Spatial Mapping of Atmospheric Precipitation Isotopes in Syria

Stable isotope ratios (δ2H and δ18O), tritium (3H). and deuterium excess (d-excess) values of atmospheric precipitation (P) at 16 stations were determined for mapping the spatial variation of oxygen and hydrogen P isotopes in Syria. The major geographical parameters (longitude east, LE; latitude north, LN; altitude, H; and inland distance from the coast, DFC) were used to find out the best fitting models for the spatial mapping of atmospheric P isotopes in Syria. The highest correlation coefficients (r>0.73) were found for the relationships δ18O-H and δ2H-H. The impacts of LE and LN factors were rather moderate (0.30.7) was found for the relationship d-excess-LN. The increase of d-excess value from 23‰ to 24‰ in southern Syria is due to the Mediterranean Sea (MS) moisture, likely as a high percentage (>65%) of prevailed winds usually comes over the MS. The concentrations of 3H in P samples(4-14TU) during the period 1989–1993 were higher than those (3-10TU) for the period 2004–2006, indicating thus a return back toward the levels of typical 3H production in nature. The high correlation coefficients (r>0.59) that were found for the relationships 3H-DFC and 3H-LE, suggest a continuous exchange between the low tritium moisture from the MS and the higher tritium moisture from the inland areas. Produced gridded isotopic values are quite satisfactory for covering certain areas in Syria and the neighboring Arab countries.


INTRODUCTION
Determination of the environmental isotopes ( 2 H, 3 H and 18 O) in atmospheric precipitation (P) has been proven to be a useful tool in many hydrological, hydrogeological, meteorological and climate change studies (Criss, 1999;Kendall and McDonnell, 1998;Clark and Fritz, 1997;Fritz and Fontes, 1980).The importance of such isotopes in the above mentioned domains is mainly due to the fact that all these atoms enter in the formulation of the water molecule, and thus they can remarkably respond to all modifications occurring within the water cycle, but most frequently unaffected by geological processes (Mook, 2001;Gat, 1980).The stable isotopes ( 2 H and 18 O) are both conservative constituents, thus they are classified as the most ideal natural tracers, especially for assessing the hydrological processes in relation to condensation, evaporation and mixing betw een the different water sys-tems (Liu et al., 2014;Aggar wal et al., 2005;Cappa et al., 2003;Mook, 2001;Gat, 1996).
Tritium ( 3 H) as the unique radioactive isotope of hydro gen, incorporated into the natural water cycle either naturally or artificially (Michel et al., 2015;Mook, 2001;Eriksson, 1983), is also a powerful radiotracer.This isotope, which has a half-life of 12.32 years (Lucas and Unter weger, 2000), was largely and successively applied in many hydrological studies, mostly as a dating tool for solving problems related to determination of the transit time of young surface and groundwaters (Fontes, 1983;Yurt sever, 1983), or tracing the water flow paths in small catchments (Herrmann et al., 1990;Maloszewski et al., 1983).
Based on the importance of such environmental isotopes in the diverse hydrological domains, a distinct international monitoring network, the so-called: "Global Network for Isotopes in Precipitation (GNIP)" was initiated in the early 1960s through a bilateral cooperation between the International Atomic Energy Agency (IAEA) and the World Meteorological Organization (IAEA- WMO, 2006).Since then a large number of studies all over the world were started, mostly for analyzing the temporal and spatial variation trends of P isotopes either on Global scales (Terzer et al., 2013;Bowen et al., 2005;Dutton et al., 2005;Schmidt et al., 2005;Bowen and Revenaugh, 2003;Bowen and Wilkinson, 2002;Rozanski et al., 1993;Dansgaard, 1964), or on smaller regional scales (Lykoudis et al., 2010;Lykoudis and Argiriou, 2007;Dutton et al., 2005;Sturm et al., 2005;Longinelli and Selmo, 2003).
Analysis of the spatial variability patterns of the stable isotopes in meteoric water globally has shown that the changes in the isotopic composition are usually because of fractionation of oxygen and hydrogen isotopes between liquid and vapor phases, likely as humid air masses move from one place to another (Bowen et al., 2012;Aggar wal et al., 2004;Kendall and Coplen, 2001;Araguás-Araguás et al., 2000;Rozanski et al., 1993).
Information on the variation patterns of P isotopes in the eastern Mediterranean areas is mostly referred to the datasets of the Bet-Dagan station, disposing the longest isotopic records in this region (IAEA-WMO, 2006).
Although a number of previous isotopic studies were devo ted to characterizing the environmental P isotopes in Syria (Al-Charideh and Abou Zakhem, 2010;Abou Zakhem and Hafez, 2007;Kattan, 1997b) and in some of the neighboring Arab countries (Bajjali, 2012;Saad et al., 2005), the available data on P isotopes in this region is regrettably still limited, likely because of the short monitoring periods of these investigations.How ever, Kattan (2019) recently characterized the factors that control the stable isotope compositions of P in Syria.
In this paper the objectives are: (1) to explore the impact of the key geographical parameters, such as the longitude east (LE, degrees), latitude north (LN, degrees), altitude (H, m above sea level (a.s.l)), and inland distance from the coast (DFC, km), on the induced spatial changes in the isotopic composition of P in Syria; and (2) to provide further insights on the spatial simulation of P isotopes in Syria.
Constituently, this work will promote the better understanding of the role of the geographical factors that can affect the spatial variations of P isotopes in Syria and certain areas in the neighboring eastern Mediterranean coun tries, such as Lebanon and northern Jordan.This study, which may help to produce new theoretical isotope knowledge that could be considered as a valid basis for covering certain areas in this region that lack similar determinations, is definitely important to apply properly these natural isotopes in the diverse hydrological, hydrogeological and climate change studies in Syria.

STUDY AREA
Syria is located at the eastern edges of the MS.From a topographical point of view, this country is dominated by the following structures: the coastal plain (width <10 km) and the coastal mountains (H up to 1575 m a.s.l), expanding in parallel to the MS; the Mt Hermon (2814 m a.s.l) and Mt Anti-Lebanon (2466 m a.s.l), both located on the border with Lebanon; the interior plains and interior mountain series of the southern and northern Pal myrides (1308 m a.s.l), situated in central Syria; the Mt Al-Arab (1790 m a.s.l), located in southern Syria; and the Syrian desert region, extending near the borders with Iraq and Jordan (Fig. 1a).
Climatically, Syria is particularly influenced by a Mediterranean type of climate, with mild wet winters, and long, hot and mostly dry summers.The air temperature (T) in the coastal areas usually ranges in winter between -3°C and 18°C, while in summer it is frequently higher than 22°C.The variation in air T value is generally small, likely because of the small size of the country.The annual average of air T value in the farther inland areas is usually >18°C, with a quite hot desert climate in south-eastern Syria (www.weatheronline.co.uk).
The highest annual P-amounts (>1000 mm) mostly occur over the coastal mountains.The rainfalls over the interior mountains are rather high (400-800 mm), while comparatively lower than 200 mm around the Damascus region.In the farther inland desert areas of south-eastern Syria, the annual P-amounts are usually in the range of 25-125 mm (Fig. 1b).The rainy season in Syria is rather short (from October to May), with an average number of 35-92 for the frequent rainy days in this country (www.climate-data.org).
The relative air humidity (RH) value in the coastal areas in winter is mostly around 60-70%, while it is frequently higher than 70-80% in summer.In the inland areas, the RH is also high in winter (60-80%), but very much lower (20-50%) in summer (www.weatheronline.co.uk).
Recent analysis of the backward trajectory tracks of air masses producing rains in Syria revealed that the genesis of most rainy events were from the following sources: the Atlantic Ocean, via western Europe and the MS; the Northern Pole, via eastern Europe; the Siberian plateau, via Turkey and eastern Europe; North Africa, via the MS; and the Arabian Sea, via the Arabian Peninsula and the Red Sea (Kattan, 2019).

1 Rainfall Sampling
A total of 290 rainfall samples were collected from a network covering 16 stations, mainly distributed in western Syria (Fig. 1b).Table 1 summarizes the number (n) of RW samples collected from each site, the geographical coordinates of the sampling stations (LE, LN, H and DFC), and the average air T, P-amount, and RH values measured at each station during the period 1989-2006.Rainfall sampling was made on a monthly basis (monthly composite), with systematic collection of RW samples twice a day (8 AM and 8 PM), by using suitable plastic containers with catch-diameters of 30-80 cm, according to the P-amount falling over the considered station.The   daily collected water from each station was then moved into a tightly-sealed container, large enough to store the cumulated amounts of monthly P.After returning back from the field to the laboratory, all collected samples were instantly stored in a refrigerated room (T below 5°C) until the time of analysis.More details on the sampling procedures were given by Kattan (2019).

2 Isotopic Analyses
At the end of each month in the rainy season (Nov-May) two rinsed bottles (polyethylene or glass) were filled for conducting the isotopic determinations.A first small bottle of 50 mL was filled for the determination of δ 18 O and δ 2 H, while the second bottle (one liter size) was filled for tritium ( 3 H) determination.The subsequent analyses of δ 18 O and δ 2 H in the RW samples were performed by using a Finnigan Mat (Delta E or Delta Plus) mass spectrometer.Measurement accuracy for δ 2 H and δ 18 O are better than ±0.1‰ and ±1.0‰, respectively.The subsequent analyses of tritium in the RW samples were performed, after electrolytic enrichment of the water samples, either by using a Packard-3253 liquid scintillation counter or by an ultra low level liquid scintillation counter (Quantulus 1220).Detection limit for 3 H measurements in both instruments is lower than 1 TU (1 TU = 1 3 H atom per 10 18 1 H atoms).More details on the analytical procedures of these isotopes in water samples were given by Kattan (2006 and2019).
Analytical results of the stable isotope values are commonly expressed in δ notation in terms per mil (‰), defined as: where R is the isotopic ratio ( 2 H/ 1 H or 18 O/ 16 O), and the indices sam and std denote the sample and the standard, respectively.The international standard largely used for the calibration of 2 H and 18 O isotopes in RW samples is called "Vienna Standard Mean Ocean Water (VSMOW)", precisely prepared and adopted by the IAEA for this purpose (Coplen, 1996;Gonfiantini, 1978).Specific req uirements and quality assurance procedures according to the ISO/IEC 17025 standard are strictly applied in all the labs of the Atomic Energy Commission of Syria (AECS) in order to achieve the production of competent analysis results.Verification of the isotopic analysis results are periodically insured through the participation in the different analysis comparison programs organized either by the IAEA or by other regional laboratories.

3 Statistical Analysis and Calculations
Descriptive statistics (average and standard deviation) for the entire dataset, together with those for the RW samples collected from each station were calculated by using the softwares Excel 2007 and Statview 5.The amount-weighted mean value (C w ), that takes into consideration the effect of rainfall intensity during any monitoring period, was calculated by using the following formula (Yurtsever and Gat, 1981): where P i is the monthly P-amount and C i is the isotopic composition of RW sample for the month i.
The deuterium excess value (d-excess or d) in each RW sample was calculated by using the formula, defined by Dansgaard (1964) as: As the accurate modeling of the spatial distribution of P isotopes requires at least a gridded network of about 10 × 10 unit scale (Lykoudis and Argiriou, 2007), the map sheet of Syria and some parts of the neighboring countries was divided into 4590 grid-units.The software Arc-GIS-9 was first used to generate the values of the geographical coordinates (LE, LN and H) for each of the grid-units, and then to produce the simulated maps for the spatial distributions of P stable isotopes (δ 18 O and δ 2 H), tritium and d-excess values across Syria.

1 Stable Isotopes in Precipitation
Table 1 illustrates the average stable isotope (δ 18 O and δ 2 H) ratios and d-excess values, compared to the amountweighted mean values calculated for the RW sam ples collected at all the stations.The stable isotope ratios in the RW samples were in the range from -14.6‰ to + 7.4‰ and from -101‰ to + 35‰, with average values of -6 ±3‰ and -32±20‰ for δ 18 O and δ 2 H, respectively (Kattan, 2019).The d-excess value for these dataset varied from ≤0‰ in the highly evaporated samples to + 30‰ in the RW samples collected from southern Syria.The amount-weighted mean δ 18 O and δ 2 H values were always lower than the average values at all the stations, likely as heavy stable isotopes are almost more concentrated in the evaporated RW samples (P-amounts<5 mm).The stable isotope ratios in the RW samples at the Arneh and Bloudan stations (H>1500 m a.s.l) were markedly the most depleted (δ 18 O<-6.5‰and δ 2 H<-35‰), while those at the Tartous, Homs and Palmyra stations (H<400 m a.s.l) were distinctly higher (δ 18 O>-5.7‰and δ 2 H>-32‰), mostly because of the evaporation of raindrops during their falling on the earth surface (Managave et al., 2016).
The relationship δ 2 H-δ 18 O for the entire RW samples (Fig. 2a) clearly shows that most of the sample points are scattered between the Global Meteoric Water Line (GMWL), defined by Craig (1961) as: δ 2 H = 8 δ 18 O + 10, and the Mediterranean Meteoric Water Line (MM WL), given by Nir (1967), as: δ 2 H = 8 δ 18 O + 22.However, a few sample points, with depleted stable isotope ratios, are regrouped close to the MMWL.The second part of the samples, with enriched isotopic ratios (δ 18 O> -4‰ and δ 2 H>-20‰), are located very close to the GMWL.The equation of the least squares regression line fitting all the sample points is given as: This distribution reflects not only the impacts of the T effect (Kattan, 2019;Rozanski et al., 1993) and secondary evaporation during the rain falling on the land surface (Managave et al., 2016;Froehlich et al., 2005;Araguás-Araguás et al., 2000), but also the exchange with the enriched moisture (d>22‰) produced by evaporation from the MS (Gat et al., 2003).

2 Tritium in Precipitation
The concentrations of 3 H in the RW samples during the period 1989-1993 were in the range of 4-14 TU, while during the period 2004-2006 they were in the range of 3-10 TU.This change, which means that the concentration of 3 H is annually decreasing by ≈5-6%, clearly suggests a return back toward the natural cosmogenic production levels in the upper atmosphere before starting the nuclear weapons testing (Michel et al., 2015).Also, it can be observed that the amount-weighted mean 3 H values in the RW samples at all the stations were lower than the average 3 H values, with a difference of 0.2-1 TU (Table 1).The concentrations of 3 H in the RW samples at the coastal station (Tartous) and Al-Kounietra and Qatana stations, located nearly close to the MS, were remarkably <5 TU, while at the Palmyra station, located  at a farther distance from the MS coast, were higher than 9 TU.The amount-weighted mean 3 H values in the RW samples at the Al-Suwieda and Izraa stations were relatively higher than 5 TU, most probably because of the short monitoring period (1989)(1990)(1991)(1992)(1993).The measured 3 H concentrations in the RW samples at both stations during that period (8-9 TU) were clearly very close to those reported for the Damascus and Aleppo stations that have nearly a similar DFC value (Kattan, 1997b).
The temporal variation of average monthly 3 H concentrations in the RW samples during the rainy season (Fig. 3), clearly shows a progressive increase from ≈3 TU in November up to ≈12 TU during April to May.This trend, which indicates that more than 2/3 of the total 3 H load was occurred during the spring season, clearly reflects the tropospheric injection, or the so-called "Spring Leak" phenomenon (Harms et al., 2016;Rozanski et al., 1991).

3 Meteorological Controls on Stable Isotopes
Variations of the stable isotope ratios in meteoric water are known to be highly influenced by the ambient local weather conditions, specifically the common meteorological (T, P-amount and RH) parameters (Araguás- Araguás et al., 1998;Rozanski et al., 1993;Dansgaard, 1964).Pressure of the water vapor in the atmosphere, air depressions, weather cold fronts and backward trajectory tracks of humid air masses are also decisive factors that play certain roles in the formulation of the isotopic signature of meteoric water (Li et al., 2015;Muller et al., 2015;Baldini et al., 2010;IAEA, 2005).
Recently it was shown that the T effect was the major factor that strongly controls the temporal variation of P stable isotopes in Syria (Kattan, 2019).This effect was remarkably reflected by enrichments of ≈0.34‰/°C and ≈1.9‰/°C for δ 18 O and δ 2 H, respectively, accompanied by a decrease of ≈-0.8‰/°C for the d-excess value with T increasing.Fig. 4a shows the relationship between the monthly δ 18 O values and the average monthly T values   for all the RW samples collected in Syria.
The effect of P-amount is also an important factor.This effect, which highly depends on the local weather conditions during the time of rain falling, was reflected by a significant decline in the stable isotope ratios (Fig. 4b), accompanied by a remarkable increase in the d-excess value with the increasing of monthly P-amount.However, it can be observed that the gradient in d-excess value in the case of the mountainous stations located close to the MS was generally small, while it was remarkably much higher in the remaining stations.
The RH effect was reflected by a decrease in the stable isotope ratios, accompanied by a comparative enrichment in the d-excess value with RH increasing.Fig. 4c shows that the low d-excess values were usually associated with low RH values (≈40%), whereas the higher d-excess values were typically coupled with high RH values (>80%).

4 Geographical Controls on Precipitation
Isotopes The common geographical factors that have strong controls on the spatial variability of stable isotopes in meteoric water are the altitude, latitude and longitude coordinates (Wang et al., 2016;Terzer et al., 2013;Lykou -dis et al., 2010;Bowen et al., 2005).The DFC factor may also affect the spatial variation of P isotopes, likely the con centration of 3 H, that tends to increase with the increasing of the inland distance (Harms et al., 2016;Michel et al., 2015;Kattan, 1997b;Michel, 1989;Gat and Carmi, 1970).Table 2 summarizes the regression parameters of the fitted straight lines and second-order polynomial curves for the different relationships between the amount weighted-mean δ 18 O, δ 2 H, d-excess and 3 H values in the RW samples at all the stations in function with the corresponding values of LE, LN, H and DFC parameters.Accordingly, it can be observed that the correlation coefficients for the fitted second-order polynomial curves are almost similar to those of the fitted straight lines.The high correlation coefficients (r> 0.73), that were found for the relationships δ 18 O-H and δ 2 H-H, strongly indicate that the so-called "altitude effect" was the most important factor controlling the spatial variability of both stable isotopes.The negative slope values for both relationships also show that this effect was shown up by progressive depletion gradients of -0.002±4E-4‰ and -0.009±0.002‰for δ 18 O and δ 2 H, respectively (Fig. 5a).
Although the altitude factor was with strong effect on the spatial variation of P stable isotopes, its impact on the spatial variation of the d-excess parameter was rather moderate (r≈0.5-0.52).The correlation coefficients for the relationships between the stable isotope ratios (δ 18 O and δ 2 H) and the geographical coordinates (LE and LN) were found also to be rather moderate (0.3<r<0.6), meaning that these factors have minor controls on the spatial variation of P stable isotopes, most likely because of the small geographical extent of the country (Fig. 5b).Similarly, it was observed that the correlation coefficients for the relationships (δ 18 O-DFC and δ 2 H-DFC) were also low (0.3<r<0.4).The reason is because the MS coast is typically extending in parallel to the global longitudinal lines (LE), and thus the impact of the DFC factor must be rather similar to that of the LE factor.However, it can be observed that the correlation coefficients for the relationship d-excess-LN for the linear and the second-order polynomial functions (Fig. 5c) were remarkably high (0.7<r<0.8 and P-value<0.001).The reason that the correlation coefficients for this relationship were rem arkably high, and with a negative slope value (-1.39±0.3), is clearly due to the extensive exchange with the enriched MS moisture (d≈22‰), likely as a high percentage of prevalent winds (>50%) usually coming from the directions of SW and S. A low correlation coefficient (r≈0.26) was found for the relationship d-excess-DFC, suggesting also a weak impact of the DFC factor on the spatial variability of dexcess value.
In the case of 3 H spatial variation, the best correlation coefficient (r>0.63) was found for the relationship 3 H-DFC (Fig. 5c).This later coefficient, which was rather close to the coefficient (r>0.59)calculated for the relationship 3 H-LE, clearly indicates that both factors have somewhat similar controls on the spatial variability  of 3 H in Syria.The correlation coefficients for the relationships 3 H-LN and 3 H-H were both low (r<0.3),indicating thus weak influences of the H and LN factors on the spatial variation of this radio isotope.

5 Modeling the Spatial Isotope Variability
Modeling the spatial variation of P isotopes is a smart alternative tool for mapping the approximate isotopic composition of P at any location on the Earth Globe (Bowen et al., 2005).Hence, extensive efforts were made to fit between the P isotopic data with the key geographical parameters (LE, LN and H), either on global scales (Bowen et al., 2005;Bowen and Revenaugh, 2003;Bowen and Wilkinson, 2002), or on regional scales (Lykoudis et al., 2010;Liu et al., 2010;Lykoudis and Argiriou, 2007;Longinelli and Selmo, 2003).However, some other works were focused on the development of certain algorithms and models that can determine, not only the average annual isotopic composition of P at a certain site, but also the seasonal, or even the daily isotopic values (Sturm et al., 2005).
In this study, attempts were made to simulate the spatial variation of P isotopes in Syria by using the amount weighted-mean δ 18 O, δ 2 H, 3 H, and d-excess values in the RW samples collected at the different stations in this country, in conjunction with the corresponding key geographical parameter (LE, LN and H) values.The purpose behind such essays was to extract mathematically the appropriate empirical relationships that fit between the measured isotopic values with the corresponding geographical parameters.The final aim was to produce a highresolution gridded isotopic data that can be used later for covering certain areas lacking similar isotopic data.Such simulated data can also be used to map the spatial distributions of P isotopes in this country, by using the traditional geographical information system (GIS), largely applied in this domain (Johnston et al., 2004).
By excluding the isotopic data of the Al-Nabk station,  The correlation coefficients for the fitted polynomial models vary between r = 0.67 in the case of 3 H and r = 0.89 in the case of δ 18 O.These coefficients are generally good, and thus they can safely be used to generate the isotopic data for the different gridded-units or cells for the selected sheet map of Syria.
Fig. 6 illustrates the relationships that were found between the calculated (simulated) and the observed (mea sured) amount weighted-mean δ 18 O, δ 2 H, 3 H and d-excess values in P at the different Syrian stations.Accordingly, the above mentioned models were all used to generate the appropriate isotopic data for the gridded cells of the Syrian map, by assuming that the cell centre is representative for the simulated isotopic values.
The modeled spatial distributions for the stable isotopes (δ 18 O and δ 2 H) imply that their values change remarkably in accordance with the topographical features in the country (Figs.7 and 8).This distribution highly reflects the significant role of the altitude factor on the spatial variability of both isotopes, with clearly depleted isotopic ratios (δ 18 O<-7.5‰and δ 2 H<-45‰) in the high altitude areas (Mts Hermon, Anti-Lebanon and Al-Arab), and enriched isotopic ratios (δ 18 O>-5.5‰and δ 2 H>-33‰) in the interior inland areas, most specifically in the low altitude areas in north-eastern Syria.However, it is noteworthy to observe that the produced map for the distribution of d-excess values in P in this country is rather of a particular spatial variation trend (Fig. 9), with a remarkable high d-excess value (d>20‰) in southern and south-western Syria (Mts Al-Arab, Hermon and Anti-Lebanon), and gradually lower d-excess value (d<20‰) in the eastern and north-eastern parts of Syria.
Although the correlation coefficient for the relationship between predicted and measured amount weightedmean 3 H values in P at the different stations (r = 0.67) was lower than those found for the stable isotopes and d-excess parameters (r>0.84), the simulated map for the spatial variability of this radioisotope was also remarkable (Fig. 10).This map permits to observe a progressive build up or 3 H enrichment with the increasing distance from the Syrian coast, from a value of about 5 TU along  the MS coast to values higher than 9 TU in the farther inland areas in the eastern and north-eastern parts of the country.This distribution clearly explains the continuous mixing between the low tritium of the MS moisture and the higher tritium of the continental moisture (Michel et al., 2015;Mook, 2001).

CONCLUSIONS
This paper explores the impacts of the key geographical parameters (LE, LN, H and DFC) that supposed to have certain controls on the spatial variations of atmospheric P isotopes in Syria.The major conclusions that can be drawn from this particular study are the following: • The stable isotope ratios in the RW samples at the high mountainous regions (H>1500 m a.s.l), such as the Mts Hermon, Anti-Lebanon and Al-Arab, were generally the most depleted, while those at the low elevation areas, such as the coastal station (Tartous) and the farther inland stations ( Jarablous and Palmyra) were the most enriched, likely because of the secondary evaporation during rain falling.The isotopic composition of rainfalls at the stations of in-between elevations (Aleppo, Damascus, Qatana, Homs and Al-Raqqa), were clearly of intermediate values between the above mentioned two groups.• The high correlation coefficients (r>0.73), that were found for the relationships (δ 18 O-H and δ 2 H-H), clearly indicate that the T effect, or indirectly the H factor, was the primary cause for the vertical spatial variation of both stable isotopes.• The rather moderate correlation coefficient (r<0.52), that was found for the relationship d-excess-H, clearly suggests that the d-excess parameter was more dependent on the initial isotopic composition of the water vapor and kinetic isotope fractionation during secondary evaporation rather than on the T effect.• The rainfalls over the mountainous stations (Bloudan, Arneh and Bab-Janeh) and southern Syria (Izraa, Al-Suwieda, and Al-Kounietra stations), where the altitude is rather high and the DFC parameter is generally short (<50 km), tend to have a high d-excess value very close to the value (22‰), mostly because of the strong exchange with the enriched moisture pro duced from the MS.• The high percentage of prevalent winds (>65%), frequently comes from the W, SW and S directions, strongly support the interactions between the humid air masses passing over the MS and the enriched MS moisture.This also means that the impact of the MS moisture is progressively decreasing towards the in land plains and the northern areas, where the exchange with the continental moisture of a lower d-excess value (≈10‰) becomes more frequent.The RW at the Idleb and Al-Nabk stations, located in specific interior inland areas, tend to be more influenced by predominance exchange with the continental moisture.• The impacts of LE and LN factors on the spatial variation of P stable isotopes were almost weak to moder ate (0.3<r<0.6), revealing that both factors have minor  controls on the spatial variability of these isotopes, likely because of the small geographical extent of this country.However, it was noteworthy to observe that the stable isotope ratios clearly increase with the increasing of LN values in an opposite trend to the usual spatial distribution on a Global scale.This trend strongly reflects the effect of mixing with the enriched moisture originated from the MS, likely as most of the rainy clouds that pass over the MS proceed farther northward towards the north-eastern plains.• The measured concentration of 3 H in the RW samples collected during the period 1989-1993 (4-14 TU) were higher than those (3-10 TU) observed during the period 2004-2006.This change means that the activity of this radioisotope is returning back toward the natural production levels before the starting of the nuclear weapon testing.• The correlation coefficients for the relationships 3 H-DFC and 3 H-LE were rather similar (r>0.63 and r>0.59, respectively), but distinctly higher than those found for the relationships 3 H-LN and 3 H-H (r<0.3).Physically, this reflects the impact of the boundary of the eastern MS coast, which stands in parallel to the LE lines.This permits to conclude that the continuous mixing between the low tritium moisture from the MS with the higher tritium moisture produced in the inland areas was the primary reason for the spatial decline of 3 H concentration toward the MS coast.
• Incorporation of other meteorological factors, such as RH, water vapor pressure and evaporation may help to improve the spatial mapping of atmospheric P isotopes in such a semi-arid to arid country.

Fig. 1 .
Fig. 1.Location maps of Syria showing major topographical structures (a) and distribution of average annual P-amounts and locations of P sampling (b).

Fig. 2 .
Fig. 2. Relationships between monthly δ 2 H and δ 18 O values for all the RW samples collected in Syria (a) and between amountweighted mean δ 2 H and δ 18 O values for all the stations (b).Black dark circles denote RW samples or station points with their errorbars.Black lines are the linear fittings for the RW samples or the station points.Red, olive and blue lines are arbitrary lines, with a slope of 8 and different Y-intercept values (17‰, 20‰ and 23‰, respectively), schematically plotted for approximate matching with the different station points.

Fig. 3 .
Fig. 3. Temporal variation of average monthly 3 H value in the RW samples during the rainy season extended between November and May.

Fig. 4 .
Fig. 4. Relationships between monthly δ 18 O and T values (a), monthly δ 2 H and P-amount values (b) and monthly d-excess and RH values (c) for all the RW samples in Syria.Black lines denote the linear fittings and their equations.

Fig. 5 .
Fig. 5. Relationships between amount-weighted mean δ 18 O ratios and LN values (a), amount-weighted mean δ 2 H and H values (b), amount-weighted mean d-excess and LN values (c), and amount-weighted mean 3 H and DFC values (d) for all the stations in Syria.Black lines and red curves denote the linear and the second order polynomial fittings and their equations, respectively.

Table 1 .
Summary data on the geographical coordinates (LE and LN), altitude (H), distance from the coast (DFC), average values of the meteorological parameter (P, T and RH), num- *Errors of the amount-weighted mean values were calculated according to Gatzand Smith (1995).

Table 2 .
Regression parameters of the linear straight lines and the second-order polynomial curves calculated for the different relationships between the amount weighted-mean isotopic values and the geographical site parameters (LE, LN, DFC and H).
Boldface r and P-values indicate significant relationships.