In this study, rainfall data from 19 stations in Saudi Arabia (SA) for the period 1985-2019 was utilized to investigate interannual, monthly, and seasonal rainfall variations and trends. The magnitudes of these trends were characterized and tested using Mann-Kendall (MK) rank statistics at different significance levels. During this study period, the mean rainfall in SA showed a slight and significant decreasing trend by about 2 mm/35 years. Investigation of seasonal trends of rainfall revealed that Winter and Spring rainfall decreased significantly by 2.7 mm/35 years and 5.4 mm/35 years respectively. Three months showed very slight significant decreasing trends of rainfall. These were the months of February, March and April. Mann-Kendall analyses were carried out to investigate the annual trends of rainfall during three sub-periods, i.e., 1985-1996, 1997-2008, and 2009-2019. The results revealed that while rainfall increased by 5.3 mm/12 years and 7.8 mm/11 years for the first and the third periods respectively, it decreased by about 11 mm/12 years during the second period. While trends of rainfall in Saudi Arabia are affected by large scale circulations and local factors, the effect of extraterrestrial factors, such as solar activity and its consequent effects on the climate may, additionally, play a potential role in affecting the pattern of rainfall in Saudi Arabia.
Climate is one of the important drivers of ecosystems, composition and other Earth systems. Precipitation and temperature are two of the most important variables in climate science, often used to track the magnitude and variability of climate change. Climate variability will have unintended consequences in many parts of the world in terms of the frequency and magnitude of precipitation and temperature fluctuations (e.g., Hunt, 2000; Easterling et al., 2000; Goswami et al., 2006; Almazroui et al., 2017a; Myhre et al., 2019 ). Consequently, climate variability tends to affect specific sectors, such as agriculture, public health, water supply, energy production and use, and land use and development (Bell et al., 2018; Sung et al., 2015) .
There are different sources of water among which, rainfall, in semi-arid and arid areas, is considered as the main source. Precipitation is a vital part of the hydrologic cycle and changes in its pattern would directly influence the water resources of a given area. Knowledge and understanding of the distribution, variations, and trends of rainfall are of vital importance, and have several applications (Augustine & McNaughton, 2006; Olayide et al., 2016; Zhang et al., 2007) .
The pattern and trend of rainfall over different spatial and temporal scales have been investigated. It was found that global rainfall trend increased throughout the last century, although large areas were characterized by negative trends(e.g., Zhang et al., 2007; Funk et al., 2014; Bari et al., 2016; Wang et al., 2016; Tan et al., 2019 ).
The pattern of rainfall and its associated atmospheric circulation over Saudi Arabia have been discussed extensively in several articles (Hasanean & Almazroui, 2015; Almazroui & Saeed, 2020; Almazroui et al., 2017b; Almazroui, 2020) .
The annual mean rainfall over Saudi Arabia varies considerably and displays a large spatial and temporal variability over the entire country. Also, it has strong seasonality, as the influence of both tropical and extra-tropical drivers causes extreme rainfall events at different times through the annual cycle. The southern region of the country receives the highest rainfall compared to the northwest territory, which is considered the driest region with the lowest average annual rainfall (e.g., Almazroui, 2020 and references therein).
The pattern of rainfall in SA is mainly affected by three major air masses; 1) the monsoon fronts during late Autumn, 2) the continental tropical air mass coming from the Atlantic Ocean through the central and northern parts of the African continent, and 3) the maritime polar air mass coming from the eastern Mediterranean Sea (Alyamani, 2001) . For instance, the winter rainfall pattern is affected throughout North Africa, Europe, West Asia, and Central Asia, while Spring rainfall is mainly linked to the westerly systems vis-à-vis the central landmass of the Arabian Peninsula except in the northern Red Sea coastal area (Hasanean & Almazroui, 2015; Almazroui et al., 2017b; Almazroui, 2020) .
Rainfall variability over large scales is affected by regional and/or local variations (Farnsworth et al., 2011, Olayide et al., 2016) . Therefore, it is important to analyze rainfall variability on the local scale for better assessment of its variability and distribution.
In this study, the longest historical records of rainfall from 19 meteorological stations across Saudi Arabia were used to investigate monthly, seasonal, and annual, long-term rainfall trends and variations for the last 35 years (1985-2019). Mann Kendall test was used to investigate long-term trends at various significance levels. Sen’s estimate was used to show the increasing/decreasing magnitudes of each trend.
Saudi Arabia is a large country with an area of about 2,250,000 km2, and it is stretches from 15.5˚N - 32.5˚N in latitude. It has a diverse climate which changes from region to region but, most of it lies within the subtropical desert climate zone. It is considered to be a dry country due to its low annual rainfall. However, precipitation varies over space and time, and the amount of precipitation also varies from region to region.
Monthly mean precipitation data series from 19 weather stations for the period between 1985-2019 was analyzed for the purpose of this study. This data was provided by the Saudi Arabian Ministry of Environmental Protection (MEP). The considered stations were selected according to the length and completeness of their records, and keeping in mind that a reasonable spatial and temporal coverage of the country is achieved (
The data was subjected to extensive quality control procedures including the non‑parametric Kruskal-Wallis test (Essenwanger, 1986) so as to retain the homogeneity of the data, and to eliminate multiple random errors found in the original data. Furthermore, linear interpolation methods were used to fill gaps in the missing data.
The mean rainfall recorded by 19 stations during the period of study indicated that Abha had the highest precipitation while the lowest was recorded in Tabuk.
In general, relatively heavy precipitation was received over the south-western region. This is due to the orographic features of the region (Subyani, 2004) . Around 38% of the country’s rainfall was received by the south-western region (Abha, Khamis Mushait, Albaha, Gizan, Bisha), and it decreased moving north where least precipitation occurred.
To study the variability of rainfall for the whole country, the monthly mean values from each station for the considered period were utilized. The monthly distribution of rainfall was classified into 12 bins, and the seasonal variations were divided into 4 groups. Winter included the months of December, January and February (DJF); Spring included March, April and May (MAM); June, July and August (JJA) made up Summer; and Autumn included September, October and November (SON).
Mann-Kendall (MK) tests were carried out and Sen’s slopes were determined to assess the trends of rainfall. These are non-parametric statistical tests that are widely used for analyzing trends in climatologic time series (e.g., Yue & Wang, 2004, Maghrabi, 2019 ).
The MK test can be applied to a time series x i ranked from i = 1 , 2 , ⋯ , n − 1 and x j ranked from j = i + 1 , 2 , ⋯ , n such that
sgn ( x j − x i ) = { 1 if ( x j − x i ) > 0 0 if ( x j − x i ) = 0 − 1 if ( x j − x i ) < 0 (1)
The Kendall test statistic S can be computed as
S = ∑ i = 1 n − 1 ∑ j = i + 1 n sgn ( x j − x i ) (2)
where xj and xj are data values at times j and i, j > i, respectively, and sgn(xj − xi) is the signum function. S is assumed to be asymptotically normal, with the expectation value E(S) = 0 for a sample size of n ≥ 8 and variance as follows (Yue & Pilon, 2004) :
V ( S ) = 1 18 [ n ( n − 1 ) ( 2 n + 5 ) − ∑ p = 1 q t p ( t p − 1 ) ( 2 t p + 5 ) ] (3)
Here, q is the number of tied groups, and t p is the number of data values in group p.
The values of S and V(S) are used to compute the test statistic Z as follows:
Z = { S − 1 V ( S ) if S > 0 0 if S = 0 S + 1 V ( S ) if S < 0 (4)
The standardized MK test statistic Z follows the standard normal distribution with a mean of zero and a variance of one and is used to measure the significance of the trend. If |Z| is greater than Zcrit (Zcrit is the (100 × (1 − α/2))th percentile of the standard normal distribution with a chosen significance level α (e.g., 5% with Z0.025 = ±1.96)), then the null hypothesis is invalid, implying that the trend is significant.
The mean rainfall was 550 mm, 811 mm, 235 mm, and 447 mm for Winter, Spring, Summer and Fall respectively. Monthly rainfall received over the country during Spring was about 40% of the country’s total compared to other seasons, with the highest contribution made in April which made up for about 44% of the Spring total. Winter rainfall was about 27% of the country’s total, in which the highest contribution was made in January, which made up for 43% of the Winter total. The lowest rainfall was received during Summer (236 mm), which made up for 11.5% of the country’s total, with the lowest contribution made in June, making up for 16.7% of the Summer total.
Close inspection of the data shows that there were three distinctive periods in which rainfall changed substantially. These were the period between 1985 and 1996, between 1997 and 2008, and between 2009 and 2019. Hence, MK analyses were performed for each period separately.
The MK results revealed that for the first period (12 years), there was non-significant increasing trend in the rainfall with about of 5.73 mm/12 years. For the third period between 2009 and 2019, MK results indicated non-significant
increase in the rainfall by about of 7.8 mm/11 years. On the other hand, significant decreasing trend with 11.32 mm/12 years at 0.01 significant level for the period between 1997-2008 was detected.
For seasonal rainfall during the sub-periods, rainfall showed increasing trends in one period and decreasing trends in another. For instance, for the period between 1997-2008, rainfall for all the seasons showed decreasing trends, whereas the periods between 2009-2019 showed increasing trends. For the former period, there was significant decrease by 15.18 mm/12 years with a significance level of 95% in Fall, and significant decrease by 7.21 mm/12 years (confidence level 90%) in Winter. For the period between 1985-1996, while the rainfall showed decreasing trend in Spring, the opposite happened for the remaining seasons.
The trend of rainfall in Saudi Arabia for each month was computed using MK tests and the results for the same are given in
| Time series | 1985-2019 | 1985-1996 | 1997-2008 | 2009-2019 | ||||
|---|---|---|---|---|---|---|---|---|
| n | Q mm/35 yrs | n | Q mm/12 yrs. | n | Q mm/12 yrs | n | Q mm/11 yrs | |
| Annual | 35 | −1.78* | 12 | 5.73 | 12 | −11.32** | 11 | 7.80 |
| January | 35 | −0.07 | 12 | 0.52 | 12 | −0.62 | 11 | −0.07 |
| February | 35 | −0.11+ | 12 | 0.72 | 12 | −0.02 | 11 | 0.34 |
| March | 35 | −0.37** | 12 | −0.05 | 12 | −3.57** | 11 | −0.01 |
| April | 35 | −0.37+ | 12 | −0.85 | 12 | 1.24 | 11 | 0.10 |
| May | 35 | 0.02 | 12 | 0.65 | 12 | 0.16 | 11 | 0.12 |
| June | 35 | 0.00 | 12 | 0.04 | 12 | −0.17+ | 11 | −0.03 |
| July | 35 | 0.00 | 12 | 0.34+ | 12 | −0.08 | 11 | −0.20 |
| August | 35 | 0.01 | 12 | 0.08 | 12 | −0.86+ | 11 | 0.05 |
| September | 35 | −0.01 | 12 | 0.01 | 12 | −0.12+ | 11 | 0.15 |
| October | 35 | −0.08 | 12 | 1.00 | 12 | −0.42 | 11 | 0.63 |
| November | 35 | 0.20 | 12 | 0.40 | 12 | −1.98* | 11 | 1.31 |
| December | 35 | −0.18 | 12 | 0.08 | 12 | −0.58 | 11 | −0.07 |
| Winter | 35 | −2.72* | 12 | 7.78 | 12 | −7.21+ | 11 | 4.13 |
| Spring | 35 | −5.42** | 12 | −7.24 | 12 | −13.20 | 11 | 8.28 |
| Summer | 35 | 0.135 | 12 | 6.94 | 12 | −5.84 | 11 | 2.16 |
| Fall | 35 | −0.255 | 12 | 8.41 | 12 | −15.18* | 11 | 14.95 |
***trend at α = 0.001 level of significance; **trend at α = 0.01 level of significance. *trend at α = 0.05 level of significance; +trend at α = 0.1 level of significance.
February, by 0.37 mm/35 years (α = 0.01) in March, and with the same decreasing rate at α = 0.1 in April.
While non-statistically significant decreasing trend was found in January, December, September, and October, increasing trends were found for the remaining months.
By considering the sub-periods, MK results for the period 1985-1996 revealed that rainfall in Saudi Arabia for each month increased slightly with different magnitudes. Rainfall was only significant at 0.1 significance level (0.34/12 years) in the month of July. For the period between 1997-2008 and at 0.1 significance level, rainfall decreased in June, August, and September by 0.17 mm, 0.86 mm, and 0.12 mm per 12 years, respectively. For the same period, and at 0.05 significance level, the rainfall decreased by 1.98/12 years in November. In addition, MK test indicated a decreasing trend by −3.57/12 years at 0.01 significance level in March. For the rest of the months in this period, and for the period between 2009-2019, monthly trends of rainfall in Saudi Arabia varied between increasing or decreasing trends.
Saudi Arabia is categorized mainly as a semi-arid to arid climatic zone (Köppen & Geiger, 1930) middle to northern regions receives moderate rainfall while southern areas have the highest rainfall. The arid and semi-arid regions have an extreme, continental climate with low annual rainfall, high temperatures, a dry Summer, and a very cold Winter, especially in the central regions of the country (e.g., Almazroui et al., 2012; Edgell, 2006, Maghrabi & Al Dajani, 2013 ).
In addition to large-scale circulation, local factors such as the geographical position, local surface heating and topography for each individual site play a potential role in determining the spatial distribution and intensity of rainfall in Saudi Arabia (Alyamani, 2001; Al Senafi & Anis, 2015; De Vries et al., 2016) . These factors may explain some of the trends in rainfall that are observed during some periods in some regions, but not all of them.
Moreover, the impact of extraterrestrial factors such as meteor showers and solar activity, and their consequential effect on the Earth’s climate, are also expected to affect rainfall patterns in several locations around the world including Saudi Arabia (Bowen, 1956; Laurenz et al., 2019; Bhattacharyya & Narasimha, 2005) .
For instance, it has been proposed that the El Niño Southern Oscillation (ENSO) significantly affects the global temperature patterns during certain periods, which consequentially, influences rainfall variability through large changes in moisture availability (Kirov & Georgieva, 2002; Emori & Brown, 2005; Foster & Rahmstorf, 2011) . It has been found that the ENSO is significantly correlated with sunspots, which, in turn, are considered as solar activity indicators (Kirov & Georgieva, 2002) .
Additionally, solar activity and associated magnetic activities may also have a potential role in influencing the pattern of rainfall, or cause severe drought/rainfall episodes observed in Saudi Arabia (Bucha, 1991) . This arises from the fact that solar activity and geomagnetic disturbances can modulate the flux of high-energy cosmic rays, which are considered as the main source of ionization in the atmosphere. These variations in the ionization can directly or indirectly affect the physical and chemical properties of the atmosphere; for example, they can change the rate of cloud formation, water vapor content, and atmospheric aerosols (Svensmark & Friis-Christensen, 1997; Carslaw et al., 2002; Svensmark et al., 2009; Maghrabi & Kudela, 2019; Maghrabi, 2019) which, in turn, affect the rate and pattern of rainfall. Research on this subject is ongoing.
The longest, historical rainfall data from 19 stations in Saudi Arabia for the period 1985-2019 was used to investigate interannual, monthly, and seasonal rainfall variations and trends.
The mean rainfall in Saudi Arabia was reported to be about 170 mm. The highest rainfall occurred in March and April whereas, the lowest was recorded in June and September. The monthly rainfall in Spring was about 811 mm which accounted for about 40% of the country’s total compared to the other seasons. On the other hand, Summer, the lowest rainfall was about 235 mm, was recorded.
Non-parametric MK tests were performed to detect rainfall magnitude and trends which gave the following results:
1) The mean rainfall in Saudi Arabia showed a decreasing trend during the study period. This decrease was about 2 mm/35 years with a 95% confidence level.
2) The MK analyses for three sub-periods 1985-1996, 1997-2008, and 2009-2019 revealed that rainfall increased by 5.3 mm/12 years and 7.8 mm/11 years for the first and the third periods respectively.
3) Significant decrease of about 11 mm/12 years was detected during the second period.
4) Seasonal trends of rainfall showed that Winter and Spring rainfalls decreased significantly by 2.7 mm/35 years and 5.4 mm/35 years respectively.
5) Three months showed very slight significant decreasing trends in their rainfall. These are February, March and April.
Rainfall trends and distributions are affected by global and regional factors, such as large scale circulations, and local factors such as topography of the sites and geographical positions. However, extraterrestrial factors such as solar activity and their consequential effects on the climate may affect the pattern of rainfall in Saudi Arabia. These factors continue to be subjects research.
The authors would like to thank King Abdulaziz University for supporting this work.
All research was conducted according to the ethical guidelines provided by the host institutions the authors are affiliated with.
All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by [Maghrabi Abdullrahman], [Alruhaili Aied], and [Hadeel A. Alamoudi]. The first draft of the manuscript was written by [Maghrabi Abdullrahman] and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.
The datasets generated and analysed during the current study are not publicly available due the fact that they constitute an excerpt of research in progress but are available from the corresponding author on reasonable request.
The authors have no relevant financial or non-financial interests to disclose.
Maghrabi, A. H., Alamoudi, H. A., & Alruhaili, A. S. (2023). Long-Term Rainfall Trends in South West Asia—Saudi Arabia. American Journal of Climate Change, 12, 204-217. https://doi.org/10.4236/ajcc.2023.121010