Stable Isotope Ratios in Meteoric Waters in El Kef Region, Northwestern Tunisia: Implications for Changes of Moisture Sources

The stable isotopes ¹⁸O(H₂O) and ²H(H₂O) have many applications in hydrogeological investigations, meteorological, hydrologic, ecological and paleoclimatic studies [1-5]. Due to their high abundance and the simplicity of their analytical determination, spatial and temporal variations in the ¹⁸O and ²H contents of precipitation are caused by the isotope fractionation effect accompanying evaporation from the ocean and condensation during the atmospheric transport of water vapor [6,7]. Isotope fractionation is a thermo-dependent reaction, as proved by the existence of a correlation between δ18O in rainwater and environmental temperature [6,8]. Thus, more the temperature is higher; more the heavy isotope is higher. The existence of thermo-dependence in the fractionation of ¹⁸O and ²H implies a correlation between the two parameters, which has been defined by different authors:

Craig [9] and Rozanski [10]: Global Meteoric Water Line (GMWL: δ²H=8 δ¹⁸O+10);

Zouari [11]: Local Meteoric Water Line of the Tunis-Carthage (LMWL: δ²H=8 δ¹⁸O+12.4).

The Global Meteoric Water Line (GMWL) corresponds to the averaging of numerous local meteoric straight lines (regional data), each of which is influenced by the above mentioned geographic and climatic factors. The intercept of the GMWL is termed the deuterium excess (d=δ²H-8 δ¹⁸O after Clark [12]. The value of this parameter is acquired during evaporation, and does not vary significantly during the later history of the cloud mass. It is thus a valuable indicator of the source area of the water vapor Rindsberger [13]; Cruz [14]; Celle-Jeanton [15]: d values close to 10% indicate waters of Atlantic origin, values close to 22% are characteristic of waters from the Eastern Mediterranean and d values close to 14%, intermediate between the first 2, are detected in rainwater falling on the Western Mediterranean.

In this study, we report the correlation between δ¹⁸O, δD and D-excess value in precipitations and their relationship with temperature. More importantly, their responses to changes in moisture sources will be discussed. This may prove beneficial in understanding the response of precipitation δ¹⁸O and δD to climatic/environmental variables in Northwestern Tunisia (El Kef area) and in understanding the palaeoclimatic records associated with precipitation δ¹⁸O and δD in monsoon climate studies.

Geography

The study area is located in the North Africa and in the NW part of Tunisia, a site located at the southern margin of the Tethys. The El Kef area is located between 40°30′ and 39°45′N, and 6°30′ and 7°30′E, near the Atlantic–Mediterranean confluence (Figure 1). It extends 60 km in the N-S direction and 60-80 km in the E-W direction, covering an area of 420 km². The elevation of the El Kef Plateau decreases from 400–500 m above sea level (m a.s.l.) (Figure 1).

Figure 1: The localization map of El Kef region (NW Tunisia).

Climate

Meteorological data (precipitation and temperature) obtained at the nearest meteorological station (El Kef meteorological station: 7G09′00”N, 40G20′70”E, asl 10 m) around the sampling site includes only the period from 1924 to 2001. During this period, the maximum of precipitation was observed in January (58.6 mm) and a minimum in July (9.4 mm) (Figure 2); the maximum and minimum monthly mean temperatures that were recorded was approximately 36°C in August and 3°C in January, respectively (Figure 3). Relative humidity was maximum in spring (˜65%) and minimum in summer (˜26%). Over 80% of the annual precipitation occurs during the rainy season from December to May. The precipitation that occurs during dry season from June to November accounts for about 20%. The potential evapo-transpiration exceeds 1,300 mm [16].

Figure 2: The averages of monthly precipitation from 1924 to 2001 in El Kef area.

Figure 3: The averages of monthly temperature from 1924 to 2001 in El Kef area.

Owing to financial constraints, a limited number of modern meteoric water samples were selected for oxygen-18 and deuterium [16]. Hydrogen and oxygen isotope analyses were performed in the Laboratory of the International Agency of Atomic Energy (IAEA) in Vienna, by employing, respectively the standard CO₂ equilibration [17,18] and the zinc reduction techniques [19], followed by analysis on a mass spectrometer. Oxygen and hydrogen isotopes analyses were reported to δ notation relative to Vienna-Standard Mean Oceanic Water (V_SMOW), where δ=[(R_S/R_SMOW)-1] x 1,000; R_S represents either the ¹⁸O/¹⁶O or the ²H/¹H ratio of the sample, and RSMOW is ¹⁸O/¹⁶O or the ²H/¹H ratio of the SMOW. Typical precisions are ± 0.1 and ± 1.0% for oxygen-18 and deuterium, respectively. The isotopic composition of oxygen (δ¹⁸O) was measured after equilibration with reference CO₂ at 25°C for 24 h [17], while reduction on Cr at 800°C was used to determine the isotopic composition of hydrogen (δD) in water [20]. Both measurements were performed on a Varian MAT 250 mass spectrometer.

Precipitation was captured using a dual bucket system in which a smaller bucket with a small hole at the bottom was placed snugly inside a larger bucket. During a rain event, the water drained through the hole, where it would remain inside the larger bucket until sample retrieval. After a snowfall, the sampler was brought inside where the water was melted and drained into the lower bucket. The precipitation samples were then transferred to 5 ml glass vials with poly-seal cone closures. These samples were collected in 1-L Nalgene bottles and then transferred after melting into 5 ml glass vials.

Regional isotopic composition of precipitation

The results for meteoric waters are given in Table 1. The isotopic compositions of collected meteoric waters range from -8.45 to -4.12% (δ¹⁸O) and -55.40 to -22.60‰ (δ²H) for El Kef area in 2001 [16]. These results are plotted in Figure 6 and in the Table 1. This variation is controlled by local climatic parameters, including the origin of the vapor mass, re-evaporation during rainfall, the seasonality and monthly of precipitation; and suggestive of different atmospheric sources during these months as compared to the other rainy season months of July through September [12,16,21-23]. Generally, the negative precipitation isotopic values are attributed to the air temperature gradient and the massifs that surrounding the study area (Algerian and Tunisian Atlas: “1,000 ≤ altitude ≤ 1,500m”) (Figure 4). One of the factors that can be identified as influential is the distance of the stations from the Ocean and/or the sea.

Figure 4: The δ¹⁸O/ δ ²H diagram of precipitation in El Kef area.

Figure 5: Plot δ¹⁸O vs Precipitation vs Months in El Kef area.

Figure 6: Plot δ¹⁸O vs Temperature vs Months of precipitation in El Kef area.

These negative residuals of the observed δ¹⁸O values in precipitation most likely result from strong convective precipitation during the period of Atlantic Ocean monsoon and the Mediterranean Sea activities. Furthermore, positive residuals are observed in the desert area of North Africa by several scientific studies [24-27]. These positive residuals can probably be attributed to the moisture forming precipitation coming from inland vapor cycle. The dry continental air masses usually result in ¹⁸O enrichment in precipitation [28,29]. Simultaneously, strong evaporation in these regions also causes raindrops to become increasingly enriched in ¹⁸O of precipitation. Additionally, it is noticeable that observed stations of these regions are relatively sparse (Tunis-Carthage and Sfax stations belonging to the GNIP network), which imparts influence on the precision of interpolation results to some extent.

By concerning the amount of precipitation and temperature effects, the values of δ¹⁸O and δD in precipitation (Figure 5) decrease with increasing precipitation. At a monthly timescale, the precipitation oxygen isotope composition co-varies with the amount of precipitation “amount effect” (showing negative relationship) and temperature (showing positive relationship) “temperature effect” at the study area (Figure 6). Although a longer time series in Northwestern Tunisia (El Kef basin) would be needed in order to confirm exactly the link of amount of rainfall and temperature with the oxygen isotope composition of precipitation, the available data (only 6 months) and mainly the geological and the climatological regional context support that these variables are direct controls of the isotope signal of precipitation. Over this period, each precipitation event higher than 2 mm (excluding potential precipitation altered by partial evaporation of droplets during fall) was collected and analyzed for deuterium and oxygen-18. The variability of oxygen isotopes in precipitation explained by amount of precipitation (discarding months with upper 40 mm of precipitation) (Table 1).

In other words, the evaporation during passage of raindrops through a warm dry atmosphere typical of semi-arid areas causes a greater enrichment in heavy isotopes in precipitation during small rainfall events than in large rainfall events [6]. This ‘‘amount effect’’ has also been observed in the northern Sahara, where heavy winter rainfalls are generally depleted in heavy isotopes [24,30]. It is also possible that air masses coming from central Europe can contribute to the depletion of isotopic composition (continental effect). The distance between the moisture source and the site of precipitation has been suggested to impact the oxygen isotope composition of precipitation [31,32]. According to the Rayleigh distillation process governing this relationship, the depletion of moisture vapour due to precipitation along the air mass trajectory results in a progressive ¹⁸O depletion of the remaining moisture fraction in the atmosphere and the subsequent precipitation events [6,33-37] was the first to recognize that the δD and δ¹⁸O composition of precipitation was negatively correlated with temperature, latitude, altitude, distance from the coast, and the amount of precipitation. The change in moisture sources is an important contributor to the precipitation isotope variability [13]. Thus, the paleoclimate oxygen isotope records depending on the isotope composition of precipitation are potentially impacted by moisture sources (continental moisture originates from evaporation of soils, surface waters and plants transpiration), the origin of the precipitation “precipitation from Mediterranean in relation to Atlantic sources, the atmospheric dynamic” [22,23,33,38-41].

Isotopic composition of precipitation in semi-arid northwestern Tunisia (El Kef area) is highly variable and barely seasonally controlled. This variability is at least in part explained by different pathways of the rain-bringing air masses. We were able to show that air masses coming from southern and western directions in general have significantly lower “d” values than rains from northern and eastern directions due to rainout effects of air-masses. Evaporation effects dominate the isotopic composition of many rainfall events with less than 2 mm, as indicated by their low “d” values.

The relationship between δ²H and δ¹⁸O for local precipitation in El Kef region shows that precipitation originating from cold and wet to hot and dry and the differing contribution of vapor derived from the closed marine basin of the Mediterranean Sea and the Atlantic Ocean vapor source that contains a maritime aerosol. However, in Tunisia geographical diversity influences the complicated mixing of continental and maritime air masses and the sources of air masses even change during a rain event. A mixed isotopic signature of Atlantic and Mediterranean origin was also observed in the aquifers of south Tunisia, Algeria, Morocco and Spain [21,22,47,59-61]. In reality, this seasonal variability of δ¹⁸O and δD values apparent in precipitation in the El Kef area is likely to be reflected in the isotopic composition of the surface waters and the ground waters [23].

The author is thankful to the Editor and the anonymous reviewers for their valuable suggestions and would like to thank the International Agency of Atomic Energy (IAEA) for financial support to carry out this work, to the Radio-Anlaysis Laboratory of the National School of Engineers of Sfax and to the Water, Energy and Environmental Laboratory (L3E)-ENIS-Sfax-Tunisia. Special thanks are directed to the staff members of El Kef Water Resources Division/Agriculture Ministry. The author should like also to thank sincerely the numerous people who helped me in the preparation of this paper.