Aridity trend in the Middle East and the adjacent areas

Available water resources in the Middle East, as one of the most water-scarce regions of the world, have undergone extra pressure due to climatic change, population growth, and economic development during the past decades. The objective of this study is to detect the trends and quantify the changes in aridity with respect to precipitation and potential evapotranspiration in 20 countries of the Middle East and the adjacent area. A Pixel-wised trend analysis was conducted on precipitation, potential evapotranspiration, and aridity index for 71 years from 1948 to 2018. A nonparametric Mann-Kendall test was used over 14106 points in the study area to detect the trends at monthly and annual time scales. Results showed statistically significant (|Z| >1.96) upward trends in aridity (a downward trend in aridity index) up to 96 percent from December through September in most parts of the region. Aridity in October and November had a downward tendency in most parts of the study area. At the annual time scale, 62.5 percent of the statistically significant trends in aridity were found to be upward (up to 96 percent increase in aridity) due to the combined effects of the decrease in precipitation and the increase in potential evapotranspiration and 37.5 percent of the detected trends were downward (up to 61 percent decrease in aridity). The highest and the lowest trends in aridity were found in the north of Sudan (96 percent increase in aridity) and Eastern Arabia (61 percent decrease in aridity), respectively.

mitigate the harmful impacts of climatic changes through adapting appropriate strategies for new climatic scenarios. Trend analysis of climate variables is a useful tool to investigate the behavior and changes associated with climate variables. Numerous studies have been conducted to study the trend analysis of precipitation and evaporation all over the globe (Partal and Kahua, 2006;Dinpashoh et al., 2011;Irmak et al., 2012;Jhajharia et al., 2012 andFathian et al., 2015;D'Oria et al., 2017;Xing et al., 2018;Tehrani et al., 2019;Güçlü 2020).
Time series analysis of precipitation, evaporation, temperature, and aridity index have been conducted in numerous case studies in the Middle East. Tabari et al., 2012, investigated the temporal pattern of monthly and annual aridity index in the north and northwest of Iran for a period of 40 years (1966 -2005) and concluded that the aridity in the region increased and the trends were more pronounced in the semi-arid sections of the study area. Some'e, et al., 2012 conducted a spatiotemporal trend analysis in arid and semiarid regions of Iran for 40 years (1966 -2005). They found statistically significant negative trends in aridity index (more dryness) in 55 percent of the stations and a significant positive trend in two stations. Patal and Kahya, 2005 used a non-parametric method to detect the trends in mean annual and monthly precipitation in Turkey for a 65-year study period (1929 to 1993). They found a noticeable decrease in the mean annual precipitation in western and southern parts of Turkey as well as along the coast of the Black sea. In Saudi Arabia, trend analysis was conducted on extreme temperature indices for 30 years , and results showed a significant increase in the annual occurrence of warm days/nights and a decrease in the annual occurrence of cold days/nights (Almazroui et al., 2013). Trend analysis of average annual temperature, annual precipitation, and annual aridity index in Israel showed positive trends in temperature and aridity among the 12 investigated stations for the period 1970-2002(Kafle and Bruins, 2007. In Egypt, the trend analysis of rainfall in 31 stations showed that 77 percent of the detected trends were negative, concluding a decrease in precipitation in Egypt (Gado et al., 2019).
In all the research mentioned above, trend analyses were applied to local field-based data, which reflect the values for those specific points and not for the area between the stations. The issue of spatial gaps between the stations was addressed in previous studies (e.g., Some'e, et al., 2012;Tehrani et al., 2019) using Inverse distance weighting (IDW) method to transfer data from multiple locations. However, the assigned values to unknown points in the IDW method are only estimated with a weighted average of the values, and it does not consider the other crucial factors such as heterogeneity of landscapes and the elevation change between stations (Livneh et al., 2014). To better understand the patterns and effects of climate change on water resources, spatially, and temporally continuous data at a fine resolution grid are needed (Hasan et al., 2019;Sahour et al., 2020). In this research, we conducted a pixel-wise trend analysis of aridity index over the entire Middle East and adjacent regions for 71 years (1948 to 2018), based on a detailed review of monthly precipitation and annual potential evapotranspiration. The trend analysis in this study was performed at the pixel level in the gridded data. The time series for individual pixel (data point) were extracted, and trend analysis was conducted for each pixel independently. Our climatic trend analysis evaluates spatiotemporal trend analysis of the following variables: monthly and annual precipitation (P), monthly and annual potential evapotranspiration (PET), and monthly and annual aridity index (P/PET).

Study area
The study area (Figure 1) is the Middle East (hereafter referred to as ME) including 20 countries (Armenia, Azerbaijan, Bahrain, Cyprus, Egypt, Eritrea, Iran, Iraq, Israel, Jordan, Kuwait, Lebanon, Oman, Qatar, Saudi Arabia, Sudan, Syria, Turkey, United Arab Emirates, and Yemen) with a population of 432 million (Roudi-Fahimi and Kent, 2007). The region is diverse in its climate and landscape, from the snowy summits of the Eastern Anatolia down to the empty quarter of the Arabian Peninsula (AP). Considering the aridity index, which is the ratio of precipitation to potential evapotranspiration, in previous studies, the hyper-arid area was found in Egypt, Saudi Arabia, and Sudan (Terink et al., 2013). Iran, the northern coast of Egypt, southwest of Saudi Arabia, and the western coastal region of the ME were defined as arid to semi-arid. The humid area was found in some parts of Turkey, Iran, Iraq, western Syria, and Lebanon (Terink et al., 2013).
The major challenge in the ME region is to manage the low annual precipitation rate and limited available water resources for the water demands in agricultural, industrial, and urban populations sectors.

5
The ME has one of the fastest-growing populations in the world, with an average annual growth rate of 2.1 percent between 1990and 2003(The World Bank, 2005. The population was growing about 100 million Figure1. Location of the study area. in 1960, 311 million in 2006, and projected to be more than 430 million in 2025 (Joffe, 2005), bringing the average amount of water per capita in the region far lower the scarcity level. According to the World Bank (2007), most of the countries in ME cannot meet their current water demands, and the situation is getting worse due to climate change and the increase in population. The water resources in the ME are mostly used in agricultural sectors, which helps the ME regional economies. Iran, Lebanon, Turkey have a relatively better situation in terms of renewable water; however, within-country and within-year variation in precipitation are problematically significant. The majority of the ME countries with scarce renewable water resources are highly dependent on non-renewable groundwater sources and seawater desalination. These countries include Bahrain, Jordan, Kuwait, Libya, Oman, Qatar, Saudi Arabia, the United Arab Emirates, and Yemen. Some others, including Syria, Iraq, and Egypt, are dependent on the inflow of trans-boundary rivers such as the Nile, the Tigris, and the Euphrates (World Bank, 2007).

Data
The NOAH output of the Global Land Data Assimilation System (GLDAS-NOAH) provides global precipitation and potential evapotranspiration data. In this research, the time series of monthly precipitation (P) and potential evapotranspiration (PET) were used to calculate the monthly aridity index (P/PET) for 71 years (1948 to 2018). The variables are available in a gridded format at a 0.25º × 0.25º spatial resolution.
GLDAS uses advanced numerical models of physical processes to integrate data from multiple groundbased and space-based observing systems to produce spatially and temporally continuous water and energy states and fluxes data. The data is available at https://disc.gsfc.nasa.gov/datasets?keywords=GLDAS.
Potential evapotranspiration (PET) is described as the amount of water that could be released to the atmosphere where it available. It is a function of surface and air temperatures, insolation, and wind, all of which affect water-vapor concentrations immediately above the evaporating surface. The broad definition of dryland is a place where the annual PET exceeds annual precipitation. The potential evapotranspiration can be calculated through the following equation: Precipitation in this research refers to the total amount of rainfall and snow water equivalent. The monthly aridity index (P/PET) was calculated for each month (UNEP, 1992). Based on the aridity index, climates can be classified into four groups (Table 1). The more precipitation, the higher the aridity index. In other words, a drier area has a lower aridity index. Regarding that, a negative trend in the aridity index means that the climate is moving toward aridity and vice versa.

Methodology
In this section, we describe the methods we used for monthly and annual trend analysis to detect the trends in precipitation, potential evapotranspiration, and the aridity index.

Mann-Kendall (MK) test for monotonic trend
The primary purpose of the Mann-Kendall (Mann, 1945;Kendall, 1975)  Then at the type I error rate , is rejected and is accepted if .
The trends with values of |Z| > 1.96 were considered to be statistically significant.

Sen's Slope
This section shows how to estimate the true slope (change per unit time) by using a non-parametric procedure developed by Sen (1968). The procedure in sen's slope is an extension of a test by Theil (1950), which is illustrated by Hollander and Wolfe (1973 Eq.5

Pre-Whitening
As noted above (Mann-Kendall test section), the Mann-Kendall test requires the time series to be serially independent (not serially correlated). Von Storch (1999) proposed a procedure called pre-whitening to eliminate this serial correlation effecting the Mann-Kendall test. The idea of the pre-whitening process is to remove the serial correlation assuming a lag-one autoregressive model, and then apply the MK test to the serially independent residuals. So that a serially correlated time series ( ) is pre-whitened by where is the lag-one serial correlation coefficient calculated by (Eq.6) Eq. 6 Where

Relative change
The relative change of the annual and monthly aridity indices were calculated using the following equation Where L is the length of trend (number of months of years), is the slope of the trend, and the | � | is the absolute average value of the time series.
All methods described above for the trend analysis were conducted in a pixel level. There were 14106 pixels (locations) for each variable (P, PET and AI) covering the study. Each pixel had a unique time series representing the values for that specific variable over time. Therefore, the time series analysis for each pixel was independent from adjacent pixels.

Monthly trends for precipitation (P)
Trend analysis of monthly precipitation ( figure 3 and figure 4) was conducted for each month of the year for the study period . For the January, February, and March, statistically significant (|Z| > 1.96) upward trend (from 0.3 to 0.5 mm/ month) were observed in the northwest of Turkey, and southwest of Iran  (Figures 3 and 4).
Also for December, upward trends were detected in the north of Turkey and west of Iran, and statistically significant downward trends were detected in the southeast of Iran, north of Iraq and Syria, Israel, southwest of Turkey, Lebanon and Israel, Jordan and Cyprus ( figure 3 and 4).

Monthly trends for potential evapotranspiration (PET)
Trend analysis of monthly PET was conducted for each month for the 71 years study period (Figures 5 and   6). The higher the PET values indicates the potential of the area for releasing water to the atmosphere and getting drier.

Monthly trends for aridity index (AI)
The monthly aridity index (AI) was calculated (P/PET) for each month (Figures 7 and 8). As we mentioned earlier, drier areas have a lower aridity index. The negative (downward) AI trends indicate that the area is getting drier. The upward AI trends suggest that the area is getting wetter.

Annual trends for precipitation, potential evapotranspiration, and aridity index
The annual trend for precipitation and statistical significance (figure 9) of the annual trends were calculated for the entire study area. The relative changes were also calculated to show the changes in percentage.
Statistically significant (|Z| < 1.96) downward trends were detected in eastern (2 to  The detected downward trends are consistent with the results from previous studies by Patal and Kahya, 2005 in Turkey; Kafle and Bruins, 2007, in Israel;and Tabari and Talaee, 2011, in Table 2 shows the percentage of the area with positive and negative trends relative to all detected significant trends and also the maximum and the minimum detected relative change for each variable.    Turkey. The areas of increased aridity were correlated with the areas with negative precipitation trends. The sources of errors in our study are related to (1) Uncertainty in the variables (precipitation, potential evapotranspiration, and calculated aridity index) that were used for the spatiotemporal trend analysis; and (2) errors introduced by the applied statistical methods. In this research, we did not consider the errors related to the dataset. Those errors could be related to measurements, and/or the interpolation techniques were used to generate the precipitation and potential evapotranspiration data.
According to the World Bank (2007), the majority of countries in ME cannot meet their current water demands. This study shows that many of those countries are moving toward severe aridity due to a