Spatio-Temporal Trend Analysis of Groundwater Levels in Sharjah , UAE

United Arab Emirates (UAE) is a water stressed country due to limited rainfall. Moreover, rapid population growth, increase in agricultural and industrial activities are causing additional stress to the water resources in UAE. Groundwater is the most important fresh water source in UAE, and sustainable and efficient management of groundwater is vital for the country. Trend analysis of groundwater levels is an essential component of groundwater management, since it provides significant information about the direction and characteristics of the groundwater level trend, allowable discharge limit and the cause of groundwater decline. In this study, monthly groundwater level trends from eleven well fields were investigated over the period of 15 years in Sharjah, UAE. Groundwater spatiotemporal data analysis tool (GWSDAT) was used for the spatial and temporal trend analysis for groundwater levels. Mann-Kendall non-parametric test was used for the temporal trend analysis. Statistically significant decreasing trends were identified almost in all well fields in Sharjah. In parallel to decreasing groundwater levels, the groundwater production has fallen substantially (more than 90% decrease in some well fields over 15 years). It was found that the well fields close to the coastline experienced significant decrease in groundwater levels, which induces sea water intrusion and salinization of groundwater.


I. INTRODUCTION
Groundwater is an important fresh water source that can meet the domestic, agricultural and industrial water demand. Statistical assessment of groundwater quality and levels is an important task in sustainable management of groundwater, especially in the arid and semi-arid regions, where there is limited surface fresh water resources. The water demand in these regions are met mainly by groundwater and non-conventional water resources such as desalinated water.
Trend analysis of groundwater levels is an essential component of groundwater management, since it provides significant information about the direction and characteristics of the groundwater level trend, allowable discharge limit and the cause of groundwater decline ( [1]; [2]). Several studies (e.g. [3]; [4]) showed that trend analysis is very powerful in planning management strategies for development and utilization of groundwater resources. As an example of groundwater level trend analysis studies, [5] studied Manuscript  groundwater level trends by using non-parametric tests in Ardabil plain of Iran, and they found significant decreasing trends in groundwater levels in the study area. [6] used Mann-Kendall (MK) and Sen's slope tests to detect groundwater level trends in Bangladesh, decreasing trends in the maximum and minimum depths of groundwater levels (during 1991-2010) were reported. [7] investigated spatio-temporal trends in groundwater in Lake Umria region. [8] used regional Kendall method to analyze groundwater levels and their annual decline rates in Beijing, Tianjin and Hebei in China. This study showed a continuing declining regional trend in groundwater levels. [9] reported no significant trend in groundwater levels (with few exceptions) in Raipur city at 95% significance level by analyzing monthly groundwater level data using MK test and Sen's slope estimator.
United Arab Emirates (UAE) is a water stressed country due to limited rainfall. Moreover, rapid population growth, increase in agricultural and industrial activities are causing additional stress to the water resources in UAE [10]. There is very limited surface fresh water available in UAE due to rare rainfall events (about 110 mm/year) [11]. Groundwater is used to meet 70% of total water demand in UAE, and agriculture sector is the highest groundwater consumer with 83% of total groundwater use [12]. The over-extraction of groundwater and lack of recharge led to a serious decline in the groundwater level and caused sea water intrusion in UAE [13], [14]. In this study, spatio-temporal trends in monthly groundwater levels were investigated in Sharjah, UAE by analyzing data from eleven well fields. This study is expected to provide significant contribution to sustainable groundwater management in Sharjah, and to assist groundwater policymakers in the region for future planning.

II. STUDY AREA AND DATA
The study area, Emirate of Sharjah, is located in UAE with central coordinates of 25.3° N and 55.5° E approximately. Sharjah occupies around 2,590 km² . Fig. 1 shows location of Sharjah along with location of well fields used in this study. The climate in Sharjah is considered hot, dry and humid. Rainy season in Sharjah is from November to March. The available capacity of the groundwater wells is around 21.50 million gallons per day, and daily average water production from groundwater wells is around 15.46 million gallons. 70% of the total groundwater consumption is for irrigation purposes in Sharjah [15].
In this study, monthly groundwater level data from eleven well fields (i.e., Al Sadiyah, Aweed, Bedai, Filli, Ghuraifa, Hamdah, Madam, Mahafis, Sajaa, Seih Aqareb and Seih Hermal) are considered. As seen in Fig. 1, the well fields are approximately evenly distributed between the west and the south parts of Sharjah. The data period ranges between 2004 and 2018 showing variation for different well fields as shown in Table I.

III. METHODOLOGY
Groundwater spatiotemporal data analysis tool (GWSDAT) was used for the spatial and temporal trend analysis for groundwater levels in this study. Since the groundwater level data is mostly non-normally distributed, MK non-parametric test was used for the temporal trend analysis, whereas penalized splines (P-Splines) smoothing was adopted for spatial trend analysis.

A. Mann-Kendall Test
The MK is a non-parametric rank-based test and used in several hydro-meteorological trend analysis studies (e.g. [16]- [24]). The z-statistics for this test can be calculated by: The S value can be computed by: where are the sequential data values, n is the number of data The Var(S) can be calculated using: Positive test statistics indicate increasing trend, whereas negative test statistic shows decreasing trend in MK test. After calculating z-statistics, normal distribution function is used to calculate the value p (P(Z z)).
p values equal or smaller than 0.1 indicate a significant trend at 90% significance level, whereas p values equal or smaller than 0.05 and 0.01 indicate a significant trend at 95% and 99% significance levels, respectively.
B. Spatial Analysis Spatial analysis of the groundwater level was conducted to understand the variation of groundwater level over space. Non-parametric P-Splines method, which enables estimation of regression functions contrast with more traditional parametric models with greater flexibility in fitting data [25], was used for spatial analysis. More details of P-Splines method are explained in [26], [27]. Table II shows MK p-values and z statistics for eleven well fields.  In Table II, bold numbers indicate statistically significant trends at 99% significance level. As shown in Table II, decreasing groundwater level trends were detected in all well fields. All detected trends were statistically significant at 99% significance level except the trends in Bedai well field. Supporting the results in Table II, Fig. 2 shows decreasing trend in groundwater levels over time. As can be seen in Table II, the strongest decreasing trend in groundwater level was detected in Aweed well field followed by Hamdah and Seih-Hermal well fields. The significant decreasing trends can be explained by large agricultural water demand around these three well fields. The groundwater production from these well fields were used to meet irrigation water demand in particular for date farms in the region.

A. Temporal Trend Analysis
Although the most significant decreasing trends (over time) was found at Aweed well field, the maximum groundwater level decrease occurred in Seih-Hermal well field. Fig. 3 shows the groundwater production values over the period of 2004 -2018 in well fields including Aweed, Ghuraifa, Seih-Aqareb and Seih-Hermal.

B. Spatial Analysis
Spatial analysis was conducted to understand the variation of groundwater levels over the study are. Fig. 4 shows an example of spatial analysis results for the years 2006 ( Fig. 4 (a)) and 2009 ( Fig. 4(b)). These years were selected, since almost all well fields had groundwater level data for these two years (which made spatial interpolation more reliable). Fig. 4 shows that highest groundwater levels were observed in southeast to east Sharjah, whereas lowest groundwater levels were detected in west to northwest Sharjah. It should be noted that scales in Fig. 4 represents the distance from  Science and Development, Vol. 11, No. 1, January 2020 groundwater table to the ground surface. Therefore, lower values in the scale (represented by blue color) indicate higher groundwater level. It can be recognized from Fig. 4 that groundwater levels decreased in all well fields between the years 2006 and 2009. The groundwater level in the well fields close to the coastline (i.e., Bedai and Sajaa) are very low. Also, temporal trend analysis showed significant groundwater level decrease in Sajaa. If the groundwater levels keep decreasing around coastal region, the sea water will intrude to the groundwater aquifer and the groundwater will be saline. This will require advanced treatment of groundwater before using it to satisfy water demands.
Reference [28] studied on assessment of groundwater quality in Sharjah, and they reported that salinity related water quality parameters including Electrical Conductivity (EC), Chloride (Cl -) and Total Dissolved Solids (TDS) are very high in well fields close to the coastal region. Fig. 5 illustrates well fields very close to sea cost (Bedia, Sajaa and Muhadhab). Table III shows the average, minimum and maximum values of EC, Cland TDS in groundwater samples from Bedia and Muhadhab well fields along with permissible limits of these parameters according to Abu Dhabi Water Quality Regulations (ADWQR), standard deviation (std) and number of samples (count). It should be noted that groundwater quality data is not available for Sajaa well field. As shown in Table III, Table IV. Positive z statistics in Table IV shows that Cl -, TDS and EC values in well fields close to sea are increasing, and these trends are statistically significant at 99% confidence level (as p values are smaller than 0.01).
This means that if groundwater levels keep decreasing in these well fields, the groundwater will be saltier, and groundwater will not be a fresh water resource.

V. CONCLUSIONS
In this study, monthly groundwater level data obtained from eleven well fields were used to conduct spatio-temporal trend analysis of groundwater levels in Sharjah, UAE. The spatial analysis of groundwater levels in Sharjah was conducted by P-spline method, whereas MK non-parametric test was used for the temporal trend analysis. Followings are the major conclusions from this study:  Decreasing groundwater level trends were detected in all well fields in Sharjah.  All detected trends were statistically significant at 99% significance level except the trend in Bedai well field.  The highest decreasing trend was found in Aweed well field, and the highest decrease in groundwater level was detected in Seih-Hermal well field.  In parallel to decreasing groundwater levels, the groundwater production has decreased significantly (more than 90% decrease in some well fields).  It was found that groundwater levels in well fields close to coastline experienced a significant decrease over time leading seawater intrusion and salinization of groundwater.  Analysis of salinity related water quality parameters including Cl -, TDS and EC at well fields close to sea showed that these parameters have very high values (very well above the permissible limits).  Statistically significant increasing trends were detected in Cl -, TDS and EC parameters at well fields close to sea.  Groundwater level decreases (and increased salinity) in particular at well fields close to sea require immediate action by groundwater managers to protect fresh groundwater. Findings of this study provide significant input for groundwater management in Sharjah. As a next step, solutions to groundwater level decreases such as artificial recharge of the aquifers, or harvesting the limited rainfall in small dams and natural recharge of the aquifers will be investigated and discussed in further studies.

CONFLICT OF INTEREST
The authors declare no conflict of interest. AUTHOR CONTRIBUTIONS Abdullah G. Yilmaz supervised the analysis and wrote the paper, Abdallah Shanableh and Rami I. Al-Ruzouq contributed data collection and analysis, and reviewed the paper. Naseraldin Kayemah assisted data analysis.