) and (b) GWS (mm) during 2003–2020.

for monthly and seasonal time series. Both observed and GRACE-derived and reaches higher depths. Evaluations show that the domain of aquifers
GWS show negative anomalies. The results of calculating the ground­ has not been updated for several years. In addition, the non-uniform
water balance for the aquifers show that the renewable water resources distribution of piezometers and observation wells in some aquifers
have decreased from 786 to 250 MCM between the water year leads to a difference between observed and GRACE-derived GWS
2009–2010 and 2019–2020. The results show that between 2009 and anomalies.
2021, EC increased from about 8500 to 12,500 μS/cm. The correlation
coefficient between EC and observed GWS is around − 0.50, which in­ CRediT authorship contribution statement
dicates the average negative correlation of these two trends. On the
other hand, the correlation coefficient between EC and GRACE-derived Vahab Amiri: Conceptualization, Software. Shoaib Ali: Methodol­
GWS shows a strong negative correlation (r = − 0.79) without consid­ ogy, Software. Nassim Sohrabi: Data analysis.
ering the years 2017 and 2021 (due to the data gap). This means that old
high-density water may be pumped as the groundwater level decreases

11
V. Amiri et al. Journal of Hydrology 620 (2023) 129416

Fig. 12. The trend of (a) GRACE-derived GWS and (b) reported groundwater storage deficit changes from 2003 to 2020.

Fig. 13. The cumulative GSD during 2009–2020.

Declaration of Competing Interest References

The authors declare that they have no known competing financial Ahamed, A., Knight, R., Alam, S., Pauloo, R., Melton, F., 2022. Assessing the utility of
remote sensing data to accurately estimate changes in groundwater storage. Sci.
interests or personal relationships that could have appeared to influence Total Environ. 807, 150635 https://doi.org/10.1016/j.scitotenv.2021.150635.
the work reported in this paper. Akhtar, F., Nawaz, R.A., Hafeez, M., Awan, U.K., Borgemeister, C., Tischbein, B., 2022.
Evaluation of GRACE derived groundwater storage changes in different agro-
ecological zones of the Indus Basin. J. Hydrol. 605, 127369 https://doi.org/
Data availability 10.1016/j.jhydrol.2021.127369.
Ali, S., Liu, D., Fu, Q., Cheema, M.J.M., Pham, Q.B., Rahaman, M., Dang, T.D., Anh, D.T.,
Data will be made available on request. 2021. Improving the resolution of GRACE data for spatio-temporal groundwater
storage assessment. Remote Sens. 13 (17), 3513. https://doi.org/10.3390/
rs13173513.
Acknowledgments Ali, S., Liu, D., Fu, Q., Cheema, M.J.M., Chandra Pal, S., Arshad, A., Pham, Q.B.,
Zhang, L., 2022a. Constructing high-resolution groundwater drought at spatio-
temporal scale using GRACE satellite data based on machine learning in the Indus
The authors would like to thank the Yazd Regional Water Authority
Basin. J. Hydrol. 612, 128295 https://doi.org/10.1016/j.jhydrol.2022.128295.
for providing part of the data. Ali, S., Wang, Q., Liu, D., Fu, Q., Rahaman, M.M., Faiz, M.A., Cheema, M.J.M., 2022b.
Estimation of spatio-temporal groundwater storage variations in the Lower
Appendix A. Supplementary data Transboundary Indus Basin using GRACE satellite. J. Hydrol. 605, 127315 https://
doi.org/10.1016/j.jhydrol.2021.127315.
Ali, S., Khorrami, B., Jehanzaib, M., Tariq, A., Ajmal, M., Arshad, A., Shafeeque, M.,
Supplementary data to this article can be found online at https://doi. Dilawar, A., Basit, I., Zhang, L., Sadri, S., Niaz, M.A., Jamil, A., Khan, S.N., 2023.
org/10.1016/j.jhydrol.2023.129416. Spatial Downscaling of GRACE Data Based on XGBoost Model for Improved
Understanding of Hydrological Droughts in the Indus Basin Irrigation System (IBIS).
Remote Sens. 15 (4), 873.
Amiri, V., Bhattacharya, P., Nakhaei, M., 2021a. The hydrogeochemical evaluation of
groundwater resources and their suitability for agricultural and industrial uses in an

12
V. Amiri et al. Journal of Hydrology 620 (2023) 129416

arid area of Iran. Groundw. Sustain. Dev. 12, 100527 https://doi.org/10.1016/j. groundwater drought. Agric. For. Meteorol. 307, 108476 https://doi.org/10.1016/j.
gsd.2020.100527. agrformet.2021.108476.
Amiri, V., Kamrani, S., Ahmad, A., Bhattacharya, P., Mansoori, J., 2021b. Groundwater Herrera-Garcia, G., Ezquerro, P., Tomas, R., Bejar-Pizarro, M., Lopez-Vinielles, J., et al.,
quality evaluation using Shannon information theory and human health risk 2021. Mapping the global threat of land subsidence. Science 371 (6524), 34–36.
assessment in Yazd province, central plateau of Iran. Environ. Sci. Pollut. Res. 28 (1), https://doi.org/10.1126/science. abb8549.
1108–1130. https://doi.org/10.1007/s11356-020-10362-6. Houborg, R., Rodell, M., Li, B., Reichle, R., Zaitchik, B., 2012. Drought indicators based
Amiri, V., Nakhaei, M., Lak, R., Li, P., 2021c. An integrated statistical-graphical approach on model assimilated GRACE terrestrial water storage observations. Wat. Resour.
for the appraisal of the natural background levels of some major ions and potentially Res. 48, W07525. https://doi.org/10.1029/2011WR011291.
toxic elements in the groundwater of Urmia aquifer. Iran. Environ. Earth. Sci. 80 Huffman, G.J., Adler, R.F., Bolvin, D.T., Gu, G.J., Nelkin, E.J., Bowman, K.P., Hong, Y.,
https://doi.org/10.1007/s12665-021-09733-0. Stocker, E.F., Wolff, D.B., 2007. The TRMM multisatellite precipitation analysis
Amiri, V., Sohrabi, N., Li, P., Amiri, F., 2022a. Groundwater Quality for Drinking and (TMPA): quasi-global, multiyear, combined-sensor precipitation estimates at fine
Non-Carcinogenic Risk of Nitrate in Urban and Rural Areas of Fereidan. Iran. Expos. scales. J. Hydrometeorol. 8, 38–55. https://doi.org/10.1175/JHM560.1.
Health. https://doi.org/10.1007/s12403-022-00525-w. Jacob, T., Wahr, J., Pfeffer, W.T., Swenson, S., 2012. Recent contributions of glaciers and
Amiri, V., Sohrabi, N., Li, P., Shukla, S., 2022b. Estimation of hydraulic conductivity and ice caps to sea-level rise. Nature 482, 514–518. https://doi.org/10.1038/
porosity of a heterogeneous porous aquifer by combining transition probability nature10847.
geostatistical simulation, geophysical survey, and pumping test data. Environ. Dev. Johnson, A.I., Prill, R.C., Morris, D.A., 1963. Specific Yield- Column Drainage and
Sustain. https://doi.org/10.1007/s10668-022-02368-6. Centrifuge Moisture Content. U.S. Geological Survey.
Arshad, A., Mirchi, A., Samimi, M., Ahmad, B., 2022. Combining downscaled-GRACE Joodaki, G., Wahr, J., Swenson, S., 2014. Estimating the human contribution to
data with SWAT to improve the estimation of groundwater storage and depletion groundwater depletion in the Middle East, from GRACE data, land surface models,
variations in the Irrigated Indus Basin (IIB). Sci. Total Environ. 838, 156044 https:// and well observations. Water Resour. Res. 50, 2679–2692. https://doi.org/10.1002/
doi.org/10.1016/j.scitotenv.2022.156044. 2013WR014633.
Ashraf, S., Nazemi, A., Agha Kouchak, A., 2021. Anthropogenic drought dominates Kendall, M., 1975. Multivariate Analysis. Charles Griffin b Co. LTD. London, High
groundwater depletion in Iran. Scientific Reports 11, 9135. https://doi.org/ Wycombe.
10.1038/s41598-021-88522-y. Khadri, S.F.R., Moharir, K., 2016. Characterization of aquifer parameter in basaltic hard
Beaudoing, H., Rodell, M., 2019. NASA/GSFC/HSL: GLDAS Noah Land Surface Model L4 rock region through pumping test methods: a case study of Man River basin in Akola
monthly 0.25×0.25 degree V2. 0. Goddard Earth Sciences Data and Information and Buldhana districts Maharashtra India. Model. Earth syst. Environ. 2, 33. https://
Services Center (GES DISC), Greenbelt, Maryland, USA. doi.org/10.1007/s40808-015-0047-9.
Cao, Y.P., Nan, Z.T., Cheng, G.D., 2015. GRACE gravity satellite observations of errestrial Khaki, M., Forootan, E., Kuhn, M., Awange, J., van Dijk, A.I.J.M., Schumacher, M.,
water storage changes for drought characterization in the arid land of Northwestern Sharifi, M.A., 2018. Determining Water Storage Depletion within Iran by
China. Remote Sens 7, 1021–1047. https://doi.org/10.3390/rs70101021. Assimilating GRACE data into the W3RA Hydrological Model. Adv. Water Resour.
Castellazzi, P., Longuevergne, L., Martel, R., Rivera, A., Brouard, C., Chaussard, E., 2018. 114, 1–18. https://doi.org/10.1016/j.advwatres.2018.02.008.
Quantitative mapping of groundwater depletion at the water management scale Khorrami, B., Gunduz, O., 2021. An enhanced water storage deficit index (EWSDI) for
using a combined GRACE/InSAR approach. Remote Sens. Environ. 205, 408–418. drought detection using GRACE gravity estimates. J. Hydrol. 603, 126812 https://
https://doi.org/10.1016/j.rse.2017.11.025. doi.org/10.1016/j.jhydrol.2021.126812.
Chambers, D.P., 2006. Observing seasonal steric sea level variations with GRACE and Landerer, F.W., Flechtner, F.M., Save, H., Webb, F.H., Bandikova, T., Bertiger, W.I.,
satellite altimetry. J. Geophys. Res. 111 (C3), C03010. https://doi.org/10.1029/ Bettadpur, S.V., Byun, S.H., Dahle, C., Dobslaw, H., Fahnestock, E., Harvey, N.,
2005JC002914. Kang, Z., Kruizinga, G.L.H., Loomis, B.D., McCullough, C., Murböck, M., Nagel, P.,
Chawla, I., Karthikeyan, L., Mishra, A.K., 2020. A review of remote sensing applications Paik, M., Pie, N., Poole, S., Strekalov, D., Tamisiea, M.E., Wang, F., Watkins, M.M.,
for water security: Quantity, quality, and extremes. J. Hydrol. 585, 124826 https:// Wen, H.-Y., Wiese, D.N., Yuan, D.-N., 2020. Extending the global mass change data
doi.org/10.1016/j.jhydrol.2020.124826. record: GRACE follow-on instrument and science data performance. Geophys. Res.
Cheema, M.J.M., Bastiaanssen, W.G.M., 2012. Local calibration of remotely sensed Lett. 47 (12) https://doi.org/10.1029/2020GL088306.
rainfall from the TRMM satellite for different periods and spatial scales in the Indus Li, X., Zhong, B., Li, J., Liu, R., 2022. Analysis of terrestrial water storage changes in the
Basin. Int. J. Remote Sens. 33, 2603–2627. https://doi.org/10.1080/ Shaan-Gan-Ning Region using GPS and GRACE/GFO. Geod. Geodyn. 13, 179–188.
01431161.2011.617397. https://doi.org/10.1016/j.geog.2021.11.001.
Chen, H., Zhang, W., Nie, N., Guo, Y., 2019. Long-term groundwater storage variations Liu, D., Mishra, A.K., Yu, Z., Lu, H., Li, Y., 2021. Support vector machine and data
estimated in the Songhua River Basin by using GRACE products, land surface assimilation framework for Groundwater Level Forecasting using GRACE satellite
models, and in-situ observations. Sci. Total Environ. 649, 372–387. https://doi.org/ data. J. Hydrol. 603, 126929 https://doi.org/10.1016/j.jhydrol.2021.126929.
10.1016/j.scitotenv.2018.08.352. Long, D., Yang, W., Scanlon, B.R., Zhao, J., Liu, D., Burek, P., Pan, Y., You, L., Wada, Y.,
de Graaf, I.E.M., Gleeson, T., van Beek, L.P.H., Sutanudjaja, E.H., Bierkens, M.F.P., 2019. 2020. South-to-North Water Diversion stabilizing Beijing’s groundwater levels. Nat.
Environmental flow limits to global groundwater pumping. Nature 574, 90–94. Commun. 11, 3665. https://doi.org/10.1038/s41467-020-17428-6.
https://doi.org/10.5683/SP2/D7I7CC. Lv, M., Xu, Z., Yang, Z.-L., Lu, H., Lv, M., 2021. A Comprehensive Review of Specific
de Linage, C., Rivera, L., Hinderer, J., Boy, J.-P., Rogister, Y., Lambotte, S., Biancale, R., Yield in Land Surface and Groundwater Studies. J. Adv. Model. Earth Syst. 13 (2).
2009. Separation of coseismic and postseismic gravity changes for the 2004 Maliva, R.G., 2016. Aquifer Pumping Tests. In: Aquifer Characterization Techniques.
Sumatra-Andaman earthquake from 4.6 yr of GRACE observations and modelling of Springer Hydrogeology. Springer, Cham. https://doi.org/10.1007/978-3-319-
the coseismic change by normal-modes summation. Geophys. J. Int. 176, 695–714. 32137-0_7.
https://doi.org/10.1111/j.1365-246X.2008.04025. Mann, H.B., 1945. Nonparametric Tests Against Trend. Description. Econometrica 13,
Famiglietti, J.S., Rodell, M., 2013. Water in the Balance. Science 340, 1300–1301. 163–171. https://doi.org/10.2307/1907187.
https://doi.org/10.1126/science.1236460. Mansouri Daneshvar, M.R., Ebrahimi, M., Nejadsoleymani, H., 2019. An overview of
Fatolazadeh, F., Goïta, K., 2022. Reconstructing groundwater storage variations from climate change in Iran: facts and statistics. Environ. Syst. Res. 8, 7. https://doi.org/
GRACE observations using a new Gaussian-Han-Fan (GHF) smoothing approach. 10.1186/s40068-019-0135-3.
J. Hydrol. 604, 127234 https://doi.org/10.1016/j.jhydrol.2021.127234. Mehrnegar, N., Jones, O., Singer, M.B., et al., 2021. Exploring groundwater and soil
Fatolazadeh, F., Voosoghi, B., Raoofian Naeeni, M., 2016. Wavelet and Gaussian water storage changes across the CONUS at 12.5 km resolution by a Bayesian
Approaches for Estimation of Groundwater Variations Using GRACE Data. integration of GRACE data into W3RA. Sci. Total Environ. 758, 143579 https://doi.
Groundwater 54 (1), 74–81. https://doi.org/10.1111/gwat.12325. org/10.1016/j.scitotenv.2020.143579.
Forootan, E., Rietbroek, R., Kusche, J., Sharifi, M.A., Awange, J., Schmidt, M., Mirzadeh, S.M.J., Jin, S., Parizi, E., Chaussard, E., Bürgmann, R., Blasco, J.M.D.,
Omondi, P., Famiglietti, J., 2014. Separation of large scale water storage patterns Amani, M., Bao, H., Mirzadeh, S.H., 2021. Characterization of Irreversible Land
over Iran using GRACE, altimetry and hydrological data. Remote Sens. Environ. 140, Subsidence in the Yazd Ardakan Plain, Iran from 2003–2020 InSAR Time Series.
580–595. https://doi.org/10.1016/j.rse.2013.09.025. Earth and Space Science Open Archive. https://doi.org/10.1002/
Forootan, E., Safari, A., Mostafaie, A., Schumacher, M., Delavar, M., Awange, J.L., 2017. essoar.10506902.1.
Large-Scale Total Water Storage and Water Flux Changes over the Arid and Semiarid Mishra, A.K., Ines, A.V.M., Das, N.N., Prakash Khedun, C., Singh, V.P., Sivakumar, B.,
Parts of the Middle East from GRACE and Reanalysis Products. Surv. Geophys. 38, Hansen, J.W., 2015. Anatomy of a local-scale drought: Application of assimilated
591–615. https://doi.org/10.1007/s10712-016-9403-1. remote sensing products, crop model, and statistical methods to an agricultural
Freedman, F.R., Pitts, K.L., Bridger, A.F.C., 2014. Evaluation of CMIP climate model drought study. J. Hydro. 526, 15–29.
hydrological output for the Mississippi River Basin using GRACE satellite Motagh, M., Walter, T.R., Sharifi, M.A., Fielding, E., Schenk, A., Anderssohn, J., et al.,
observations. J. Hydrol. 519, 3566–3577. https://doi.org/10.1016/j. 2008. Land subsidence in Iran caused by widespread water reservoir
jhydrol.2014.10.036. overexploitation. Geophys. Res. Lett. 35, L16403. https://doi.org/10.1029/
Gleeson, T., Wang-Erlandsson, L., Porkka, M., Zipper, S.C., Jaramillo, F., Gerten, D., 2008GL033814.
Fetzer, I., Cornell, S.E., Piemontese, L., Gordon, L.J., Rockström, J., Oki, T., Mousavi, A., Ardalan, A., Takian, A., Ostadtaghizadeh, A., Naddafi, K., Bavani, A.M.,
Sivapalan, M., Wada, Y., Brauman, K.A., Flörke, M., Bierkens, M.F.P., Lehner, B., 2020. Climate change and health in Iran: a narrative review. J. Environ. Health Sci.
Keys, P., Kummu, M., Wagener, T., Dadson, S., Troy, T.J., Steffen, W., Eng. 18 (1), 367–378.
Falkenmark, M., Famiglietti, J.S., 2020. Illuminating water cycle modifications and Nagale, D.S., Kannaujiya, S., Gautam, P.K., Taloor, A.K., Sarkar, T., 2022. Impact
Earth system resilience in the Anthropocene. Water Resour. Res. 56 (4) https://doi. assessment of the seasonal hydrological loading on geodetic movement and
org/10.1029/2019WR024957. seismicity in Nepal himalaya using GRACE and GNSS measurements. Geodesy and
Han, Z., Huang, S., Huang, Q., Leng, G., Liu, Y., Bai, Q., He, P., Liang, H., Shi, W., 2021. Geodynamics 13 (5), 445–455.
GRACE-based high-resolution propagation threshold from meteorological to Neves, M.C., Nunes, L.M., Monteiro, J.P., 2020. Evaluation of GRACE data for water
resource management in Iberia: a case study of groundwater storage monitoring in

13
V. Amiri et al. Journal of Hydrology 620 (2023) 129416

the Algarve region. J. Hydrol. Reg. Stud. 32, 100734 https://doi.org/10.1016/j. Taylor, C. J., Alley, W.M., 2001. Ground-water-level monitoring and the importance of
ejrh.2020.100734. long-term water-level data (Vol. 1217). US Geological Survey Denver, CO, USA.
Noori, R., Maghrebi, M., Mirchi, A., Tang, Q., Bhattarai, R., Sadegh, M., Noury, M., Thaw, M., GebreEgziabher, M., Villafañe-Pagán, J.Y., Jasechko, S., 2022. Modern
Torabi Haghighi, A., Kløve, B., Madani, K., 2021. Anthropogenic depletion of Iran’s groundwater reaches deeper depths in heavily pumped aquifer systems. Nat.
aquifers. PNAS 118 (25). Commun. https://doi.org/10.1038/s41467-022-32954-1.
Robson, S.G., 1993. Techniques for estimating specific yield and specific retention from Vaghefi, S.A., Keykhai, M., Jahanbakhshi, F., Sheikholeslami, J., Ahmadi, A., Yang, H.,
grain-size data and geophysical logs from clastic bedrock aquifers. U.S. Geological Abbaspour, K.C., 2019. The future of extreme climate in Iran. Sci. Rep. 9, 1464.
Survey, Water-Resources Investigations Report 93-4198. https://doi.org/10.1038/s41598-018-38071-8.
Rodell, M., Chen, J., Kato, H., Famiglietti, J.S., Nigro, J., Wilson, C.R., 2007. Estimating Wang, F., Lai, H., Li, Y., Feng, K., Zhang, Z., Tian, Q., Zhu, X., Yang, H., 2022. Identifying
groundwater storage changes in the Mississippi River Basin (USA) using GRACE. the status of groundwater drought from a GRACE mascon model perspective across
Hydrogeol. J. 15, 159–166. https://doi.org/10.1007/s10040-006-0103-7. China during 2003–2018. Agric. Water Manag. 260, 107251 https://doi.org/
Rodell, M., Velicogna, I., Famiglietti, J.S., 2009. Satellite-based estimates of groundwater 10.1016/j.agwat.2021.107251.
depletion in India. Nature 460 (7258), 999–1002. Wu, Q., Si, B., He, H., Wu, P., 2019. Determining regional-scale groundwater recharge
Saemian, P., Tourian, M.J., AghaKouchak, A., Madani, K., Sneeuw, N., 2022. How much with GRACE and GLDAS. Remote Sens. 11 (2), 154. https://doi.org/10.3390/
water did Iran lose over the last two decades? J. Hydrol. Reg. Stud. 41, 101095 rs11020154.
https://doi.org/10.1016/j.ejrh.2022.101095. Xiong, J., Yin, J., Guo, S., Gu, L., Xiong, F., Li, N., 2021. Integrated flood potential index
Scanlon, B.R., Fakhreddine, S., Rateb, A. et al., 2023. Global water resources and the role for flood monitoring in the GRACE era. J. Hydrol. 603, 127115 https://doi.org/
of groundwater in a resilient water future. Nat. Rev. Earth Environ. https://doi.org/ 10.1016/j.jhydrol.2021.127115.
10.1038/s43017-022-00378-6. Yan, X., Zhang, B., Yao, Y., Yang, Y., Li, J., Ran, Q., 2021. GRACE and land surface
Sen, P.K., 1968. Estimates of the regression coefficient based on Kendall’s Tau. J. Am. models reveal severe drought in eastern China in 2019. J. Hydrol. 601, 126640
Stat. Assoc. 63, 324. https://doi.org/10.2307/2285891. https://doi.org/10.1016/j.jhydrol.2021.126640.
Sjöberg, L.E., Bagherbandi, M., 2020. Upper mantle density and surface gravity change in Yan, X., Zhang, B., Yao, Y., Yin, J., Wang, H., Ran, Q., 2022. Jointly using the GLDAS 2.2
Fennoscandia, determined from GRACE monthly data. Tectonophysics 782–783, model and GRACE to study the severe Yangtze flooding of 2020. J. Hydrol. 610,
228428. https://doi.org/10.1016/j.tecto.2020.228428. 127927 https://doi.org/10.1016/j.jhydrol.2022.127927.
Smith, M., Cross, K., Paden, M., Laban, P., 2016. Spring -Managing groundwater Yang, P., Chen, Y.N., 2015. An analysis of terrestrial water storage variations from
sustainability. IUCN, Gland, Switzerland. GRACE and GLDAS: the Tianshan Mountains and its adjacent areas, central Asia.
Srivastava, S., Dikshit, O., 2022. Analysis of groundwater storage (GWS) dynamics and Quat. Int. 358, 106–112. https://doi.org/10.1016/j.quaint.2014.09.077.
its temporal evolution for The Indo-Gangetic plain using GRACE data. Remote Sens. YRWA, 2021a. Annual and cumulative storage deficit of aquifers in Yazd province.
Appl.: Soc. Environ. 25 (100685) https://doi.org/10.1016/j.rsase.2021.100685. Internal report (in Persian).
Swenson, S.C., Chambers, D.P., Wahr, J., 2008. Estimating geocenter variations from a YRWA, 2021b. Precipitation report of Yazd province. Internal report (in Persian).
combination of GRACE and ocean model output. J Geophys. Res.-Solid. Earth 113 YRWA (Yazd Regional Water Authority) 2022. The status of water resources in Yazd
(B8). https://doi.org/10.1029/2007JB005338. province. Office of Basic Studies of Water Resources. Internal report (in Persian).
Tangdamrongsub, N., Han, S.-C., Yeo, I.-Y., Dong, J., Steele-Dunne, S.C., Willgoose, G., Zhang, X., Li, J., Dong, Q., Wang, Z., Zhang, H., Liu, X., 2022a. Bridging the gap between
Walker, J.P., 2020. Multivariate data assimilation of GRACE, SMOS, SMAP GRACE and GRACE-FO using a hydrological model. Sci. Total Environ. 822, 153659
measurements for improved regional soil moisture and groundwater storage https://doi.org/10.1016/j.scitotenv.2022.153659.
estimates. Adv. Water Resour. 135, 103477 https://doi.org/10.1016/j. Zhang, X., Ren, L., Feng, W., 2022b. Comparison of the shallow groundwater storage
advwatres.2019.103477. change estimated by a distributed hydrological model and GRACE satellite
Tapley, B.D., Watkins, M.M., Flechtner, F., Reigber, C., Bettadpur, S., Rodell, M., gravimetry in a well-irrigated plain of the Haihe River basin. China. J. Hydrol. 610,
Sasgen, I., Famiglietti, J.S., Landerer, F.W., Chambers, D.P., Reager, J.T., Gardner, A. 127799 https://doi.org/10.1016/j.jhydrol.2022.127799.
S., Save, H., Ivins, E.R., Swenson, S.C., Boening, C., Dahle, C., Wiese, D.N., Zhong, Y., Zhong, M., Feng, W., Zhang, Z., Shen, Y., Wu, D., 2018. Groundwater
Dobslaw, H., Tamisiea, M.E., Velicogna, I., 2019. Contributions of GRACE to depletion in the West Liaohe River Basin, China and its implications revealed by
understanding climate change. Nat. Clim. Chang. 9 (5), 358–369. GRACE and in situ measurements. Remote Sens. 10 (4), 493. https://doi.org/
10.3390/rs10040493.

14