1 Introduction

Investigation of hydrogen (2H and 3H) and oxygen (18O) isotopes of the water molecule in current precipitation (P) is a useful approach to assess and characterize a variety of hydrological and geochemical processes with relation to the water cycle (Gat, 1996; Clark & Fritz, 1997; Criss, 1999; Mook, 2001; IAEA, 2005a; Hoefs, 2009). These isotopes are greatly studied for characterizing the global and regional scale variations (Rozanski et al., 1993; Bowen & Revenaugh, 2003) and widely applied for tracing the circulation and transport of water vapour and humid air masses in the atmosphere (Aggarwal et al., 2004; Strong et al., 2007; Pfahl & Wernli, 2008). Tritium (3H), as a radioactive isotope, with a half-life of ≈12.32 years (Lucas & Unterweger, 2000), naturally formed by interactions between the fast cosmic rays and nitrogen atoms in the upper atmosphere (Eriksson, 1983), and anthropogenically produced by nuclear experiment programmes (Akata et al., 2011), is also widely and successfully used as a radiotracer for differentiating between old and renewable waters and estimation of groundwater residence time, mostly through mathematical modelling (Zuber, 1986; Kattan, 1997a).

Information on the temporal and spatial variability of the stable isotopes (δ2H and δ18O) and 3H levels in present P are also important for investigating and modelling the surface water flow in catchment areas (Kendall & Coplen, 2001; Karim & Veizer, 2002) and characterizing the groundwater in the different systems (Gonfiantini et al., 2001; Bhat & Jeelani, 2015). Determinations of P stable isotopes are also useful to interpret the paleo-climate data archives and understanding of the past, present and future climate dynamics (Johnsen et al., 1989; Swart et al., 1993; Petit et al., 1999; Aizen et al., 2005; Klein et al., 2016).

It is well documented that the major meteorological parameters, such as air temperature (T), relative humidity (RH) and P-amount values, play certain roles in the variability of P stable isotopes (Lee & Fung, 2008; Terzer et al., 2013). This is because fractionation of the stable isotopes strongly depends on these parameters, which tend to characterize the isotopic signature of the water molecule or water vapour at its initial occurrence under different climate conditions (Majoube, 1971; Gat, 1996; Criss, 1999; Terzer-Wassmuth et al., 2021; Vystavna et al., 2021). Therefore, it is chiefly important to understand the role of such parameters on the temporal and spatial variability of P stable isotopes, especially to better reconstruct the different climatic parameters in the different environments and assess the provenance of water and its modification processes in the various aquatic systems (Clark & Fritz, 1997; Leketa & Abiye, 2020).

However, the chief role is still the change in air T values (Rozanski et al., 1993) and in a second step the cause so-called P-amount effect (Risi et al., 2008). Hence, the isotopic composition of P during the cold periods is usually depleted in heavy stable isotopes while more enriched in the warmer seasons (Angert et al., 2008; Akers et al., 2017). The rain falling over higher elevation areas is steadily more depleted compared to lower elevation areas (Liu et al., 2010). Further inland, this composition is very sensitive to oscillation in air T values (Rozanski et al., 1993) whereas weakly dependent on this parameter in coastal areas (Hendricks et al., 2000). Evaporation of the rainwater drops during their falling on the ground often involves enrichment in the heavy stable isotopes (Scholl et al., 2009), and thus, any change in P isotopes because of secondary evaporation may mask the original isotopic signature of the single rainstorms (Liotta et al., 2008). In temperate regions, variations of P stable isotopes can frequently be correlated with the mean surface air T values (Aggarwal et al., 2012), whereas in equatorial zone rainfall, this relationship is almost feeble and thus unsuitable for monthly or seasonal predictions (Allen et al., 2019).

The “amount effect”, which defines the tendency of stable isotopes to become more depleted with increasing P-amount values (Dansgaard, 1964; Rozanski et al., 1993), is generally not yet enough clear and mostly not consistent because of some superimposed processes, such as moisture recycling, orographic rainout and secondary evaporation (Yang et al., 2019; Vystavna et al., 2021; Mahindawansha et al., 2022). Based on studies conducted in tropical areas, it is revealed that this effect is the major controlling factor (Cruz et al., 2005). Also, it was shown that the amount effect is significant during the monsoon season (Li et al., 2022), and that variations of P stable isotopes can even occur within extreme short time during the single rainstorms (Munksgaard et al., 2012). In all cases, the effects of air T- and P-amount on P stable isotopes are reflected by linear relationships over long-term monitoring periods. Thus, studying the temporal variability of P isotopes during long-time periods would give more in-depth details on the potential regional and global climate changes (Rozanski et al., 1992; Terzer-Wassmuth et al., 2021).

Assessments of the regional climate change in the Mediterranean region have shown that this region is expected to be affected strongly by the increased regional warming (Mariotti et al., 2015). Over the twentieth century, this region will be subjected to strong warming in summer and typically lower rainfall in winter (Giorgi & Lionello, 2008; Ozturk et al., 2015). According to Önol and Semazzi (2009), the seasonal surface air T averages in Syria during 2071–2100 will rise by 2.8 °C and 4.4 °C in winter and autumn, respectively, relative to the period (1961–1990). This increase in climate warming will be accompanied by decline in the seasonal P-amount averages by −28.8% and −18.6% in winter and spring, respectively, and an increase of +4.9% and +90.5% in summer and autumn, respectively. The annual P-amounts are expected to decline by 5–25% during 2040–2069 and 5–30% during 2070–2099, relative to the period (1961–1990), together with a significant rain shortage (24–32%) in winter (Lelieveld et al., 2012).

The effect of climate change at the daytime regional warming (3.5 °C) will considerably amplify the rainfall deficit and thus increasing the risks of drought occurrences in Syria (Guiot & Cramer, 2016; Lehner et al., 2017). A shifting from 1.5 to 2 °C in the regional warming will be reflected in strong increases in dryness and decreases in water availability, likely the renewable waters (Davis et al., 2017). Hence, as Syria is situated in an arid to semiarid region, and severely subjected to extensive changes in its climate under increased warming, the expected substantial future decrease in rainfall will result in high risks to the natural ecosystems, national human activities and country economy (Smiatek et al., 2013; Kelley et al., 2015).

This paper, which aims primarily to characterize the seasonal and annual variability patterns of P isotopes in the Damascus rainfall, has the following main objectives: (1) characterizing the isotopic signatures of P at this station during a period of 22 years, (2) quantifying the temporal variations of P isotopes and exploring the controls of the meteorological parameters on variability of these isotopes, and (3) assessing the vulnerability to climate change in this area. Implications for the regional climate change were extracted by analysing concurrently the linear relationships between annual average and annual amount-weighted means (AWM) values of P isotopes over time, as well as the variability of the meteorological parameters during long-term periods.

2 Methods

2.1 Site description

The station selected for rainwater (RW) sampling is situated on the roof of Soumaria building, located to the west of the Damascus city, the capital of Syria (Fig. 1). The geographical coordinates for this station are as follows: 36.10°E, 33.35°N and ≈675-m elevation (H) above the sea level. This station, nearly placed in a flat plain area at distance of ≈80 km from the Mediterranean Sea (MS), is surrounded by several high elevated structures, such as the Mt. Hermon (2814 m), Mt. anti-Lebanon (2466 m), Mt. Al-Arab (1790 ml) and the southern Palmyrides series (up to 1308 m), situated close to the Syrian Desert (H < 600 m).

Fig. 1
figure 1

Location map of south-western Syria showing the main topographical structures and site of P sampling

Full size image

The weather in Damascus is influenced by a Mediterranean type of climate, with warm to hot and dry summers and moderate rainy winters (Meslmani et al., 2010). The annual average air T values range between 16.3 ± 7.6 °C (Euroweather) and 18.4 ± 6.0 °C (stat.world), while annual average RH values vary between 50 ± 14% and 56 ± 13% (climatemps.com; NOAA). Analysis of the wind trends in the Damascus area shows that the major prevailed winds (≈75%) were coming from the south, southwest and west directions (weatheronline.co.uk).

The rainy season in Damascus is usually short, extending from November to May, with a completely dry period from June to August. The opportunity for occurring light rain showers in September and October is also possible. The number of the rainy days in Damascus (19–68) is usually much higher than the snowy (≈11) days (weather–atlas.com). The distribution of annual average P-amount values in southern Syria shows a gradual increase from east to west, from < 100 mm in the Syrian Desert up to 800 mm on the Mt. Hermon summit (Fig. 2).

Fig. 2
figure 2

Distribution of annual P-amounts in south-western Syria

Full size image

2.2 Precipitation sampling

RW sampling from the Damascus-Soumaria station was started in December 1989. The datasets belonging to this study cover three monitoring periods (December 1989 to April 1993, December 2003 to April 2006 and October 2007 to March 2022). The data of the first two periods (up to April 2006) refer to samples collected as a monthly mixture, whereas those of the last period (2007–2022) refer to samples collected either as daily (single rainstorms) or monthly composite (mixture) samples. A total of 224 daily and 95 monthly composite samples were collected during this study. Only 2 daily samples were collected in summer (June and August).

Daily sampling is based on systematic collection twice a day (morning and evening), whereas monthly sampling is made by accumulating the daily P-amounts, and storing all quantities collected during the entire month in a tightly sealed container, large enough to contain the month P-amount. Both daily and monthly samples were promptly kept in a special refrigerated room (T < 5 °C) until the time of analysis.

A simple dish funnel (diameter ≈80 cm) designed to fit with the frequent P-amounts at the Damascus area was used for P collection. This funnel is securely connected to a plastic container of 20 L, by using a plastic tube wrapped several turns inside the container to prevent from water evaporation. A simple stainless steel rain gauge device (precision ≈0.1 mm) was used for P-amount measurements.

2.3 Isotopic analyses

Two rinsed glass or polyethylene bottles were filled for conducting the isotopic analyses of RW samples. A small bottle of 50 mL was filled for the determination of the stable isotope ratios (δ18O/δ16O and δ2H/δ1H), for which the subsequent analyses were conducted by using a Finnigan Mat mass spectrometer, with measurement accuracy better than ± 0.1‰ and ± 1.0‰, for δ18O and δ2H, respectively. The Vienna Standard Mean Ocean Water (VSMOW) and the Greenland Ice Sheet Precipitation (GISP) standards, carefully prepared and widely distributed by the International Atomic Energy Agency (IAEA), were used for calibrating the analytical results of δ18O and δ2H ratios (Gonfiantini, 1978; Coplen, 1996).

A second bottle of 1 l was also filled for 3H determination, precisely when the P-amounts collected were sufficient to conduct this analysis. The subsequent 3H analyses were performed, after electrolytic enrichment of the RW samples, by using a liquid scintillation counter with detection limit below 0.5 TU (1 TU is equal to 1 3H atom per 1018 atoms of 1H). More information on P sampling and analytical procedures of these isotopes were given by Kattan (2019).

Quality assurance procedures, relative to the ISO/IEC 17025 system, are strictly applied in all laboratories of the Atomic Energy Commission of Syria (AECS), and validity of the analytical results is verified by participation in the diverse inter-comparison tests, organized by either the IAEA (WICO and TRIC), or other regional authorities.

2.4 Statistical analysis and calculations

The softwares Excel 2007 and StatView 5 were used for conducting the descriptive statistics (min, max, mean, standard deviation (SD) and median) of isotopic and meteorological data.

Measured ratios of the stable hydrogen (2H) and oxygen (18O) isotopes in the RW samples were expressed in the δ notation (‰), relative to VSMOW, defined as follows:

$$\mathrm\delta=\lbrack({\mathrm R}_{\mathrm{sam}}-{\mathrm R}_{\mathrm{std}})/{\mathrm R}_{\mathrm{std}}\rbrack\times1000$$
(1)

where Rsam and Rstd refer to the ratio (2H/1H or 18O/16O) in the samples and VSMOW standard, respectively.

The deuterium excess (d-excess or d) value in the RW sample was calculated by applying the Dansgaard (1964) formula, defined as follows:

$$d - excess = {\updelta }^{{2}} {\text{H}} - 8{\updelta }^{{{18}}} {\text{O}}$$
(2)

The AWM value that considers the “P-amount effect” on the isotopic composition during any monitoring period (weeks, months or years) was calculated by using the formula of Yurtsever and Gat (1981):

$$\mathrm{AWM}=\sum\nolimits_{\mathrm i}^{\mathrm n}{({\mathrm P}_{\mathrm i}\times{\mathrm C}_{\mathrm i})}/\sum\nolimits_{\mathrm i}^{\mathrm n}{({\mathrm P}_{\mathrm i})}$$
(3)

where Pi is the daily or monthly P-amount value and Ci is the isotopic composition of RW sample for the period i (days or months).

As the standardised P index (SPI) is a statistical marker largely used for the assessment of drought stress, mostly the wetter and drier climate time series, this index was computed using the following equation (Angelidis et al., 2012; Sarker et al., 2021).

$$\mathrm{SPI}=\frac{\mathrm{Pi}-\mathrm{Pm}}{\mathrm\sigma}$$
(4)

where Pi is the annual P-amount of year i, Pm is the long-term mean annual P-amount value and σ is the SD of annual P-amounts.

In order to identify similarities and compare between the time-series variability results of the meteorological and isotopic data, the paired t-test analysis was carried out using the software StatView 5 by adapting the hypothesized mean difference of 0 and confidential interval of 95% (Bevans, 2022).

3 Results and discussion

3.1 Variability of stable isotopes in RW samples

To identify the variability of the stable isotopes (δ18O, δ2H) and d-excess values in the daily and monthly RW samples collected at the Soumaria station during the period 2007–2022, a detail descriptive statistic was performed (Table 1). Accordingly, the stable isotope ratios in the daily samples varied from −16.6 to +4.1‰ and from −116 to +30‰, with average values of −5.3 ± 3.5‰ and −29 ± 25‰ for δ18O and δ2H, respectively. These isotopes changed in the monthly composite samples from −12.7 to +2.7‰ and from −80 to +29‰, with average values of −5.5 ± 3‰ and −31 ± 20‰ for δ18O and δ2H, respectively. Thus, the spread in isotope ratios in the daily samples is greater compared to the monthly samples. However, both ranges are rather comparable to the ranges (−14.6 to +7.4‰ and −101 to +35‰ for δ18O and δ2H, respectively), previously determined in the RW samples collected all over Syria during the period 1990–2006 (Kattan, 2019).

Table 1 Descriptive statistics of environmental isotope values in the daily and monthly composite RW samples collected at the Damascus-Soumaria station during the period 2007–2022
Full size table

The variability of d-excess values in both daily and monthly samples is rather similar, varying from < 0‰ in the samples highly evaporated or originated from hot and dry sources, up to 27–30‰ in the samples strongly affected by exchange with the enriched moisture (d≈30‰), produced from the MS (Gat et al., 2003). Both ranges are also identical to the range (< 0 to + 30‰), previously reported for the d-excess values in the Syrian rainfall (Kattan, 2019).

The average d-excess values in the daily and monthly samples are rather identical (13 ± 8‰), and both are lower than the value (22‰), given by Nir (1967) for the Mediterranean meteoric water line (MMWL). The AWM δ18O and AWM δ2H values are constantly lower than the average and median δ18O and δ2H values, likely as the isotopic composition of P associated with high rain events (p > 10 mm) is usually much depleted compared to low rain events. Oppositely, the AWM d-excess value (15–16‰) in both daily and monthly samples is higher by ≈1–3‰ compared to the median and average d-excess values. The average d-excess and AWM d-excess values (15 ± 9‰ and 19 ± 1‰, respectively), previously calculated in the monthly composite samples collected at the Damascus station (Kattan, 2019), were much closer to the value (22‰), given by Nir (1967).

3.2 Variability of tritium in RW samples

The descriptive statistics of 3H concentrations in the monthly RW samples collected at Damascus during the period from December 1989 to May 2019 are shown in Table 1. The concentrations of 3H varied in the range from 1.9 to 14.9 TU, with an average value of 4.8 ± 3.6 TU, slightly higher than the median (3.9 TU) or the AWM (4.5 TU) 3H values. The current 3H average is thus higher by ≈1.1 TU compared to the value (5.9 ± 3.5), previously calculated as an average value for the period 1989–2006. This difference clearly means that the concentration of 3H in current P is mostly due to the natural cosmogenic production of this radioisotope in the upper atmosphere before the start of the nuclear weapons tests (Eriksson, 1983). Previous assessments of spatial distribution of 3H content in the rainfall in this region showed a build-up in 3H values with increasing distance from the MS coast, likely because of the continuous mixing between the low 3H moisture produced from the MS and the high 3H moisture derived from the inland areas (Gat & Carmi, 1970; Kattan, 1997b, 2021).

3.3 Local meteoric water line

In order to define the commonly so-called local meteoric water line (LMWL), the relationships between δ2H and δ18O values in the daily and monthly composite RW samples collected at the Soumaria station during the period 1990–2022 were determined (Fig. 3). Distribution of the sample points in this graph permits to observe that all the sample points are distributed between two lines: the global meteoric water line (GMWL), defined by Craig (1961) as δ2H = 8δ18O + 10, and the MMWL, given by Nir (1967) as δ2H = 8δ18O + 22. Most of the sample points representing the depleted isotopic ratios are rather scattered close to the MMWL, while those representing the enriched or evaporated samples are located close to the GMWL. All samples fit nicely with a least squares regression line and defines the so-called LMWL, approximated by the following equations:

$$\mathrm\delta{}^2\mathrm H=(6.77\pm0.14)\mathrm\delta{}^{18}\mathrm O+(6.8\pm0.9)$$
(5)

with [R2 = 0.93, p-value < 0.0001 and N = 224] for the daily data, and

$$\mathrm\delta{}^2\mathrm H=(6\text{.68}\pm0.20)\mathrm\delta{}^{18}\mathrm O+(5.8\pm1.3)$$
(6)

with [R2 = 0.94, p-value < 0.0001 and N = 95] for the monthly data.

Fig. 3
figure 3

The relationships between δ2H and δ18O values in daily and monthly RW samples (a) and between annual average δ2H and annual δO values (daily data) and annual AWM δ2H and annual AWM δ18O values (monthly data) in the RW samples (b) collected at the Damascus station during 1990–2022

Full size image

The line slope values for daily and monthly data (≈6.7–6.8) are rather similar and lower than the value 8, determined for the isotopic fractionation under equilibrium condition (Craig, 1961; Rozanski et al., 1993). The sample points representing the most depleted isotopic ratios match remarkably with a line slope of ≈8. These samples can isotopically be considered as slightly or not affected by evaporation. However, as the re-evaporation of the raindrops is usually associated with a line slope value lower than 8, the sample points scattering around a line with a slope lower than 8 means that the isotopically enriched samples (δ18O >  − 4‰ and δ2H >  − 20‰) were subjected to evaporation during their falling on the land surface (Managave et al., 2016). The line slope values calculated for the relationship δ2H-δ18O at each month of the rainy season (Table 2), varied from ≈6.3 ± 1 (May) up to ≈7.3 ± 0.2 (January). The line slope values calculated for the months from December to March are higher than 7, whereas those for the remaining months are lower (≈6.3–6.8), meaning thus that the rainfall in winter is less affected by evaporation, compared to those falling in spring and autumn.

Table 2 Regression parameters of the relationship δ2H-δ18O calculated for the Damascus rainfall at each month of the rainy season
Full size table

The relationships (δ2H-δ18O), established on the basis of annual average values (daily data) and annual AWM values (monthly data) for the diverse rain seasons (22 years), permit to observe a different scatter between the distribution of the points representing daily and monthly data (Fig. 3b). The points representing the monthly data reflect generally more depleted isotopic ratios compared to those of the daily data. Although this difference is between daily and monthly data, all points representing both data fit nicely with two least squares regression lines having nearly a similar line slope and an identical Y-intercept value, as approximated by the following equations (Fig. 3b):

$$\mathrm\delta{}^2\mathrm H=(6.11\pm0.48)\mathrm\delta{}^{18}\mathrm O+(3.5\pm2.6)$$
(7)

with [R2 = 0.93, p-value < 0.0001 and N = 15] for daily data, and

$$\mathrm\delta{}^2\mathrm H=(6\text{.19}\pm0.46)\mathrm\delta{}^{18}\mathrm O+(4.0\pm3.2)$$
(8)

with [R2 = 0.92, o-value < 0.0001 and N = 22] for monthly data.

When comparing the isotopic ratios of P collected in daily or monthly composite samples with those calculated as annual average or annual AWM isotopic values for the different rain seasons, the line slope values in the case of annual average and annual AWM values are always lower than the line slopes calculated on the basis of all samples (Table 2). The differences that were found between the two methods, precisely in the line slope (≈0.5–0.7) and the Y-intercept value (≈1.8–3.3‰), generally mean that the annual average and annual AWM values conjunctly reflect the effects of P-amount and re-evaporation of raindrops during their falling. In all cases, the line slope values estimated for the relationship δ2H-δ18O in this study (6.1–7.3 ± 1) completely agree with the slope values reported for the rainfall in the Eastern Mediterranean region, changing from 5.48 for the Bet-Dagan station (H≈36 m), the unique station that belongs to the Global Network of Isotopes in Precipitation (GNIP) in this region (IAEA-WMO, 2021), up to 7.3 ± 0.4 for the Arneh station (H ≈1440 m), located in Syria (Kattan, 2019). The line slope values calculated also fit with the range (6.2–6.6), previously reported for the LMWLs in Jordan (Bajjali, 2012), Lebanon (IAEA, 2005b), and Syria (Abou Zakhem & Hafez, 2007).

3.4 Seasonal variability of meteorological parameters

With the aim to study the seasonal variability of the key meteorological parameters (P-amount, air T, and RH) recorded at the Soumaria station during the period 1990–2022, the monthly averages and SD values ( ±) for these parameters were calculated (Table 3). The highest monthly average P-amount values (> 30 mm) usually occur during the period from November to February (Fig. 4a). December is the most humid month (p > 40 mm), while October and April are the less wet (p < 20 mm). However, the month of May showed an unusually high monthly average P-amount (≈34 ± 32 mm), specifically because of the intensive rain events (≈70 mm) occurred in May 2018. The monthly average P-amount values calculated for the months of May, October, November, and December are clearly higher than the long-term average monthly values reported for these months (climatemps.com). The monthly average P-amount values that were recorded for the period from January to March are remarkably lower than the long-term average values, reported for these months (Fig. 4a). The decrease of P-amount in winter and its increasing in autumn is likely because of the intense climate change affecting this area.

Table 3 Monthly averages and SD ( ±) values of P-amount, air T, RH, δ18O, δ2H, d-excess and 3H values in the daily and monthly RW samples collected at the Damascus-Soumaria station during the period 1990–2022
Full size table
Fig. 4
figure 4

Temporal variations in monthly averages and SD values for P-amount (a), air T (b) and RH (c) values during the rainy season

Full size image

Variations in the monthly average air T values during the rainy season clearly show that the values calculated during this study are systematically higher than the long-term average monthly values (Fig. 4b), reported for the Damascus region (stat.world). The difference in the air T value during the period from January to April, compared to the long-term average values, is rather high (> 3 °C), whereas it is much lower in remaining months. The seasonal variations in the monthly average air T values are frequently marked by low T values (< 15 °C) during the period from December to March and higher T values (> 15 °C) during remaining months. The monthly average air T values in May and October are usually the highest (> 20 °C).

The seasonal variation in the average monthly RH values is reflected by remarkably high RH values (> 60%) during the period from November to February and lower values (35% < RH < 55%) during remaining months (Fig. 4c). The highest monthly average RH values are associated with the wettest months (December and January), whereas the lowest values (< 55%) usually occur at the beginning (October) and at the end of the rainy season (April and May). The monthly average RH values recorded within this study for the period from December to May are distinctly lower than the long-term monthly average values reported for this parameter for the same period (weatheronline.co.uk), while they are higher in October and November (Fig. 4c). The variations of the diverse meteorological parameters agree well with the strong regional warming and the typically lower rainfall in winter expected for this region as a result of climate change (Giorgi & Lionello, 2008; Ozturk et al., 2015).

3.5 Seasonal variability of P isotopes

The seasonal variations in the monthly average δ18O and δ2H values, calculated on the basis of daily and monthly data (Table 3), are remarkably identical (Fig. 5a and b). Both isotopes have to some extent similar variations to those of the monthly average air T values (Fig. 4b), with depleted isotopic ratios in the cold months from December to February and enriched ratios in the warmer or less wettest months (October and May). However, the monthly average δ18O and δ2H values, calculated on the basis of monthly data during the cold period (December to March), are clearly more depleted compared to those of the daily data, with the lowest isotopic ratios during the coldest month (January). Changes in the air T values in the atmosphere are generally the principal cause behind the seasonal variations of the stable isotope ratios in the Damascus rainfall. This trend of isotope variations strongly suggests that the depleted isotopic ratios usually occur when the monthly P-amount and RH values are both elevated.

Fig. 5
figure 5

Temporal variations in monthly averages and SD values for δ18O (a), δ2H (b), d-excess (c) and 3H (d) values during the rainy season

Full size image

The seasonal variation in the monthly average d-excess value is completely opposite to those of δ18O and δ2H values (Fig. 5c). The monthly average d-excess values calculated on the basis of monthly data are generally equal or slightly different from those of the daily data, especially during the cold period from December to March. However, the monthly average d-excess values, calculated on the basis of daily data during the warmer months (October, November, April and May), are clearly higher than those calculated for all monthly data. A good relationship is found between the seasonal variations in the monthly average d-excess and monthly average RH values, with high d-excess values (> 14‰) during the wettest months (December to March), and lower monthly average d-excess values (< 10‰) at the beginning and at the end of the rainy season. This evolution trend strongly reflects the influence of the enriched moisture (d≈30‰), produced from the MS (Gat et al., 2003), likely as a result of the strong interactions occurring between the enriched moisture existing over this sea with the humid air masses passing over the MS, specifically during the period from December to February. This isotopic exchange also explains the significant increase in the d-excess value in the rains falling all around the Mediterranean region (IAEA, 2005b).

The seasonal variations in the monthly average 3H concentrations in the Damascus rainfall calculated on the basis of both daily and monthly data are marked by slight fluctuations around ≈3–6 TU during the period from October to February (Fig. 5d). Then, by the beginning of March, a gradual build-up in this parameter is observed, up to values greater than 9 TU during April and May (monthly data) or 15 TU in April (daily data). This evolution clearly reflects the phenomenon so-called spring leak, which usually occurs during the spring season, likely as a result of the continuous injection of high 3H loads from the troposphere to the ambient atmosphere (Rozanski et al., 1991; Ehhalt et al., 2002; Harms et al., 2016).

3.6 Annual variability of meteorological parameters

Distribution of the long-term annual P-amount values recorded at Mazzeh station, the closest station to the sampling site, during the period from 1919 to 2022 shows a range from ≈50 mm up to ≈350 mm, with an average value of ≈208 ± 70 mm and a coefficient of variation of 34% (Fig. 6a). A dozen of water years (P > 350 mm) can be considered as wet ones, while only 8 years (P < 100 mm) can be classified as dry ones. Although the annual P-amount values recorded during 1919–1969 varied in a rather similar range (53–354 mm), and have an identical coefficient of variation (34%) compared to the period 1919–2022, the average P-amount value (≈213 ± 72 mm), calculated for the period 1919–1969 is more elevated compared to the long-term average value (208 mm), calculated for the period 1919–2022. The distribution of P-amounts over a period of ≈1 century permits to observe a decrease of ≈ − 3.1 mm/decade (Table 4). However, when considering the distribution of this parameter at the sampling station during the period 1990–2019, a decreasing gradient of ≈ − 9 mm/decade is found (Fig. 6b). This variation trend could be considered as a representative evolution of the rainfall shortage associated with the climate change. The annual P-amount values during 1990–2022, which varied nearly in a similar range to those of the long-term data (86–337 mm), and have a similar coefficient of variation (34%), have an average of ≈185 ± 67 mm. This means that the annual average P-amount during the last three decades has decreased by ≈23 mm, relative to the period 1919–1989 (Table 4), and by about ≈28 mm, relative to the period 1919–1969. This significant decrease in P strongly proves the effect of climate change, as previously also shown in the unusually high rain showers in October and May and low rain events in January and February (Fig. 4a).

Fig. 6
figure 6

Distribution of long-term annual P-amount values recorded at Mazzeh station during 1919–2022 (a) and at the sampling Soumaria station during 1990–2022 (b)

Full size image
Table 4 Regression parameters of the linear relationships between annual P-amount, air T and RH values versus time and between annual δ18O, δ2H, d-excess and 3H values versus time for the rainfall collected at the Damascus-Soumaria station during the period 1990–2022
Full size table

In order to assess the drought stress on P time series, the index (SPI), as the standard deviations from the median (Sarker et al., 2021), or as the probability distribution function of P record during a long-term period of at least 30 years, was computed (Fig. 7). Accordingly, the variation in this index during the studied period shows a range from − 2.2 (1960) to 2.1 (1953 and 1945). This means thus that the intensity of the drought stress (negative SPI values) has changed between normal (0 to − 0.5) and severe (− 1 to − 2) classes (Hao & Singh, 2015). By adapting a polynomial curve of 4th order to fit with all SPI values during the period 1919–2022, it can be observed that the fitted curve over the period 1970–2022 is marked by a negative SPI trend, whereas it is positive over the previous period (Fig. 7). However, by plotting a second polynomial curve of the same order for the period 1990–2022, it was found that the correlation coefficient of the fitted last curve (R2 = 0.08 and p-value < 0.0001) was higher than the first one (R2 = 0.02 and p-value < 0.0001), with also more severe drought stress (higher SPI values) during the period 1994–2012.

Fig. 7
figure 7

Distribution of long-term annual SPI values calculated for Mazzeh station data during 1919–2022 and plots of the fitted polynomial curves of the 4th order during the period 1919–2022 (blue line) and the period 1990–2022 (red line)

Full size image

In all cases, the effect of climate change on the rain shortage in the Damascus region is shown by a decrease of ≈11–13%, relative to the period 1919–1969. However, the decline estimated in this study in P-amounts is clearly still lower than the average value (17%), previously given by Abou Zakhem and Kattaa (2016), based on data covering the period 1941–2000. A significant reduction in the annual P-amount values (5–30%) is expected during 2040–2099, relative to the period 1961–1990 (Lelieveld et al., 2012), and a similar decrease is also expected at a number of stations in the vicinity of the Figeh spring near Damascus during 2040–2069 (Smiatek et al., 2013).

The distribution of the long-term annual air T values recorded at Mazzeh station during 1919–2022 (Fig. 8a) shows a range from 15.2 °C up to 19.6 °C, with an average value of 17.0 ± 0.9 °C and a coefficient of variation of ≈5%. The temporal evolution of annual air T values during 1990–2019 is marked by variations from 15.9 to 19.6 °C, with an average value of 17.9 ± 0.8 °C and a coefficient of variation of ≈4% (Fig. 8b). The temporal evolution of annual air T values during 1919–1969 is marked by a slight continuous increase (≈ + 0.01 °C/decade), while it is clearly much higher (≈ + 0.64 °C/decade) during 1990–2022. The increasing gradient in the annual air T values during 1919–1969 fluctuated between 15.2 and 18.0 °C, with an average value of 16.6 ± 0.6 °C, and a coefficient of variation of ≈4%. This distribution strongly proves that the annual average air T value during the last three decades has increased by ≈ + 1.2 °C, compared to that recorded one century before. If this parameter continues increasing at the same gradients (Table 4), the annual air T value by the end of the twenty-first century is expected to attain ≈23 °C, i.e. higher by ≈5 °C, compared to the current air T values. Scenarios of climate change in the Mediterranean regions predicted an increase of ≈1.5–2.8 °C in the annual air T value in 2050 (Dai, 2011). The annual average air T value at Damascus during 2070–2099 was simulated to increase by + 4.8 °C, relative to the period 1961–1990 (Smiatek et al., 2013).

Fig. 8
figure 8

Distribution of long-term annual average air T values recorded at Mazzeh station during 1919–2022 (a) and at the sampling Soumaria station during 1990–2022 (b)

Full size image

The annual average RH values recorded at the sampling station during 1990–2022 varied within the range of 42–56%, with an average annual RH value of ≈49 ± 3% (Fig. 9). The temporal variation in this parameter generally reflects a small reduction (≈ − 0.3%/decade). If this parameter continues to decline at the same rate, the annual average RH value at the end of this century is expected to decrease by ≈2%, relative to the current RH values, meaning thus that this parameter is approximately the least affected by the consequences associated with climate change.

Fig. 9
figure 9

Temporal variation in annual average RH values recorded at the sampling Soumaria station during 1990–2022

Full size image

3.7 Annual variability of P isotopes

The regression parameters for the linear relationships established between the annual values in δ18O, δ2H, d-excess and 3H versus time, relative to the RW samples collected at the sampling station during 1990–2022, are shown in Table 4. Accordingly, the points representing the annual values in δ18O and δ2H, calculated as arithmetic averages, are clearly higher than those calculated as AWM values. The temporal variations in both stable isotopes over time show generally a detectable isotopic increase in the RW composition with increasing time (Fig. 10a and b). The average increasing gradients are ≈ + 0.3–0.5‰ and ≈ + 1.1–3.2‰ per decade for δ18O and δ2H, respectively. This means that the annual δ18O and δ2H values in the RW samples at the beginning of 2050 will be increased by ≈ + 1.0–1.5‰ (≈15–20%) and ≈ + 3–10‰ (≈10–25%) for δ18O and δ2H, respectively, compared to the current annual δ18O and δ2H ratios.

Fig. 10
figure 10

Temporal variations in annual δ18O (a), δ2H (b), d-excess (c) and 3H (d) values measured in the RW samples collected at the Soumaria station during 1990–2022

Full size image

The temporal variations in the annual d-excess values and annual 3H concentrations show oppositely a significant decrease with increasing time (Fig. 10c and d). Although the annual AWM d-excess values are slightly higher than the annual average d-excess values, the temporal variations in the d-excess values in both cases are rather similar. The annual average d-excess values vary in the range of 7–22‰ whereas fluctuate from 11 to 23‰ in the case of annual AWM d-excess values. The correlation coefficient calculated for the relationship between annual AWM d-excess values versus time is relatively high (R = 0.35) compared to remaining isotopic parameters (Table 4). This relationship indicates a decrease of ≈ − 1.1‰ per decade in the AWM d-excess value, while a reduction of ≈ − 0.9‰ is found in the case of the annual average d-excess values. This decrease clearly means that after three decades, the d-excess values will be lower by ≈ − 2.7–3.3‰, compared to the current d-excess values. This decline also means that the predicted d-excess values during 2050 will be around 11–13‰, which are much lower than the value 22‰, given by Nir (1967) for this region, approaching the value 10‰, given as an average for the Globe (Craig, 1961).

The annual average concentrations of 3H in the RW samples in the Damascus area during the study period varied from 8.5–8.7 TU (1989–1990) to 2.3–2.5 TU (2012–2013), with an average value of ≈4.8 ± 2.6 TU. The gradients of 3H degradation over time (− 0.40 and − 0.42 TU/decade for annual AWM and annual average 3H values, respectively) are rather similar (Fig. 10d). However, the annual 3H values calculated for the year 2018 was relatively high (> 7 TU), likely because of the heavy rains occurred in May 2018, likely because of the “spring leak” effect, usually associated with injection of high 3H loads and bilateral exchange among the stratosphere, troposphere and ambient atmosphere (Ehhalt et al., 2002; Harms et al., 2016). Due to this effect, the concentration of 3H in April 1990 was also elevated (13.6 TU), compared to the value (4.3 TU) measured in December 1989 (Kattan, 1997b). The annual variability in this radioisotope during the last three decades shows a progressive decline towards low 3H levels (< 5 TU), which is identical to those observed before the start of the nuclear weapons tests (Eriksson, 1983). Hence, as the nuclear weapon experiments had ceased completely since the 1980s, the previous 3H loads due to anthropogenic sources are currently considered to be of minimal levels, likely because of 3H decay. The decline of 3H concentration in the current P towards low levels strongly suggests that this radioactive isotope is primarily produced by the natural cosmogenic reactions in the upper atmosphere.

3.8 Comparison between the annual time-series variability of studied parameters

To identify the similarities and compare between the time-series variability of the different meteorological parameters and assess their possible controls on the isotopic data, the paired t-test analysis was carried out by adapting the null hypothesis, i.e. no variation exists between variables, or that a single variable is no different than zero (Bevans, 2022). The t-values, calculated based on the multivariate analysis between the different parameters of the entire dataset (Table 5), show that the true difference in means is different from 0, and that most of the p-values are < 0.0001. In all cases, the p-values are much smaller than 0.03. Only a small number of p-values are between 0.03 and 0.07. The t-values calculated for all variables with respect to time range from 2.4 to 25.2, with negative values in the case of P and RH parameters, as both variables tend to decrease with increasing time. The results also prove that the variability trends in the stable isotopes with time are rather similar (15 < t-value < 25) but distinctly different from those of air T and d-excess (t-value < 5). Also, it can be observed that the variability trends in P and RH are of the same order (− 13.3 < t-value <  − 12.6). When considering the variability of all parameters against P, it is found that the t-values are almost of the same order, with slight significant differences in the t-value between 8.7 (RH) and 13.6 (annual AWM 2H). When comparing the variability trends of all parameters with respect to air T parameter, more significant differences are found in the t-value, changing between 2.0 (annual AWM d-excess) and 93.7 (annual AWM 18O), meaning thus that the role of air T variability is accompanied by different effects on the isotopic variables. When comparing the similarities of the different variables with respect to RH parameter, it is found that the t-values are fluctuating between 36.1 (annual average 2H) and 62.2 (annual AWM 18O), meaning thus that the RH parameter has a moderate influence on most isotopes.

Table 5 Matrix of t-values calculated based on the multivariate analysis between the different parameters of the rainfall collected at the Damascus-Soumaria station during the period 1990–2022
Full size table

3.9 Meteorological controls on variability of P stable isotopes

Changes in the ambient weather conditions, specifically air T, P-amount and RH values, are usually the principle causes for the temporal and spatial variations of the isotopic composition of meteoric water (Dansgaard, 1964; Kumar et al., 2010; Terzer et al., 2013). However, these factors are not always the only causes for the variability of such isotopes, as other factors such as the isotopic composition of the water vapour in the atmosphere, the origins and trajectories of humid air masses may contribute altogether in the formulation of the isotopic signature of rainfall (Lawrence et al., 1982; Liotta et al., 2008; Liu et al., 2011; Jeelani et al., 2018; Vystavna et al., 2021).

The regression parameters for the linear relationships, calculated between the annual average and annual AWM values of δ18O, δ2H and d-excess versus the annual average values of P-amount, air T and RH values recorded for the Damascus rainfall during 1990–2022 (Table 6), show that the correlation coefficients (R2) calculated on the basis of annual average values are mostly higher or equal to those of the annual AWM values. Although the R2 values for the relationships δ18O-T, δ2H-T and d-excess-T are moderate (0.07 < R2 < 0.29), they are much higher than those of the other meteorological parameters, meaning thus that the effect of air T is the primary controlling factor affecting the annual variability of P stable isotopes. The difference in the R2 value between the two methods is mostly below 3%.

Table 6 Regression parameters of the linear relationships between annual δ18O, δ2H and d-excess values versus annul P-amount, annual air T and annual RH values for the rainfall collected at the Damascus-Soumaria station during the period 1990–2022
Full size table

The effect of P-amount on the annual variability of P stable isotopes is shown by small isotopic degradations (≈0.4–0.6‰ and 1.5–3.3‰ for δ18O and δ2H, respectively) and weak correlations (0.02 < R2 < 0.16), with the increasing of annual P-amount value by 100 mm (Fig. 11a and b). However, this effect is relatively better in the case of variations in the annual d-excess values (R2 = 0.16) and reflected by an increase of ≈1.7‰ (AWM data) and ≈1.8‰ (average data), with the increasing of annual P-amount value by 100 mm (Fig. 11c). Previous estimates of this effect on the monthly variations of P stable isotopes during 1989–2006 showed a decrease of ≈ − 0.7 ± 0.3‰ and ≈ − 2.9 ± 2.1‰ for δ18O and δ2H, respectively, accompanied with an increase of + 2.7 ± 0.6‰ in the d-excess value, with the increasing of monthly P-amount value by 10 mm (Kattan, 2019).

Fig. 11
figure 11

Relationships between annual δ18O and annual P-amount values (a), annual δ2H and annual P-amount values (b) and annual d-excess and annual P-amount values (c) recorded at the Soumaria station during 1990–2022

Full size image

The T effect on the annual variability of δ18O and δ2H values is reflected by enrichments of ≈0.50–0.58‰/°C and ≈ + 2.2–2.7‰/°C, respectively, with the increasing of annual air T value by 1 °C (Fig. 12a and b). The R2 values in the case of the relationship δ18O-T (0.16 < R2 < 0.17) are generally higher than those calculated for the relationship δ2H-T (0.07 < R2 < 0.10). However, the strongest correlations are found for the relationship d-excess-T (0.28 < R2 < 0.29). This relationship reflects a decrease of ≈1.8–1.9‰/°C in the annual d-excess value with the increasing of annual air T value by 1 °C (Fig. 12c).

Fig. 12
figure 12

Relationships between annual δ18O and annual air T values (a), annual δ2H and annual air T values (b) and annual d-excess and annual air T values (c) recorded at the Soumaria station during 1990–2022

Full size image

The enrichments in the annual δ18O and annual δ2H values are slightly higher than the monthly enrichment gradients (≈0.5 ± 0.12‰/°C and ≈2.7 ± 0.9‰/°C for δ18O and δ2H, respectively), previously given for the Damascus rainfall during 1989–2006, with the increasing of monthly air T value by 1 °C (Kattan, 2019). The average decrease in the monthly d-excess value in the case of the aforementioned rainfall was ≈1.3 ± 0.3‰/°C, with the increasing of monthly air T value by 1 °C. In all cases, the enrichments in δ18O value calculated for the Damascus rainfall are generally in a good accordance with the published δ18O/ΔT gradients (> 0.4‰/°C), given for rainfall in the Mediterranean countries (IAEA, 2005b). However, the monthly enrichments in δ18O value are relatively much higher compared to the average monthly enrichment (0.4 ± 0.02‰/°C), given for the GNIP data (Rozanski et al., 1993).

The effect of RH has also a decisive influence on the variations of P stable isotopes (Gat, 1996). Changes in this parameter can principally affect the isotopic fractionation occurring either under equilibrium or kinetic conditions (Majoube, 1971). This effect is usually reflected by enrichments in heavy stable isotopes due to evaporation, as the lower the RH value, the higher the isotopic fractionation and the higher the evaporation (Clark & Fritz, 1997). The RH effect on the annual variations of P stable isotopes is marked by small depletions (≈0.8–1.3‰ and ≈2–6‰ for δ18O and δ2H, respectively), accompanied by an increase of ≈4–5‰ in the annual d-excess value, with the increasing of annual RH value by 10% (Fig. 13). Previous estimates of this effect on the monthly P stable isotopes in the Damascus rainfall during 1989–2006 showed depletions of ≈2.2 ± 0.5‰ and ≈12.6 ± 3.6‰ for δ18O and δ2H, respectively, accompanied by an increase of ≈5.3 ± 1.2‰ in the monthly d-excess values, with the increasing of monthly RH value by 10% (Kattan, 2019).

Fig. 13
figure 13

Relationships between annual δ18O and annual RH values (a), annual δ2H and annual RH values (b) and annual d-excess and annual RH values (c) recorded at the Soumaria station during 1990–2022

Full size image

4 Conclusions

Analysis of 22-year records of δ18O and δ2H, d-excess and 3H values in the RW samples at the Damascus area during 1990–2022 and long-term time series of air T- and P-amount values during 1919–2022 is discussed. Daily and monthly composite RW samples are conjunctly used to characterize the seasonal and annual variability patterns of P isotopes at the Damascus station. The meteorological controls on the temporal variations of P isotopes are also quantified during the last three decades to assess the vulnerability to climate change in this region. The major conclusions that can be extracted from this study are the following:

  • The seasonal variations in the monthly average δ18O and δ2H values, calculated on the basis of both daily and monthly data, are remarkably similar, with depleted isotopic ratios in the coldest months (December to February), specifically when P-amount and RH values are both elevated. However, the enriched isotopic ratios distinctly occur in the warmer and less wettest months (October and May) and strongly prove that the change in air T value is the principal cause for the variability of P stable isotopes. The high 3H concentrations (> 9 TU), markedly measured in April and May, clearly reflect the “spring leak” phenomena, typically distinguished by injection of high 3H loads, associated with bilateral exchange among the stratosphere, troposphere and atmosphere.

  • The seasonal variation in the monthly average d-excess values is completely opposite to those of δ18O and δ2H values, with high d-excess values (d > 14‰) during the wettest months, when P-amount and RH values are both high, and lower values (d < 10‰) at the beginning and the end of the rainy season. The good relationship between RH and d-excess values proves the efficacy of using the d-excess parameter as a marker for assessing secondary evaporation. The seasonal variation in the d-excess values strongly reflects the effect of interactions occurring between the humid air masses passing over the MS in winter and the enriched moisture (d≈30‰), produced by evaporation from this sea.

  • The long-term annual variation of P-amount values recorded at the Damascus area during 1919–2022 clearly shows that the average P-amount value during the last three decades (185 ± 67 mm) is lower by ≈28 mm compared to the value (213 ± 72 mm), recorded during 1919–1969. The effect of climate change on rain shortage at the Damascus area is marked by a decrease of ≈11–13%, relative to the period 1919–1969, which is clearly lower than the rate (17%), given for the period 1941–2000.

  • The distribution of the long-term annual air T values recorded at the Damascus area during 1919–2022 is marked by an increase gradient of ≈ + 0.2 °C/decade, which is remarkably much lower than the gradient (+ 0.64 °C/decade), determined for the period 1990–2022. This temporal evolution clearly reflects the impact of climate change affecting this region since the 1970s. The annual variation in the RH values during the period 1990–2022, which shows a small reduction gradient (≈ − 0.3%/decade), reveals that this parameter is the less affected by the consequences related to climate change.

  • The linear relationships between annual average δ18O and δ2H values versus time indicate an increase of ≈0.3–0.5‰ and ≈1.1–3.2‰ per decade for δ18O and δ2H, respectively, associated by a decrease of ≈0.9–1.1‰/decade in the d-excess value.

  • Variability of the annual average 3H concentrations during the last three decades towards low levels (< 5 TU) strongly suggests the return back towards the cosmogenic production levels of 3H in the upper atmosphere before the start of the nuclear experiment tests.

  • The principal factor that strongly controls the seasonal and annual variability of P stable isotopes is the change in air T over time. This factor is reflected by significant isotopic enrichments (0.5–0.6‰ and 2.2–2.7‰ for δ18O and δ2H values, respectively) with the increase of annual air T value by 1 °C. The impacts of P-amount and RH are of minor importance. However, further causes, such as the secondary evaporation and enriched moisture produced from the MS, can conjunctly affect the variability of P isotopes in this semi-arid area.

  • As P is an important component of the water cycle and the overall regional water availability is highly dependent on the cold P, which has continuously influenced by climate change, the findings gained from this study will certainly help to better improve the knowledge on the temporal variability patterns of P isotopes and properly use these natural tracers in assessing the risks due to climate change on the quality and quantity of renewable surface and groundwater resources in the Eastern Mediterranean region.