Investigation of Groundwater Depletion in the State of Qatar and Its Implication to Energy Water and Food Nexus

: Groundwater is a precious freshwater resource heavily relied upon by agricultural activities in many parts of the world, and especially by countries with limited water resources located in arid regions. Groundwater resources are under severe pressures due to population increase, urbanisation and socio-economic development, with potential for causing long-term threats to human life and natural ecosystems. This study attempts to investigate the impacts of local and regional climatic trends, and establish key forcing functions that have changed local groundwater resources. The main questions answered through this study include: Are these changes beneﬁcial or detrimental? If they are detrimental, what is the future outlook for impacts on the ecosystem? What are the corrective actions needed to avert the long-term risks in arid environments? In view of this, the methodology developed in this study focuses on a joint time-series statistical analysis using ground data as well as Gravity Recovery and Climate Experiment (GRACE) satellite data. Results show a substantial depletion in the groundwater thickness (0.24 ± 0.20 cm/year) during the period of observation (2002–2020). Long-term temperature data indicates that the annual mean temperature increased signiﬁcantly by 1.02 ◦ C between 1987 and 2016, while total rainfall exhibited a slight decreasing trend. In addition to groundwater extraction, ﬂuctuations in monthly rainfall, soil moisture, evapotranspiration and relative humidity support the groundwater thickness reduction of GRACE datasets. The use of desalinated water and wastewater reuse in the agriculture sector may reduce the pressure on groundwater resources. Optimization, adaptation and mitigation in the EWF nexus will further improve the sustainability of the EWF resources.


Introduction
Groundwater is the second most abundant freshwater supply after the cryosphere, and it is a critical resource for human existence and ecological function. Over a third of the world's freshwater comes from groundwater sources [1]. Demands for freshwater are increasing due to population expansion, socio-economic development, and land-use changes in dry and semiarid regions with insufficient surface water availability [2]. In the last decades, groundwater has been depleted significantly in dry and semiarid regions raising concerns for groundwater sustainability [3]. In addition to extraction, climate change is inflicting further pressure on groundwater resources [4]. Climate change may affect the natural balance of recharge/discharge, alter demand and supply in water-dependent sectors and may put further pressure on groundwater systems through land use management practices [5].
In addition, climate change can also reduce the infiltration capacity of soil and enhance the rate of evapotranspiration, causing a decline in groundwater recharge over time [3]. An increase in mean temperature also increases evaporation, which affects the quantity of water available to recharge groundwater [6]. Furthermore, groundwater is also affected by a decrease in total rainfall. Rainfall and temperature are the main parameters of the Water 2021, 13, x FOR PEER REVIEW 3 of 18 desert environment and erratic annual rainfall. Qatar is often hit by Shamals, which are predominantly north-westerly dust storms that occur during the summer. .

Datasets
According to the Qatar meteorological department, seven weather stations provide daily data on different climatic parameters such as temperature, rainfall, relative humidity, wind and solar radiation. However, due to less availability of long-term data, only six weather station data were used for long-term analysis (Table 1). Daily mean temperature and daily total rainfall data were acquired from 1987 to 2016. Although many of the weather stations have been installed recently, those weather stations were selected with data collected for over 30 years. ArcMap software version 10.4 was used to plot weather stations on a map using the respective longitude and latitude of each station. Raw data were analysed in Excel for outliers and missing data. In addition, annual groundwater extraction data from 1990-2019 was acquired from the Planning and Statistical Authority. Table 1 below provides a detailed summary of the weather stations.

Datasets
According to the Qatar meteorological department, seven weather stations provide daily data on different climatic parameters such as temperature, rainfall, relative humidity, wind and solar radiation. However, due to less availability of long-term data, only six weather station data were used for long-term analysis (Table 1). Daily mean temperature and daily total rainfall data were acquired from 1987 to 2016. Although many of the weather stations have been installed recently, those weather stations were selected with data collected for over 30 years. ArcMap software version 10.4 was used to plot weather stations on a map using the respective longitude and latitude of each station. Raw data were analysed in Excel for outliers and missing data. In addition, annual groundwater extraction data from 1990-2019 was acquired from the Planning and Statistical Authority. Table 1 below provides a detailed summary of the weather stations.

GRACE and GRACE-FO
The Gravity Recovery and Climate Experiment (GRACE) satellite mission was cooperatively established and operated by the National Aeronautics and Space Administration (NASA) and the German Aerospace Centre (DLR). Gravity Recovery and Climate Experiment (GRACE) satellites operated from 2002-2017, and their replacement pair, the GRACE Follow-On, were started in 2018. The operation comprises two twin satellites in identical circular polar orbits at 500 km of altitude with a distance of 220 km between the two satellites. They measure differences in the earth's gravitational field, which cause small fluctuation in the distance between both satellites as a result of different masses distributed on earth. These fluctuations are calculated by employing a GPS and a microwave sensing system giving spatiotemporal gravitational maps. Changes in the planet's mass, shown as alterations in the gravitational field, are primarily triggered by fluctuations in surface and groundwater [25]. GRACE-based monthly gravity product produced by Jet Propulsion Laboratory (JPL) Monthly Mass Grids-Global mascons (JPL RL06_v02) were used in the present study. This monthly global mascon dataset contains gridded monthly global water storage information developed from GRACE and GRACE-FO. This dataset is available at https://podaac.jpl.nasa.gov/dataset/TELLUS_GRAC-GRFO_MASCON_CRI_GRID_ RL06_V2 (accessed 27 March 2021) [26] in a single data file in netCDF format. This dataset has a spatial resolution of 0.5-degrees and a temporal resolution of one month. The current version of the data uses a Coastal Resolution Improvement (CRI) filter that decreases signal leakage errors across coastlines [27]. The water storage/height anomalies are given in equivalent water thickness units (cm) (NASA-JPL). Further details on the Mascon solution, involving the mathematical derivation, validation and realization of geophysical constraints can be found in [28].

GLDAS
The Global Land Data Assimilation System (GLDAS) is a program to utilise satelliteand ground-based observational data using sophisticated land surface modelling and data integration methods to produce the best possible land surface states and fluxes. GLDAS has been established by NASA and the National Oceanic and Atmospheric Administration (NOAA) and utilizes in situ and ex situ observations of the land surface interactions by the use of various models. The model used in this study was the monthly GLDAS NOAH Land Surface Model L4 monthly 0.25 × 0.25 degree V2.1, 2000-2020 [29]. Monthly soil moisture and evapotranspiration data were acquired from the GLDAS model using the Bash command line. GLDAS data have been validated against the observation data. A significant positive correlation exists between the observed total monthly precipitation data (r = 0.79) and GLDAS total monthly precipitation dataset. Similarly, a strong positive relationship with monthly mean temperature (r = 0.95) occurs in comparison to the GLDAS dataset.

Digital Elevation Model
A Shuttle Radar Topography Mission SRTM 30m digital elevation model was used for the delineation of the watershed. Thirty-meter (1 arc-second) resolution SRTM DEM data were downloaded from https://earthexplorer.usgs.gov/ (accessed on 21 September 2020) [30]. Downloaded tiles of the SRTM DEM were mosaicked and the study area was extracted using ArcMap 10.4.

Mann-Kendall Trend Analysis
The Mann-Kendall (MK) test [31,32] was used to identify trends in the time series of climatic and hydrological data. The MK statistic indicated whether a trend was present in the time series data and whether the trend was positive or negative. The MK test was also used to identify stations where variations were significant or of large magnitude [33]. The MK test is a nonparametric test, which means that it does not depend on the knowledge of the distribution of the sampled population. The null hypothesis (H 0 ) was that the data had no trend. The alternative hypothesis (Ha) was that the data follow a monotonic trend. A monotonic trend means that the trend is consistently decreasing or increasing over time. Kendall's tau (t) is a measure of the strength of a trend. A trend is reported as statistically significant if the p-value is smaller than the significance level alpha (=0.05). Similarly, a null hypothesis is rejected or accepted if the p-value greater or smaller than the significance level alpha = 0.05.
The S statistic used for the Mann Kendall test is: where X j and X k are the adjacent data values, S is the sum of positive or negative signs, n is the number of observations and variance can be calculated as in [31]: Moreover, the extent of a time series trend was assessed by a nonparametric test know as Sen's slope [34]. The slope of the trend is calculated by where β is the Sen's slope estimate. A β value greater than zero means an upward trend and β < zero indicates a downward trend.

Mann-Whitney-Pettitt Change Point Analysis
Pettitt's test is a nonparametric test proposed by Pettitt in 1979 [35]. Pettitt's test is used to identify the time at which a shift occurs. It is commonly applied to detect a single change-point in hydrological and climatic data series [36]. This test not only detects occurrence time but also the significance of the change point. The null hypothesis (H 0 ) is that the T variables follow one or more distributions that have the same location parameter (no change). The alternative hypothesis (H a ) is that a change point exists in the data. The nonparametric statistic is defined as: Xi and Xj are the adjacent data values and n is the number of observations. The changepoint of the series is situated at K T , if that the statistic is significant.
The significance probability of K T is approximated for p ≤ 0.05 as:

State-Wide Annual and Seasonal Climatic Trends
Mean temperature data from six weather stations from 1987-2016 were averaged to establish a state-wide representative mean annual temperature. Long-term temperature data indicated that the mean annual temperature in Qatar is 27.42 • C with significant variation between diurnal and seasonal temperature. Mean annual data were analysed using the Mann Kendall trend test with a confidence level of 95%. Mann Kendal trend test analysis indicated a significant (p < 0.005) increase in the state-wide mean annual temperature over the period of observation ( Figure 2). The Sen's slope estimated that mean temperature had increased by 1.02 • C over the last 30 years with a 0.034 • C per year increment. These warming trends are in agreement with previous studies in Qatar [21], Saudi Arabia [8] Iran [37], and Kuwait [9]. The Mann Pettit test indicated a significant change in the annual temperature data in 1997. According to The Oceanic Niño Index

State-Wide Annual and Seasonal Climatic Trends
Mean temperature data from six weather stations from 1987-2016 were averaged to establish a state-wide representative mean annual temperature. Long-term temperature data indicated that the mean annual temperature in Qatar is 27.42 °C with significant variation between diurnal and seasonal temperature. Mean annual data were analysed using the Mann Kendall trend test with a confidence level of 95%. Mann Kendal trend test analysis indicated a significant (p < 0.005) increase in the state-wide mean annual temperature over the period of observation ( Figure 2). The Sen's slope estimated that mean temperature had increased by 1.02 °C over the last 30 years with a 0.034 °C per year increment. These warming trends are in agreement with previous studies in Qatar [21], Saudi Arabia [8] Iran [37], and Kuwait [9]. The Mann Pettit test indicated a significant change in the annual temperature data in 1997. According to The Oceanic Niño Index (ONI) by the Na-  Rainfall data from six weather stations were also analysed for long-term trends using the Mann Kendall trend test with 95% confidence. State-wide long-term total annual rainfall data indicated rainfall is decreasing over the state of Qatar ( Figure 3). It was determined that rainfall during the period of observation decreased by −0.42 mm per year. However, this decreasing trend was not statistically significant (p < 0.05). These drying trends are consistent with the subtropical drier projection [38]. More recently, similar dry episodes were observed in the MENA region [39,40] and particularly in Saudi Arabia [41]. Rainfall data from six weather stations were also analysed for long-term trends using the Mann Kendall trend test with 95% confidence. State-wide long-term total annual rainfall data indicated rainfall is decreasing over the state of Qatar ( Figure 3). It was determined that rainfall during the period of observation decreased by −0.42 mm per year. However, this decreasing trend was not statistically significant (p < 0.05). These drying trends are consistent with the subtropical drier projection [38]. More recently, similar dry episodes were observed in the MENA region [39,40] and particularly in Saudi Arabia [41].

State-Wide Seasonal Climatic Trends
Mean temperature and total monthly rainfall data from six weather stations were divided into seasonal subsets; for example, winter (December, January, February = DJF), summer (

State-Wide Seasonal Climatic Trends
Mean temperature and total monthly rainfall data from six weather stations were divided into seasonal subsets; for example, winter (December, January, February = DJF), summer (   Similarly, total monthly rainfall was analysed for seasonal trends using the Mann Kendall trend test with 95% confidence. Annual rainfall in the State of Qatar is 70 mm based on the average data from 1987-2016. Most of the rainfall occurs during DJF and MAM followed by SON and JJA. JJA is usually dry with zero rainfall; however, in the recent decade, rainfall during JJA was increasing. Based on the Mann Kendall trend test, a statistically significant increment in the total monthly rainfall data was observed during JJA (τ = 0.453, S = 0, p < 0.0001) and SON (τ = 0.314, S = 0, p < 0.0001) ( Figure 5). In contrast a nonsignificant downward trend in the total monthly rainfall was found during DJF (τ = −0.109, S = −0.045, p > 0.05), while a nonsignificant increasing trend was observed for MAM (τ = 0.047, S = 0, p > 0.05). Similarly, total monthly rainfall was analysed for seasonal trends using the Mann Kendall trend test with 95% confidence. Annual rainfall in the State of Qatar is 70 mm based on the average data from 1987-2016. Most of the rainfall occurs during DJF and MAM followed by SON and JJA. JJA is usually dry with zero rainfall; however, in the recent decade, rainfall during JJA was increasing. Based on the Mann Kendall trend test, a statistically significant increment in the total monthly rainfall data was observed during JJA (τ = 0.453, S = 0, p < 0.0001) and SON (τ = 0.314, S = 0, p < 0.0001) ( Figure 5). In contrast a nonsignificant downward trend in the total monthly rainfall was found during DJF Similarly, total monthly rainfall was analysed for seasonal trends using the Mann Kendall trend test with 95% confidence. Annual rainfall in the State of Qatar is 70 mm based on the average data from 1987-2016. Most of the rainfall occurs during DJF and MAM followed by SON and JJA. JJA is usually dry with zero rainfall; however, in the recent decade, rainfall during JJA was increasing. Based on the Mann Kendall trend test, a statistically significant increment in the total monthly rainfall data was observed during JJA (τ = 0.453, S = 0, p < 0.0001) and SON (τ = 0.314, S = 0, p < 0.0001) ( Figure 5). In contrast a nonsignificant downward trend in the total monthly rainfall was found during DJF (τ = −0.109, S = −0.045, p > 0.05), while a nonsignificant increasing trend was observed for MAM (τ = 0.047, S = 0, p > 0.05).

Station-Wise Annual Trends
Rainfall and temperature data of six weather stations were analysed for long-term trends and spatial variability. Long-term data analysis indicated that four (66%) weather stations exhibited statistically significant increasing trends, while two weather stations namely, Dukhan and Mukanis, demonstrated increasing but nonsignificant upward trends in mean annual temperature ( Figure 6). It was determined that weather stations situated at the eastern southern coastal areas experienced significant increment in mean annual temperature, while the temperature at the western coastal areas such as Dukhan is increasing; however, the trend is not statistically significant.
In contrast to the increasing mean annual temperature, the annual total rainfall is decreasing at all weather stations, though the only significant decrease in the total annual rainfall was found at Al-khor weather station ( Figure 6).

Seasonal Station Wise Spatial Variability
Data from six weather stations were analysed for seasonal trends in winter = DJF, spring = MAM, summer = JJA and autumn = SON. All weather stations indicated an increased temperature in the seasonal mean monthly temperature. However, 50% of the weather stations, namely Doha, Al-ruwais and Mukanis, indicated a significant increase during DJF. The number of weather stations with a significant increase in the monthly temperature further increased to 66.67% during MAM, JJA and SON (Figure 7). Most of these weather stations are installed in the eastern and southern coastal areas. In contrast, both the Al-Khor station installed at the eastern coastal line, and the Dukhan weather trends and spatial variability. Long-term data analysis indicated that four (66%) weather stations exhibited statistically significant increasing trends, while two weather stations namely, Dukhan and Mukanis, demonstrated increasing but nonsignificant upward trends in mean annual temperature ( Figure 6). It was determined that weather stations situated at the eastern southern coastal areas experienced significant increment in mean annual temperature, while the temperature at the western coastal areas such as Dukhan is increasing; however, the trend is not statistically significant.
In contrast to the increasing mean annual temperature, the annual total rainfall is decreasing at all weather stations, though the only significant decrease in the total annual rainfall was found at Al-khor weather station ( Figure 6).

Seasonal Station Wise Spatial Variability
Data from six weather stations were analysed for seasonal trends in winter = DJF, spring = MAM, summer = JJA and autumn = SON. All weather stations indicated an increased temperature in the seasonal mean monthly temperature. However, 50% of the weather stations, namely Doha, Al-ruwais and Mukanis, indicated a significant increase during DJF. The number of weather stations with a significant increase in the monthly temperature further increased to 66.67% during MAM, JJA and SON (Figure 7). Most of these weather stations are installed in the eastern and southern coastal areas. In contrast, both the Al-Khor station installed at the eastern coastal line, and the Dukhan weather station situated at the western coastal line, exhibited nonsignificant increasing trends. This could be attributed to the geographic, climatic and anthropogenic changes in those areas. All six weather stations were also investigated for seasonal rainfall variation. The Mann Kendall trend test indicated great variability in seasonal temperature in DJF and MAM compared to JJA and SON. In DJF, none of the weather stations indicated a statistically significant increase in the monthly rainfall; however, a general nonsignificant decreasing trend was observed at five stations. As the temperature increases with the early warming in later February, monthly rainfall tends to increase with a significant increase at Dukhan weather station. In JJA, 80% of stations demonstrated a statistically significant increase in the total monthly rainfall. Similarly, 83% of stations indicated a significant upward tendency in the total monthly rainfall during SON. Though monthly rainfall increased significantly in JJA and SON, a significant amount of rainfall was received in DJF. Thus, decreasing trends in DJF is of importance compared to the observed significant increases in MAM, JJA and SON (Figure 8). Water 2021, 13, x FOR PEER REVIEW 10 of 18 All six weather stations were also investigated for seasonal rainfall variation. The Mann Kendall trend test indicated great variability in seasonal temperature in DJF and MAM compared to JJA and SON. In DJF, none of the weather stations indicated a statistically significant increase in the monthly rainfall; however, a general nonsignificant decreasing trend was observed at five stations. As the temperature increases with the early warming in later February, monthly rainfall tends to increase with a significant increase at Dukhan weather station. In JJA, 80% of stations demonstrated a statistically significant increase in the total monthly rainfall. Similarly, 83% of stations indicated a significant upward tendency in the total monthly rainfall during SON. Though monthly rainfall

Groundwater Assessment
Groundwater is one of the most important water resources of Qatar. Groundwater extraction and desalination of seawater fulfil national water needs. Based on the longterm average of GRACE-FO data analysis there is a considerable interannual variation (± 2 cm) in the groundwater depletion ( Figure 9). Groundwater depletion reaches the highest in June as increased temperature increases the rate of evapotranspiration and the overall

Groundwater Assessment
Groundwater is one of the most important water resources of Qatar. Groundwater extraction and desalination of seawater fulfil national water needs. Based on the long-term average of GRACE-FO data analysis there is a considerable interannual variation (±2 cm) in the groundwater depletion ( Figure 9). Groundwater depletion reaches the highest in June as increased temperature increases the rate of evapotranspiration and the overall water requirements for agriculture, horticulture and other recreational activities increases. Rainfall during winter is one of the major sources of the aquifer. Around 71.6 million cubic meters are contributed by annual rainfall to the groundwater recharge while about 2.2 million cubic meters are added by inflow from Saudi Arabia [19]. Temperature and rainfall play an important role in the variation of soil moisture, evapotranspiration, relative humidity, and groundwater recharge. Most of the rainfall occurs during DJF and that is why the rate of groundwater depletion is lower during these months. In contrast, groundwater depletion is highest in June due to zero rainfall.
water requirements for agriculture, horticulture and other recreational activities increases. Rainfall during winter is one of the major sources of the aquifer. Around 71.6 million cubic meters are contributed by annual rainfall to the groundwater recharge while about 2.2 million cubic meters are added by inflow from Saudi Arabia [19]. Temperature and rainfall play an important role in the variation of soil moisture, evapotranspiration, relative humidity, and groundwater recharge. Most of the rainfall occurs during DJF and that is why the rate of groundwater depletion is lower during these months. In contrast, groundwater depletion is highest in June due to zero rainfall. GRACE-FO monthly water equivalent thickness data analysis indicated a significant depletion in the groundwater from 2002-2020. The GRACE-FO JPL data showed a statistically significant (t = −0.260, p < 0.05) decrease in the groundwater thickness. Sen's slope estimated a depletion trend of 0.24 ± 0.20 cm per year ( Figure 10). This reduction in the groundwater thickness is in agreement with the previous study conducted in the UAE where an annual decline of 0.5 cm per year was reported [22]. A recent study in Saudi Arabia found an annual decreasing trend of 1 cm per year in the groundwater thickness using GRACE data [42]. Similarly, a groundwater thickness reduction of 0.76 cm per year was found in Bakhtegan catchment of Iran [24]. Groundwater is mainly used for agriculture purposes, as the majority of the domestic and industrial water demand is fulfilled by desalination. According to the Planning and Statistical Authority, agriculture is the biggest consumer of freshwater resources, with an annual use of 309.97 million cubic meters in 2019 [43]. About 250 million cubic meters of groundwater were extracted by the agriculture sector in 2019. Long-term groundwater extraction data indicated a gradual increase in the water consumption pattern by the agriculture sector ( Figure 10). Groundwater extraction by the agriculture sector could be one of the reasons for the observed significant groundwater thickness loss observed by GRACE-FO. Moreover, correlation analysis indicated a significant direct correlation only between groundwater and evapotranspiration (r = 0.196) and groundwater and relative humidity (r = 0.287). This indicates that rainfall plays a significant role in groundwater recharge. Relative humidity increases with an increase in rainfall as rainwater is evaporated from the soil into the atmosphere.  Figure 10). This reduction in the groundwater thickness is in agreement with the previous study conducted in the UAE where an annual decline of 0.5 cm per year was reported [22]. A recent study in Saudi Arabia found an annual decreasing trend of 1 cm per year in the groundwater thickness using GRACE data [42]. Similarly, a groundwater thickness reduction of 0.76 cm per year was found in Bakhtegan catchment of Iran [24]. Groundwater is mainly used for agriculture purposes, as the majority of the domestic and industrial water demand is fulfilled by desalination. According to the Planning and Statistical Authority, agriculture is the biggest consumer of freshwater resources, with an annual use of 309.97 million cubic meters in 2019 [43]. About 250 million cubic meters of groundwater were extracted by the agriculture sector in 2019. Long-term groundwater extraction data indicated a gradual increase in the water consumption pattern by the agriculture sector ( Figure 10). Groundwater extraction by the agriculture sector could be one of the reasons for the observed significant groundwater thickness loss observed by GRACE-FO. Moreover, correlation analysis indicated a significant direct correlation only between groundwater and evapotranspiration (r = 0.196) and groundwater and relative humidity (r = 0.287). This indicates that rainfall plays a significant role in groundwater recharge. Relative humidity increases with an increase in rainfall as rainwater is evaporated from the soil into the atmosphere. Rainfall is one of the main sources of groundwater recharge in the State of Qatar. As mentioned earlier ( Figure 3) the long-term rainfall trend exhibited a slight negative trend from 1987-2016. However, state-wide total monthly rainfall data exhibited a significant positive trend from 2000 onwards, which was primarily due to anomalous extreme rainfall, especially in 201,8 which resulted in flash floods. Average total monthly rainfall data from the six weather stations indicated a positive trend and with an increase of 0.072 mm per year. Rainfall and temperature play a great role in the hydrological cycle. Rainfall in the State of Qatar is erratic and falls mostly during the winter months. Correlation analysis indicates that a significant negative correlation (r = −0.421) exists between rainfall and temperature. As the rain falls it causes a cooling effect and thus a reduction in the mean temperature. This cooling effect is a result of soil moisture being evaporated from the land into the atmosphere. A significant inverse correlation of mean temperature with soil moisture and relative humidity further explains the phenomenon that as the temperature increases it reduces soil moisture and relative humidity.
Data analysis indicated a slight loss in soil moisture with an annual decrease of −0.012 mm per year ( Figure 11). The soil moisture-holding capacity depends on the type of soil in addition to other environmental parameters. The major soil type of the state of Qatar is sandy loam with a high salt content. Sandy soil tends to store a greater proportion of water at depth ( > ~50 cm) than finer textured soils [44]. This variation in the soil moisture along with the soil depth shows that most of the precipitated water is stored in the lower layers of the soil. A strong negative correlation (r = −851) exists between temperature and soil moisture. As the temperature increases more amount of soil moisture evaporates into the atmosphere causing a reduction in the soil moisture.
Evapotranspiration is the combined water loss by evaporation from the soil and the transpiration of vegetation. A shift in climate and weather parameters causes a significant change in evapotranspiration [45]. Changes in evaporation affect the soil moisture content and groundwater recharge. In addition, rising temperature increases evaporative demand over land, which limits the amount of water available to replenish groundwater [6]. Evapotranspiration data analysis indicated a significant increment (t = 0.272, p < 0.05) in evapotranspiration with an increase of 0.096 mm per year. The state of Qatar has sandy soil characterised by lower water-holding capacity. In addition, the significant increasing trend in mean temperature could be one of the main reasons for the observed decreasing Rainfall is one of the main sources of groundwater recharge in the State of Qatar. As mentioned earlier ( Figure 3) the long-term rainfall trend exhibited a slight negative trend from 1987-2016. However, state-wide total monthly rainfall data exhibited a significant positive trend from 2000 onwards, which was primarily due to anomalous extreme rainfall, especially in 201,8 which resulted in flash floods. Average total monthly rainfall data from the six weather stations indicated a positive trend and with an increase of 0.072 mm per year. Rainfall and temperature play a great role in the hydrological cycle. Rainfall in the State of Qatar is erratic and falls mostly during the winter months. Correlation analysis indicates that a significant negative correlation (r = −0.421) exists between rainfall and temperature. As the rain falls it causes a cooling effect and thus a reduction in the mean temperature. This cooling effect is a result of soil moisture being evaporated from the land into the atmosphere. A significant inverse correlation of mean temperature with soil moisture and relative humidity further explains the phenomenon that as the temperature increases it reduces soil moisture and relative humidity.
Data analysis indicated a slight loss in soil moisture with an annual decrease of −0.012 mm per year ( Figure 11). The soil moisture-holding capacity depends on the type of soil in addition to other environmental parameters. The major soil type of the state of Qatar is sandy loam with a high salt content. Sandy soil tends to store a greater proportion of water at depth (>~50 cm) than finer textured soils [44]. This variation in the soil moisture along with the soil depth shows that most of the precipitated water is stored in the lower layers of the soil. A strong negative correlation (r = −851) exists between temperature and soil moisture. As the temperature increases more amount of soil moisture evaporates into the atmosphere causing a reduction in the soil moisture.
Evapotranspiration is the combined water loss by evaporation from the soil and the transpiration of vegetation. A shift in climate and weather parameters causes a significant change in evapotranspiration [45]. Changes in evaporation affect the soil moisture content and groundwater recharge. In addition, rising temperature increases evaporative demand over land, which limits the amount of water available to replenish groundwater [6]. Evapotranspiration data analysis indicated a significant increment (t = 0.272, p < 0.05) in evapotranspiration with an increase of 0.096 mm per year. The state of Qatar has sandy soil characterised by lower water-holding capacity. In addition, the significant increasing trend in mean temperature could be one of the main reasons for the observed decreasing trend in topsoil (0-10 cm) moisture. Furthermore, data analysis of relative humidity indicated a significant (p < 0.05) reduction over the last three decades with an annual decrease rate of −0.216 % (Figure 11). A significant inverse correlation (r = −763) was found between relative humidity and temperature. Relative humidity has an inverse relationship with air temperature as the temperature increases the water holding capacity of the air increases and the relative humidity decreases. This significant decrease in the relative humidity could be one of the reasons for the observed significant increment in mean temperature. Similar decreasing trends in relative humidity with an increase in the mean temperature were reported in Iran [46].
Water 2021, 13, x FOR PEER REVIEW 14 of 18 trend in topsoil (0-10 cm) moisture. Furthermore, data analysis of relative humidity indicated a significant (p < 0.05) reduction over the last three decades with an annual decrease rate of −0.216 % (Figure 11). A significant inverse correlation (r = −763) was found between relative humidity and temperature. Relative humidity has an inverse relationship with air temperature as the temperature increases the water holding capacity of the air increases and the relative humidity decreases. This significant decrease in the relative humidity could be one of the reasons for the observed significant increment in mean temperature. Similar decreasing trends in relative humidity with an increase in the mean temperature were reported in Iran [46].

Implications for the Energy Water Food Nexus in Qatar
The interdependent and interlinked nature of the EWF nexus is strong in the State of Qatar due to limited renewable freshwater options, erratic annual rainfall, and lack of arable land due to harsh weather conditions. The current freshwater production system is heavily dependent on the continuous supply of energy for desalination, pumping of groundwater, transportation, storage and wastewater treatment. While a considerable amount of water is used for cooling purposes during energy production, desalination and food processing. Agricultural practices consume 33.8% of the total water produced, which is mainly extracted from groundwater [44]. The energy sector plays a bonding and compensatory role in the EWF nexus. For instance, Qatar fossil-based energy-intensive desalination technologies compensate for the nonavailability of surface water and brackish groundwater through desalination to fulfil domestic, industrial, and agricultural water consumption needs.
Groundwater depletion as a consequence of direct anthropogenic actions such as over-extraction of groundwater, and indirect depletion by climate change, will enhance intersectoral competition in the EWF nexus. Groundwater depletion affects water supply to the food sector in two ways. First, it reduces the amount of water to be extracted, and second it reduces groundwater quality through saltwater intrusion of coastal shallow

Implications for the Energy Water Food Nexus in Qatar
The interdependent and interlinked nature of the EWF nexus is strong in the State of Qatar due to limited renewable freshwater options, erratic annual rainfall, and lack of arable land due to harsh weather conditions. The current freshwater production system is heavily dependent on the continuous supply of energy for desalination, pumping of groundwater, transportation, storage and wastewater treatment. While a considerable amount of water is used for cooling purposes during energy production, desalination and food processing. Agricultural practices consume 33.8% of the total water produced, which is mainly extracted from groundwater [44]. The energy sector plays a bonding and compensatory role in the EWF nexus. For instance, Qatar fossil-based energy-intensive desalination technologies compensate for the nonavailability of surface water and brackish groundwater through desalination to fulfil domestic, industrial, and agricultural water consumption needs.
Groundwater depletion as a consequence of direct anthropogenic actions such as over-extraction of groundwater, and indirect depletion by climate change, will enhance intersectoral competition in the EWF nexus. Groundwater depletion affects water supply to the food sector in two ways. First, it reduces the amount of water to be extracted, and second it reduces groundwater quality through saltwater intrusion of coastal shallow aquifers. Groundwater uptake, especially in the dry seasons, encourages saltwater intrusion to fill the gap created by the over-extraction of the groundwater. Annual groundwater extraction is currently more than 250 million cubic meters per year, with a deficit of 98 million cubic meters. If the present groundwater depletion trends persist, along with the reduction in annual total rainfall, the groundwater deficit is expected to increase in the coming years. This reduction will increase the amount of energy consumed to reach deeper aquifers and to use more energy on groundwater desalination as the water quality deteriorates due to saltwater intrusion.
Rainfall is the prime source of groundwater recharge which is responsible for adding 71.6 million cubic meters of water per year to the groundwater. The observed long-term declining trends are significant for the sustainability of groundwater resources. Reduction in the total annual rainfall is likely to affect groundwater recharge, affecting demand and supply of water and energy in the food sector. Dry weather conditions in the future will put further pressure on groundwater to produce more water for outdoor and indoor irrigation, either through extraction or seawater desalination. Excessive groundwater extraction and high saline water is often associated with increased energy consumption for pumping and saltwater treatment costs. Thus, reduction in rainfall not only affects the groundwater resources but also affects food production and energy generation systems. In addition, less frequent rainfall will affect the efficiency of solar power, as rainfall washes off dust from solar panels. Dust is considered one of the main limitations of renewable solar energy in dry and dusty regions like the Middle East and Africa.
The observed significant increase in the mean temperature will affect the rate of evapotranspiration and soil moisture, which are vital parameters for crop growth. Variation in the mean temperature affects plants water requirements, as the temperature is directly related to evaporation and soil moisture loss. In dry conditions, plants require more water as soil moisture is evaporated into the atmosphere. This not only affects the soil moisture, which is vital for plant growth, but also groundwater recharge by reducing soil infiltration. In addition, a considerable amount of water is used for cooling during energy production. According to International Energy Agency, about 10% of the global freshwater withdrawals are for power generation [47]. Therefore, water demand for cooling purposes will increase in power plants and desalination plants if the increasing trends in mean temperature prevail in the future. Furthermore, power plants, solar cells and desalination power plant efficiency are likely to be affected by changes in mean temperature ( Figure 12). Though the state of Qatar imports the majority of its food, fluctuations in the mean temperature is expected to affect local crop rotation, agricultural yield, and local meat and fish farming.

Conclusions
In this study, groundwater variation based on ground data and satellite (GRACE) data was measured and compared to climatic parameters from weather stations and model data. The results showed good agreement between satellite-based groundwater levels and climatic parameters such as rainfall, temperature, soil moisture loss, evapotranspiration, and relative humidity. During the period of study, the groundwater level

Conclusions
In this study, groundwater variation based on ground data and satellite (GRACE) data was measured and compared to climatic parameters from weather stations and model data. The results showed good agreement between satellite-based groundwater levels and climatic parameters such as rainfall, temperature, soil moisture loss, evapotranspiration, and relative humidity. During the period of study, the groundwater level dropped by an average of 0.24 ± 0.20 cm per year. This was mainly the result of an increase in groundwater exploitation, particularly for irrigation as groundwater is mainly used by agricultural activities. The imbalance between aquifer recharge and groundwater extraction is expected to increase in the future as water demand will increase with the increases in population, economic growth and agricultural development. Desalination powered by renewable energy such as wind and solar can play a great role in compensating for the imbalance caused by groundwater extraction for agricultural practices. In addition, the use of wastewater for irrigation and industrial use will further reduce pressure on the groundwater resources of the State of Qatar.
With increasing population, urbanisation and economic and technological development, there is continuous pressure on energy, water and food resources. In addition, climate change, which can cause variations in temperature and total annual rainfall, is one of the greatest risks to energy, water and food security. A decline in the groundwater is of great concern for the sustainability of water-dependent sectors such as agriculture and aquaculture. If groundwater depletion persists it will cause saltwater intrusion and land subsidence. In addition, an observed increase in the mean temperature will further affect the evapotranspiration rate and soil moisture loss in outdoor farming, which will likely increase water supply-demand for irrigation. Moreover, as water demand in agriculture and recreational activities increases due to increased evapotranspiration, energy demand will also increase to extract more groundwater and to treat wastewater. Similarly, water demand will increase in power plants and desalination plants for cooling if the increasing trend in mean temperature prevails in the future. Furthermore, power plants, solar cells and desalination power plant efficiency are likely to be affected by changes in temperature and rainfall. Changes in the mean temperature and total rainfall will likely affect local crop rotation and agricultural yield. On the other hand, a reduction in the annual rainfall will affect natural groundwater recharge making groundwater more saline. As a result, energy will be needed to treat groundwater before making it available for irrigation and recreational purposes Thus, informed smart decision-making needs to be introduced in the planning, policy and implementation stages of future projects dealing with the EWF nexus. Though the present study was conducted using long-term 30 year ground-based data and GRACE data from 18 years, this can be further improved by increasing the temporal resolution of the data as more data is made available. There is also a need to investigate and project variations in climatic parameters such as rainfall and temperature under different scenarios, and the possible impacts on groundwater resources using high-resolution regional climate models.