Abstract

In the current literature, little is known about the role of Siberian High (SbH) on precipitation variability over the Kingdom of Saudi Arabia. The technique of centers of action (COA) has been employed here to study the possible impact of SbH on interannual variations in winter precipitation over the kingdom because the COA methodology helps us to find changes in interannual variations of pressure intensity, latitudinal and longitudinal positions of the pressure system using sea-level pressure data. The results show that there are two regions of the kingdom whose precipitation is significantly affected by SbH, the southeast region and the region from northeast to central Saudi Arabia. The precipitation over the southeastern region is significantly influenced by the intensity of SbH as well as its meridional displacement whereas the precipitation variability from northeast to the central kingdom is significantly influenced by the meridional displacement of SbH. The effect of remote forcings of North Atlantic Oscillation and El-Nino Southern Oscillation on precipitation of the kingdom has also been discussed in the paper. The empirical results can be understood by the mechanism of changes and circulations in the atmosphere.

INTRODUCTION

Bruce (1994) pointed out that extreme hydro-meteorological occurrences, such as heavy rainfall, floods, storms and typhoons, are regarded as being the costliest natural disaster risk and leading research hotspots (Bruce 1994; Kar et al. 2017). Thus, rainfall magnitude and frequency distribution related information are very important for the management of water resources, design and operations of water system (Kar et al. 2017; Arab Amiri & Mesgari 2018; Chowdhury et al. 2019). The importance of study of climate variability over various regions of the Middle East has also been discussed in the current literature (e.g. Arab Amiri et al. 2016). Arab Amiri et al. (2016, 2018), Arab Amiri & Conoscenti (2017) and Arab Amiri & Mesgari (2018) used mean monthly rainfall measured at four synoptic stations over the QaraQum catchment to explain temporal behavior and spatial distribution of mean monthly precipitation. Similarly, Arab Amiri & Mesgari (2018) carried out the spatial analysis of precipitation extremes in northwest Iran. Moreover, Arab Amiri & Conoscenti (2017) discussed the linkage between the precipitation and shallow landslides in Mazandaran province, Iran. Furthermore, Babu et al. (2011) carried out a very comprehensive analysis on the rainfall climatology of the Middle East. The results obtained from the study showed the daily and monthly pattern of rainfall in the region. The areas that receive less rainfall and the areas that receive more rainfall were indicated in their study. Moreover, Almazroui et al. (2012a, 2012b) studied and analyzed the rainfall and temperature climatology of the Kingdom of Saudi Arabia (KSA). The rainfall exhibits the mixed, insignificantly increasing and significantly decreasing trend from 1979 to 2009 (Almazroui et al. 2012b). The region of Rub Al-Khali receives the lightest rainfall during the winter season (November–April) and the relatively heavy rainfall is inclined southwest to northeast (Almazroui et al. 2012b).

Hasanean & Almazroui (2015) observed that the climatic variation in the region of KSA is greatly influenced by El-Nino Southern Oscillation (ENSO). ENSO plays an important role in pulling dry and hot masses of air in the region of KSA. Also, Hameed & Riemer (2012) observed that Sahel precipitation is considerably affected by different atmospheric pressures, namely South Asia low and Azores High. Similarly, Al-Ahmadi & Al-Ahmadi (2013b) showed the spatial variations in the relationship between topography and mean annual and seasonal rainfalls in southwestern Saudi Arabia. The results show that the rainfall in the foothills of the Asir Mountains increases during the winter season (Al-Ahmadi & Al-Ahmadi 2013b). Furthermore, Abdullah & Almazroui (1998) studied climatological analysis of rainfall over the southwestern region of KSA. The southwest (SW) part of KSA can be classified as arid and semi-arid (Abdullah & Almazroui 1998). Their study showed that the rainfall distribution over SW KSA is not uniform due to many factors such as elevation, topographic configuration and orientation.

Cohen et al. (2001) observed that there are mainly three semi-permanent centers of action (COA) in the northern hemisphere (NH) from middle to high latitude during the winter season that describe the configuration of seasonal weather, among these three two of them, namely the Aleutian Lows and Icelandic, are inherent in major ocean basins of the NH, and the third is the SbH that is inherent in the region of Asia. SbH is linked with the densest and coldest air masses in the NH (Lydolf 1977). It is a semi-permanent and quasi-stationary atmospheric center of action, which is more prominent in the boreal winter season. The SbH forms generally in the month of October which is the result of continuous and strong radiative cooling in the lower troposphere located above the snow-covered surface of Asia and continues until around the end of April (Panagiotopoulos et al. 2005). The possible effect of SbH in the climatic fluctuations over the region of Europe, the Arctic region and Middle East region have received relatively less importance. Roger (1997) studied the changes in the North Atlantic storm-track and proposed that the westward extension of the SbH into Europe is related to southwesterly advection of warm air into northern Europe (Cohen et al. 2001). The SbH is surrounded by regions of high terrain north of the Tibetan Plateau (Jones & Cohen 2011).

Some recent studies (Panagiotopoulos et al. 2005; Jeong et al. 2011; Hasanean et al. 2013) focused on analysis of the intensity of SbH. Panagiotopoulos et al. (2005) specifically studied the fluctuations in the intensity of SbH. On the other hand, Jeong et al. (2011) mentioned the issue of decreasing and recovery in trend of the SbH intensity and claimed that the recovery of the trend in recent years is much faster as compared to the decreasing trend in the 1970s and 1980s. Similarly, Hasanean et al. (2013) also discussed in detail the movements and behavior of the SbH Index (SbHI).

Hasanean et al. (2013) examined the teleconnection of SbH with wintertime surface-air temperatures (SATs) over KSA. They observed a deep connection between SAT and SbHI during the winter season over western and northeastern areas of Saudi Arabia. Cohen et al. (2001) also explained the influence of SbH in fluctuations of climate over the region of NH. The significance of SbH in determining the fluctuations of temperature and precipitation over the region of Eurasia has been deeply examined by Gong & Ho (2002). Likewise, Iqbal et al. (2013b) and Riaz et al. (2018) discussed that during the winter season, precipitation and temperature over most regions of Asia are significantly influenced by the strong Siberian anticyclone.

Rossby et al. (1939) introduced ‘center of action’ to be a considerable and noticeable semi-permanent low and high pressure system dominant over the contour map of mean sea level pressure (MSLP). These COA have a connection with the fluctuations in the atmospheric circulations on both global and regional scales. In this paper, some new parameters accountable for the change in wintertime precipitation for December, January, February and March (DJFM) over the regions of Arabian Peninsula, KSA in particular, have been observed using ‘COA’ methodology as proposed by Hameed et al. (1995). These include Icelandic Lows and Aleutian which are low pressure systems and Hawaiian High, Azores High and SbH which are high pressure systems. Rossby et al. (1939) pointed out a very significant point that the regional circulation is not only affected by the intensity of pressure, but it is also affected by the position of the COA. There are three indices in the COA approach, namely pressure index, longitudinal index and the latitudinal index. The pressure index is used to measure the anomaly of atmospheric mass over the study region, whereas the longitudinal index measures the pressure weighted mean longitudinal position of the center and latitudinal index measures the pressure weighted mean latitudinal position of the center.

Principal Component Analysis (PCA) is one of the widely used methodologies in atmospheric science research (e.g. Wallace et al. 1992; Dommenget & Latif 2002) that identifies the linear transformations of the dataset that concentrate as much of the variance as possible into a small number of variables. In the current literature, PCA has been successfully used in atmospheric sciences research (e.g. Wallace et al. 1992; Dommenget & Latif 2002; Arab Amiri & Mesgari 2016, 2017, 2019; Arab Amiri et al. 2017). Bretherton et al. (1992) on the other hand, suggested using Singular Value Decomposition Analysis for geophysical fields because it is easier to apply and involves simple matrix application. Moreover, Kar et al. (2017) suggested that the L-moments and other statistical information derived from the study can be useful for important hydrological design considerations. Furthermore, the methodology of the COA approach has also been widely used in the current literature (e.g. Bakalian et al. 2007; Hameed & Riemer 2012; Iqbal & Ilyas 2013; Iqbal et al. 2013a, 2013b; Riaz et al. 2018) because it provides better insight in explaining the climate variability (Usmani et al. 2019).

The KSA is one of the largest tourist places in the Middle East. The KSA has received abrupt and heavy rainfall in recent years so it is important to study the precipitation distribution and the factors that are affecting the precipitation of KSA. This study was carried out to study the possible effects of SbH and remote forcings of North Atlantic Oscillation (NAO) and ENSO on winter precipitation of KSA. Thus, the aim of this study was to study the possible impacts of SbH and remote forcings of NAO and ENSO on winter precipitation using the COA approach. We have used the COA approach because it focuses not only on the influence of intensity of the pressure but also the meridional and zonal movement of the pressure on climate variability. The explanation of winter precipitation variability by the remote forcings, namely NAO and ENSO, is weak compared to SbH indices. Therefore, an attempt has been made in this paper to show that SbH is the most dominant center of action over the kingdom as compared to the remote forcings.

DATA AND METHODOLOGY

Climate of KSA

The primary factors affecting the climate of KSA region is its geographical location. The subtropical latitude ranges from 16 to 32 °N and its location is very near to the circum global latitudinal belt of high pressure. The topography of the region can be understood by the Digital Elevation Model (DEM) as shown in Figure 1 (Al-Ahmadi & Al-Ahmadi 2013a). The geographical location of the kingdom extends over an area of 2,250,000 km2 and more than 90% of the land is covered with deserts and steppes (Alyamani & Sen 1993) making it one of the sunniest and hottest countries in the world. The nearby Persian Gulf and Red Sea are too narrow to fulfill the demand of water for the entire region. These limited water resources become too hot during the summer season which therefore restricts the drop in air temperature during the night making the nights very humid and uncomfortable. The minimum temperature during summer nights is 29–30 °C and it is very humid during the daytime.

Figure 1

Digital elevation model of KSA. Reproduced from Al-Ahmadi & Al-Ahmadi (2013a). Reprinted courtesy of the Copyright Holder under a Creative Commons Licence CC by 3.0 (https://creativecommons.org/licenses/by/3.0/).

Figure 1

Digital elevation model of KSA. Reproduced from Al-Ahmadi & Al-Ahmadi (2013a). Reprinted courtesy of the Copyright Holder under a Creative Commons Licence CC by 3.0 (https://creativecommons.org/licenses/by/3.0/).

The temperature of KSA is mainly controlled by the altitude and to a small extent by its proximity to the sea. Along the mountainous chain from the region of northwest to the region of southeast the temperatures are more comfortable and not so hot. To the western side of these mountains there is a very thin coastal plain of Red Sea and to the eastern side of the mountains there is a very high plateau that slowly drops down to a broader eastern coastal plain. Excluding the mountainous region, mostly the daytime temperature for the months of May to September are between 38 and 43 °C (or even several degrees higher) whereas in the mountainous region it is 30–32 °C. However, during the night the temperature falls, especially in the interior, whereas in the northwest mountains snow and frost sometimes occur in the winter season. Annual mean temperatures range from 25 °C in Makkah and Jizan and 30–31 °C in Dhahran.

The KSA receives most rainfall in boreal winter, which is a result of Mediterranean cyclones passing over the region. The annual gridded dataset of precipitation over KSA, which is spanned by 80% of the Arabian Peninsula, has been analyzed in this study. The rainfall over the region of the Arabian Peninsula is unpredictable and annual average totals are usually around 100 mm. The wettest area of the region is the far south west mountainous region of KSA, where during the spring and summer most of the rain is recorded, increasing the annual total to 141 mm in the region of Jizan and 199 mm in Khamis Mushait. The rainfall pattern in the northern part of Saudi Arabia is mainly during the months from November to April and is the result of a very weak weather system moving eastwards from North Africa or the Mediterranean. In the southern region of the country rain can occur in any season. Furthermore, Hasanean & Almazroui (2015) studied the total annual rainfall in most of the region of KSA which decreases with an exception of some increase in the southern side of KSA.

Sources of data

The monthly mean precipitation dataset used in the study was obtained from the Climate Research Unit (CRU), University of East Anglia, UK because this dataset has a resolution of 0.5 × 0.5° and is available from 1901 to 2016 (CRU TS 3.25, Harris et al. 2014). The monthly mean precipitation dataset was also obtained from National Centre for Environmental Prediction (NCEP) with a resolution of 1 × 1°. The high resolution data of mean sea-level pressure (MSLP) of resolution 2.5 × 2.5° have also been used in this study which was obtained from NCEP and has been available since 1948 (Kalnay et al. 1996). The sources form where the datasets were obtained and their latitude × longitude resolution are summarized in Table 1. We have also used NAO and Southern Oscillation Index (SOI) monthly indices which are available at the Climate Prediction Center (CPC), NCEP, USA.

Table 1

Data used, its sources and latitude × longitude resolution

Data Source Resolution 
Precipitation Climate Research Unit (CRU), University of East Anglia 0.5 × 0.5° 
Precipitation National Centre of Environmental Prediction (NCEP) 1 × 1° 
Sea level pressure National Centre of Environmental Prediction (NCEP) 2.5 × 2.5° 
500 mb winds National Centre of Environmental Prediction (NCEP) 2.5 × 2.5° 
Data Source Resolution 
Precipitation Climate Research Unit (CRU), University of East Anglia 0.5 × 0.5° 
Precipitation National Centre of Environmental Prediction (NCEP) 1 × 1° 
Sea level pressure National Centre of Environmental Prediction (NCEP) 2.5 × 2.5° 
500 mb winds National Centre of Environmental Prediction (NCEP) 2.5 × 2.5° 

Hameed et al. (1995) described how fluctuations in the COA are examined using SLP and this concept is used to compute objective indices of pressure, latitudinal and longitudinal positions of the semi-permanent low/high pressure systems. In this study, we made composite indices of the winter precipitation anomalies over KSA to observe interannual climatic fluctuations from 1981 to 2016.

Methods

Hameed & Riemer (2012) studied the phenomenon of movement of atmospheric centers that can be used to study fluctuations in the regional circulation. This phenomenon is used to observe the changes in high pressure systems (Iqbal et al. 2013a, 2013b). These changes are helpful in approximating its impact on variability of winter precipitation. Moreover, the impact of atmospheric pressure variability over the regions of KSA on winter precipitation can be successfully obtained through quantitative calculation of the oscillations in the pressure of SbH. These variations are useful in approximating its effect on wintertime precipitation variability.

Hameed et al. (1995) and Hameed & Riemer (2012) defined the pressure, latitudinal and longitudinal indices. The pressure index Ip of a high-pressure system is defined as an area-weighted pressure departure from a threshold value over the domain (I, J): 
formula
(1)
Similarly, the latitudinal index is defined as: 
formula
(2)
and the longitudinal index Iλ,Δt is constructed in the same manner. The reader is referred to Hameed et al. (1995) and Hameed & Riemer (2012) for details.

In the above equations, Pijt is the SLP value at grid point (i, j) averaged over a time interval Δt, Pt is the threshold SLP value and is equal to 1,015 mb for SbH, ϕij is the latitude of the grid point and λij is the longitude of the grid point . We use in these equations as test factors to guarantee that the difference in pressure is due to a high pressure system. If the value of the factor is positive, we take the value of δ to be one and if it is negative, we take the value of δ to be zero. Iqbal et al. (2013a, 2013b) observed that the intensity is thus the measurement of the anomaly of the atmospheric pressure. The boundaries of SbH and its threshold value, , are taken by examining its geographical ranges in NCEP reanalysis composite mean from the years 1948–2006, the selected domain for the SbH is thus taken as 20–55 °N and 50–170 °E. We define SbHI as mean SLP during winter (December–March) averaged over the domain of SbH. Area-averaged indices are usually more reliable and can provide more insight than single point indices, such as those used by Sahsamanoglou et al. (1991) and Mokhov & Petukhov (2000). This is because the area-averaged indices represent the variability of the COA rather than a single location only (Panagiotopoulos et al. 2005).

The COA method has been used in this study to observe the effect of SbH on the winter precipitation over the region of Saudi Arabia. We formulated the winter precipitation from December to March (DJFM). It has been observed that the study region receives more rainfall in the month of March. To examine the possible effect of SbH on the winter precipitation, the precipitation over the two regions are correlated with the different indices of SbH. Iqbal et al. (2013a, 2013b) observed that the changes explained by each of the indices are equally important in formulating regional variability. The significantly correlated indices are then correlated with each other to observe which indices are linearly independent. All the collinearities between the independent variables are not taken into consideration in order to identify the regions having a significant impact. In the next step, a linear model has been constructed by using multiple linear regressions between the winter rainfall over the study region and the independent indices.

For comparison purposes, we also employed PCA on a covariance matrix of gridded SLP data over the Siberian region. We then correlated leading Principal Components (PCs) with the two regions of KSA precipitation.

We plotted correlation maps of DJFM precipitation over KSA with the indices of SbH as shown in Figure 2. Figure 2 shows the correlation between DJFM precipitation and COA indices of SbH (Figure 2(a) and 2(b)) and SbHI (Figure 2(c)) at the 5% level. It is clear from Figure 2 that there are two regions of KSA that are significantly correlated with SbH indices, the southeast region and from northeast to central KSA. For this purpose, the KSA winter precipitation index is calculated by averaging DJFM total precipitation over the regions 26.25–48.75 °E, named region 1, and 37.50–42.50 °N, named region 2. Figure 2(c) shows the contour map of correlation between DJFM precipitation and SbHI. It can be seen from Figure 2(c) that DJFM precipitation has significant correlation with SbHI over the two regions of KSA at the 5% level.

Figure 2

Correlation map between DJFM precipitation and (a) SbHPS, (b) SbHLT and (c) SbHI. Solid black line represents regions where correlations are significant at 5% level and 1 and 2 represent the regions 1 and 2 of KSA respectively.

Figure 2

Correlation map between DJFM precipitation and (a) SbHPS, (b) SbHLT and (c) SbHI. Solid black line represents regions where correlations are significant at 5% level and 1 and 2 represent the regions 1 and 2 of KSA respectively.

We calculated winters when there were higher pressure values from its mean value by one standard deviation. These years are called high years and there were six high years. The six high years were 1984, 1985, 1996, 2005, 2006 and 2012. Similarly, low years were in winters when there were lower pressure values of SbH from its mean position. There were six low years and those were 1989, 1992, 1993, 1997, 2002 and 2007. We also determined the northward movement of SbH when the location of SbH was northern from its mean location by one standard deviation. These years are called SbH N years and they were seven years, namely 1985, 2001, 2006, 2009, 2010, 2012 and 2013. Similarly, in 1989, 1992, 1993 and 2002 the location of SbH was more southern from its mean location by one standard deviation and these years are called SbH S years.

RESULTS ON INTERANNUAL VARIATIONS IN WINTER PRECIPITATION

Interannual variations of KSA precipitation during the winter season, DJFM, have been considered in this paper. The COA indices, NAO and SOI have also been calculated for the DJFM season. We correlated the DJFM rainfall of the two regions of KSA with NAO, SOI and three indices of SbH, SbH pressure index (SbHPS), SbH latitudinal index (SbHLT), SbH longitudinal index (SbHLN), SbH index (SbHI) and the two leading PCs (PC1 and PC2) as shown in Table 2.

Table 2

Pearson's correlation coefficients between DJFM precipitation for KSA regions and COA variables (significant correlations are in bold)

  CRU observed data
 
NCEP/NCAR reanalyis
 
Parameters Region 1 De-trended Region 1 Region 2 Region 1 De-trended Region 1 Region 2 
Time −0.45 0.00 −0.28 −0.41 0.00 −0.26 
SbHPS −0.23 −0.25 −0.43 −0.19 −0.22 −0.46 
SbHLT −0.44 −0.35 −0.38 −0.48 −0.37 −0.42 
SbHLN −0.13 −0.16 0.17 −0.11 −0.15 0.20 
SbHI −0.47 −0.42 −0.44 −0.45 −0.48 −0.41 
PC1 0.42 0.33 0.49 0.46 0.29 0.51 
PC2 0.12 −0.02 −0.13 0.15 −0.1 −0.2 
NAO −0.004 −0.14 −0.04 −0.03 −0.19 −0.1 
SOI −0.34 −0.20 −0.28 −0.39 0.35 −0.21 
  CRU observed data
 
NCEP/NCAR reanalyis
 
Parameters Region 1 De-trended Region 1 Region 2 Region 1 De-trended Region 1 Region 2 
Time −0.45 0.00 −0.28 −0.41 0.00 −0.26 
SbHPS −0.23 −0.25 −0.43 −0.19 −0.22 −0.46 
SbHLT −0.44 −0.35 −0.38 −0.48 −0.37 −0.42 
SbHLN −0.13 −0.16 0.17 −0.11 −0.15 0.20 
SbHI −0.47 −0.42 −0.44 −0.45 −0.48 −0.41 
PC1 0.42 0.33 0.49 0.46 0.29 0.51 
PC2 0.12 −0.02 −0.13 0.15 −0.1 −0.2 
NAO −0.004 −0.14 −0.04 −0.03 −0.19 −0.1 
SOI −0.34 −0.20 −0.28 −0.39 0.35 −0.21 

PCA is employed over SLP over the Siberian region to construct the leading PCs which serve as indices of SbH. PCA is employed over the correlation matrix of SLP over the Siberian region. PC1 explains 65% of the total variance of the SLP which tells us the importance of the first leading PC. The second leading PC (PC2) explains only 14% of the total variance of the SLP which means that PC2 is very much less important than PC1. It can be seen from Table 2 that DJFM precipitation over region 1 is significantly correlated with SbHI (−0.47) and PC1 (0.42) at the 5% level, we also find that region 1 DJFM precipitation is possibly influenced by meridional movement of SbH and remote forcing of ENSO. The correlation of DJFM precipitation over region 1 with SbH latitude (−0.44) and SOI (−0.34) are statistically significant at 5%. There is also a statistically significant drying trend in region 1 precipitation, as shown in Figure 3. We therefore correlated again the de-trended DFM precipitation of region 1 with all of the above-mentioned indices as shown in the third column of Table 2. Thus, it can be seen from Table 2 that significant correlation with SbHLT is not due to the trend in region 1 precipitation since it is still significantly correlated with SbHLT in the absence of trend. This means that SbHLT has a possible influence on the variability of region 1 precipitation.

Figure 3

Time series comparison between (a) observed precipitation over region 1 and the regression model with SbHLT and SOI taken as independent variables, (b) observed precipitation over region 1 and the regression model with SbHI and SOI taken as independent variables and (c) observed precipitation over region 1 and the regression model with PC1 taken as independent variables.

Figure 3

Time series comparison between (a) observed precipitation over region 1 and the regression model with SbHLT and SOI taken as independent variables, (b) observed precipitation over region 1 and the regression model with SbHI and SOI taken as independent variables and (c) observed precipitation over region 1 and the regression model with PC1 taken as independent variables.

From Table 2 it can be seen that SbH latitude influences DJFM precipitation of region 1. The correlation coefficients between the indices is shown in Table 3. SbHLT and SOI are statistically uncorrelated (, as shown in Table 3. We thus constructed a multiple regression equation of region 1 precipitation with SbHLT and SOI which is given in Equation (3). The regression explains 26% of variance of precipitation during 1981–2016. The comparison between the observed precipitation over region 1 and the regression model is shown in Figure 3(a). Similarly, Table 3 suggests that SbHI and SOI are statistically independent of each other. Thus, the regression equation of DJFM precipitation over region 1 is constructed by taking SbHI and SOI as independent variables and is given in Equation (4). This equation explains 29% variability of DJFM precipitation over region 1, as shown in Figure 3(b). Furthermore, the first principal component of SLP over the Siberian region (PC1) is not statistically independent of any of the significant indices, as shown in Table 3. Thus, simple linear regression of region 1 precipitation with PC1 has been constructed which explains 18% variability of the precipitation over region 1. The comparison between the observed precipitation and the regression model is shown in Figure 3(c) and Equation (5). It can be seen from Figure 3 that region 1 received extreme precipitation of above 20 mm in 1997 as compared to other winters between 1981 and 2016. On the other hand, region 1 received less precipitation (around 5 mm) in 2009 as compared to other winters in the same period (Figure 2). 
formula
(3)
 
formula
(4)
 
formula
(5)
Table 3

Pearson's cross-correlations between the indices (significant correlations are in bold)

Indices SbHI PC1 PC2 SbHPS SbHLT SbHLN NAO SOI 
SbHI −0.93 0.024 0.79 0.87 −0.13 −0.28 0.14 
PC1 −0.93 −0.007 −0.81 −0.93 0.23 0.38 −0.34 
PC2 0.024 −0.007 0.53 −0.24 −0.55 0.35 0.09 
SbHPS 0.79 −0.81 0.53 0.61 −0.49 −0.15 0.26 
SbHLT 0.87 −0.93 −0.24 0.61 −0.08 −0.42 0.28 
SbHLN −0.13 0.23 −0.55 −0.49 −0.08 0.19 −0.29 
NAO −0.28 0.38 0.35 −0.15 −0.42 0.19 0.02 
SOI 0.14 −0.34 0.09 0.26 0.28 −0.29 0.02 
Indices SbHI PC1 PC2 SbHPS SbHLT SbHLN NAO SOI 
SbHI −0.93 0.024 0.79 0.87 −0.13 −0.28 0.14 
PC1 −0.93 −0.007 −0.81 −0.93 0.23 0.38 −0.34 
PC2 0.024 −0.007 0.53 −0.24 −0.55 0.35 0.09 
SbHPS 0.79 −0.81 0.53 0.61 −0.49 −0.15 0.26 
SbHLT 0.87 −0.93 −0.24 0.61 −0.08 −0.42 0.28 
SbHLN −0.13 0.23 −0.55 −0.49 −0.08 0.19 −0.29 
NAO −0.28 0.38 0.35 −0.15 −0.42 0.19 0.02 
SOI 0.14 −0.34 0.09 0.26 0.28 −0.29 0.02 
Correlation coefficients for DJFM precipitation over region 2 have also been computed with NAO, SOI, SbHI, the three COA indices of SbH, and the first two leading PCs (PC1 and PC2) of SLP over the Siberian region as shown in Table 2. It can be seen from Table 2 that there is no significant trend in region 2 precipitation. There is a significant correlation found between SbH pressure and region 2 precipitation . Also, the correlation between SbHLT and the region 2 precipitation is significant at 5%. The region 2 precipitation also has significant correlations with SbHI (−0.44) and PC1 (0.49). The correlation coefficient with SbHLT is lesser than other significant correlations. It can be seen from Table 3 that there is a significant correlation present between these significant indices which means that that all these significant indices are not statistically independent of each other. Thus, we have made simple linear regression equations with these significant indices. The regression equation of region 2 precipitation with SbHPS has also been constructed and is shown in Equation (6). Thus, SbHPS alone explains 18% of the variability of region 2 precipitation as shown in Figure 3(a). Similarly, the regression equation with SbHLT explains 15% of variability of region 2 precipitation and is given in Equation (7). The regression equations with SbHI and PC1 explain 20 and 24% of variance of region 2 precipitation respectively and are given in Equations (8) and (9) respectively. It can be seen from Figure 4 that region 2 received heavy precipitation of above 15–20 mm during 1987, 1997 and 2009. Similarly, region 2 received less precipitation (below 5 mm) during 1983, 1993, 2000 and 2011 (Figure 4). The time-series comparisons between the observed precipitation and the regression models are shown in Figure 4(a)–4(d). 
formula
(6)
 
formula
(7)
 
formula
(8)
 
formula
(9)
Figure 4

Time series comparison between (a) observed precipitation over region 2 and the regression model with SbHPS taken as independent variables, (b) observed precipitation over region 2 and the regression model with SbHLT taken as independent variables, (c) observed precipitation over region 2 and the regression model with SbHI taken as independent variables and (d) observed precipitation over region 2 and the regression model with PC1 taken as independent variables.

Figure 4

Time series comparison between (a) observed precipitation over region 2 and the regression model with SbHPS taken as independent variables, (b) observed precipitation over region 2 and the regression model with SbHLT taken as independent variables, (c) observed precipitation over region 2 and the regression model with SbHI taken as independent variables and (d) observed precipitation over region 2 and the regression model with PC1 taken as independent variables.

The NAO index is the most frequently used index to observe fluctuations in the atmospheric pressure over the North Atlantic Ocean (Hurrell et al. 2001). However, it has been observed that the remote forcing of NAO does not have a statistically significant contribution in the variability of winter precipitation over the regions of KSA.

From the calculations above, we infer that northward displacement of SbH corresponds to an increase in region 1 DJFM precipitation and vice-versa. Similarly, northward displacement of SbH and a decrease in intensity of SbH pressure corresponds to an increase in region 2 DJFM precipitation and vice-versa. To show that physical mechanisms are consistent with existing correlations, we constructed composite maps of monthly averaged fields of SLP, winds and precipitation using NCEP/NCAR reanalysis.

Figure 5 shows the mean SLP distribution during the DJFM season averaged for 1981–2016. The SbH can be clearly seen in the figure. SLP greater than 1,015 mb is located above 30 °N with high pressure of 1,030 mb at its center. Figure 6(a) shows the SLP anomaly during the DJFM season when pressure values are greater than its mean value by one standard deviation. We can see in Figure 6(a) that the high pressure system is extended over the map compared to the mean SLP distribution in Figure 5. Figure 6(b) shows the SLP anomaly during the DJFM season when pressure values over the Siberian region are less than its mean value by one standard deviation. We can see from Figure 6(b) that the high pressure system is reduced over the map. Similarly, Figure 7(a) shows the SLP anomaly during winters when the meridional movement of SbH towards the north was higher than its mean value by one standard deviation. It can be seen in Figure 7(a) that SbH is displaced towards the north, slightly above its mean position. In Figure 7(b), SLP anomaly during the winters when SbH was displaced towards south from its mean location by one standard deviation is shown. It can be seen from Figure 7(b) that the area of high pressure (greater than 1,015 mb) is smaller than that of its mean (Figure 5).

Figure 5

Mean SLP distribution for DJFM season from 1981 to 2016.

Figure 5

Mean SLP distribution for DJFM season from 1981 to 2016.

Figure 6

SLP distribution anomaly over Siberian region for: (a) high years and (b) low years.

Figure 6

SLP distribution anomaly over Siberian region for: (a) high years and (b) low years.

Figure 7

SLP distribution anomaly over Siberian region for: (a) SbH N years and (b) SbH S years.

Figure 7

SLP distribution anomaly over Siberian region for: (a) SbH N years and (b) SbH S years.

The mean wind at 500 mb over the KSA region during the DJFM season for 1981–2010 is displayed in Figure 8(a). The winds are westerly with speeds of over 14–15 ms–1 over KSA. The composites in Figure 8(b) and 8(c) show the impact of SbHPS and its latitudinal displacement on wind distribution over KSA. In Figure 8(b), the wind anomaly for the difference between high and low years is shown. It can be seen in Figure 8(b) that a strong anticyclone encompasses the whole KSA with wind speeds of about 8 ms–1 across the southern KSA below 24 °N, and wind speeds of about 4 ms–1 above 24 °N. Similarly, the wind anomaly for the difference between SbH N years and SbH S years is shown in Figure 8(c). The wind speed across the southern KSA below 28 °N is around 10 ms–1 whereas it is around 5 ms–1 above 28 °N.

Figure 8

Composite vector wind at 500 hPa: (a) climatological mean from 1981 to 2010, (b) vector wind difference for high years minus low years and (c) vector wind difference for SbH N years minus SbH S years.

Figure 8

Composite vector wind at 500 hPa: (a) climatological mean from 1981 to 2010, (b) vector wind difference for high years minus low years and (c) vector wind difference for SbH N years minus SbH S years.

The wind anomaly distributions in Figure 8(b) and 8(c) can be understood by the pressure distribution anomalies shown in Figure 9(a) and 9(b). Figure 9(a) shows an SLP anomaly for high years. The pressures over regions 1 and 2 are low. In Figure 9(b), the SLP anomaly for low years is shown and it can be seen that the pressures over regions 1 and 2 are high. The high and low pressure anomalies over the two regions under study for SbH N and SbH S years can be seen in Figure 9(c) and 9(d) respectively.

Figure 9

SLP anomaly maps for: (a) high years, (b) low years, (c) SbH N years and (d) SbH S years.

Figure 9

SLP anomaly maps for: (a) high years, (b) low years, (c) SbH N years and (d) SbH S years.

Figure 10(a) shows the climatological mean of DJFM precipitation over KSA from 1981 to 2010. Figure 10(b) shows the difference in precipitation for high and low years. The precipitation increased to 5–8 mm over the northwest and southeast regions of KSA. Similarly, Figure 10(c) shows the difference in precipitation for SbH N and SbH S years. The precipitation increased to around 5 mm over the central and southeastern KSA regions.

Figure 10

(a) Mean precipitation (measured in mm) from 1981 to 2010, (b) difference of mean precipitation for high and low years and (c) difference of mean precipitation for SbH N and SbH S years.

Figure 10

(a) Mean precipitation (measured in mm) from 1981 to 2010, (b) difference of mean precipitation for high and low years and (c) difference of mean precipitation for SbH N and SbH S years.

CONCLUSIONS

This study examines the relationship between the precipitation of KSA during winters with SbH which is the dominant center of action over NH during winters. The methodology employed in our study is the COA approach. Correlation maps between the DJFM precipitation and the three indices of SbH show that there are only two regions of KSA that are significantly influenced by SbH, named region 1 and region 2. We calculated the average precipitation from December to March (DJFM) for the period 1981–2016. The correlation coefficients between the winter precipitation over region 1 and the three indices of SbH show that the poleward displacement of SbH is more responsible for the precipitation variability over region 1 than the other parameters. Similarly, the correlation coefficients between the winter precipitation over region 2 and the SbH pressure and latitudinal indices are statistically significant at 5% but the correlation with SbH pressure is more than that of its latitudinal index. There is also a weak correlation present with NAO for the two regions under study. However, region 1 precipitation is significantly correlated with ENSO in the presence of a trend and the correlation became weak in the absence of the trend.

These empirical results show that by considering the intensities and positions of the SbH, the primary influence on interannual variation of winter precipitation over the two regions under study is the intensity and the meridional position of SbH. Thus, in winters when the intensity of SbH is more than its mean, there is more precipitation observed over KSA and the opposite anomaly is experienced when the intensity of SbH is less than its mean. Similarly, when SbH is displaced towards the north, KSA receives more precipitation and when it is displaced towards the south, KSA receives less precipitation. Thus, by employing the COA approach, we conclude that the SbH contributes in the interannual variation of the winter precipitation over the KSA regions.

REFERENCES

REFERENCES
Al-Ahmadi
K.
&
Al-Ahmadi
S.
2013a
Rainfall-altitude relationship in Saudi Arabia
.
Adv. Meteorol.
2013
,
1
14
.
Almazroui
M.
,
Islam
M. N.
,
Athar
H.
,
Jones
P.
&
Rahman
M. A.
2012a
Recent climate change in the Arabian Peninsula: annual rainfall and temperature analysis of Saudi Arabia for 1978–2009
.
Int. J. Climatol.
32
,
953
966
.
Arab Amiri
M.
&
Mesgari
M.
2016
Spatial variability analysis of precipitation in northwest Iran
.
Arab. J. Geosci.
9
(
11
),
1
10
.
Arab Amiri
M.
&
Mesgari
M.
2017
Modeling the spatial and temporal variability of precipitation in northwest Iran
.
Atmosphere
8
(
12
),
1
14
.
Arab Amiri
M.
&
Mesgari
M.
2018
Analyzing the spatial variabilty of precipitation extremes along longitude and latitude, northwest Iran
.
Kuwait J. Sci.
45
(
1
),
121
127
.
Arab Amiri
M.
,
Mesgari
M.
&
Conoscenti
C.
2017
Detection of homogeneous precipitation regions at seasonal and annual time scales, northwest Iran
.
J. Water Clim. Change
8
(
4
),
701
714
.
Arab Amiri
M.
,
Conoscenti
C.
&
Mesgari
M.
2018
Improving the accuracy of rainfall prediction using a regionalization approach and neural networks
.
Kuwait J. Sci.
45
(
4
),
66
75
.
Babu
C.
,
Samah
A.
&
Varikoden
H.
2011
Rainfall climatology over Middle East region and its variabilty
.
Int. J. Water Resour. Arid Environ.
1
,
180
192
.
Bretherton
C. S.
,
Smith
C.
&
Wallace
J. M.
1992
An intercomparison of methods for finding coupled patterns in the climate data
.
J. Clim.
5
,
541
560
.
Bruce
J. P.
1994
Natural disaster reduction and global change
.
Bull. Am. Meteorol. Soc.
75
(
10
),
1831
1835
.
Chowdhury
A. K.
,
Kar
K. K.
,
Shahid
S.
,
Chowdhury
R.
&
Rashid
M. M.
2019
Evaluation of spatio-temporal rainfall variability and performance of a stochastic rainfall model in Bangladesh
.
Int. J. Climatol.
39
,
4256
4273
.
Cohen
J.
,
Saito
K.
&
Entekhabi
D.
2001
The role of the Siberian High in Northern Hemisphere climate variability
.
Geophys. Res. Lett.
28
(
2
),
299
302
.
Dommenget
D.
&
Latif
M.
2002
A cautionary note on the interpretation of EOFs
.
J. Clim.
15
,
216
225
.
Hameed
S.
,
Shi
W.
,
Boyle
J.
&
Santer
B.
1995
Investigation of the centers of action in the North Atlantic and North Pacific in the ECHAM AMIP simulation
. In:
1st International AMIP Scientific Conference Monterey
,
WRCP 92
,
California
, p.
221
.
Harris
I.
,
Jones
P.
,
Osborn
T.
&
Lister
D.
2014
Updated high-resolution grids of monthly climatic observations- the CR TS 3.10 dataset
.
Int. J. Climatol.
34
,
623
642
.
Hurrell
J. W.
,
Kushnir
Y.
&
Visbeck
M.
2001
The north Atlantic oscillation
.
Science
291
(
5504
),
603
605
.
Iqbal
M. J.
,
Hameed
S.
&
Khan
F.
2013a
Influence of Azores high pressure on Middle Eastern rainfall
.
Theor. Appl. Climatol
111
,
211
221
.
Iqbal
M. J.
,
Riaz
S. M.
&
Ghauri
B. M.
2013b
Impact of Siberian high on rainfall variability over northern part of Indo-Pak region
.
Arab. J. Geosci.
6
,
3087
3092
.
Jeong
J.-H.
,
Ou
T.
,
Linderholm
H. W.
,
Kim
B.-M.
,
Kim
S.-J.
,
Kug
J.-S.
&
Chen
D.
2011
Recent recovery of Siberian high intensity
.
J. Geophys. Res.
116
,
1
9
.
Kalnay
E.
,
Kanamitsu
M.
,
Kistler
R.
,
Collins
W.
,
Deaven
D.
,
Gandin
L.
,
Iredell
M.
,
Saha
S.
,
White
G.
,
Woollen
J.
,
Zhu
Y.
,
Chelliah
M.
,
Ebisuzaki
W.
,
Higgins
W.
,
Janowiak
J.
,
Mo
K. C.
,
Ropelewski
C.
,
Wang
J.
,
Leetmaa
A.
,
Reynolds
R.
,
Jenne
R.
&
Joseph
D.
1996
The NCEP/NCAR 40-year reanalysis project
.
Bull. Am. Meteorol. Soc.
77
(
3
),
437
472
.
Kar
K. K.
,
Yang
S. K.
,
Lee
J. H.
&
Khadim
F. K.
2017
Regional frequency analysis for consecutive hour rainfall using L-moments approach in Jeju Island, Korea
.
Geoenviron. Disasters
18
,
1
13
.
Lydolf
P.
1977
Climates of the Soviet Union
.
Elsevier
,
Amsterdam
, p.
443
.
Mokhov
I.
&
Petukhov
V.
2000
Atmospheric centers of action and tendencies of their change
.
Izvest. Atmos. Ocean. Phys.
36
,
292
299
.
Panagiotopoulos
F.
,
Shahgedanova
M.
,
Hannachi
A.
&
Stephenson
D. B.
2005
Observed trends and teleconnections of the Siberian high: a recently declining center of action
.
J. Clim.
18
,
1411
1422
.
Sahsamanoglou
H.
,
Makrogiannis
T.
&
Kallimopoulos
P.
1991
Some aspects of the basic characteristics of the Siberian anticyclone
.
Int. J. Climatol.
11
,
827
839
.
Usmani
B. A.
,
Rehman
S.
,
Hussain
S. A.
,
Khan
K.
&
Elahi
A.
2019
Regional precipitation variability associated with zonal movement of Indian Ocean high pressure system
.
Asia Pac. J. Atmos. Sci
.
https://doi.org/10.1007/s13143-019-00119-8
.